libvirt/daemon/remote.c

7074 lines
233 KiB
C
Raw Normal View History

/*
* remote.c: handlers for RPC method calls
*
* Copyright (C) 2007-2015 Red Hat, Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see
* <http://www.gnu.org/licenses/>.
*
* Author: Richard W.M. Jones <rjones@redhat.com>
*/
#include <config.h>
#include "virerror.h"
#include "remote.h"
#include "libvirtd.h"
#include "libvirt_internal.h"
#include "datatypes.h"
2012-12-12 18:06:53 +00:00
#include "viralloc.h"
2012-12-12 17:59:27 +00:00
#include "virlog.h"
#include "stream.h"
2012-12-13 18:01:25 +00:00
#include "viruuid.h"
#include "vircommand.h"
#include "intprops.h"
#include "virnetserverservice.h"
#include "virnetserver.h"
#include "virfile.h"
#include "virtypedparam.h"
#include "virdbus.h"
#include "virprocess.h"
#include "remote_protocol.h"
#include "qemu_protocol.h"
#include "lxc_protocol.h"
#include "virstring.h"
#include "object_event.h"
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
#include "domain_conf.h"
#include "network_conf.h"
#include "virprobe.h"
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
#include "viraccessapicheck.h"
#include "viraccessapicheckqemu.h"
#include "virpolkit.h"
#include "virthreadjob.h"
#define VIR_FROM_THIS VIR_FROM_RPC
VIR_LOG_INIT("daemon.remote");
#if SIZEOF_LONG < 8
# define HYPER_TO_TYPE(_type, _to, _from) \
do { \
if ((_from) != (_type)(_from)) { \
virReportError(VIR_ERR_OVERFLOW, \
_("conversion from hyper to %s overflowed"), \
#_type); \
goto cleanup; \
} \
(_to) = (_from); \
} while (0)
# define HYPER_TO_LONG(_to, _from) HYPER_TO_TYPE(long, _to, _from)
# define HYPER_TO_ULONG(_to, _from) HYPER_TO_TYPE(unsigned long, _to, _from)
#else
# define HYPER_TO_LONG(_to, _from) (_to) = (_from)
# define HYPER_TO_ULONG(_to, _from) (_to) = (_from)
#endif
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
struct daemonClientEventCallback {
virNetServerClientPtr client;
int eventID;
int callbackID;
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
bool legacy;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
};
static virDomainPtr get_nonnull_domain(virConnectPtr conn, remote_nonnull_domain domain);
static virNetworkPtr get_nonnull_network(virConnectPtr conn, remote_nonnull_network network);
static virInterfacePtr get_nonnull_interface(virConnectPtr conn, remote_nonnull_interface iface);
static virStoragePoolPtr get_nonnull_storage_pool(virConnectPtr conn, remote_nonnull_storage_pool pool);
static virStorageVolPtr get_nonnull_storage_vol(virConnectPtr conn, remote_nonnull_storage_vol vol);
static virSecretPtr get_nonnull_secret(virConnectPtr conn, remote_nonnull_secret secret);
static virNWFilterPtr get_nonnull_nwfilter(virConnectPtr conn, remote_nonnull_nwfilter nwfilter);
static virDomainSnapshotPtr get_nonnull_domain_snapshot(virDomainPtr dom, remote_nonnull_domain_snapshot snapshot);
static virNodeDevicePtr get_nonnull_node_device(virConnectPtr conn, remote_nonnull_node_device dev);
static void make_nonnull_domain(remote_nonnull_domain *dom_dst, virDomainPtr dom_src);
static void make_nonnull_network(remote_nonnull_network *net_dst, virNetworkPtr net_src);
static void make_nonnull_interface(remote_nonnull_interface *interface_dst, virInterfacePtr interface_src);
static void make_nonnull_storage_pool(remote_nonnull_storage_pool *pool_dst, virStoragePoolPtr pool_src);
static void make_nonnull_storage_vol(remote_nonnull_storage_vol *vol_dst, virStorageVolPtr vol_src);
static void make_nonnull_node_device(remote_nonnull_node_device *dev_dst, virNodeDevicePtr dev_src);
static void make_nonnull_secret(remote_nonnull_secret *secret_dst, virSecretPtr secret_src);
static void make_nonnull_nwfilter(remote_nonnull_nwfilter *net_dst, virNWFilterPtr nwfilter_src);
static void make_nonnull_domain_snapshot(remote_nonnull_domain_snapshot *snapshot_dst, virDomainSnapshotPtr snapshot_src);
static int
remoteSerializeDomainDiskErrors(virDomainDiskErrorPtr errors,
int nerrors,
remote_domain_disk_error **ret_errors_val,
u_int *ret_errors_len);
#include "remote_dispatch.h"
#include "qemu_dispatch.h"
#include "lxc_dispatch.h"
/* Prototypes */
static void
remoteDispatchObjectEventSend(virNetServerClientPtr client,
virNetServerProgramPtr program,
int procnr,
xdrproc_t proc,
void *data);
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
static void
remoteEventCallbackFree(void *opaque)
{
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
daemonClientEventCallbackPtr callback = opaque;
if (!callback)
return;
virObjectUnref(callback->client);
VIR_FREE(callback);
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static bool
remoteRelayDomainEventCheckACL(virNetServerClientPtr client,
virConnectPtr conn, virDomainPtr dom)
{
virDomainDef def;
virIdentityPtr identity = NULL;
bool ret = false;
/* For now, we just create a virDomainDef with enough contents to
* satisfy what viraccessdriverpolkit.c references. This is a bit
* fragile, but I don't know of anything better. */
memset(&def, 0, sizeof(def));
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
def.name = dom->name;
memcpy(def.uuid, dom->uuid, VIR_UUID_BUFLEN);
if (!(identity = virNetServerClientGetIdentity(client)))
goto cleanup;
if (virIdentitySetCurrent(identity) < 0)
goto cleanup;
ret = virConnectDomainEventRegisterAnyCheckACL(conn, &def);
cleanup:
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
ignore_value(virIdentitySetCurrent(NULL));
virObjectUnref(identity);
return ret;
}
static bool
remoteRelayNetworkEventCheckACL(virNetServerClientPtr client,
virConnectPtr conn, virNetworkPtr net)
{
virNetworkDef def;
virIdentityPtr identity = NULL;
bool ret = false;
/* For now, we just create a virNetworkDef with enough contents to
* satisfy what viraccessdriverpolkit.c references. This is a bit
* fragile, but I don't know of anything better. */
def.name = net->name;
memcpy(def.uuid, net->uuid, VIR_UUID_BUFLEN);
if (!(identity = virNetServerClientGetIdentity(client)))
goto cleanup;
if (virIdentitySetCurrent(identity) < 0)
goto cleanup;
ret = virConnectNetworkEventRegisterAnyCheckACL(conn, &def);
cleanup:
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
ignore_value(virIdentitySetCurrent(NULL));
virObjectUnref(identity);
return ret;
}
static bool
remoteRelayStoragePoolEventCheckACL(virNetServerClientPtr client,
virConnectPtr conn,
virStoragePoolPtr pool)
{
virStoragePoolDef def;
virIdentityPtr identity = NULL;
bool ret = false;
/* For now, we just create a virStoragePoolDef with enough contents to
* satisfy what viraccessdriverpolkit.c references. This is a bit
* fragile, but I don't know of anything better. */
def.name = pool->name;
memcpy(def.uuid, pool->uuid, VIR_UUID_BUFLEN);
if (!(identity = virNetServerClientGetIdentity(client)))
goto cleanup;
if (virIdentitySetCurrent(identity) < 0)
goto cleanup;
ret = virConnectStoragePoolEventRegisterAnyCheckACL(conn, &def);
cleanup:
ignore_value(virIdentitySetCurrent(NULL));
virObjectUnref(identity);
return ret;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static bool
remoteRelayNodeDeviceEventCheckACL(virNetServerClientPtr client,
virConnectPtr conn,
virNodeDevicePtr dev)
{
virNodeDeviceDef def;
virIdentityPtr identity = NULL;
bool ret = false;
/* For now, we just create a virNodeDeviceDef with enough contents to
* satisfy what viraccessdriverpolkit.c references. This is a bit
* fragile, but I don't know of anything better. */
def.name = dev->name;
if (!(identity = virNetServerClientGetIdentity(client)))
goto cleanup;
if (virIdentitySetCurrent(identity) < 0)
goto cleanup;
ret = virConnectNodeDeviceEventRegisterAnyCheckACL(conn, &def);
cleanup:
ignore_value(virIdentitySetCurrent(NULL));
virObjectUnref(identity);
return ret;
}
static bool
remoteRelaySecretEventCheckACL(virNetServerClientPtr client,
virConnectPtr conn,
virSecretPtr secret)
{
virSecretDef def;
virIdentityPtr identity = NULL;
bool ret = false;
/* For now, we just create a virSecretDef with enough contents to
* satisfy what viraccessdriverpolkit.c references. This is a bit
* fragile, but I don't know of anything better. */
memcpy(def.uuid, secret->uuid, VIR_UUID_BUFLEN);
def.usage_type = secret->usageType;
def.usage_id = secret->usageID;
if (!(identity = virNetServerClientGetIdentity(client)))
goto cleanup;
if (virIdentitySetCurrent(identity) < 0)
goto cleanup;
ret = virConnectSecretEventRegisterAnyCheckACL(conn, &def);
cleanup:
ignore_value(virIdentitySetCurrent(NULL));
virObjectUnref(identity);
return ret;
}
static bool
remoteRelayDomainQemuMonitorEventCheckACL(virNetServerClientPtr client,
virConnectPtr conn, virDomainPtr dom)
{
virDomainDef def;
virIdentityPtr identity = NULL;
bool ret = false;
/* For now, we just create a virDomainDef with enough contents to
* satisfy what viraccessdriverpolkit.c references. This is a bit
* fragile, but I don't know of anything better. */
def.name = dom->name;
memcpy(def.uuid, dom->uuid, VIR_UUID_BUFLEN);
if (!(identity = virNetServerClientGetIdentity(client)))
goto cleanup;
if (virIdentitySetCurrent(identity) < 0)
goto cleanup;
ret = virConnectDomainQemuMonitorEventRegisterCheckACL(conn, &def);
cleanup:
ignore_value(virIdentitySetCurrent(NULL));
virObjectUnref(identity);
return ret;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventLifecycle(virConnectPtr conn,
virDomainPtr dom,
int event,
int detail,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_lifecycle_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
return -1;
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
VIR_DEBUG("Relaying domain lifecycle event %d %d, callback %d legacy %d",
event, detail, callback->callbackID, callback->legacy);
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
data.event = event;
data.detail = detail;
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_LIFECYCLE,
(xdrproc_t)xdr_remote_domain_event_lifecycle_msg,
&data);
} else {
remote_domain_event_callback_lifecycle_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE,
(xdrproc_t)xdr_remote_domain_event_callback_lifecycle_msg,
&msg);
}
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventReboot(virConnectPtr conn,
virDomainPtr dom,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_reboot_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain reboot event %s %d, callback %d legacy %d",
dom->name, dom->id, callback->callbackID, callback->legacy);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_REBOOT,
(xdrproc_t)xdr_remote_domain_event_reboot_msg, &data);
} else {
remote_domain_event_callback_reboot_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_REBOOT,
(xdrproc_t)xdr_remote_domain_event_callback_reboot_msg, &msg);
}
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventRTCChange(virConnectPtr conn,
virDomainPtr dom,
long long offset,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_rtc_change_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain rtc change event %s %d %lld, callback %d legacy %d",
dom->name, dom->id, offset,
callback->callbackID, callback->legacy);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
data.offset = offset;
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_RTC_CHANGE,
(xdrproc_t)xdr_remote_domain_event_rtc_change_msg, &data);
} else {
remote_domain_event_callback_rtc_change_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_RTC_CHANGE,
(xdrproc_t)xdr_remote_domain_event_callback_rtc_change_msg, &msg);
}
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventWatchdog(virConnectPtr conn,
virDomainPtr dom,
int action,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_watchdog_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain watchdog event %s %d %d, callback %d",
dom->name, dom->id, action, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
data.action = action;
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_WATCHDOG,
(xdrproc_t)xdr_remote_domain_event_watchdog_msg, &data);
} else {
remote_domain_event_callback_watchdog_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_WATCHDOG,
(xdrproc_t)xdr_remote_domain_event_callback_watchdog_msg, &msg);
}
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventIOError(virConnectPtr conn,
virDomainPtr dom,
const char *srcPath,
const char *devAlias,
int action,
void *opaque)
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
{
daemonClientEventCallbackPtr callback = opaque;
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
remote_domain_event_io_error_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
return -1;
VIR_DEBUG("Relaying domain io error %s %d %s %s %d, callback %d",
dom->name, dom->id, srcPath, devAlias, action,
callback->callbackID);
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
/* build return data */
memset(&data, 0, sizeof(data));
if (VIR_STRDUP(data.srcPath, srcPath) < 0 ||
VIR_STRDUP(data.devAlias, devAlias) < 0)
goto error;
make_nonnull_domain(&data.dom, dom);
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
data.action = action;
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_IO_ERROR,
(xdrproc_t)xdr_remote_domain_event_io_error_msg, &data);
} else {
remote_domain_event_callback_io_error_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_IO_ERROR,
(xdrproc_t)xdr_remote_domain_event_callback_io_error_msg, &msg);
}
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
return 0;
error:
VIR_FREE(data.srcPath);
VIR_FREE(data.devAlias);
return -1;
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventIOErrorReason(virConnectPtr conn,
virDomainPtr dom,
const char *srcPath,
const char *devAlias,
int action,
const char *reason,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_io_error_reason_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain io error %s %d %s %s %d %s, callback %d",
dom->name, dom->id, srcPath, devAlias, action, reason,
callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
if (VIR_STRDUP(data.srcPath, srcPath) < 0 ||
VIR_STRDUP(data.devAlias, devAlias) < 0 ||
VIR_STRDUP(data.reason, reason) < 0)
goto error;
data.action = action;
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_IO_ERROR_REASON,
(xdrproc_t)xdr_remote_domain_event_io_error_reason_msg, &data);
} else {
remote_domain_event_callback_io_error_reason_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_IO_ERROR_REASON,
(xdrproc_t)xdr_remote_domain_event_callback_io_error_reason_msg, &msg);
}
return 0;
error:
VIR_FREE(data.srcPath);
VIR_FREE(data.devAlias);
VIR_FREE(data.reason);
return -1;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventGraphics(virConnectPtr conn,
virDomainPtr dom,
int phase,
virDomainEventGraphicsAddressPtr local,
virDomainEventGraphicsAddressPtr remote,
const char *authScheme,
virDomainEventGraphicsSubjectPtr subject,
void *opaque)
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
{
daemonClientEventCallbackPtr callback = opaque;
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
remote_domain_event_graphics_msg data;
size_t i;
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
return -1;
VIR_DEBUG("Relaying domain graphics event %s %d %d - %d %s %s - %d %s %s - %s, callback %d",
dom->name, dom->id, phase,
local->family, local->service, local->node,
remote->family, remote->service, remote->node,
authScheme, callback->callbackID);
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
VIR_DEBUG("Subject %d", subject->nidentity);
for (i = 0; i < subject->nidentity; i++)
VIR_DEBUG(" %s=%s", subject->identities[i].type, subject->identities[i].name);
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
/* build return data */
memset(&data, 0, sizeof(data));
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
data.phase = phase;
data.local.family = local->family;
data.remote.family = remote->family;
if (VIR_STRDUP(data.authScheme, authScheme) < 0 ||
VIR_STRDUP(data.local.node, local->node) < 0 ||
VIR_STRDUP(data.local.service, local->service) < 0 ||
VIR_STRDUP(data.remote.node, remote->node) < 0 ||
VIR_STRDUP(data.remote.service, remote->service) < 0)
goto error;
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
data.subject.subject_len = subject->nidentity;
if (VIR_ALLOC_N(data.subject.subject_val, data.subject.subject_len) < 0)
goto error;
for (i = 0; i < data.subject.subject_len; i++) {
if (VIR_STRDUP(data.subject.subject_val[i].type, subject->identities[i].type) < 0 ||
VIR_STRDUP(data.subject.subject_val[i].name, subject->identities[i].name) < 0)
goto error;
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
}
make_nonnull_domain(&data.dom, dom);
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_GRAPHICS,
(xdrproc_t)xdr_remote_domain_event_graphics_msg, &data);
} else {
remote_domain_event_callback_graphics_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_GRAPHICS,
(xdrproc_t)xdr_remote_domain_event_callback_graphics_msg, &msg);
}
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
return 0;
error:
VIR_FREE(data.authScheme);
VIR_FREE(data.local.node);
VIR_FREE(data.local.service);
VIR_FREE(data.remote.node);
VIR_FREE(data.remote.service);
if (data.subject.subject_val != NULL) {
for (i = 0; i < data.subject.subject_len; i++) {
VIR_FREE(data.subject.subject_val[i].type);
VIR_FREE(data.subject.subject_val[i].name);
}
VIR_FREE(data.subject.subject_val);
}
return -1;
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventBlockJob(virConnectPtr conn,
virDomainPtr dom,
const char *path,
int type,
int status,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_block_job_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain block job event %s %d %s %i, %i, callback %d",
dom->name, dom->id, path, type, status, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
if (VIR_STRDUP(data.path, path) < 0)
return -1;
data.type = type;
data.status = status;
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_BLOCK_JOB,
(xdrproc_t)xdr_remote_domain_event_block_job_msg, &data);
} else {
remote_domain_event_callback_block_job_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_BLOCK_JOB,
(xdrproc_t)xdr_remote_domain_event_callback_block_job_msg, &msg);
}
return 0;
}
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventControlError(virConnectPtr conn,
virDomainPtr dom,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_control_error_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain control error %s %d, callback %d",
dom->name, dom->id, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CONTROL_ERROR,
(xdrproc_t)xdr_remote_domain_event_control_error_msg, &data);
} else {
remote_domain_event_callback_control_error_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_CONTROL_ERROR,
(xdrproc_t)xdr_remote_domain_event_callback_control_error_msg, &msg);
}
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventDiskChange(virConnectPtr conn,
virDomainPtr dom,
const char *oldSrcPath,
const char *newSrcPath,
const char *devAlias,
int reason,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_disk_change_msg data;
char **oldSrcPath_p = NULL, **newSrcPath_p = NULL;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain %s %d disk change %s %s %s %d, callback %d",
dom->name, dom->id, oldSrcPath, newSrcPath, devAlias, reason,
callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
if (oldSrcPath &&
((VIR_ALLOC(oldSrcPath_p) < 0) ||
VIR_STRDUP(*oldSrcPath_p, oldSrcPath) < 0))
goto error;
if (newSrcPath &&
((VIR_ALLOC(newSrcPath_p) < 0) ||
VIR_STRDUP(*newSrcPath_p, newSrcPath) < 0))
goto error;
data.oldSrcPath = oldSrcPath_p;
data.newSrcPath = newSrcPath_p;
if (VIR_STRDUP(data.devAlias, devAlias) < 0)
goto error;
data.reason = reason;
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_DISK_CHANGE,
(xdrproc_t)xdr_remote_domain_event_disk_change_msg, &data);
} else {
remote_domain_event_callback_disk_change_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_DISK_CHANGE,
(xdrproc_t)xdr_remote_domain_event_callback_disk_change_msg, &msg);
}
return 0;
error:
VIR_FREE(oldSrcPath_p);
VIR_FREE(newSrcPath_p);
return -1;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventTrayChange(virConnectPtr conn,
virDomainPtr dom,
const char *devAlias,
int reason,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_tray_change_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain %s %d tray change devAlias: %s reason: %d, callback %d",
dom->name, dom->id, devAlias, reason, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
if (VIR_STRDUP(data.devAlias, devAlias) < 0)
return -1;
data.reason = reason;
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_TRAY_CHANGE,
(xdrproc_t)xdr_remote_domain_event_tray_change_msg, &data);
} else {
remote_domain_event_callback_tray_change_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_TRAY_CHANGE,
(xdrproc_t)xdr_remote_domain_event_callback_tray_change_msg, &msg);
}
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventPMWakeup(virConnectPtr conn,
virDomainPtr dom,
event: pass reason for PM events Commit 57ddcc23 (v0.9.11) introduced the pmwakeup event, with an optional 'reason' field reserved for possible future expansion. But it failed to wire the field through RPC, so even if we do add a reason in the future, we will be unable to get it back to the user. Worse, commit 7ba5defb (v1.0.0) repeated the same mistake with the pmsuspend_disk event. As long as we are adding new RPC calls, we might as well fix the events to actually match the signature so that we don't have to add yet another RPC in the future if we do decide to start using the reason field. * src/remote/remote_protocol.x (remote_domain_event_callback_pmwakeup_msg) (remote_domain_event_callback_pmsuspend_msg) (remote_domain_event_callback_pmsuspend_disk_msg): Add reason field. * daemon/remote.c (remoteRelayDomainEventPMWakeup) (remoteRelayDomainEventPMSuspend) (remoteRelayDomainEventPMSuspendDisk): Pass reason to client. * src/conf/domain_event.h (virDomainEventPMWakeupNewFromDom) (virDomainEventPMSuspendNewFromDom) (virDomainEventPMSuspendDiskNewFromDom): Require additional parameter. * src/conf/domain_event.c (virDomainEventPMClass): New class. (virDomainEventPMDispose): New function. (virDomainEventPMWakeupNew*, virDomainEventPMSuspendNew*) (virDomainEventPMSuspendDiskNew*) (virDomainEventDispatchDefaultFunc): Use new class. * src/remote/remote_driver.c (remoteDomainBuildEvent*PM*): Pass reason through. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-29 00:41:34 +00:00
int reason,
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_pmwakeup_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain %s %d system pmwakeup, callback %d",
dom->name, dom->id, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_PMWAKEUP,
(xdrproc_t)xdr_remote_domain_event_pmwakeup_msg, &data);
} else {
remote_domain_event_callback_pmwakeup_msg msg = { callback->callbackID,
event: pass reason for PM events Commit 57ddcc23 (v0.9.11) introduced the pmwakeup event, with an optional 'reason' field reserved for possible future expansion. But it failed to wire the field through RPC, so even if we do add a reason in the future, we will be unable to get it back to the user. Worse, commit 7ba5defb (v1.0.0) repeated the same mistake with the pmsuspend_disk event. As long as we are adding new RPC calls, we might as well fix the events to actually match the signature so that we don't have to add yet another RPC in the future if we do decide to start using the reason field. * src/remote/remote_protocol.x (remote_domain_event_callback_pmwakeup_msg) (remote_domain_event_callback_pmsuspend_msg) (remote_domain_event_callback_pmsuspend_disk_msg): Add reason field. * daemon/remote.c (remoteRelayDomainEventPMWakeup) (remoteRelayDomainEventPMSuspend) (remoteRelayDomainEventPMSuspendDisk): Pass reason to client. * src/conf/domain_event.h (virDomainEventPMWakeupNewFromDom) (virDomainEventPMSuspendNewFromDom) (virDomainEventPMSuspendDiskNewFromDom): Require additional parameter. * src/conf/domain_event.c (virDomainEventPMClass): New class. (virDomainEventPMDispose): New function. (virDomainEventPMWakeupNew*, virDomainEventPMSuspendNew*) (virDomainEventPMSuspendDiskNew*) (virDomainEventDispatchDefaultFunc): Use new class. * src/remote/remote_driver.c (remoteDomainBuildEvent*PM*): Pass reason through. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-29 00:41:34 +00:00
