Commit Graph

4759 Commits

Author SHA1 Message Date
Jim Meyering
0641f0f72c build: avoid compile failure on linux kernels older than 2.6.19
* configure.ac: Check for <linux/magic.h>.
* src/util/storage_file.c: Include <linux/magic.h> only if present.
Linux kernels prior to 2.6.19 lacked it.
[__linux__] (NFS_SUPER_MAGIC): Define if not already defined.
2010-05-17 16:50:36 +02:00
Jim Meyering
258d59cff9 x86ModelHasFeature: avoid NULL-dereference for unmatched CPU "feature"
* src/cpu/cpu_x86.c (x86ModelHasFeature): Do not dereference the pointer
returned by x86cpuidFind without first ensuring it is non-NULL.
2010-05-17 16:50:36 +02:00
Cole Robinson
83be64034a qemu: Report cmdline output if VM dies early
qemuReadLogOutput early VM death detection is racy and won't always work.
Startup then errors when connecting to the VM monitor. This won't report
the emulator cmdline output which is typically the most useful diagnostic.

Check if the VM has died at the very end of the monitor connection step,
and if so, report the cmdline output.

See also: https://bugzilla.redhat.com/show_bug.cgi?id=581381
2010-05-17 10:15:53 -04:00
Cole Robinson
d536f6b177 qemu: Fix previous commit, use comparision in if() 2010-05-17 09:39:11 -04:00
Jim Meyering
560758c900 qemu_driver: avoid NULL dereference
* src/qemu/qemu_driver.c (qemudDomainStart): After setting vm to NULL,
goto cleanup, rather than dereferencing the NULL pointer.
2010-05-17 13:47:45 +02:00
Daniel P. Berrange
2d665c9e2d Remove debugging fprintf() calls
* src/qemu/qemu_driver.c: Remove debugging fprintf() calls
  accidentally left in code
2010-05-17 10:44:49 +01:00
Jim Meyering
b48fb801dd qemudDomainSetVcpus: avoid NULL-deref
* src/qemu/qemu_driver.c (qemudDomainSetVcpus): Avoid NULL-deref
upon unknown UUID.  Call qemuDomainObjBeginJob(vm) only after
ensuring that vm != NULL, not before.  This potential NULL-deref
was introduced by commit 2c555d87b0.
2010-05-15 09:02:54 +02:00
Eric Blake
39b3845fd7 Revert "build: fix cygwin build"
This reverts commit 2d550542ee.

The patch worked for incremental builds, but broke fresh
builds, because it interfered with automake's automatic
dependency generation.  Until I figure out how to make
automake do what we want, I'd rather leave cygwin broken
but fresh Linux builds working.
2010-05-14 17:46:47 -06:00
Eric Blake
2d550542ee build: fix cygwin build
make[3]: *** No rule to make target `-lxml2', needed by `libvirt.la'.  Stop.

Due to treating the wrong string as a dependency.

* src/Makefile.am (libvirt_la_DEPENDENCIES): Depend only on
locally-built file, not on strings that might resolve as '-lxml2'.
2010-05-14 16:03:57 -06:00
Stefan Berger
ba99a1b637 nwfilter: Add missing driver lock in qemu driver
This adds a missing driver lock in the qemu driver to protect
the list of domains.
2010-05-14 14:22:39 -04:00
Ryota Ozaki
d6644013d2 Fix a misuse of virAsprintf in qemudDomainMemoryPeek
The code specifies driver->cacheDir as the format string,
but it usually doesn't contain '%s', so the subsequent
argument, "/qemu.mem.XXXXXX", is always ignored.

