- Devel - libvirt List Archives

[libvirt] [PATCH] qemu: Substitute VIR_ERR_NO_SUPPORT with VIR_ERR_OPERATION_INVALID
by Osier Yang 23 Aug '11

23 Aug '11

* src/qemu/qemu_monitor_text.c: Error like "this function is not supported by the connection driver" is confused obviously. --- src/qemu/qemu_monitor_text.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index d296675..4a7c242 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -713,7 +713,7 @@ int qemuMonitorTextGetBlockStatsInfo(qemuMonitorPtr mon, * to detect if qemu supports the command. */ if (strstr(info, "\ninfo ")) { - qemuReportError(VIR_ERR_NO_SUPPORT, + qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", _("'info blockstats' not supported by this qemu")); goto cleanup; @@ -1388,7 +1388,7 @@ int qemuMonitorTextMigrate(qemuMonitorPtr mon, /* If the command isn't supported then qemu prints: * unknown command: migrate" */ if (strstr(info, "unknown command:")) { - qemuReportError(VIR_ERR_NO_SUPPORT, + qemuReportError(VIR_ERR_OPERATION_INVALID, _("migration to '%s' not supported by this qemu: %s"), dest, info); goto cleanup; } @@ -1652,7 +1652,7 @@ int qemuMonitorTextAddPCIHostDevice(qemuMonitorPtr mon, } if (strstr(reply, "invalid type: host")) { - qemuReportError(VIR_ERR_NO_SUPPORT, "%s", + qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", _("PCI device assignment is not supported by this version of qemu")); goto cleanup; } -- 1.7.6

2 4

[libvirt] [PATCH] virsh: Print error if specified bandwidth is invalid for blockjob
by Osier Yang 23 Aug '11

23 Aug '11

It's strange that the command fails but without any error if one specifies as not a number. --- tools/virsh.c | 21 ++++++++++++++++----- 1 files changed, 16 insertions(+), 5 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index ffb4ced..97512b5 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -5209,17 +5209,28 @@ blockJobImpl(vshControl *ctl, const vshCmd *cmd, if (vshCommandOptString(cmd, "path", &path) < 0) goto out; - if (vshCommandOptUL(cmd, "bandwidth", &bandwidth) < 0) + if (vshCommandOptUL(cmd, "bandwidth", &bandwidth) < 0) { + vshError(ctl, "%s", _("bandwidth must be a number")); goto out; + } - if (mode == VSH_CMD_BLOCK_JOB_ABORT) + if (mode == VSH_CMD_BLOCK_JOB_ABORT) { + vshPrint(ctl, "abort"); ret = virDomainBlockJobAbort(dom, path, 0); - else if (mode == VSH_CMD_BLOCK_JOB_INFO) + } + else if (mode == VSH_CMD_BLOCK_JOB_INFO) { + vshPrint(ctl, "info\n"); ret = virDomainGetBlockJobInfo(dom, path, info, 0); - else if (mode == VSH_CMD_BLOCK_JOB_SPEED) + vshPrint(ctl, "ret = %d\n", ret); + } + else if (mode == VSH_CMD_BLOCK_JOB_SPEED) { + vshPrint(ctl, "speed"); ret = virDomainBlockJobSetSpeed(dom, path, bandwidth, 0); - else if (mode == VSH_CMD_BLOCK_JOB_PULL) + } + else if (mode == VSH_CMD_BLOCK_JOB_PULL) { + vshPrint(ctl, "pull"); ret = virDomainBlockPull(dom, path, bandwidth, 0); + } out: virDomainFree(dom); -- 1.7.6

2 3

[libvirt] Adding migration speed to domainXML
by Jim Fehlig 23 Aug '11

23 Aug '11

I finally have some time to investigate Eric's suggestion [1] to add virDomainMigrateGetMaxSpeed API and track current value in domainXML. WRT the XML representation, it seems this best fits under the existing lifecycle controls, e.g. a new "on_migrate" element. I'm not sure if this should be kept simple and rather flat such as <on_migrate max-speed='1000000000' /> or a little more expressive and extensible with something like <on_migrate> <bandwidth peak='1000000000' /> </on_migrate> Any comments on these representations or perhaps alternative suggestions? Regards, Jim [1] https://www.redhat.com/archives/libvir-list/2011-August/msg00224.html

1 0

[libvirt] Fwd: [virt-tools-list] virt-manager, dependency question
by Jason Helfman 23 Aug '11

23 Aug '11

Can anyone please confirm this for me? Thanks! Jason -- Jason Helfman System Administrator experts-exchange.com http://www.experts-exchange.com/M_4830110.html E4AD 7CF1 1396 27F6 79DD 4342 5E92 AD66 8C8C FBA5

3 2

[libvirt] [PATCH] qemu: Report error if qemu monitor command not found for BlockJob
by Osier Yang 23 Aug '11

