- Devel - libvirt List Archives

[libvirt] [PATCH 0/8 v2] Glue domain and storage
by Osier Yang 08 Apr '13

08 Apr '13

This is the 4th part to implement NPIV migration support [1]. Part 1 and part 2 are pushed. Part 3 (v3: pending for review): https://www.redhat.com/archives/libvir-list/2013-March/msg01440.html ====== This introduces new XMLs to specify the disk source with storage like <disk type='volume' device='disk'> <driver name='qemu' type='raw' cache='none'/> <source pool="$pool_name" volume='$vol_name'/> <seclabel relabel='no'/> </source> <target dev='vdb' bus='virtio'/> </disk> And before domain starting, and disk attaching/updating, the source represented by storage is translated into the real underlying storage volume. Future work: * Support network volume * Support disk backing chain? v1 - v2: * Invoke storage APIs to translate the source directly in qemu driver * 1/8 in v1 is pushed. * 2/8 in v1 is splitted (the code refactoring now is a standalone patch) * Support shared disk for volume type disk * Support sgio setting for volume type disk * No network volume support in v2 RFC - v1: * The XMLs are more simpler - only using pool name and volume name to specify disk source. * Support network pool (rbd, and sheepdog) * Support startupPolicy for volume type disk * Support seclabels for volume type disk * Fix bugs on disk source formating Osier Yang (8): conf: New helper virDomainDiskSourceDefFormat to format the disk source Introduce new XMLs to specify disk source using libvirt storage qemu: Translate the pool disk source when building drive string Support startupPolicy for 'volume' disk Support seclabels for volume type disk qemu: Translate the pool disk source earlier qemu: Support shareable volume type disk qemu: Support sgio setting for volume type disk docs/formatdomain.html.in | 28 ++- docs/schemas/domaincommon.rng | 24 ++ src/conf/domain_conf.c | 274 ++++++++++++++------- src/conf/domain_conf.h | 10 + src/qemu/qemu_command.c | 32 +++ src/qemu/qemu_command.h | 1 - src/qemu/qemu_conf.c | 68 ++++- src/qemu/qemu_conf.h | 3 + src/qemu/qemu_domain.c | 4 +- src/qemu/qemu_driver.c | 16 +- src/qemu/qemu_process.c | 14 +- .../qemuxml2argv-disk-source-pool.xml | 37 +++ tests/qemuxml2xmltest.c | 1 + 13 files changed, 406 insertions(+), 106 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-source-pool.xml -- 1.8.1.4

4 34

[libvirt] [PATCH 0/7 v3] Persistent vHBA support in storage driver
by Osier Yang 08 Apr '13

08 Apr '13

v2: https://www.redhat.com/archives/libvir-list/2013-February/msg00118.html v2 - v3: * 6/8 of v1 is ACKed/pushed. * Use virStringToLong_ui instead of sscanf * Fixed various nits pointed out by John This is the 3rd part to implement NPIV migration support [1]. Part 1: (pushed) https://www.redhat.com/archives/libvir-list/2013-February/msg00112.html Part 2: (pushed) https://www.redhat.com/archives/libvir-list/2013-January/msg00859.html Osier Yang (7): New XML attributes for storage pool source adapter storage: Make the adapter name be consistent with node device driver storage: Move virStorageBackendSCSIGetHostNumber into iscsi backend phyp: Prohibit fc_host adapter for phyp driver util: Add helper to get the scsi host name by iterating over sysfs storage: Add startPool and stopPool for scsi backend storage: Guess the parent if it's not specified for vHBA docs/formatstorage.html.in | 17 +- docs/schemas/storagepool.rng | 33 +++- src/conf/storage_conf.c | 132 +++++++++++++- src/conf/storage_conf.h | 23 ++- src/libvirt_private.syms | 4 + src/phyp/phyp_driver.c | 15 +- src/storage/storage_backend_iscsi.c | 39 +++- src/storage/storage_backend_scsi.c | 202 +++++++++++++++------ src/storage/storage_backend_scsi.h | 3 - src/util/virutil.c | 202 +++++++++++++++++++++ src/util/virutil.h | 11 +- .../pool-scsi-type-fc-host.xml | 15 ++ .../pool-scsi-type-scsi-host.xml | 15 ++ .../pool-scsi-type-fc-host.xml | 18 ++ .../pool-scsi-type-scsi-host.xml | 18 ++ tests/storagepoolxml2xmlout/pool-scsi.xml | 2 +- tests/storagepoolxml2xmltest.c | 2 + 17 files changed, 669 insertions(+), 82 deletions(-) create mode 100644 tests/storagepoolxml2xmlin/pool-scsi-type-fc-host.xml create mode 100644 tests/storagepoolxml2xmlin/pool-scsi-type-scsi-host.xml create mode 100644 tests/storagepoolxml2xmlout/pool-scsi-type-fc-host.xml create mode 100644 tests/storagepoolxml2xmlout/pool-scsi-type-scsi-host.xml -- 1.8.0.1

3 23

[libvirt] [PATCH 1/2] cgroup: export virCgroupRemoveRecursively
by Gao feng 08 Apr '13

08 Apr '13

We will use virCgroupRemoveRecursively to remove cgroup directories in the coming patch. Signed-off-by: Gao feng <gaofeng(a)cn.fujitsu.com> --- src/libvirt_private.syms | 1 + src/util/vircgroup.c | 4 ++-- src/util/vircgroup.h | 1 + 3 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index dc01bfa..8cc50c4 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1118,6 +1118,7 @@ virCgroupKillRecursive; virCgroupMounted; virCgroupMoveTask; virCgroupPathOfController; +virCgroupRemoveRecursively; virCgroupRemove; virCgroupSetBlkioDeviceWeight; virCgroupSetBlkioWeight; diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index 532e704..6998f13 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -686,7 +686,7 @@ cleanup: #endif #if defined _DIRENT_HAVE_D_TYPE -static int virCgroupRemoveRecursively(char *grppath) +int virCgroupRemoveRecursively(char *grppath) { DIR *grpdir; struct dirent *ent; @@ -735,7 +735,7 @@ static int virCgroupRemoveRecursively(char *grppath) return rc; } #else -static int virCgroupRemoveRecursively(char *grppath ATTRIBUTE_UNUSED) +int virCgroupRemoveRecursively(char *grppath ATTRIBUTE_UNUSED) { /* Claim no support */ return -ENXIO; diff --git a/src/util/vircgroup.h b/src/util/vircgroup.h index 2ed6ff9..ea42fa2 100644 --- a/src/util/vircgroup.h +++ b/src/util/vircgroup.h @@ -157,6 +157,7 @@ int virCgroupGetCpusetMems(virCgroupPtr group, char **mems); int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus); int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus); +int virCgroupRemoveRecursively(char *grppath); int virCgroupRemove(virCgroupPtr group); void virCgroupFree(virCgroupPtr *group); -- 1.7.11.7

5 10

[libvirt] [PATCH v3 1/2] Optimize machine option to set more options with it
by Li Zhang 07 Apr '13