The patch fixes the misuse.
2010-05-14 10:45:58 -06:00
Daniel P. Berrange
de4d70873a Make domain save work when dynamic_ownership=0
Setting dynamic_ownership=0 in /etc/libvirt/qemu.conf prevents
libvirt's DAC security driver from setting uid/gid on disk
files when starting/stopping QEMU, allowing the admin to manage
this manually. As a side effect it also stopped setting of
uid/gid when saving guests to a file, which completely breaks
save when QEMU is running non-root. Thus saved state labelling
code must ignore the dynamic_ownership parameter

* src/qemu/qemu_security_dac.c: Ignore dynamic_ownership=0 when
  doing save/restore image labelling
2010-05-14 09:21:33 -04:00
Daniel P. Berrange
02ddaddfa8 Don't reset user/group/security label on shared filesystems during migrate
When QEMU runs with its disk on NFS, and as a non-root user, the
disk is chownd to that non-root user. When migration completes
the last step is shutting down the QEMU on the source host. THis
normally resets user/group/security label. This is bad when the
VM was just migrated because the file is still in use on the dest
host. It is thus neccessary to skip the reset step for any files
found to be on a shared filesystem

* src/libvirt_private.syms: Export virStorageFileIsSharedFS
* src/util/storage_file.c, src/util/storage_file.h: Add a new
  method virStorageFileIsSharedFS() to determine if a file is
  on a shared filesystem (NFS, GFS, OCFS2, etc)
* src/qemu/qemu_driver.c: Tell security driver not to reset
  disk labels on migration completion
* src/qemu/qemu_security_dac.c, src/qemu/qemu_security_stacked.c,
  src/security/security_selinux.c, src/security/security_driver.h,
  src/security/security_apparmor.c: Add ability to skip disk
  restore step for files on shared filesystems.
2010-05-14 09:21:24 -04:00
Daniel P. Berrange
117d04fb1d Fix handling of disk backing stores with cgroups
The cgroups ACL code was only allowing the primary disk image.
It is possible to chain images together, so we need to search
for backing stores and add them to the ACL too. Since the ACL
only handles block devices, we ignore the EINVAL we get from
plain files. In addition it was missing code to teardown the
cgroup when hot-unplugging a disk

* src/qemu/qemu_driver.c: Allow backing stores in cgroup ACLs
  and add missing teardown code in unplug path
2010-05-14 09:20:13 -04:00
Daniel P. Berrange
abb7694211 Fix possible crash in handling IO Error event
If the IO error event does not include a reason, then there
is a possible crash dispatching the event

* src/conf/domain_event.c: Missing check for a NULL reason before
  strduping allows for a crash
2010-05-14 09:18:51 -04:00
Daniel P. Berrange
ff45b4c26f Add support for NIC hotplug using netdev_add in QEMU
QEMU is gaining a new monitor command netdev_add for hotplugging
NICs using the netdev backend code. We already support this on
the command this, though it is disabled. This adds support for
hotplug too, also to remain disabled until 0.13 QEMU is released

* src/qemu/qemu_driver.c: Support netdev hotplug for NICs
* src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h,
  src/qemu/qemu_monitor_json.c, src/qemu/qemu_monitor_json.h,
  src/qemu/qemu_monitor_text.c, src/qemu/qemu_monitor_text.h: Add
  support for netdev_add and netdev_remove commands
2010-05-14 09:16:36 -04:00
Eric Blake
152ccceb61 datatypes: fix comment typo
* src/datatypes.c: Use correct word.
2010-05-12 12:18:22 -06:00
Jiri Denemark
d84bb6d6a3 Fix monitor ref counting when adding event handle
When closing a monitor using qemuMonitorClose(), we are aware of
the possibility the monitor is still being used somewhere:

    /* NB: ordinarily one might immediately set mon->watch to -1
     * and mon->fd to -1, but there may be a callback active
     * that is still relying on these fields being valid. So
     * we merely close them, but not clear their values and
     * use this explicit 'closed' flag to track this state */

but since we call virEventAddHandle() on that monitor without increasing
its ref counter, the monitor is still freed which makes possible users
of it quite unhappy. The unhappiness can lead to a hang if qemuMonitorIO
tries to lock mutex which no longer exists.
2010-05-12 16:07:42 +02:00
Jiri Denemark
6ef9d9da5e Remove watches before calling REMOTE_PROC_CLOSE
First calling REMOTE_PROC_CLOSE and then removing watches might lead to
a hang as HANGUP event can be triggered before the watches are actually
removed but after virConnectPtr is already freed. As a result of that
remoteDomainEventFired() would try to lock uninitialized mutex, which
would hang for ever.
2010-05-12 16:07:08 +02:00
Jim Meyering
c2c4abb43b tests: use GPLv2+, not GPLv3
* tests/cpuset: Change from GPLv3 to GPLv2+
* tests/read-bufsiz: Likewise.
* tests/read-non-seekable: Likewise.
* tests/start: Likewise.
* tests/undefine: Likewise.
* tests/vcpupin: Likewise.
* tests/virsh-all: Likewise.
* tests/virsh-schedinfo: Likewise.
* tests/virsh-synopsis: Likewise.
2010-05-12 08:41:10 +02:00
Eric Blake
ff1d6f859c libvirt_proxy: link with -lpthread if needed
Continuation of earlier patches to fix LIB_PTHREAD, only
triggered by ./configure --with-xen-proxy (a la autobuild.sh).

* proxy/Makefile.am (libvirt_proxy_LDADD): Add LIB_PTHREAD.
2010-05-11 16:13:16 -06:00
Cole Robinson
74c7a3463d node_device: udev: Fix PCI product/vendor swappage
Product and vendor values were swapped in the XML, which made virt-manager
PCI device listing kinda useless.
2010-05-11 16:55:56 -04:00
Eric Blake
e8a1a730fe build: update gnulib
* .gnulib: Update to latest.
* bootstrap.conf (gnulib_modules): Import netdb.
* src/esx/esx_util.c (AI_ADDRCONFIG): Rely on gnulib.
* src/remote/remote_driver.c (AI_ADDRCONFIG): Likewise.
* tools/virsh.c (WEXITSTATUS, O_SYNC): Likewise.
2010-05-11 10:03:48 -06:00
Eric Blake
03ae900f02 build: allow older gettext
* bootstrap.conf (gnulib_modules): Use gettext-h, not gettext,
since the latter drags in a depedency on gettext 0.18.
Suggested by Bruno Haible.
2010-05-11 10:03:48 -06:00
Jim Meyering
e915962a04 tests: correct PATH in new test, for when running manually
* tests/virsh-schedinfo: This test sets PATH internally, just in
case you're running it manually.  Normally, the PATH setting from
tests/Makefile.am's TESTS_ENVIRONMENT is sufficient.  Prepend the
correct directory, and take advantage of the PATH setting in one
more case.
2010-05-11 18:01:11 +02:00
Daniel P. Berrange
36a03bd2ee Add env variable for debugging gnutls usage
Allow debugging of GNUTLS interactions by setting