23 Aug '11

* src/qemu/qemu_monitor_json.c: Handle error "CommandNotFound" and report the error. * src/qemu/qemu_monitor_text.c: If a sub info command is not found, it prints the output of "help info", for other commands, "unknown command" is printed. Without this patch, libvirt always report: An error occurred, but the cause is unknown --- src/qemu/qemu_monitor_json.c | 34 ++++++++++++++++++----------- src/qemu/qemu_monitor_text.c | 47 +++++++++++++++++++++++++++++++++--------- 2 files changed, 58 insertions(+), 23 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 7adfb26..715b26e 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -2909,20 +2909,25 @@ int qemuMonitorJSONBlockJob(qemuMonitorPtr mon, int ret = -1; virJSONValuePtr cmd = NULL; virJSONValuePtr reply = NULL; - - if (mode == BLOCK_JOB_ABORT) - cmd = qemuMonitorJSONMakeCommand("block_job_cancel", - "s:device", device, NULL); - else if (mode == BLOCK_JOB_INFO) - cmd = qemuMonitorJSONMakeCommand("query-block-jobs", NULL); - else if (mode == BLOCK_JOB_SPEED) - cmd = qemuMonitorJSONMakeCommand("block_job_set_speed", - "s:device", device, - "U:value", bandwidth * 1024ULL * 1024ULL, + const char *cmd_name = NULL; + + if (mode == BLOCK_JOB_ABORT) { + cmd_name = "block_job_cancel"; + cmd = qemuMonitorJSONMakeCommand(cmd_name, "s:device", device, NULL); + } else if (mode == BLOCK_JOB_INFO) { + cmd_name = "query-block-jobs"; + cmd = qemuMonitorJSONMakeCommand(cmd_name, NULL); + } else if (mode == BLOCK_JOB_SPEED) { + cmd_name = "block_job_set_speed"; + cmd = qemuMonitorJSONMakeCommand(cmd_name, "s:device", + device, "U:value", + bandwidth * 1024ULL * 1024ULL, NULL); - else if (mode == BLOCK_JOB_PULL) - cmd = qemuMonitorJSONMakeCommand("block_stream", - "s:device", device, NULL); + } else if (mode == BLOCK_JOB_PULL) { + cmd_name = "block_stream"; + cmd = qemuMonitorJSONMakeCommand(cmd_name, "s:device", + device, NULL); + } if (!cmd) return -1; @@ -2939,6 +2944,9 @@ int qemuMonitorJSONBlockJob(qemuMonitorPtr mon, else if (qemuMonitorJSONHasError(reply, "NotSupported")) qemuReportError(VIR_ERR_OPERATION_INVALID, _("Operation is not supported for device: %s"), device); + else if (qemuMonitorJSONHasError(reply, "CommandNotFound")) + qemuReportError(VIR_ERR_OPERATION_INVALID, + _("Command '%s' is not found"), cmd_name); else qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Unexpected error")); diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 52d924a..d296675 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -3031,18 +3031,24 @@ int qemuMonitorTextBlockJob(qemuMonitorPtr mon, char *cmd = NULL; char *reply = NULL; int ret; - - if (mode == BLOCK_JOB_ABORT) - ret = virAsprintf(&cmd, "block_job_cancel %s", device); - else if (mode == BLOCK_JOB_INFO) - ret = virAsprintf(&cmd, "info block-jobs"); - else if (mode == BLOCK_JOB_SPEED) - ret = virAsprintf(&cmd, "block_job_set_speed %s %llu", device, + const char *cmd_name = NULL; + + if (mode == BLOCK_JOB_ABORT) { + cmd_name = "block_job_cancel"; + ret = virAsprintf(&cmd, "%s %s", cmd_name, device); + } else if (mode == BLOCK_JOB_INFO) { + cmd_name = "info block-jobs"; + ret = virAsprintf(&cmd, "%s", cmd_name); + } else if (mode == BLOCK_JOB_SPEED) { + cmd_name = "block_job_set_speed"; + ret = virAsprintf(&cmd, "%s %s %llu", cmd_name, device, bandwidth * 1024ULL * 1024ULL); - else if (mode == BLOCK_JOB_PULL) - ret = virAsprintf(&cmd, "block_stream %s", device); - else + } else if (mode == BLOCK_JOB_PULL) { + cmd_name = "block_stream"; + ret = virAsprintf(&cmd, "%s %s", cmd_name, device); + } else { return -1; + } if (ret < 0) { virReportOOMError(); @@ -3056,6 +3062,27 @@ int qemuMonitorTextBlockJob(qemuMonitorPtr mon, goto cleanup; } + /* Check error: Command not found, for "info block-jobs" command. + * If qemu monitor command "info" doesn't support one sub-command, + * it will print the output of "help info" instead, so simply check + * if "info version" is in the output. + */ + if (mode == BLOCK_JOB_INFO) { + if (strstr(reply, "info version")) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + _("Command '%s' is not found"), cmd_name); + ret = -1; + goto cleanup; + } + } else { + if (strstr(reply, "unknown command")) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + _("Command '%s' is not found"), cmd_name); + ret = -1; + goto cleanup; + } + } + ret = qemuMonitorTextParseBlockJob(reply, device, info); cleanup: -- 1.7.6

