Doug Goldstein [Mon, 12 Aug 2013 22:29:40 +0000 (17:29 -0500)]
VMX: Some serial ports are not actually connected
Sometimes a serial port might not be actually wired to a device when the
user does not have the VM powered on and we should not consider this a
fatal error.
Cole Robinson [Thu, 1 Aug 2013 01:40:35 +0000 (21:40 -0400)]
qemu: Support virtio-mmio transport for virtio on ARM
Starting with qemu 1.6, the qemu-system-arm vexpress-a9 model has a
hardcoded virtio-mmio transport which enables attaching all virtio
devices.
On the command line, we have to use virtio-XXX-device rather than
virtio-XXX-pci, thankfully s390 already set the precedent here so
it's fairly straight forward.
At the XML level, this adds a new device address type virtio-mmio.
The controller and addressing don't have any subelements at the
moment because we they aren't needed for this usecase, but could
be added later if needed.
Add a test case for an ARM guest with one of every virtio device
enabled.
Similar to the chardev bit, ARM boards depend on the old style '-net nic'
for actually instantiating net devices. But we can't block out
-netdev altogether since it's needed for upcoming virtio support.
And add tests for working ARM XML with console, disk, and networking.
This corresponds to '-sd' and '-drive if=sd' on the qemu command line.
Needed for many ARM boards which don't provide any other way to
pass in storage.
Cole Robinson [Sat, 17 Aug 2013 19:30:47 +0000 (15:30 -0400)]
qemu: Set QEMU_AUDIO_DRV=none with -nographic
On my machine, a guest fails to boot if it has a sound card, but not
graphical device/display is configured, because pulseaudio fails to
initialize since it can't access $HOME.
A workaround is removing the audio device, however on ARM boards there
isn't any option to do that, so -nographic always fails.
Set QEMU_AUDIO_DRV=none if no <graphics> are configured. Unfortunately
this has massive test suite fallout.
Add a qemu.conf parameter nographics_allow_host_audio, that if enabled
will pass through QEMU_AUDIO_DRV from sysconfig (similar to
vnc_allow_host_audio)
Fred A. Kemp [Fri, 23 Aug 2013 10:38:11 +0000 (12:38 +0200)]
qemu: Support setting the 'removable' flag for USB disks
Add an attribute named 'removable' to the 'target' element of disks,
which controls the removable flag. For instance, on a Linux guest it
controls the value of /sys/block/$dev/removable. This option is only
valid for USB disks (i.e. bus='usb'), and its default value is 'off',
which is the same behaviour as before.
To achieve this, 'removable=on' (or 'off') is appended to the '-device
usb-storage' parameter sent to qemu when adding a USB disk via
'-disk'. A capability flag QEMU_CAPS_USB_STORAGE_REMOVABLE was added
to keep track if this option is supported by the qemu version used.
Bug: https://bugzilla.redhat.com/show_bug.cgi?id=922495 Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Fred A. Kemp [Fri, 23 Aug 2013 10:38:10 +0000 (12:38 +0200)]
qemu: Add capability flag for usb-storage
Allow use of the usb-storage device only if the new capability flag
QEMU_CAPS_DEVICE_USB_STORAGE is set, which it is for qemu(-kvm)
versions >= 0.12.1.2-rhel62-beta.
Doug Goldstein [Wed, 28 Aug 2013 18:37:06 +0000 (13:37 -0500)]
VMX: Create virVMXFormatDisk() from HD and CD-ROM
virVMXFormatHardDisk() and virVMXFormatCDROM() duplicated a lot of code
from each other and made a lot of nested if checks to build each part of
the VMX file. This hopefully simplifies the code path while combining
the two functions with no net difference.
Eric Blake [Sat, 31 Aug 2013 02:16:06 +0000 (20:16 -0600)]
build: fix virtlockd file distribution
Since virtlockd is only built when libvirtd is built, we should
not install its auxiliary files unconditionally. This solves
two failures. 1. 'make distcheck' complains:
rm -f Makefile
ERROR: files left in build directory after distclean:
./src/virtlockd.8
when simulating the case of a user doing a VPATH build from a
read-only source tree. The culprit? BUILT_SOURCES are _always_
built, and so must NOT be built into srcdir and need not be part
of the tarball. On the other hand, shipped files must never
depend on files in the builddir. While it would be possible to
fix the problem by generating aclperms.htmlinc into builddir,
we then have the problem that we ship acl.html - we'd have to
rejigger a lot of things to not ship pre-built html. So this
patch goes the other direction - we don't need BUILT_SOURCES,
but instead ensure that we have proper dependencies so that
all files in srcdir are up-to-date at the time the tarball is
created. And because we ship html files in the tarball, that
implies we don't expect users to be able to rebuild them, so
we must not clean any files that would trigger a rebuild except
under the maintainer rules.