07 Apr '13

From: Li Zhang <zhlcindy(a)linux.vnet.ibm.com> Currently, -machine option is used only when dump-guest-core is set. To use options defined in machine option for newer version of QEMU, it needs to use -machine xxx, and to be compatible with older version -M, this patch addes QEMU_CAPS_MACHINE_OPT capability for newer version which supports -machine option. Signed-off-by: Li Zhang <zhlcindy(a)linux.vnet.ibm.com> --- v3 -> v2: * Set QEMU_CAPS_MACHINE_OPT with help string * Set QEMU_CAPS_DUMP_GUEST_CORE with QMP v2 -> v1: * Split the patch to 2 parts suggested by Daniel P.Berrange * Rename QEMU_CAPS_MACH_OPT to QEMU_CAPS_MACHINE_OPT * Remove version 1.1 assertion for QEMU_CAPS_MACHINE_OPT src/qemu/qemu_capabilities.c | 11 +++++++++++ src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 32 ++++++++++++++++++-------------- tests/qemuhelptest.c | 4 ++++ tests/qemuxml2argvtest.c | 6 +++--- 5 files changed, 37 insertions(+), 17 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 861d3c4..7d459db 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -213,6 +213,8 @@ VIR_ENUM_IMPL(virQEMUCaps, QEMU_CAPS_LAST, "virtio-ccw", "dtb", "megasas", + + "machine-opt", /* 135 */ ); struct _virQEMUCaps { @@ -1091,6 +1093,9 @@ virQEMUCapsComputeCmdFlags(const char *help, if (strstr(help, "-dtb")) virQEMUCapsSet(qemuCaps, QEMU_CAPS_DTB); + if (strstr(help, "-machine")) + virQEMUCapsSet(qemuCaps, QEMU_CAPS_MACHINE_OPT); + /* * Handling of -incoming arg with varying features * -incoming tcp (kvm >= 79, qemu >= 0.10.0) @@ -2442,6 +2447,12 @@ virQEMUCapsInitQMP(virQEMUCapsPtr qemuCaps, virQEMUCapsInitQMPBasic(qemuCaps); + /* machine option is supported for newer version */ + virQEMUCapsSet(qemuCaps, QEMU_CAPS_MACHINE_OPT); + + /* -dump-guest-core is supported for newer version*/ + virQEMUCapsSet(qemuCaps, QEMU_CAPS_DUMP_GUEST_CORE); + if (!(archstr = qemuMonitorGetTargetArch(mon))) goto cleanup; diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index 7101f67..2595868 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -174,6 +174,7 @@ enum virQEMUCapsFlags { QEMU_CAPS_VIRTIO_CCW = 132, /* -device virtio-*-ccw */ QEMU_CAPS_DTB = 133, /* -dtb file */ QEMU_CAPS_SCSI_MEGASAS = 134, /* -device megasas */ + QEMU_CAPS_MACHINE_OPT = 135, /* -machine xxxx*/ QEMU_CAPS_LAST, /* this must always be the last item */ }; diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index a0c278f..0ee634b 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -5197,6 +5197,8 @@ qemuBuildMachineArgStr(virCommandPtr cmd, const virDomainDefPtr def, virQEMUCapsPtr qemuCaps) { + virBuffer buf = VIR_BUFFER_INITIALIZER; + /* This should *never* be NULL, since we always provide * a machine in the capabilities data for QEMU. So this * check is just here as a safety in case the unexpected @@ -5204,27 +5206,29 @@ qemuBuildMachineArgStr(virCommandPtr cmd, if (!def->os.machine) return 0; - if (!def->mem.dump_core) { + if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_MACHINE_OPT)) { /* if no parameter to the machine type is needed, we still use * '-M' to keep the most of the compatibility with older versions. */ virCommandAddArgList(cmd, "-M", def->os.machine, NULL); } else { - if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_DUMP_GUEST_CORE)) { - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - "%s", _("dump-guest-core is not available " - " with this QEMU binary")); - return -1; - } - /* However, in case there is a parameter to be added, we need to - * use the "-machine" parameter because qemu is not parsing the - * "-M" correctly */ virCommandAddArg(cmd, "-machine"); - virCommandAddArgFormat(cmd, - "%s,dump-guest-core=%s", - def->os.machine, - virDomainMemDumpTypeToString(def->mem.dump_core)); + virBufferAsprintf(&buf, "%s", def->os.machine); + + if (def->mem.dump_core) { + if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_DUMP_GUEST_CORE)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + "%s", _("dump-guest-core is not available " + " with this QEMU binary")); + return -1; + } + + virBufferAsprintf(&buf, ",dump-guest-core=%s", + virDomainMemDumpTypeToString(def->mem.dump_core)); + } + + virCommandAddArg(cmd, virBufferContentAndReset(&buf)); } return 0; diff --git a/tests/qemuhelptest.c b/tests/qemuhelptest.c index a28109a..0c7e607 100644 --- a/tests/qemuhelptest.c +++ b/tests/qemuhelptest.c @@ -718,6 +718,7 @@ mymain(void) QEMU_CAPS_SCSI_LSI, QEMU_CAPS_BLOCKIO, QEMU_CAPS_VNC, + QEMU_CAPS_MACHINE_OPT, QEMU_CAPS_DEVICE_QXL, QEMU_CAPS_DEVICE_VGA, QEMU_CAPS_DEVICE_CIRRUS_VGA, @@ -806,6 +807,7 @@ mymain(void) QEMU_CAPS_VIRTIO_SCSI, QEMU_CAPS_BLOCKIO, QEMU_CAPS_VNC, + QEMU_CAPS_MACHINE_OPT, QEMU_CAPS_DEVICE_QXL, QEMU_CAPS_DEVICE_VGA, QEMU_CAPS_DEVICE_CIRRUS_VGA, @@ -903,6 +905,7 @@ mymain(void) QEMU_CAPS_SECCOMP_SANDBOX, QEMU_CAPS_DUMP_GUEST_CORE, QEMU_CAPS_VNC, + QEMU_CAPS_MACHINE_OPT, QEMU_CAPS_USB_REDIR_BOOTINDEX, QEMU_CAPS_USB_HOST_BOOTINDEX, QEMU_CAPS_DEVICE_QXL, @@ -1009,6 +1012,7 @@ mymain(void) QEMU_CAPS_SECCOMP_SANDBOX, QEMU_CAPS_DUMP_GUEST_CORE, QEMU_CAPS_VNC, + QEMU_CAPS_MACHINE_OPT, QEMU_CAPS_USB_REDIR_BOOTINDEX, QEMU_CAPS_USB_HOST_BOOTINDEX, QEMU_CAPS_DEVICE_QXL, diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 38787ac..4b45c55 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -363,9 +363,9 @@ mymain(void) DO_TEST("minimal-s390", QEMU_CAPS_NAME); DO_TEST("machine-aliases1", NONE); DO_TEST("machine-aliases2", QEMU_CAPS_KVM); - DO_TEST("machine-core-on", QEMU_CAPS_DUMP_GUEST_CORE); - DO_TEST("machine-core-off", QEMU_CAPS_DUMP_GUEST_CORE); - DO_TEST_FAILURE("machine-core-on", NONE); + DO_TEST("machine-core-on", QEMU_CAPS_MACHINE_OPT, QEMU_CAPS_DUMP_GUEST_CORE); + DO_TEST("machine-core-off", QEMU_CAPS_MACHINE_OPT, QEMU_CAPS_DUMP_GUEST_CORE); + DO_TEST_FAILURE("machine-core-on", QEMU_CAPS_MACHINE_OPT, NONE); DO_TEST("boot-cdrom", NONE); DO_TEST("boot-network", NONE); DO_TEST("boot-floppy", NONE); -- 1.7.10.1