reason, data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_PMWAKEUP,
(xdrproc_t)xdr_remote_domain_event_callback_pmwakeup_msg, &msg);
}
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventPMSuspend(virConnectPtr conn,
virDomainPtr dom,
event: pass reason for PM events Commit 57ddcc23 (v0.9.11) introduced the pmwakeup event, with an optional 'reason' field reserved for possible future expansion. But it failed to wire the field through RPC, so even if we do add a reason in the future, we will be unable to get it back to the user. Worse, commit 7ba5defb (v1.0.0) repeated the same mistake with the pmsuspend_disk event. As long as we are adding new RPC calls, we might as well fix the events to actually match the signature so that we don't have to add yet another RPC in the future if we do decide to start using the reason field. * src/remote/remote_protocol.x (remote_domain_event_callback_pmwakeup_msg) (remote_domain_event_callback_pmsuspend_msg) (remote_domain_event_callback_pmsuspend_disk_msg): Add reason field. * daemon/remote.c (remoteRelayDomainEventPMWakeup) (remoteRelayDomainEventPMSuspend) (remoteRelayDomainEventPMSuspendDisk): Pass reason to client. * src/conf/domain_event.h (virDomainEventPMWakeupNewFromDom) (virDomainEventPMSuspendNewFromDom) (virDomainEventPMSuspendDiskNewFromDom): Require additional parameter. * src/conf/domain_event.c (virDomainEventPMClass): New class. (virDomainEventPMDispose): New function. (virDomainEventPMWakeupNew*, virDomainEventPMSuspendNew*) (virDomainEventPMSuspendDiskNew*) (virDomainEventDispatchDefaultFunc): Use new class. * src/remote/remote_driver.c (remoteDomainBuildEvent*PM*): Pass reason through. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-29 00:41:34 +00:00
int reason,
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_pmsuspend_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain %s %d system pmsuspend, callback %d",
dom->name, dom->id, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_PMSUSPEND,
(xdrproc_t)xdr_remote_domain_event_pmsuspend_msg, &data);
} else {
remote_domain_event_callback_pmsuspend_msg msg = { callback->callbackID,
event: pass reason for PM events Commit 57ddcc23 (v0.9.11) introduced the pmwakeup event, with an optional 'reason' field reserved for possible future expansion. But it failed to wire the field through RPC, so even if we do add a reason in the future, we will be unable to get it back to the user. Worse, commit 7ba5defb (v1.0.0) repeated the same mistake with the pmsuspend_disk event. As long as we are adding new RPC calls, we might as well fix the events to actually match the signature so that we don't have to add yet another RPC in the future if we do decide to start using the reason field. * src/remote/remote_protocol.x (remote_domain_event_callback_pmwakeup_msg) (remote_domain_event_callback_pmsuspend_msg) (remote_domain_event_callback_pmsuspend_disk_msg): Add reason field. * daemon/remote.c (remoteRelayDomainEventPMWakeup) (remoteRelayDomainEventPMSuspend) (remoteRelayDomainEventPMSuspendDisk): Pass reason to client. * src/conf/domain_event.h (virDomainEventPMWakeupNewFromDom) (virDomainEventPMSuspendNewFromDom) (virDomainEventPMSuspendDiskNewFromDom): Require additional parameter. * src/conf/domain_event.c (virDomainEventPMClass): New class. (virDomainEventPMDispose): New function. (virDomainEventPMWakeupNew*, virDomainEventPMSuspendNew*) (virDomainEventPMSuspendDiskNew*) (virDomainEventDispatchDefaultFunc): Use new class. * src/remote/remote_driver.c (remoteDomainBuildEvent*PM*): Pass reason through. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-29 00:41:34 +00:00
reason, data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_PMSUSPEND,
(xdrproc_t)xdr_remote_domain_event_callback_pmsuspend_msg, &msg);
}
return 0;
}
static int
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
remoteRelayDomainEventBalloonChange(virConnectPtr conn,
virDomainPtr dom,
unsigned long long actual,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_balloon_change_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain balloon change event %s %d %lld, callback %d",
dom->name, dom->id, actual, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
data.actual = actual;
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_BALLOON_CHANGE,
(xdrproc_t)xdr_remote_domain_event_balloon_change_msg, &data);
} else {
remote_domain_event_callback_balloon_change_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_BALLOON_CHANGE,
(xdrproc_t)xdr_remote_domain_event_callback_balloon_change_msg, &msg);
}
return 0;
}
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
static int
remoteRelayDomainEventPMSuspendDisk(virConnectPtr conn,
virDomainPtr dom,
event: pass reason for PM events Commit 57ddcc23 (v0.9.11) introduced the pmwakeup event, with an optional 'reason' field reserved for possible future expansion. But it failed to wire the field through RPC, so even if we do add a reason in the future, we will be unable to get it back to the user. Worse, commit 7ba5defb (v1.0.0) repeated the same mistake with the pmsuspend_disk event. As long as we are adding new RPC calls, we might as well fix the events to actually match the signature so that we don't have to add yet another RPC in the future if we do decide to start using the reason field. * src/remote/remote_protocol.x (remote_domain_event_callback_pmwakeup_msg) (remote_domain_event_callback_pmsuspend_msg) (remote_domain_event_callback_pmsuspend_disk_msg): Add reason field. * daemon/remote.c (remoteRelayDomainEventPMWakeup) (remoteRelayDomainEventPMSuspend) (remoteRelayDomainEventPMSuspendDisk): Pass reason to client. * src/conf/domain_event.h (virDomainEventPMWakeupNewFromDom) (virDomainEventPMSuspendNewFromDom) (virDomainEventPMSuspendDiskNewFromDom): Require additional parameter. * src/conf/domain_event.c (virDomainEventPMClass): New class. (virDomainEventPMDispose): New function. (virDomainEventPMWakeupNew*, virDomainEventPMSuspendNew*) (virDomainEventPMSuspendDiskNew*) (virDomainEventDispatchDefaultFunc): Use new class. * src/remote/remote_driver.c (remoteDomainBuildEvent*PM*): Pass reason through. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-29 00:41:34 +00:00
int reason,
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_pmsuspend_disk_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain %s %d system pmsuspend-disk, callback %d",
dom->name, dom->id, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_PMSUSPEND_DISK,
(xdrproc_t)xdr_remote_domain_event_pmsuspend_disk_msg, &data);
} else {
remote_domain_event_callback_pmsuspend_disk_msg msg = { callback->callbackID,
event: pass reason for PM events Commit 57ddcc23 (v0.9.11) introduced the pmwakeup event, with an optional 'reason' field reserved for possible future expansion. But it failed to wire the field through RPC, so even if we do add a reason in the future, we will be unable to get it back to the user. Worse, commit 7ba5defb (v1.0.0) repeated the same mistake with the pmsuspend_disk event. As long as we are adding new RPC calls, we might as well fix the events to actually match the signature so that we don't have to add yet another RPC in the future if we do decide to start using the reason field. * src/remote/remote_protocol.x (remote_domain_event_callback_pmwakeup_msg) (remote_domain_event_callback_pmsuspend_msg) (remote_domain_event_callback_pmsuspend_disk_msg): Add reason field. * daemon/remote.c (remoteRelayDomainEventPMWakeup) (remoteRelayDomainEventPMSuspend) (remoteRelayDomainEventPMSuspendDisk): Pass reason to client. * src/conf/domain_event.h (virDomainEventPMWakeupNewFromDom) (virDomainEventPMSuspendNewFromDom) (virDomainEventPMSuspendDiskNewFromDom): Require additional parameter. * src/conf/domain_event.c (virDomainEventPMClass): New class. (virDomainEventPMDispose): New function. (virDomainEventPMWakeupNew*, virDomainEventPMSuspendNew*) (virDomainEventPMSuspendDiskNew*) (virDomainEventDispatchDefaultFunc): Use new class. * src/remote/remote_driver.c (remoteDomainBuildEvent*PM*): Pass reason through. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-29 00:41:34 +00:00
reason, data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_PMSUSPEND_DISK,
(xdrproc_t)xdr_remote_domain_event_callback_pmsuspend_disk_msg, &msg);
}
return 0;
}
static int
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
remoteRelayDomainEventDeviceRemoved(virConnectPtr conn,
virDomainPtr dom,
const char *devAlias,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_device_removed_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain device removed event %s %d %s, callback %d",
dom->name, dom->id, devAlias, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
if (VIR_STRDUP(data.devAlias, devAlias) < 0)
return -1;
make_nonnull_domain(&data.dom, dom);
if (callback->legacy) {
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_DEVICE_REMOVED,
(xdrproc_t)xdr_remote_domain_event_device_removed_msg,
&data);
} else {
remote_domain_event_callback_device_removed_msg msg = { callback->callbackID,
data };
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_DEVICE_REMOVED,
(xdrproc_t)xdr_remote_domain_event_callback_device_removed_msg,
&msg);
}
return 0;
}
blockjob: use stable disk string in job event When the block job event was first added, it was for block pull, where the active layer of the disk remains the same name. It was also in a day where we only cared about local files, and so we always had a canonical absolute file name. But two things have changed since then: we now have network disks, where determining a single absolute string does not really make sense; and we have two-phase jobs (copy and active commit) where the name of the active layer changes between the first event (ready, on the old name) and second (complete, on the pivoted name). Adam Litke reported that having an unstable string between events makes life harder for clients. Furthermore, all of our API that operate on a particular disk of a domain accept multiple strings: not only the absolute name of the active layer, but also the destination device name (such as 'vda'). As this latter name is stable, even for network sources, it serves as a better string to supply in block job events. But backwards-compatibility demands that we should not change the name handed to users unless they explicitly request it. Therefore, this patch adds a new event, BLOCK_JOB_2 (alas, I couldn't think of any nicer name - but at least Migrate2 and Migrate3 are precedent for a number suffix). We must double up on emitting both old-style and new-style events according to what clients have registered for (see also how IOError and IOErrorReason emits double events, but there the difference was a larger struct rather than changed meaning of one of the struct members). Unfortunately, adding a new event isn't something that can easily be broken into pieces, so the commit is rather large. * include/libvirt/libvirt.h.in (virDomainEventID): Add a new id for VIR_DOMAIN_EVENT_ID_BLOCK_JOB_2. (virConnectDomainEventBlockJobCallback): Document new semantics. * src/conf/domain_event.c (_virDomainEventBlockJob): Rename field, to ensure we catch all clients. (virDomainEventBlockJobNew): Add parameter. (virDomainEventBlockJobDispose) (virDomainEventBlockJobNewFromObj) (virDomainEventBlockJobNewFromDom) (virDomainEventDispatchDefaultFunc): Adjust clients. (virDomainEventBlockJob2NewFromObj) (virDomainEventBlockJob2NewFromDom): New functions. * src/conf/domain_event.h: Add new prototypes. * src/libvirt_private.syms (domain_event.h): Export new functions. * src/qemu/qemu_driver.c (qemuDomainBlockJobImpl): Generate two different events. * src/qemu/qemu_process.c (qemuProcessHandleBlockJob): Likewise. * src/remote/remote_protocol.x (remote_domain_event_block_job_2_msg): New struct. (REMOTE_PROC_DOMAIN_EVENT_BLOCK_JOB_2): New RPC. * src/remote/remote_driver.c (remoteDomainBuildEventBlockJob2): New handler. (remoteEvents): Register new event. * daemon/remote.c (remoteRelayDomainEventBlockJob2): New handler. (domainEventCallbacks): Register new event. * tools/virsh-domain.c (vshEventCallbacks): Likewise. (vshEventBlockJobPrint): Adjust client. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-06-14 13:18:04 +00:00
static int
remoteRelayDomainEventBlockJob2(virConnectPtr conn,
virDomainPtr dom,
const char *dst,
int type,
int status,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_block_job_2_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain block job 2 event %s %d %s %i, %i, callback %d",
dom->name, dom->id, dst, type, status, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
data.callbackID = callback->callbackID;
if (VIR_STRDUP(data.dst, dst) < 0)
return -1;
blockjob: use stable disk string in job event When the block job event was first added, it was for block pull, where the active layer of the disk remains the same name. It was also in a day where we only cared about local files, and so we always had a canonical absolute file name. But two things have changed since then: we now have network disks, where determining a single absolute string does not really make sense; and we have two-phase jobs (copy and active commit) where the name of the active layer changes between the first event (ready, on the old name) and second (complete, on the pivoted name). Adam Litke reported that having an unstable string between events makes life harder for clients. Furthermore, all of our API that operate on a particular disk of a domain accept multiple strings: not only the absolute name of the active layer, but also the destination device name (such as 'vda'). As this latter name is stable, even for network sources, it serves as a better string to supply in block job events. But backwards-compatibility demands that we should not change the name handed to users unless they explicitly request it. Therefore, this patch adds a new event, BLOCK_JOB_2 (alas, I couldn't think of any nicer name - but at least Migrate2 and Migrate3 are precedent for a number suffix). We must double up on emitting both old-style and new-style events according to what clients have registered for (see also how IOError and IOErrorReason emits double events, but there the difference was a larger struct rather than changed meaning of one of the struct members). Unfortunately, adding a new event isn't something that can easily be broken into pieces, so the commit is rather large. * include/libvirt/libvirt.h.in (virDomainEventID): Add a new id for VIR_DOMAIN_EVENT_ID_BLOCK_JOB_2. (virConnectDomainEventBlockJobCallback): Document new semantics. * src/conf/domain_event.c (_virDomainEventBlockJob): Rename field, to ensure we catch all clients. (virDomainEventBlockJobNew): Add parameter. (virDomainEventBlockJobDispose) (virDomainEventBlockJobNewFromObj) (virDomainEventBlockJobNewFromDom) (virDomainEventDispatchDefaultFunc): Adjust clients. (virDomainEventBlockJob2NewFromObj) (virDomainEventBlockJob2NewFromDom): New functions. * src/conf/domain_event.h: Add new prototypes. * src/libvirt_private.syms (domain_event.h): Export new functions. * src/qemu/qemu_driver.c (qemuDomainBlockJobImpl): Generate two different events. * src/qemu/qemu_process.c (qemuProcessHandleBlockJob): Likewise. * src/remote/remote_protocol.x (remote_domain_event_block_job_2_msg): New struct. (REMOTE_PROC_DOMAIN_EVENT_BLOCK_JOB_2): New RPC. * src/remote/remote_driver.c (remoteDomainBuildEventBlockJob2): New handler. (remoteEvents): Register new event. * daemon/remote.c (remoteRelayDomainEventBlockJob2): New handler. (domainEventCallbacks): Register new event. * tools/virsh-domain.c (vshEventCallbacks): Likewise. (vshEventBlockJobPrint): Adjust client. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-06-14 13:18:04 +00:00
data.type = type;
data.status = status;
make_nonnull_domain(&data.dom, dom);
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_BLOCK_JOB_2,
(xdrproc_t)xdr_remote_domain_event_block_job_2_msg, &data);
return 0;
}
static int
remoteRelayDomainEventTunable(virConnectPtr conn,
virDomainPtr dom,
virTypedParameterPtr params,
int nparams,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_callback_tunable_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain tunable event %s %d, callback %d, params %p %d",
dom->name, dom->id, callback->callbackID, params, nparams);
/* build return data */
memset(&data, 0, sizeof(data));
data.callbackID = callback->callbackID;
make_nonnull_domain(&data.dom, dom);
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &data.params.params_val,
&data.params.params_len,
VIR_TYPED_PARAM_STRING_OKAY) < 0) {
VIR_FREE(data.dom.name);
return -1;
}
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_TUNABLE,
(xdrproc_t)xdr_remote_domain_event_callback_tunable_msg,
&data);
return 0;
}
static int
remoteRelayDomainEventAgentLifecycle(virConnectPtr conn,
virDomainPtr dom,
int state,
int reason,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_callback_agent_lifecycle_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain agent lifecycle event %s %d, callback %d, "
" state %d, reason %d",
dom->name, dom->id, callback->callbackID, state, reason);
/* build return data */
memset(&data, 0, sizeof(data));
data.callbackID = callback->callbackID;
make_nonnull_domain(&data.dom, dom);
data.state = state;
data.reason = reason;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_AGENT_LIFECYCLE,
(xdrproc_t)xdr_remote_domain_event_callback_agent_lifecycle_msg,
&data);
return 0;
}
static int
remoteRelayDomainEventDeviceAdded(virConnectPtr conn,
virDomainPtr dom,
const char *devAlias,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_callback_device_added_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain device added event %s %d %s, callback %d",
dom->name, dom->id, devAlias, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
if (VIR_STRDUP(data.devAlias, devAlias) < 0)
return -1;
make_nonnull_domain(&data.dom, dom);
data.callbackID = callback->callbackID;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_DEVICE_ADDED,
(xdrproc_t)xdr_remote_domain_event_callback_device_added_msg,
&data);
return 0;
}
static int
remoteRelayDomainEventMigrationIteration(virConnectPtr conn,
virDomainPtr dom,
int iteration,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_callback_migration_iteration_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain migration pass event %s %d, "
"callback %d, iteration %d",
dom->name, dom->id, callback->callbackID, iteration);
/* build return data */
memset(&data, 0, sizeof(data));
data.callbackID = callback->callbackID;
make_nonnull_domain(&data.dom, dom);
data.iteration = iteration;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_MIGRATION_ITERATION,
(xdrproc_t)xdr_remote_domain_event_callback_migration_iteration_msg,
&data);
return 0;
}
static int
remoteRelayDomainEventJobCompleted(virConnectPtr conn,
virDomainPtr dom,
virTypedParameterPtr params,
int nparams,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_callback_job_completed_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain migration completed event %s %d, "
"callback %d, params %p %d",
dom->name, dom->id, callback->callbackID, params, nparams);
/* build return data */
memset(&data, 0, sizeof(data));
data.callbackID = callback->callbackID;
make_nonnull_domain(&data.dom, dom);
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &data.params.params_val,
&data.params.params_len,
VIR_TYPED_PARAM_STRING_OKAY) < 0) {
VIR_FREE(data.dom.name);
return -1;
}
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_JOB_COMPLETED,
(xdrproc_t)xdr_remote_domain_event_callback_job_completed_msg,
&data);
return 0;
}
static int
remoteRelayDomainEventDeviceRemovalFailed(virConnectPtr conn,
virDomainPtr dom,
const char *devAlias,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_callback_device_removal_failed_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain device removal failed event %s %d %s, callback %d",
dom->name, dom->id, devAlias, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
if (VIR_STRDUP(data.devAlias, devAlias) < 0)
return -1;
make_nonnull_domain(&data.dom, dom);
data.callbackID = callback->callbackID;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_DEVICE_REMOVAL_FAILED,
(xdrproc_t)xdr_remote_domain_event_callback_device_removal_failed_msg,
&data);
return 0;
}
static int
remoteRelayDomainEventMetadataChange(virConnectPtr conn,
virDomainPtr dom,
int type,
const char *nsuri,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_callback_metadata_change_msg data;
char **nsurip;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain metadata change %s %d %d %s, callback %d",
dom->name, dom->id, type, NULLSTR(nsuri), callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
data.type = type;
if (nsuri) {
if (VIR_ALLOC(nsurip) < 0)
return -1;
if (VIR_STRDUP(*nsurip, nsuri) < 0) {
VIR_FREE(nsurip);
return -1;
}
data.nsuri = nsurip;
}
make_nonnull_domain(&data.dom, dom);
data.callbackID = callback->callbackID;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_CALLBACK_METADATA_CHANGE,
(xdrproc_t)xdr_remote_domain_event_callback_metadata_change_msg,
&data);
return 0;
}
static int
remoteRelayDomainEventBlockThreshold(virConnectPtr conn,
virDomainPtr dom,
const char *dev,
const char *path,
unsigned long long threshold,
unsigned long long excess,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_domain_event_block_threshold_msg data;
if (callback->callbackID < 0 ||
!remoteRelayDomainEventCheckACL(callback->client, conn, dom))
return -1;
VIR_DEBUG("Relaying domain block threshold event %s %d %s %s %llu %llu, callback %d",
dom->name, dom->id, dev, NULLSTR(path), threshold, excess, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
data.callbackID = callback->callbackID;
if (VIR_STRDUP(data.dev, dev) < 0)
goto error;
if (path) {
if (VIR_ALLOC(data.path) < 0)
goto error;
if (VIR_STRDUP(*(data.path), path) < 0)
goto error;
}
data.threshold = threshold;
data.excess = excess;
make_nonnull_domain(&data.dom, dom);
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_DOMAIN_EVENT_BLOCK_THRESHOLD,
(xdrproc_t)xdr_remote_domain_event_block_threshold_msg, &data);
return 0;
error:
VIR_FREE(data.dev);
return -1;
}
static virConnectDomainEventGenericCallback domainEventCallbacks[] = {
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventLifecycle),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventReboot),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventRTCChange),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventWatchdog),
