Commit Graph

4389 Commits

Author SHA1 Message Date
Matthias Bolte
67f34cd67b xen: Remove virConnectPtr from virXenStoreError 2010-04-06 01:46:02 +02:00
Matthias Bolte
68d5d96bce xen: Remove virConnectPtr from virXenError/virXenErrorFunc 2010-04-06 01:45:47 +02:00
Matthias Bolte
f2192637b0 xen: Remove virConnectPtr from virXMError 2010-04-06 01:44:51 +02:00
Matthias Bolte
5cec7af469 xen: Remove virConnectPtr from virXendError 2010-04-06 01:43:51 +02:00
Matthias Bolte
2127b42b46 proxy: Remove virConnectPtr from virProxyError 2010-04-06 01:43:23 +02:00
Matthias Bolte
a3884f8705 vbox: Remove virConnectPtr from vboxError 2010-04-06 01:41:58 +02:00
Matthias Bolte
1c91a0b44e test: Remove virConnectPtr from testError 2010-04-06 01:41:58 +02:00
Matthias Bolte
49411f029a Remove unnecessary trailing \n in log messages 2010-04-06 01:41:58 +02:00
Matthias Bolte
d2e0bd1779 Fix compiler warning about non-literal format string 2010-04-06 01:38:53 +02:00
Stefan Berger
45274d4121 This patch removes the virConnectPtr parameter from all functions where it's not necessary starting out with removing it as a parameter to the error reporting function. 2010-04-05 12:34:55 -04:00
Chris Lalancette
88151c7cbc Snapshot virsh implementation.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-05 10:25:02 -04:00
Jiri Denemark
c572222a14 Snapshots for VBox 2010-04-05 10:24:57 -04:00
Chris Lalancette
462c74c38a Snapshot QEMU driver.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-05 10:24:51 -04:00
Chris Lalancette
32c3c1f0b7 Only assign newDef when we have a new def.
While playing around with def/newDef with the qemu code,
I noticed that newDef was *always* getting set to a value,
even when I didn't redefine the domain.  I think the problem
is the virDomainLoadConfig is always doing virDomainAssignDef
regardless of whether the domain already exists in the hashtable.
In turn, virDomainAssignDef is assigning the definition (which
is actually a duplicate) to newDef.  Fix this so that newDef stays
NULL until we actually have a new def.

Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-05 10:24:44 -04:00
Chris Lalancette
9b55a52b4f Snapshot internal methods.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-05 10:24:38 -04:00
Chris Lalancette
2f992d4be4 Snapshot API framework.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-05 10:24:34 -04:00
David Allan
1bbe12de44 Clarified error message
* Since the file pattern matches RNG schemas as well as C sources, the error message should mention both.
2010-04-05 10:24:13 -04:00
Laine Stump
5dac5970a5 Eliminate compiler warning about non-const format string 2010-04-04 22:58:33 -04:00
Matthias Bolte
905709500a xenapi: Add managedsave entries to the driver struct
Commit 15c647a91e added the new
managedsave entries to all driver structs except the XenAPI one.
2010-04-04 18:23:52 +02:00
Stefan Berger
a44b23ba63 Get rid of the regular expressions when evaluating variable names and
values. Rather use the strspn() function. Along with this cleanup the
initialization function for the code that used the regular expression
can also be removed.
2010-04-04 10:34:52 -04:00
Matthias Bolte
5288b2ad8e nwfilter: Fix random index in virNWFilterRuleDefDetailsFormat
An uninitialized int value was used to index an array. This can
result in a segfault in nwfilterxml2xmltest.
2010-04-04 16:06:56 +02:00
Matthias Bolte
1b747f0ded xenapi: Fix uninitialized variable warning 2010-04-04 16:06:39 +02:00
Daniel Veillard
d068f9609c Add a managedsave command to virsh
This command implements the managed save operation

* tools/virsh.c: new command
* tools/virsh.pod: documentation
2010-04-04 11:36:59 +02:00
Daniel Veillard
6e41f30efc Implement managed save operations for qemu driver
The images are saved in /var/lib/libvirt/qemu/save/
and named $domainname.save . The directory is created appropriately
at daemon startup. When a domain is started while a saved image is
available, libvirt will try to load this saved image, and start the
domain as usual in case of failure. In any case the saved image is
discarded once the domain is created.