4 10

[libvirt] RFC: dhcp <option> element revisited
by Laine Stump 07 Apr '13

07 Apr '13

Before 1.0.3, Pieter Hollants kindly implemented the XML I had earlier suggested to support configure-by-number dhcp options in libvirt networks. https://www.redhat.com/archives/libvir-list/2013-February/msg01251.html I also posted a followup patch for a "force" attribute to options: https://www.redhat.com/archives/libvir-list/2013-February/msg01349.html I actually pushed a slightly fixed version of Pieter's patch, and was looking for an ACK to mine when Eric questioned whether it was all really ready to push since we hadn't implemented *named* options (although there is no official RFC standard for the names of options, dnsmasq uses a pretty good set of lowercase-dash-separated names that make sense, and would certainly make it easier to decode the config). At the time I had just started trying to write such a patch, and was realizing that the XML we'd come up with didn't have any provision to easily deal with an option that took a list of values (multiple IP addresses, multiple domain names, etc); for numbered options we could just punt on this and say "put the multiple values in the single 'value' attribute, separated by commas", but we all know that's a cop-out, so I decided to temporarily pull Pieter's patch to avoid regrets over getting something suboptimal in an official release. That's all the history. Now on to discussing how we should modify these patches to be exactly what we want. I tried digging down into the data of the options to try and make an exhaustive XML representation, with the results below (these would all go directly inside the <dhcp> element of a libvirt network). But it's starting to seem like the deeper I go, the deeper it gets (the classless-static-route option is the current "deepest", and I haven't even attempted to do anything about vendor-specific options), so I'm starting to wonder where to draw the line - see the "STEPPING BACK" section at the bottom. 1) numeric option, single value: <option code='119'> <value type='domain' data='a.example.com'/> </option> (Note that I changed "number" to "code" after seeing that suggestion in [1] because that's what it's called in the RFC). (Also note that the actual encoding of option 119's data is more complicated than described here [2], but it's complicated enough that I think whatever dhcp server implementation is underneath libvirt will need to know the details anyway and decided what we should go for with the "type" attribute is enough for libvirt to validate the input, not necessarily to encode it into a DHCP response packet; any translation from the validated input will be handled internally to the backend driver (currently bridge_driver.c which uses dnsmasq) (I'm actually wondering if type is needed in the XML at all, since the type of any piece of data is always either implicit, or unknown (in which case we just accept any text and pass it through to the backend driver).) 2) numeric option, multiple values: <option code='119' force='yes'> <value type='domain' data='a.example.com'/> <value type='domain' data='b.example.com'/> <value type='domain' data='c.example.com'/> </option> 3) named option, single simple value: <option name='ip-forward-enable'> <value type='boolean' data='yes'/> </option> 4) named option, multiple simple values: <option name='dns-server'> <value type='ipv4Addr' data='10.1.1.1'/> <value type='ipv4Addr' data='10.1.1.2'/> </option> <option (for all named options and known numeric options, the "type" field would be optional, as libvirt will already have an internal table associating option names/code with the type of data, as well as whether or not multiples are allowed) 5) named option, multiple compound values: <option name='classless-static-route'> <value type='compound'> <value type='ipv4Addr' data='1.2.3.0'/> <value type='ipv4Prefix' data='24'/> <value type='ipv4Addr' data='192.168.122.5'/> </value> <value type='compound'> <value type='ipv4Addr' data='1.2.4.0'/> <value type='ipv4Prefix' data='24'/> <value type='ipv4Addr' data='192,.168.122.6'/> </value> </option> (I suppose this could also be done as a flat list, since every static route needs all three) The types I can see that we need (to describe everything in RFC2132 (except vendor-specific options) and the other options I know about) are: ipv4Addr ipv4Prefix text path domain boolean number ANOTHER POSSIBILITY =================== Another way of dealing with options that permit lists of multiple values - just specify the option multiple times, e.g.: <option name='dnsServer' type='ipv4Addr' data='10.1.1.1'/> <option name='dnsServer' type='ipv4Addr' data='10.1.1.2'/> ============================ BTW - another point; I had earlier said that we didn't need to worry about an option number blacklist/whitelist. After looking through the options in RFC 2132, I've changed my mind - many of the option codes (particularly those in the 50's) aren't really general purpose options, but are used by dhcp as integral parts of implementing the exchange sequence between the server and client. So I think what we need to do is have a whitelist that has an entry for every dhcp option we are willing to accept. It would contain: uint8 code; const char *name; /* could be "", in which case we don't recognize a name for it */ unsigned int flags; /* whether or not multiple values are allowed, name recognized by dnsmasq */ /* some representation of type here - what to do about the "compound" type? */ STEPPING BACK ------------- So are we going too far with this? Or should we dial it back to the original proposal in Pieter's patch and just say that 1) if <option> is used, the data is passed through unqualified to the dhcp implementation 2) only "simple", known option types with a single no-repeating value can be specified this way (and data type is implied)(this would be enforced with the "whitelist") 3) then implement separate elements for the complicated/repeating options, e.g.: <route destination='1.2.3.0' prefix='24' nextHop='192.168.122.5'/> <route destination='1.2.4.0' prefix='24' nextHop='192.168.122.6'/> ... <dnsServer address='10.1.1.1'/> <dnsServer address='10.1.1.2'/> Or maybe to simplify the parser this would work as: <option name='classless-static-route' destination='1.2.3.0' prefix='24' nextHop='192.168.122.5'/> <option name='classless-static-route' destination='1.2.4.0' prefix='24' nextHop='192.168.122.6'/> ... <option name='dns-server address='10.1.1.1'/> <option name='dns-server address='10.1.1.2'/> Either of these definitely seems more compact and easy to understand than all that ranting and raving at the top of this message... Any other suggestions? === [1] https://bugzilla.redhat.com/show_bug.cgi?id=666556 [2] https://tools.ietf.org/rfc/rfc3397.txt

2 2

[libvirt] Patch set to add InteractiveContainer support to virt-sandbox
by Dan Walsh 06 Apr '13

06 Apr '13

Resend of patches with lots of cleanups. This patch set adds InteractiveContainer support to virt-sandbox-service command. This is needed if we want to support the OpenShift model of containers. There are also some bug fix patches in the set. [sandbox PATCH 01/15] virt-sandbox-service-util needs to free [sandbox PATCH 02/15] Add support for InteractiveContainers to [sandbox PATCH 03/15] Move virt-sandbox-service bash completion [sandbox PATCH 04/15] Add -u UNITFILE option to virt-sandbox-service [sandbox PATCH 05/15] Change virt-sandbox-service-create.pod to use [sandbox PATCH 06/15] Internationalize all output strings in [sandbox PATCH 07/15] Remove distinction from Internal vs External [sandbox PATCH 08/15] Make CONFIG_PATH external to the Container [sandbox PATCH 09/15] Add exception handler GlibGerror to [sandbox PATCH 10/15] Change variable config to config_path to avoid [sandbox PATCH 11/15] Refactor Container class into Container and [sandbox PATCH 12/15] Add support for InteractiveContainer [sandbox PATCH 13/15] Use args.uri rather then hard coding lxc:/// [sandbox PATCH 14/15] Check for LXC if virt-sandbox-service execute [sandbox PATCH 15/15] Create new /etc/rc.d directory to bind mount

3 31

[libvirt] [PATCH v3] util: Add a helper to check if all bits of a bitmap are clear
by Osier Yang 06 Apr '13

06 Apr '13

--- src/libvirt_private.syms | 1 + src/util/virbitmap.c | 17 +++++++++++++++++ src/util/virbitmap.h | 3 +++ tests/virbitmaptest.c | 33 ++++++++++++++++++++++++++++++++- 4 files changed, 53 insertions(+), 1 deletion(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index b7b9631..6b831b3 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1046,6 +1046,7 @@ virBitmapEqual; virBitmapFormat; virBitmapFree; virBitmapGetBit; +virBitmapIsAllClear; virBitmapIsAllSet; virBitmapNew; virBitmapNewCopy; diff --git a/src/util/virbitmap.c b/src/util/virbitmap.c index 21509ac..998c302 100644 --- a/src/util/virbitmap.c +++ b/src/util/virbitmap.c @@ -591,6 +591,23 @@ bool virBitmapIsAllSet(virBitmapPtr bitmap) } /** + * virBitmapIsAllClear: + * @bitmap: the bitmap to check + * + * check if all bits in @bitmap are clear + */ +bool virBitmapIsAllClear(virBitmapPtr bitmap) +{ + int i; + + for (i = 0; i < bitmap->map_len; i++) + if (bitmap->map[i] != 0) + return false; + + return true; +} + +/** * virBitmapNextSetBit: * @bitmap: the bitmap * @pos: the position after which to search for a set bit diff --git a/src/util/virbitmap.h b/src/util/virbitmap.h index 044c7a6..b682523 100644 --- a/src/util/virbitmap.h +++ b/src/util/virbitmap.h @@ -100,6 +100,9 @@ void virBitmapClearAll(virBitmapPtr bitmap) bool virBitmapIsAllSet(virBitmapPtr bitmap) ATTRIBUTE_NONNULL(1); +bool virBitmapIsAllClear(virBitmapPtr bitmap) + ATTRIBUTE_NONNULL(1); + ssize_t virBitmapNextSetBit(virBitmapPtr bitmap, ssize_t pos) ATTRIBUTE_NONNULL(1); diff --git a/tests/virbitmaptest.c b/tests/virbitmaptest.c index 95d010a..3067f6e 100644 --- a/tests/virbitmaptest.c +++ b/tests/virbitmaptest.c @@ -125,6 +125,8 @@ static int test2(const void *data ATTRIBUTE_UNUSED) goto error; virBitmapClearAll(bitmap); + if (!virBitmapIsAllClear(bitmap)) + goto error; if (testBit(bitmap, 0, size - 1, false) < 0) goto error; if (virBitmapCountBits(bitmap) != 0) @@ -154,6 +156,9 @@ static int test3(const void *data ATTRIBUTE_UNUSED) if (!virBitmapIsAllSet(bitmap)) goto error; + virBitmapClearAll(bitmap); + if (!virBitmapIsAllClear(bitmap)) + goto error; ret = 0; error: @@ -196,6 +201,9 @@ static int test4(const void *data ATTRIBUTE_UNUSED) if (virBitmapNextClearBit(bitmap, i) != -1) goto error; + if (!virBitmapIsAllClear(bitmap)) + goto error; + virBitmapFree(bitmap); bitmap = NULL; @@ -406,6 +414,10 @@ static int test7(const void *v ATTRIBUTE_UNUSED) if (!virBitmapIsAllSet(bitmap)) goto error; + virBitmapClearAll(bitmap); + if (!virBitmapIsAllClear(bitmap)) + goto error; + virBitmapFree(bitmap); } @@ -416,6 +428,24 @@ error: return -1; } +static int test8(const void *v ATTRIBUTE_UNUSED) +{ + virBitmapPtr bitmap = NULL; + char data[108] = {0x00,}; + + bitmap = virBitmapNewData(data, sizeof(data)); + if (!bitmap) + goto error; + + if (!virBitmapIsAllClear(bitmap)) + goto error; + + return 0; +error: + virBitmapFree(bitmap); + return -1; +} + static int mymain(void) { @@ -435,7 +465,8 @@ mymain(void) ret = -1; if (virtTestRun("test7", 1, test7, NULL) < 0) ret = -1; - + if (virtTestRun("test8", 1, test8, NULL) < 0) + ret = -1; return ret; } -- 1.8.1.4