Add support for an explicit IO error event This introduces a new event type VIR_DOMAIN_EVENT_ID_IO_ERROR This event includes the action that is about to be taken as a result of the watchdog triggering typedef enum { VIR_DOMAIN_EVENT_IO_ERROR_NONE = 0, VIR_DOMAIN_EVENT_IO_ERROR_PAUSE, VIR_DOMAIN_EVENT_IO_ERROR_REPORT, } virDomainEventIOErrorAction; In addition it has the source path of the disk that had the error and its unique device alias. It does not include the target device name (/dev/sda), since this would preclude triggering IO errors from other file backed devices (eg serial ports connected to a file) Thus there is a new callback definition for this event type typedef void (*virConnectDomainEventIOErrorCallback)(virConnectPtr conn, virDomainPtr dom, const char *srcPath, const char *devAlias, int action, void *opaque); This is currently wired up to the QEMU block IO error events * daemon/remote.c: Dispatch IO error events to client * examples/domain-events/events-c/event-test.c: Watch for IO error events * include/libvirt/libvirt.h.in: Define new IO error event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle IO error events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for block IO errors and emit a libvirt IO error event * src/remote/remote_driver.c: Receive and dispatch IO error events to application * src/remote/remote_protocol.x: Wire protocol definition for IO error events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for BLOCK_IO_ERROR event from QEMU monitor
2010-03-18 19:37:44 +00:00
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventIOError),
Add domain events for graphics network clients This introduces a new event type VIR_DOMAIN_EVENT_ID_GRAPHICS The same event can be emitted in 3 scenarios typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_CONNECT = 0, VIR_DOMAIN_EVENT_GRAPHICS_INITIALIZE, VIR_DOMAIN_EVENT_GRAPHICS_DISCONNECT, } virDomainEventGraphicsPhase; Connect/disconnect are triggered at socket accept/close. The initialize phase is immediately after the protocol setup and authentication has completed. ie when the client is authorized and about to start interacting with the graphical desktop This event comes with *a lot* of potential information - IP address, port & address family of client - IP address, port & address family of server - Authentication scheme (arbitrary string) - Authenticated subject identity. A subject may have multiple identities with some authentication schemes. For example, vencrypt+sasl results in a x509dname and saslUsername identities. This results in a very complicated callback :-( typedef enum { VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV4, VIR_DOMAIN_EVENT_GRAPHICS_ADDRESS_IPV6, } virDomainEventGraphicsAddressType; struct _virDomainEventGraphicsAddress { int family; const char *node; const char *service; }; typedef struct _virDomainEventGraphicsAddress virDomainEventGraphicsAddress; typedef virDomainEventGraphicsAddress *virDomainEventGraphicsAddressPtr; struct _virDomainEventGraphicsSubject { int nidentity; struct { const char *type; const char *name; } *identities; }; typedef struct _virDomainEventGraphicsSubject virDomainEventGraphicsSubject; typedef virDomainEventGraphicsSubject *virDomainEventGraphicsSubjectPtr; typedef void (*virConnectDomainEventGraphicsCallback)(virConnectPtr conn, virDomainPtr dom, int phase, virDomainEventGraphicsAddressPtr local, virDomainEventGraphicsAddressPtr remote, const char *authScheme, virDomainEventGraphicsSubjectPtr subject, void *opaque); The wire protocol is similarly complex struct remote_domain_event_graphics_address { int family; remote_nonnull_string node; remote_nonnull_string service; }; const REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX = 20; struct remote_domain_event_graphics_identity { remote_nonnull_string type; remote_nonnull_string name; }; struct remote_domain_event_graphics_msg { remote_nonnull_domain dom; int phase; remote_domain_event_graphics_address local; remote_domain_event_graphics_address remote; remote_nonnull_string authScheme; remote_domain_event_graphics_identity subject<REMOTE_DOMAIN_EVENT_GRAPHICS_IDENTITY_MAX>; }; This is currently implemented in QEMU for the VNC graphics protocol, but designed to be usable with SPICE graphics in the future too. * daemon/remote.c: Dispatch graphics events to client * examples/domain-events/events-c/event-test.c: Watch for graphics events * include/libvirt/libvirt.h.in: Define new graphics event ID and callback signature * src/conf/domain_event.c, src/conf/domain_event.h, src/libvirt_private.syms: Extend API to handle graphics events * src/qemu/qemu_driver.c: Connect to the QEMU monitor event for VNC events and emit a libvirt graphics event * src/remote/remote_driver.c: Receive and dispatch graphics events to application * src/remote/remote_protocol.x: Wire protocol definition for graphics events * src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h, src/qemu/qemu_monitor_json.c: Watch for VNC_CONNECTED, VNC_INITIALIZED & VNC_DISCONNETED events from QEMU monitor
2010-03-19 13:27:45 +00:00
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventGraphics),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventIOErrorReason),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventControlError),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventBlockJob),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventDiskChange),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventTrayChange),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventPMWakeup),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventPMSuspend),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventBalloonChange),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventPMSuspendDisk),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventDeviceRemoved),
blockjob: use stable disk string in job event When the block job event was first added, it was for block pull, where the active layer of the disk remains the same name. It was also in a day where we only cared about local files, and so we always had a canonical absolute file name. But two things have changed since then: we now have network disks, where determining a single absolute string does not really make sense; and we have two-phase jobs (copy and active commit) where the name of the active layer changes between the first event (ready, on the old name) and second (complete, on the pivoted name). Adam Litke reported that having an unstable string between events makes life harder for clients. Furthermore, all of our API that operate on a particular disk of a domain accept multiple strings: not only the absolute name of the active layer, but also the destination device name (such as 'vda'). As this latter name is stable, even for network sources, it serves as a better string to supply in block job events. But backwards-compatibility demands that we should not change the name handed to users unless they explicitly request it. Therefore, this patch adds a new event, BLOCK_JOB_2 (alas, I couldn't think of any nicer name - but at least Migrate2 and Migrate3 are precedent for a number suffix). We must double up on emitting both old-style and new-style events according to what clients have registered for (see also how IOError and IOErrorReason emits double events, but there the difference was a larger struct rather than changed meaning of one of the struct members). Unfortunately, adding a new event isn't something that can easily be broken into pieces, so the commit is rather large. * include/libvirt/libvirt.h.in (virDomainEventID): Add a new id for VIR_DOMAIN_EVENT_ID_BLOCK_JOB_2. (virConnectDomainEventBlockJobCallback): Document new semantics. * src/conf/domain_event.c (_virDomainEventBlockJob): Rename field, to ensure we catch all clients. (virDomainEventBlockJobNew): Add parameter. (virDomainEventBlockJobDispose) (virDomainEventBlockJobNewFromObj) (virDomainEventBlockJobNewFromDom) (virDomainEventDispatchDefaultFunc): Adjust clients. (virDomainEventBlockJob2NewFromObj) (virDomainEventBlockJob2NewFromDom): New functions. * src/conf/domain_event.h: Add new prototypes. * src/libvirt_private.syms (domain_event.h): Export new functions. * src/qemu/qemu_driver.c (qemuDomainBlockJobImpl): Generate two different events. * src/qemu/qemu_process.c (qemuProcessHandleBlockJob): Likewise. * src/remote/remote_protocol.x (remote_domain_event_block_job_2_msg): New struct. (REMOTE_PROC_DOMAIN_EVENT_BLOCK_JOB_2): New RPC. * src/remote/remote_driver.c (remoteDomainBuildEventBlockJob2): New handler. (remoteEvents): Register new event. * daemon/remote.c (remoteRelayDomainEventBlockJob2): New handler. (domainEventCallbacks): Register new event. * tools/virsh-domain.c (vshEventCallbacks): Likewise. (vshEventBlockJobPrint): Adjust client. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-06-14 13:18:04 +00:00
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventBlockJob2),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventTunable),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventAgentLifecycle),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventDeviceAdded),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventMigrationIteration),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventJobCompleted),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventDeviceRemovalFailed),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventMetadataChange),
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventBlockThreshold),
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
};
verify(ARRAY_CARDINALITY(domainEventCallbacks) == VIR_DOMAIN_EVENT_ID_LAST);
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
static int
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
remoteRelayNetworkEventLifecycle(virConnectPtr conn,
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
virNetworkPtr net,
int event,
int detail,
void *opaque)
{
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
daemonClientEventCallbackPtr callback = opaque;
remote_network_event_lifecycle_msg data;
event: move event filtering to daemon (regression fix) https://bugzilla.redhat.com/show_bug.cgi?id=1058839 Commit f9f56340 for CVE-2014-0028 almost had the right idea - we need to check the ACL rules to filter which events to send. But it overlooked one thing: the event dispatch queue is running in the main loop thread, and therefore does not normally have a current virIdentityPtr. But filter checks can be based on current identity, so when libvirtd.conf contains access_drivers=["polkit"], we ended up rejecting access for EVERY event due to failure to look up the current identity, even if it should have been allowed. Furthermore, even for events that are triggered by API calls, it is important to remember that the point of events is that they can be copied across multiple connections, which may have separate identities and permissions. So even if events were dispatched from a context where we have an identity, we must change to the correct identity of the connection that will be receiving the event, rather than basing a decision on the context that triggered the event, when deciding whether to filter an event to a particular connection. If there were an easy way to get from virConnectPtr to the appropriate virIdentityPtr, then object_event.c could adjust the identity prior to checking whether to dispatch an event. But setting up that back-reference is a bit invasive. Instead, it is easier to delay the filtering check until lower down the stack, at the point where we have direct access to the RPC client object that owns an identity. As such, this patch ends up reverting a large portion of the framework of commit f9f56340. We also have to teach 'make check' to special-case the fact that the event registration filtering is done at the point of dispatch, rather than the point of registration. Note that even though we don't actually use virConnectDomainEventRegisterCheckACL (because the RegisterAny variant is sufficient), we still generate the function for the purposes of documenting that the filtering takes place. Also note that I did not entirely delete the notion of a filter from object_event.c; I still plan on using that for my upcoming patch series for qemu monitor events in libvirt-qemu.so. In other words, while this patch changes ACL filtering to live in remote.c and therefore we have no current client of the filtering in object_event.c, the notion of filtering in object_event.c is still useful down the road. * src/check-aclrules.pl: Exempt event registration from having to pass checkACL filter down call stack. * daemon/remote.c (remoteRelayDomainEventCheckACL) (remoteRelayNetworkEventCheckACL): New functions. (remoteRelay*Event*): Use new functions. * src/conf/domain_event.h (virDomainEventStateRegister) (virDomainEventStateRegisterID): Drop unused parameter. * src/conf/network_event.h (virNetworkEventStateRegisterID): Likewise. * src/conf/domain_event.c (virDomainEventFilter): Delete unused function. * src/conf/network_event.c (virNetworkEventFilter): Likewise. * src/libxl/libxl_driver.c: Adjust caller. * src/lxc/lxc_driver.c: Likewise. * src/network/bridge_driver.c: Likewise. * src/qemu/qemu_driver.c: Likewise. * src/remote/remote_driver.c: Likewise. * src/test/test_driver.c: Likewise. * src/uml/uml_driver.c: Likewise. * src/vbox/vbox_tmpl.c: Likewise. * src/xen/xen_driver.c: Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-28 21:50:02 +00:00
if (callback->callbackID < 0 ||
!remoteRelayNetworkEventCheckACL(callback->client, conn, net))
return -1;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
VIR_DEBUG("Relaying network lifecycle event %d, detail %d, callback %d",
event, detail, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_network(&data.net, net);
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
data.callbackID = callback->callbackID;
data.event = event;
data.detail = detail;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_NETWORK_EVENT_LIFECYCLE,
(xdrproc_t)xdr_remote_network_event_lifecycle_msg, &data);
return 0;
}
static virConnectNetworkEventGenericCallback networkEventCallbacks[] = {
VIR_NETWORK_EVENT_CALLBACK(remoteRelayNetworkEventLifecycle),
};
verify(ARRAY_CARDINALITY(networkEventCallbacks) == VIR_NETWORK_EVENT_ID_LAST);
static int
remoteRelayStoragePoolEventLifecycle(virConnectPtr conn,
virStoragePoolPtr pool,
int event,
int detail,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_storage_pool_event_lifecycle_msg data;
if (callback->callbackID < 0 ||
!remoteRelayStoragePoolEventCheckACL(callback->client, conn, pool))
return -1;
VIR_DEBUG("Relaying storage pool lifecycle event %d, detail %d, callback %d",
event, detail, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_storage_pool(&data.pool, pool);
data.callbackID = callback->callbackID;
data.event = event;
data.detail = detail;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_STORAGE_POOL_EVENT_LIFECYCLE,
(xdrproc_t)xdr_remote_storage_pool_event_lifecycle_msg,
&data);
return 0;
}
static int
remoteRelayStoragePoolEventRefresh(virConnectPtr conn,
virStoragePoolPtr pool,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_storage_pool_event_refresh_msg data;
if (callback->callbackID < 0 ||
!remoteRelayStoragePoolEventCheckACL(callback->client, conn, pool))
return -1;
VIR_DEBUG("Relaying storage pool refresh event callback %d",
callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_storage_pool(&data.pool, pool);
data.callbackID = callback->callbackID;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_STORAGE_POOL_EVENT_REFRESH,
(xdrproc_t)xdr_remote_storage_pool_event_refresh_msg,
&data);
return 0;
}
static virConnectStoragePoolEventGenericCallback storageEventCallbacks[] = {
VIR_STORAGE_POOL_EVENT_CALLBACK(remoteRelayStoragePoolEventLifecycle),
VIR_STORAGE_POOL_EVENT_CALLBACK(remoteRelayStoragePoolEventRefresh),
};
verify(ARRAY_CARDINALITY(storageEventCallbacks) == VIR_STORAGE_POOL_EVENT_ID_LAST);
static int
remoteRelayNodeDeviceEventLifecycle(virConnectPtr conn,
virNodeDevicePtr dev,
int event,
int detail,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_node_device_event_lifecycle_msg data;
if (callback->callbackID < 0 ||
!remoteRelayNodeDeviceEventCheckACL(callback->client, conn, dev))
return -1;
VIR_DEBUG("Relaying node device lifecycle event %d, detail %d, callback %d",
event, detail, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_node_device(&data.dev, dev);
data.callbackID = callback->callbackID;
data.event = event;
data.detail = detail;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_NODE_DEVICE_EVENT_LIFECYCLE,
(xdrproc_t)xdr_remote_node_device_event_lifecycle_msg,
&data);
return 0;
}
static int
remoteRelayNodeDeviceEventUpdate(virConnectPtr conn,
virNodeDevicePtr dev,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_node_device_event_update_msg data;
if (callback->callbackID < 0 ||
!remoteRelayNodeDeviceEventCheckACL(callback->client, conn, dev))
return -1;
VIR_DEBUG("Relaying node device update event callback %d",
callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_node_device(&data.dev, dev);
data.callbackID = callback->callbackID;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_NODE_DEVICE_EVENT_UPDATE,
(xdrproc_t)xdr_remote_node_device_event_update_msg,
&data);
return 0;
}
static virConnectNodeDeviceEventGenericCallback nodeDeviceEventCallbacks[] = {
VIR_NODE_DEVICE_EVENT_CALLBACK(remoteRelayNodeDeviceEventLifecycle),
VIR_NODE_DEVICE_EVENT_CALLBACK(remoteRelayNodeDeviceEventUpdate),
};
verify(ARRAY_CARDINALITY(nodeDeviceEventCallbacks) == VIR_NODE_DEVICE_EVENT_ID_LAST);
static int
remoteRelaySecretEventLifecycle(virConnectPtr conn,
virSecretPtr secret,
int event,
int detail,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_secret_event_lifecycle_msg data;
if (callback->callbackID < 0 ||
!remoteRelaySecretEventCheckACL(callback->client, conn, secret))
return -1;
VIR_DEBUG("Relaying node secretice lifecycle event %d, detail %d, callback %d",
event, detail, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_secret(&data.secret, secret);
data.callbackID = callback->callbackID;
data.event = event;
data.detail = detail;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_SECRET_EVENT_LIFECYCLE,
(xdrproc_t)xdr_remote_secret_event_lifecycle_msg,
&data);
return 0;
}
static int
remoteRelaySecretEventValueChanged(virConnectPtr conn,
virSecretPtr secret,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
remote_secret_event_value_changed_msg data;
if (callback->callbackID < 0 ||
!remoteRelaySecretEventCheckACL(callback->client, conn, secret))
return -1;
VIR_DEBUG("Relaying node secret value changed callback %d",
callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
make_nonnull_secret(&data.secret, secret);
data.callbackID = callback->callbackID;
remoteDispatchObjectEventSend(callback->client, remoteProgram,
REMOTE_PROC_SECRET_EVENT_VALUE_CHANGED,
(xdrproc_t)xdr_remote_secret_event_value_changed_msg,
&data);
return 0;
}
static virConnectSecretEventGenericCallback secretEventCallbacks[] = {
VIR_SECRET_EVENT_CALLBACK(remoteRelaySecretEventLifecycle),
VIR_SECRET_EVENT_CALLBACK(remoteRelaySecretEventValueChanged),
};
verify(ARRAY_CARDINALITY(secretEventCallbacks) == VIR_SECRET_EVENT_ID_LAST);
static void
remoteRelayDomainQemuMonitorEvent(virConnectPtr conn,
virDomainPtr dom,
const char *event,
long long seconds,
unsigned int micros,
const char *details,
void *opaque)
{
daemonClientEventCallbackPtr callback = opaque;
qemu_domain_monitor_event_msg data;
char **details_p = NULL;
if (callback->callbackID < 0 ||
!remoteRelayDomainQemuMonitorEventCheckACL(callback->client, conn,
dom))
return;
VIR_DEBUG("Relaying qemu monitor event %s %s, callback %d",
event, details, callback->callbackID);
/* build return data */
memset(&data, 0, sizeof(data));
data.callbackID = callback->callbackID;
if (VIR_STRDUP(data.event, event) < 0)
goto error;
data.seconds = seconds;
data.micros = micros;
if (details &&
((VIR_ALLOC(details_p) < 0) ||
VIR_STRDUP(*details_p, details) < 0))
goto error;
data.details = details_p;
make_nonnull_domain(&data.dom, dom);
remoteDispatchObjectEventSend(callback->client, qemuProgram,
QEMU_PROC_DOMAIN_MONITOR_EVENT,
(xdrproc_t)xdr_qemu_domain_monitor_event_msg,
&data);
return;
error:
VIR_FREE(data.event);
VIR_FREE(details_p);
}
static
void remoteRelayConnectionClosedEvent(virConnectPtr conn ATTRIBUTE_UNUSED, int reason, void *opaque)
{
virNetServerClientPtr client = opaque;
VIR_DEBUG("Relaying connection closed event, reason %d", reason);
remote_connect_event_connection_closed_msg msg = { reason };
remoteDispatchObjectEventSend(client, remoteProgram,
REMOTE_PROC_CONNECT_EVENT_CONNECTION_CLOSED,
(xdrproc_t)xdr_remote_connect_event_connection_closed_msg,
&msg);
}
#define DEREG_CB(conn, eventCallbacks, neventCallbacks, deregFcn, name) \
do { \
size_t i; \
for (i = 0; i < neventCallbacks; i++) { \
int callbackID = eventCallbacks[i]->callbackID; \
if (callbackID < 0) { \
VIR_WARN("unexpected incomplete %s callback %zu", name, i); \
continue; \
} \
VIR_DEBUG("Deregistering remote %s event relay %d", \
name, callbackID); \
eventCallbacks[i]->callbackID = -1; \
if (deregFcn(conn, callbackID) < 0) \
VIR_WARN("unexpected %s event deregister failure", name); \
} \
VIR_FREE(eventCallbacks); \
} while (0);
/*
* You must hold lock for at least the client
* We don't free stuff here, merely disconnect the client's
* network socket & resources.
* We keep the libvirt connection open until any async
* jobs have finished, then clean it up elsewhere
*/
void remoteClientFreeFunc(void *data)
{
struct daemonClientPrivate *priv = data;
/* Deregister event delivery callback */
if (priv->conn) {
virIdentityPtr sysident = virIdentityGetSystem();
virIdentitySetCurrent(sysident);
DEREG_CB(priv->conn, priv->domainEventCallbacks,
priv->ndomainEventCallbacks,
virConnectDomainEventDeregisterAny, "domain");
DEREG_CB(priv->conn, priv->networkEventCallbacks,
priv->nnetworkEventCallbacks,
virConnectNetworkEventDeregisterAny, "network");
DEREG_CB(priv->conn, priv->storageEventCallbacks,
priv->nstorageEventCallbacks,
virConnectStoragePoolEventDeregisterAny, "storage");
DEREG_CB(priv->conn, priv->nodeDeviceEventCallbacks,
priv->nnodeDeviceEventCallbacks,
virConnectNodeDeviceEventDeregisterAny, "node device");
DEREG_CB(priv->conn, priv->secretEventCallbacks,
priv->nsecretEventCallbacks,
virConnectSecretEventDeregisterAny, "secret");
DEREG_CB(priv->conn, priv->qemuEventCallbacks,
priv->nqemuEventCallbacks,
virConnectDomainQemuMonitorEventDeregister, "qemu monitor");
if (priv->closeRegistered) {
if (virConnectUnregisterCloseCallback(priv->conn,
remoteRelayConnectionClosedEvent) < 0)
VIR_WARN("unexpected close callback event deregister failure");
}
virConnectClose(priv->conn);
virIdentitySetCurrent(NULL);
virObjectUnref(sysident);
}
VIR_FREE(priv);
}
#undef DEREG_CB
static void remoteClientCloseFunc(virNetServerClientPtr client)
{
struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client);
daemonRemoveAllClientStreams(priv->streams);
}
void *remoteClientInitHook(virNetServerClientPtr client,
void *opaque ATTRIBUTE_UNUSED)
{
struct daemonClientPrivate *priv;
if (VIR_ALLOC(priv) < 0)
return NULL;
if (virMutexInit(&priv->lock) < 0) {
VIR_FREE(priv);
virReportSystemError(errno, "%s", _("unable to init mutex"));
return NULL;
}
virNetServerClientSetCloseHook(client, remoteClientCloseFunc);
return priv;
}
/*----- Functions. -----*/
static int
remoteDispatchConnectOpen(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
struct remote_connect_open_args *args)
{
const char *name;
unsigned int flags;
struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client);
int rv = -1;
VIR_DEBUG("priv=%p conn=%p", priv, priv->conn);
virMutexLock(&priv->lock);
/* Already opened? */
if (priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection already open"));
goto cleanup;
}
name = args->name ? *args->name : NULL;
/* If this connection arrived on a readonly socket, force
* the connection to be readonly.
