- Devel - libvirt List Archives

[libvirt] FYI: valgrind output from libvirtd when running the libguestfs test suite
by Richard W.M. Jones 18 Sep '12

18 Sep '12

./run \ valgrind --leak-check=full --child-silent-after-fork=yes \ --log-file=/tmp/valgrind.log \ daemon/libvirtd --timeout=30 Then run '../libvirt/run make check' in another window. This is with libvirt from git, updated a few hours ago. I'm always dubious about these strlen problems since they are often caused by glibc, but could be worth looking into. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-top is 'top' for virtual machines. Tiny program with many powerful monitoring features, net stats, disk stats, logging, etc. http://et.redhat.com/~rjones/virt-top

1 0

[libvirt] [libvirt-glib][PATCH] autobuild: Init AUTOBUILD_INSTALL_ROOT variable
by Michal Privoznik 18 Sep '12

18 Sep '12

Otherwise autobuild sets prefix to / resulting in Permission denied when building as non-root as 'make install' tries to write to /lib. We need to set this var to accessible path like $HOME/builder. --- autobuild.sh | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/autobuild.sh b/autobuild.sh index 3675249..053dc1f 100755 --- a/autobuild.sh +++ b/autobuild.sh @@ -4,6 +4,7 @@ set -e set -v test -n "$1" && RESULTS=$1 || RESULTS=results.log +: ${AUTOBUILD_INSTALL_ROOT=$HOME/builder} # Make things clean. test -f Makefile && make -k distclean || : -- 1.7.8.6

