Jiri Denemark [Sun, 22 May 2011 14:05:07 +0000 (17:05 +0300)]
Replace all remaining setgid/setuid calls with virSetUIDGID
Two additional places need initgroups call to properly work in an
environment where the UID is allowed to open/create stuff through its
supplementary groups.
Michal Privoznik [Mon, 23 May 2011 12:41:00 +0000 (14:41 +0200)]
python: Don't free must-not-free variables
py_str() function call PyString_AsString(). As written in documentation,
the caller must not free the returned value, because it points to some
internal structures.
Alon Levy [Mon, 23 May 2011 15:16:42 +0000 (18:16 +0300)]
spice: support streaming-video parameter
This adds a streaming-video=filter|all|off attribute. It is used to change
the behavior of video stream detection in spice, the default is filter (the
default for libvirt is not to specify it - the actual default is defined in
libspice-server.so).
The capabilities code only sets the flag to allow use of vhost-net if
kvm is detected (set if the help string contains "(qemu-kvm-" or
"(kvm-"), but actually vhost-net is available in some qemu builds that
don't have kvm in their name, so just checking for ",vhost=" is enough.
Doug Goldstein [Fri, 20 May 2011 20:18:09 +0000 (15:18 -0500)]
Use per-user TLS certificates when possible
When using TLS authentication and operating as the non-root user,
initially attempt to use that specific user's TLS certificates before
attempting to use the system wide TLS certificates.
Eric Blake [Thu, 12 May 2011 15:29:51 +0000 (09:29 -0600)]
build: require newer gettext
Now that RHEL 5.6 ships with gettext 0.17, we can get out of the
stone age and use the newer gettext capabilities and improved
macros for certain configure tests.
* configure.ac (AM_GNU_GETTEXT_VERSION): Bump to 0.17.
(MKINSTALLDIRS): Delete hack that is no longer needed.
* bootstrap.conf (buildreq): Check for minimum gettext version.
Based on a report by Wen Congyang.
Don't generate cookies with v2 migration protocol.
The v2 migration protocol had a limit on cookie length that was
too small to be useful for QEMU. Avoid generating cookies with
v2 protocol, so that old libvirtd can still reliably migrate a
guest to new libvirtd uses v2 protocol.
* src/qemu/qemu_driver.c: Avoid migration cookies with v2
migration
Fix QEMU migration cookie crash for guests with no graphics
When generating a cookie for a guest with no data, the
QEMU_MIGRATION_COOKIE_GRAPHICS flag was set even if no
graphics data was added. Avoid setting the flag unless
it was needed, also add a safety check for mig->graphics
being non-NULL
* src/qemu/qemu_migration.c: Avoid cookie crash for guest
with no graphics
Ensure p2p and direct migration use the new v3 protocol if available
The internal virDomainMigratePeer2Peer and virDomainMigrateDirect
helper methods were not checking whether the target supports the
v3 migration protocol.
* src/libvirt.c: Use v3 migration protocol for p2p/direct
migration if available.
Blank out the 'listenAddr' parameter if empty string
Some bogus apps are generating a VNC/SPICE/RFB listen attribute
with no content. This then causes a failure with the graphics
migration cookie parsing. Blank out the 'listenAddr' parameter
after parsing domain XML if it is the empty string, so the host
default takes over
* src/qemu/qemu_migration.c: Blank out listenAddr parameter
if empty
Eric Blake [Wed, 11 May 2011 22:48:35 +0000 (16:48 -0600)]
remote: remove special case for getting version
The on-the-wire protocol is identical; XDR guarantees that
both 'hyper' and 'unsigned hyper' are transmitted as 8 bytes.
* src/remote/remote_protocol.x (remote_get_version_ret)
(remote_get_lib_version_ret): Match public API.
* daemon/remote_generator.pl: Drop special case.
* src/remote_protocol-structs: Reflect updated type.
Eric Blake [Thu, 12 May 2011 01:41:46 +0000 (19:41 -0600)]
build: silence clang false positive
Clang couldn't quite see that the same condition of
(flags & VIR_DOMAIN_MEM_CONFIG) is used twice, such that
the second block is guaranteed that def was assigned in
the first block.
* src/libxl/libxl_driver.c (libxlDomainSetMemoryFlags): Add a hint
for clang.
Eric Blake [Mon, 14 Mar 2011 20:30:24 +0000 (14:30 -0600)]
virsh: optimize creation of default connection
Ramon de Carvalho Valle reported a problem with:
virsh connect qemu:///system
as a non-root user. The real root problem appears to be a regression
in libvirtd being auto-started on the default qemu:///session URI;
however, the symptom points to an independent flaw in virsh - we
shouldn't be wasting efforts on making a connection if we aren't going
to be using that connection. Fixing virsh avoids Ramon's issue, while
I work in the meantime to fix the real libvirtd regression.
