This was also found while investigating
https://bugzilla.redhat.com/show_bug.cgi?id=670848
An EOF on a domain's monitor socket results in an event being queued
to handle the EOF. The handler calls qemuProcessHandleMonitorEOF. If
it is a transient domain, this leads to a call to
virDomainRemoveInactive, which removes the domain from the driver's
hashtable and unref's it. Nowhere in this code is the qemu driver lock
acquired.
However, all modifications to the driver's domain hashtable *must* be
done while holding the driver lock, otherwise the hashtable can become
corrupt, and (even more likely) another thread could call a different
hashtable function and acquire a pointer to the domain that is in the
process of being destroyed.
To prevent such a disaster, qemuProcessHandleMonitorEOF must get the
qemu driver lock *before* it gets the DomainObj's lock, and hold it
until it is finished with the DomainObj. This guarantees that nobody
else modifies the hashtable at the same time, and that anyone who had
already gotten the DomainObj from the hashtable prior to this call has
finished with it before we remove/destroy it.
This was found while researching the root cause of:
https://bugzilla.redhat.com/show_bug.cgi?id=670848
virDomainUnref should only be called with the lock held for the
virDomainObj in question. However, when a transient qemu domain gets
EOF on its monitor socket, it queues an event which frees the monitor,
which unref's the virDomainObj without first locking it. If another
thread has already locked the virDomainObj, the modification of the
refcount could potentially be corrupted. In an extreme case, it could
also be potentially unlocked by virDomainObjFree, thus left open to
modification by anyone else who would have otherwise waited for the
lock (not to mention the fact that they would be accessing freed
data!).
The solution is to have qemuMonitorFree lock the domain object right
before unrefing it. Since the caller to qemuMonitorFree doesn't expect
this lock to be held, if the refcount doesn't go all the way to 0,
qemuMonitorFree must unlock it after the unref.
qemudDomainSaveImageStartVM was evil - it closed the incoming fd
argument on some, but not all, code paths, without informing the
caller about that action. No wonder that this resulted in
double-closes: https://bugzilla.redhat.com/show_bug.cgi?id=672725
* src/qemu/qemu_driver.c (qemudDomainSaveImageStartVM): Alter
signature, to avoid double-close.
(qemudDomainRestore, qemudDomainObjRestore): Update callers.
When a SPICE or VNC graphics controller is present, and sound is
piggybacked over a channel to the graphics device rather than
directly accessing host hardware, then there is no need to grant
host hardware access to that qemu process.
* src/qemu/qemu_cgroup.c (qemuSetupCgroup): Prevent sound with
spice, and with vnc when vnc_allow_host_audio is 0.
Reported by Daniel Berrange.
Relax the restriction that the hash table key must be a string
by allowing an arbitrary hash code generator + comparison func
to be provided
* util/hash.c, util/hash.h: Allow any pointer as a key
* internal.h: Include stdbool.h as standard.
* conf/domain_conf.c, conf/domain_conf.c,
conf/nwfilter_params.c, nwfilter/nwfilter_gentech_driver.c,
nwfilter/nwfilter_gentech_driver.h, nwfilter/nwfilter_learnipaddr.c,
qemu/qemu_command.c, qemu/qemu_driver.c,
qemu/qemu_process.c, uml/uml_driver.c,
xen/xm_internal.c: s/char */void */ in hash callbacks
Since the deallocator is passed into the constructor of
a hash table it is not desirable to pass it into each
function again. Remove it from all functions, but provide
a virHashSteal to allow a item to be removed from a hash
table without deleteing it.
* src/util/hash.c, src/util/hash.h: Remove deallocator
param from all functions. Add virHashSteal
* src/libvirt_private.syms: Add virHashSteal
* src/conf/domain_conf.c, src/conf/nwfilter_params.c,
src/nwfilter/nwfilter_learnipaddr.c,
src/qemu/qemu_command.c, src/xen/xm_internal.c: Update
for changed hash API
This is done for two reasons:
- we are getting very close to 64 flags which is the maximum we can use
with unsigned long long
- by using LL constants in enum we already violates C99 constraint that
enum values have to fit into int
Steps to reproduce this bug:
1. virsh attach-disk domain --source imagefile --target sdb --sourcetype file --driver qemu --subdriver raw
2. virsh detach-device controller.xml # remove scsi controller 0
3. virsh detach-disk domain sdb
error: Failed to detach disk
error: operation failed: detaching scsi0-0-1 device failed: Device 'scsi0-0-1' not found
I think we should not detach a controller when it is used by some other device.
