May 2015 - Devel - libvirt List Archives

[libvirt] [PATCH v2] sysinfo: Fix reports on ARM
by Michal Privoznik 22 May '15

22 May '15

Due to a kernel commit (b4b8f770e), cpuinfo format has changed on ARMs. Firstly, 'Processor: ...' may not be reported, it's replaced by 'model name: ...'. Secondly, the "Processor" string may occur in CPU name, e.g. 'ARMv7 Processor rev 5 (v7l)'. Therefore, we must firstly look for 'model name' and then for 'Processor' if not found. Moreover, lines in the cpuinfo file are shuffled, so we better not manipulate the pointer to start of internal buffer as we may lost some info. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- diff to v1: - Jan's comments worked in - added new test case (yes, there really is a space at EOL in cpuinfo on my ARM) src/util/virsysinfo.c | 9 +++---- tests/sysinfodata/arm-rpi2cpuinfo.data | 43 ++++++++++++++++++++++++++++++++ tests/sysinfodata/arm-rpi2sysinfo.expect | 18 +++++++++++++ tests/sysinfotest.c | 22 ++++++++++++---- 4 files changed, 82 insertions(+), 10 deletions(-) create mode 100644 tests/sysinfodata/arm-rpi2cpuinfo.data create mode 100644 tests/sysinfodata/arm-rpi2sysinfo.expect diff --git a/src/util/virsysinfo.c b/src/util/virsysinfo.c index 8bb17f0..40390ab 100644 --- a/src/util/virsysinfo.c +++ b/src/util/virsysinfo.c @@ -289,16 +289,15 @@ virSysinfoParseProcessor(const char *base, virSysinfoDefPtr ret) virSysinfoProcessorDefPtr processor; char *processor_type = NULL; - if (!(tmp_base = strstr(base, "Processor"))) + if (!(tmp_base = strstr(base, "model name")) && + !(tmp_base = strstr(base, "Processor"))) return 0; - base = tmp_base; - eol = strchr(base, '\n'); - cur = strchr(base, ':') + 1; + eol = strchr(tmp_base, '\n'); + cur = strchr(tmp_base, ':') + 1; virSkipSpaces(&cur); if (eol && VIR_STRNDUP(processor_type, cur, eol - cur) < 0) goto error; - base = cur; while ((tmp_base = strstr(base, "processor")) != NULL) { base = tmp_base; diff --git a/tests/sysinfodata/arm-rpi2cpuinfo.data b/tests/sysinfodata/arm-rpi2cpuinfo.data new file mode 100644 index 0000000..41fde00 --- /dev/null +++ b/tests/sysinfodata/arm-rpi2cpuinfo.data @@ -0,0 +1,43 @@ +processor : 0 +model name : ARMv7 Processor rev 5 (v7l) +BogoMIPS : 38.40 +Features : half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm +CPU implementer : 0x41 +CPU architecture: 7 +CPU variant : 0x0 +CPU part : 0xc07 +CPU revision : 5 + +processor : 1 +model name : ARMv7 Processor rev 5 (v7l) +BogoMIPS : 38.40 +Features : half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm +CPU implementer : 0x41 +CPU architecture: 7 +CPU variant : 0x0 +CPU part : 0xc07 +CPU revision : 5 + +processor : 2 +model name : ARMv7 Processor rev 5 (v7l) +BogoMIPS : 38.40 +Features : half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm +CPU implementer : 0x41 +CPU architecture: 7 +CPU variant : 0x0 +CPU part : 0xc07 +CPU revision : 5 + +processor : 3 +model name : ARMv7 Processor rev 5 (v7l) +BogoMIPS : 38.40 +Features : half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm +CPU implementer : 0x41 +CPU architecture: 7 +CPU variant : 0x0 +CPU part : 0xc07 +CPU revision : 5 + +Hardware : BCM2709 +Revision : a01041 +Serial : 00000000c9e9323d diff --git a/tests/sysinfodata/arm-rpi2sysinfo.expect b/tests/sysinfodata/arm-rpi2sysinfo.expect new file mode 100644 index 0000000..acb3ad9 --- /dev/null +++ b/tests/sysinfodata/arm-rpi2sysinfo.expect @@ -0,0 +1,18 @@ +<sysinfo type='smbios'> + <processor> + <entry name='socket_destination'>0</entry> + <entry name='type'>ARMv7 Processor rev 5 (v7l)</entry> + </processor> + <processor> + <entry name='socket_destination'>1</entry> + <entry name='type'>ARMv7 Processor rev 5 (v7l)</entry> + </processor> + <processor> + <entry name='socket_destination'>2</entry> + <entry name='type'>ARMv7 Processor rev 5 (v7l)</entry> + </processor> + <processor> + <entry name='socket_destination'>3</entry> + <entry name='type'>ARMv7 Processor rev 5 (v7l)</entry> + </processor> +</sysinfo> diff --git a/tests/sysinfotest.c b/tests/sysinfotest.c index d8266a7..74e5f71 100644 --- a/tests/sysinfotest.c +++ b/tests/sysinfotest.c @@ -166,11 +166,23 @@ VIRT_TEST_MAIN(test_x86) static int test_arm(void) { - return sysinfotest_run("arm sysinfo", - NULL, - NULL, - "/sysinfodata/armcpuinfo.data", - "/sysinfodata/armsysinfo.expect"); + int ret = EXIT_SUCCESS; + + if (sysinfotest_run("arm sysinfo", + NULL, + NULL, + "/sysinfodata/armcpuinfo.data", + "/sysinfodata/armsysinfo.expect") != EXIT_SUCCESS) + ret = EXIT_FAILURE; + + if (sysinfotest_run("Raspberry Pi 2 sysinfo", + NULL, + NULL, + "/sysinfodata/arm-rpi2cpuinfo.data", + "/sysinfodata/arm-rpi2sysinfo.expect") != EXIT_SUCCESS) + ret = EXIT_FAILURE; + + return ret; } VIRT_TEST_MAIN(test_arm) -- 2.3.6