2 2

[libvirt] [PATCH v2] qemu: Support multiple queue virtio-scsi
by Osier Yang 06 Apr '13

06 Apr '13

This introduce a new attribute "num_queues" (same with the good name QEMU uses) for virtio-scsi controller. An example of the XML: <controller type='scsi' index='0' model='virtio-scsi' num_queues='8'/> The corresponding QEMU command line: -device virtio-scsi-pci,id=scsi0,num_queues=8,bus=pci.0,addr=0x3 \ --- docs/formatdomain.html.in | 18 +++++++++------ docs/schemas/domaincommon.rng | 5 +++++ src/conf/domain_conf.c | 13 +++++++++++ src/conf/domain_conf.h | 1 + src/qemu/qemu_command.c | 11 +++++++++ .../qemuxml2argv-disk-virtio-scsi-num_queues.args | 8 +++++++ .../qemuxml2argv-disk-virtio-scsi-num_queues.xml | 26 ++++++++++++++++++++++ tests/qemuxml2argvtest.c | 3 +++ tests/qemuxml2xmltest.c | 1 + 9 files changed, 79 insertions(+), 7 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index cf382e8..2973ce2 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -2092,13 +2092,17 @@ controller. A "scsi" controller has an optional attribute <code>model</code>, which is one of "auto", "buslogic", "ibmvscsi", "lsilogic", "lsisas1068", "lsisas1078", "virtio-scsi" or - "vmpvscsi". A "usb" controller has an optional attribute - <code>model</code>, which is one of "piix3-uhci", "piix4-uhci", "ehci", - "ich9-ehci1", "ich9-uhci1", "ich9-uhci2", "ich9-uhci3", - "vt82c686b-uhci", "pci-ohci" or "nec-xhci". Additionally, - since 0.10.0, if the USB bus needs to be - explicitly disabled for the guest, <code>model='none'</code> may be used. - The PowerPC64 "spapr-vio" addresses do not have an associated controller. + "vmpvscsi". The attribute <code>num_queues</code> + (1.0.5 (QEMU and KVM only)) is to specify + the number of queues for the controller, it's recommended to specify + the value matching the number of vCPUs for best performance. A "usb" + controller has an optional attribute <code>model</code>, which is one + of "piix3-uhci", "piix4-uhci", "ehci", "ich9-ehci1", "ich9-uhci1", + "ich9-uhci2", "ich9-uhci3", "vt82c686b-uhci", "pci-ohci" or "nec-xhci". + Additionally, since 0.10.0, if the USB bus + needs to be explicitly disabled for the guest, <code>model='none'</code> + may be used. The PowerPC64 "spapr-vio" addresses do not have an + associated controller. diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng index 63ba7d1..d486ae8 100644 --- a/docs/schemas/domaincommon.rng +++ b/docs/schemas/domaincommon.rng @@ -1427,6 +1427,11 @@ </attribute> </optional> <optional> + <attribute name="num_queues"> + <ref name="unsignedInt"/> + </attribute> + </optional> + <optional> <ref name="usbmaster"/> </optional> <optional> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 5ad3080..f50a964 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -5034,6 +5034,7 @@ virDomainControllerDefParseXML(xmlNodePtr node, char *type = NULL; char *idx = NULL; char *model = NULL; + char *num_queues = NULL; if (VIR_ALLOC(def) < 0) { virReportOOMError(); @@ -5069,6 +5070,14 @@ virDomainControllerDefParseXML(xmlNodePtr node, def->model = -1; } + if ((num_queues = virXMLPropString(node, "num_queues"))) { + if (virStrToLong_ui(num_queues, NULL, 10, &def->num_queues) < 0) { + virReportError(VIR_ERR_XML_ERROR, + _("Malformed 'num_queues' value '%s'"), num_queues); + goto error; + } + } + if (virDomainDeviceInfoParseXML(node, NULL, &def->info, flags) < 0) goto error; @@ -5146,6 +5155,7 @@ cleanup: VIR_FREE(type); VIR_FREE(idx); VIR_FREE(model); + VIR_FREE(num_queues); return def; @@ -13195,6 +13205,9 @@ virDomainControllerDefFormat(virBufferPtr buf, virBufferEscapeString(buf, " model='%s'", model); } + if (def->num_queues) + virBufferAsprintf(buf, " num_queues='%u'", def->num_queues); + switch (def->type) { case VIR_DOMAIN_CONTROLLER_TYPE_VIRTIO_SERIAL: if (def->opts.vioserial.ports != -1) { diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 700f03f..75e3f15 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -728,6 +728,7 @@ struct _virDomainControllerDef { int type; int idx; int model; /* -1 == undef */ + unsigned int num_queues; union { virDomainVirtioSerialOpts vioserial; } opts; diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 493e5f8..53caba5 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -3533,6 +3533,14 @@ qemuBuildControllerDevStr(virDomainDefPtr domainDef, virBuffer buf = VIR_BUFFER_INITIALIZER; int model; + if (def->num_queues && + !(def->type == VIR_DOMAIN_CONTROLLER_TYPE_SCSI && + def->model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("'num_queues' is only supported by virtio-scsi controller")); + return NULL; + } + switch (def->type) { case VIR_DOMAIN_CONTROLLER_TYPE_SCSI: model = def->model; @@ -3616,6 +3624,9 @@ qemuBuildControllerDevStr(virDomainDefPtr domainDef, goto error; } + if (def->num_queues) + virBufferAsprintf(&buf, ",num_queues=%u", def->num_queues); + if (qemuBuildDeviceAddressStr(&buf, &def->info, qemuCaps) < 0) goto error; diff --git a/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.args b/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.args new file mode 100644 index 0000000..8d46799 --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.args @@ -0,0 +1,8 @@ +LC_ALL=C PATH=/bin HOME=/home/test USER=test LOGNAME=test \ +/usr/bin/qemu -S -M pc -m 214 -smp 8 -nographic -nodefconfig -nodefaults \ +-monitor unix:/tmp/test-monitor,server,nowait -no-acpi -boot c \ +-device virtio-scsi-pci,id=scsi0,num_queues=8,bus=pci.0,addr=0x3 \ +-usb \ +-drive file=/dev/HostVG/QEMUGuest1,if=none,id=drive-scsi0-0-0-0 \ +-device scsi-disk,bus=scsi0.0,channel=0,scsi-id=0,lun=0,drive=drive-scsi0-0-0-0,id=scsi0-0-0-0 \ +-device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x4 diff --git a/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml b/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml new file mode 100644 index 0000000..dfa9cf1 --- /dev/null +++ b/tests/qemuxml2argvdata/qemuxml2argv-disk-virtio-scsi-num_queues.xml @@ -0,0 +1,26 @@ +<domain type='qemu'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory unit='KiB'>219136</memory> + <currentMemory unit='KiB'>219136</currentMemory> + <vcpu placement='static'>8</vcpu> + <os> + <type arch='i686' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <on_crash>destroy</on_crash> + <devices> + <emulator>/usr/bin/qemu</emulator> + <disk type='block' device='disk'> + <source dev='/dev/HostVG/QEMUGuest1'/> + <target dev='sdb' bus='scsi'/> + <address type='drive' controller='0' bus='0' target='0' unit='0'/> + </disk> + <controller type='usb' index='0'/> + <controller type='scsi' index='0' model='virtio-scsi' num_queues='8'/> + <memballoon model='virtio'/> + </devices> +</domain> diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 099fb36..d6575e7 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -553,6 +553,9 @@ mymain(void) DO_TEST("disk-scsi-virtio-scsi", QEMU_CAPS_DRIVE, QEMU_CAPS_DEVICE, QEMU_CAPS_NODEFCONFIG, QEMU_CAPS_VIRTIO_SCSI); + DO_TEST("disk-virtio-scsi-num_queues", + QEMU_CAPS_DRIVE, QEMU_CAPS_DEVICE, QEMU_CAPS_NODEFCONFIG, + QEMU_CAPS_VIRTIO_SCSI); DO_TEST("disk-scsi-megasas", QEMU_CAPS_DRIVE, QEMU_CAPS_DEVICE, QEMU_CAPS_NODEFCONFIG, QEMU_CAPS_SCSI_MEGASAS); diff --git a/tests/qemuxml2xmltest.c b/tests/qemuxml2xmltest.c index aa058bd..10a6ea2 100644 --- a/tests/qemuxml2xmltest.c +++ b/tests/qemuxml2xmltest.c @@ -179,6 +179,7 @@ mymain(void) DO_TEST("disk-scsi-device"); DO_TEST("disk-scsi-vscsi"); DO_TEST("disk-scsi-virtio-scsi"); + DO_TEST("disk-virtio-scsi-num_queues"); DO_TEST("disk-scsi-megasas"); DO_TEST_FULL("disk-mirror", false, WHEN_ACTIVE); DO_TEST_FULL("disk-mirror", true, WHEN_INACTIVE); -- 1.8.1.4