*/
flags = args->flags;
if (virNetServerClientGetReadonly(client))
flags |= VIR_CONNECT_RO;
priv->conn =
flags & VIR_CONNECT_RO
? virConnectOpenReadOnly(name)
: virConnectOpen(name);
if (priv->conn == NULL)
goto cleanup;
rv = 0;
2008-12-04 22:12:53 +00:00
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectClose(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED)
{
virNetServerClientDelayedClose(client);
2008-12-04 22:12:53 +00:00
return 0;
}
static int
remoteDispatchDomainGetSchedulerType(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_scheduler_type_args *args,
remote_domain_get_scheduler_type_ret *ret)
{
virDomainPtr dom = NULL;
char *type;
int nparams;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (!(type = virDomainGetSchedulerType(dom, &nparams)))
goto cleanup;
ret->type = type;
ret->nparams = nparams;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetSchedulerParameters(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_scheduler_parameters_args *args,
remote_domain_get_scheduler_parameters_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->nparams > REMOTE_DOMAIN_SCHEDULER_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetSchedulerParameters(dom, params, &nparams) < 0)
goto cleanup;
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
0) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetSchedulerParametersFlags(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_scheduler_parameters_flags_args *args,
remote_domain_get_scheduler_parameters_flags_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->nparams > REMOTE_DOMAIN_SCHEDULER_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetSchedulerParametersFlags(dom, params, &nparams,
args->flags) < 0)
goto cleanup;
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
args->flags) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMemoryStats(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_memory_stats_args *args,
remote_domain_memory_stats_ret *ret)
{
virDomainPtr dom = NULL;
virDomainMemoryStatPtr stats = NULL;
int nr_stats;
size_t i;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->maxStats > REMOTE_DOMAIN_MEMORY_STATS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("maxStats > REMOTE_DOMAIN_MEMORY_STATS_MAX"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
/* Allocate stats array for making dispatch call */
if (VIR_ALLOC_N(stats, args->maxStats) < 0)
goto cleanup;
libvirt: do not mix internal flags into public API There were two API in driver.c that were silently masking flags bits prior to calling out to the drivers, and several others that were explicitly masking flags bits. This is not forward-compatible - if we ever have that many flags in the future, then talking to an old server that masks out the flags would be indistinguishable from talking to a new server that can honor the flag. In general, libvirt.c should forward _all_ flags on to drivers, and only the drivers should reject unknown flags. In the case of virDrvSecretGetValue, the solution is to separate the internal driver callback function to have two parameters instead of one, with only one parameter affected by the public API. In the case of virDomainGetXMLDesc, it turns out that no one was ever mixing VIR_DOMAIN_XML_INTERNAL_STATUS with the dumpxml path in the first place; that internal flag was only used in saving and restoring state files, which happened to be in functions internal to a single file, so there is no mixing of the internal flag with a public flags argument. Additionally, virDomainMemoryStats passed a flags argument over RPC, but not to the driver. * src/driver.h (VIR_DOMAIN_XML_FLAGS_MASK) (VIR_SECRET_GET_VALUE_FLAGS_MASK): Delete. (virDrvSecretGetValue): Separate out internal flags. (virDrvDomainMemoryStats): Provide missing flags argument. * src/driver.c (verify): Drop unused check. * src/conf/domain_conf.h (virDomainObjParseFile): Delete declaration. (virDomainXMLInternalFlags): Move... * src/conf/domain_conf.c: ...here. Delete redundant include. (virDomainObjParseFile): Make static. * src/libvirt.c (virDomainGetXMLDesc, virSecretGetValue): Update clients. (virDomainMemoryPeek, virInterfaceGetXMLDesc) (virDomainMemoryStats, virDomainBlockPeek, virNetworkGetXMLDesc) (virStoragePoolGetXMLDesc, virStorageVolGetXMLDesc) (virNodeNumOfDevices, virNodeListDevices, virNWFilterGetXMLDesc): Don't mask unknown flags. * src/interface/netcf_driver.c (interfaceGetXMLDesc): Reject unknown flags. * src/secret/secret_driver.c (secretGetValue): Update clients. * src/remote/remote_driver.c (remoteSecretGetValue) (remoteDomainMemoryStats): Likewise. * src/qemu/qemu_process.c (qemuProcessGetVolumeQcowPassphrase): Likewise. * src/qemu/qemu_driver.c (qemudDomainMemoryStats): Likewise. * daemon/remote.c (remoteDispatchDomainMemoryStats): Likewise.
2011-07-13 21:31:56 +00:00
nr_stats = virDomainMemoryStats(dom, stats, args->maxStats, args->flags);
if (nr_stats < 0)
goto cleanup;
/* Allocate return buffer */
if (VIR_ALLOC_N(ret->stats.stats_val, args->maxStats) < 0)
goto cleanup;
/* Copy the stats into the xdr return structure */
for (i = 0; i < nr_stats; i++) {
ret->stats.stats_val[i].tag = stats[i].tag;
ret->stats.stats_val[i].val = stats[i].val;
}
ret->stats.stats_len = nr_stats;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
VIR_FREE(stats);
return rv;
}
static int
remoteDispatchDomainBlockPeek(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_block_peek_args *args,
remote_domain_block_peek_ret *ret)
{
virDomainPtr dom = NULL;
char *path;
unsigned long long offset;
size_t size;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
path = args->path;
offset = args->offset;
size = args->size;
flags = args->flags;
if (size > REMOTE_DOMAIN_BLOCK_PEEK_BUFFER_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("size > maximum buffer size"));
goto cleanup;
}
ret->buffer.buffer_len = size;
if (VIR_ALLOC_N(ret->buffer.buffer_val, size) < 0)
goto cleanup;
if (virDomainBlockPeek(dom, path, offset, size,
ret->buffer.buffer_val, flags) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(ret->buffer.buffer_val);
}
virObjectUnref(dom);
return rv;
}
2011-09-05 08:20:03 +00:00
static int
remoteDispatchDomainBlockStatsFlags(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
2011-09-05 08:20:03 +00:00
virNetMessageErrorPtr rerr,
remote_domain_block_stats_flags_args *args,
remote_domain_block_stats_flags_ret *ret)
{
virTypedParameterPtr params = NULL;
virDomainPtr dom = NULL;
const char *path = args->path;
int nparams = 0;
2011-09-05 08:20:03 +00:00
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
2011-09-05 08:20:03 +00:00
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
flags = args->flags;
if (args->nparams > REMOTE_DOMAIN_BLOCK_STATS_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
2011-09-05 08:20:03 +00:00
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
2011-09-05 08:20:03 +00:00
goto cleanup;
nparams = args->nparams;
2011-09-05 08:20:03 +00:00
if (virDomainBlockStatsFlags(dom, path, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
/* Serialize the block stats. */
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
args->flags) < 0)
2011-09-05 08:20:03 +00:00
goto cleanup;
success:
2011-09-05 08:20:03 +00:00
rv = 0;
cleanup:
if (rv < 0)
2011-09-05 08:20:03 +00:00
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
2011-09-05 08:20:03 +00:00
return rv;
}
static int
remoteDispatchDomainMemoryPeek(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_memory_peek_args *args,
remote_domain_memory_peek_ret *ret)
{
virDomainPtr dom = NULL;
unsigned long long offset;
size_t size;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
offset = args->offset;
size = args->size;
flags = args->flags;
if (size > REMOTE_DOMAIN_MEMORY_PEEK_BUFFER_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR,
"%s", _("size > maximum buffer size"));
goto cleanup;
}
ret->buffer.buffer_len = size;
if (VIR_ALLOC_N(ret->buffer.buffer_val, size) < 0)
goto cleanup;
if (virDomainMemoryPeek(dom, offset, size,
ret->buffer.buffer_val, flags) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(ret->buffer.buffer_val);
}
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetSecurityLabel(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_security_label_args *args,
remote_domain_get_security_label_ret *ret)
{
virDomainPtr dom = NULL;
virSecurityLabelPtr seclabel = NULL;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (VIR_ALLOC(seclabel) < 0)
goto cleanup;
if (virDomainGetSecurityLabel(dom, seclabel) < 0)
goto cleanup;
ret->label.label_len = strlen(seclabel->label) + 1;
if (VIR_ALLOC_N(ret->label.label_val, ret->label.label_len) < 0)
goto cleanup;
strcpy(ret->label.label_val, seclabel->label);
ret->enforcing = seclabel->enforcing;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
VIR_FREE(seclabel);
return rv;
}
static int
remoteDispatchDomainGetSecurityLabelList(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_security_label_list_args *args,
remote_domain_get_security_label_list_ret *ret)
{
virDomainPtr dom = NULL;
virSecurityLabelPtr seclabels = NULL;
int len, rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if ((len = virDomainGetSecurityLabelList(dom, &seclabels)) < 0) {
ret->ret = len;
ret->labels.labels_len = 0;
ret->labels.labels_val = NULL;
goto done;
}
if (VIR_ALLOC_N(ret->labels.labels_val, len) < 0)
goto cleanup;
for (i = 0; i < len; i++) {
size_t label_len = strlen(seclabels[i].label) + 1;
remote_domain_get_security_label_ret *cur = &ret->labels.labels_val[i];
if (VIR_ALLOC_N(cur->label.label_val, label_len) < 0)
goto cleanup;
if (virStrcpy(cur->label.label_val, seclabels[i].label, label_len) == NULL) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("failed to copy security label"));
goto cleanup;
}
cur->label.label_len = label_len;
cur->enforcing = seclabels[i].enforcing;
}
ret->labels.labels_len = ret->ret = len;
done:
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
VIR_FREE(seclabels);
return rv;
}
static int
remoteDispatchNodeGetSecurityModel(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_get_security_model_ret *ret)
{
virSecurityModel secmodel;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
memset(&secmodel, 0, sizeof(secmodel));
if (virNodeGetSecurityModel(priv->conn, &secmodel) < 0)
goto cleanup;
ret->model.model_len = strlen(secmodel.model) + 1;
if (VIR_ALLOC_N(ret->model.model_val, ret->model.model_len) < 0)
goto cleanup;
strcpy(ret->model.model_val, secmodel.model);
ret->doi.doi_len = strlen(secmodel.doi) + 1;
if (VIR_ALLOC_N(ret->doi.doi_val, ret->doi.doi_len) < 0)
goto cleanup;
strcpy(ret->doi.doi_val, secmodel.doi);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
return rv;
}
static int
remoteDispatchDomainGetVcpuPinInfo(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_vcpu_pin_info_args *args,
remote_domain_get_vcpu_pin_info_ret *ret)
{
virDomainPtr dom = NULL;
unsigned char *cpumaps = NULL;
int num;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (args->ncpumaps > REMOTE_VCPUINFO_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("ncpumaps > REMOTE_VCPUINFO_MAX"));
goto cleanup;
}
if (INT_MULTIPLY_OVERFLOW(args->ncpumaps, args->maplen) ||
args->ncpumaps * args->maplen > REMOTE_CPUMAPS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("maxinfo * maplen > REMOTE_CPUMAPS_MAX"));
goto cleanup;
}
/* Allocate buffers to take the results. */
if (args->maplen > 0 &&
VIR_ALLOC_N(cpumaps, args->ncpumaps * args->maplen) < 0)
goto cleanup;
if ((num = virDomainGetVcpuPinInfo(dom,
args->ncpumaps,
cpumaps,
args->maplen,
args->flags)) < 0)
goto cleanup;
ret->num = num;
/* Don't need to allocate/copy the cpumaps if we make the reasonable
* assumption that unsigned char and char are the same size.
* Note that remoteDispatchClientRequest will free.
*/
ret->cpumaps.cpumaps_len = args->ncpumaps * args->maplen;
ret->cpumaps.cpumaps_val = (char *) cpumaps;
cpumaps = NULL;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
VIR_FREE(cpumaps);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainPinEmulator(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_pin_emulator_args *args)
{
int rv = -1;
virDomainPtr dom = NULL;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainPinEmulator(dom,
(unsigned char *) args->cpumap.cpumap_val,
args->cpumap.cpumap_len,
args->flags) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetEmulatorPinInfo(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_emulator_pin_info_args *args,
remote_domain_get_emulator_pin_info_ret *ret)
{
virDomainPtr dom = NULL;
unsigned char *cpumaps = NULL;
int r;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
/* Allocate buffers to take the results */
if (args->maplen > 0 &&
VIR_ALLOC_N(cpumaps, args->maplen) < 0)
goto cleanup;
if ((r = virDomainGetEmulatorPinInfo(dom,
cpumaps,
args->maplen,
args->flags)) < 0)
goto cleanup;
ret->ret = r;
ret->cpumaps.cpumaps_len = args->maplen;
ret->cpumaps.cpumaps_val = (char *) cpumaps;
cpumaps = NULL;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
VIR_FREE(cpumaps);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetVcpus(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_vcpus_args *args,
remote_domain_get_vcpus_ret *ret)
{
virDomainPtr dom = NULL;
virVcpuInfoPtr info = NULL;
unsigned char *cpumaps = NULL;
int info_len;
size_t i;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (args->maxinfo > REMOTE_VCPUINFO_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("maxinfo > REMOTE_VCPUINFO_MAX"));
goto cleanup;
}
if (INT_MULTIPLY_OVERFLOW(args->maxinfo, args->maplen) ||
args->maxinfo * args->maplen > REMOTE_CPUMAPS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("maxinfo * maplen > REMOTE_CPUMAPS_MAX"));
goto cleanup;
}
/* Allocate buffers to take the results. */
if (VIR_ALLOC_N(info, args->maxinfo) < 0)
goto cleanup;
if (args->maplen > 0 &&
VIR_ALLOC_N(cpumaps, args->maxinfo * args->maplen) < 0)
goto cleanup;
if ((info_len = virDomainGetVcpus(dom,
info, args->maxinfo,
cpumaps, args->maplen)) < 0)
goto cleanup;
/* Allocate the return buffer for info. */
ret->info.info_len = info_len;
if (VIR_ALLOC_N(ret->info.info_val, info_len) < 0)
goto cleanup;
for (i = 0; i < info_len; ++i) {
ret->info.info_val[i].number = info[i].number;
ret->info.info_val[i].state = info[i].state;
ret->info.info_val[i].cpu_time = info[i].cpuTime;
ret->info.info_val[i].cpu = info[i].cpu;
}
/* Don't need to allocate/copy the cpumaps if we make the reasonable
* assumption that unsigned char and char are the same size.
* Note that remoteDispatchClientRequest will free.
*/
ret->cpumaps.cpumaps_len = args->maxinfo * args->maplen;
ret->cpumaps.cpumaps_val = (char *) cpumaps;
cpumaps = NULL;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(ret->info.info_val);
}
VIR_FREE(cpumaps);
VIR_FREE(info);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetIOThreadInfo(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_iothread_info_args *args,
remote_domain_get_iothread_info_ret *ret)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client);
virDomainIOThreadInfoPtr *info = NULL;
virDomainPtr dom = NULL;
remote_domain_iothread_info *dst;
int ninfo = 0;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if ((ninfo = virDomainGetIOThreadInfo(dom, &info, args->flags)) < 0)
goto cleanup;
if (ninfo > REMOTE_IOTHREAD_INFO_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many IOThreads in info: %d for limit %d"),
ninfo, REMOTE_IOTHREAD_INFO_MAX);
goto cleanup;
}
if (ninfo) {
if (VIR_ALLOC_N(ret->info.info_val, ninfo) < 0)
goto cleanup;
ret->info.info_len = ninfo;
for (i = 0; i < ninfo; i++) {
dst = &ret->info.info_val[i];
dst->iothread_id = info[i]->iothread_id;
/* No need to allocate/copy the cpumap if we make the reasonable
* assumption that unsigned char and char are the same size.
*/
dst->cpumap.cpumap_len = info[i]->cpumaplen;
dst->cpumap.cpumap_val = (char *)info[i]->cpumap;
info[i]->cpumap = NULL;
}
} else {
ret->info.info_len = 0;
ret->info.info_val = NULL;
}
ret->ret = ninfo;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
if (ninfo >= 0)
for (i = 0; i < ninfo; i++)
virDomainIOThreadInfoFree(info[i]);
VIR_FREE(info);
return rv;
}
static int
remoteDispatchDomainMigratePrepare(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_prepare_args *args,
remote_domain_migrate_prepare_ret *ret)
{
char *cookie = NULL;
int cookielen = 0;
char *uri_in;
char **uri_out;
char *dname;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
uri_in = args->uri_in == NULL ? NULL : *args->uri_in;
dname = args->dname == NULL ? NULL : *args->dname;
/* Wacky world of XDR ... */
if (VIR_ALLOC(uri_out) < 0)
goto cleanup;
if (virDomainMigratePrepare(priv->conn, &cookie, &cookielen,
uri_in, uri_out,
args->flags, dname, args->resource) < 0)
goto cleanup;
/* remoteDispatchClientRequest will free cookie, uri_out and
* the string if there is one.
*/
ret->cookie.cookie_len = cookielen;
ret->cookie.cookie_val = cookie;
if (*uri_out == NULL) {
ret->uri_out = NULL;
} else {
ret->uri_out = uri_out;
uri_out = NULL;
}
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
VIR_FREE(uri_out);
return rv;
}
static int
remoteDispatchDomainMigratePrepare2(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_prepare2_args *args,
remote_domain_migrate_prepare2_ret *ret)
{
char *cookie = NULL;
int cookielen = 0;
char *uri_in;
char **uri_out;
char *dname;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
uri_in = args->uri_in == NULL ? NULL : *args->uri_in;
dname = args->dname == NULL ? NULL : *args->dname;
/* Wacky world of XDR ... */
if (VIR_ALLOC(uri_out) < 0)
goto cleanup;
if (virDomainMigratePrepare2(priv->conn, &cookie, &cookielen,
uri_in, uri_out,
args->flags, dname, args->resource,
args->dom_xml) < 0)
goto cleanup;
/* remoteDispatchClientRequest will free cookie, uri_out and
* the string if there is one.
*/
ret->cookie.cookie_len = cookielen;
ret->cookie.cookie_val = cookie;
ret->uri_out = *uri_out == NULL ? NULL : uri_out;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(uri_out);
}
return rv;
}
static int
remoteDispatchDomainGetMemoryParameters(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_memory_parameters_args *args,
remote_domain_get_memory_parameters_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
if (args->nparams > REMOTE_DOMAIN_MEMORY_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetMemoryParameters(dom, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
args->flags) < 0)
goto cleanup;
success:
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetNumaParameters(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_numa_parameters_args *args,
remote_domain_get_numa_parameters_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
if (args->nparams > REMOTE_DOMAIN_NUMA_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetNumaParameters(dom, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
flags) < 0)
goto cleanup;
success:
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetBlkioParameters(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_blkio_parameters_args *args,
remote_domain_get_blkio_parameters_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
if (args->nparams > REMOTE_DOMAIN_BLKIO_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetBlkioParameters(dom, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
args->flags) < 0)
goto cleanup;
success:
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchNodeGetCPUStats(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_get_cpu_stats_args *args,
remote_node_get_cpu_stats_ret *ret)
{
virNodeCPUStatsPtr params = NULL;
size_t i;
int cpuNum = args->cpuNum;
int nparams = 0;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
if (args->nparams > REMOTE_NODE_CPU_STATS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (virNodeGetCPUStats(priv->conn, cpuNum, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of stats
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
/* Serialise the memory parameters. */
ret->params.params_len = nparams;
if (VIR_ALLOC_N(ret->params.params_val, nparams) < 0)
goto cleanup;
for (i = 0; i < nparams; ++i) {
/* remoteDispatchClientRequest will free this: */
if (VIR_STRDUP(ret->params.params_val[i].field, params[i].field) < 0)
goto cleanup;
ret->params.params_val[i].value = params[i].value;
}
success:
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
if (ret->params.params_val) {
for (i = 0; i < nparams; i++)
VIR_FREE(ret->params.params_val[i].field);
VIR_FREE(ret->params.params_val);
}
}
VIR_FREE(params);
return rv;
}
static int
remoteDispatchNodeGetMemoryStats(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_get_memory_stats_args *args,
remote_node_get_memory_stats_ret *ret)
{
virNodeMemoryStatsPtr params = NULL;
size_t i;
int cellNum = args->cellNum;
int nparams = 0;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
if (args->nparams > REMOTE_NODE_MEMORY_STATS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (virNodeGetMemoryStats(priv->conn, cellNum, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
/* Serialise the memory parameters. */
ret->params.params_len = nparams;
if (VIR_ALLOC_N(ret->params.params_val, nparams) < 0)
goto cleanup;
for (i = 0; i < nparams; ++i) {
/* remoteDispatchClientRequest will free this: */
if (VIR_STRDUP(ret->params.params_val[i].field, params[i].field) < 0)
goto cleanup;
ret->params.params_val[i].value = params[i].value;
}
success:
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
if (ret->params.params_val) {
for (i = 0; i < nparams; i++)
VIR_FREE(ret->params.params_val[i].field);
VIR_FREE(ret->params.params_val);
}
}
VIR_FREE(params);
return rv;
}
static int
remoteDispatchDomainGetPerfEvents(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_perf_events_args *args,
remote_domain_get_perf_events_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetPerfEvents(dom, &params, &nparams, args->flags) < 0)
goto cleanup;
if (nparams > REMOTE_DOMAIN_PERF_EVENTS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
0) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetBlockJobInfo(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_block_job_info_args *args,
remote_domain_get_block_job_info_ret *ret)
{
virDomainPtr dom = NULL;
virDomainBlockJobInfo tmp;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
rv = virDomainGetBlockJobInfo(dom, args->path, &tmp, args->flags);
if (rv <= 0)
goto cleanup;
ret->type = tmp.type;
ret->bandwidth = tmp.bandwidth;
ret->cur = tmp.cur;
ret->end = tmp.end;
ret->found = 1;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetBlockIoTune(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr hdr ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_block_io_tune_args *args,
remote_domain_get_block_io_tune_ret *ret)
{
virDomainPtr dom = NULL;
int rv = -1;
virTypedParameterPtr params = NULL;
int nparams = 0;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->nparams > REMOTE_DOMAIN_BLOCK_IO_TUNE_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetBlockIoTune(dom, args->disk ? *args->disk : NULL,
params, &nparams, args->flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
/* Serialize the block I/O tuning parameters. */
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
args->flags) < 0)
goto cleanup;
success:
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
/*-------------------------------------------------------------*/
static int
remoteDispatchAuthList(virNetServerPtr server,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_list_ret *ret)
{
int rv = -1;
int auth = virNetServerClientGetAuth(client);
uid_t callerUid;
gid_t callerGid;
pid_t callerPid;
unsigned long long timestamp;
/* If the client is root then we want to bypass the
* policykit auth to avoid root being denied if
* some piece of polkit isn't present/running
*/
if (auth == VIR_NET_SERVER_SERVICE_AUTH_POLKIT) {
if (virNetServerClientGetUNIXIdentity(client, &callerUid, &callerGid,
&callerPid, &timestamp) < 0) {
/* Don't do anything on error - it'll be validated at next
* phase of auth anyway */
virResetLastError();
} else if (callerUid == 0) {
char *ident;
if (virAsprintf(&ident, "pid:%lld,uid:%d",
(long long) callerPid, (int) callerUid) < 0)
goto cleanup;
VIR_INFO("Bypass polkit auth for privileged client %s", ident);
virNetServerClientSetAuth(client, 0);
virNetServerTrackCompletedAuth(server);
auth = VIR_NET_SERVER_SERVICE_AUTH_NONE;
VIR_FREE(ident);
}
}
ret->types.types_len = 1;
if (VIR_ALLOC_N(ret->types.types_val, ret->types.types_len) < 0)
goto cleanup;
switch (auth) {
case VIR_NET_SERVER_SERVICE_AUTH_NONE:
ret->types.types_val[0] = REMOTE_AUTH_NONE;
break;
case VIR_NET_SERVER_SERVICE_AUTH_POLKIT:
ret->types.types_val[0] = REMOTE_AUTH_POLKIT;
break;
case VIR_NET_SERVER_SERVICE_AUTH_SASL:
ret->types.types_val[0] = REMOTE_AUTH_SASL;
break;
default:
ret->types.types_val[0] = REMOTE_AUTH_NONE;
}
2008-12-04 22:16:40 +00:00
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
return rv;
}
#ifdef WITH_SASL
/*
* Initializes the SASL session in prepare for authentication
* and gives the client a list of allowed mechanisms to choose
*/
static int
remoteDispatchAuthSaslInit(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_sasl_init_ret *ret)
{
virNetSASLSessionPtr sasl = NULL;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virMutexLock(&priv->lock);
2008-12-04 22:16:40 +00:00
VIR_DEBUG("Initialize SASL auth %d", virNetServerClientGetFD(client));
if (virNetServerClientGetAuth(client) != VIR_NET_SERVER_SERVICE_AUTH_SASL ||
priv->sasl != NULL) {
VIR_ERROR(_("client tried invalid SASL init request"));
2008-12-04 22:16:40 +00:00
goto authfail;
}
sasl = virNetSASLSessionNewServer(saslCtxt,
"libvirt",
virNetServerClientLocalAddrStringSASL(client),
virNetServerClientRemoteAddrStringSASL(client));
if (!sasl)
2008-12-04 22:16:40 +00:00
goto authfail;
2013-01-08 21:02:05 +00:00
# if WITH_GNUTLS
/* Inform SASL that we've got an external SSF layer from TLS */
if (virNetServerClientHasTLSSession(client)) {
int ssf;
if ((ssf = virNetServerClientGetTLSKeySize(client)) < 0)
2008-12-04 22:16:40 +00:00
goto authfail;
ssf *= 8; /* key size is bytes, sasl wants bits */
VIR_DEBUG("Setting external SSF %d", ssf);
if (virNetSASLSessionExtKeySize(sasl, ssf) < 0)
2008-12-04 22:16:40 +00:00
goto authfail;
}
# endif
if (virNetServerClientIsSecure(client))
/* If we've got TLS or UNIX domain sock, we don't care about SSF */
virNetSASLSessionSecProps(sasl, 0, 0, true);
else
/* Plain TCP, better get an SSF layer */
virNetSASLSessionSecProps(sasl,
56, /* Good enough to require kerberos */
100000, /* Arbitrary big number */
false); /* No anonymous */
if (!(ret->mechlist = virNetSASLSessionListMechanisms(sasl)))
2008-12-04 22:16:40 +00:00
goto authfail;
VIR_DEBUG("Available mechanisms for client: '%s'", ret->mechlist);
priv->sasl = sasl;
virMutexUnlock(&priv->lock);
return 0;
2008-12-04 22:16:40 +00:00
authfail:
virResetLastError();
virReportError(VIR_ERR_AUTH_FAILED, "%s",
_("authentication failed"));
virNetMessageSaveError(rerr);
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_FAIL,
"client=%p auth=%d",
client, REMOTE_AUTH_SASL);
virObjectUnref(sasl);
virMutexUnlock(&priv->lock);
2008-12-04 22:16:40 +00:00
return -1;
}
/*
* Returns 0 if ok, -1 on error, -2 if rejected
*/
static int
remoteSASLFinish(virNetServerPtr server,
virNetServerClientPtr client)
{
virIdentityPtr clnt_identity = NULL;
const char *identity;
struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client);
int ssf;
/* TLS or UNIX domain sockets trivially OK */
if (!virNetServerClientIsSecure(client)) {
if ((ssf = virNetSASLSessionGetKeySize(priv->sasl)) < 0)
goto error;
VIR_DEBUG("negotiated an SSF of %d", ssf);
if (ssf < 56) { /* 56 is good for Kerberos */
VIR_ERROR(_("negotiated SSF %d was not strong enough"), ssf);
return -2;
}
}
if (!(identity = virNetSASLSessionGetIdentity(priv->sasl)))
return -2;
if (!virNetSASLContextCheckIdentity(saslCtxt, identity))
return -2;
if (!(clnt_identity = virNetServerClientGetIdentity(client)))
goto error;
virNetServerClientSetAuth(client, 0);
virNetServerTrackCompletedAuth(server);
virNetServerClientSetSASLSession(client, priv->sasl);
virIdentitySetSASLUserName(clnt_identity, identity);
VIR_DEBUG("Authentication successful %d", virNetServerClientGetFD(client));
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_ALLOW,
"client=%p auth=%d identity=%s",
client, REMOTE_AUTH_SASL, identity);
virObjectUnref(clnt_identity);
virObjectUnref(priv->sasl);
priv->sasl = NULL;
return 0;
error:
return -1;
}
/*
* This starts the SASL authentication negotiation.