* src/qemu/qemu_conf.h: adds an extra save path to the driver config
* src/qemu/qemu_driver.c: implement the 3 new operations and handling
  of the image directory
2010-04-04 11:36:46 +02:00
Daniel Veillard
7a3da17a58 Implement remote protocol for managed save
* src/remote/remote_protocol.x src/remote/remote_protocol.h
  src/remote/remote_protocol.c src/remote/remote_driver.c: add the entry
  points in the remote driver
* daemon/remote.c daemon/remote_dispatch_args.h
  daemon/remote_dispatch_prototypes.h daemon/remote_dispatch_table.h:
  and implement the daemon counterpart
2010-04-04 11:36:33 +02:00
Daniel Veillard
15c647a91e Add managed save API entry points
virDomainManagedSave() is to be run on a running domain. Once the call
complete, as in virDomainSave() the domain is stopped upon completion,
but there is no restore counterpart as any order to start the domain
from the API would load the state from the managed file, similary if
the domain is autostarted when libvirtd starts.
Once a domain has restarted his managed save image is destroyed,
basically managed save image can only exist for a stopped domain,
for a running domain that would be by definition outdated data.

* include/libvirt/libvirt.h.in src/libvirt.c src/libvirt_public.syms:
  adds the new entry points virDomainManagedSave(),
  virDomainHasManagedSaveImage() and virDomainManagedSaveRemove()
* src/driver.h src/esx/esx_driver.c src/lxc/lxc_driver.c
  src/opennebula/one_driver.c  src/openvz/openvz_driver.c
  src/phyp/phyp_driver.c src/qemu/qemu_driver.c src/vbox/vbox_tmpl.c
  src/remote/remote_driver.c src/test/test_driver.c src/uml/uml_driver.c
  src/xen/xen_driver.c: add corresponding new internal drivers entry
  points
2010-04-04 11:07:48 +02:00
Eric Blake
d002c250cd build: improve check for out-of-date .gnulib submodule
git reset --hard 96e5a2d4d5
./autogen.sh
make -s
git pull
make -s    <-- expecting auto-bootstrap here, doesn't happen

Use git diff to expose whether the submodule has untracked changes,
which are typical on an incremental pull if .gnulib was updated but
the user did not manually run 'git submodule update'.

After this patch is applied, I encountered a new problem when
following the reproducing pattern.  Basically, the change to .gnulib
between libvirt's commit 96e5a2d4 and this patch introduced a change
to sys_ioctl.in.h, but gnulib (intentionally) does not make the
replacement headers depend on Makefile changes.  Therefore, I ended up
with the generated replacement header being broken:
gnulib/lib/sys/ioctl.h complained about a use of @.  But that seems
like something that should be fixed upstream in gnulib's bootstrap
script (that is, when doing a gnulib update, all files created from
.in.h file should probably be deleted).  Without the benefit of that
proposed gnulib fix, I worked around the problem by manually removing
the stale gnulib/lib/sys/ioctl.h.

* autogen.sh (t): Also run bootstrap if the gnulib submodule needs
to be updated.
* cfg.mk (_autogen): Likewise.
Reported by Matthias Bolte.
2010-04-02 15:49:32 -06:00
Stefan Berger
8d30e5f74d Use the virStrToLong_ui() function rather than the virStrToLong_i()
where possible.
2010-04-02 15:02:27 -04:00
Stefan Berger
d9292cfefb The attached patch optimizes the validation of the name of an interface. 2010-04-02 14:57:01 -04:00
Stefan Berger
8459218cbc This patch adds a couple of test cases for the XML parsing test suite covering various filterable protocols. For each test case an input XML and an output XML is provided checking the input XML after parsing and converting back into XML against the exepcted output XML.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
2010-04-02 13:28:28 -04:00
Stefan Berger
b3e8f9f415 The following issues are fixed in the patch below:
- ebtables requires that some of the command line parameters are passed as hex numbers; so have those attributes call a function that prints 16 and 8 bit integers as hex nunbers.

- ip6tables requires '--icmpv6-type' rather than '--icmp-type'

- ebtables complains about protocol identifiers lower than 0x600, so already discard anything lower than 0x600 in the parser

- make the protocol entry types more readable using a #define for its entries

- continue parsing a filtering rule even if a faulty entry is encountered; return an error value at the end and let the caller decide what to do with the rule's object

- fix an error message
2010-04-02 13:21:10 -04:00
Eric Blake
e7168f7965 build: import latest gnulib
A lot of syntax check rules have to be rewritten, but the
result is easier to maintain.  I tested each syntax rule
by intentionally introducing a temporary violation of the rule.
Additionally, some false positives for unmarked_diagnostics
crept in, and an improved copyright_format test caught some bugs.