2 2

[libvirt] [PATCH 0/7] Fix multiple compiler warnings on ARMv7
by Daniel P. Berrange 06 Apr '13

06 Apr '13

The ARMv7 builds of libvirt generate a number of warnings, mostly about cast alignment http://arm.koji.fedoraproject.org/packages/libvirt/1.0.4/1.fc19/data/logs/a… this patch series fixes as many as possible, and then disables warnings for the rest using a pragma

3 17

[libvirt] [PATCH 0/3] Prohibit passing all clear cpumap for vcpupin/emulatorpin
by Osier Yang 06 Apr '13

06 Apr '13

*** NONE *** Osier Yang (3): util: Add a helper to check if all bits of a bitmap are clear qemu: Error out if the bitmap for pinning is all clear virsh: Prohibit all clear cpumap src/libvirt_private.syms | 1 + src/qemu/qemu_driver.c | 12 ++++++++++++ src/util/virbitmap.c | 30 ++++++++++++++++++++++++++++++ src/util/virbitmap.h | 3 +++ tools/virsh-domain.c | 35 ++++++++++++++++++++++++++++++----- 5 files changed, 76 insertions(+), 5 deletions(-) -- 1.8.1.4

4 15

[libvirt] [PATCH] build: check correct protocol.o file
by Eric Blake 06 Apr '13

06 Apr '13

By default, libtool builds two .o files for every .lo rule: src/foo.o - static builds src/.libs/foo.o - shared library builds But since commit ad42b34b disabled static builds, src/foo.o is no longer built by default. On a fresh checkout, this means our protocol check rules using pdwtags were testing a missing file, and thanks a lousy behavior of pdwtags happily giving no output and 0 exit status (http://bugzilla.redhat.com/949034) we were merely claiming that "dwarves is too old" and skipping the test. However, if you swap between branches and do incremental builds, such as building v0.10.2-maint and then switching back to master, you end up with src/foo.o being leftover from its 0.10.2 state, and then 'make check' fails because the .o file does not match the protocol-structs file due to API additions in the meantime. A simpler fix would be to always look in .libs for the .o to be parsed; but since it is possible to pass ./configure options to tell libtool to do a static-only build with no shared .o, I went with the approach of finding the newest of the two files, whenever both exist. * src/Makefile.am (PDWTAGS): Ensure we test just-built file. --- Pushing under the build-breaker rule. src/Makefile.am | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/src/Makefile.am b/src/Makefile.am index 78b4ab6..5dc58f9 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -341,9 +341,16 @@ r1 = /\* \d+ \*/ r2 = /\* <[[:xdigit:]]+> \S+:\d+ \*/ struct_prefix = (remote_|qemu_|lxc_|virNet|keepalive_) +# Depending on configure options, libtool creates one or both of +# {,.libs/}libvirt_remote_driver_la-remote_protocol.o. We want the +# newest of the two, in case configure options changed and a stale +# file is left around from an earlier build. PDWTAGS = \ $(AM_V_GEN)if (pdwtags --help) > /dev/null 2>&1; then \ - pdwtags --verbose $(<:.lo=.$(OBJEXT)) > $(@F)-t1 2> $(@F)-t2; \ + file=`ls -t $(<:.lo=.$(OBJEXT)) .libs/$(<:.lo=.$(OBJEXT)) \ + 2>/dev/null | sed -n 1p`; \ + test -f $$file || { echo ".o for $< not found" >&2; exit 1; };\ + pdwtags --verbose $$file > $(@F)-t1 2> $(@F)-t2; \ if test -s $(@F)-t2; then \ rm -rf $(@F)-t?; \ echo 'WARNING: pdwtags appears broken; skipping the $@ test' >&2;\ -- 1.8.1.4

1 0

[libvirt] [PATCH] Add a test suite for keycode mapping functions
by Daniel P. Berrange 06 Apr '13

06 Apr '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> Validate that translations between different keycode sets are functioning. Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- tests/virkeycodetest.c | 107 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 107 insertions(+) create mode 100644 tests/virkeycodetest.c diff --git a/tests/virkeycodetest.c b/tests/virkeycodetest.c new file mode 100644 index 0000000..d3c5975 --- /dev/null +++ b/tests/virkeycodetest.c @@ -0,0 +1,107 @@ +/* + * Copyright (C) 2013 Red Hat, Inc. + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; If not, see + * <http://www.gnu.org/licenses/>. + * + * Author: Daniel P. Berrange <berrange(a)redhat.com> + */ + +#include <config.h> + +#include <stdlib.h> + +#include "testutils.h" + +#include "virkeycode.h" +#include "virutil.h" +#include "virerror.h" +#include "viralloc.h" +#include "virlog.h" + +#include "virlockspace.h" + +#define VIR_FROM_THIS VIR_FROM_NONE + + +static int testKeycodeMapping(const void *data ATTRIBUTE_UNUSED) +{ + int ret = -1; + int got; + +#define TRANSLATE(from, to, val, want) \ + do { \ + if ((got = virKeycodeValueTranslate(VIR_KEYCODE_SET_##from, \ + VIR_KEYCODE_SET_##to, \ + val)) != want) { \ + fprintf(stderr, "Translating %d from %s to %s, got %d want %d\n", \ + val, #from, #to, got, want); \ + goto cleanup; \ + } \ + } while (0) + + TRANSLATE(LINUX, LINUX, 111, 111); + TRANSLATE(LINUX, USB, 111, 76); + TRANSLATE(LINUX, RFB, 88, 88); + TRANSLATE(LINUX, RFB, 160, 163); + TRANSLATE(ATSET2, ATSET3, 259, 55); + +#undef TRANSLATE + + ret = 0; +cleanup: + return ret; +} + + +static int testKeycodeStrings(const void *data ATTRIBUTE_UNUSED) +{ + int ret = -1; + int got; + +#define TRANSLATE(from, str, want) \ + do { \ + if ((got = virKeycodeValueFromString(VIR_KEYCODE_SET_##from, \ + str)) != want) { \ + fprintf(stderr, "Converting %s from %s, got %d want %d\n", \ + str, #from, got, want); \ + goto cleanup; \ + } \ + } while (0) + + TRANSLATE(LINUX, "KEY_DELETE", 111); + TRANSLATE(OSX, "Function", 0x3f); + TRANSLATE(WIN32, "VK_UP", 0x26); + +#undef TRANSLATE + + ret = 0; +cleanup: + return ret; +} + +static int +mymain(void) +{ + int ret = 0; + + if (virtTestRun("Keycode mapping ", 1, testKeycodeMapping, NULL) < 0) + ret = -1; + if (virtTestRun("Keycode strings ", 1, testKeycodeStrings, NULL) < 0) + ret = -1; + + return ret==0 ? EXIT_SUCCESS : EXIT_FAILURE; +} + +VIRT_TEST_MAIN(mymain) -- 1.8.1.4