*/
static int
remoteDispatchAuthSaslStart(virNetServerPtr server,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_sasl_start_args *args,
remote_auth_sasl_start_ret *ret)
{
const char *serverout;
size_t serveroutlen;
int err;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
const char *identity;
virMutexLock(&priv->lock);
2008-12-04 22:16:40 +00:00
VIR_DEBUG("Start SASL auth %d", virNetServerClientGetFD(client));
if (virNetServerClientGetAuth(client) != VIR_NET_SERVER_SERVICE_AUTH_SASL ||
priv->sasl == NULL) {
VIR_ERROR(_("client tried invalid SASL start request"));
2008-12-04 22:16:40 +00:00
goto authfail;
}
VIR_DEBUG("Using SASL mechanism %s. Data %d bytes, nil: %d",
args->mech, args->data.data_len, args->nil);
err = virNetSASLSessionServerStart(priv->sasl,
args->mech,
/* NB, distinction of NULL vs "" is *critical* in SASL */
args->nil ? NULL : args->data.data_val,
args->data.data_len,
&serverout,
&serveroutlen);
if (err != VIR_NET_SASL_COMPLETE &&
err != VIR_NET_SASL_CONTINUE)
2008-12-04 22:16:40 +00:00
goto authfail;
if (serveroutlen > REMOTE_AUTH_SASL_DATA_MAX) {
VIR_ERROR(_("sasl start reply data too long %d"), (int)serveroutlen);
2008-12-04 22:16:40 +00:00
goto authfail;
}
/* NB, distinction of NULL vs "" is *critical* in SASL */
if (serverout) {
if (VIR_ALLOC_N(ret->data.data_val, serveroutlen) < 0)
goto authfail;
memcpy(ret->data.data_val, serverout, serveroutlen);
} else {
ret->data.data_val = NULL;
}
ret->nil = serverout ? 0 : 1;
ret->data.data_len = serveroutlen;
VIR_DEBUG("SASL return data %d bytes, nil; %d", ret->data.data_len, ret->nil);
if (err == VIR_NET_SASL_CONTINUE) {
ret->complete = 0;
} else {
/* Check username whitelist ACL */
if ((err = remoteSASLFinish(server, client)) < 0) {
if (err == -2)
goto authdeny;
else
goto authfail;
}
ret->complete = 1;
}
virMutexUnlock(&priv->lock);
return 0;
2008-12-04 22:16:40 +00:00
authfail:
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_FAIL,
"client=%p auth=%d",
client, REMOTE_AUTH_SASL);
goto error;
authdeny:
identity = virNetSASLSessionGetIdentity(priv->sasl);
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_DENY,
"client=%p auth=%d identity=%s",
client, REMOTE_AUTH_SASL, identity);
goto error;
error:
virObjectUnref(priv->sasl);
priv->sasl = NULL;
virResetLastError();
virReportError(VIR_ERR_AUTH_FAILED, "%s",
_("authentication failed"));
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
2008-12-04 22:16:40 +00:00
return -1;
}
static int
remoteDispatchAuthSaslStep(virNetServerPtr server,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_sasl_step_args *args,
remote_auth_sasl_step_ret *ret)
{
const char *serverout;
size_t serveroutlen;
int err;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
const char *identity;
2008-12-04 22:16:40 +00:00
virMutexLock(&priv->lock);
VIR_DEBUG("Step SASL auth %d", virNetServerClientGetFD(client));
if (virNetServerClientGetAuth(client) != VIR_NET_SERVER_SERVICE_AUTH_SASL ||
priv->sasl == NULL) {
VIR_ERROR(_("client tried invalid SASL start request"));
2008-12-04 22:16:40 +00:00
goto authfail;
}
VIR_DEBUG("Step using SASL Data %d bytes, nil: %d",
args->data.data_len, args->nil);
err = virNetSASLSessionServerStep(priv->sasl,
/* NB, distinction of NULL vs "" is *critical* in SASL */
args->nil ? NULL : args->data.data_val,
args->data.data_len,
&serverout,
&serveroutlen);
if (err != VIR_NET_SASL_COMPLETE &&
err != VIR_NET_SASL_CONTINUE)
2008-12-04 22:16:40 +00:00
goto authfail;
if (serveroutlen > REMOTE_AUTH_SASL_DATA_MAX) {
2009-01-06 18:32:03 +00:00
VIR_ERROR(_("sasl step reply data too long %d"),
(int)serveroutlen);
2008-12-04 22:16:40 +00:00
goto authfail;
}
/* NB, distinction of NULL vs "" is *critical* in SASL */
if (serverout) {
if (VIR_ALLOC_N(ret->data.data_val, serveroutlen) < 0)
goto authfail;
memcpy(ret->data.data_val, serverout, serveroutlen);
} else {
ret->data.data_val = NULL;
}
ret->nil = serverout ? 0 : 1;
ret->data.data_len = serveroutlen;
VIR_DEBUG("SASL return data %d bytes, nil; %d", ret->data.data_len, ret->nil);
if (err == VIR_NET_SASL_CONTINUE) {
ret->complete = 0;
} else {
/* Check username whitelist ACL */
if ((err = remoteSASLFinish(server, client)) < 0) {
if (err == -2)
goto authdeny;
else
goto authfail;
}
ret->complete = 1;
}
virMutexUnlock(&priv->lock);
return 0;
2008-12-04 22:16:40 +00:00
authfail:
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_FAIL,
"client=%p auth=%d",
client, REMOTE_AUTH_SASL);
goto error;
authdeny:
identity = virNetSASLSessionGetIdentity(priv->sasl);
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_DENY,
"client=%p auth=%d identity=%s",
client, REMOTE_AUTH_SASL, identity);
goto error;
error:
virObjectUnref(priv->sasl);
priv->sasl = NULL;
virResetLastError();
virReportError(VIR_ERR_AUTH_FAILED, "%s",
_("authentication failed"));
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return -1;
}
#else
static int
remoteDispatchAuthSaslInit(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_sasl_init_ret *ret ATTRIBUTE_UNUSED)
{
VIR_WARN("Client tried unsupported SASL auth");
virReportError(VIR_ERR_AUTH_FAILED, "%s",
_("authentication failed"));
virNetMessageSaveError(rerr);
return -1;
}
static int
remoteDispatchAuthSaslStart(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_sasl_start_args *args ATTRIBUTE_UNUSED,
remote_auth_sasl_start_ret *ret ATTRIBUTE_UNUSED)
{
VIR_WARN("Client tried unsupported SASL auth");
virReportError(VIR_ERR_AUTH_FAILED, "%s",
_("authentication failed"));
virNetMessageSaveError(rerr);
return -1;
}
static int
remoteDispatchAuthSaslStep(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_sasl_step_args *args ATTRIBUTE_UNUSED,
remote_auth_sasl_step_ret *ret ATTRIBUTE_UNUSED)
{
VIR_WARN("Client tried unsupported SASL auth");
virReportError(VIR_ERR_AUTH_FAILED, "%s",
_("authentication failed"));
virNetMessageSaveError(rerr);
return -1;
}
#endif
static int
remoteDispatchAuthPolkit(virNetServerPtr server,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_auth_polkit_ret *ret)
{
pid_t callerPid = -1;
gid_t callerGid = -1;
uid_t callerUid = -1;
unsigned long long timestamp;
const char *action;
char *ident = NULL;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
int rv;
virMutexLock(&priv->lock);
action = virNetServerClientGetReadonly(client) ?
"org.libvirt.unix.monitor" :
"org.libvirt.unix.manage";
VIR_DEBUG("Start PolicyKit auth %d", virNetServerClientGetFD(client));
if (virNetServerClientGetAuth(client) != VIR_NET_SERVER_SERVICE_AUTH_POLKIT) {
VIR_ERROR(_("client tried invalid PolicyKit init request"));
goto authfail;
}
if (virNetServerClientGetUNIXIdentity(client, &callerUid, &callerGid,
&callerPid, &timestamp) < 0) {
goto authfail;
}
if (timestamp == 0) {
VIR_WARN("Failing polkit auth due to missing client (pid=%lld) start time",
(long long)callerPid);
goto authfail;
}
VIR_INFO("Checking PID %lld running as %d",
(long long) callerPid, callerUid);
rv = virPolkitCheckAuth(action,
callerPid,
timestamp,
callerUid,
NULL,
true);
if (rv == -1)
goto authfail;
else if (rv == -2)
goto authdeny;
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_ALLOW,
"client=%p auth=%d identity=%s",
client, REMOTE_AUTH_POLKIT, ident);
VIR_INFO("Policy allowed action %s from pid %lld, uid %d",
action, (long long) callerPid, callerUid);
ret->complete = 1;
virNetServerClientSetAuth(client, 0);
virNetServerTrackCompletedAuth(server);
virMutexUnlock(&priv->lock);
2008-12-04 22:16:40 +00:00
return 0;
error:
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return -1;
authfail:
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_FAIL,
"client=%p auth=%d",
client, REMOTE_AUTH_POLKIT);
goto error;
authdeny:
Rewrite all the DTrace/SystemTAP probing The libvirtd daemon had a few crude system tap probes. Some of these were broken during the RPC rewrite. The new modular RPC code is structured in a way that allows much more effective tracing. Instead of trying to hook up the original probes, define a new set of probes for the RPC and event code. The master probes file is now src/probes.d. This contains probes for virNetServerClientPtr, virNetClientPtr, virSocketPtr virNetTLSContextPtr and virNetTLSSessionPtr modules. Also add probes for the poll event loop. The src/dtrace2systemtap.pl script can convert the probes.d file into a libvirt_probes.stp file to make use from systemtap much simpler. The src/rpc/gensystemtap.pl script can generate a set of systemtap functions for translating RPC enum values into printable strings. This works for all RPC header enums (program, type, status, procedure) and also the authentication enum The PROBE macro will automatically generate a VIR_DEBUG statement, so any place with a PROBE can remove any existing manual DEBUG statements. * daemon/libvirtd.stp, daemon/probes.d: Remove obsolete probing * daemon/libvirtd.h: Remove probe macros * daemon/Makefile.am: Remove all probe buildings/install * daemon/remote.c: Update authentication probes * src/dtrace2systemtap.pl, src/rpc/gensystemtap.pl: Scripts to generate STP files * src/internal.h: Add probe macros * src/probes.d: Master list of probes * src/rpc/virnetclient.c, src/rpc/virnetserverclient.c, src/rpc/virnetsocket.c, src/rpc/virnettlscontext.c, src/util/event_poll.c: Insert probe points, removing any DEBUG statements that duplicate the info
2011-09-30 13:40:23 +00:00
PROBE(RPC_SERVER_CLIENT_AUTH_DENY,
"client=%p auth=%d identity=%s",
client, REMOTE_AUTH_POLKIT, ident);
goto error;
}
/***************************************************************
* NODE INFO APIS
**************************************************************/
static int
remoteDispatchNodeDeviceGetParent(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_device_get_parent_args *args,
remote_node_device_get_parent_ret *ret)
{
virNodeDevicePtr dev = NULL;
const char *parent = NULL;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dev = virNodeDeviceLookupByName(priv->conn, args->name)))
goto cleanup;
parent = virNodeDeviceGetParent(dev);
if (parent == NULL) {
ret->parent = NULL;
} else {
/* remoteDispatchClientRequest will free this. */
char **parent_p;
if (VIR_ALLOC(parent_p) < 0)
goto cleanup;
if (VIR_STRDUP(*parent_p, parent) < 0) {
VIR_FREE(parent_p);
goto cleanup;
}
ret->parent = parent_p;
}
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dev);
return rv;
}
static int
remoteDispatchConnectRegisterCloseCallback(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr)
{
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virMutexLock(&priv->lock);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (virConnectRegisterCloseCallback(priv->conn,
remoteRelayConnectionClosedEvent,
client, NULL) < 0)
goto cleanup;
priv->closeRegistered = true;
rv = 0;
cleanup:
virMutexUnlock(&priv->lock);
if (rv < 0)
virNetMessageSaveError(rerr);
return rv;
}
static int
remoteDispatchConnectUnregisterCloseCallback(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr)
{
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virMutexLock(&priv->lock);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (virConnectUnregisterCloseCallback(priv->conn,
remoteRelayConnectionClosedEvent) < 0)
goto cleanup;
priv->closeRegistered = false;
rv = 0;
cleanup:
virMutexUnlock(&priv->lock);
if (rv < 0)
virNetMessageSaveError(rerr);
return rv;
}
/***************************
* Register / deregister events
***************************/
static int
remoteDispatchConnectDomainEventRegister(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_domain_event_register_ret *ret ATTRIBUTE_UNUSED)
{
int callbackID;
int rv = -1;
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; or you can use VIR_APPEND_ELEMENT_COPY to avoid
* clearing 'callback' and having to juggle the pointer
* between 'ref' and 'callback'.
*/
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
callback->eventID = VIR_DOMAIN_EVENT_ID_LIFECYCLE;
callback->callbackID = -1;
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
callback->legacy = true;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->domainEventCallbacks,
priv->ndomainEventCallbacks,
callback) < 0)
goto cleanup;
if ((callbackID = virConnectDomainEventRegisterAny(priv->conn,
NULL,
VIR_DOMAIN_EVENT_ID_LIFECYCLE,
VIR_DOMAIN_EVENT_CALLBACK(remoteRelayDomainEventLifecycle),
ref,
remoteEventCallbackFree)) < 0) {
VIR_SHRINK_N(priv->domainEventCallbacks,
priv->ndomainEventCallbacks, 1);
callback = ref;
goto cleanup;
}
ref->callbackID = callbackID;
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectDomainEventDeregister(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_domain_event_deregister_ret *ret ATTRIBUTE_UNUSED)
{
int callbackID = -1;
int rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
for (i = 0; i < priv->ndomainEventCallbacks; i++) {
if (priv->domainEventCallbacks[i]->eventID == VIR_DOMAIN_EVENT_ID_LIFECYCLE) {
callbackID = priv->domainEventCallbacks[i]->callbackID;
break;
}
}
if (callbackID < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("domain event %d not registered"),
VIR_DOMAIN_EVENT_ID_LIFECYCLE);
goto cleanup;
}
if (virConnectDomainEventDeregisterAny(priv->conn, callbackID) < 0)
goto cleanup;
VIR_DELETE_ELEMENT(priv->domainEventCallbacks, i,
priv->ndomainEventCallbacks);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static void
remoteDispatchObjectEventSend(virNetServerClientPtr client,
virNetServerProgramPtr program,
int procnr,
xdrproc_t proc,
void *data)
{
virNetMessagePtr msg;
2011-08-31 16:42:58 +00:00
if (!(msg = virNetMessageNew(false)))
goto cleanup;
msg->header.prog = virNetServerProgramGetID(program);
msg->header.vers = virNetServerProgramGetVersion(program);
msg->header.proc = procnr;
msg->header.type = VIR_NET_MESSAGE;
msg->header.serial = 1;
msg->header.status = VIR_NET_OK;
if (virNetMessageEncodeHeader(msg) < 0)
goto cleanup;
if (virNetMessageEncodePayload(msg, proc, data) < 0)
goto cleanup;
VIR_DEBUG("Queue event %d %zu", procnr, msg->bufferLength);
virNetServerClientSendMessage(client, msg);
xdr_free(proc, data);
return;
cleanup:
virNetMessageFree(msg);
xdr_free(proc, data);
}
static int
remoteDispatchSecretGetValue(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_secret_get_value_args *args,
remote_secret_get_value_ret *ret)
{
virSecretPtr secret = NULL;
size_t value_size;
unsigned char *value;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(secret = get_nonnull_secret(priv->conn, args->secret)))
goto cleanup;
if (!(value = virSecretGetValue(secret, &value_size, args->flags)))
goto cleanup;
ret->value.value_len = value_size;
ret->value.value_val = (char *)value;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(secret);
return rv;
}
static int
remoteDispatchDomainGetState(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_state_args *args,
remote_domain_get_state_ret *ret)
{
virDomainPtr dom = NULL;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetState(dom, &ret->state, &ret->reason, args->flags) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
/* Due to back-compat reasons, two RPC calls map to the same libvirt
* API of virConnectDomainEventRegisterAny. A client should only use
* the new call if they have probed
* VIR_DRV_SUPPORTS_FEATURE(VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK),
* and must not mix the two styles. */
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
static int
remoteDispatchConnectDomainEventRegisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_domain_event_register_any_args *args)
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
{
int callbackID;
int rv = -1;
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
/* We intentionally do not use VIR_DOMAIN_EVENT_ID_LAST here; any
* new domain events added after this point should only use the
* modern callback style of RPC. */
if (args->eventID > VIR_DOMAIN_EVENT_ID_DEVICE_REMOVED ||
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
args->eventID < 0) {
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR, _("unsupported event ID %d"),
args->eventID);
goto cleanup;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
}
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; but since VIR_APPEND_ELEMENT clears 'callback' on
* success, we use 'ref' to save a copy of the pointer. */
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
callback->eventID = args->eventID;
callback->callbackID = -1;
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
callback->legacy = true;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->domainEventCallbacks,
priv->ndomainEventCallbacks,
callback) < 0)
goto cleanup;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
if ((callbackID = virConnectDomainEventRegisterAny(priv->conn,
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
NULL,
args->eventID,
domainEventCallbacks[args->eventID],
ref,
remoteEventCallbackFree)) < 0) {
VIR_SHRINK_N(priv->domainEventCallbacks,
priv->ndomainEventCallbacks, 1);
callback = ref;
goto cleanup;
}
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
ref->callbackID = callbackID;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
}
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
static int
remoteDispatchConnectDomainEventCallbackRegisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_domain_event_callback_register_any_args *args,
remote_connect_domain_event_callback_register_any_ret *ret)
{
int callbackID;
int rv = -1;
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virDomainPtr dom = NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
if (args->dom &&
!(dom = get_nonnull_domain(priv->conn, *args->dom)))
goto cleanup;
if (args->eventID >= VIR_DOMAIN_EVENT_ID_LAST || args->eventID < 0) {
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR, _("unsupported event ID %d"),
args->eventID);
goto cleanup;
}
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; but since VIR_APPEND_ELEMENT clears 'callback' on
* success, we use 'ref' to save a copy of the pointer. */
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
callback->eventID = args->eventID;
callback->callbackID = -1;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->domainEventCallbacks,
priv->ndomainEventCallbacks,
callback) < 0)
goto cleanup;
if ((callbackID = virConnectDomainEventRegisterAny(priv->conn,
dom,
args->eventID,
domainEventCallbacks[args->eventID],
ref,
remoteEventCallbackFree)) < 0) {
VIR_SHRINK_N(priv->domainEventCallbacks,
priv->ndomainEventCallbacks, 1);
callback = ref;
goto cleanup;
}
ref->callbackID = callbackID;
ret->callbackID = callbackID;
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
virMutexUnlock(&priv->lock);
return rv;
}
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
static int
remoteDispatchConnectDomainEventDeregisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_domain_event_deregister_any_args *args)
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
{
int callbackID = -1;
int rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
/* We intentionally do not use VIR_DOMAIN_EVENT_ID_LAST here; any
* new domain events added after this point should only use the
* modern callback style of RPC. */
if (args->eventID > VIR_DOMAIN_EVENT_ID_DEVICE_REMOVED ||
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
args->eventID < 0) {
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR, _("unsupported event ID %d"),
args->eventID);
goto cleanup;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
}
for (i = 0; i < priv->ndomainEventCallbacks; i++) {
if (priv->domainEventCallbacks[i]->eventID == args->eventID) {
callbackID = priv->domainEventCallbacks[i]->callbackID;
break;
}
}
if (callbackID < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("domain event %d not registered"), args->eventID);
goto cleanup;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
}
if (virConnectDomainEventDeregisterAny(priv->conn, callbackID) < 0)
goto cleanup;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
VIR_DELETE_ELEMENT(priv->domainEventCallbacks, i,
priv->ndomainEventCallbacks);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
Remote driver & daemon impl of new event API This wires up the remote driver to handle the new events APIs. The public API allows an application to request a callback filters events to a specific domain object, and register multiple callbacks for the same event type. On the wire there are two strategies for this - Register multiple callbacks with the remote daemon, each with filtering as needed - Register only one callback per event type, with no filtering Both approaches have potential inefficiency. In the first scheme, the same event gets sent over the wire many times if multiple callbacks are registered. With the second scheme, unneccessary events get sent over the wire if a per-domain filter is set on the client. The second scheme is far easier to implement though, so this patch takes that approach. * daemon/dispatch.h: Don't export remoteRelayDomainEvent since it is no longer needed for unregistering callbacks, instead the unique callback ID is used * daemon/libvirtd.c, daemon/libvirtd.h: Track and unregister callbacks based on callback ID, instead of function pointer * daemon/remote.c: Switch over to using virConnectDomainEventRegisterAny instead of legacy virConnectDomainEventRegister function. Refactor remoteDispatchDomainEventSend() to cope with arbitrary event types * src/driver.h, src/driver.c: Move verify() call into source file instead of header, to avoid polluting the global namespace with the verify function name * src/remote/remote_driver.c: Implement new APIs for event registration. Refactor processCallDispatchMessage() to cope with arbitrary incoming event types. Merge remoteDomainQueueEvent() into processCallDispatchMessage() to avoid duplication of code. Rename remoteDomainReadEvent() to remoteDomainReadEventLifecycle() * src/remote/remote_protocol.x: Define wire format for the new virConnectDomainEventRegisterAny and virConnectDomainEventDeregisterAny functions
2010-03-18 14:56:56 +00:00
}
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
static int
remoteDispatchConnectDomainEventCallbackDeregisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_domain_event_callback_deregister_any_args *args)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
for (i = 0; i < priv->ndomainEventCallbacks; i++) {
if (priv->domainEventCallbacks[i]->callbackID == args->callbackID)
break;
}
if (i == priv->ndomainEventCallbacks) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("domain event callback %d not registered"),
args->callbackID);
goto cleanup;
}
if (virConnectDomainEventDeregisterAny(priv->conn, args->callbackID) < 0)
goto cleanup;
VIR_DELETE_ELEMENT(priv->domainEventCallbacks, i,
priv->ndomainEventCallbacks);
rv = 0;
cleanup:
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
qemuDispatchDomainMonitorCommand(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
qemu_domain_monitor_command_args *args,
qemu_domain_monitor_command_ret *ret)
{
virDomainPtr dom = NULL;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainQemuMonitorCommand(dom, args->cmd, &ret->result,
args->flags) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMigrateBegin3(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_begin3_args *args,
remote_domain_migrate_begin3_ret *ret)
{
char *xml = NULL;
virDomainPtr dom = NULL;
char *dname;
char *xmlin;
char *cookieout = NULL;
int cookieoutlen = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
xmlin = args->xmlin == NULL ? NULL : *args->xmlin;
dname = args->dname == NULL ? NULL : *args->dname;
if (!(xml = virDomainMigrateBegin3(dom, xmlin,
&cookieout, &cookieoutlen,
args->flags, dname, args->resource)))
goto cleanup;
/* remoteDispatchClientRequest will free cookie and
* the xml string if there is one.
