September 2013 - Devel - libvirt List Archives

[libvirt] test failure on rawhide
by Eric Blake 12 Sep '13

12 Sep '13

Just noticed this on latest rawhide (Fedora 21): Looks like the bug might not be libvirt's fault (as the assertion failure is not in our source file), but I'm out of time to investigate further today. $ VIR_TEST_DEBUG=1 ./virdrivermoduletest TEST: virdrivermoduletest 1) Test driver "network" ... OK 2) Test driver "storage" ... OK 3) Test driver "nodedev" ... OK 4) Test driver "secret" ... OK 5) Test driver "nwfilter" ... OK 6) Test driver "interface" ... lt-virdrivermoduletest: route/tc.c:973: rtnl_tc_register: Assertion `0' failed. Aborted -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org

3 6

[libvirt] [v0.9.12-maint 00/11] Debian's 0.9.12 patches
by Guido Günther 12 Sep '13

12 Sep '13

These are the patches Debian is currently carrying on 0.9.12. Most are straight cherry-picks. Since we're maintaining 0.9.12 for our current stable release I'm happy to push these to v0.9.12-maint. Ferenc Wágner (1): Fix race condition when destroying guests Guido Günther (1): Include stdint.h for uint32_t Jiri Denemark (3): daemon: Fix crash in virTypedParameterArrayClear Revert "rpc: Discard non-blocking calls only when necessary" qemu: Add support for -no-user-config Luca Tettamanti (1): Make sure regree is called close to it's usage Martin Kletzander (1): security: Fix libvirtd crash possibility Peter Krempa (4): qemu: Fix off-by-one error while unescaping monitor strings rpc: Fix crash on error paths of message dispatching conf: Remove callback from stream when freeing entries in console hash conf: Remove console stream callback only when freeing console helper cfg.mk | 3 +- daemon/remote.c | 16 +- src/conf/virconsole.c | 13 ++ src/qemu/qemu_capabilities.c | 7 +- src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 11 +- src/qemu/qemu_driver.c | 21 ++- src/qemu/qemu_monitor.c | 11 +- src/rpc/virnetclient.c | 21 +-- src/rpc/virnetserverclient.c | 3 + src/rpc/virnetserverprogram.c | 11 +- src/storage/storage_backend_logical.c | 5 +- src/util/virnetlink.h | 2 + tests/qemuhelpdata/qemu-1.1 | 268 ++++++++++++++++++++++++++++++++++ tests/qemuhelpdata/qemu-1.1-device | 160 ++++++++++++++++++++ tests/qemuhelptest.c | 75 ++++++++++ 16 files changed, 586 insertions(+), 42 deletions(-) create mode 100644 tests/qemuhelpdata/qemu-1.1 create mode 100644 tests/qemuhelpdata/qemu-1.1-device -- 1.8.4.rc3

2 1

[libvirt] [v0.9.12-maint 11/11] Fix race condition when destroying guests
by Ferenc Wágner 12 Sep '13

12 Sep '13

Backport of 81621f3e6e45e8681cc18ae49404736a0e772a11 and f1b4021b38f9485c50d386af6f682ecfc8025af5 to fix a race (resulting in a segfault) when destroying domains. --- src/qemu/qemu_driver.c | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 0053ed1..c0b4707 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -1827,6 +1827,12 @@ qemuDomainDestroyFlags(virDomainPtr dom, qemuDomainSetFakeReboot(driver, vm, false); + + /* We need to prevent monitor EOF callback from doing our work (and sending + * misleading events) while the vm is unlocked inside BeginJob/ProcessKill API + */ + priv->beingDestroyed = true; + /* Although qemuProcessStop does this already, there may * be an outstanding job active. We want to make sure we * can kill the process even if a job is active. Killing @@ -1834,19 +1840,20 @@ qemuDomainDestroyFlags(virDomainPtr dom, */ if (flags & VIR_DOMAIN_DESTROY_GRACEFUL) { if (qemuProcessKill(driver, vm, 0) < 0) { + priv->beingDestroyed = false; qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", _("failed to kill qemu process with SIGTERM")); goto cleanup; } } else { - ignore_value(qemuProcessKill(driver, vm, VIR_QEMU_PROCESS_KILL_FORCE)); + if (qemuProcessKill(driver, vm, VIR_QEMU_PROCESS_KILL_FORCE) < 0) { + priv->beingDestroyed = false; + qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", + _("failed to kill qemu process with SIGTERM")); + goto cleanup; + } } - /* We need to prevent monitor EOF callback from doing our work (and sending - * misleading events) while the vm is unlocked inside BeginJob API - */ - priv->beingDestroyed = true; - if (qemuDomainObjBeginJobWithDriver(driver, vm, QEMU_JOB_DESTROY) < 0) goto cleanup; -- 1.8.4.rc3

