- Devel - libvirt List Archives

[libvirt] [PATCH] Remove bogus log warning lines when launching QEMU
by Daniel P. Berrange 28 Jan '11

28 Jan '11

The refactoring of QEMU command startup was comitted with a couple of VIR_WARN lines left in from debugging. * src/qemu/qemu_driver.c: Remove log warning lines --- src/qemu/qemu_driver.c | 2 -- 1 files changed, 0 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 98eb91f..2e8f4f9 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2865,7 +2865,6 @@ static int qemudStartVMDaemon(virConnectPtr conn, if (driver->clearEmulatorCapabilities) virCommandClearCaps(cmd); - VIR_WARN("Executing %s", vm->def->emulator); virCommandSetPreExecHook(cmd, qemudSecurityHook, &hookData); virCommandSetOutputFD(cmd, &logfile); @@ -2875,7 +2874,6 @@ static int qemudStartVMDaemon(virConnectPtr conn, virCommandDaemonize(cmd); ret = virCommandRun(cmd, NULL); - VIR_WARN("Executing done %s", vm->def->emulator); VIR_FREE(pidfile); /* wait for qemu process to to show up */ -- 1.7.3.4

1 0

[libvirt] beginner's project
by Eric Blake 28 Jan '11

28 Jan '11

This idea sprang from an IRC conversation, but I don't have time to do it myself at the moment. If you're looking for a great way to help development, it would be nice to teach tests/testutils.c:virtTestLoadFile how to read a file with backslash-newline pairs and flatten those into a single line. Then rewrite all the tests/*data/*.args files to use line continuations, rather than being monstrous 1000-character lines. That way, git send-email will quit griping about the inability to send patches to those files. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 0

[libvirt] [PATCH] docs: expand the man page text for virsh setmaxmem
by Justin Clift 28 Jan '11

28 Jan '11

Addresses BZ # 622534: https://bugzilla.redhat.com/show_bug.cgi?id=622534 --- tools/virsh.pod | 17 ++++++++++++++--- 1 files changed, 14 insertions(+), 3 deletions(-) diff --git a/tools/virsh.pod b/tools/virsh.pod index 4e8b295..0d37512 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -584,9 +584,20 @@ update the XML <currentMemory> element. =item B<setmaxmem> I<domain-id> B<kilobytes> -Change the maximum memory allocation limit in the guest domain. This should -not change the current memory use. The memory limit is specified in -kilobytes. +Change the maximum memory allocation limit for an inactive guest domain. + +This command works for at least the Xen and vSphere/ESX hypervisors, +but not for QEMU/KVM. + +Some hypervisors require a larger granularity than kilobytes, rounding down +or rejecting requests that are not an even multiple of the desired amount. +vSphere/ESX is one of these, requiring the parameter to be evenly divisible +by 4MB. For example with vSphere/ESX, 262144 (256MB) is valid as it's a +multiple of 4MB; 263168 (257MB) is not valid as it's not a multiple of 4MB; +266240 (260MB) is also valid, as it's a multiple of 4MB. + +Note, to change the maximum memory allocation for a QEMU/KVM guest domain, +use the virsh B<edit> command instead to update its XML <memory> element. =item B<memtune> I<domain-id> optional I<--hard-limit> B<kilobytes> optional I<--soft-limit> B<kilobytes> optional I<--swap-hard-limit> -- 1.7.3.5