*/
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
ret->xml = xml;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMigratePrepare3(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_prepare3_args *args,
remote_domain_migrate_prepare3_ret *ret)
{
char *cookieout = NULL;
int cookieoutlen = 0;
char *uri_in;
char **uri_out;
char *dname;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
uri_in = args->uri_in == NULL ? NULL : *args->uri_in;
dname = args->dname == NULL ? NULL : *args->dname;
/* Wacky world of XDR ... */
if (VIR_ALLOC(uri_out) < 0)
goto cleanup;
if (virDomainMigratePrepare3(priv->conn,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
&cookieout, &cookieoutlen,
uri_in, uri_out,
args->flags, dname, args->resource,
args->dom_xml) < 0)
goto cleanup;
/* remoteDispatchClientRequest will free cookie, uri_out and
* the string if there is one.
*/
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
ret->uri_out = *uri_out == NULL ? NULL : uri_out;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(uri_out);
}
return rv;
}
static int
remoteDispatchDomainMigratePerform3(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_perform3_args *args,
remote_domain_migrate_perform3_ret *ret)
{
virDomainPtr dom = NULL;
char *xmlin;
char *dname;
char *uri;
char *dconnuri;
char *cookieout = NULL;
int cookieoutlen = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
xmlin = args->xmlin == NULL ? NULL : *args->xmlin;
dname = args->dname == NULL ? NULL : *args->dname;
uri = args->uri == NULL ? NULL : *args->uri;
dconnuri = args->dconnuri == NULL ? NULL : *args->dconnuri;
if (virDomainMigratePerform3(dom, xmlin,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
&cookieout, &cookieoutlen,
dconnuri, uri,
args->flags, dname, args->resource) < 0)
goto cleanup;
/* remoteDispatchClientRequest will free cookie
*/
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMigrateFinish3(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_finish3_args *args,
remote_domain_migrate_finish3_ret *ret)
{
virDomainPtr dom = NULL;
char *cookieout = NULL;
int cookieoutlen = 0;
char *uri;
char *dconnuri;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
uri = args->uri == NULL ? NULL : *args->uri;
dconnuri = args->dconnuri == NULL ? NULL : *args->dconnuri;
if (!(dom = virDomainMigrateFinish3(priv->conn, args->dname,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
&cookieout, &cookieoutlen,
dconnuri, uri,
args->flags,
args->cancelled)))
goto cleanup;
make_nonnull_domain(&ret->dom, dom);
/* remoteDispatchClientRequest will free cookie
*/
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(cookieout);
}
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMigrateConfirm3(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_confirm3_args *args)
{
virDomainPtr dom = NULL;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainMigrateConfirm3(dom,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
args->flags, args->cancelled) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int remoteDispatchConnectSupportsFeature(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_connect_supports_feature_args *args,
remote_connect_supports_feature_ret *ret)
{
int rv = -1;
int supported;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
/* This feature is checked before opening the connection, thus we must
* check it first.
*/
if (args->feature == VIR_DRV_FEATURE_PROGRAM_KEEPALIVE) {
if (virNetServerClientStartKeepAlive(client) < 0)
goto cleanup;
supported = 1;
goto done;
}
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
switch (args->feature) {
case VIR_DRV_FEATURE_FD_PASSING:
event: server RPC protocol tweaks for domain lifecycle events This patch adds some new RPC call numbers, but for ease of review, they sit idle until a later patch adds the client counterpart to drive the new RPCs. Also for ease of review, I limited this patch to just the lifecycle event; although converting the remaining 15 domain events will be quite mechanical. On the server side, we have to have a function per RPC call, largely with duplicated bodies (the key difference being that we store in our callback opaque pointer whether events should be fired with old or new style); meanwhile, a single function can drive multiple RPC messages. With a strategic choice of XDR struct layout, we can make the event generation code for both styles fairly compact. I debated about adding a tri-state witness variable per connection (values 'unknown', 'legacy', 'modern'). It would start as 'unknown', move to 'legacy' if any RPC call is made to a legacy event call, and move to 'modern' if the feature probe is made; then the event code could issue an error if the witness state is incorrect (a legacy RPC call while in 'modern', a modern RPC call while in 'unknown' or 'legacy', and a feature probe while in 'legacy' or 'modern'). But while it might prevent odd behavior caused by protocol fuzzing, I don't see that it would prevent any security holes, so I considered it bloat. Note that sticking @acl markers on the new RPCs generates unused functions in access/viraccessapicheck.c, because there is no new API call that needs to use the new checks; however, having a consistent .x file is worth the dead code. * src/libvirt_internal.h (VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK): New feature. * src/remote/remote_protocol.x (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_REGISTER_ANY) (REMOTE_PROC_CONNECT_DOMAIN_EVENT_CALLBACK_DEREGISTER_ANY) (REMOTE_PROC_DOMAIN_EVENT_CALLBACK_LIFECYCLE): New RPCs. * daemon/remote.c (daemonClientCallback): Add field. (remoteDispatchConnectDomainEventCallbackRegisterAny) (remoteDispatchConnectDomainEventCallbackDeregisterAny): New functions. (remoteDispatchConnectDomainEventRegisterAny) (remoteDispatchConnectDomainEventDeregisterAny): Mark legacy use. (remoteRelayDomainEventLifecycle): Change message based on legacy or new use. (remoteDispatchConnectSupportsFeature): Advertise new feature. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:59:35 +00:00
case VIR_DRV_FEATURE_REMOTE_EVENT_CALLBACK:
case VIR_DRV_FEATURE_REMOTE_CLOSE_CALLBACK:
supported = 1;
break;
default:
if ((supported = virConnectSupportsFeature(priv->conn, args->feature)) < 0)
goto cleanup;
break;
}
done:
ret->supported = supported;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
return rv;
}
static int
remoteDispatchDomainOpenGraphics(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg,
virNetMessageErrorPtr rerr,
remote_domain_open_graphics_args *args)
{
virDomainPtr dom = NULL;
int rv = -1;
int fd = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if ((fd = virNetMessageDupFD(msg, 0)) < 0)
goto cleanup;
if (virDomainOpenGraphics(dom,
args->idx,
fd,
args->flags) < 0)
goto cleanup;
rv = 0;
cleanup:
VIR_FORCE_CLOSE(fd);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainOpenGraphicsFd(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg,
virNetMessageErrorPtr rerr,
remote_domain_open_graphics_fd_args *args)
{
virDomainPtr dom = NULL;
int rv = -1;
int fd = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if ((fd = virDomainOpenGraphicsFD(dom,
args->idx,
args->flags)) < 0)
goto cleanup;
if (virNetMessageAddFD(msg, fd) < 0)
goto cleanup;
/* return 1 here to let virNetServerProgramDispatchCall know
* we are passing a FD */
rv = 1;
cleanup:
VIR_FORCE_CLOSE(fd);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetInterfaceParameters(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_interface_parameters_args *args,
remote_domain_get_interface_parameters_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
const char *device = args->device;
int nparams = 0;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
if (args->nparams > REMOTE_DOMAIN_INTERFACE_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetInterfaceParameters(dom, device, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
flags) < 0)
goto cleanup;
success:
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetCPUStats(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr hdr ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_cpu_stats_args *args,
remote_domain_get_cpu_stats_ret *ret)
{
virDomainPtr dom = NULL;
struct daemonClientPrivate *priv;
virTypedParameterPtr params = NULL;
int rv = -1;
int percpu_len = 0;
priv = virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->nparams > REMOTE_NODE_CPU_STATS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->ncpus > REMOTE_DOMAIN_GET_CPU_STATS_NCPUS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("ncpus too large"));
goto cleanup;
}
if (args->nparams > 0 &&
VIR_ALLOC_N(params, args->ncpus * args->nparams) < 0)
goto cleanup;
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
percpu_len = virDomainGetCPUStats(dom, params, args->nparams,
args->start_cpu, args->ncpus,
args->flags);
if (percpu_len < 0)
goto cleanup;
/* If nparams == 0, the function returns a single value */
if (args->nparams == 0)
goto success;
if (virTypedParamsSerialize(params, args->nparams * args->ncpus,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
args->flags) < 0)
goto cleanup;
success:
rv = 0;
ret->nparams = percpu_len;
rpc: allow truncated return for virDomainGetCPUStats The RPC code assumed that the array returned by the driver would be fully populated; that is, ncpus on entry resulted in ncpus * return value on exit. However, while we don't support holes in the middle of ncpus, we do want to permit the case of ncpus on entry being longer than the array returned by the driver (that is, it should be safe for the caller to pass ncpus=128 on entry, and the driver will stop populating the array when it hits max_id). Additionally, a successful return implies that the caller will then use virTypedParamArrayClear on the entire array; for this to not free uninitialized memory, the driver must ensure that all skipped entries are explicitly zeroed (the RPC driver did this, but not the qemu driver). There are now three cases: server 0.9.10 and client 0.9.10 or newer: No impact - there were no hypervisor drivers that supported cpu stats server 0.9.11 or newer and client 0.9.10: if the client calls with ncpus beyond the max, then the rpc call will fail on the client side and disconnect the client, but the server is no worse for the wear server 0.9.11 or newer and client 0.9.11: the server can return a truncated array and the client will do just fine I reproduced the problem by using a host with 2 CPUs, and doing: virsh cpu-stats $dom --start 1 --count 2 * daemon/remote.c (remoteDispatchDomainGetCPUStats): Allow driver to omit tail of array. * src/remote/remote_driver.c (remoteDomainGetCPUStats): Accommodate driver that omits tail of array. * src/libvirt.c (virDomainGetCPUStats): Document this. * src/qemu/qemu_driver.c (qemuDomainGetPercpuStats): Clear all unpopulated entries.
2012-03-07 04:36:53 +00:00
if (args->nparams && !(args->flags & VIR_TYPED_PARAM_STRING_OKAY)) {
size_t i;
rpc: allow truncated return for virDomainGetCPUStats The RPC code assumed that the array returned by the driver would be fully populated; that is, ncpus on entry resulted in ncpus * return value on exit. However, while we don't support holes in the middle of ncpus, we do want to permit the case of ncpus on entry being longer than the array returned by the driver (that is, it should be safe for the caller to pass ncpus=128 on entry, and the driver will stop populating the array when it hits max_id). Additionally, a successful return implies that the caller will then use virTypedParamArrayClear on the entire array; for this to not free uninitialized memory, the driver must ensure that all skipped entries are explicitly zeroed (the RPC driver did this, but not the qemu driver). There are now three cases: server 0.9.10 and client 0.9.10 or newer: No impact - there were no hypervisor drivers that supported cpu stats server 0.9.11 or newer and client 0.9.10: if the client calls with ncpus beyond the max, then the rpc call will fail on the client side and disconnect the client, but the server is no worse for the wear server 0.9.11 or newer and client 0.9.11: the server can return a truncated array and the client will do just fine I reproduced the problem by using a host with 2 CPUs, and doing: virsh cpu-stats $dom --start 1 --count 2 * daemon/remote.c (remoteDispatchDomainGetCPUStats): Allow driver to omit tail of array. * src/remote/remote_driver.c (remoteDomainGetCPUStats): Accommodate driver that omits tail of array. * src/libvirt.c (virDomainGetCPUStats): Document this. * src/qemu/qemu_driver.c (qemuDomainGetPercpuStats): Clear all unpopulated entries.
2012-03-07 04:36:53 +00:00
for (i = 0; i < percpu_len; i++) {
if (params[i].type == VIR_TYPED_PARAM_STRING)
ret->nparams--;
}
}
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, args->ncpus * args->nparams);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainGetDiskErrors(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_disk_errors_args *args,
remote_domain_get_disk_errors_ret *ret)
{
int rv = -1;
virDomainPtr dom = NULL;
virDomainDiskErrorPtr errors = NULL;
int len = 0;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (args->maxerrors > REMOTE_DOMAIN_DISK_ERRORS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("maxerrors too large"));
goto cleanup;
}
if (args->maxerrors &&
VIR_ALLOC_N(errors, args->maxerrors) < 0)
goto cleanup;
if ((len = virDomainGetDiskErrors(dom, errors,
args->maxerrors,
args->flags)) < 0)
goto cleanup;
ret->nerrors = len;
if (errors &&
remoteSerializeDomainDiskErrors(errors, len,
&ret->errors.errors_val,
&ret->errors.errors_len) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
if (errors && len > 0) {
size_t i;
for (i = 0; i < len; i++)
VIR_FREE(errors[i].disk);
}
VIR_FREE(errors);
return rv;
}
static int
remoteDispatchNodeGetMemoryParameters(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_get_memory_parameters_args *args,
remote_node_get_memory_parameters_ret *ret)
{
virTypedParameterPtr params = NULL;
int nparams = 0;
unsigned int flags;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
if (args->nparams > REMOTE_NODE_MEMORY_PARAMETERS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("nparams too large"));
goto cleanup;
}
if (args->nparams && VIR_ALLOC_N(params, args->nparams) < 0)
goto cleanup;
nparams = args->nparams;
if (virNodeGetMemoryParameters(priv->conn, params, &nparams, flags) < 0)
goto cleanup;
/* In this case, we need to send back the number of parameters
* supported
*/
if (args->nparams == 0) {
ret->nparams = nparams;
goto success;
}
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
args->flags) < 0)
goto cleanup;
success:
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
return rv;
}
static int
remoteDispatchNodeGetCPUMap(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_get_cpu_map_args *args,
remote_node_get_cpu_map_ret *ret)
{
unsigned char *cpumap = NULL;
unsigned int online = 0;
unsigned int flags;
int cpunum;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
flags = args->flags;
cpunum = virNodeGetCPUMap(priv->conn, args->need_map ? &cpumap : NULL,
args->need_online ? &online : NULL, flags);
if (cpunum < 0)
goto cleanup;
/* 'serialize' return cpumap */
if (args->need_map) {
ret->cpumap.cpumap_len = VIR_CPU_MAPLEN(cpunum);
ret->cpumap.cpumap_val = (char *) cpumap;
cpumap = NULL;
}
ret->online = online;
ret->ret = cpunum;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
VIR_FREE(cpumap);
return rv;
}
static int
lxcDispatchDomainOpenNamespace(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
lxc_domain_open_namespace_args *args)
{
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
int *fdlist = NULL;
int ret;
virDomainPtr dom = NULL;
size_t i;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
ret = virDomainLxcOpenNamespace(dom,
&fdlist,
args->flags);
if (ret < 0)
goto cleanup;
/* We shouldn't have received any from the client,
* but in case they're playing games with us, prevent
* a resource leak
*/
for (i = 0; i < msg->nfds; i++)
VIR_FORCE_CLOSE(msg->fds[i]);
VIR_FREE(msg->fds);
msg->nfds = 0;
msg->fds = fdlist;
msg->nfds = ret;
rv = 1;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
2013-01-30 09:49:28 +00:00
static int
remoteDispatchDomainGetJobStats(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_job_stats_args *args,
remote_domain_get_job_stats_ret *ret)
{
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetJobStats(dom, &ret->type, &params,
&nparams, args->flags) < 0)
goto cleanup;
if (nparams > REMOTE_DOMAIN_JOB_STATS_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many job stats '%d' for limit '%d'"),
nparams, REMOTE_DOMAIN_JOB_STATS_MAX);
goto cleanup;
}
if (virTypedParamsSerialize(params, nparams,
(virTypedParameterRemotePtr *) &ret->params.params_val,
&ret->params.params_len,
0) < 0)
2013-01-30 09:49:28 +00:00
goto cleanup;
rv = 0;
cleanup:
2013-01-30 09:49:28 +00:00
if (rv < 0)
virNetMessageSaveError(rerr);
virTypedParamsFree(params, nparams);
virObjectUnref(dom);
2013-01-30 09:49:28 +00:00
return rv;
}
static int
remoteDispatchDomainMigrateBegin3Params(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_begin3_params_args *args,
remote_domain_migrate_begin3_params_ret *ret)
{
char *xml = NULL;
virDomainPtr dom = NULL;
virTypedParameterPtr params = NULL;
int nparams = 0;
char *cookieout = NULL;
int cookieoutlen = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->params.params_len > REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many migration parameters '%d' for limit '%d'"),
args->params.params_len, REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX);
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virTypedParamsDeserialize((virTypedParameterRemotePtr) args->params.params_val,
args->params.params_len,
0, &params, &nparams) < 0)
goto cleanup;
if (!(xml = virDomainMigrateBegin3Params(dom, params, nparams,
&cookieout, &cookieoutlen,
args->flags)))
goto cleanup;
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
ret->xml = xml;
rv = 0;
cleanup:
virTypedParamsFree(params, nparams);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMigratePrepare3Params(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_prepare3_params_args *args,
remote_domain_migrate_prepare3_params_ret *ret)
{
virTypedParameterPtr params = NULL;
int nparams = 0;
char *cookieout = NULL;
int cookieoutlen = 0;
char **uri_out;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->params.params_len > REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many migration parameters '%d' for limit '%d'"),
args->params.params_len, REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX);
goto cleanup;
}
if (virTypedParamsDeserialize((virTypedParameterRemotePtr) args->params.params_val,
args->params.params_len,
0, &params, &nparams) < 0)
goto cleanup;
/* Wacky world of XDR ... */
if (VIR_ALLOC(uri_out) < 0)
goto cleanup;
if (virDomainMigratePrepare3Params(priv->conn, params, nparams,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
&cookieout, &cookieoutlen,
uri_out, args->flags) < 0)
goto cleanup;
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
ret->uri_out = !*uri_out ? NULL : uri_out;
rv = 0;
cleanup:
virTypedParamsFree(params, nparams);
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(uri_out);
}
return rv;
}
static int
remoteDispatchDomainMigratePrepareTunnel3Params(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg,
virNetMessageErrorPtr rerr,
remote_domain_migrate_prepare_tunnel3_params_args *args,
remote_domain_migrate_prepare_tunnel3_params_ret *ret)
{
virTypedParameterPtr params = NULL;
int nparams = 0;
char *cookieout = NULL;
int cookieoutlen = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virStreamPtr st = NULL;
daemonClientStreamPtr stream = NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->params.params_len > REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many migration parameters '%d' for limit '%d'"),
args->params.params_len, REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX);
goto cleanup;
}
if (virTypedParamsDeserialize((virTypedParameterRemotePtr) args->params.params_val,
args->params.params_len,
0, &params, &nparams) < 0)
goto cleanup;
if (!(st = virStreamNew(priv->conn, VIR_STREAM_NONBLOCK)) ||
!(stream = daemonCreateClientStream(client, st, remoteProgram,
&msg->header)))
goto cleanup;
if (virDomainMigratePrepareTunnel3Params(priv->conn, st, params, nparams,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
&cookieout, &cookieoutlen,
args->flags) < 0)
goto cleanup;
if (daemonAddClientStream(client, stream, false) < 0)
goto cleanup;
ret->cookie_out.cookie_out_val = cookieout;
ret->cookie_out.cookie_out_len = cookieoutlen;
rv = 0;
cleanup:
virTypedParamsFree(params, nparams);
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(cookieout);
if (stream) {
virStreamAbort(st);
daemonFreeClientStream(client, stream);
} else {
virObjectUnref(st);
}
}
return rv;
}
static int
remoteDispatchDomainMigratePerform3Params(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_perform3_params_args *args,
remote_domain_migrate_perform3_params_ret *ret)
{
virTypedParameterPtr params = NULL;
int nparams = 0;
virDomainPtr dom = NULL;
char *cookieout = NULL;
int cookieoutlen = 0;
char *dconnuri;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->params.params_len > REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many migration parameters '%d' for limit '%d'"),
args->params.params_len, REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX);
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virTypedParamsDeserialize((virTypedParameterRemotePtr) args->params.params_val,
args->params.params_len,
0, &params, &nparams) < 0)
goto cleanup;
dconnuri = args->dconnuri == NULL ? NULL : *args->dconnuri;
if (virDomainMigratePerform3Params(dom, dconnuri, params, nparams,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
&cookieout, &cookieoutlen,
args->flags) < 0)
goto cleanup;
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
rv = 0;
cleanup:
virTypedParamsFree(params, nparams);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMigrateFinish3Params(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_finish3_params_args *args,