2 2

[libvirt] [PATCH] build: use proper pod for nested bulleted VIRSH_DEBUG list
by Eric Blake 06 Apr '13

06 Apr '13

Newer pod (hello rawhide) complains if you attempt to mix bullets and non-bullets in the same list: virsh.pod around line 3177: Expected text after =item, not a bullet As our intent was to nest an inner list, we make that explicit to keep pod happy. * tools/virsh.pod (ENVIRONMENT): Use correct pod syntax. --- Pushing under the build-breaker rule. tools/virsh.pod | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/tools/virsh.pod b/tools/virsh.pod index 1c8f9ee..89467f6 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -3174,6 +3174,8 @@ of C<virsh> Turn on verbose debugging of virsh commands. Valid levels are +=over 4 + =item * VIRSH_DEBUG=0 DEBUG - Messages at ALL levels get logged @@ -3194,6 +3196,8 @@ WARNING - Logs messages at levels WARNING and ERROR ERROR - Messages at only ERROR level gets logged. +=back + =item VIRSH_LOG_FILE=C<LOGFILE> The file to log virsh debug messages. -- 1.8.1.4

1 0

[libvirt] [PATCH] qemu: Don't require a block or file when looking for an alias.
by Wido den Hollander 05 Apr '13

05 Apr '13

From: root <root(a)kvm01.(none)> This for example prohibits you to use iotune for Ceph or Sheepdog devices. Signed-off-by: Wido den Hollander <wido(a)widodh.nl> --- src/qemu/qemu_driver.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 552a81b..464d30a 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -13032,10 +13032,6 @@ qemuDiskPathToAlias(virDomainObjPtr vm, const char *path, int *idx) if (idx) *idx = i; - if (disk->type != VIR_DOMAIN_DISK_TYPE_BLOCK && - disk->type != VIR_DOMAIN_DISK_TYPE_FILE) - goto cleanup; - if (disk->src) { if (virAsprintf(&ret, "drive-%s", disk->info.alias) < 0) { virReportOOMError(); -- 1.7.9.5

2 2

[libvirt] [PATCH] Make guest OS bootable when hardware failure happens in log disk
by Seiji Aguchi 05 Apr '13

05 Apr '13

[Problem] Currently, guest OS's messages can be logged to a local disk of host OS by creating chadevs with options below. -chardev file,id=charserial0,path=<log file's path> -device isa-serial,chardev=chardevserial0,id=serial0 When a hardware failure happens in the disk, qemu-kvm can't create the chardevs. In this case, guest OS doesn't boot up. Actually, there are users who don't desire that guest OS goes down due to a hardware failure of a log disk only.Therefore, libvirt should offer some way to boot guest OS up even if the log disk is broken. [Solution] This patch changes a destination to /dev/null in case a log file can't be opened. Signed-off-by: Seiji Aguchi <seiji.aguchi(a)hds.com> --- src/qemu/qemu_process.c | 20 +++++++++++++++++++- 1 files changed, 19 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 8c4bfb7..fdf26ad 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -2440,7 +2440,25 @@ qemuProcessPrepareChardevDevice(virDomainDefPtr def ATTRIBUTE_UNUSED, virReportSystemError(errno, _("Unable to pre-create chardev file '%s'"), dev->source.data.file.path); - return -1; + VIR_FREE(dev->source.data.file.path); + /* + * Change a destination to /dev/null to boot guest OS up + * even if a log disk is broken. + */ + dev->source.data.file.path = strdup("/dev/null"); + + if (!dev->source.data.file.path) { + virReportOOMError(); + return -1; + } + + if ((fd = open(dev->source.data.file.path, + O_CREAT | O_APPEND, S_IRUSR|S_IWUSR)) < 0) { + virReportSystemError(errno, + _("Unable to pre-create chardev file '%s'"), + dev->source.data.file.path); + return -1; + } } VIR_FORCE_CLOSE(fd); -- 1.7.1

2 2

[libvirt] [PATCH] qemu: Remove maximum cpu limit when setting processor count using the API
by Peter Krempa 05 Apr '13

05 Apr '13

When setting processor count for a domain using the API libvirt enforced a maximum processor count, while it isn't enforced when taking the XML path. This patch removes the check to match the API. --- src/qemu/qemu_driver.c | 24 +++--------------------- 1 file changed, 3 insertions(+), 21 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 3f62414..833bf7f 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -3737,8 +3737,6 @@ qemuDomainSetVcpusFlags(virDomainPtr dom, unsigned int nvcpus, virQEMUDriverPtr driver = dom->conn->privateData; virDomainObjPtr vm = NULL; virDomainDefPtr persistentDef; - const char * type; - int max; int ret = -1; bool maximum; virQEMUDriverConfigPtr cfg = NULL; @@ -3778,27 +3776,11 @@ qemuDomainSetVcpusFlags(virDomainPtr dom, unsigned int nvcpus, goto endjob; } - if (!(type = virDomainVirtTypeToString(vm->def->virtType))) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("unknown virt type in domain definition '%d'"), - vm->def->virtType); - goto endjob; - } - - if ((max = qemuGetMaxVCPUs(NULL, type)) < 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, "%s", - _("could not determine max vcpus for the domain")); - goto endjob; - } - - if (!maximum && vm->def->maxvcpus < max) { - max = vm->def->maxvcpus; - } - - if (nvcpus > max) { + if (!maximum && nvcpus > vm->def->maxvcpus) { virReportError(VIR_ERR_INVALID_ARG, _("requested vcpus is greater than max allowable" - " vcpus for the domain: %d > %d"), nvcpus, max); + " vcpus for the domain: %d > %d"), + nvcpus, vm->def->maxvcpus); goto endjob; } -- 1.8.1.5

2 4

[libvirt] [PATCH 0/5] Drop virReportOOMReport from almost everywhere
by Michal Privoznik 05 Apr '13

05 Apr '13

3 13

[libvirt] [PATCH v2 2/2] Implement support for <hostdev caps=net>
by Bogdan Purcareata 05 Apr '13

05 Apr '13

This allows a container-type domain to have exclusive access to one of the host's NICs. Wire <hostdev caps=net> with the lxc_controller - when moving the newly created veth devices into a new namespace, also look for any hostdev devices that should be moved. Note: once the container domain has been destroyed, there is no code that moves the interfaces back to the original namespace. This does happen, though, probably due to default cleanup on namespace destruction. Signed-off-by: Bogdan Purcareata <bogdan.purcareata(a)freescale.com> --- src/lxc/lxc_container.c | 4 +++- src/lxc/lxc_controller.c | 16 ++++++++++++++++ src/lxc/lxc_hostdev.c | 1 + 3 files changed, 20 insertions(+), 1 deletion(-) diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index 002ba9e..e59bfdf 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -1551,7 +1551,6 @@ cleanup: return ret; } - static int lxcContainerSetupHostdevSubsys(virDomainDefPtr vmDef, virDomainHostdevDefPtr def, const char *dstprefix, @@ -1582,6 +1581,9 @@ static int lxcContainerSetupHostdevCaps(virDomainDefPtr vmDef, case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_MISC: return lxcContainerSetupHostdevCapsMisc(vmDef, def, dstprefix, securityDriver); + case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_NET: + return 0; // case is handled in virLXCControllerMoveInterfaces + default: virReportError(VIR_ERR_CONFIG_UNSUPPORTED, _("Unsupported host device mode %s"), diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index cede445..edd99bf 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -1050,12 +1050,28 @@ cleanup2: static int virLXCControllerMoveInterfaces(virLXCControllerPtr ctrl) { size_t i; + virDomainDefPtr def = ctrl->def; for (i = 0 ; i < ctrl->nveths ; i++) { if (virNetDevSetNamespace(ctrl->veths[i], ctrl->initpid) < 0) return -1; } + for (i = 0; i < def->nhostdevs; i ++) { + virDomainHostdevDefPtr hdev = def->hostdevs[i]; + + if (hdev->mode != VIR_DOMAIN_HOSTDEV_MODE_CAPABILITIES) + continue; + + virDomainHostdevCaps hdcaps = hdev->source.caps; + + if (hdcaps.type != VIR_DOMAIN_HOSTDEV_CAPS_TYPE_NET) + continue; + + if (virNetDevSetNamespace(hdcaps.u.net.iface, ctrl->initpid) < 0) + return -1; + } + return 0; } diff --git a/src/lxc/lxc_hostdev.c b/src/lxc/lxc_hostdev.c index 33b0b60..53a1a31 100644 --- a/src/lxc/lxc_hostdev.c +++ b/src/lxc/lxc_hostdev.c @@ -307,6 +307,7 @@ int virLXCPrepareHostDevices(virLXCDriverPtr driver, switch (dev->source.subsys.type) { case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_STORAGE: case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_MISC: + case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_NET: break; default: virReportError(VIR_ERR_CONFIG_UNSUPPORTED, -- 1.7.11.7