3 6

[libvirt] [PATCH] event: fix two event-handling bugs
by Eric Blake 28 Jan '11

28 Jan '11

Regression introduced in commit e6b68d7. Prior to that point, handlesAlloc was always a multiple of EVENT_ALLOC_EXTENT, and was an integer (so even if the subtraction wrapped, a negative value was less than the count and did not try to free the handles array). But after that point, VIR_RESIZE_N made handlesAlloc grow geometrically, so handlesAlloc could be 49 when handlesCount was 47, while still freeing off only 10 at a time, and eventually reach handlesAlloc 7 and handlesCount 1. Since (size_t)(7 - 10) is indeed greater than 1, this then tried to free 10 elements, which had the awful effect of nuking the handles array while there were still live handles. Which leads to crashes like this: https://bugzilla.redhat.com/show_bug.cgi?id=670848 * daemon/event.c (virEventDispatchHandles): Cache watch while the lock is still held, as eventLoop.handles might be relocated outside the lock. (virEventCleanupHandles): Avoid integer wrap-around causing us to delete the entire handles array while handles are still active. --- Thanks to Dave Allan for letting me use his faqemu setup for testing this. Basically, starting 60 faqemu followed by stopping them all was a reliable way to trigger the handle cleanup integer wraparound. daemon/event.c | 12 +++++++----- 1 files changed, 7 insertions(+), 5 deletions(-) diff --git a/daemon/event.c b/daemon/event.c index 89ca9f0..9cad466 100644 --- a/daemon/event.c +++ b/daemon/event.c @@ -1,7 +1,7 @@ /* * event.c: event loop for monitoring file handles * - * Copyright (C) 2007, 2010 Red Hat, Inc. + * Copyright (C) 2007, 2010-2011 Red Hat, Inc. * Copyright (C) 2007 Daniel P. Berrange * * This library is free software; you can redistribute it and/or @@ -435,6 +435,7 @@ static int virEventDispatchTimeouts(void) { */ static int virEventDispatchHandles(int nfds, struct pollfd *fds) { int i, n; + int watch; DEBUG("Dispatch %d", nfds); /* NB, use nfds not eventLoop.handlesCount, because new @@ -463,9 +464,9 @@ static int virEventDispatchHandles(int nfds, struct pollfd *fds) { EVENT_DEBUG("Dispatch n=%d f=%d w=%d e=%d %p", i, fds[n].fd, eventLoop.handles[i].watch, fds[n].revents, eventLoop.handles[i].opaque); + watch = eventLoop.handles[i].watch; virMutexUnlock(&eventLoop.lock); - (cb)(eventLoop.handles[i].watch, - fds[n].fd, hEvents, opaque); + (cb)(watch, fds[n].fd, hEvents, opaque); virMutexLock(&eventLoop.lock); } } @@ -542,9 +543,10 @@ static int virEventCleanupHandles(void) { } /* Release some memory if we've got a big chunk free */ - if ((eventLoop.handlesAlloc - EVENT_ALLOC_EXTENT) > eventLoop.handlesCount) { + if (eventLoop.handlesAlloc - eventLoop.handlesCount > EVENT_ALLOC_EXTENT) { EVENT_DEBUG("Releasing %zu out of %zu handles slots used, releasing %d", - eventLoop.handlesCount, eventLoop.handlesAlloc, EVENT_ALLOC_EXTENT); + eventLoop.handlesCount, eventLoop.handlesAlloc, + EVENT_ALLOC_EXTENT); VIR_SHRINK_N(eventLoop.handles, eventLoop.handlesAlloc, EVENT_ALLOC_EXTENT); } -- 1.7.3.4

3 8

[libvirt] [PATCH] Prevent overfilling of self-pipe in python event loop
by Daniel P. Berrange 27 Jan '11