remote_domain_migrate_finish3_params_ret *ret)
{
virTypedParameterPtr params = NULL;
int nparams = 0;
virDomainPtr dom = NULL;
char *cookieout = NULL;
int cookieoutlen = 0;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->params.params_len > REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many migration parameters '%d' for limit '%d'"),
args->params.params_len, REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX);
goto cleanup;
}
if (virTypedParamsDeserialize((virTypedParameterRemotePtr) args->params.params_val,
args->params.params_len,
0, &params, &nparams) < 0)
goto cleanup;
dom = virDomainMigrateFinish3Params(priv->conn, params, nparams,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
&cookieout, &cookieoutlen,
args->flags, args->cancelled);
if (!dom)
goto cleanup;
make_nonnull_domain(&ret->dom, dom);
ret->cookie_out.cookie_out_len = cookieoutlen;
ret->cookie_out.cookie_out_val = cookieout;
rv = 0;
cleanup:
virTypedParamsFree(params, nparams);
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(cookieout);
}
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchDomainMigrateConfirm3Params(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_migrate_confirm3_params_args *args)
{
virTypedParameterPtr params = NULL;
int nparams = 0;
virDomainPtr dom = NULL;
int rv = -1;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->params.params_len > REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many migration parameters '%d' for limit '%d'"),
args->params.params_len, REMOTE_DOMAIN_MIGRATE_PARAM_LIST_MAX);
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virTypedParamsDeserialize((virTypedParameterRemotePtr) args->params.params_val,
args->params.params_len,
0, &params, &nparams) < 0)
goto cleanup;
if (virDomainMigrateConfirm3Params(dom, params, nparams,
args->cookie_in.cookie_in_val,
args->cookie_in.cookie_in_len,
args->flags, args->cancelled) < 0)
goto cleanup;
rv = 0;
cleanup:
virTypedParamsFree(params, nparams);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchConnectGetCPUModelNames(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client ATTRIBUTE_UNUSED,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_connect_get_cpu_model_names_args *args,
remote_connect_get_cpu_model_names_ret *ret)
{
int len, rv = -1;
char **models = NULL;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
len = virConnectGetCPUModelNames(priv->conn, args->arch,
args->need_results ? &models : NULL,
args->flags);
if (len < 0)
goto cleanup;
if (len > REMOTE_CONNECT_CPU_MODELS_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many CPU models '%d' for limit '%d'"),
len, REMOTE_CONNECT_CPU_MODELS_MAX);
goto cleanup;
}
if (len && models) {
ret->models.models_val = models;
ret->models.models_len = len;
models = NULL;
} else {
ret->models.models_val = NULL;
ret->models.models_len = 0;
}
ret->ret = len;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virStringListFree(models);
return rv;
}
static int
remoteDispatchDomainCreateXMLWithFiles(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_create_xml_with_files_args *args,
remote_domain_create_xml_with_files_ret *ret)
{
int rv = -1;
virDomainPtr dom = NULL;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
int *files = NULL;
unsigned int nfiles = 0;
size_t i;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (VIR_ALLOC_N(files, msg->nfds) < 0)
goto cleanup;
for (i = 0; i < msg->nfds; i++) {
if ((files[i] = virNetMessageDupFD(msg, i)) < 0)
goto cleanup;
nfiles++;
}
if ((dom = virDomainCreateXMLWithFiles(priv->conn, args->xml_desc,
nfiles, files,
args->flags)) == NULL)
goto cleanup;
make_nonnull_domain(&ret->dom, dom);
rv = 0;
cleanup:
for (i = 0; i < nfiles; i++)
VIR_FORCE_CLOSE(files[i]);
VIR_FREE(files);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int remoteDispatchDomainCreateWithFiles(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_create_with_files_args *args,
remote_domain_create_with_files_ret *ret)
{
int rv = -1;
virDomainPtr dom = NULL;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
int *files = NULL;
unsigned int nfiles = 0;
size_t i;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (VIR_ALLOC_N(files, msg->nfds) < 0)
goto cleanup;
for (i = 0; i < msg->nfds; i++) {
if ((files[i] = virNetMessageDupFD(msg, i)) < 0)
goto cleanup;
nfiles++;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainCreateWithFiles(dom,
nfiles, files,
args->flags) < 0)
goto cleanup;
make_nonnull_domain(&ret->dom, dom);
rv = 0;
cleanup:
for (i = 0; i < nfiles; i++)
VIR_FORCE_CLOSE(files[i]);
VIR_FREE(files);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchConnectNetworkEventRegisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_network_event_register_any_args *args,
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
remote_connect_network_event_register_any_ret *ret)
{
int callbackID;
int rv = -1;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
virNetworkPtr net = NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
if (args->net &&
!(net = get_nonnull_network(priv->conn, *args->net)))
goto cleanup;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
if (args->eventID >= VIR_NETWORK_EVENT_ID_LAST || args->eventID < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unsupported network event ID %d"), args->eventID);
goto cleanup;
}
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; but since VIR_APPEND_ELEMENT clears 'callback' on
* success, we use 'ref' to save a copy of the pointer. */
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
callback->eventID = args->eventID;
callback->callbackID = -1;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->networkEventCallbacks,
priv->nnetworkEventCallbacks,
callback) < 0)
goto cleanup;
if ((callbackID = virConnectNetworkEventRegisterAny(priv->conn,
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
net,
args->eventID,
networkEventCallbacks[args->eventID],
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
ref,
remoteEventCallbackFree)) < 0) {
VIR_SHRINK_N(priv->networkEventCallbacks,
priv->nnetworkEventCallbacks, 1);
callback = ref;
goto cleanup;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
}
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
ref->callbackID = callbackID;
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
ret->callbackID = callbackID;
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(net);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectNetworkEventDeregisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
remote_connect_network_event_deregister_any_args *args)
{
int rv = -1;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
for (i = 0; i < priv->nnetworkEventCallbacks; i++) {
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
if (priv->networkEventCallbacks[i]->callbackID == args->callbackID)
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
break;
}
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
if (i == priv->nnetworkEventCallbacks) {
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR,
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
_("network event callback %d not registered"),
args->callbackID);
goto cleanup;
}
event: wire up RPC for server-side network event filtering We haven't had a release with network events yet, so we are free to fix the RPC so that it actually does what we want. Doing client-side filtering of per-network events is inefficient if a connection is only interested in events on a single network out of hundreds available on the server. But to do server-side per-network filtering, the server needs to know which network to filter on - so we need to pass an optional network over on registration. Furthermore, it is possible to have a client with both a global and per-network filter; in the existing code, the server sends only one event and the client replicates to both callbacks. But with server-side filtering, the server will send the event twice, so we need a way for the client to know which callbackID is sending an event, to ensure that the client can filter out events from a registration that does not match the callbackID from the server. Likewise, the existing style of deregistering by eventID alone is fine; but in the new style, we have to remember which callbackID to delete. This patch fixes the RPC wire definition to contain all the needed pieces of information, and hooks into the server and client side improvements of the previous patches, in order to switch over to full server-side filtering of network events. Also, since we fixed this in time, all released versions of libvirtd that support network events also support per-network filtering, so we can hard-code that assumption into network_event.c. Converting domain events to server-side filtering will require the introduction of new RPC numbers, as well as a server feature bit that the client can use to tell whether to use old-style (server only supports global events) or new-style (server supports filtered events), so that is deferred to a later set of patches. * src/conf/network_event.c (virNetworkEventStateRegisterClient): Assume server-side filtering. * src/remote/remote_protocol.x (remote_connect_network_event_register_any_args): Add network argument. (remote_connect_network_event_register_any_ret): Return callbackID instead of count. (remote_connect_network_event_deregister_any_args): Pass callbackID instead of eventID. (remote_connect_network_event_deregister_any_ret): Drop unused type. (remote_network_event_lifecycle_msg): Add callbackID. * daemon/remote.c (remoteDispatchConnectNetworkEventDeregisterAny): Drop unused arg, and deal with callbackID from client. (remoteRelayNetworkEventLifecycle): Pass callbackID. (remoteDispatchConnectNetworkEventRegisterAny): Likewise, and recognize non-NULL network. * src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny): Pass network, and track server side id. (remoteConnectNetworkEventDeregisterAny): Deregister by callback id. (remoteNetworkBuildEventLifecycle): Pass remote id to event queue. * src/remote_protocol-structs: Regenerate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-08 20:24:22 +00:00
if (virConnectNetworkEventDeregisterAny(priv->conn, args->callbackID) < 0)
goto cleanup;
event: track callbackID on daemon side of RPC Right now, the daemon side of RPC events is hard-coded to at most one callback per eventID. But when there are hundreds of domains or networks coupled and multiple conections, then sending every event to every connection that wants an event, even for the connections that only care about events for a particular object, is inefficient. In order to track more than one callback in the server, we need to store callbacks by more than just their eventID. This patch rearranges the daemon side to store network callbacks in a dynamic array, which can eventually be used for multiple callbacks of the same eventID, although actual behavior is unchanged without further patches to the RPC protocol. For ease of review, domain events are saved for a later patch, as they touch more code. While at it, fix a bug where a malicious client could send a negative eventID to cause network event registration to access outside of array bounds (thankfully not a CVE, since domain events were already doing the bounds check, and since network events have not been released). * daemon/libvirtd.h (daemonClientPrivate): Alter the tracking of network events. * daemon/remote.c (daemonClientEventCallback): New struct. (remoteEventCallbackFree): New function. (remoteClientInitHook, remoteRelayNetworkEventLifecycle) (remoteClientFreeFunc) (remoteDispatchConnectNetworkEventRegisterAny): Track network callbacks differently. (remoteDispatchConnectNetworkEventDeregisterAny): Enforce bounds. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-05 19:37:17 +00:00
VIR_DELETE_ELEMENT(priv->networkEventCallbacks, i,
priv->nnetworkEventCallbacks);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectStoragePoolEventRegisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_storage_pool_event_register_any_args *args,
remote_connect_storage_pool_event_register_any_ret *ret)
{
int callbackID;
int rv = -1;
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virStoragePoolPtr pool = NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
if (args->pool &&
!(pool = get_nonnull_storage_pool(priv->conn, *args->pool)))
goto cleanup;
if (args->eventID >= VIR_STORAGE_POOL_EVENT_ID_LAST || args->eventID < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unsupported storage pool event ID %d"), args->eventID);
goto cleanup;
}
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; but since VIR_APPEND_ELEMENT clears 'callback' on
* success, we use 'ref' to save a copy of the pointer. */
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
callback->eventID = args->eventID;
callback->callbackID = -1;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->storageEventCallbacks,
priv->nstorageEventCallbacks,
callback) < 0)
goto cleanup;
if ((callbackID = virConnectStoragePoolEventRegisterAny(priv->conn,
pool,
args->eventID,
storageEventCallbacks[args->eventID],
ref,
remoteEventCallbackFree)) < 0) {
VIR_SHRINK_N(priv->storageEventCallbacks,
priv->nstorageEventCallbacks, 1);
callback = ref;
goto cleanup;
}
ref->callbackID = callbackID;
ret->callbackID = callbackID;
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(pool);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectStoragePoolEventDeregisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_storage_pool_event_deregister_any_args *args)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
for (i = 0; i < priv->nstorageEventCallbacks; i++) {
if (priv->storageEventCallbacks[i]->callbackID == args->callbackID)
break;
}
if (i == priv->nstorageEventCallbacks) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("storage pool event callback %d not registered"),
args->callbackID);
goto cleanup;
}
if (virConnectStoragePoolEventDeregisterAny(priv->conn, args->callbackID) < 0)
goto cleanup;
VIR_DELETE_ELEMENT(priv->storageEventCallbacks, i,
priv->nstorageEventCallbacks);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectNodeDeviceEventRegisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_node_device_event_register_any_args *args,
remote_connect_node_device_event_register_any_ret *ret)
{
int callbackID;
int rv = -1;
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virNodeDevicePtr dev = NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
if (args->dev &&
!(dev = get_nonnull_node_device(priv->conn, *args->dev)))
goto cleanup;
if (args->eventID >= VIR_NODE_DEVICE_EVENT_ID_LAST || args->eventID < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unsupported node device event ID %d"), args->eventID);
goto cleanup;
}
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; but since VIR_APPEND_ELEMENT clears 'callback' on
* success, we use 'ref' to save a copy of the pointer. */
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
callback->eventID = args->eventID;
callback->callbackID = -1;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->nodeDeviceEventCallbacks,
priv->nnodeDeviceEventCallbacks,
callback) < 0)
goto cleanup;
if ((callbackID = virConnectNodeDeviceEventRegisterAny(priv->conn,
dev,
args->eventID,
nodeDeviceEventCallbacks[args->eventID],
ref,
remoteEventCallbackFree)) < 0) {
VIR_SHRINK_N(priv->nodeDeviceEventCallbacks,
priv->nnodeDeviceEventCallbacks, 1);
callback = ref;
goto cleanup;
}
ref->callbackID = callbackID;
ret->callbackID = callbackID;
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dev);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectNodeDeviceEventDeregisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_node_device_event_deregister_any_args *args)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
for (i = 0; i < priv->nnodeDeviceEventCallbacks; i++) {
if (priv->nodeDeviceEventCallbacks[i]->callbackID == args->callbackID)
break;
}
if (i == priv->nnodeDeviceEventCallbacks) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("node device event callback %d not registered"),
args->callbackID);
goto cleanup;
}
if (virConnectNodeDeviceEventDeregisterAny(priv->conn, args->callbackID) < 0)
goto cleanup;
VIR_DELETE_ELEMENT(priv->nodeDeviceEventCallbacks, i,
priv->nnodeDeviceEventCallbacks);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectSecretEventRegisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_secret_event_register_any_args *args,
remote_connect_secret_event_register_any_ret *ret)
{
int callbackID;
int rv = -1;
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virSecretPtr secret = NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
if (args->secret &&
!(secret = get_nonnull_secret(priv->conn, *args->secret)))
goto cleanup;
if (args->eventID >= VIR_SECRET_EVENT_ID_LAST || args->eventID < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unsupported secret event ID %d"), args->eventID);
goto cleanup;
}
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; but since VIR_APPEND_ELEMENT clears 'callback' on
* success, we use 'ref' to save a copy of the pointer. */
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
callback->eventID = args->eventID;
callback->callbackID = -1;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->secretEventCallbacks,
priv->nsecretEventCallbacks,
callback) < 0)
goto cleanup;
if ((callbackID = virConnectSecretEventRegisterAny(priv->conn,
secret,
args->eventID,
secretEventCallbacks[args->eventID],
ref,
remoteEventCallbackFree)) < 0) {
VIR_SHRINK_N(priv->secretEventCallbacks,
priv->nsecretEventCallbacks, 1);
callback = ref;
goto cleanup;
}
ref->callbackID = callbackID;
ret->callbackID = callbackID;
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(secret);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchConnectSecretEventDeregisterAny(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
remote_connect_secret_event_deregister_any_args *args)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
for (i = 0; i < priv->nsecretEventCallbacks; i++) {
if (priv->secretEventCallbacks[i]->callbackID == args->callbackID)
break;
}
if (i == priv->nsecretEventCallbacks) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("node device event callback %d not registered"),
args->callbackID);
goto cleanup;
}
if (virConnectSecretEventDeregisterAny(priv->conn, args->callbackID) < 0)
goto cleanup;
VIR_DELETE_ELEMENT(priv->secretEventCallbacks, i,
priv->nsecretEventCallbacks);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
qemuDispatchConnectDomainMonitorEventRegister(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
qemu_connect_domain_monitor_event_register_args *args,
qemu_connect_domain_monitor_event_register_ret *ret)
{
int callbackID;
int rv = -1;
daemonClientEventCallbackPtr callback = NULL;
daemonClientEventCallbackPtr ref;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
virDomainPtr dom = NULL;
const char *event = args->event ? *args->event : NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
if (args->dom &&
!(dom = get_nonnull_domain(priv->conn, *args->dom)))
goto cleanup;
/* If we call register first, we could append a complete callback
* to our array, but on OOM append failure, we'd have to then hope
* deregister works to undo our register. So instead we append an
* incomplete callback to our array, then register, then fix up
* our callback; but since VIR_APPEND_ELEMENT clears 'callback' on
* success, we use 'ref' to save a copy of the pointer. */
if (VIR_ALLOC(callback) < 0)
goto cleanup;
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
callback->client = virObjectRef(client);
callback->callbackID = -1;
ref = callback;
if (VIR_APPEND_ELEMENT(priv->qemuEventCallbacks,
priv->nqemuEventCallbacks,
callback) < 0)
goto cleanup;
if ((callbackID = virConnectDomainQemuMonitorEventRegister(priv->conn,
dom,
event,
remoteRelayDomainQemuMonitorEvent,
ref,
remoteEventCallbackFree,
args->flags)) < 0) {
VIR_SHRINK_N(priv->qemuEventCallbacks,
priv->nqemuEventCallbacks, 1);
callback = ref;
goto cleanup;
}
ref->callbackID = callbackID;
ret->callbackID = callbackID;
rv = 0;
cleanup:
remote: Fix possible use-after-free when sending event message Based upon an idea and some research by Wang King <king.wang@huawei.com> and xinhua.Cao <caoxinhua@huawei.com>. Since we're assigning the 'client' to our callback event lookaside list, it's imperative that we grab a reference to the object; otherwise, when the object is unref'd during virNetServerProcessClients when it's determined that the virNetServerClientIsClosed and the memory is free'd before perhaps the object event state callbacks are run. When a virObjectLock() is run, before sending the message the following trace occurs; #0 0x00007fda223d66d8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fda24c81b40) at util/virobject.c:169 #1 0x00007fda223d6a1e in virObjectIsClass (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>) at util/virobject.c:365 #2 0x00007fda223d6a44 in virObjectLock (anyobj=0x7fd9e575b400) at util/virobject.c:317 #3 0x00007fda22507f71 in virNetServerClientSendMessage (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90) at rpc/virnetserverclient.c:1422 #4 0x00007fda230d714d in remoteDispatchObjectEventSend (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348, proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>, data=0x7ffc3857fdb0) at remote.c:3803 #5 0x00007fda230dd71b in remoteRelayDomainEventTunable (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0, nparams=1,opaque=0x7fd9e6c99e00) at remote.c:1033 #6 0x00007fda224484cb in virDomainEventDispatchDefaultFunc (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610 <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00) at conf/domain_event.c:1910 #7 0x00007fda22446871 in virObjectEventStateDispatchCallbacks (callbacks=<optimized out>, callbacks=<optimized out>, event=0x7fda2736ea00,state=0x7fda24ca3960) at conf/object_event.c:722 #8 virObjectEventStateQueueDispatch (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960) at conf/object_event.c:736 #9 virObjectEventStateFlush (state=0x7fda24ca3960) at conf/object_event.c:814 #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960) at conf/object_event.c:560 #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts () at util/vireventpoll.c:458 #12 virEventPollRunOnce () at util/vireventpoll.c:654 #13 0x00007fda223ad1d2 in virEventRunDefaultImpl () at util/virevent.c:314 #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0) at rpc/virnetdaemon.c:818 #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>) at libvirtd.c:1623 Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-03-27 16:47:37 +00:00