2 2

[libvirt] [PATCH 00/10] new virNetworkUpdate API
by Laine Stump 18 Sep '12

18 Sep '12

This patchset implements a new API function called virNetworkUpdate which enables updating certain parts of a libvirt network's definition without the need to destroy/re-start the network. This is especially useful, for example, to add/remove hosts from the dhcp static hosts table, or change portgroup settings. This was previously discussed in this thread: https://www.redhat.com/archives/libvir-list/2012-August/msg01535.html continuing here in September: https://www.redhat.com/archives/libvir-list/2012-September/msg00328.html with the final form here: https://www.redhat.com/archives/libvir-list/2012-September/msg00465.html In short, the single function has a "section" specifier which tells the part of the network definition to be updated, a "parentIndex" that gives the index of the *parent* element containing this section (when there are multiples - in particular in the case of the <ip> element), and a fully formed XML element which will be added as-is in the case of VIR_NETWORK_UPDATE_ADD_* (after checking for a duplicate), used to search for the specific element to delete in case of VIR_NETWORK_UPDATE_DELETE, and used both to find the existing element and replace its current contents in the case of VIR_UPDATE_EXISTING (this implies that you can't change the change the attribute used for indexing, e.g. the name of a portgroup, or mac address of a dhcp host entry). An example of use: to add a dhcp host entry to network "net", you would do this: virNetworkUpdate(net, VIR_NETWORK_SECTION_IP_DHCP_HOST, -1, "<host mac='00:11:22:33:44:55' ip='192.168.122.5'/>", VIR_NETWORK_UPDATE_AFFECT_LIVE | VIR_NETWORK_UPDATE_AFFECT_CONFIG | VIR_NETWORK_UPDATE_ADD_LAST); To delete that same entry: virNetworkUpdate(net, VIR_NETWORK_SECTION_IP_DHCP_HOST, -1, "<host mac='00:11:22:33:44:55'/>", VIR_NETWORK_UPDATE_AFFECT_LIVE | VIR_NETWORK_UPDATE_AFFECT_CONFIG | VIR_NETWORK_UPDATE_DELETE); If you wanted to force any of these to affect the dhcp host list in the 3rd <ip> element of the network, you would replace "-1" with "2". Another example: to modify the portgroup named "engineering" (e.g. to increase the inbound average bandwidth from 1000 to 2000): virNetworkUpdate(net, VIR_NETWORK_SECTION_PORTGROUP, -1, "<portgroup name='engineering' default='yes'>" " <virtualport type='802.1Qbh'>" " <parameters profileid='test'/>" " </virtualport>" " <bandwidth>" " <inbound average='2000' peak='5000' burst='5120'/>" " <outbound average='1000' peak='5000' burst='5120'/>" " </bandwidth>" "</portgroup>", VIR_NETWORK_UPDATE_EXISTING | VIR_NETWORK_UPDATE_LIVE | VIR_NETWORK_UPDATE_CONFIG) (note that parentIndex is irrelevant for PORTGROUP, since they are in the toplevel of <network>, so there aren't multiple instances of parents. In such cases, the caller *must* set parentIndex to -1 or 0 - any other value indicates that they don't understand the purpose/usage of parentIndex, so it must result in an error. Also note that the above function would fail if it couldn't find an existing portgroup with name='engineering' (i.e. it wouldn't automatically add a new one).) Adding support for each of the different sections has been reduced to a single function that handles the update of a virNetworkDef; all the logic to determine which virNetworkDef (def or newDef) and to restart/SIGHUP the appropriate daemons is in higher levels and is 100% complete. The low level functions aren't yet finished, although the function for IP_DHCP_HOST is nearly done. As usual, several of the patches are re-factoring existing code, and a couple are bugfixes that are only peripherally related: 1/10 - splits out parsing of dhcp related elements to separate functions 2/10 - fixes a small bug that would probably never be encountered IRL anyway 3/10+4/10 - actual API 5/10 - utility functions to simplify API implementation 6/10 - framework for backend that updates the virNetworkDef 7/10 - refactoring in bridge_driver 8/10 - virNetworkUpdate for bridge_driver 9/10 - virNetworkUpdate for test_driver 10/10 - simple troubleshooting aid - restart dnsmasq/radvd when libvirtd is restarted (if its process is missing). I'll try to have the following additional patches posted later today: 11/10 - backend for IP_DHCP_HOST update 12/10 - backend for PORTGROUP update 13/10 - virsh "net-update" command Differences from email thread: 1) Aside from being actual code instead of just pseudo code, I've changed the "index" arg into "parentIndex" to avoid confusion between the index of the element we're looking at (e.g. which host in the list of hosts in a <dhcp> element) and the index of the parent (e.g. which <ip> element will contain the <dhcp> with the list of <hosts> we want to update). I also changed this from a unsigned into to an int so that it can hold the special value "-1", which means "pick the most useful/only item". Again, in the case of VIR_NETWORK_SECTION_IP_DHCP_HOST, this would mean to pick the one <ip> element that actually has a <dhcp> section (may not be the first, but there can only be one). 2) added VIR_NETWORK_UPDATE_AFFECT_CURRENT similar to VIR_DOMAIN_AFFECT_CURRENT

3 20

[libvirt] [PATCH] rpc: Fix name of member in remote_protocol-structs
by Peter Krempa 18 Sep '12

18 Sep '12

Commit 7a99b0abafb69e1686198ac3473892a9aaeb8255 adds a new RPC struct but one of the members has different names in remote_protocol.x and remote_protocol-struct breaking make check. --- Pushing under build-breaker rule. src/remote_protocol-structs | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/remote_protocol-structs b/src/remote_protocol-structs index 0a9beff..4d2627a 100644 --- a/src/remote_protocol-structs +++ b/src/remote_protocol-structs @@ -981,7 +981,7 @@ struct remote_network_update_args { remote_nonnull_network net; u_int command; u_int section; - int index; + int parentIndex; remote_nonnull_string xml; u_int flags; }; -- 1.7.12

1 0

[libvirt] Add a ./run script for running programs from the local directory.
by Richard W.M. Jones 18 Sep '12

18 Sep '12

In libguestfs we successfully use a ./run script which sets up the environment so you can run libguestfs programs (internal and external ones, in C and non-C) without needing to install anything. This adds a similar script for libvirt. Here is the documentation included in the script: # With this script you can run libvirt programs without needing to # install them first. You just have to do for example: # # ./run ./tools/virsh [args ...] # # If you are already in the tools/ subdirectory, then the following # command will also work: # # ../run ./virsh [...] # # You can also run the C programs under valgrind like this: # # ./run valgrind [valgrind opts...] ./program # # or under gdb: # # ./run gdb --args ./program # # This also works with sudo (eg. if you need root access for libvirt): # # sudo ./run ./tools/virsh list --all