27 Jan '11

If the event loop takes a very long time todo something, it is possible for the 'self pipe' buffer to become full at which point the entire event loop + remote driver deadlock. Use a boolean flag to ensure we have strict one-in, one-out behaviour on writes/reads of the 'self pipe' --- examples/domain-events/events-python/event-test.py | 10 +++++++++- 1 files changed, 9 insertions(+), 1 deletions(-) diff --git a/examples/domain-events/events-python/event-test.py b/examples/domain-events/events-python/event-test.py index 903f934..c149ed9 100644 --- a/examples/domain-events/events-python/event-test.py +++ b/examples/domain-events/events-python/event-test.py @@ -92,6 +92,8 @@ class virEventLoopPure: def __init__(self): self.poll = select.poll() self.pipetrick = os.pipe() + self.pendingWakeup = False + self.runningPoll = False self.nextHandleID = 1 self.nextTimerID = 1 self.handles = [] @@ -166,6 +168,7 @@ class virEventLoopPure: # these pointless repeated tiny sleeps. def run_once(self): sleep = -1 + self.runningPoll = True next = self.next_timeout() debug("Next timeout due at %d" % next) if next > 0: @@ -184,6 +187,7 @@ class virEventLoopPure: # telling us to wakup. if so, then discard # the data just continue if fd == self.pipetrick[0]: + self.pendingWakeup = False data = os.read(fd, 1) continue @@ -206,6 +210,8 @@ class virEventLoopPure: t.set_last_fired(now) t.dispatch() + self.runningPoll = False + # Actually the event loop forever def run_loop(self): @@ -214,7 +220,9 @@ class virEventLoopPure: self.run_once() def interrupt(self): - os.write(self.pipetrick[1], 'c') + if self.runningPoll and not self.pendingWakeup: + self.pendingWakeup = True + os.write(self.pipetrick[1], 'c') # Registers a new file handle 'fd', monitoring for 'events' (libvirt -- 1.7.3.4

2 1

[libvirt] [PATCH] docs: fix incorrect XML element mentioned by setmem text
by Justin Clift 27 Jan '11

27 Jan '11

--- Pushed this under the trivial rule, as it's definitely wrong. The setmaxmem text can follow when ready. tools/virsh.pod | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/tools/virsh.pod b/tools/virsh.pod index 1f15fef..4e8b295 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -580,7 +580,7 @@ paravirtualized or running the PV balloon driver. Note, this command only works on active guest domains. To change the memory allocation for an inactive guest domain, use the virsh B<edit> command to -update the XML <memory> element. +update the XML <currentMemory> element. =item B<setmaxmem> I<domain-id> B<kilobytes> -- 1.7.3.5

1 0

[libvirt] Planning for next release and discussions on project process
by Daniel Veillard 27 Jan '11

27 Jan '11

So we are around the end of the month, i.e. the time where I usually come out and suggest to enter a freeze within a few days for the next release. However, looking at the git history I note that we have only around 125 patches commited since 0.8.7 was released and while I'm fairly convinced that time based releases are a really good thing (allowing people to see their patches out within a predictable delay, and allowing the whole ecosystem of distros/users to plan accordingly), it's also time I think to open up the process to the growing (or at least changing) set of people involved in building libvirt. I usually ask if everybody agrees on the next freeze but I have to make the initial selection based on the calendar and my own agenda (or Red Hat one). Other projects I know use things like weekly teleconferences to discuss ther development plans, or IRC meetings. My opinion is that phone is not alway convenient/eficient, IRC has the advantage of being more open (no spoken english barrier), less intrusive, and it's nearly trivial to extract and publish some kind of minutes. But both suffer from the timezone problem, basically if there is meeting time it's gonna be during insane hours for someone somewhere (we have a large set of developpers in the US, Matthias and Dan are in Europe, Justin is in Australia, I'm in China now and I think the Fujitsu developpers too) One possibility could be to use a mix of IRC channel gathering and an Etherpad file (very convenient tool try it), I just created one at http://ietherpad.com/oETnXKdcFf to try this out and put my own sugegstions, then others can come and put their own feedback even if they happen to not be online at the time of the meeting. To come back to the release, maybe we can do 2 release for Jan-Mar since we didn't released early in Jan and Feb is short, so I would suggest to plan a release before Feb 15, entering a freeze around the 7th Feb. This should allow to put in some of the code already reviewed but not yet commited that I spotted: + migration enhancements from Hu Tao and Wen Congyang + integration of locking in the QEmu driver from Dan + cgroup blkio support (lxc and qemu) from Gui Jianfeng + smartcard support from Eric but there is certainly things I forgot, and well people may have specific things they want in ASAP, so let's use IRC (irc.oftc.net #virt) and http://ietherpad.com/oETnXKdcFf and share opinions about next release and features to come, well unless people have a better idea on how to organize ourselves, Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