2 1

[libvirt] [v0.9.12-maint 10/11] conf: Remove console stream callback only when freeing console helper
by Peter Krempa 12 Sep '13

12 Sep '13

Commit ba226d334acbc49f6751b430e0c4e00f69eef6bf tried to fix crash of the daemon when a domain with an open console was destroyed. The fix was wrong as it tried to remove the callback also when the stream was aborted, where at that point the fd stream driver was already freed and removed. This patch clears the callbacks with a helper right before the hash is freed, so that it doesn't interfere with other codepaths where the stream object is freed. --- src/conf/virconsole.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) diff --git a/src/conf/virconsole.c b/src/conf/virconsole.c index e665149..01f1c84 100644 --- a/src/conf/virconsole.c +++ b/src/conf/virconsole.c @@ -222,9 +222,6 @@ static void virConsoleHashEntryFree(void *data, const char *pty = name; virStreamPtr st = data; - /* remove callback from stream */ - virFDStreamSetInternalCloseCb(st, NULL, NULL, NULL); - /* free stream reference */ virStreamFree(st); @@ -293,6 +290,18 @@ error: } /** + * Helper to clear stream callbacks when freeing the hash + */ +static void virConsoleFreeClearCallbacks(void *payload, + const void *name ATTRIBUTE_UNUSED, + void *data ATTRIBUTE_UNUSED) +{ + virStreamPtr st = payload; + + virFDStreamSetInternalCloseCb(st, NULL, NULL, NULL); +} + +/** * Free structures for handling open console streams. * * @cons Pointer to the private structure. @@ -303,6 +312,7 @@ void virConsoleFree(virConsolesPtr cons) return; virMutexLock(&cons->lock); + virHashForEach(cons->hash, virConsoleFreeClearCallbacks, NULL); virHashFree(cons->hash); virMutexUnlock(&cons->lock); virMutexDestroy(&cons->lock); -- 1.8.4.rc3

2 1

[libvirt] [v0.9.12-maint 09/11] conf: Remove callback from stream when freeing entries in console hash
by Peter Krempa 12 Sep '13

12 Sep '13

When a domain has a active console connection and is destroyed the callback is called on private data that no longer exist causing a segfault. --- src/conf/virconsole.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/conf/virconsole.c b/src/conf/virconsole.c index 443d80d..e665149 100644 --- a/src/conf/virconsole.c +++ b/src/conf/virconsole.c @@ -222,6 +222,9 @@ static void virConsoleHashEntryFree(void *data, const char *pty = name; virStreamPtr st = data; + /* remove callback from stream */ + virFDStreamSetInternalCloseCb(st, NULL, NULL, NULL); + /* free stream reference */ virStreamFree(st); -- 1.8.4.rc3

2 1

[libvirt] [v0.9.12-maint 03/11] Include stdint.h for uint32_t
by Guido Günther 12 Sep '13

12 Sep '13

to fix the build on kFreeBSD. References: https://buildd.debian.org/status/fetch.php?pkg=libvirt&arch=kfreebsd-amd64&… --- src/util/virnetlink.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/util/virnetlink.h b/src/util/virnetlink.h index bafe8ca..acd88b7 100644 --- a/src/util/virnetlink.h +++ b/src/util/virnetlink.h @@ -21,8 +21,10 @@ # define __VIR_NETLINK_H__ # include "config.h" +# include <stdint.h> # include "internal.h" + # if defined(__linux__) && defined(HAVE_LIBNL) # include <netlink/msg.h> -- 1.8.4.rc3

2 1

[libvirt] [PATCH] Fix polkit permission names for storage pools, vols & node devices
by Daniel P. Berrange 12 Sep '13