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Done mechanically with:
$ git grep -l '\bDEBUG0\? *(' | xargs -L1 sed -i 's/\bDEBUG0\? *(/VIR_&/'
followed by manual deletion of qemudDebug in daemon/libvirtd.c, along
with a single 'make syntax-check' fallout in the same file, and the
actual deletion in src/util/logging.h.
* src/util/logging.h (DEBUG, DEBUG0): Delete.
* daemon/libvirtd.h (qemudDebug): Likewise.
* global: Change remaining clients over to VIR_DEBUG counterpart.
Now that the virHash handling functions call virReportOOMError by
themselves when needed, users of the virHash API no longer need to
do it by themselves. Since users of the virHash API were not
consistently calling virReportOOMError after memory failures from
the virHash code, this has the added benefit of making OOM
reporting from this code more consistent and reliable.
When we attach a disk, but we specify a wrong format of disk image,
qemu monitor command drive_add will fail, but libvirt does not detect
this error.
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
An upcoming patch has a use for a tap device to be created that
doesn't need to be actually put into the "up" state, and keeping it
"down" keeps the output of ifconfig from being unnecessarily cluttered
(ifconfig won't show down interfaces unless you add "-a").
bridge.[ch]: add "up" as an arg to brAddTap()
uml_conf.c, qemu_command.c: add "up" (set to "true") to brAddTap() call.
This is in response to:
https://bugzilla.redhat.com/show_bug.cgi?id=629662
Explanation
qemu's virtio-net-pci driver allows setting the algorithm used for tx
packets to either "bh" or "timer". This is done by adding ",tx=bh" or
",tx=timer" to the "-device virtio-net-pci" commandline option.
'bh' stands for 'bottom half'; when this is set, packet tx is all done
in an iothread in the bottom half of the driver. (In libvirt, this
option is called the more descriptive "iothread".)
'timer' means that tx work is done in qemu, and if there is more tx
data than can be sent at the present time, a timer is set before qemu
moves on to do other things; when the timer fires, another attempt is
made to send more data. (libvirt retains the name "timer" for this
option.)
The resulting difference, according to the qemu developer who added
the option is:
bh makes tx more asynchronous and reduces latency, but potentially
causes more processor bandwidth contention since the cpu doing the
tx isn't necessarily the cpu where the guest generated the
packets.
Solution
This patch provides a libvirt domain xml knob to change the option on
the qemu commandline, by adding a new attribute "txmode" to the
<driver> element that can be placed inside any <interface> element in
a domain definition. It's use would be something like this:
<interface ...>
...
<model type='virtio'/>
<driver txmode='iothread'/>
...
</interface>
I chose to put this setting as an attribute to <driver> rather than as
a sub-element to <tune> because it is specific to the virtio-net
driver, not something that is generally usable by all network drivers.
(note that this is the same placement as the "driver name=..."
attribute used to choose kernel vs. userland backend for the
virtio-net driver.)
Actually adding the tx=xxx option to the qemu commandline is only done
if the version of qemu being used advertises it in the output of
qemu -device virtio-net-pci,?
If a particular txmode is requested in the XML, and the option isn't
listed in that help output, an UNSUPPORTED_CONFIG error is logged, and
the domain fails to start.
When the <driver> element (and its "name" attribute) was added to the
domain XML's interface element, a "backend" enum was simply added to
the toplevel of the virDomainNetDef struct.
Ignoring the naming inconsistency ("name" vs. "backend"), this is fine
when there's only a single item contained in the driver element of the
XML, but doesn't scale well as we add more attributes that apply to
the backend of the virtio-net driver, or add attributes applicable to
other drivers.
This patch changes virDomainNetDef in two ways:
1) Rename the item in the struct from "backend" to "name", so that
it's the same in the XML and in the struct, hopefully avoiding
confusion for someone unfamiliar with the function of the
attribute.
2) Create a "driver" union within virDomainNetDef, and a "virtio"
struct in that struct, which contains the "name" enum value.
3) Move around the virDomainNetParse and virDomainNetFormat functions
to allow for simple plugin of new attributes without disturbing
existing code. (you'll note that this results in a seemingly
redundant if() in the format function, but that will no longer be
the case as soon as a 2nd attribute is added).
In the future, new attributes for the virtio driver backend can be
added to the "virtio" struct, and any other network device backend that
needs an attribute will have its own struct added to the "driver"
union.
The introduction of the v3 migration protocol, along with
support for migration cookies, will significantly expand
the size of the migration code. Move it all to a separate
file to make it more manageable
The functions are not moved 100%. The API entry points
remain in the main QEMU driver, but once the public
virDomainPtr is resolved to the internal virDomainObjPtr,
all following code is moved.