3 2

[libvirt] [PATCH] util: Avoid shadow of 'ulong' in virMemoryMaxValue
by Peter Krempa 21 May '15

21 May '15

Old compilers whine: src/util/virutil.c: In function 'virMemoryMaxValue': src/util/virutil.c:2612: error: declaration of 'ulong' shadows a global declaration [-Wshadow] /usr/include/sys/types.h:151: error: shadowed declaration is here [-Wshadow] s/ulong/capped/ to work around the problem ---1 Pushed under the trivial && build-breaker rules. src/util/virutil.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/util/virutil.c b/src/util/virutil.c index 638d6e2..e479cce 100644 --- a/src/util/virutil.c +++ b/src/util/virutil.c @@ -2603,17 +2603,17 @@ virMemoryLimitIsSet(unsigned long long value) /** * virMemoryMaxValue * - * @ulong: whether the value must fit into unsigned long + * @capped: whether the value must fit into unsigned long * (long long is assumed otherwise) * * Returns the maximum possible memory value in bytes. */ unsigned long long -virMemoryMaxValue(bool ulong) +virMemoryMaxValue(bool capped) { /* On 32-bit machines, our bound is 0xffffffff * KiB. On 64-bit * machines, our bound is off_t (2^63). */ - if (ulong && sizeof(unsigned long) < sizeof(long long)) + if (capped && sizeof(unsigned long) < sizeof(long long)) return 1024ull * ULONG_MAX; else return LLONG_MAX; -- 2.3.5

1 0

[libvirt] [PATCH 0/3] Support changing user passwords in the guest
by Ján Tomko 21 May '15

21 May '15

The guest-set-user-password command was added to qemu-agent in the 2.3.0 release by Daniel Berrange (some of you might have heard of him): commit 215a2771a7b6b29037ee8deba484815d816b6fdd qga: add guest-set-user-password command http://git.qemu.org/?p=qemu.git;a=commitdiff;h=215a277 This mini-series introduces a libvirt API, set_password ACL, a virsh command and wires up the API in the QEMU driver. Any bikesh^Wsuggestions regarding the API naming and the order of the virsh parameters are welcome. Ján Tomko (3): Introduce virDomainSetUserPassword API virsh: add set-user-password command qemu: wire up virDomainSetUserPassword include/libvirt/libvirt-domain.h | 9 +++++ src/access/viraccessperm.c | 2 +- src/access/viraccessperm.h | 6 ++++ src/driver-hypervisor.h | 7 ++++ src/libvirt-domain.c | 47 +++++++++++++++++++++++++ src/libvirt_public.syms | 5 +++ src/qemu/qemu_agent.c | 39 +++++++++++++++++++++ src/qemu/qemu_agent.h | 4 +++ src/qemu/qemu_driver.c | 55 +++++++++++++++++++++++++++++ src/remote/remote_driver.c | 1 + src/remote/remote_protocol.x | 15 +++++++- src/remote_protocol-structs | 7 ++++ tools/virsh-domain.c | 76 ++++++++++++++++++++++++++++++++++++++++ tools/virsh.pod | 11 ++++++ 14 files changed, 282 insertions(+), 2 deletions(-) -- 2.0.5