12 Sep '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> The polkit access driver used the wrong permission names for checks on storage pools, volumes and node devices. This led to them always being denied access. Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/access/viraccessdriverpolkit.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/access/viraccessdriverpolkit.c b/src/access/viraccessdriverpolkit.c index 4c76e64..b472bc3 100644 --- a/src/access/viraccessdriverpolkit.c +++ b/src/access/viraccessdriverpolkit.c @@ -248,7 +248,7 @@ virAccessDriverPolkitCheckNodeDevice(virAccessManagerPtr manager, }; return virAccessDriverPolkitCheck(manager, - "nodedevice", + "node-device", virAccessPermNodeDeviceTypeToString(perm), attrs); } @@ -355,7 +355,7 @@ virAccessDriverPolkitCheckStoragePool(virAccessManagerPtr manager, virUUIDFormat(pool->uuid, uuidstr); return virAccessDriverPolkitCheck(manager, - "pool", + "storage-pool", virAccessPermStoragePoolTypeToString(perm), attrs); } @@ -379,7 +379,7 @@ virAccessDriverPolkitCheckStorageVol(virAccessManagerPtr manager, virUUIDFormat(pool->uuid, uuidstr); return virAccessDriverPolkitCheck(manager, - "vol", + "storage-vol", virAccessPermStorageVolTypeToString(perm), attrs); } -- 1.8.3.1

2 3

[libvirt] [PATCH] Fix launching of VMs on when only logind part of systemd is present
by Daniel P. Berrange 12 Sep '13