1 0

[libvirt] libvirt_lxc: implement <hostdev caps=net> device isolation in containers
by Bogdan Purcareata 05 Apr '13

05 Apr '13

This set of patches implements <hostdev caps=net> interface isolation in containers, thus allowing an interface NIC to be assigned exclusively to a container-domain. This is done like moving veth devices in container namespaces, only this time it is actual host devices.

2 4

[libvirt] [PATCH v5.1 00/11] Introduce driver specific XML parsing callbacks
by Peter Krempa 05 Apr '13

05 Apr '13

In this respin the renaming patch is moved to be first in the series. The argument rearrangement patch was too complicated to move. Unfortunately this rebase operation caused my "git notes" to be lost for the individual patches, so please see v5 for more specific comments. Patches 5 to 11 except for 10 were ACKed in previous versions. Peter Krempa (11): maint: Rename xmlconf to xmlopt and virDomainXMLConfing to virDomainXMLOption conf: Add post XML parse callbacks and prepare for cleaning of virCaps conf callback: Rearrange function parameters qemu: Record the default NIC model in the domain XML virCaps: get rid of "defaultInitPath" value in the virCaps struct virCaps: get rid of defaultDiskDriverName virCaps: get rid of emulatorRequired virCaps: get rid of defaultDiskDriverType virCaps: get rid of hasWideScsiBus virCaps: get rid of macPrefix field virCaps: get rid of defaultConsoleTargetType callback src/conf/capabilities.c | 24 - src/conf/capabilities.h | 23 - src/conf/domain_conf.c | 660 ++++++++++++--------- src/conf/domain_conf.h | 102 ++-- src/conf/snapshot_conf.c | 4 +- src/conf/snapshot_conf.h | 2 +- src/esx/esx_driver.c | 32 +- src/esx/esx_private.h | 2 +- src/libvirt_private.syms | 10 +- src/libvirt_vmx.syms | 1 + src/libxl/libxl_conf.c | 13 - src/libxl/libxl_conf.h | 2 +- src/libxl/libxl_driver.c | 103 ++-- src/lxc/lxc_conf.c | 21 +- src/lxc/lxc_conf.h | 4 +- src/lxc/lxc_controller.c | 10 +- src/lxc/lxc_domain.c | 34 ++ src/lxc/lxc_domain.h | 1 + src/lxc/lxc_driver.c | 71 ++- src/lxc/lxc_process.c | 8 +- src/openvz/openvz_conf.c | 16 +- src/openvz/openvz_conf.h | 2 +- src/openvz/openvz_driver.c | 64 +- src/parallels/parallels_driver.c | 37 +- src/parallels/parallels_utils.h | 2 +- src/phyp/phyp_driver.c | 24 +- src/phyp/phyp_driver.h | 2 +- src/qemu/qemu_capabilities.c | 19 - src/qemu/qemu_command.c | 27 +- src/qemu/qemu_command.h | 6 +- src/qemu/qemu_conf.c | 18 +- src/qemu/qemu_conf.h | 5 +- src/qemu/qemu_domain.c | 98 ++- src/qemu/qemu_domain.h | 1 + src/qemu/qemu_driver.c | 147 +++-- src/qemu/qemu_migration.c | 21 +- src/qemu/qemu_process.c | 34 +- src/security/virt-aa-helper.c | 19 +- src/test/test_driver.c | 67 +-- src/uml/uml_conf.c | 9 - src/uml/uml_conf.h | 2 +- src/uml/uml_driver.c | 69 ++- src/vbox/vbox_tmpl.c | 33 +- src/vmware/vmware_conf.c | 21 +- src/vmware/vmware_conf.h | 2 +- src/vmware/vmware_driver.c | 24 +- src/vmx/vmx.c | 39 +- src/vmx/vmx.h | 12 +- src/xen/xen_driver.c | 37 +- src/xen/xen_driver.h | 4 +- src/xen/xen_hypervisor.c | 13 - src/xen/xend_internal.c | 20 +- src/xen/xm_internal.c | 16 +- src/xenapi/xenapi_driver.c | 40 +- src/xenapi/xenapi_driver_private.h | 2 +- tests/domainsnapshotxml2xmltest.c | 6 +- tests/lxcxml2xmldata/lxc-hostdev.xml | 1 + tests/lxcxml2xmldata/lxc-systemd.xml | 1 + tests/lxcxml2xmltest.c | 8 +- tests/qemuargv2xmltest.c | 6 +- tests/qemumonitorjsontest.c | 28 +- tests/qemumonitortestutils.c | 4 +- tests/qemumonitortestutils.h | 2 +- ...qemuxml2argv-disk-drive-network-nbd-export.args | 3 +- .../qemuxml2argv-disk-drive-network-nbd-export.xml | 1 + ...ml2argv-disk-drive-network-nbd-ipv6-export.args | 3 +- ...xml2argv-disk-drive-network-nbd-ipv6-export.xml | 1 + .../qemuxml2argv-disk-drive-network-nbd-ipv6.args | 3 +- .../qemuxml2argv-disk-drive-network-nbd-ipv6.xml | 1 + .../qemuxml2argv-disk-drive-network-nbd-unix.args | 3 +- .../qemuxml2argv-disk-drive-network-nbd-unix.xml | 1 + .../qemuxml2argv-disk-drive-network-nbd.args | 5 +- .../qemuxml2argv-disk-drive-network-nbd.xml | 1 + .../qemuxml2argv-disk-drive-network-rbd-auth.args | 2 +- .../qemuxml2argv-disk-drive-network-rbd-ipv6.args | 2 +- .../qemuxml2argv-disk-drive-network-rbd-ipv6.xml | 1 + .../qemuxml2argv-disk-drive-network-rbd.args | 2 +- .../qemuxml2argv-disk-drive-network-rbd.xml | 1 + .../qemuxml2argv-disk-drive-network-sheepdog.args | 3 +- .../qemuxml2argv-disk-drive-network-sheepdog.xml | 1 + .../qemuxml2argv-net-bandwidth.xml | 1 + .../qemuxml2argvdata/qemuxml2argv-net-client.args | 4 +- .../qemuxml2argv-net-eth-ifname.args | 4 +- .../qemuxml2argv-net-eth-ifname.xml | 1 + .../qemuxml2argv-net-eth-names.args | 8 +- tests/qemuxml2argvdata/qemuxml2argv-net-eth.args | 4 +- tests/qemuxml2argvdata/qemuxml2argv-net-eth.xml | 1 + .../qemuxml2argvdata/qemuxml2argv-net-hostdev.xml | 1 + tests/qemuxml2argvdata/qemuxml2argv-net-mcast.args | 4 +- .../qemuxml2argv-net-openvswitch.xml | 1 + .../qemuxml2argvdata/qemuxml2argv-net-server.args | 4 +- tests/qemuxml2argvdata/qemuxml2argv-net-user.args | 3 +- tests/qemuxml2argvdata/qemuxml2argv-net-user.xml | 1 + .../qemuxml2argv-net-virtio-network-portgroup.xml | 2 + tests/qemuxml2argvtest.c | 6 +- .../qemuxml2xmlout-graphics-spice-timeout.xml | 1 + tests/qemuxml2xmltest.c | 6 +- tests/qemuxmlnstest.c | 6 +- tests/securityselinuxlabeltest.c | 6 +- tests/testutilslxc.c | 9 - tests/testutilsqemu.c | 11 - tests/testutilsqemu.h | 2 +- tests/testutilsxen.c | 16 - tests/testutilsxen.h | 2 - tests/vmx2xmltest.c | 17 +- tests/xmconfigtest.c | 8 +- tests/xml2sexprtest.c | 9 +- tests/xml2vmxtest.c | 20 +- 108 files changed, 1275 insertions(+), 1052 deletions(-) -- 1.8.1.5