* docs/Makefile.am (BUILT_SOURCES): Delete.
(CLEANFILES): Downgrade aclperms.htmlinc cleanup...
(maintainer-clean-local): ...and move hvsupport.html.in...
(MAINTAINERCLEANFILES): ...to a maintainer action.
(hvsupport.html.in): Write into srcdir.
(hvsupport.html): Ensure files are built in order.
(aclperms.htmlinc): Honor silent make.
(EXTRA_DIST): Ship aclperms.htmlinc.
Eric Blake [Fri, 30 Aug 2013 21:03:52 +0000 (15:03 -0600)]
build: fix 'make distcheck' out of the box
With the 1.1.1 tarball, if a user does 'make && make distcheck',
things pass, but if they do 'make distcheck' after 'make clean',
there is an odd failure:
GEN ../../docs/devhelp/index.html
I/O error : Permission denied
I/O error : Permission denied
runtime error: file ../../docs/devhelp/devhelp.xsl line 43 element document
xsltDocumentElem: unable to save to ../../docs/devhelp/libvirt-virterror.html
I/O error : Permission denied
I/O error : Permission denied
This implies that the rules for 'make dist' are missing a
dependency - the generated documentation needs to be up-to-date
before creating the tarball, or else the tarball will be missing
files, where the end user will end up trying to rebuild files in
srcdir, and that fails when srcdir is read-only.
1.1.1 plus this patch now works without issues (other issues have
crept in to 1.1.2-rc1 that prevent 'make distcheck' from working,
but those will be cleaned up in later patches).
Eric Blake [Fri, 30 Aug 2013 19:58:59 +0000 (13:58 -0600)]
build: only create virt-login-shell for lxc builds
I noticed from an ./autobuild.sh run that we were installing a
virt-login-shell.exe binary when cross-building for mingw,
even though such a binary is necessarily worthless since the
code depends on lxc which is a Linux-only concept.
* tools/Makefile.am (conf_DATA, bin_PROGRAMS, dist_man1_MANS):
Make virt-login-shell installation conditional.
Cole Robinson [Thu, 1 Aug 2013 01:37:40 +0000 (21:37 -0400)]
qemu: Only setup vhost if virtType == "kvm"
vhost only works in KVM mode at the moment, and is infact compiled
out if the emulator is built for non-native architecture. While it
may work at some point in the future for plain qemu, for now it's
just noise on the command line (and which contributes to arm cli
breakage).
Ubuntu libdbus.so links with -Bsymbolic-functions, which means
that we can only LD_PRELOAD functions that we directly call.
Functions which libdbus.so calls internally can not be replaced.
Thus we cannot use dbus_message_new_error or dbus_message_new_method_return
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Eric Blake [Thu, 29 Aug 2013 23:03:34 +0000 (17:03 -0600)]
random: don't mix RAND_MAX with random_r
FreeBSD 10 recently changed their definition of RAND_MAX, to try
and cover the fact that their evenly distributed results of rand()
really are a smaller range than a full power of 2. As a result,
I did some investigation, and learned:
1. POSIX requires random() to be evenly distributed across exactly
31 bits. glibc also guarantees this for rand(), but the two are
unrelated, and POSIX only associates RAND_MAX with rand().
Avoiding RAND_MAX altogether thus avoids a build failure on
FreeBSD 10.
2. Concatenating random bits from a PRNG will NOT provide uniform
coverage over the larger value UNLESS the period of the original
PRNG is at least as large as the number of bits being concatenated.
Simple example: suppose that RAND_MAX were 1 with a period of 2**1
(which means that the PRNG merely alternates between 0 and 1).
Concatenating two successive rand() calls would then invariably
result in 01 or 10, which is a rather non-uniform distribution
(00 and 11 are impossible) and an even worse period (2**0, since
our second attempt will get the same number as our first attempt).
But a RAND_MAX of 1 with a period of 2**2 (alternating between
0, 1, 1, 0) provides sane coverage of all four values, if properly
tempered. (Back-to-back calls would still only see half the values
if we don't do some tempering). We therefore want to guarantee a
period of at least 2**64, preferably larger (as a tempering factor);
POSIX only makes this guarantee for random() with 256 bytes of info.
* src/util/virrandom.c (virRandomBits): Use constants that are
accurate for the PRNG we are using, not an unrelated PRNG.
