Commit Graph

6675 Commits

Author SHA1 Message Date
Wen Congyang
847efb32de qemu: avoid qemu_driver being unlocked twice when virThreadPoolNew() failed
We do not lock qemu_driver when calling virThreadPoolNew(). If it failed,
we will unlock qemu_driver. It is dangerous.

We may use this pool during auto starting domains. So we must create it before
calling qemuAutostartDomains(). Otherwise, libvirtd will crash.
2011-04-18 09:38:45 +08:00
Richard Laager
cf7124946a Fix two out-of-date comments in LVM backend 2011-04-17 08:37:31 +02:00
Matthias Bolte
12459d1ea7 xen: Replace statsErrorFunc with a macro
Also mark error messages in block_stats.c for translation, add the
new macro to the msg_gen functions in cfg.mk and add block_stats.c
to po/POTFILES.in
2011-04-17 07:53:26 +02:00
Matthias Bolte
60d769a13a Remove virConnectPtr from virRaiseErrorFull
And from all related macros and functions.
2011-04-17 07:22:23 +02:00
Jiri Denemark
b0d28307c8 tests: Unit tests for internal hash APIs
This is a basic set of tests for testing removals of hash entries during
iteration.
2011-04-16 11:31:52 +02:00
Wen Congyang
ffbdf3e144 build: include esx_vi.generated.* into dist file
commit d4601696 introduces two more generated files: esx_vi.generated.h
and esx_vi.generated.h. But we do not include them into dist file.
It will break building if using dist file to build.
2011-04-16 07:09:51 +02:00
Eric Blake
c2d92f6a9e tests: test recent virsh option parsing changes
* tests/virsh-optparse: New file.
* tests/Makefile.am (test_scripts): Use it.
2011-04-15 16:19:22 -06:00
Eric Blake
b9973f526c virsh: fix regression in parsing optional integer
Regression introduced in 0.8.5, commit c1564268.  The command
'virsh freecell 0' quit working when it changed from an optional
string to an optional integer.

This patch introduces a slight change that specifying an option
twice is now detected as an error.  It also changes things so
that a command that has more than 1 required option will not
complain about missing options if one but not all of the options
were given in long format, as in 'virsh vol-create --pool p file',
as well as making positional parsing work for all optional
options (each positional argument is associated with the earliest
option that has not yet been seen by name).

Optional boolean options can appear before required argument
options, because they don't affect positional argument parsing,
and obviously a required boolean option makes no sense.

Technically, this patch renders VSH_OT_STRING and VSH_OT_DATA
redundant; but cleaning that up can be a separate patch.

No command should ever need more than 32 options, right? :)

* tools/virsh.c (vshCmddefGetData, vshCmddefGetOption)
(vshCommandCheckOpts): Alter parameters to use bitmaps.
(vshCmddefOptParse): New function.
(vshCommandParse): Update for better handling of positional
arguments.
(vshCmddefHelp): Allow unit tests to validate options.
2011-04-15 16:08:18 -06:00
Eric Blake
6b75a1a5b0 virsh: list required options first
The current state of virsh parsing is that:

$ virsh vol-info /path/to/image
$ virsh vol-info --pool default /path/to/image
$ virsh vol-info --pool default --vol /path/to/image

all lookup the volume by path (technically, the last two also attempt
a name lookup within a pool, whereas the first skips that step, but
the end result is the same); meanwhile:

$ virsh vol-info default /path/to/image

complains about unexpected data.  Why?  Because the --pool option is
optional, so default was parsed as the --vol argument, and
/path/to/image.img doesn't match up with any remaining options that
require an argument.  For proof, note that:

$ virsh vol-info default --vol /path/to/image

complains about looking up 'default' - the parser mis-associated both
arguments with --vol.  Given the above, the only way to specify pool
is with an explicit "--pool" argument (you can't specify it
positionally).  However, named arguments can appear in any order, so:

$ virsh vol-info /path/to/image --pool default
$ virsh vol-info --vol /path/to/image --pool default

have also always worked.  Therefore, this patch has no functional
change on vol-info option parsing, but only on 'virsh help vol-info'
synopsis layout.  However, it also allows the next patch to 1) enforce
that required options are always first (without this patch, the next
patch would fail the testsuite), and 2) allow the user to omit the
"--pool" argument.  That is, the next patch makes it possible to do:

$ virsh vol-info /path/to/image default

which to date was not possible.