3 3

[libvirt] [PATCH] blockjob: Report correct error when missing qemu support
by Adam Litke 22 Aug '11

22 Aug '11

When executing blockjob monitor commands, we are not checking for CommandNotFound errors. This results in 'unexpected error' messages when using a qemu without support for the required commands. Fix this up by checking for CommandNotFound errors for QMP and detecting this condition for the text monitor. This fixes the problem reported in Bug 727502. Signed-off-by: Adam Litke <agl(a)us.ibm.com> --- src/qemu/qemu_monitor_json.c | 3 +++ src/qemu/qemu_monitor_text.c | 20 ++++++++++++++++++++ 2 files changed, 23 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 7adfb26..4c8a08b 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -2939,6 +2939,9 @@ int qemuMonitorJSONBlockJob(qemuMonitorPtr mon, else if (qemuMonitorJSONHasError(reply, "NotSupported")) qemuReportError(VIR_ERR_OPERATION_INVALID, _("Operation is not supported for device: %s"), device); + else if (qemuMonitorJSONHasError(reply, "CommandNotFound")) + qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", + _("Operation is not supported")); else qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Unexpected error")); diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 52d924a..d9cad2d 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -271,6 +271,19 @@ qemuMonitorTextCommandWithFd(qemuMonitorPtr mon, scm_fd, reply); } +/* Check monitor output for evidence that the command was not recognized. + * For 'info' commands, qemu returns help text. For other commands, qemu + * returns 'unknown command:'. + */ +static int +qemuMonitorTextCommandNotFound(const char *reply) +{ + if (strstr(reply, "unknown command:")) + return 1; + if (strstr(reply, "info version -- show the version of QEMU")) + return 1; + return 0; +} static int qemuMonitorSendDiskPassphrase(qemuMonitorPtr mon, @@ -3056,6 +3069,13 @@ int qemuMonitorTextBlockJob(qemuMonitorPtr mon, goto cleanup; } + if (qemuMonitorTextCommandNotFound(reply)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", + _("Operation is not supported")); + ret = -1; + goto cleanup; + } + ret = qemuMonitorTextParseBlockJob(reply, device, info); cleanup: -- 1.7.3

2 1

[libvirt] [RFC 0/4] Report disk latency information
by Osier Yang 22 Aug '11

22 Aug '11

The QEMU side patches are not posted yet, but considering what these patche series need is only a fixed interface provided by QEMU, posting it for early reviewing, won't push it before QEMU patches are got pushed even it got ACK. Osier

6 13

[libvirt] [PATCH] remote: Switch to TLS when connecting to remote socket
by Michal Privoznik 22 Aug '11

22 Aug '11

If users wants to connect to remote unix socket, e.g. 'qemu+unix://<remote>/system' currently the <remote> part is ignored, ending up connecting to localhost. This patch solves this issue by creating TLS connection. --- src/remote/remote_driver.c | 11 ++++++++--- 1 files changed, 8 insertions(+), 3 deletions(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index e5bfa4b..630c8ea 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -325,9 +325,14 @@ doRemoteOpen (virConnectPtr conn, } else { if (STRCASEEQ (transport_str, "tls")) transport = trans_tls; - else if (STRCASEEQ (transport_str, "unix")) - transport = trans_unix; - else if (STRCASEEQ (transport_str, "ssh")) + else if (STRCASEEQ (transport_str, "unix")) { + /* When connecting to unix socket on remote + * machine use TLS to connect to it */ + if (conn->uri->server) + transport = trans_tls; + else + transport = trans_unix; + } else if (STRCASEEQ (transport_str, "ssh")) transport = trans_ssh; else if (STRCASEEQ (transport_str, "ext")) transport = trans_ext; -- 1.7.3.4

2 1

[libvirt] [PATCH] [BUGFIX][PATCH] send-key: fix scan keycode map
by KAMEZAWA Hiroyuki 22 Aug '11

22 Aug '11

Now, bad key-code in send-key can cause segmentation fault in libvirt. (example) % virsh send-key --codeset win32 12 error: End of file while reading data: Input/output error This is caused by overrun at scanning keycode array. Fix it. Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> --- src/util/virkeycode.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/util/virkeycode.c b/src/util/virkeycode.c index 4d54060..0d42767 100644 --- a/src/util/virkeycode.c +++ b/src/util/virkeycode.c @@ -118,7 +118,7 @@ static int __virKeycodeValueTranslate(unsigned int from_offset, { int i; - for (i = 0; ARRAY_CARDINALITY(virKeycodes); i++) { + for (i = 0; i < ARRAY_CARDINALITY(virKeycodes); i++) { if (getfield(virKeycodes + i, unsigned short, from_offset) == key_value) return getfield(virKeycodes + i, unsigned short, to_offset); } -- 1.7.4.1