(randomState): Ensure the period of our PRNG exceeds our usage.
Eric Blake [Fri, 23 Aug 2013 15:30:42 +0000 (09:30 -0600)]
security: provide supplemental groups even when parsing label (CVE-2013-4291)
Commit 29fe5d7 (released in 1.1.1) introduced a latent problem
for any caller of virSecurityManagerSetProcessLabel and where
the domain already had a uid:gid label to be parsed. Such a
setup would collect the list of supplementary groups during
virSecurityManagerPreFork, but then ignores that information,
and thus fails to call setgroups() to adjust the supplementary
groups of the process.
Upstream does not use virSecurityManagerSetProcessLabel for
qemu (it uses virSecurityManagerSetChildProcessLabel instead),
so this problem remained latent until backporting the initial
commit into v0.10.2-maint (commit c061ff5, released in 0.10.2.7),
where virSecurityManagerSetChildProcessLabel has not been
backported. As a result of using a different code path in the
backport, attempts to start a qemu domain that runs as qemu:qemu
will end up with supplementary groups unchanged from the libvirtd
parent process, rather than the desired supplementary groups of
the qemu user. This can lead to failure to start a domain
(typical Fedora setup assigns user 107 'qemu' to both group 107
'qemu' and group 36 'kvm', so a disk image that is only readable
under kvm group rights is locked out). Worse, it is a security
hole (the qemu process will inherit supplemental group rights
from the parent libvirtd process, which means it has access
rights to files owned by group 0 even when such files should
not normally be visible to user qemu).
LXC does not use the DAC security driver, so it is not vulnerable
at this time. Still, it is better to plug the latent hole on
the master branch first, before cherry-picking it to the only
vulnerable branch v0.10.2-maint.
* src/security/security_dac.c (virSecurityDACGetIds): Always populate
groups and ngroups, rather than only when no label is parsed.
The use of <> is a security issue for RPC parameters, since a
malicious client can set a huge array length causing arbitrary
memory allocation in the daemon.
It is also a robustness issue for RPC return values, because if
the stream is corrupted, it can cause the client to also allocate
arbitrary memory.
Use a syntax-check rule to prohibit any use of <>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virConnectListAllSecrets RPC call
The return values for the virConnectListAllSecrets call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virConnectListAllNWFilters RPC call
The return values for the virConnectListAllNWFilters call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virConnectListAllNodeDevices RPC call
The return values for the virConnectListAllNodeDevices call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virConnectListAllInterfaces RPC call
The return values for the virConnectListAllInterfaces call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virConnectListAllNetworks RPC call
The return values for the virConnectListAllNetworks call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virStoragePoolListAllVolumes RPC call
The return values for the virStoragePoolListAllVolumes call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virConnectListAllStoragePools RPC call
The return values for the virConnectListAllStoragePools call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virConnectListAllDomains RPC call
The return values for the virConnectListAllDomains call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virDomain{SnapshotListAllChildren,ListAllSnapshots} RPC calls
The return values for the virDomain{SnapshotListAllChildren,ListAllSnapshots}
calls were not bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virDomainGetJobStats RPC call
The return values for the virDomainGetJobStats call were not
bounds checked. This is a robustness issue for clients if
something where to cause corruption of the RPC stream data.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Add bounds checking on virDomainMigrate*Params RPC calls (CVE-2013-4292)
The parameters for the virDomainMigrate*Params RPC calls were
not bounds checks, meaning a malicious client can cause libvirtd
to consume arbitrary memory
This issue was introduced in the 1.1.0 release of libvirt
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Michal Privoznik [Thu, 29 Aug 2013 11:19:45 +0000 (13:19 +0200)]
autogen.sh: Correctly detect .git as a file
One of my previous patches 5cfe0d37cd0be tried to handle the case when
libvirt is a submodule of another project. In that case, the .git is
just a link to the parent .git directory (which the autogen.sh script
didn't count on). The fix was missing 'test' though.
Peter Krempa [Tue, 27 Aug 2013 17:06:18 +0000 (19:06 +0200)]
qemu: Remove hostdev entry when freeing the depending network entry
When using a <interface type="network"> that points to a network with
hostdev forwarding mode a hostdev alias is created for the network. This
allias is inserted into the hostdev list, but is backed with a part of
the network object that it is connected to.
When a VM is being stopped qemuProcessStop() calls
networkReleaseActualDevice() which eventually frees the memory for the
hostdev object. Afterwards when the domain definition is being freed by
virDomainDefFree() an invalid pointer is accessed by
virDomainHostdevDefFree() and may cause a crash of the daemon.