2 2

[libvirt] [v2] remote: Add extra parameter pkipath for URI
by Osier Yang 27 Jan '11

27 Jan '11

This new parameter allows user specifies where the client cerficate, client key, CA certificate of x509 is, instead of hardcoding it. If 'pkipath' is not specified, and the user is not root, try to find files in $HOME/.pki, as long as one of client cerficate, client key, CA certificate can not be found, use default global location (LIBVIRT_CACERT, LIBVIRT_CLIENTCERT, LIBVIRT_CLIENTKEY, see src/remote/remote_driver.h) e.g. [root@Osier client]# virsh -c qemu+tls://10.66.93.111/system?pkipath=/tmp/pki/client error: Cannot access CA certificate '/tmp/pki/client/cacert.pem': No such file or directory error: failed to connect to the hypervisor [root@Osier client]# ls -l total 24 -rwxrwxr-x. 1 root root 6424 Jan 24 21:35 a.out -rw-r--r--. 1 root root 1245 Jan 23 19:04 clientcert.pem -rw-r--r--. 1 root root 132 Jan 23 19:04 client.info -rw-r--r--. 1 root root 1679 Jan 23 19:04 clientkey.pem [root@Osier client]# cp /tmp/cacert.pem . [root@Osier client]# virsh -c qemu+tls://10.66.93.111/system?pkipath=/tmp/pki/client Welcome to virsh, the virtualization interactive terminal. Type: 'help' for help with commands 'quit' to quit virsh # * src/remote/remote_driver.c --- src/remote/remote_driver.c | 130 ++++++++++++++++++++++++++++++++++++++------ 1 files changed, 112 insertions(+), 18 deletions(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index ea119c6..01f1995 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -268,7 +268,7 @@ void remoteDomainEventQueueFlush(int timer, void *opaque); static char *get_transport_from_scheme (char *scheme); /* GnuTLS functions used by remoteOpen. */ -static int initialize_gnutls(void); +static int initialize_gnutls(char *pkipath, int flags); static gnutls_session_t negotiate_gnutls_on_connection (virConnectPtr conn, struct private_data *priv, int no_verify); #ifdef WITH_LIBVIRTD @@ -430,6 +430,7 @@ doRemoteOpen (virConnectPtr conn, char *port = NULL, *authtype = NULL, *username = NULL; int no_verify = 0, no_tty = 0; char **cmd_argv = NULL; + char *pkipath = NULL; /* Return code from this function, and the private data. */ int retcode = VIR_DRV_OPEN_ERROR; @@ -509,9 +510,14 @@ doRemoteOpen (virConnectPtr conn, priv->debugLog = stdout; else priv->debugLog = stderr; - } else + } else if (STRCASEEQ(var->name, "pkipath")) { + pkipath = strdup(var->value); + if (!pkipath) goto out_of_memory; + var->ignore = 1; + } else { DEBUG("passing through variable '%s' ('%s') to remote end", var->name, var->value); + } } /* Construct the original name. */ @@ -577,7 +583,7 @@ doRemoteOpen (virConnectPtr conn, /* Connect to the remote service. */ switch (transport) { case trans_tls: - if (initialize_gnutls() == -1) goto failed; + if (initialize_gnutls(pkipath, flags) == -1) goto failed; priv->uses_tls = 1; priv->is_secure = 1; @@ -947,6 +953,7 @@ doRemoteOpen (virConnectPtr conn, } VIR_FREE(cmd_argv); } + VIR_FREE(pkipath); return retcode; @@ -1139,11 +1146,17 @@ static void remote_debug_gnutls_log(int level, const char* str) { } static int -initialize_gnutls(void) +initialize_gnutls(char *pkipath, int flags) { static int initialized = 0; int err; char *gnutlsdebug; + char *libvirt_cacert = NULL; + char *libvirt_clientkey = NULL; + char *libvirt_clientcert = NULL; + int ret = -1; + char *userdir = NULL; + char *user_pki_path = NULL; if (initialized) return 0; @@ -1166,43 +1179,124 @@ initialize_gnutls(void) return -1; } + if (pkipath) { + if ((virAsprintf(&libvirt_cacert, "%s/%s", pkipath, + "cacert.pem")) < 0) + goto out_of_memory; - if (check_cert_file("CA certificate", LIBVIRT_CACERT) < 0) - return -1; - if (check_cert_file("client key", LIBVIRT_CLIENTKEY) < 0) - return -1; - if (check_cert_file("client certificate", LIBVIRT_CLIENTCERT) < 0) - return -1; + if ((virAsprintf(&libvirt_clientkey, "%s/%s", pkipath, + "clientkey.pem")) < 0) + goto out_of_memory; + + if ((virAsprintf(&libvirt_clientcert, "%s/%s", pkipath, + "clientcert.pem")) < 0) + goto out_of_memory; + } else if (flags & VIR_DRV_OPEN_REMOTE_USER) { + userdir = virGetUserDirectory(getuid()); + + if (!userdir) + goto out_of_memory; + + if (virAsprintf(&user_pki_path, "%s/.pki", userdir) < 0) + goto out_of_memory; + + if ((virAsprintf(&libvirt_cacert, "%s/%s", user_pki_path, + "cacert.pem")) < 0) + goto out_of_memory; + + if ((virAsprintf(&libvirt_clientkey, "%s/%s", user_pki_path, + "clientkey.pem")) < 0) + goto out_of_memory; + + if ((virAsprintf(&libvirt_clientcert, "%s/%s", user_pki_path, + "clientcert.pem")) < 0) + goto out_of_memory; + + /* Use default locations as long as one of CA certificate, + * client key, and client certificate can not be found in + * $HOME/.pki, we don't want to make user confused with one + * file is here, and the other is there. + */ + if (!virFileExists(libvirt_cacert) || + !virFileExists(libvirt_clientkey) || + !virFileExists(libvirt_clientcert)) { + VIR_FREE(libvirt_cacert); + VIR_FREE(libvirt_clientkey); + VIR_FREE(libvirt_clientcert); + + libvirt_cacert = strdup(LIBVIRT_CACERT); + if (!libvirt_cacert) goto out_of_memory; + + libvirt_clientkey = strdup(LIBVIRT_CLIENTKEY); + if (!libvirt_clientkey) goto out_of_memory; + + libvirt_clientcert = strdup(LIBVIRT_CLIENTCERT); + if (!libvirt_clientcert) goto out_of_memory; + } + } else { + libvirt_cacert = strdup(LIBVIRT_CACERT); + if (!libvirt_cacert) goto out_of_memory; + + libvirt_clientkey = strdup(LIBVIRT_CLIENTKEY); + if (!libvirt_clientkey) goto out_of_memory; + + libvirt_clientcert = strdup(LIBVIRT_CLIENTCERT); + if (!libvirt_clientcert) goto out_of_memory; + } + + if (check_cert_file("CA certificate", libvirt_cacert) < 0) + goto error; + if (check_cert_file("client key", libvirt_clientkey) < 0) + goto error; + if (check_cert_file("client certificate", libvirt_clientcert) < 0) + goto error; /* Set the trusted CA cert. */ - DEBUG("loading CA file %s", LIBVIRT_CACERT); + DEBUG("loading CA file %s", libvirt_cacert); err = - gnutls_certificate_set_x509_trust_file (x509_cred, LIBVIRT_CACERT, + gnutls_certificate_set_x509_trust_file (x509_cred, libvirt_cacert, GNUTLS_X509_FMT_PEM); if (err < 0) { remoteError(VIR_ERR_GNUTLS_ERROR, _("unable to load CA certificate: %s"), gnutls_strerror (err)); - return -1; + goto error; } /* Set the client certificate and private key. */ DEBUG("loading client cert and key from files %s and %s", - LIBVIRT_CLIENTCERT, LIBVIRT_CLIENTKEY); + libvirt_clientcert, libvirt_clientkey); err = gnutls_certificate_set_x509_key_file (x509_cred, - LIBVIRT_CLIENTCERT, - LIBVIRT_CLIENTKEY, + libvirt_clientcert, + libvirt_clientkey, GNUTLS_X509_FMT_PEM); if (err < 0) { remoteError(VIR_ERR_GNUTLS_ERROR, _("unable to load private key/certificate: %s"), gnutls_strerror (err)); - return -1; + goto error; } initialized = 1; - return 0; + ret = 0; + +cleanup: + VIR_FREE(libvirt_cacert); + VIR_FREE(libvirt_clientkey); + VIR_FREE(libvirt_clientcert); + VIR_FREE(userdir); + VIR_FREE(user_pki_path); + return ret; + +error: + ret = -1; + goto cleanup; + +out_of_memory: + ret = -1; + virReportOOMError(); + goto cleanup; } static int verify_certificate (virConnectPtr conn, struct private_data *priv, gnutls_session_t session); -- 1.7.3.2