2 1

[libvirt] portability issue on Centos 6.0
by Daniel Veillard 22 Aug '11

22 Aug '11

I just tred compiling 0.9.4 and git head on Centos 6.0, for the possible libvirt.org server replacement and it fails in both cases with: CC libvirtd-remote.o remote.c: In function 'remoteDispatchAuthPolkit': remote.c:2161: error: invalid initializer This correspond to the following macro PROBE(CLIENT_AUTH_ALLOW, "fd=%d, auth=%d, username=%s", virNetServerClientGetFD(client), REMOTE_AUTH_POLKIT, ident); I'm a bit lost how the expansion could lead to this and why, unless the DTrace support in Centos 6 is slightly different than on RHEL, I will look but if someone has an idea :-) 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 3

[libvirt] How to avoid the disk still persistent after no persistent hotplug and with restarting libvirtd then.
by Osier Yang 22 Aug '11

22 Aug '11

Libvirt loads the domain conf from the status file (/var/run/libvirt/qemu/$dom.xml) if the domain is running, the problem if one restart libvirtd just after some changing on the domain conf, and then destroy/start the guest. libvirt will never known what the original domain conf is then. (vm->newdef is NULL at this moment, generally libvirt restore the domain conf to the original conf (vm->newDef) when destroying and shutdown). Following is a demo of the problem: 1) hotplug a disk to a running domain without "--persistent" specified. 2) restart libvirtd 3) Destroy/start the guest 4) you will see the hotpluged disk is still there. The solution in my mind is to add some attribute to the status XML, such as "live_updated=1", and do the checking when loading the domain conf from status XML: * if "live_updated=1", then loads the domain conf from the persistent domain XML. But the problem is: For a persistent domain, if one made changes on the domain conf both persistently and transiently, we have no way to known which part are changed. Loading from the persistent domain conf is not correct either. For a transient domain, we still have no way to get the original domain conf, that means one will still be surprised when seeing the the disk is still there. Any idea on this? Thanks Osier

2 3

[libvirt] Proposal: A place to store metadata about guests
by Richard W.M. Jones 22 Aug '11

22 Aug '11

I think we need a place to store general metadata about guests. Here are some uses: - Store data about available installer ISOs, guests which are archived and not known to libvirt. - Replace the guest inspection thread in virt-manager. virt-manager would just collect the data from the store. The inspection data would be generated by another process. - Expand on <description/> field in the libvirt XML, and satisfy various requests we've had for a place to store metadata and extend libvirt XML. - Store data on behalf of applications, esp. virt-manager. I'm not envisaging a database or daemon for this. Just a simple directory per host containing flat files. Perhaps we can add some tools or a library to make creation/querying of these files easy. It would be nice if the database was made available over the network (Avahi?). You could also use this database to answer questions like: - "Do I have a Debian wheezy 32 bit guest anywhere on my network?" - "Does any guest have libpng <= <vulnerable-version> installed?" - "Bring up a Fedora 14 instance" (the tools would have to locate a suitable install ISO and put together the correct virt-install command line) There is some crossover with VDSM and even RHEV-M here. Thoughts? Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones libguestfs lets you edit virtual machines. Supports shell scripting, bindings from many languages. http://libguestfs.org

3 8

[libvirt] [libvirt-snmp][PATCH 1/3] subagent: Avoid dead code
by Michal Privoznik 22 Aug '11

22 Aug '11

Detected by Coverity --- src/libvirtMib_subagent.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/libvirtMib_subagent.c b/src/libvirtMib_subagent.c index f5749be..ee99fb9 100644 --- a/src/libvirtMib_subagent.c +++ b/src/libvirtMib_subagent.c @@ -60,7 +60,7 @@ main (int argc, char **argv) { /* Defs for arg-handling code: handles setting of policy-related variables */ int ch; extern char *optarg; - int dont_fork = 0, use_syslog = 0; + int dont_fork = 0, use_syslog = 1; char *agentx_socket = NULL; while ((ch = getopt(argc, argv, "D:fHLMx:")) != EOF) -- 1.7.3.4

1 2

[libvirt] [PATCH v2] qemu: Allow graceful domain destroy
by Michal Privoznik 22 Aug '11