* .gnulib: Update to latest.
* cfg.mk (sc_prohibit_test_minus_ao): Delete, it was moved into
gnulib's maint.mk.
(sc_avoid_write, sc_prohibit_strcmp_and_strncmp)
(sc_prohibit_asprintf, sc_prohibit_strncpy, sc_prohibit_readlink)
(sc_prohibit_gethostname, sc_prohibit_gettext_noop)
(sc_prohibit_VIR_ERR_NO_MEMORY, sc_prohibit_nonreentrant)
(sc_prohibit_ctype_h, sc_TAB_in_indentation)
(sc_avoid_ctype_macros)
(sc_prohibit_virBufferAdd_with_string_literal)
(sc_prohibit_gethostby, sc_copyright_format): Rewrite in terms of
new maint.mk macros.
(sc_libvirt_unmarked_diagnostics): Fix whitespace.
* .x-sc_unmarked_diagnostics: New file.
* tests/object-locking.ml: Fix copyright.
* tools/virt-pki-validate.in: Likewise.
* tools/virt-xml-validate.in: Likewise.
2010-04-02 10:18:55 -06:00
Chris Lalancette
96e5a2d4d5 Add a missing break statement to nwfilter errors.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-02 10:13:44 -04:00
Chris Lalancette
fba3dfd156 Make virDomainLoadConfig static.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-02 10:13:07 -04:00
Jiri Denemark
f76c3493d8 VBox: Fix use of uninitialized value 2010-04-02 15:59:23 +02:00
Laine Stump
4ee2b31804 Changes to clock timer XML to match final design.
The clock timer XML is being updated in the following ways (based on
further off-list discussion that was missed during the initial
implementation):

1) 'wallclock' is changed to 'track', and the possible values are 'boot'
 (corresponds to old 'host'), 'guest', and 'wall'.

2) 'mode' has an additional value 'smpsafe'

3) when tickpolicy='catchup', there can be an optional sub-element of
   timer called 'catchup':

   <catchup threshold=123 slew=120 limit=10000/>

Those three values are all longs, always optional, and if they are present,
they are positive. Internally, 0 indicates "unspecified".

* docs/schemas/domain.rng: updated RNG definition to account for changes

* src/conf/domain_conf.h: change the C struct and enums to match changes.

* src/conf/domain_conf.c: timer parse and format functions changed to
                          handle the new selections and new element.

* src/libvirt_private.syms: *TimerWallclock* changes to *TimerTrack*

* src/qemu/qemu_conf.c: again, account for Wallclock --> Track change.
2010-04-02 09:50:29 -04:00
Laine Stump
8a7b4be5ab Allow domain disk images on root-squash NFS to coexist with security driver.
(suggested by Daniel Berrange, tested by Dan Kenigsberg)

virStorageFileGetMetadata will fail for disk images that are stored on
a root-squash NFS share that isn't world-readable.
SELinuxSetSecurityImageLabel is called during the startup of every
domain (as long as security_driver != "none"), and it will propogate
the error from virStorageFileGetMetadata, causing the domain startup
to fail. This is, however, a common scenario when qemu is run as a
non-root user and the disk image is stored on NFS.

Ignoring this failure (which doesn't matter in this case, since the
next thing done by SELinuxSetSecurityImageLabel - setting the file
context - will also fail (and that function already ignores failures
due to root-squash NFS) will allow us to continue bringing up the
domain. The result is that we don't need to disable the entire
security driver just because a domain's disk image is stored on
root-squashed NFS.
2010-04-02 09:39:20 -04:00
Laine Stump
78151ec96e Eliminate compile warnings in nwfilter error log calls 2010-04-02 09:39:00 -04:00
Chris Lalancette
d62f4c36c2 Don't use virFileReadLimFD in qemuDomainRestore.
virFileReadLimFD is a poor fit for reading the header
of the restore file.  The problem is that virFileReadLimFD
returns an error when there is more data after the amount
you ask to read, but that is *expected* in this case.

This patch is essentially a revert of
1a4d5c9543, but I don't think
that commit does what it says anyway.  It purports to prevent
an unwarranted OOM error, but since virFileReadLimFD will
allocate memory up to the maximum anyway, the upper limit
on the total amount of memory allocated is the same for either
the old version or the new version.  Since the old saferead
actually works and virFileReadLimFD does not, revert to
using saferead.