2 8

[libvirt] [PATCH] virBitmap: fix build without HAVE_NUMACTL
by Ján Tomko 18 Sep '12

18 Sep '12

Commit 75b198b3e7b2c8b4106afbd42f8bb795c4773511 forgot to change arguments of dummy qemuProcessInitNumaMemoryPolicy from char* to virBitmapPtr. --- src/qemu/qemu_process.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 5095381..c28f5a5 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -1798,7 +1798,7 @@ cleanup: #else static int qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm, - const char *nodemask ATTRIBUTE_UNUSED) + virBitmapPtr nodemask ATTRIBUTE_UNUSED) { if (vm->def->numatune.memory.nodemask) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", -- 1.7.8.6

3 2

[libvirt] [PATCHv3 0/4] qemu: Add seccomp sandbox support
by Ján Tomko 18 Sep '12

18 Sep '12

This series adds support to run QEMU with seccomp sandbox enabled. It can be configured in qemu.conf to on, off, or the QEMU default, which is off in 1.2. Default value is the QEMU default. V1 of the patches supports tri-state configuration and includes a test for detecting the capability. V2 includes libvirtd_qemu.aug, renames vague 'sandbox' to 'seccompSandbox' and prints an error if the sandbox was requested but QEMU doesn't have the capability. V3 is a rebase, needed after the addition of usb-redir.filter and the change of virBitmapPtr to qemuCapsPtr. Ján Tomko (4): qemu: add capability flag for seccomp sandbox qemu: conf: add seccomp_sandbox option qemu: add -sandbox to command line if requested tests: add qemu-1.2.0 help data src/qemu/libvirtd_qemu.aug | 1 + src/qemu/qemu.conf | 8 + src/qemu/qemu_capabilities.c | 3 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 11 ++ src/qemu/qemu_conf.c | 5 + src/qemu/qemu_conf.h | 1 + tests/qemuhelpdata/qemu-1.2.0 | 270 ++++++++++++++++++++++++++++++++++ tests/qemuhelpdata/qemu-1.2.0-device | 181 +++++++++++++++++++++++ tests/qemuhelptest.c | 82 ++++++++++ 10 files changed, 563 insertions(+), 0 deletions(-) create mode 100644 tests/qemuhelpdata/qemu-1.2.0 create mode 100644 tests/qemuhelpdata/qemu-1.2.0-device -- 1.7.8.6

2 5

[libvirt] [PATCH] qemu: Avoid deadlock on HandleAgentEOF
by Michal Privoznik 18 Sep '12

18 Sep '12

On agent EOF the qemuProcessHandleAgentEOF() callback is called which locks virDomainObjPtr. Then qemuAgentClose() is called (with domain object locked) which eventually calls qemuAgentFree() and qemuProcessHandleAgentDestroy(). This tries to lock the domain object again. Hence the deadlock. --- src/qemu/qemu_process.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 5ac1d2b..8f42c83 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -115,7 +115,7 @@ extern struct qemud_driver *qemu_driver; * performed */ static void -qemuProcessHandleAgentEOF(qemuAgentPtr agent ATTRIBUTE_UNUSED, +qemuProcessHandleAgentEOF(qemuAgentPtr agent, virDomainObjPtr vm) { struct qemud_driver *driver = qemu_driver; @@ -127,12 +127,12 @@ qemuProcessHandleAgentEOF(qemuAgentPtr agent ATTRIBUTE_UNUSED, virDomainObjLock(vm); priv = vm->privateData; - - qemuAgentClose(agent); priv->agent = NULL; virDomainObjUnlock(vm); qemuDriverUnlock(driver); + + qemuAgentClose(agent); } -- 1.7.8.6