This will allow the new v3 API entry points to call into the
same shared internal migration functions
* src/qemu/qemu_domain.c, src/qemu/qemu_domain.h: Add
qemuDomainFormatXML helper method
* src/qemu/qemu_driver.c: Remove all migration code
* src/qemu/qemu_migration.c, src/qemu/qemu_migration.h: Add
all migration code.
Move the qemudStartVMDaemon and qemudShutdownVMDaemon
methods into a separate file, renaming them to
qemuProcessStart, qemuProcessStop. All helper methods
called by these are also moved & renamed to match
* src/Makefile.am: Add qemu_process.c/.h
* src/qemu/qemu_command.c: Add qemuDomainAssignPCIAddresses
* src/qemu/qemu_command.h: Add VNC port min/max
* src/qemu/qemu_domain.c, src/qemu/qemu_domain.h: Add
domain event queue helpers
* src/qemu/qemu_driver.c, src/qemu/qemu_driver.h: Remove
all QEMU process startup/shutdown functions
* src/qemu/qemu_process.c, src/qemu/qemu_process.h: Add
all QEMU process startup/shutdown functions
"qemudDomainSaveFlag" goto wrong label "endjob", which will cause
error when security manager trying to restore label (regression).
As it's more reasonable to check if vm is shutoff immediately, and
return right away if it is, remove the checking in "qemudDomainSaveFlag",
and add checking in "qemudDomainSave".
* src/qemu/qemu_driver.c
The code expected that host CPU architecture matches the architecture on
which libvirt runs. This is normally true but not in tests, where host
CPU is faked to produce consistent results.
The processWatchdogEvent fix is real, although it can only trigger
on OOM, since bad things happen if doCoreDump is called with a NULL
pathname argument. The other fixes silence clang, but aren't a real
bug because virReportErrorHelper tolerates a NULL format string even
though *printf does not.
* src/qemu/qemu_driver.c (processWatchdogEvent): Exit on OOM.
(qemuDomainIsActive, qemuDomainIsPersistent, qemuDomainIsUpdated):
Provide valid message.
Commit 9962e406c664ed5521f5aca500c860a331cb3979 introduced a
problem where if the VM failed to startup, it would not be
correctly cleaned up. Amongst other things the SELinux
security label would not be removed, which prevents the VM
from ever starting again.
The virDomainIsActive() check at the start of qemudShutdownVMDaemon
checks for vm->def->id not being -1. By moving the assignment of the
VM id to the start of qemudStartVMDaemon, we can ensure cleanup will
occur on failure
* src/qemu/qemu_driver.c: Move initialization of 'vm->def->id'
so that qemudShutdownVMDaemon() will process the shutdown
Suspending a VM which contains shell meta characters doesn't work with
libvirt-0.8.7:
/var/log/libvirt/qemu/andreas_231-ne\ doch\ nicht.log:
sh: -c: line 0: syntax error near unexpected token `doch'
sh: -c: line 0: `cat | { dd bs=4096 seek=1 if=/dev/null && dd bs=1048576; }
Although target="andreas_231-ne doch nicht" contains shell meta
characters (here: blanks), they are not properly escaped by
src/qemu/qemu_monitor_{json,text}.c#qemuMonitor{JSON,Text}MigrateToFile()
First, the filename needs to be properly escaped for the shell, than
this command line has to be properly escaped for qemu again.
For this to work, remove the old qemuMonitorEscapeArg() wrapper, rename
qemuMonitorEscape() to it removing the handling for shell=TRUE, and
implement a new qemuMonitorEscapeShell() returning strings using single
quotes.
Using double quotes or escaping special shell characters with backslashes
would also be possible, but the set of special characters heavily
depends on the concrete shell (dsh, bash, zsh) and its setting (history
expansion, interactive use, ...)
Signed-off-by: Philipp Hahn <hahn@univention.de>
QEMUD_CMD_FLAG_PCI_MULTIBUS should be set in the function
qemuCapsExtractVersionInfo()
The flag QEMUD_CMD_FLAG_PCI_MULTIBUS is used in the function
qemuBuildDeviceAddressStr(). All callers get qemuCmdFlags
by the function qemuCapsExtractVersionInfo() except that
testCompareXMLToArgvFiles() in qemuxml2argvtest.c.
So we should set QEMUD_CMD_FLAG_PCI_MULTIBUS in the function
qemuCapsExtractVersionInfo() instead of qemuBuildCommandLine()
because the function qemuBuildCommandLine() does not be called
when we attach a pci device.
tests: set QEMUD_CMD_FLAG_PCI_MULTIBUS in testCompareXMLToArgvFiles()
set QEMUD_CMD_FLAG_PCI_MULTIBUS before calling qemuBuildCommandLine()
as the flags is not set by qemuCapsExtractVersionInfo().