  LIBVIRT_GNUTLS_DEBUG=10 LIBVIRT_DEBUG=1 virsh

* src/remote/remote_driver.c: Use LIBVIRT_GNUTLS_DEBUG to
  enable gnutls debugging
2010-05-11 15:54:38 +01:00
Jim Meyering
c5be8bcb8f tests: adjust copyrights on scripts: s/FSF/Red Hat/
* tests/cpuset: Change copyright holder from FSF to Red Hat, Inc.
* tests/read-bufsiz: Likewise.
* tests/read-non-seekable: Likewise.
* tests/start: Likewise.
* tests/undefine: Likewise.
* tests/vcpupin: Likewise.
* tests/virsh-all: Likewise.
* tests/virsh-synopsis: Likewise.
2010-05-11 16:43:07 +02:00
Jim Meyering
9a641564fb virsh: schedinfo --set invalid=value would simply ignore the option
For example, virsh -c test:///default schedinfo 1 --set P=k would
mistakenly exit successfully, giving no indication that it had failed
to set the scheduling parameter "P".
* tools/virsh.c (cmdSchedinfo): Diagnose an invalid --set j=k option,
rather than silently ignoring it.
* tests/virsh-schedinfo: New test for the above.
* tests/Makefile.am (test_scripts): Add it.
Reported by Jintao Yang in http://bugzilla.redhat.com/586632
2010-05-11 16:37:15 +02:00
Jim Meyering
5603515127 virsh: fix a typo in a diagnostic
* tools/virsh.c (cmdSchedInfoUpdate): Fix typo in a diagnostic:
s/an long long/a long long/.  One in a comment, too.
2010-05-11 16:06:53 +02:00
Eric Blake
78a6af1ff9 delMacvtap: typo fix
* src/util/macvtap.c (delMacvtap): Fix documentation.
2010-05-10 17:12:22 -06:00
Eric Blake
3876e010eb maint: allow VPATH use of remote_protocol-structs
* src/Makefile.am (remote_protocol-structs): Ensure file lives in srcdir.
2010-05-10 15:17:31 -06:00
Eric Blake
1c5891204d docs/Makefile.am: remove unnecessary subshells
* docs/Makefile.am (ChangeLog.html.in, %.html.tmp, %.html)
(html/index.html, $(devhelphtml)): Avoid spurious subshells.
2010-05-10 14:59:25 -06:00
Eric Blake
7cdf26637e maint: avoid spurious output if program not present
Some shells warn about missing programs before redirection;
the idiomatic way to silence them is to run the program check
inside a subshell, with the redirections outside the subshell.
But a subshell is only needed in places where it is reasonable
to expect the use of such a noisy shell in the first place.

* src/Makefile.am (remote_protocol-structs): Use subshell, for
FreeBSD 8.0 /bin/sh.
* cfg.mk (sc_preprocessor_indentation): Avoid subshell, since the
only users running cfg.mk can be assumed to have decent tools.
2010-05-10 14:56:37 -06:00
Eric Blake
23958aedf4 storage_encryption: silence clang warning
For printf("%*s",foo,bar), clang complains if foo is not int:

warning: field width should have type 'int', but argument has
type 'unsigned int' [-Wformat]

* src/conf/storage_encryption_conf.c
(virStorageEncryptionSecretFormat, virStorageEncryptionFormat):
Use correct type.
* src/conf/storage_encryption_conf.h (virStorageEncryptionFormat):
Likewise.
2010-05-10 13:53:12 -06:00
Jim Meyering
180d4b2b3b help avoid accidental remote_protocol.x changes
Now, if you update remote_protocol.x without also updating
remote_protocol-structs to match, then "make check" will fail.
* src/Makefile.am (remote_protocol-structs): Extract list of
structs and member names from remote_protocol.o.
(check-local): Depend on it.
* src/remote_protocol-structs: New file.
2010-05-08 10:56:52 +02:00
Eric Blake
d0616e354d mingw32-libvirt.spec: bring up to date
Right now, mingw32-portablexdr is not available in Fedora, but is
present in fedora-mingw.git.  With that package, plus
redhat-rpm-config and this patch, it is once again possible to build
mingw32-libvirt from a Fedora 12 host.

* mingw32-libvirt.spec.in (__debug_install_post): Override.
(%files): Mention recent additions.
2010-05-07 13:11:25 -06:00
Matthias Bolte
cc04c4ca95 Manually revert "Fix pthread related link error for virt-aa-helper"
This reverts commit b5b8a6db69.

That commit was not necessary. The problem is fixed by commit
0e9b3a269b, but I didn't rebuild
it properly after pulling in the commit and didn't notice it.
2010-05-07 20:07:37 +02:00
Ersek Laszlo
563d7ee3a9 Fix a typo in docs 2010-05-07 19:52:35 +02:00
Eric Blake
48445ccff9 build: use LIBADD, not LDFLAGS, for adding libraries
Per automake, LDFLAGS is used early in the line, and LIBADD
(libraries) or LDADD (programs) is used late.  On platforms like
cygwin, without lazy linking, this order matters.  Therefore, libtool
commands, -L, and similar should be in LDFLAGS, but -l should be in
L*ADD.