22 Aug '11

This patch introduces support for domain destroying via 'quit' monitor command which gives qemu time to flush caches and therefore prevent disks corruption. However, qemu can be unresponsive and to prevent waiting indefinitely, execute command in a separate thread and wait reasonable time (QEMU_QUIT_WAIT_SECONDS) on a condition. If we hit timeout, qemu is qemuProcessKill'ed which causes monitor close and therefore also thread being terminable. The synchronization between qemu driver and monitor-quit thread is done through mutex and condition. However, this alone is not enough. If a condition is signalized but without anybody listening signal is lost. To prevent this a boolean variable is used that is set iff nobody is listening but condition would be signalized, or iff driver is waiting on given condition. --- diff to v1: - more debug printings - don't test (!flags) but (!(flags & VIR_DOMAIN_DESTROY_FLUSH_CACHE)) include/libvirt/libvirt.h.in | 11 ++- src/qemu/qemu_driver.c | 146 ++++++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_monitor.c | 13 ++++ src/qemu/qemu_monitor.h | 2 + src/qemu/qemu_monitor_json.c | 22 ++++++ src/qemu/qemu_monitor_json.h | 1 + src/qemu/qemu_monitor_text.c | 20 ++++++ src/qemu/qemu_monitor_text.h | 1 + tools/virsh.c | 9 ++- 9 files changed, 213 insertions(+), 12 deletions(-) diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in index aa29fb6..0610b53 100644 --- a/include/libvirt/libvirt.h.in +++ b/include/libvirt/libvirt.h.in @@ -919,10 +919,13 @@ virConnectPtr virDomainGetConnect (virDomainPtr domain); * Domain creation and destruction */ -/* - * typedef enum { - * } virDomainDestroyFlagsValues; - */ + +typedef enum { + VIR_DOMAIN_DESTROY_FLUSH_CACHE = 1 << 0, /* In hypervisors supporting this, + try to get cache flushed prior + to destroying the domain */ +} virDomainDestroyFlagsValues; + virDomainPtr virDomainCreateXML (virConnectPtr conn, const char *xmlDesc, unsigned int flags); diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index c8dda73..44f4bbd 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -1564,6 +1564,47 @@ cleanup: return ret; } +struct qemuDomainDestroyHelperData { + struct qemud_driver *driver; + virDomainObjPtr vm; + qemuDomainObjPrivatePtr priv; + virMutex lock; + virCond cond; + bool guard; +}; + +static void +qemuDomainDestroyHelper(void *opaque) +{ + struct qemuDomainDestroyHelperData *data = opaque; + struct qemud_driver *driver = data->driver; + virDomainObjPtr vm = data->vm; + qemuDomainObjPrivatePtr priv = data->priv; + + /* Job was already obtained by caller */ + VIR_DEBUG("Destroy thread started, sending 'quit' command."); + qemuDomainObjEnterMonitorWithDriver(driver, vm); + qemuMonitorQuit(priv->mon); + qemuDomainObjExitMonitorWithDriver(driver, vm); + VIR_DEBUG("Command finished."); + + /* To avoid loosing signal, we need to remember + * we tried to send one, but nobody was waiting + * for it. + */ + virMutexLock(&data->lock); + if (data->guard) { + VIR_DEBUG("Driver is listening, signalize on condition"); + virCondSignal(&data->cond); + } else { + VIR_DEBUG("Nobody's listening, setting guard variable"); + data->guard = true; + } + virMutexUnlock(&data->lock); + VIR_DEBUG("Destroy thread exiting"); +} + +#define QEMU_QUIT_WAIT_SECONDS 5 static int qemuDomainDestroyFlags(virDomainPtr dom, @@ -1574,8 +1615,13 @@ qemuDomainDestroyFlags(virDomainPtr dom, int ret = -1; virDomainEventPtr event = NULL; qemuDomainObjPrivatePtr priv; + bool use_thread = false; + bool use_kill = false; + virThread destroy_thread; + struct qemuDomainDestroyHelperData data; + unsigned long long then; - virCheckFlags(0, -1); + virCheckFlags(VIR_DOMAIN_DESTROY_FLUSH_CACHE, -1); qemuDriverLock(driver); vm = virDomainFindByUUID(&driver->domains, dom->uuid); @@ -1590,12 +1636,14 @@ qemuDomainDestroyFlags(virDomainPtr dom, priv = vm->privateData; priv->fakeReboot = false; - /* 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 - * it now means the job will be released - */ - qemuProcessKill(vm); + if (!(flags & VIR_DOMAIN_DESTROY_FLUSH_CACHE)) { + /* 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 + * it now means the job will be released + */ + qemuProcessKill(vm); + } if (qemuDomainObjBeginJobWithDriver(driver, vm, QEMU_JOB_DESTROY) < 0) goto cleanup; @@ -1606,6 +1654,78 @@ qemuDomainDestroyFlags(virDomainPtr dom, goto endjob; } + if (flags & VIR_DOMAIN_DESTROY_FLUSH_CACHE) { + /* Try to shutdown domain gracefully. + * Send 'quit' to qemu. However, qemu can be unresponsive. + * Therefore create a separate thread in which we execute + * that monitor comand. Wait max QEMU_QUIT_WAIT_SECONDS. + */ + VIR_DEBUG("Trying to gracefully shutdown %s", vm->def->name); + if (virMutexInit(&data.lock) < 0) { + virReportOOMError(); + goto endjob; + } + + if (virCondInit(&data.cond) < 0) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("cannot initialize thread condition")); + virMutexDestroy(&data.lock); + goto endjob; + } + + data.driver = driver; + data.vm = vm; + data.priv = priv; + data.guard = false; + + VIR_DEBUG("Spawning thread to run 'quit' command"); + if (virThreadCreate(&destroy_thread, true, + qemuDomainDestroyHelper, &data) < 0) { + virReportSystemError(errno, "%s", + _("unable to create destroy thread")); + ignore_value(virCondDestroy(&data.cond)); + virMutexDestroy(&data.lock); + goto endjob; + } + use_thread = true; + + if (virTimeMs(&then) < 0) + goto endjob; + + then += QEMU_QUIT_WAIT_SECONDS * 1000; + + /* How synchronization with destroy thread works: + * Basically wait on data.cond. But because conditions + * does not 'remember' that they have been signalized + * data.guard is used. Practically, data.guard says + * to destroy thread if we are waiting on condition + * and to us whether we should even try. + */ + virMutexLock(&data.lock); + if (!data.guard) { + data.guard = true; + VIR_DEBUG("Waiting for thread to exit"); + if (virCondWaitUntil(&data.cond, &data.lock, then) < 0) { + if (errno == ETIMEDOUT) { + VIR_DEBUG("Condition timeout"); + use_kill = true; + data.guard = false; + } else { + virReportSystemError(errno, "%s", _("unable to wait " + "on condition")); + virMutexUnlock(&data.lock); + goto endjob; + } + } + } else { + VIR_DEBUG("Guard variable set, thread exited"); + } + virMutexUnlock(&data.lock); + + if (use_kill) + qemuProcessKill(vm); + } + qemuProcessStop(driver, vm, 0, VIR_DOMAIN_SHUTOFF_DESTROYED); event = virDomainEventNewFromObj(vm, VIR_DOMAIN_EVENT_STOPPED, @@ -1626,6 +1746,18 @@ endjob: vm = NULL; cleanup: + if (use_thread) { + VIR_DEBUG("Thread used, clean it up"); + virMutexLock(&data.lock); + if (!data.guard) { + data.guard = true; + ignore_value(virCondWait(&data.cond, &data.lock)); + } + virMutexUnlock(&data.lock); + ignore_value(virCondDestroy(&data.cond)); + virMutexDestroy(&data.lock); + virThreadJoin(&destroy_thread); + } if (vm) virDomainObjUnlock(vm); if (event) diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index db6107c..41b9c5c 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -2475,3 +2475,16 @@ int qemuMonitorBlockJob(qemuMonitorPtr mon, ret = qemuMonitorTextBlockJob(mon, device, bandwidth, info, mode); return ret; } + +int qemuMonitorQuit(qemuMonitorPtr mon) +{ + int ret; + + VIR_DEBUG("mon=%p", mon); + + if (mon->json) + ret = qemuMonitorJSONQuit(mon); + else + ret = qemuMonitorTextQuit(mon); + return ret; +} diff --git a/src/qemu/qemu_monitor.h b/src/qemu/qemu_monitor.h index f241c9e..3fe6bb9 100644 --- a/src/qemu/qemu_monitor.h +++ b/src/qemu/qemu_monitor.h @@ -475,6 +475,8 @@ int qemuMonitorBlockJob(qemuMonitorPtr mon, virDomainBlockJobInfoPtr info, int mode); +int qemuMonitorQuit(qemuMonitorPtr mon); + /** * When running two dd process and using <> redirection, we need a * shell that will not truncate files. These two strings serve that diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 7adfb26..1f078cd 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -2956,3 +2956,25 @@ int qemuMonitorJSONBlockJob(qemuMonitorPtr mon, virJSONValueFree(reply); return ret; } + + +int qemuMonitorJSONQuit(qemuMonitorPtr mon) +{ + int ret = -1; + virJSONValuePtr cmd = NULL; + virJSONValuePtr reply = NULL; + + cmd = qemuMonitorJSONMakeCommand("quit", NULL); + + if (!cmd) + return -1; + + ret =qemuMonitorJSONCommand(mon, cmd, &reply); + + if (ret == 0) + ret = qemuMonitorJSONCheckError(cmd, reply); + + virJSONValueFree(cmd); + virJSONValueFree(reply); + return ret; +} diff --git a/src/qemu/qemu_monitor_json.h b/src/qemu/qemu_monitor_json.h index 9512793..2a7df76 100644 --- a/src/qemu/qemu_monitor_json.h +++ b/src/qemu/qemu_monitor_json.h @@ -231,4 +231,5 @@ int qemuMonitorJSONBlockJob(qemuMonitorPtr mon, virDomainBlockJobInfoPtr info, int mode); +int qemuMonitorJSONQuit(qemuMonitorPtr mon); #endif /* QEMU_MONITOR_JSON_H */ diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 52d924a..ca9c21b 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -3063,3 +3063,23 @@ cleanup: VIR_FREE(reply); return ret; } + + +int qemuMonitorTextQuit(qemuMonitorPtr mon) +{ + const char *cmd = "quit"; + char *reply = NULL; + int ret = -1; + + if (qemuMonitorHMPCommand(mon, cmd, &reply) < 0) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("cannot run monitor command")); + goto cleanup; + } + + ret = 0; + +cleanup: + VIR_FREE(reply); + return ret; +} diff --git a/src/qemu/qemu_monitor_text.h b/src/qemu/qemu_monitor_text.h index b250738..9ade938 100644 --- a/src/qemu/qemu_monitor_text.h +++ b/src/qemu/qemu_monitor_text.h @@ -224,4 +224,5 @@ int qemuMonitorTextBlockJob(qemuMonitorPtr mon, virDomainBlockJobInfoPtr info, int mode); +int qemuMonitorTextQuit(qemuMonitorPtr mon); #endif /* QEMU_MONITOR_TEXT_H */ diff --git a/tools/virsh.c b/tools/virsh.c index 7d849ec..4c7962b 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -2567,6 +2567,8 @@ static const vshCmdInfo info_destroy[] = { static const vshCmdOptDef opts_destroy[] = { {"domain", VSH_OT_DATA, VSH_OFLAG_REQ, N_("domain name, id or uuid")}, + {"flush", VSH_OT_BOOL, 0, N_("try to flush hypervisor's cache prior to " + "destroying the domain")}, {NULL, 0, 0, NULL} }; @@ -2576,6 +2578,7 @@ cmdDestroy(vshControl *ctl, const vshCmd *cmd) virDomainPtr dom; bool ret = true; const char *name; + int flags = 0; if (!vshConnectionUsability(ctl, ctl->conn)) return false; @@ -2583,7 +2586,11 @@ cmdDestroy(vshControl *ctl, const vshCmd *cmd) if (!(dom = vshCommandOptDomain(ctl, cmd, &name))) return false; - if (virDomainDestroy(dom) == 0) { + if (vshCommandOptBool(cmd, "flush")) + flags |= VIR_DOMAIN_DESTROY_FLUSH_CACHE; + + if ((!flags && virDomainDestroy(dom) == 0) || + virDomainDestroyFlags(dom, flags) == 0) { vshPrint(ctl, _("Domain %s destroyed\n"), name); } else { vshError(ctl, _("Failed to destroy domain %s"), name); -- 1.7.3.4