4 6

[libvirt] [PATCH] [v2] storage: Round up capacity for LVM volume creation
by Osier Yang 27 Jan '11

27 Jan '11

If vol->capacity is odd, the capacity will be rounded down by devision, this patch is to round it up instead of rounding down, to be safer in case of one writes to the volume with the size he used to create. * src/storage/storage_backend_logical.c --- src/storage/storage_backend_logical.c | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/src/storage/storage_backend_logical.c b/src/storage/storage_backend_logical.c index 203fe5d..2057692 100644 --- a/src/storage/storage_backend_logical.c +++ b/src/storage/storage_backend_logical.c @@ -604,7 +604,10 @@ virStorageBackendLogicalCreateVol(virConnectPtr conn, cmdargv = cmdargvsnap; } - snprintf(size, sizeof(size)-1, "%lluK", vol->capacity/1024); + unsigned long long int capacity; + capacity = (vol->capacity + 1023) /1024; + + snprintf(size, sizeof(size)-1, "%lluK", capacity); size[sizeof(size)-1] = '\0'; vol->type = VIR_STORAGE_VOL_BLOCK; -- 1.7.3.2

2 1

[libvirt] [PATCH 1/4] cgroup: Enable cgroup hierarchy for blkio cgroup
by Gui Jianfeng 27 Jan '11

27 Jan '11

Enable cgroup hierarchy for blkio cgroup. Signed-off-by: Gui Jianfeng <guijianfeng(a)cn.fujitsu.com> --- src/util/cgroup.c | 2 +- src/util/cgroup.h | 1 + 2 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/util/cgroup.c b/src/util/cgroup.c index cd9caba..309f4e9 100644 --- a/src/util/cgroup.c +++ b/src/util/cgroup.c @@ -37,7 +37,7 @@ VIR_ENUM_IMPL(virCgroupController, VIR_CGROUP_CONTROLLER_LAST, "cpu", "cpuacct", "cpuset", "memory", "devices", - "freezer"); + "freezer", "blkio"); struct virCgroupController { int type; diff --git a/src/util/cgroup.h b/src/util/cgroup.h index 964da7a..67b1299 100644 --- a/src/util/cgroup.h +++ b/src/util/cgroup.h @@ -22,6 +22,7 @@ enum { VIR_CGROUP_CONTROLLER_MEMORY, VIR_CGROUP_CONTROLLER_DEVICES, VIR_CGROUP_CONTROLLER_FREEZER, + VIR_CGROUP_CONTROLLER_BLKIO, VIR_CGROUP_CONTROLLER_LAST }; -- 1.7.1

1 0