remoteEventCallbackFree(callback);
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
virMutexUnlock(&priv->lock);
return rv;
}
static int
qemuDispatchConnectDomainMonitorEventDeregister(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr ATTRIBUTE_UNUSED,
qemu_connect_domain_monitor_event_deregister_args *args)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
virMutexLock(&priv->lock);
for (i = 0; i < priv->nqemuEventCallbacks; i++) {
if (priv->qemuEventCallbacks[i]->callbackID == args->callbackID)
break;
}
if (i == priv->nqemuEventCallbacks) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("qemu monitor event callback %d not registered"),
args->callbackID);
goto cleanup;
}
if (virConnectDomainQemuMonitorEventDeregister(priv->conn,
args->callbackID) < 0)
goto cleanup;
VIR_DELETE_ELEMENT(priv->qemuEventCallbacks, i,
priv->nqemuEventCallbacks);
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virMutexUnlock(&priv->lock);
return rv;
}
static int
remoteDispatchDomainGetTime(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_time_args *args,
remote_domain_get_time_ret *ret)
{
int rv = -1;
virDomainPtr dom = NULL;
long long seconds;
unsigned int nseconds;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if (virDomainGetTime(dom, &seconds, &nseconds, args->flags) < 0)
goto cleanup;
ret->seconds = seconds;
ret->nseconds = nseconds;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
return rv;
}
static int
remoteDispatchNodeGetFreePages(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_get_free_pages_args *args,
remote_node_get_free_pages_ret *ret)
{
int rv = -1;
int len;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->pages.pages_len * args->cellCount > REMOTE_NODE_MAX_CELLS) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("the result won't fit into REMOTE_NODE_MAX_CELLS"));
goto cleanup;
}
/* Allocate return buffer. */
if (VIR_ALLOC_N(ret->counts.counts_val,
args->pages.pages_len * args->cellCount) < 0)
goto cleanup;
if ((len = virNodeGetFreePages(priv->conn,
args->pages.pages_len,
args->pages.pages_val,
args->startCell,
args->cellCount,
(unsigned long long *) ret->counts.counts_val,
args->flags)) <= 0)
goto cleanup;
ret->counts.counts_len = len;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
VIR_FREE(ret->counts.counts_val);
}
return rv;
}
/* Copy contents of virNetworkDHCPLeasePtr to remote_network_dhcp_lease */
static int
remoteSerializeDHCPLease(remote_network_dhcp_lease *lease_dst, virNetworkDHCPLeasePtr lease_src)
{
char **mac_tmp = NULL;
char **iaid_tmp = NULL;
char **hostname_tmp = NULL;
char **clientid_tmp = NULL;
lease_dst->expirytime = lease_src->expirytime;
lease_dst->type = lease_src->type;
lease_dst->prefix = lease_src->prefix;
if (VIR_STRDUP(lease_dst->iface, lease_src->iface) < 0 ||
VIR_STRDUP(lease_dst->ipaddr, lease_src->ipaddr) < 0)
goto error;
if (lease_src->mac) {
if (VIR_ALLOC(mac_tmp) < 0 ||
VIR_STRDUP(*mac_tmp, lease_src->mac) < 0)
goto error;
}
if (lease_src->iaid) {
if (VIR_ALLOC(iaid_tmp) < 0 ||
VIR_STRDUP(*iaid_tmp, lease_src->iaid) < 0)
goto error;
}
if (lease_src->hostname) {
if (VIR_ALLOC(hostname_tmp) < 0 ||
VIR_STRDUP(*hostname_tmp, lease_src->hostname) < 0)
goto error;
}
if (lease_src->clientid) {
if (VIR_ALLOC(clientid_tmp) < 0 ||
VIR_STRDUP(*clientid_tmp, lease_src->clientid) < 0)
goto error;
}
lease_dst->mac = mac_tmp;
lease_dst->iaid = iaid_tmp;
lease_dst->hostname = hostname_tmp;
lease_dst->clientid = clientid_tmp;
return 0;
error:
if (mac_tmp)
VIR_FREE(*mac_tmp);
if (iaid_tmp)
VIR_FREE(*iaid_tmp);
if (hostname_tmp)
VIR_FREE(*hostname_tmp);
if (clientid_tmp)
VIR_FREE(*clientid_tmp);
VIR_FREE(mac_tmp);
VIR_FREE(iaid_tmp);
VIR_FREE(hostname_tmp);
VIR_FREE(clientid_tmp);
VIR_FREE(lease_dst->ipaddr);
VIR_FREE(lease_dst->iface);
return -1;
}
static int
remoteDispatchNetworkGetDHCPLeases(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_network_get_dhcp_leases_args *args,
remote_network_get_dhcp_leases_ret *ret)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client);
virNetworkDHCPLeasePtr *leases = NULL;
virNetworkPtr net = NULL;
int nleases = 0;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(net = get_nonnull_network(priv->conn, args->net)))
goto cleanup;
if ((nleases = virNetworkGetDHCPLeases(net,
args->mac ? *args->mac : NULL,
args->need_results ? &leases : NULL,
args->flags)) < 0)
goto cleanup;
if (nleases > REMOTE_NETWORK_DHCP_LEASES_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Number of leases is %d, which exceeds max limit: %d"),
nleases, REMOTE_NETWORK_DHCP_LEASES_MAX);
goto cleanup;
}
if (leases && nleases) {
if (VIR_ALLOC_N(ret->leases.leases_val, nleases) < 0)
goto cleanup;
ret->leases.leases_len = nleases;
for (i = 0; i < nleases; i++) {
if (remoteSerializeDHCPLease(ret->leases.leases_val + i, leases[i]) < 0)
goto cleanup;
}
} else {
ret->leases.leases_len = 0;
ret->leases.leases_val = NULL;
}
ret->ret = nleases;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
daemon: avoid memleak when ListAll returns nothing Commit 4f25146 (v1.2.8) managed to silence Coverity, but at the cost of a memory leak detected by valgrind: ==24129== 40 bytes in 5 blocks are definitely lost in loss record 355 of 637 ==24129== at 0x4A08B1C: realloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==24129== by 0x5084B8E: virReallocN (viralloc.c:245) ==24129== by 0x514D5AA: virDomainObjListExport (domain_conf.c:22200) ==24129== by 0x201227DB: qemuConnectListAllDomains (qemu_driver.c:18042) ==24129== by 0x51CC1B6: virConnectListAllDomains (libvirt-domain.c:6797) ==24129== by 0x14173D: remoteDispatchConnectListAllDomains (remote.c:1580) ==24129== by 0x121BE1: remoteDispatchConnectListAllDomainsHelper (remote_dispatch.h:1072) In short, every time a client calls a ListAll variant and asks for the resulting list, but there are 0 elements to return, we end up leaking the 1-entry array that holds the NULL terminator. What's worse, a read-only client can access these functions in a tight loop to cause libvirtd to eventually run out of memory; and this can be considered a denial of service attack against more privileged clients. Thankfully, the leak is so small (8 bytes per call) that you would already have some other denial of service with any guest calling the API that frequently, so an out-of-memory crash is unlikely enough that this did not warrant a CVE. * daemon/remote.c (remoteDispatchConnectListAllDomains) (remoteDispatchDomainListAllSnapshots) (remoteDispatchDomainSnapshotListAllChildren) (remoteDispatchConnectListAllStoragePools) (remoteDispatchStoragePoolListAllVolumes) (remoteDispatchConnectListAllNetworks) (remoteDispatchConnectListAllInterfaces) (remoteDispatchConnectListAllNodeDevices) (remoteDispatchConnectListAllNWFilters) (remoteDispatchConnectListAllSecrets) (remoteDispatchNetworkGetDHCPLeases): Plug leak. Signed-off-by: Eric Blake <eblake@redhat.com>
2015-03-13 23:01:43 +00:00
if (leases && nleases > 0)
for (i = 0; i < nleases; i++)
virNetworkDHCPLeaseFree(leases[i]);
daemon: avoid memleak when ListAll returns nothing Commit 4f25146 (v1.2.8) managed to silence Coverity, but at the cost of a memory leak detected by valgrind: ==24129== 40 bytes in 5 blocks are definitely lost in loss record 355 of 637 ==24129== at 0x4A08B1C: realloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==24129== by 0x5084B8E: virReallocN (viralloc.c:245) ==24129== by 0x514D5AA: virDomainObjListExport (domain_conf.c:22200) ==24129== by 0x201227DB: qemuConnectListAllDomains (qemu_driver.c:18042) ==24129== by 0x51CC1B6: virConnectListAllDomains (libvirt-domain.c:6797) ==24129== by 0x14173D: remoteDispatchConnectListAllDomains (remote.c:1580) ==24129== by 0x121BE1: remoteDispatchConnectListAllDomainsHelper (remote_dispatch.h:1072) In short, every time a client calls a ListAll variant and asks for the resulting list, but there are 0 elements to return, we end up leaking the 1-entry array that holds the NULL terminator. What's worse, a read-only client can access these functions in a tight loop to cause libvirtd to eventually run out of memory; and this can be considered a denial of service attack against more privileged clients. Thankfully, the leak is so small (8 bytes per call) that you would already have some other denial of service with any guest calling the API that frequently, so an out-of-memory crash is unlikely enough that this did not warrant a CVE. * daemon/remote.c (remoteDispatchConnectListAllDomains) (remoteDispatchDomainListAllSnapshots) (remoteDispatchDomainSnapshotListAllChildren) (remoteDispatchConnectListAllStoragePools) (remoteDispatchStoragePoolListAllVolumes) (remoteDispatchConnectListAllNetworks) (remoteDispatchConnectListAllInterfaces) (remoteDispatchConnectListAllNodeDevices) (remoteDispatchConnectListAllNWFilters) (remoteDispatchConnectListAllSecrets) (remoteDispatchNetworkGetDHCPLeases): Plug leak. Signed-off-by: Eric Blake <eblake@redhat.com>
2015-03-13 23:01:43 +00:00
VIR_FREE(leases);
virObjectUnref(net);
return rv;
}
static int
remoteDispatchConnectGetAllDomainStats(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_connect_get_all_domain_stats_args *args,
remote_connect_get_all_domain_stats_ret *ret)
{
int rv = -1;
size_t i;
struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client);
virDomainStatsRecordPtr *retStats = NULL;
int nrecords = 0;
virDomainPtr *doms = NULL;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (args->doms.doms_len) {
if (VIR_ALLOC_N(doms, args->doms.doms_len + 1) < 0)
goto cleanup;
for (i = 0; i < args->doms.doms_len; i++) {
if (!(doms[i] = get_nonnull_domain(priv->conn, args->doms.doms_val[i])))
goto cleanup;
}
if ((nrecords = virDomainListGetStats(doms,
args->stats,
&retStats,
args->flags)) < 0)
goto cleanup;
} else {
if ((nrecords = virConnectGetAllDomainStats(priv->conn,
args->stats,
&retStats,
args->flags)) < 0)
goto cleanup;
}
if (nrecords > REMOTE_CONNECT_GET_ALL_DOMAIN_STATS_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Number of domain stats records is %d, "
"which exceeds max limit: %d"),
nrecords, REMOTE_CONNECT_GET_ALL_DOMAIN_STATS_MAX);
goto cleanup;
}
if (nrecords) {
if (VIR_ALLOC_N(ret->retStats.retStats_val, nrecords) < 0)
goto cleanup;
ret->retStats.retStats_len = nrecords;
for (i = 0; i < nrecords; i++) {
remote_domain_stats_record *dst = ret->retStats.retStats_val + i;
make_nonnull_domain(&dst->dom, retStats[i]->dom);
if (virTypedParamsSerialize(retStats[i]->params,
retStats[i]->nparams,
(virTypedParameterRemotePtr *) &dst->params.params_val,
&dst->params.params_len,
VIR_TYPED_PARAM_STRING_OKAY) < 0)
goto cleanup;
}
} else {
ret->retStats.retStats_len = 0;
ret->retStats.retStats_val = NULL;
}
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virDomainStatsRecordListFree(retStats);
virObjectListFree(doms);
return rv;
}
static int
remoteDispatchNodeAllocPages(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_node_alloc_pages_args *args,
remote_node_alloc_pages_ret *ret)
{
int rv = -1;
int len;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if ((len = virNodeAllocPages(priv->conn,
args->pageSizes.pageSizes_len,
args->pageSizes.pageSizes_val,
(unsigned long long *) args->pageCounts.pageCounts_val,
args->startCell,
args->cellCount,
args->flags)) < 0)
goto cleanup;
ret->ret = len;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
return rv;
}
static int
remoteDispatchDomainGetFSInfo(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_get_fsinfo_args *args,
remote_domain_get_fsinfo_ret *ret)
{
int rv = -1;
size_t i, j;
struct daemonClientPrivate *priv = virNetServerClientGetPrivateData(client);
virDomainFSInfoPtr *info = NULL;
virDomainPtr dom = NULL;
remote_domain_fsinfo *dst;
int ninfo = 0;
size_t ndisk;
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if ((ninfo = virDomainGetFSInfo(dom, &info, args->flags)) < 0)
goto cleanup;
if (ninfo > REMOTE_DOMAIN_FSINFO_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many mountpoints in fsinfo: %d for limit %d"),
ninfo, REMOTE_DOMAIN_FSINFO_MAX);
goto cleanup;
}
if (ninfo) {
if (VIR_ALLOC_N(ret->info.info_val, ninfo) < 0)
goto cleanup;
ret->info.info_len = ninfo;
for (i = 0; i < ninfo; i++) {
dst = &ret->info.info_val[i];
if (VIR_STRDUP(dst->mountpoint, info[i]->mountpoint) < 0)
goto cleanup;
if (VIR_STRDUP(dst->name, info[i]->name) < 0)
goto cleanup;
if (VIR_STRDUP(dst->fstype, info[i]->fstype) < 0)
goto cleanup;
ndisk = info[i]->ndevAlias;
if (ndisk > REMOTE_DOMAIN_FSINFO_DISKS_MAX) {
virReportError(VIR_ERR_RPC,
_("Too many disks in fsinfo: %zd for limit %d"),
ndisk, REMOTE_DOMAIN_FSINFO_DISKS_MAX);
goto cleanup;
}
if (ndisk > 0) {
if (VIR_ALLOC_N(dst->dev_aliases.dev_aliases_val, ndisk) < 0)
goto cleanup;
for (j = 0; j < ndisk; j++) {
if (VIR_STRDUP(dst->dev_aliases.dev_aliases_val[j],
info[i]->devAlias[j]) < 0)
goto cleanup;
}
dst->dev_aliases.dev_aliases_len = ndisk;
} else {
dst->dev_aliases.dev_aliases_val = NULL;
dst->dev_aliases.dev_aliases_len = 0;
}
}
} else {
ret->info.info_len = 0;
ret->info.info_val = NULL;
}
ret->ret = ninfo;
rv = 0;
cleanup:
if (rv < 0) {
virNetMessageSaveError(rerr);
if (ret->info.info_val && ninfo > 0) {
for (i = 0; i < ninfo; i++) {
dst = &ret->info.info_val[i];
VIR_FREE(dst->mountpoint);
if (dst->dev_aliases.dev_aliases_val) {
for (j = 0; j < dst->dev_aliases.dev_aliases_len; j++)
VIR_FREE(dst->dev_aliases.dev_aliases_val[j]);
VIR_FREE(dst->dev_aliases.dev_aliases_val);
}
}
VIR_FREE(ret->info.info_val);
}
}
virObjectUnref(dom);
if (ninfo >= 0)
for (i = 0; i < ninfo; i++)
virDomainFSInfoFree(info[i]);
VIR_FREE(info);
return rv;
}
static int
remoteSerializeDomainInterface(virDomainInterfacePtr *ifaces,
unsigned int ifaces_count,
remote_domain_interface_addresses_ret *ret)
{
size_t i, j;
if (ifaces_count > REMOTE_DOMAIN_INTERFACE_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Number of interfaces, %d exceeds the max limit: %d"),
ifaces_count, REMOTE_DOMAIN_INTERFACE_MAX);
return -1;
}
if (VIR_ALLOC_N(ret->ifaces.ifaces_val, ifaces_count) < 0)
return -1;
ret->ifaces.ifaces_len = ifaces_count;
for (i = 0; i < ifaces_count; i++) {
virDomainInterfacePtr iface = ifaces[i];
remote_domain_interface *iface_ret = &(ret->ifaces.ifaces_val[i]);
if ((VIR_STRDUP(iface_ret->name, iface->name)) < 0)
goto cleanup;
if (iface->hwaddr &&
(VIR_ALLOC(iface_ret->hwaddr) < 0 ||
VIR_STRDUP(*iface_ret->hwaddr, iface->hwaddr) < 0))
goto cleanup;
if (iface->naddrs > REMOTE_DOMAIN_IP_ADDR_MAX) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Number of interfaces, %d exceeds the max limit: %d"),
iface->naddrs, REMOTE_DOMAIN_IP_ADDR_MAX);
goto cleanup;
}
if (VIR_ALLOC_N(iface_ret->addrs.addrs_val,
iface->naddrs) < 0)
goto cleanup;
iface_ret->addrs.addrs_len = iface->naddrs;
for (j = 0; j < iface->naddrs; j++) {
virDomainIPAddressPtr ip_addr = &(iface->addrs[j]);
remote_domain_ip_addr *ip_addr_ret =
&(iface_ret->addrs.addrs_val[j]);
if (VIR_STRDUP(ip_addr_ret->addr, ip_addr->addr) < 0)
goto cleanup;
ip_addr_ret->prefix = ip_addr->prefix;
ip_addr_ret->type = ip_addr->type;
}
}
return 0;
cleanup:
if (ret->ifaces.ifaces_val) {
for (i = 0; i < ifaces_count; i++) {
remote_domain_interface *iface_ret = &(ret->ifaces.ifaces_val[i]);
VIR_FREE(iface_ret->name);
if (iface_ret->hwaddr) {
VIR_FREE(*iface_ret->hwaddr);
VIR_FREE(iface_ret->hwaddr);
}
for (j = 0; j < iface_ret->addrs.addrs_len; j++) {
remote_domain_ip_addr *ip_addr =
&(iface_ret->addrs.addrs_val[j]);
VIR_FREE(ip_addr->addr);
}
}
VIR_FREE(ret->ifaces.ifaces_val);
}
return -1;
}
static int
remoteDispatchDomainInterfaceAddresses(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_domain_interface_addresses_args *args,
remote_domain_interface_addresses_ret *ret)
{
size_t i;
int rv = -1;
virDomainPtr dom = NULL;
virDomainInterfacePtr *ifaces = NULL;
int ifaces_count = 0;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(dom = get_nonnull_domain(priv->conn, args->dom)))
goto cleanup;
if ((ifaces_count = virDomainInterfaceAddresses(dom, &ifaces, args->source, args->flags)) < 0)
goto cleanup;
if (remoteSerializeDomainInterface(ifaces, ifaces_count, ret) < 0)
goto cleanup;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(dom);
if (ifaces && ifaces_count > 0) {
for (i = 0; i < ifaces_count; i++)
virDomainInterfaceFree(ifaces[i]);
}
VIR_FREE(ifaces);
return rv;
}
static int
remoteDispatchStorageVolGetInfoFlags(virNetServerPtr server ATTRIBUTE_UNUSED,
virNetServerClientPtr client,
virNetMessagePtr msg ATTRIBUTE_UNUSED,
virNetMessageErrorPtr rerr,
remote_storage_vol_get_info_flags_args *args,
remote_storage_vol_get_info_flags_ret *ret)
{
int rv = -1;
virStorageVolPtr vol = NULL;
virStorageVolInfo tmp;
struct daemonClientPrivate *priv =
virNetServerClientGetPrivateData(client);
if (!priv->conn) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("connection not open"));
goto cleanup;
}
if (!(vol = get_nonnull_storage_vol(priv->conn, args->vol)))
goto cleanup;
if (virStorageVolGetInfoFlags(vol, &tmp, args->flags) < 0)
goto cleanup;
ret->type = tmp.type;
ret->capacity = tmp.capacity;
ret->allocation = tmp.allocation;
rv = 0;
cleanup:
if (rv < 0)
virNetMessageSaveError(rerr);
virObjectUnref(vol);
return rv;
}
/*----- Helpers. -----*/
/* get_nonnull_domain and get_nonnull_network turn an on-wire
* (name, uuid) pair into virDomainPtr or virNetworkPtr object.
* virDomainPtr or virNetworkPtr cannot be NULL.
*
* NB. If these return NULL then the caller must return an error.
*/
static virDomainPtr
get_nonnull_domain(virConnectPtr conn, remote_nonnull_domain domain)
{
/* Should we believe the domain.id sent by the client? Maybe
* this should be a check rather than an assignment? XXX
*/
return virGetDomain(conn, domain.name, BAD_CAST domain.uuid, domain.id);
}
static virNetworkPtr
get_nonnull_network(virConnectPtr conn, remote_nonnull_network network)
{
return virGetNetwork(conn, network.name, BAD_CAST network.uuid);
}
static virInterfacePtr
get_nonnull_interface(virConnectPtr conn, remote_nonnull_interface iface)
{
return virGetInterface(conn, iface.name, iface.mac);
}
static virStoragePoolPtr
get_nonnull_storage_pool(virConnectPtr conn, remote_nonnull_storage_pool pool)
{
return virGetStoragePool(conn, pool.name, BAD_CAST pool.uuid,
NULL, NULL);
}
static virStorageVolPtr
get_nonnull_storage_vol(virConnectPtr conn, remote_nonnull_storage_vol vol)
{
virStorageVolPtr ret;
ret = virGetStorageVol(conn, vol.pool, vol.name, vol.key,
NULL, NULL);
return ret;
}
static virSecretPtr
get_nonnull_secret(virConnectPtr conn, remote_nonnull_secret secret)
{
return virGetSecret(conn, BAD_CAST secret.uuid, secret.usageType, secret.usageID);
}
static virNWFilterPtr
get_nonnull_nwfilter(virConnectPtr conn, remote_nonnull_nwfilter nwfilter)
{
return virGetNWFilter(conn, nwfilter.name, BAD_CAST nwfilter.uuid);
}
static virDomainSnapshotPtr
get_nonnull_domain_snapshot(virDomainPtr dom, remote_nonnull_domain_snapshot snapshot)
{
return virGetDomainSnapshot(dom, snapshot.name);
}
static virNodeDevicePtr
get_nonnull_node_device(virConnectPtr conn, remote_nonnull_node_device dev)
{
return virGetNodeDevice(conn, dev.name);
}
/* Make remote_nonnull_domain and remote_nonnull_network. */
static void
make_nonnull_domain(remote_nonnull_domain *dom_dst, virDomainPtr dom_src)
{
dom_dst->id = dom_src->id;
ignore_value(VIR_STRDUP_QUIET(dom_dst->name, dom_src->name));
memcpy(dom_dst->uuid, dom_src->uuid, VIR_UUID_BUFLEN);
}
static void
make_nonnull_network(remote_nonnull_network *net_dst, virNetworkPtr net_src)
{
ignore_value(VIR_STRDUP_QUIET(net_dst->name, net_src->name));
memcpy(net_dst->uuid, net_src->uuid, VIR_UUID_BUFLEN);
}
static void
make_nonnull_interface(remote_nonnull_interface *interface_dst,
virInterfacePtr interface_src)
{
ignore_value(VIR_STRDUP_QUIET(interface_dst->name, interface_src->name));
ignore_value(VIR_STRDUP_QUIET(interface_dst->mac, interface_src->mac));
}
static void
make_nonnull_storage_pool(remote_nonnull_storage_pool *pool_dst, virStoragePoolPtr pool_src)
{
ignore_value(VIR_STRDUP_QUIET(pool_dst->name, pool_src->name));
memcpy(pool_dst->uuid, pool_src->uuid, VIR_UUID_BUFLEN);
}
static void
make_nonnull_storage_vol(remote_nonnull_storage_vol *vol_dst, virStorageVolPtr vol_src)
{
ignore_value(VIR_STRDUP_QUIET(vol_dst->pool, vol_src->pool));
ignore_value(VIR_STRDUP_QUIET(vol_dst->name, vol_src->name));
ignore_value(VIR_STRDUP_QUIET(vol_dst->key, vol_src->key));
}
static void
make_nonnull_node_device(remote_nonnull_node_device *dev_dst, virNodeDevicePtr dev_src)
{
ignore_value(VIR_STRDUP_QUIET(dev_dst->name, dev_src->name));
}
static void
make_nonnull_secret(remote_nonnull_secret *secret_dst, virSecretPtr secret_src)
{
memcpy(secret_dst->uuid, secret_src->uuid, VIR_UUID_BUFLEN);
secret_dst->usageType = secret_src->usageType;
ignore_value(VIR_STRDUP_QUIET(secret_dst->usageID, secret_src->usageID));
}
static void
make_nonnull_nwfilter(remote_nonnull_nwfilter *nwfilter_dst, virNWFilterPtr nwfilter_src)
{
ignore_value(VIR_STRDUP_QUIET(nwfilter_dst->name, nwfilter_src->name));
memcpy(nwfilter_dst->uuid, nwfilter_src->uuid, VIR_UUID_BUFLEN);
}
static void
make_nonnull_domain_snapshot(remote_nonnull_domain_snapshot *snapshot_dst, virDomainSnapshotPtr snapshot_src)
{
ignore_value(VIR_STRDUP_QUIET(snapshot_dst->name, snapshot_src->name));
make_nonnull_domain(&snapshot_dst->dom, snapshot_src->domain);
}
static int
remoteSerializeDomainDiskErrors(virDomainDiskErrorPtr errors,
int nerrors,
remote_domain_disk_error **ret_errors_val,
u_int *ret_errors_len)
{
remote_domain_disk_error *val = NULL;
size_t i = 0;
if (VIR_ALLOC_N(val, nerrors) < 0)
goto error;
for (i = 0; i < nerrors; i++) {
if (VIR_STRDUP(val[i].disk, errors[i].disk) < 0)
goto error;
val[i].error = errors[i].error;
}
*ret_errors_len = nerrors;
*ret_errors_val = val;
return 0;
error:
if (val) {
size_t j;
for (j = 0; j < i; j++)
VIR_FREE(val[j].disk);
VIR_FREE(val);
}
return -1;
}