12 Sep '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/util/vircgroup.c | 5 +++- tests/vircgroupmock.c | 40 ++++++++++++++++++++++++--- tests/vircgrouptest.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 116 insertions(+), 4 deletions(-) diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index a260356..626bbc6 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -91,7 +91,10 @@ virCgroupAvailable(void) return false; while (getmntent_r(mounts, &entry, buf, sizeof(buf)) != NULL) { - if (STREQ(entry.mnt_type, "cgroup")) { + /* We're looking for at least one 'cgroup' fs mount, + * which is *not* a named mount. */ + if (STREQ(entry.mnt_type, "cgroup") && + !strstr(entry.mnt_opts, "name=")) { ret = true; break; } diff --git a/tests/vircgroupmock.c b/tests/vircgroupmock.c index d83496c..adc1718 100644 --- a/tests/vircgroupmock.c +++ b/tests/vircgroupmock.c @@ -124,6 +124,27 @@ const char *proccgroupsallinone = "devices 6 1 1\n" "blkio 6 1 1\n"; +const char *procmountslogind = + "none /not/really/sys/fs/cgroup tmpfs rw,rootcontext=system_u:object_r:sysfs_t:s0,seclabel,relatime,size=4k,mode=755 0 0\n" + "systemd /not/really/sys/fs/cgroup/systemd cgroup rw,nosuid,nodev,noexec,relatime,name=systemd 0 0\n"; + +const char *procselfcgroupslogind = + "1:name=systemd:/\n"; + +const char *proccgroupslogind = + "#subsys_name hierarchy num_cgroups enabled\n" + "cpuset 0 1 1\n" + "cpu 0 1 1\n" + "cpuacct 0 1 1\n" + "memory 0 1 0\n" + "devices 0 1 1\n" + "freezer 0 1 1\n" + "net_cls 0 1 1\n" + "blkio 0 1 1\n" + "perf_event 0 1 1\n"; + + + static int make_file(const char *path, const char *name, const char *value) @@ -400,18 +421,25 @@ static void init_sysfs(void) FILE *fopen(const char *path, const char *mode) { const char *mock; - bool allinone = false; + bool allinone = false, logind = false; init_syms(); mock = getenv("VIR_CGROUP_MOCK_MODE"); - if (mock && STREQ(mock, "allinone")) - allinone = true; + if (mock) { + if (STREQ(mock, "allinone")) + allinone = true; + else if (STREQ(mock, "logind")) + logind = true; + } if (STREQ(path, "/proc/mounts")) { if (STREQ(mode, "r")) { if (allinone) return fmemopen((void *)procmountsallinone, strlen(procmountsallinone), mode); + else if (logind) + return fmemopen((void *)procmountslogind, + strlen(procmountslogind), mode); else return fmemopen((void *)procmounts, strlen(procmounts), mode); } else { @@ -424,6 +452,9 @@ FILE *fopen(const char *path, const char *mode) if (allinone) return fmemopen((void *)proccgroupsallinone, strlen(proccgroupsallinone), mode); + else if (logind) + return fmemopen((void *)proccgroupslogind, + strlen(proccgroupslogind), mode); else return fmemopen((void *)proccgroups, strlen(proccgroups), mode); } else { @@ -436,6 +467,9 @@ FILE *fopen(const char *path, const char *mode) if (allinone) return fmemopen((void *)procselfcgroupsallinone, strlen(procselfcgroupsallinone), mode); + else if (logind) + return fmemopen((void *)procselfcgroupslogind, + strlen(procselfcgroupslogind), mode); else return fmemopen((void *)procselfcgroups, strlen(procselfcgroups), mode); } else { diff --git a/tests/vircgrouptest.c b/tests/vircgrouptest.c index f12587c..570e061 100644 --- a/tests/vircgrouptest.c +++ b/tests/vircgrouptest.c @@ -109,6 +109,16 @@ const char *mountsAllInOne[VIR_CGROUP_CONTROLLER_LAST] = { [VIR_CGROUP_CONTROLLER_BLKIO] = "/not/really/sys/fs/cgroup", [VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL, }; +const char *mountsLogind[VIR_CGROUP_CONTROLLER_LAST] = { + [VIR_CGROUP_CONTROLLER_CPU] = NULL, + [VIR_CGROUP_CONTROLLER_CPUACCT] = NULL, + [VIR_CGROUP_CONTROLLER_CPUSET] = NULL, + [VIR_CGROUP_CONTROLLER_MEMORY] = NULL, + [VIR_CGROUP_CONTROLLER_DEVICES] = NULL, + [VIR_CGROUP_CONTROLLER_FREEZER] = NULL, + [VIR_CGROUP_CONTROLLER_BLKIO] = NULL, + [VIR_CGROUP_CONTROLLER_SYSTEMD] = "/not/really/sys/fs/cgroup/systemd", +}; const char *links[VIR_CGROUP_CONTROLLER_LAST] = { [VIR_CGROUP_CONTROLLER_CPU] = "/not/really/sys/fs/cgroup/cpu", @@ -132,6 +142,17 @@ const char *linksAllInOne[VIR_CGROUP_CONTROLLER_LAST] = { [VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL, }; +const char *linksLogind[VIR_CGROUP_CONTROLLER_LAST] = { + [VIR_CGROUP_CONTROLLER_CPU] = NULL, + [VIR_CGROUP_CONTROLLER_CPUACCT] = NULL, + [VIR_CGROUP_CONTROLLER_CPUSET] = NULL, + [VIR_CGROUP_CONTROLLER_MEMORY] = NULL, + [VIR_CGROUP_CONTROLLER_DEVICES] = NULL, + [VIR_CGROUP_CONTROLLER_FREEZER] = NULL, + [VIR_CGROUP_CONTROLLER_BLKIO] = NULL, + [VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL, +}; + static int testCgroupNewForSelf(const void *args ATTRIBUTE_UNUSED) { @@ -466,6 +487,48 @@ cleanup: } +static int testCgroupNewForSelfLogind(const void *args ATTRIBUTE_UNUSED) +{ + virCgroupPtr cgroup = NULL; + int ret = -1; + const char *placement[VIR_CGROUP_CONTROLLER_LAST] = { + [VIR_CGROUP_CONTROLLER_CPU] = NULL, + [VIR_CGROUP_CONTROLLER_CPUACCT] = NULL, + [VIR_CGROUP_CONTROLLER_CPUSET] = NULL, + [VIR_CGROUP_CONTROLLER_MEMORY] = NULL, + [VIR_CGROUP_CONTROLLER_DEVICES] = NULL, + [VIR_CGROUP_CONTROLLER_FREEZER] = NULL, + [VIR_CGROUP_CONTROLLER_BLKIO] = NULL, + [VIR_CGROUP_CONTROLLER_SYSTEMD] = "/", + }; + + if (virCgroupNewSelf(&cgroup) < 0) { + fprintf(stderr, "Cannot create cgroup for self\n"); + goto cleanup; + } + + ret = validateCgroup(cgroup, "", mountsLogind, linksLogind, placement); + +cleanup: + virCgroupFree(&cgroup); + return ret; +} + + +static int testCgroupAvailable(const void *args) +{ + bool got = virCgroupAvailable(); + bool want = args == (void*)0x1; + + if (got != want) { + fprintf(stderr, "Expected cgroup %savailable, but state was wrong\n", + want ? "" : "not "); + return -1; + } + + return 0; +} + # define FAKESYSFSDIRTEMPLATE abs_builddir "/fakesysfsdir-XXXXXX" @@ -505,9 +568,21 @@ mymain(void) if (virtTestRun("New cgroup for domain partition escaped", 1, testCgroupNewForPartitionDomainEscaped, NULL) < 0) ret = -1; + if (virtTestRun("Cgroup available", 1, testCgroupAvailable, (void*)0x1) < 0) + ret = -1; + setenv("VIR_CGROUP_MOCK_MODE", "allinone", 1); if (virtTestRun("New cgroup for self (allinone)", 1, testCgroupNewForSelfAllInOne, NULL) < 0) ret = -1; + if (virtTestRun("Cgroup available", 1, testCgroupAvailable, (void*)0x1) < 0) + ret = -1; + unsetenv("VIR_CGROUP_MOCK_MODE"); + + setenv("VIR_CGROUP_MOCK_MODE", "logind", 1); + if (virtTestRun("New cgroup for self (logind)", 1, testCgroupNewForSelfLogind, NULL) < 0) + ret = -1; + if (virtTestRun("Cgroup available", 1, testCgroupAvailable, (void*)0x0) < 0) + ret = -1; unsetenv("VIR_CGROUP_MOCK_MODE"); if (getenv("LIBVIRT_SKIP_CLEANUP") == NULL) -- 1.8.3.1