* src/Makefile.am (*_LDFLAGS): Move libraries...
(*_LIBADD): ...to their LIBADD counterpart.
2010-05-07 11:12:09 -06:00
Eric Blake
e6125e4938 maint: whitespace cleanups
* src/Makefile.am: Fix some space-tab issues.
2010-05-07 11:11:02 -06:00
Matthias Bolte
b5b8a6db69 Fix pthread related link error for virt-aa-helper
Link virt-aa-helper explicitly with pthread. This is at least
required on Ubuntu 10.04.
2010-05-07 17:52:37 +02:00
Matthias Bolte
ea8d236b73 qemu: Fix warning about a non-literal format string 2010-05-07 17:47:16 +02:00
Jim Meyering
0e9b3a269b avoid link error in tests using libvirt_util; due to pthread_sigmask
* src/Makefile.am (libvirt_util_la_LDFLAGS): Add $(LIB_PTHREAD),
required, now that we're using gnulib's pthread module.
2010-05-07 16:55:42 +02:00
Matthew Booth
34a7f3f6be Remove unused nwfilter field from struct remote_error
Change 965466c1 added a new field to struct remote_error, which broke
the RPC protocol. Fortunately the new field is unused, so this change
simply removes it again.

* src/remote/remote_protocol.(c|h|x): Remove remote_nwfilter from struct
  remote_error
2010-05-07 16:38:05 +02:00
Wolfgang Mauerer
6bc4ea5af0 Implement SCSI disk unplugging
With the introduction of the generic qemu device model, unplugging
SCSI disks works like a charm, so support it in libvirt.

* src/qemu/qemu_driver.c: Add qemudDomainDetachSCSIDiskDevice() to do the
  unplugging, extend qemudDomainDetachDeviceAdd().

Signed-off-by: Wolfgang Mauerer <wolfgang.mauerer@siemens.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2010-05-06 15:55:29 -06:00
Eric Blake
697426e440 qemu: use better types
* src/qemu/qemu_driver.c (qemudFindDisk): Mark parameter const.
(qemudShrinkDisks): Mark parameter unsigned.
2010-05-06 15:32:56 -06:00
Wolfgang Mauerer
dc1a2fcce6 Refactor disk unplugging
We can reuse some of the code for other purposes.

Signed-off-by: Wolfgang Mauerer <wolfgang.mauerer@siemens.com>
2010-05-06 15:19:39 -06:00
Eric Blake
2de2d525fc build: simplify checks for sched.h
* configure.ac: Remove redundant checks.
2010-05-06 14:35:38 -06:00
Eric Blake
062a25f078 build: update gnulib
81 patches to gnulib, picks up several new syntax checks.

* .gnulib: Update to latest.
* .x-sc_prohibit_always_true_header_tests: New file.
* bootstrap.conf (gnulib_modules): Add sched.
* src/util/processinfo.c (includes): <sched.h> is now guaranteed.
* src/network/bridge_driver.c (includes): Drop useless
<strings.h>.
* src/openvz/openvz_conf.c (includes): Likewise.
* src/openvz/openvz_driver.c (includes): Likewise.
* src/phyp/phyp_driver.c (includes): Likewise.
* src/qemu/qemu_driver.c (includes): Likewise.
* src/uml/uml_driver.c (includes): Likewise.
2010-05-06 14:35:38 -06:00
Eric Blake
70e05fa6d0 build: drop more redundant configure checks
* configure.ac (AC_CHECK_FUNCS_ONCE, AC_SYS_LARGEFILE): Rely on
gnulib for strtok_r and large file support.
(AC_OBJEXT): Drop call now done by AC_PROG_CC.
(m4_foreach_w): Drop macro guaranteed by gnulib.
(AC_C_CONST): Drop call declared obsolete by autoconf.

Signed-off-by: Eric Blake <eblake@redhat.com>
2010-05-06 14:35:38 -06:00