1 0

[libvirt] USB2 controllers & domain XML
by Marc-André Lureau 22 Aug '11

22 Aug '11

Hi I have started looking at: https://bugzilla.redhat.com/show_bug.cgi?id=725670 My initial thought was to just add a new type of controller: <controller type='usb2' index='0'/> But it also needs the 3 companion UHCI controllers (see bug). We can either include them implicitely (not really elegant in libvirt, but doable), or with more flexibility list all the controllers in the XML: <controller type='ich9-usb-ehci1' index='0'/> <controller type='ich9-usb-uhci1' index='0' masterbus='0' firstport='0'> <controller type='ich9-usb-uhci2' index='0' masterbus='0' firstport='2'> <controller type='ich9-usb-uhci3' index='0' masterbus='0' firstport='4'> Any other idea? Daniel, have you also started working on USB2 & usb-redir support? regards -- Marc-André Lureau

5 7

[libvirt] virt-manager, debug error, question
by Jason Helfman 22 Aug '11

22 Aug '11

Hi, I am getting this error when I run virt-manager in FreeBSD: [Fri, 19 Aug 2011 09:32:14 virt-manager 24762] DEBUG (engine:555) No inspection thread because libguestfs is too old, not available, or libvirt is not thread safe. libguestfs isn't ported to FreeBSD, however I was wondering about the thread safe for libvirt. Is there a way to "enable" this, or is the software telling me it really is not thread safe, and can't be. [jhelfman@eggman ~/ports/devel/libvirt/work/libvirt-0.9.4]$ ./configure --help |grep thread --enable-threads={posix|solaris|pth|win32} specify multithreading API --disable-threads build without multithread safety --enable-test-locking thread locking tests using CIL [default=no] Thanks, Jason -- Jason Helfman System Administrator experts-exchange.com http://www.experts-exchange.com/M_4830110.html E4AD 7CF1 1396 27F6 79DD 4342 5E92 AD66 8C8C FBA5

3 2

[libvirt] status of ocaml for libvirt
by Jason Helfman 20 Aug '11

20 Aug '11

Hello All, I was looking at porting virt-top to FreeBSD, and noticed that it needed ocaml-libvirt, and I know that isn't ported. What is the status of this library, and it is up-to-date? Thanks, Jason -- Jason Helfman System Administrator experts-exchange.com http://www.experts-exchange.com/M_4830110.html E4AD 7CF1 1396 27F6 79DD 4342 5E92 AD66 8C8C FBA5

1 0

[libvirt] [PATCH] virsh: properly interleave shared stdout and stderr
by Eric Blake 20 Aug '11