* tools/virsh.c (opts_vol_create_from, opts_vol_clone)
(opts_vol_upload, opts_vol_download, opts_vol_delete)
(opts_vol_wipe, opts_vol_info, opts_vol_dumpxml, opts_vol_key)
(opts_vol_path): List optional pool parameter after required
arguments.
2011-04-15 15:39:53 -06:00
Eric Blake
0bd34a9dec phyp: avoid memory leaks in command values
* src/phyp/phyp_driver.c (phypExecBuffer): New function. Use it
throughout file for less code, and for plugging a few leaks.
2011-04-15 15:32:19 -06:00
Eric Blake
a1b46e714a phyp: use consistent return string handling
Use the name 'ret' for all phypExec results, to make it easier
to wrap phypExec.  Don't allow a possibly NULL ret through printf.

* src/phyp/phyp_driver.c (phypBuildVolume, phypDestroyStoragePool)
(phypBuildStoragePool, phypBuildLpar): Avoid NULL dereference.
(phypInterfaceDestroy): Avoid redundant free.
(phypVolumeLookupByPath, phypVolumeGetPath): Use consistent
naming.
2011-04-15 15:26:27 -06:00
Eric Blake
e00c892f0b phyp: prefer memcpy over memmove when legal
* src/phyp/phyp_driver.c (phypUUIDTable_AddLpar)
(phypGetLparUUID, phypGetStoragePoolUUID, phypVolumeGetXMLDesc)
(phypGetStoragePoolXMLDesc): Use faster method.
2011-04-15 15:26:27 -06:00
Eric Blake
f61785225d phyp: use consistent style for labels
* src/phyp/phyp_driver.c: Match label style of rest of project.
(phypExec, phypUUIDTable_Pull): Drop an extra label.
2011-04-15 15:26:26 -06:00
Eric Blake
444306d591 phyp: more return handling cleanup
* src/phyp/phyp_driver.c (phypInterfaceDestroy)
(phypInterfaceDefineXML, phypInterfaceLookupByName)
(phypInterfaceIsActive, phypListInterfaces, phypNumOfInterfaces):
Clean up return handling of recent additions.
2011-04-15 15:26:26 -06:00
Eric Blake
8f03c6e869 phyp: avoid memory leak on failure
* src/phyp/phyp_driver.c (phypUUIDTable_Init): Avoid memory leak
on error.
2011-04-15 15:26:26 -06:00
Eric Blake
ef6147c40e phyp: avoid a logic bug
Ever since commit ebc46f, the destroy function built two command
variants but only used one.  I went with the variant that matches
the idiom used in the counterpart of phypBuildStoragePool.

* src/phyp/phyp_driver.c (phypDestroyStoragePool): Avoid
clobbering cmd.  Fix error message typo.
2011-04-15 15:26:26 -06:00
Eric Blake
dbe3bad987 maint: use lighter-weight function for straight appends
It costs quite a few processor cycles to go through printf parsing
just to determine that we only meant to append.

* src/xen/xend_internal.c (xend_op_ext): Consolidate multiple
printfs into one.
* src/qemu/qemu_command.c (qemuBuildWatchdogDevStr)
(qemuBuildUSBInputDevStr, qemuBuildSoundDevStr)
(qemuBuildSoundCodecStr, qemuBuildVideoDevStr): Likewise.
(qemuBuildCpuArgStr, qemuBuildCommandLine): Prefer virBufferAdd
over virBufferVsprintf for trivial appends.
* src/phyp/phyp_driver.c (phypExec, phypUUIDTable_Push)
(phypUUIDTable_Pull): Likewise.
* src/conf/nwfilter_conf.c (macProtocolIDFormatter)
(arpOpcodeFormatter, formatIPProtocolID, printStringItems)
(virNWFilterPrintStateMatchFlags, virNWIPAddressFormat)
(virNWFilterDefFormat): Likewise.
* src/security/virt-aa-helper.c (main): Likewise.
* src/util/sexpr.c (sexpr2string): Likewise.
* src/xenxs/xen_sxpr.c (xenFormatSxprChr): Likewise.
* src/xenxs/xen_xm.c (xenFormatXMDisk): Likewise.
2011-04-15 15:26:26 -06:00
Matthias Bolte
1afaafe307 esx: Fix gcc 4.6 warning about initialized but unused variables
This warnings come from partly generated code. Therefore, the best
solution is to mark them as potentially being unused using the
ATTRIBUTE_UNUSED macro. This is suggested by the gcc documentation.