Signed-off-by: Chris Lalancette <clalance@redhat.com>
2010-04-02 09:23:41 -04:00
Jim Fehlig
320ea3a47e Only parse 'CPU XML' in virCPUDefParseXML()
Received report of user crashing libvirtd with

virsh capabilities > capabilities.xml
virsh cpu-compare capabilities.xml

While user has been informed about proper usage of cpu-compare,
segfaulting libvirt should be avoided.

Do not parse CPU definition in virCPUDefParseXML() if XML is not
a 'cpu' node.
2010-04-01 14:04:43 +02:00
Daniel P. Berrange
f983c32675 Keep build quiet for generated file
Adds $(AM_V_GEN) to many more manual makefile.am rules that
were generating files
2010-04-01 12:35:51 +01:00
Daniel P. Berrange
cd2404a3c3 Keep track of guest paused state after disk IO / watchdog events
When a watchdog/IO error occurs, one of the possible actions that
QEMU might take is to pause the guest. In this scenario libvirt
needs to update its internal state for the VM, and emit a
lifecycle event:

  VIR_DOMAIN_EVENT_SUSPENDED

with a detail being one of:

  VIR_DOMAIN_EVENT_SUSPENDED_IOERROR
  VIR_DOMAIN_EVENT_SUSPENDED_WATCHDOG

To future proof against possible QEMU support for multiple monitor
consoles, this patch also hooks into the 'STOPPED' event in QEMU
and emits a generic VIR_DOMAIN_EVENT_SUSPENDED_PAUSED event

* include/libvirt/libvirt.h.in: Add VIR_DOMAIN_EVENT_SUSPENDED_IOERROR
* src/qemu/qemu_driver.c: Update VM state to paused when IO error
  or watchdog events occurrs
* src/qemu/qemu_monitor_json.c: Fix typo in disk IO event name
2010-04-01 12:35:46 +01:00
Matthias Bolte
d13eb9fc95 Replace sscanf in PCI device address parsing
This also fixes a problem with MinGW's GCC on Windows. GCC complains
about the L modifier being unknown.

Parsing in pciIterDevices is stricter now and doesn't accept trailing
characters after the actual <domain>:<bus>:<slot>.<function> sequence
anymore.

Parsing in pciWaitForDeviceCleanup is also stricter now and expects
the <start>-<end> : <domain>:<bus>:<slot>.<function> sequence to be
terminated by \n.

Change domain from unsigned long long to unsigned int in
pciWaitForDeviceCleanup, because everywhere else domain is handled as
unsigned int too.
2010-04-01 12:53:41 +02:00
Matthias Bolte
57dab74804 xen: Use virStrToLong_i instead of sscanf for XenD port parsing
Parsing is stricter now and doesn't accept trailing characters
after the actual value anymore.
2010-04-01 12:53:41 +02:00
Matthias Bolte
82cb2e73aa xenapi: Use virStrToLong_i instead of sscanf for CPU map parsing
Parsing is stricter now and doesn't accept trailing characters
after the actual value anymore.
2010-04-01 12:53:41 +02:00
Matthias Bolte
1778da7185 openvz: Use strtok_r instead of sscanf for VPS UUID parsing
Also free 2k stack space.
2010-04-01 12:53:41 +02:00
Matthias Bolte
00e00510c5 xen: Use virParseMacAddr instead of sscanf
This also fixes a bug in xenXMDomainConfigParse where uninitialized
memory would be used as MAC address if sscanf fails.
2010-04-01 12:53:41 +02:00
Matthias Bolte
2f047d4816 vbox: Replace atoi with virStrToLong_i
Parsing is stricter now and doesn't accept trailing characters
after the actual value or non-number strings anymore. atoi just
returns 0 in case it cannot parse a number from the given string.
Now an error is reported for such a string.
2010-04-01 12:53:41 +02:00
Matthias Bolte
73b45bfbff cgroup: Replace sscanf with virStrToLong_ll
The switch from %lli to %lld in virCgroupGetValueI64 is intended,
as virCgroupGetValueU64 uses base 10 too, and virCgroupSetValueI64
uses %lld to format the number to string.

Parsing is stricter now and doesn't accept trailing characters
after the actual value anymore.
2010-04-01 12:53:41 +02:00
Matthias Bolte
56bac4a221 Refactor major.minor.micro version parsing into a function
virParseVersionString uses virStrToLong_ui instead of sscanf.

This also fixes a bug in the UML driver, that always returned 0
as version number.

Introduce STRSKIP to check if a string has a certain prefix and
to skip this prefix.
2010-04-01 12:53:41 +02:00