3 9

[libvirt] [PATCH 0/3] parallels: fix prlsdkLoadDomain and prlsdkAddDomain
by Maxim Nestratov 21 May '15

21 May '15

Maxim Nestratov (3): parallels: fix prlsdkAddDomain parallels: move up updating autostart parameter in prlsdkLoadDomain parallels: fix possible crash in case of errors in prlsdkLoadDomain src/parallels/parallels_sdk.c | 39 ++++++++++++++++++++++++++------------- 1 file changed, 26 insertions(+), 13 deletions(-) -- 2.1.0

5 15

[libvirt] [PATCH 0/2] Add and use for virDomainDiskIndexBy*
by Jiri Denemark 21 May '15

21 May '15

Sometimes the only thing we need is the pointer to virDomainDiskDef and having to call virDomainDiskIndexBy* APIs, storing the disk index, and looking it up in the disks array is ugly. After this patch, we can just call virDomainDiskBy* and get the pointer in one step. Jiri Denemark (2): Add wrappers for virDomainDiskIndexBy* Use virDomainDiskByName where appropriate src/conf/domain_conf.c | 20 +++++++++++++++ src/conf/domain_conf.h | 8 ++++++ src/libvirt_private.syms | 2 ++ src/libxl/libxl_driver.c | 4 +-- src/lxc/lxc_driver.c | 10 +++----- src/qemu/qemu_agent.c | 9 ++++--- src/qemu/qemu_blockjob.c | 5 ++-- src/qemu/qemu_driver.c | 62 ++++++++++++++++++----------------------------- src/qemu/qemu_migration.c | 6 ++--- src/qemu/qemu_process.c | 23 ++++-------------- 10 files changed, 72 insertions(+), 77 deletions(-) -- 2.4.1

4 9

[libvirt] [PATCH] threadpool: Switch to detached threads
by Jiri Denemark 21 May '15

21 May '15

Using joinable threads does not help anything, but it can lead to memory leaks. When a worker thread exits, it decreases nWorkers or nPrioWorkers and once both nWorkers and nPrioWorkers are zero (i.e., the last worker is gone), quit_cond is signaled. When freeing the pool we first tell all threads to die and then we are waiting for both nWorkers and nPrioWorkers to become zero. At this point we already know all threads are gone. So the only reason for calling virThreadJoin of all workers is to free the memory allocated for joinable threads. If we avoid allocating this memory, we don't need to take care of freeing it. Moreover, any memory associated with a worker thread which died before we asked it to die (e.g., because virCondWait failed in the thread) would be lost anyway since virThreadPoolFree calls virThreadJoin only for threads which were running at the time virThreadPoolFree was called. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/util/virthreadpool.c | 17 +++-------------- 1 file changed, 3 insertions(+), 14 deletions(-) diff --git a/src/util/virthreadpool.c b/src/util/virthreadpool.c index ed22486..f640448 100644 --- a/src/util/virthreadpool.c +++ b/src/util/virthreadpool.c @@ -201,7 +201,7 @@ virThreadPoolNewFull(size_t minWorkers, data->cond = &pool->cond; if (virThreadCreateFull(&pool->workers[i], - true, + false, virThreadPoolWorker, pool->jobFuncName, true, @@ -225,7 +225,7 @@ virThreadPoolNewFull(size_t minWorkers, data->priority = true; if (virThreadCreateFull(&pool->prioWorkers[i], - true, + false, virThreadPoolWorker, pool->jobFuncName, true, @@ -249,16 +249,11 @@ void virThreadPoolFree(virThreadPoolPtr pool) { virThreadPoolJobPtr job; bool priority = false; - size_t i; - size_t nWorkers; - size_t nPrioWorkers; if (!pool) return; virMutexLock(&pool->mutex); - nWorkers = pool->nWorkers; - nPrioWorkers = pool->nPrioWorkers; pool->quit = true; if (pool->nWorkers > 0) virCondBroadcast(&pool->cond); @@ -275,12 +270,6 @@ void virThreadPoolFree(virThreadPoolPtr pool) VIR_FREE(job); } - for (i = 0; i < nWorkers; i++) - virThreadJoin(&pool->workers[i]); - - for (i = 0; i < nPrioWorkers; i++) - virThreadJoin(&pool->prioWorkers[i]); - VIR_FREE(pool->workers); virMutexUnlock(&pool->mutex); virMutexDestroy(&pool->mutex); @@ -338,7 +327,7 @@ int virThreadPoolSendJob(virThreadPoolPtr pool, data->cond = &pool->cond; if (virThreadCreateFull(&pool->workers[pool->nWorkers - 1], - true, + false, virThreadPoolWorker, pool->jobFuncName, true, -- 2.4.1