Reported by Christophe Fergeau
2011-04-15 12:31:38 -06:00
Alexander Todorov
d934bd0a58 libvirt-guests: implement START_DELAY
Allow libvirt-guests to stage a delay between guest startups,
to avoid system load caused by back-to-back startup.
2011-04-15 11:15:06 -06:00
Eric Blake
e692352689 maint: silence cppi warnings
* src/nodeinfo.c (linuxNodeInfoCPUPopulate): Fix indentation of
last patch.
2011-04-14 13:33:24 -06:00
Laine Stump
020ad8d1a2 network: truncate bridges' dummy tap device names to IFNAMSIZ (15) chars
This patch addresses:

   https://bugzilla.redhat.com/show_bug.cgi?id=694382

In order to give each libvirt-created bridge a fixed MAC address,
commit 5754dbd56d, added code to create
a dummy tap device with guaranteed lowest MAC address and attach it to
the bridge. This tap device was given the name "${bridgename}-nic".
However, an interface device name must be IFNAMSIZ (15) characters or
less, so a valid ${bridgename} such as "verylongname123" (15
characters) would lead to an invalid tap device name
("verylongname123-nic" - 19 characters), and that in turn led to a
failure to bring up the network.

The solution is to shorten the part of the original name used to
generate the tap device name. However, simply truncating it is
insufficient, because the last few characters of an interface name are
often a number used to indicate one of a list of several similar
devices (for example, "verylongname123", "verylongname124", etc) and
simple truncation would lead to duplicate names (eg "verlongnam-nic"
and "verylongnam-nic"). So instead we take the first 8 characters of
$bridgename ("verylong" in the example), add on the final 3 bytes
("123"), then add "-nic" (so "verylong123-nic").  Not pretty, but it
is much more likely to generate a unique name, and is reproducible
(unlike, say, a random number).
2011-04-14 15:24:17 -04:00
Stefan Berger
d21f9d5a64 ppc: Enable starting of Qemu VMs on ppc host
Due to differences in /proc/cpuinfo the parsing of the cpu data is
different between architectures. On PPC /proc/cpuinfo looks like this:

[original formatting with tabs]

processor    : 0
cpu          : PPC970MP, altivec supported
clock        : 2297.700000MHz
revision     : 1.1 (pvr 0044 0101)

processor    : 1
cpu          : PPC970MP, altivec supported
clock        : 2297.700000MHz
revision     : 1.1 (pvr 0044 0101)

[..]

timebase     : 14318000
platform     : pSeries
model        : IBM,8844-AC1
machine      : CHRP IBM,8844-AC1

The patch adapts the parsing of the data found in /proc/cpuinfo.

/sys/devices/system/cpu/cpuX/topology/physical_package_id also
always returns -1. Check for it on ppc and make it '0' if found negative.
2011-04-14 14:50:22 -04:00
Stefan Berger
cf2145d546 Migrate VMs between different-endianess hosts
This patch enables the migration of Qemu VMs between hosts of different endianess. I tested this by migrating a i686 VM between a x86 and ppc64 host.

I am converting the 'int's in the VM's state header to uint32_t assuming this doesn't break compatibility with existing deployments other than Linux.
2011-04-14 14:48:03 -04:00
Christophe Fergeau
c59f3d8de4 Fix gcc 4.6 warnings in vbox_tmpl.c 2011-04-14 19:23:30 +02:00
Christophe Fergeau
454e50beee Fix gcc 4.6 warnings
gcc 4.6 warns when a variable is initialized but isn't used afterwards:

vmware/vmware_driver.c:449:18: warning: variable 'vmxPath' set but not used [-Wunused-but-set-variable]

This patch fixes these warnings. There are still 2 offending files:

- vbox_tmpl.c: the variable is used inside an #ifdef and is assigned several
  times outside of #ifdef. Fixing the warning would have required wrapping
  all the assignment inside #ifdef which hurts readability.

vbox/vbox_tmpl.c: In function 'vboxAttachDrives':
vbox/vbox_tmpl.c:3918:22: warning: variable 'accessMode' set but not used [-Wunused-but-set-variable]

- esx_vi_types.generated.c: the name implies it's generated code and I
  didn't want to dive into the code generator

esx/esx_vi_types.generated.c: In function 'esxVI_FileQueryFlags_Free':
esx/esx_vi_types.generated.c:1203:3: warning: variable 'item' set but not used [-Wunused-but-set-variable]
2011-04-14 19:09:12 +02:00
Michal Novotny
2ac455c4d2 Introduce virDomainChrDefNew()
Make: passed
Make check: passed
Make syntax-check: passed

this is the commit to introduce the function to create new character
device definition for the domain as advised by Cole Robinson
<crobinso@redhat.com>.