Signed-off-by: Wen Congyang <wency@cn.fujitsu.com>
Quite a few hosts don't have cgroups mounted and so see warnings
from libvirt logged, which then cause bug reports, etc. Reduce
the log level to INFO so they're not visible by default
* src/qemu/qemu_driver.c: Reduce log level for cgroups
When probing machine types if the QEMU binary does not exist
we get a hard to diagnose error, due to the execve() in the
child failing
error: internal error Child process exited with status 1.
Add an explicit check so that we get
error: Cannot find QEMU binary /usr/libexec/qem3u-kvm: No such file or directory
* src/qemu/qemu_capabilities.c: Check for QEMU binary
qemu 0.13.0 (at least as built for Fedora 14, and also backported to
RHEL 6.0 qemu) supported an older syntax for a spicevmc channel; it's
not as flexible (it has an implicit name and hides the chardev
aspect), but now that we support spicevmc, we might as well target
both variants.
* src/qemu/qemu_capabilities.h (QEMUD_CMD_FLAG_DEVICE_SPICEVMC):
New flag.
* src/qemu/qemu_capabilities.c (qemuCapsParseDeviceStr): Set it
correctly.
* src/qemu/qemu_command.h (qemuBuildVirtioSerialPortDevStr): Drop
declaration.
* src/qemu/qemu_command.c (qemuBuildVirtioSerialPortDevStr): Alter
signature, check flag.
(qemuBuildCommandLine): Adjust caller and check flag.
* tests/qemuhelptest.c (mymain): Update test.
* tests/qemuxml2argvtest.c (mymain): New test.
* tests/qemuxml2argvdata/qemuxml2argv-channel-spicevmc-old.xml:
New file.
* tests/qemuxml2argvdata/qemuxml2argv-channel-spicevmc-old.args:
Likewise.
Adds <smartcard mode='passthrough' type='spicevmc'/>, which uses the
new <channel name='smartcard'/> of <graphics type='spice'>.
* docs/schemas/domain.rng: Support new XML.
* docs/formatdomain.html.in: Document it.
* src/conf/domain_conf.h (virDomainGraphicsSpiceChannelName): New
enum value.
(virDomainChrSpicevmcName): New enum.
(virDomainChrSourceDef): Distinguish spicevmc types.
* src/conf/domain_conf.c (virDomainGraphicsSpiceChannelName): Add
smartcard.
(virDomainSmartcardDefParseXML): Parse it.
(virDomainChrDefParseXML, virDomainSmartcardDefParseXML): Set
spicevmc name.
(virDomainChrSpicevmc): New enum conversion functions.
* src/libvirt_private.syms: Export new functions.
* src/qemu/qemu_command.c (qemuBuildChrChardevStr): Conditionalize
name.
* tests/qemuxml2argvtest.c (domain): New test.
* tests/qemuxml2argvdata/qemuxml2argv-smartcard-passthrough-spicevmc.args:
New file.
* tests/qemuxml2argvdata/qemuxml2argv-smartcard-passthrough-spicevmc.xml:
Likewise.
Inspired by https://bugzilla.redhat.com/show_bug.cgi?id=615757
Add a new character device backend for virtio serial channels that
activates the QEMU spice agent on the main channel using the vdagent
spicevmc connection. The <target> must be type='virtio', and supports
an optional name that specifies how the guest will see the channel
(for now, name must be com.redhat.spice.0).
<channel type='spicevmc'>
<target type='virtio'/>
<address type='virtio-serial' controller='1' bus='0' port='3'/>
</channel>
* docs/schemas/domain.rng: Support new XML.
* docs/formatdomain.html.in: Document it.
* src/conf/domain_conf.h (virDomainChrType): New enum value.
* src/conf/domain_conf.c (virDomainChr): Add spicevmc.
(virDomainChrDefParseXML, virDomainChrSourceDefParseXML)
(virDomainChrDefParseTargetXML): Parse and enforce proper use.
(virDomainChrSourceDefFormat, virDomainChrDefFormat): Format.
* src/qemu/qemu_command.c (qemuBuildChrChardevStr)
(qemuBuildCommandLine): Add qemu support.
* tests/qemuxml2argvtest.c (domain): New test.
* tests/qemuxml2argvdata/qemuxml2argv-channel-spicevmc.xml: New
file.
* tests/qemuxml2argvdata/qemuxml2argv-channel-spicevmc.args:
Likewise.
Signed-off-by: Eric Blake <eblake@redhat.com>