This patch looks big, but that's because 'gcc -Wmissing-field-initializers'
gets triggered by './autobuild.sh --enable-compile-warnings=error', so I
had to add 0 initialization to everyone (rather than my preference of
just adding the non-zero flags to virshCmds and to cmdConnect).
Meanwhile, if you use 'virsh -c URI', the connection must succeed; this
patch _only_ optimizes the default connection to be deferred to a later
point where we know if a particular command to be run needs a connection.
* tools/virsh.c (VSH_CMD_FLAG_NOCONNECT): New flag.
(vshCmdDef): Add new flags field.
(vshCommandRun): Honor new flag.
(domManagementCmds, domMonitoringCmds, storagePoolCmds)
(storageVolCmds, networkCmds, nodedevCmds, ifaceCmds)
(nwfilterCmds, secretCmds, virshCmds, snapshotCmds)
(hostAndHypervisorCmds): Populate new field.
(vshReconnect): Don't warn on initial connection.
Matthias Bolte [Wed, 18 May 2011 08:52:57 +0000 (10:52 +0200)]
Clarify the semantic of virDomainGetSchedulerParameters arguments
params and nparams are essential and cannot be NULL. Check this in
libvirt.c and remove redundant checks from the drivers (e.g. xend).
Instead of enforcing that nparams must point to exact same value as
returned by virDomainGetSchedulerType relax this to a lower bound
check. This is what some drivers (e.g. xen hypervisor and esx)
already did. Other drivers (e.g. xend) didn't check nparams at all
and assumed that there is enough space in params.
Unify the behavior in all drivers to a lower bound check and update
nparams to the number of valid values in params on success.
Matthias Bolte [Wed, 18 May 2011 07:55:32 +0000 (09:55 +0200)]
Clarify semantic of nparams argument of virDomainGetSchedulerType
Some drivers assumed it can be NULL (e.g. qemu and lxc) and check it
before assigning to it, other drivers assumed it must be non-NULL
(e.g. test and esx) and just assigned to it.
Unify this to nparams being optional and document it.
Cole Robinson [Thu, 12 May 2011 21:22:29 +0000 (17:22 -0400)]
Fix messages using VIR_ERR_XML_ERROR
This error code has existed since the dawn of time, yet the messages it
generates are almost universally busted. Here's a small sampling:
src/conf/domain_conf.c:4889 : XML description for missing root element is not well formed or invalid
src/conf/domain_conf.c:4951 : XML description for unknown device type is not well formed or invalid
src/conf/domain_conf.c:5460 : XML description for maximum vcpus must be an integer is not well formed or invalid
src/conf/domain_conf.c:5468 : XML description for invalid maxvcpus %(count)lu is not well formed or invalid
Fix up the error code to instead be
XML error: <msg>
Adjust the few locations that were using the original correctly (or shouldn't
have been using the error code at all).
v2:
Fix wording of error code without a passed argument
Eric Blake [Tue, 17 May 2011 17:44:53 +0000 (11:44 -0600)]
libvirt.h: consolidate typed parameter handling
The new type is identical to the three old types that it replaces,
and by creating a common type, this allows future patches to share
common code that manipulates typed parameters.
This change is backwards-compatible in API (recompilation works
without any edits) and ABI (an older client that has not been
recompiled uses the same layout) for code using only public
names; only code using private names (those beginning with _)
will have to adapt.
* include/libvirt/libvirt.h.in (virTypedParameterType)
(VIR_TYPED_PARAM_FIELD_LENGTH, _virTypedParameter): New enum,
macro, and type.
(virSchedParameter, virBlkioParameter, virMemoryParameter):
Rewrite in terms of a common type, while keeping all old public
names for backwards compatibility.
(struct _virSchedParameter, struct _virBlkioParameter)
(struct _virMemoryParameter): Delete - these are private names.
* python/generator.py (enum): Cope with the refactoring.
Add support for 'passthru' mode for direct network interfaces
starting with kernel 2.6.38 macvtap supports a 'passthru' mode for
attaching virtual functions of a SRIOV capable network card directly to a VM.
This patch adds the capability to configure such a device.
Eric Blake [Mon, 16 May 2011 19:27:42 +0000 (13:27 -0600)]
build: drop unused <sys/syslimits.h> header
<sys/syslimits.h> is not standardized, so portable programs should
not need to rely on it. If there really is something that we need
where <sys/syslimits.h> provided the limit but <limits.h> did not,
then that would be a candidate for fixing in gnulib. But this patch
did not turn up any compilation failures on Linux.