The function is used on the relevant places and also new tests has
been added.

Signed-off-by: Michal Novotny <minovotn@redhat.com>
2011-04-14 10:29:39 -06:00
Michal Privoznik
abb1570eac Spice: support audio, images and stream compression
This extends the SPICE XML to allow variable compression settings for audio,
images and streaming:
    <graphics type='spice' port='5901' tlsPort='-1' autoport='yes'>
        <image compression='auto_glz'/>
        <jpeg compression='auto'/>
        <zlib compression='auto'/>
        <playback compression='on'/>
    </graphics>

All new elements are optional.
2011-04-14 10:23:59 -06:00
Hu Tao
eed9d69e3c free cpumask of vcpupinDef
cpumask doesn't get freed when vcpupinDef being freed, this leaks
memory.
2011-04-14 09:47:10 -06:00
Matthias Bolte
e63dfb8198 esx: Make the parsed URI part of the private connection data
This will be used to make esxVI_Context clonable.

Also move cleanup code for esxPrivate to esxFreePrivate().
2011-04-14 17:39:05 +02:00
Matthias Bolte
991f97d827 esx: Mark error message in macros for translation 2011-04-14 17:35:11 +02:00
Matthias Bolte
d460169610 esx: Extend VI generator to cover managed object types
Generate lookup functions for managed object types.
2011-04-14 17:14:21 +02:00
Matthias Bolte
ac1e6586ec esx: Cleanup and refactor CastFromAnyType macros
Add CastFromAnyType functions for the String type.
2011-04-14 17:03:42 +02:00
Matthias Bolte
a8ad5a40c6 esx: Cleanup VI generator code 2011-04-14 16:56:50 +02:00
Michal Privoznik
2444c411ca network: Fix NULL dereference during error recovery
This fixes: https://bugzilla.redhat.com/show_bug.cgi?id=696660

While starting a network, if brSetForwardDelay() fails, we go to err1
where we want to access macTapIfName variable which was just
VIR_FREE'd a few lines above. Instead, keep macTapIfName until we are
certain of success.
2011-04-14 10:56:17 -04:00
Matthias Bolte
1ef5a3d37e docs: Serial and parallel device target ports actually start from 0
Reported by Igor Galić
2011-04-14 16:36:25 +02:00
Daniel P. Berrange
97263cb115 Add missing checks for QEMU domain state in tunables APIs
The methods qemuDomain{Get,Set}{Memory,Blkio,Scheduler}Parameters
all forgot to do a check on virDomainIsActive(), resulting in bogus
error messages from later parts of their impl

* src/qemu/qemu_driver.c: Add missing checks on virDomainIsActive()
2011-04-14 14:32:34 +01:00
Matthias Bolte
28e938a9ec phyp: Fix too small buffer allocation in phypAttachDevice
sizeof(domain->name) is the wrong thing. Instead of using strdup here
rewrite escape_specialcharacters to allocate the buffer itself.

Add a contains_specialcharacters to be used in phypOpen, as phypOpen is
not interested in the escaped version.
2011-04-14 13:58:22 +02:00
Matthias Bolte
e13e1f4f5d phyp: Don't overwrite error from virDomainDeviceDefParse by OOM error 2011-04-14 13:08:30 +02:00
Matthias Bolte
e69aa73339 phyp: Don't try to use a string from a failed virAsprintf 2011-04-14 13:06:37 +02:00
Matthias Bolte
d765a6f071 phyp: Reduce code duplication in error and success paths
Also fix memory leaks along the way in phypCreateServerSCSIAdapter and
phypAttachDevice.
2011-04-14 13:01:37 +02:00
Matthias Bolte
444fd07a0a phyp: Remove stack allocating a 4kb volume key and fix related memory leaks
Don't pre-allocate 4kb per key, make phypVolumeGetKey allocate the memory.