2 4

[libvirt] [PATCH] qemu: Log error if domain uses security driver which is not loaded
by Erik Skultety 21 May '15

21 May '15

When starting a domain, if we find out domain requests security drivers we do not have loaded, we fail. However we don't check for this during reconnect, so any operation relying on security driver functionality would fail. If someone e.g. starts a domain with selinux driver loaded, then they turn off the security driver in config, restart the daemon and call dump/save/.., QEMU returns an error. As we shouldn't kill the domain, we should at least log an error to let the user know that domain reconnect wasn't completely clean. https://bugzilla.redhat.com/show_bug.cgi?id=1183893 --- src/qemu/qemu_process.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 56719eb..8da79e5 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3747,6 +3747,12 @@ qemuProcessReconnect(void *opaque) if ((qemuDomainAssignAddresses(obj->def, priv->qemuCaps, obj)) < 0) goto error; + /* if domain requests security driver we haven't loaded, report error, but + * do not kill the domain + */ + ignore_value(virSecurityManagerCheckAllLabel(driver->securityManager, + obj->def)); + if (virSecurityManagerReserveLabel(driver->securityManager, obj->def, obj->pid) < 0) goto error; -- 1.9.3

3 2

[libvirt] [PATCH 0/3] virsh: Fix and enhance handling of numeric options
by Andrea Bolognani 21 May '15

21 May '15

The first commit improves error messages by making them much more consistent and a little bit more helpful to the user. The other two commits contain smaller fixes and enhancement. Andrea Bolognani (3): virsh: Improve error message on integer value parsing failure. virsh: Fix dommemstat --period option type. virsh: Improve handling of send-process-signal --pid. tests/vcpupin | 4 +- tools/virsh-domain-monitor.c | 12 +++-- tools/virsh-domain.c | 114 ++++++++++++++++++++++++++++++------------- tools/virsh-host.c | 44 ++++++++++++----- tools/virsh-interface.c | 4 +- tools/virsh-network.c | 4 +- tools/virsh-volume.c | 16 ++++-- 7 files changed, 141 insertions(+), 57 deletions(-) -- 2.1.0

2 6

[libvirt] [PATCH 0/2] Fix virsh nodeset on shutoff domains
by Michal Privoznik 21 May '15

21 May '15

This goes on the top of Martin's (ACKed) patch: https://www.redhat.com/archives/libvir-list/2015-May/msg00585.html Michal Privoznik (2): virDomainNumatuneGetMode: Report if numatune was defined qemuDomainGetNumaParameters: Don't report spurious info src/conf/numa_conf.c | 38 ++++++++++++++++++++++-------- src/conf/numa_conf.h | 5 ++-- src/lxc/lxc_cgroup.c | 5 ++-- src/lxc/lxc_controller.c | 30 ++++++++++++------------ src/parallels/parallels_sdk.c | 5 ++-- src/qemu/qemu_cgroup.c | 15 +++++++----- src/qemu/qemu_command.c | 4 ++-- src/qemu/qemu_driver.c | 54 ++++++++++++++++++++++++++++--------------- src/qemu/qemu_process.c | 28 +++++++++++----------- 9 files changed, 113 insertions(+), 71 deletions(-) -- 2.3.6

2 6

[libvirt] [libvirt-python 0/3] reproducible builds
by Guido Günther 21 May '15

21 May '15

Debian tries to make builds reporducible[0] therefore we need predictable output (i.e. no timestamps, predictable iteration over hashes, etc.). This patch (mostly by Chris Lamb) to libvirt-python makes the necessary changes. [0] https://wiki.debian.org/ReproducibleBuilds Guido Günther (3): Sort dictionary keys Simplify sorting Sort tuples on both items generator.py | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-) -- 2.1.4

2 7