3 2

[libvirt] [PATCH 0/4] Introduce APIs to extract DHCP leases info
by Nehal J Wani 12 Sep '13

12 Sep '13

This API returns the information stored in the DHCP leases file created by dnsmasq for a given virtual network. It contacts the bridge network driver, which parses the leases file. It supports two methods: 1. Return info for all network interfaces connected to a given virtual network 2. Return information for a particular network interface in a given virtual network by providing its MAC Address * The need for these APIs were result of a RFC was proposed on the list. Refer: http://www.redhat.com/archives/libvir-list/2013-July/msg01603.html Nehal J Wani (4): net-dhcp-leases: Implement the public APIs net-dhcp-leases: Implement the remote protocol net-dhcp-leases: Private implementation inside network driver net-dhcp-leases: Add virsh support daemon/remote.c | 128 +++++++++++++++++++++++++ include/libvirt/libvirt.h.in | 22 +++++ python/generator.py | 3 + src/driver.h | 13 +++ src/libvirt.c | 166 ++++++++++++++++++++++++++++++++ src/libvirt_public.syms | 7 ++ src/network/bridge_driver.c | 222 +++++++++++++++++++++++++++++++++++++++++++ src/remote/remote_driver.c | 117 +++++++++++++++++++++++ src/remote/remote_protocol.x | 48 +++++++++- src/remote_protocol-structs | 27 ++++++ src/rpc/gendispatch.pl | 1 + tools/virsh-network.c | 92 ++++++++++++++++++ tools/virsh.pod | 6 ++ 13 files changed, 851 insertions(+), 1 deletion(-) -- 1.7.11.7

3 11

[libvirt] cgroups issues when logind is running without systemd acting as init system
by Laurent Bigonville 12 Sep '13

12 Sep '13

Hi, As explained in a bug on the debian BTS[0], I'm getting the following backtrace on my machine when trying to start a domain: Erreur lors du démarrage du domaine: At least one cgroup controller is required: No such device or address Traceback (most recent call last): File "/usr/share/virt-manager/virtManager/asyncjob.py", line 96, in cb_wrapper callback(asyncjob, *args, **kwargs) File "/usr/share/virt-manager/virtManager/asyncjob.py", line 117, in tmpcb callback(*args, **kwargs) File "/usr/share/virt-manager/virtManager/domain.py", line 1160, in startup self._backend.create() File "/usr/lib/python2.7/dist-packages/libvirt.py", line 697, in create if ret == -1: raise libvirtError ('virDomainCreate() failed', dom=self) libvirtError: At least one cgroup controller is required: No such device or address This is only happening when /sys/fs/cgroup/systemd is being mounted by logind while systemd is not acting as the init system. I've been asked by danpb to report the following information on this mailing list: $ cat /proc/mounts|grep cgroup none /sys/fs/cgroup tmpfs rw,rootcontext=system_u:object_r:sysfs_t:s0,seclabel,relatime,size=4k,mode=755 0 0 systemd /sys/fs/cgroup/systemd cgroup rw,nosuid,nodev,noexec,relatime,name=systemd 0 0 $ cat /proc/cgroups #subsys_name hierarchy num_cgroups enabled cpuset 0 1 1 cpu 0 1 1 cpuacct 0 1 1 memory 0 1 0 devices 0 1 1 freezer 0 1 1 net_cls 0 1 1 blkio 0 1 1 perf_event 0 1 1 $ cat /proc/$(pidof libvirtd)/cgroup 1:name=systemd:/ I hope this will be enough to fix this issue. Kind regards, Laurent Bigonville [0]http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979

1 0