This patch removes the entry in the hostdev list before freeing the
depending memory to avoid this issue.
Eric Blake [Fri, 16 Aug 2013 22:07:31 +0000 (16:07 -0600)]
virsh: detect programming errors with option parsing
Noticed while reviewing another patch that had an accidental
mismatch due to refactoring. An audit of the code showed that
very few callers of vshCommandOpt were expecting a return of
-2, indicating programmer error, and of those that DID check,
they just propagated that status to yet another caller that
did not check. Fix this by making the code blatantly warn
the programmer, rather than silently ignoring it and possibly
doing the wrong thing downstream.
I know that we frown on assert()/abort() inside libvirtd
(libraries should NEVER kill the program that linked them),
but as virsh is an app rather than the library, and as this
is not the first use of assert() in virsh, I think this
approach is okay.
* tools/virsh.h (vshCommandOpt): Drop declaration.
* tools/virsh.c (vshCommandOpt): Make static, and add a
parameter. Abort on programmer errors rather than making callers
repeat that logic.
(vshCommandOptInt, vshCommandOptUInt, vshCommandOptUL)
(vshCommandOptString, vshCommandOptStringReq)
(vshCommandOptLongLong, vshCommandOptULongLong)
(vshCommandOptBool): Adjust callers.
Jiri Denemark [Wed, 28 Aug 2013 11:50:10 +0000 (13:50 +0200)]
virt-sanlock-cleanup; Fix augtool usage
Surprisingly, augtool get (or print) returns "path = value" while we are
only interested in the value. We need to remove the "path = " part from
the augtool's output. The following is an example of the augtool command
as used in virt-sanlock-cleanup script:
$ augtool get /files/etc/libvirt/qemu-sanlock.conf/disk_lease_dir
/files/etc/libvirt/qemu-sanlock.conf/disk_lease_dir = /var/lib/libvirt/sanlock
Commit a0b6a36f "fixed" what abfff210 broke (URI precedence), but
there was still one more thing missing to fix. When using virsh
parameters to setup debugging, those weren't honored, because at the
time debugging was initializing, arguments weren't parsed yet. To
make ewerything work as expected, we need to initialize the debugging
twice, once before debugging (so we can debug option parsing properly)
and then again after these options are parsed.
As a side effect, this patch also fixes a leak when virsh is ran with
multiple '-l' parameters.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Michal Privoznik [Wed, 28 Aug 2013 07:25:59 +0000 (09:25 +0200)]
virsh-pool.c: Don't jump over variable declaration
Since 785ff34bf8 we are using the outputStr variable in cleanup label.
However, there is a possibility to jump to the label before the variable
has been declared:
virsh-pool.c: In function 'cmdPoolList':
virsh-pool.c:1121:25: error: jump skips variable initialization [-Werror=jump-misses-init]
goto asprintf_failure;
^
virsh-pool.c:1308:1: note: label 'asprintf_failure' defined here
asprintf_failure:
^
virsh-pool.c:1267:11: note: 'outputStr' declared here
char *outputStr = NULL;
Ján Tomko [Tue, 27 Aug 2013 11:47:57 +0000 (13:47 +0200)]
virsh: free the caps list properly if one of them is invalid
VIR_FREE(caps) is not enough to free an array allocated
by vshStringToArray.
==17== 4 bytes in 1 blocks are definitely lost in loss record 4 of 728
==17== by 0x4EFFC44: virStrdup (virstring.c:554)
==17== by 0x128B10: _vshStrdup (virsh.c:125)
==17== by 0x129164: vshStringToArray (virsh.c:218)
==17== by 0x157BB3: cmdNodeListDevices (virsh-nodedev.c:409)
Ján Tomko [Tue, 27 Aug 2013 11:34:09 +0000 (13:34 +0200)]
virsh: free the formatting string when listing pool details
==23== 41 bytes in 1 blocks are definitely lost in loss record 626 of 727
==23== by 0x4F0099F: virAsprintfInternal (virstring.c:358)
==23== by 0x15D2C9: cmdPoolList (virsh-pool.c:1268)
Ján Tomko [Tue, 27 Aug 2013 11:27:50 +0000 (13:27 +0200)]
virsh: free the list from ListAll APIs even for 0 items
virsh secret-list leak when no secrets are defined:
==27== 8 bytes in 1 blocks are definitely lost in loss record 6 of 726
==27== by 0x4E941DD: virAllocN (viralloc.c:183)
==27== by 0x5037F1A: remoteConnectListAllSecrets (remote_driver.c:3076)
==27== by 0x5004EC6: virConnectListAllSecrets (libvirt.c:16298)
==27== by 0x15F813: vshSecretListCollect (virsh-secret.c:397)
==27== by 0x15F0E1: cmdSecretList (virsh-secret.c:532)
Ján Tomko [Tue, 27 Aug 2013 11:07:27 +0000 (13:07 +0200)]
virsh: free messages after logging them to a file
The messages were only freed on error.