Make phypBuildVolume return the volume key instead of using pre-allocated
memory to store it.

Also fix a memory leak in phypVolumeLookupByName when phypVolumeGetKey
fails. Fix another memory leak in phypVolumeLookupByPath in the success
path. Fix phypVolumeGetXMLDesc leaking voldef.key.
2011-04-14 12:54:36 +02:00
Daniel P. Berrange
4bfab4a0d3 Remove C99 variable declare in PHYP network driver
Move the virInterfacePtr declaration to the top of the
function to avoid jump uninitialized variable warnings

* src/phyp/phyp_driver.c: Fix var declaration
2011-04-13 19:15:22 +01:00
Daniel P. Berrange
d93ef4466b Replace REMOTE_DEBUG with VIR_DEBUG in daemon dispatcher
The daemon dispatcher code had an obsolete macro

  #define REMOTE_DEBUG(fmt, ...) VIR_DEBUG(fmt, __VA_ARGS__)

This can be trivially removed

* daemon/remote.c: s/REMOTE_DEBUG/VIR_DEBUG/
2011-04-13 19:01:24 +01:00
Daniel P. Berrange
aaa6d7eb39 Add missing checks for whether the connection is open in dispatcher
Many functions did not check for whether a connection was
open. Replace the macro which obscures control flow, with
explicit checks, and ensure all dispatcher code has checks.

* daemon/remote.c: Add connection checks
2011-04-13 19:01:11 +01:00
Daniel P. Berrange
36b652138b Standard on error variable name in libvirtd dispatcher
Some dispatcher methods have a parameter

            remote_error *err,

Instead of the more normal

            remote_error *rerr,

* daemon/remote.c: s/err/rerr/
2011-04-13 19:00:46 +01:00
Daniel P. Berrange
8005048447 Remove all whitespace before function brackets in daemon dispatcher
A lot of code in libvirtd's dispatcher used the style

    dom = get_nonnull_domain (conn, args->dom);

Instead of the normal libvirt style

    dom = get_nonnull_domain(conn, args->dom);

* daemon/remote.c: Remove all whitelist before function brackets
2011-04-13 19:00:21 +01:00
Eduardo Otubo
67eecd1610 PHYP: Adding network interface
This is the implementation of the previous patch now using virInterface*
API. Ended up this patch got much more simpler, smaller and easier to
review. Here is some details:

  * MAC size and interface name are fixed due to specifications on HMC,
    both are created automatically and CAN'T be specified from user. They
    have the following format:

     * MAC: 122980003002
     * Interface name: U9124.720.067BE8B-V3-C0

  * I did replaced all the |grep|sed following the comments Eric Blake
    did on the last patch.

  * According to my last email, It's not possible to create a network
    interface without assigning it to a specific lpar. Then, I am using
    this very minimalistic XML file for testing:

     <interface type='ethernet' name='LPAR01'>
     </interface>

    In this file I am using "name" as the lpar name which I am going to
    assign the new network interface. I couldn't find a better way to
    refer to it. Comments are welcome.

  * Regarding the fact I am sleeping one second waiting for the HMC to
    complete creation of the interface, I don't have means to check
    if the whole process is done. All I do is execute a command, wait
    until is complete (which is not enough in this case) check
    the return and the exit status. The process of actually creating
    a networking interface seems to take a little longer than just the
    return of the ssh control.
2011-04-13 11:18:22 -06:00
Hu Tao
30c551ab4e qemu: fix a dead-lock problem
In qemuDomainObjBeginJobWithDriver, when virCondWaitUntil timeouts,
the function tries to call qemuDriverLock with virDomainObj locked,
this causes the dead-lock problem. This patch fixes this.
2011-04-13 09:43:16 +08:00
Jiri Denemark
5c53160a2f util: Fix crash when removing entries during hash iteration
Commit 9677cd33ee made it possible to
remove current entry when iterating through all hash entries. However,
it didn't properly handle a special case of removing first entry
assigned to a given key which contains several entries in its collision
list.
2011-04-12 19:18:08 +02:00
Michal Privoznik
50e4b9195d Fix possible infinite loop in remote driver
When we take out completed calls from queue we might end up
in circular pointer. We don't want pointer to previous item
point to element taken out.
2011-04-12 09:18:40 -06:00