2 2

[libvirt] [PATCH 0/4 v3] Support to set disk wwn
by Osier Yang 18 Sep '12

18 Sep '12

v2 - v3: * Just rebase on the top. v1 - v2: * Validate wwn while parsing * Error out earlier "if (disk->device == "lun" && disk->wwn)", no logic change. This introduces new element <wwn> for disk, to allow to set wwn (just like setting serial number) for the virtual disk (Only QEMU devices like ide-drive, ide-hd, ide-cd, scsi-disk, scsi-hd, and scsi-cd support it). Osier Yang (4): schema: Add schema for disk <wwn> conf: Parse and format disk <wwn> qemu: Add caps to indentify if setting wwn is supported by qemu qemu: Use disk wwn in qemu command line docs/formatdomain.html.in | 6 +++ docs/schemas/basictypes.rng | 6 +++ docs/schemas/domaincommon.rng | 5 +++ docs/schemas/nodedev.rng | 6 --- src/conf/domain_conf.c | 12 +++++++ src/conf/domain_conf.h | 1 + src/libvirt_private.syms | 1 + src/qemu/qemu_capabilities.c | 8 ++++ src/qemu/qemu_capabilities.h | 2 + src/qemu/qemu_command.c | 33 ++++++++++++++++++ src/util/util.c | 17 +++++++++ src/util/util.h | 2 + .../qemuxml2argv-disk-ide-wwn.args | 6 +++ .../qemuxml2argvdata/qemuxml2argv-disk-ide-wwn.xml | 28 ++++++++++++++++ .../qemuxml2argv-disk-scsi-disk-wwn.args | 10 ++++++ .../qemuxml2argv-disk-scsi-disk-wwn.xml | 35 ++++++++++++++++++++ tests/qemuxml2argvtest.c | 7 ++++ 17 files changed, 179 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-ide-wwn.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-ide-wwn.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-scsi-disk-wwn.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-scsi-disk-wwn.xml -- 1.7.7.3

2 9

[libvirt] [PATCH 0/3] block commit support
by Eric Blake 18 Sep '12

18 Sep '12

I'm still working on the qemu implementation of this API, but want to get the API reviewed and included in 0.10.2 before adding piecewise support for more and more of the API. My biggest issues for current implementations are: online qemu: Jeff's block-commit API is still not part of qemu.git, and might be subject to change, in particular based on review from Paolo's drive-mirror changes (since both patch series will be sharing related code on how to handle error reporting). offline qemu: qemu-img is a blocking command, but this API promises to be asynchronous. I'll have to run qemu-img in a worker thread, and I still need to figure out whether qemu-img can display progress for virDomainBlockJobInfo (can you send it a SIGUSR1, like for dd?), otherwise, the job will look like it is making no progress. But as neither of those issues invalidate the API, I'm comfortable with committing the API now (after all, that's what we did back in 0.9.12 with the virDomainBlockRebase API - we committed code in preparation for qemu to implement drive-mirror, and we are STILL waiting on qemu to follow through with that promise...) Obviously, there will be more patches to come as I get more things working at the qemu layer (I'm not sure whether online or offline qemu will be done first, as I have branches working on both approaches while I am testing against Jeff's qemu patch proposals). Eric Blake (3): blockjob: add virDomainBlockCommit blockjob: add virsh blockcommit blockjob: add blockcommit support to rpc docs/apibuild.py | 1 + include/libvirt/libvirt.h.in | 20 ++++++ src/driver.h | 5 ++ src/libvirt.c | 105 +++++++++++++++++++++++++++++ src/libvirt_public.syms | 1 + src/remote/remote_driver.c | 1 + src/remote/remote_protocol.x | 12 +++- src/remote_protocol-structs | 9 +++ src/rpc/gendispatch.pl | 1 + tools/virsh-domain.c | 153 ++++++++++++++++++++++++++++++++++++++++++- tools/virsh.pod | 30 +++++++++ 11 files changed, 336 insertions(+), 2 deletions(-) -- 1.7.11.4

2 9