==12== 1,100 bytes in 1 blocks are definitely lost in loss record 698 of 729
==12== by 0x4E98C22: virBufferAsprintf (virbuffer.c:294)
==12== by 0x12C950: vshOutputLogFile (virsh.c:2440)
==12== by 0x12880B: vshError (virsh.c:2254)
==12== by 0x131957: vshCommandOptDomainBy (virsh-domain.c:109)
==12== by 0x14253E: cmdStart (virsh-domain.c:3333)
Ján Tomko [Mon, 12 Aug 2013 11:48:34 +0000 (13:48 +0200)]
Build QEMU command line for pcihole64
QEMU commit 3984890 introduced the "pci-hole64-size" property,
to i440FX-pcihost and q35-pcihost with a default setting of 2 GB.
Translate <pcihole64>x<pcihole64/> to:
-global q35-pcihost.pci-hole64-size=x for q35 machines and
-global i440FX-pcihost.pci-hole64-size=x for i440FX-based machines.
Error out on other machine types or if the size was specified
but the pcihost device lacks 'pci-hole64-size' property.
It can be used to adjust (or disable) the size of the 64-bit
PCI hole. The size attribute is in kilobytes (different unit
can be specified on input), but it gets rounded up to
the nearest GB by QEMU.
Disabling it will be needed for guests that crash with the
64-bit PCI hole (like Windows XP), see:
https://bugzilla.redhat.com/show_bug.cgi?id=990418
Ján Tomko [Tue, 20 Aug 2013 15:37:08 +0000 (17:37 +0200)]
Always specify qcow2 compat level on qemu-img command line
qemu-img is going to switch the default for QCOW2
to QCOW2v3 (compat=1.1)
Extend the probing for qemu-img command line options to check
if -o compat is supported. If the volume definition specifies
the qcow2 format but no compat level and -o compat is supported,
specify -o compat=0.10 to create a QCOW2v2 image.
Michal Privoznik [Mon, 26 Aug 2013 14:59:03 +0000 (16:59 +0200)]
qemuDomainAttachHostPciDevice: Fall back to mem balloon if there's no hard_limit
If there's no hard_limit set and domain uses VFIO we still must lock
the guest memory (prerequisite from qemu). Hence, we should compute
the amount to be locked from max_balloon.
Claudio Bley [Thu, 22 Aug 2013 09:16:03 +0000 (11:16 +0200)]
Test for object identity when checking for None in Python
Consistently use "is" or "is not" to compare variables to None,
because doing so is preferrable, as per PEP 8
(http://www.python.org/dev/peps/pep-0008/#programming-recommendations):
> Comparisons to singletons like None should always be done with is or
> is not, never the equality operators.
Set security label on FD for virDomainOpenGraphics
The virDomainOpenGraphics method accepts a UNIX socket FD from
the client app. It must set the label on this FD otherwise QEMU
will be prevented from receiving it with recvmsg.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Michal Privoznik [Thu, 22 Aug 2013 08:19:08 +0000 (10:19 +0200)]
qemuBuildNicDevStr: Add mq=on for multiqueue networking
If user requested multiqueue networking, beside multiple /dev/tap and
/dev/vhost-net openings, we forgot to pass mq=on onto the -device
virtio-net-pci command line. This is advised at:
Peter Krempa [Mon, 19 Aug 2013 14:08:24 +0000 (16:08 +0200)]
virBitmapParse: Fix behavior in case of error and fix up callers
Re-arrange the code so that the returned bitmap is always initialized to
NULL even on early failures and return an error message as some callers
are already expecting it. Fix up the rest not to shadow the error.
Commit abfff210 changed the order of vshParseArgv() and vshInit() in
order to make fix debugging of parameter parsing. However, vshInit()
did a vshReconnect() even though ctl->name wasn't set according to the
'-c' parameter yet. In order to keep both issues fixed, I've split
the vshInit() into vshInitDebug() and vshInit().
One simple memleak of ctl->name is fixed as a part of this patch,
since it is related to the issue it's fixing.
projects / libvirt.git / log