* src/internal.h (includes): Drop unused header.
* daemon/libvirtd.h (includes): Likewise.
* configure.ac (AC_CHECK_HEADERS): Likewise.
Based on a report by Matthias Bolte.
Eric Blake [Mon, 16 May 2011 21:37:15 +0000 (15:37 -0600)]
build: tolerate unlimited group size
POSIX allows sysconf(_SC_GETPW_R_SIZE_MAX) to return -1 if there
is no fixed limit, and requires ERANGE errors to track real size.
Model our behavior after the example in POSIX itself:
http://pubs.opengroup.org/onlinepubs/9699919799/functions/getpwuid_r.html
Also, on error for get*_r functions, errno is undefined, and the
real error was the return value.
* src/util/util.c (virGetUserEnt, virGetUserID, virGetGroupID)
(virSetUIDGID): Cope with sysconf failure or too small buffer.
Reported by Matthias Bolte.
Matthias Bolte [Sun, 1 May 2011 19:57:43 +0000 (21:57 +0200)]
esx: Fix race condition in esxVI_EnsureSession
When the session has expired then multiple threads can race while
reestablishing it.
This race condition is not that critical as it requires a special usage
pattern to be triggered. It can only happen when an application doesn't
do API calls for quite some time (the session expires after 30 min
inactivity) and then multiple threads doing simultaneous API calls and
end up doing simultaneous calls to esxVI_EnsureSession.
Jiri Denemark [Mon, 9 May 2011 11:57:09 +0000 (13:57 +0200)]
virsh: Correctly initialize libvirt
virsh didn't call virInitialize(), which (among other things)
initializes virLastErr thread local variable. As a result of that, virsh
could just segfault in virEventRegisterDefaultImpl() since that is the
first call that touches (resets) virLastErr.
I have no idea what lucky coincidence made this bug visible but I was
able to reproduce it in 100% cases but only in one specific environment
which included building in sandbox.
By running the doTunnelSendAll code in a separate thread, the
main thread can do qemuMigrationWaitForCompletion as with
normal migration. This in turn ensures that job signals work
correctly and that progress monitoring can be done
* src/qemu/qemu_migration.c: Run tunnelled migration in
separate thread
Close all sockets before cancelling QEMU migration
Cancelling the QEMU migration may cause QEMU to flush pending
data on the migration socket. This may in turn block QEMU if
nothing reads from the other end of the socket. Closing the
socket before cancelling QEMU migration avoids this possible
deadlock.
* src/qemu/qemu_migration.c: Close sockets before cancelling
migration on failure
Ensure we always read a full buffer in tunnelled migration
The 'nbytes' variable was not re-initialized to the
buffer size on each iteration of the tunnelled migration
loop. While saferead() will ensure a full read, except
on EOF, it is clearer to use the real buffer size
* src/qemu/qemu_migration.c: Always read full buffer of data
Refactor migration completion loop to allow code reuse
The qemuMigrationWaitForCompletion method contains a loop which
repeatedly queries QEMU to check migration progress, and also
processes job signals (pause, setspeed, setbandwidth, cancel).
The tunnelled migration loop does not currently support this
functionality, but should. Refactor the code to allow it to
be used with tunnelled migration.
Implement the v3 migration protocol, which has two extra
steps, 'begin' on the source host and 'confirm' on the
source host. All other methods also gain both input and
output cookies to allow bi-directional data passing at
all stages.
The QEMU peer2peer migration method gains another impl
to provide the v3 migration. This finally allows migration
cookies to work with tunnelled migration, which is required
for Spice seamless migration & the lock manager transfer
* src/qemu/qemu_driver.c: Wire up migrate v3 APIs
* src/qemu/qemu_migration.c, src/qemu/qemu_migration.h: Add
begin & confirm methods, and peer2peer impl of v3
Merge the doNonTunnelMigrate2 and doTunnelMigrate2 methods
into one doPeer2PeerMigrate2 method, since they are substantially
the same. With the introduction of v3 migration, this will be
even more important, to avoid massive code duplication.
To facilitate the introduction of the v3 migration protocol,
the doTunnelMigrate method is refactored into two pieces. One
piece is intended to mirror the flow of virDomainMigrateVersion2,
while the other is the helper for setting up sockets and processing
the data.
Previously socket setup would be done before the 'prepare' step,
so errors could be dealt with immediately, avoiding need to shut
off the destination QEMU. In the new split, socket setup is done
after the 'prepare' step. This is not a serious problem, since
the control flow already requires calling 'finish' to tear down
the destination QEMU upon several errors.