5 32

[libvirt] Device backend in another domain (xen)
by Marek Marczykowski 05 Apr '13

05 Apr '13

Hi all, Xen have (AFAIK unique) feature of having backend device in other domU (sometimes called driver domain) instead of dom0. Currently libxl support it for disk and network devices, but in general it is valid for any device (perhaps excluding PCI). Current implementation requires ID of backend domain to be specified during device attach/domain startup. There are some upcoming patches to allow also use domain name. I'd like to add support for it into libvirt. The question is how specify domain reference. The simplest approach is to use direct domain ID, logically it fits in <source/> tag, so something like: <disk type='file'> <source file='/path/to/file/inside/domain/7' domid='7'/> <target dev='xvdb'/> </disk> <interface type='bridge'> <source bridge='name-of-bridge-in-domiain-7' domid='7'/> </interface> This have obvious limitation that domain ID changes with each domain restart. But perhaps more correct approach is to use name or UUID? Extending <source/> tag is right thing to do? Or should I add a new tag (<backenddomain uuid='...'/> ? )? Regarding implementation I need some hints how to do the value verification (check if domain exists at least). If it has to be done in domain_conf.c (during XML parsing) it will require predetermined domain load order, which IMHO isn't acceptable. So perhaps the check should be done at domain startup/device attach (by driver)? In any case: is it possible at all to check if domain with given name/UUID exists in scope of domain_conf.c? It looks like it requires connection reference, which isn't passed to virDomain*DefParseXML. -- Best Regards, Marek Marczykowski Invisible Things Lab

2 1

[libvirt] [PATCH 1/3] Remove support for old kernels lacking private devpts
by Daniel P. Berrange 05 Apr '13

05 Apr '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> Early on kernel support for private devpts was not widespread, so we had compatibiltiy codepaths. Such old kernels are not seriously used for LXC these days, so the compat code can go away Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/lxc/lxc_controller.c | 28 +++++++++------------------- 1 file changed, 9 insertions(+), 19 deletions(-) diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index 1d1443c..cede445 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -1252,8 +1252,9 @@ virLXCControllerSetupDevPTS(virLXCControllerPtr ctrl) } if (access(ctrl->devptmx, R_OK) < 0) { - VIR_WARN("Kernel does not support private devpts, using shared devpts"); - VIR_FREE(ctrl->devptmx); + virReportSystemError(ENOSYS, "%s", + _("Kernel does not support private devpts")); + goto cleanup; } ret = 0; @@ -1278,24 +1279,13 @@ virLXCControllerSetupConsoles(virLXCControllerPtr ctrl, size_t i; for (i = 0 ; i < ctrl->nconsoles ; i++) { - if (ctrl->devptmx) { - VIR_DEBUG("Opening tty on private %s", ctrl->devptmx); - if (lxcCreateTty(ctrl->devptmx, - &ctrl->consoles[i].contFd, - &containerTTYPaths[i]) < 0) { - virReportSystemError(errno, "%s", - _("Failed to allocate tty")); - return -1; - } - } else { - VIR_DEBUG("Opening tty on shared /dev/ptmx"); - if (virFileOpenTty(&ctrl->consoles[i].contFd, - &containerTTYPaths[i], - 0) < 0) { - virReportSystemError(errno, "%s", + VIR_DEBUG("Opening tty on private %s", ctrl->devptmx); + if (lxcCreateTty(ctrl->devptmx, + &ctrl->consoles[i].contFd, + &containerTTYPaths[i]) < 0) { + virReportSystemError(errno, "%s", _("Failed to allocate tty")); - return -1; - } + return -1; } } return 0; -- 1.7.11.7

3 11

[libvirt] [RFC] Per machine type qemu capability detection
by Peter Krempa 05 Apr '13

05 Apr '13

Hi, qemu recently added a qmp/hmp command ("get-cpu-max") to determine maximum number of supported CPUs for a VM. The output of this command depends on the machine type that is used when it is called. The problem here is that libvirt uses "-M none" while detecting capabilities and with that machine type the get-cpu-max command returns always 1. My question is: Is it worthwhile to do per machine type capability detection? It definitely isn't worth doing just for the get-cpu-max support, but it might be needed for Q35 or other stuff that may depend on the actual machine type used. Pros of this approach: ---------------------- - (possibly) more accurate capabilities Cons: ----- - ~10x more qemus started on libvirtd startup (one per machine type) Thanks for your opinions. Peter

4 6

[libvirt] [RFC 0/5] X86CPU "feature-words" & "filtered-features" properties (v2)
by Eduardo Habkost 05 Apr '13

05 Apr '13

Changes v2: * Added "filtered-features" properties * Removed the non-qapi version and kept only the qapi version of the code * Moved the new types to qapi-schema.json instead of creating cpu-qapi-schema.json * Rebased on top of "feature words array" series Git tree: https://github.com/ehabkost/qemu-hacks/tree/work/cpu-raw-feature-props-v2 git://github.com/ehabkost/qemu-hacks.git work/cpu-raw-feature-props-v2 Description: As the work to get X86CPU subclasses static properties defined is taking very long to get reviewed and we risk not getting it finished in 1.5, I am sending an alternative mechanism to allow libvirt to probe for CPU model information. It's interesting to note that this alternative solution is actually _easier_ to use for libvirt, because libvirt logic is already based on CPUID bit values, not "feature names", so it should be very easy to convert the current libvirt code that (incorrectly) query the host CPU directly for available features to instead start QEMU with "-cpu host" and use the "feature-words" array information to find out which CPU features are supported by the host. The "feature-words" property will have two main use cases: * Checking host capabilities, by checking the features of the "host" CPU model; * Checking which features are enabled on each CPU model. The "filtered-features" property will have two main use cases: * Allowing libvirt to emulate the "check" & "enforce" flags while collecting more detailted data about the missing features; * Allowing libvirt to probe for CPU model information even if host features are missing (it can simply combine the "feature-words" and "filtered-features" values to get the full CPU model definition) Example output: $ ./QMP/qmp --path=/tmp/m qom-get --path=/machine/unattached/device[1] --property=feature-words item[0].cpuid-register: EDX item[0].cpuid-input-eax: 1 item[0].features: 126614521 item[1].cpuid-register: ECX item[1].cpuid-input-eax: 1 item[1].features: 2155880449 item[2].cpuid-register: EBX item[2].cpuid-input-eax: 7 item[2].features: 0 item[2].cpuid-input-ecx: 0 item[3].cpuid-register: EDX item[3].cpuid-input-eax: 2147483649 item[3].features: 563346425 item[4].cpuid-register: ECX item[4].cpuid-input-eax: 2147483649 item[4].features: 101 item[5].cpuid-register: EDX item[5].cpuid-input-eax: 3221225473 item[5].features: 0 item[6].cpuid-register: EAX item[6].cpuid-input-eax: 1073741825 item[6].features: 0 item[7].cpuid-register: EDX item[7].cpuid-input-eax: 2147483658 item[7].features: 0 Eduardo Habkost (5): target-i386: Add ECX information to FeatureWordInfo target-i386: Add "feature-words" property target-i386: Use FeatureWord loop on filter_features_for_kvm() target-i386: Introduce X86CPU.filtered_features field target-i386: Add "filtered-features" property to X86CPU .gitignore | 2 + Makefile.objs | 7 +++- qapi-schema.json | 32 +++++++++++++++ target-i386/cpu-qom.h | 3 ++ target-i386/cpu.c | 110 +++++++++++++++++++++++++++++++++++--------------- 5 files changed, 121 insertions(+), 33 deletions(-) -- 1.8.1.4

1 5

[libvirt] libvirt_lxc: implement <hostdev caps=net> device isolation in containers
by Bogdan Purcareata 05 Apr '13

05 Apr '13

>From Bogdan Purcareata <bogdan.purcareata(a)freescale.com> # This line is ignored. From: Bogdan Purcareata <bogdan.purcareata(a)freescale.com> Subject: libvirt_lxc: implement <hostdev caps=net> device isolation in containers In-Reply-To:

1 2