Wire up SPICE client relocation with QEMU migration
Use the graphics information from the QEMU migration cookie to
issue a 'client_migrate_info' monitor command to QEMU. This causes
the SPICE client to automatically reconnect to the target host
when migration completes
* src/qemu/qemu_migration.c: Set data for SPICE client relocation
before starting migration on src
* 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
new qemuMonitorGraphicsRelocate() command
Pass graphics setup from dst back to src via migration cookies
Extend the QEMU migration cookie structure to allow information
about the destination host graphics setup to be passed by to
the source host. This will enable seamless migration of any
connected graphics clients
* src/qemu/qemu_migration.c: Add graphics info to migration
cookies
* daemon/libvirtd.c: Always initialize gnutls to enable
x509 cert parsing in QEMU
The migration protocol has support for a 'cookie' parameter which
is an opaque array of bytes as far as libvirt is concerned. Drivers
may use this for passing around arbitrary extra data they might
need during migration. The QEMU driver needs to do a few things:
- Pass hostname/uuid to allow strict protection against localhost
migration attempts
- Pass SPICE/VNC server port from the target back to the source to
allow seamless relocation of client sessions
- Pass lock driver state from source to destination
This patch introduces the basic glue for handling cookies
but only includes the host/guest UUID & name.
* src/libvirt_private.syms: Export virXMLParseStrHelper
* src/qemu/qemu_migration.c, src/qemu/qemu_migration.h: Parsing
and formatting of migration cookies
* src/qemu/qemu_driver.c: Pass in cookie parameters where possible
* src/remote/remote_protocol.h, src/remote/remote_protocol.x: Change
cookie max length to 16384 bytes
Migration just seems to go from bad to worse. We already had to
introduce a second migration protocol when adding the QEMU driver,
since the one from Xen was insufficiently flexible to cope with
passing the data the QEMU driver required.
It turns out that this protocol still has some flaws that we
need to address. The current sequence is
* Src: DumpXML
- Generate XML to pass to dst
* Dst: Prepare
- Get ready to accept incoming VM
- Generate optional cookie to pass to src
* Src: Perform
- Start migration and wait for send completion
- Kill off VM if successful, resume if failed
* Dst: Finish
- Wait for recv completion and check status
- Kill off VM if unsuccessful
The problems with this are:
- Since the first step is a generic 'DumpXML' call, we can't
add in other migration specific data. eg, we can't include
any VM lease data from lock manager plugins
- Since the first step is a generic 'DumpXML' call, we can't
emit any 'migration begin' event on the source, or have
any hook that runs right at the start of the process
- Since there is no final step on the source, if the Finish
method fails to receive all migration data & has to kill
the VM, then there's no way to resume the original VM
on the source
This patch attempts to introduce a version 3 that uses the
improved 5 step sequence
* Src: Begin
- Generate XML to pass to dst
- Generate optional cookie to pass to dst
* Dst: Prepare
- Get ready to accept incoming VM
- Generate optional cookie to pass to src
* Src: Perform
- Start migration and wait for send completion
- Generate optional cookie to pass to dst
* Dst: Finish
- Wait for recv completion and check status
- Kill off VM if failed, resume if success
- Generate optional cookie to pass to src
* Src: Confirm
- Kill off VM if success, resume if failed
The API is designed to allow both input and output cookies
in all methods where applicable. This lets us pass around
arbitrary extra driver specific data between src & dst during
migration. Combined with the extra 'Begin' method this lets
us pass lease information from source to dst at the start of
migration
Moving the killing of the source VM out of Perform and
into Confirm, means we can now recover if the dst host
can't successfully Finish receiving migration data.
Automatically generate the hvsupport.html.in file from source files
The hvsupport.html.in file is forever out of date. By annotating
the driver struct tables in each driver with version information,
we can auto-generate the hvsupport.html.in file. Annotating the
drivers will be mandatory for new patches, ensuring hvsupport.html.in
is never out of date again.
Convert all driver struct intializers to C99 style
Change all the driver struct initializers to use the
C99 style, leaving out unused fields. This will make
it possible to add new APIs without changing every
driver. eg change:
Jiri Denemark [Wed, 4 May 2011 09:07:01 +0000 (11:07 +0200)]
Implement domain state reason
Only in drivers which use virDomainObj, drivers that query hypervisor
for domain status need to be updated separately in case their hypervisor
supports this functionality.
The reason is also saved into domain state XML so if a domain is not
running (i.e., no state XML exists) the reason will be lost by libvirtd
restart. I think this is an acceptable limitation.