20 Aug '11

Without this patch, invoking 'virsh >file 2>&1' results in error messages appearing before normal output, even if they occurred later in time than the normal output (since stderr is unbuffered, but stdout waits until a full buffer). * tools/virsh.c (print_job_progress, vshError): Flush between stream transitions. * tests/undefine: Test it. --- This fixes the bug I mentioned here: https://www.redhat.com/archives/libvir-list/2011-August/msg00891.html tests/undefine | 5 +---- tools/virsh.c | 8 ++++++++ 2 files changed, 9 insertions(+), 4 deletions(-) diff --git a/tests/undefine b/tests/undefine index 6c821ad..22d6c14 100755 --- a/tests/undefine +++ b/tests/undefine @@ -58,17 +58,14 @@ compare exp out || fail=1 # Succeed, now: first shut down, then undefine, both via name. $abs_top_builddir/tools/virsh -q -c test:///default \ - 'shutdown test; undefine test; dominfo test' > out 2> err + 'shutdown test; undefine test; dominfo test' > out 2>&1 test $? = 1 || fail=1 cat <<\EOF > expout || fail=1 Domain test is being shutdown Domain test has been undefined -EOF -cat <<\EOF > experr || fail=1 error: failed to get domain 'test' error: Domain not found EOF compare expout out || fail=1 -compare experr err || fail=1 (exit $fail); exit $fail diff --git a/tools/virsh.c b/tools/virsh.c index 1c67150..7d849ec 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -4956,7 +4956,10 @@ print_job_progress(const char *label, unsigned long long remaining, } } + /* see comments in vshError about why we must flush */ + fflush(stdout); fprintf(stderr, "\r%s: [%3d %%]", label, progress); + fflush(stderr); } static bool @@ -14336,6 +14339,10 @@ vshError(vshControl *ctl, const char *format, ...) va_end(ap); } + /* Most output is to stdout, but if someone ran virsh 2>&1, then + * printing to stderr will not interleave correctly with stdout + * unless we flush between every transition between streams. */ + fflush(stdout); fputs(_("error: "), stderr); va_start(ap, format); @@ -14345,6 +14352,7 @@ vshError(vshControl *ctl, const char *format, ...) va_end(ap); fprintf(stderr, "%s\n", NULLSTR(str)); + fflush(stderr); VIR_FREE(str); } -- 1.7.4.4

2 2

[libvirt] Fwd: [virt-tools-list] virt-manager, dependency question
by Jason Helfman 20 Aug '11

20 Aug '11

Any thoughts on this to this list? Figure I would get a quicker response here, possibly. Thanks, Jason -- Jason Helfman System Administrator experts-exchange.com http://www.experts-exchange.com/M_4830110.html E4AD 7CF1 1396 27F6 79DD 4342 5E92 AD66 8C8C FBA5

1 0

[libvirt] xml style?
by Eric Blake 20 Aug '11

20 Aug '11

When writing an attribute in xml, which style is preferred? <element attr='single'/> <element attr="double"/> We are currently rather split-brain in libvirt, among three general areas: test programs under tests/*data/*, documentation under docs/*.html.in, and RNG schemas under docs/schemas/*. In all cases, both styles are present, but in different proportions: $ git grep '="' tests/*data |wc 48 163 4276 $ git grep "='" tests/*data |wc 5994 21336 526402 $ git grep '="' docs/schemas/ |wc 1522 4599 103050 $ git grep "='" docs/schemas/ |wc 566 1699 33005 $ git grep '="' docs/*.in |wc 1288 6436 114146 $ git grep "='" docs/*.in |wc 611 2242 42950 Is it worth a cleanup patch that settles on one particular style, or even a syntax check that forces style compliance? -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 1

[libvirt] [PATCHv2 1/2] virsh: add virsh snapshot-current --name
by Eric Blake 20 Aug '11

20 Aug '11

Sometimes, full XML is too much; since most snapshot commands operate on a snapshot name, there should be an easy way to get at the current snapshot's name. * tools/virsh.c (cmdSnapshotCurrent): Add an option. * tools/virsh.pod (snapshot-current): Document it. --- v2: use xpath instead of strstr; replaces: https://www.redhat.com/archives/libvir-list/2011-August/msg00329.html tools/virsh.c | 29 ++++++++++++++++++++++++++--- tools/virsh.pod | 4 +++- 2 files changed, 29 insertions(+), 4 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index bffec1d..fd64020 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -12034,6 +12034,7 @@ static const vshCmdInfo info_snapshot_current[] = { static const vshCmdOptDef opts_snapshot_current[] = { {"domain", VSH_OT_DATA, VSH_OFLAG_REQ, N_("domain name, id or uuid")}, + {"name", VSH_OT_BOOL, 0, N_("list the name, rather than the full xml")}, {NULL, 0, 0, NULL} }; @@ -12044,6 +12045,7 @@ cmdSnapshotCurrent(vshControl *ctl, const vshCmd *cmd) bool ret = false; int current; virDomainSnapshotPtr snapshot = NULL; + char *xml = NULL; if (!vshConnectionUsability(ctl, ctl->conn)) goto cleanup; @@ -12056,7 +12058,7 @@ cmdSnapshotCurrent(vshControl *ctl, const vshCmd *cmd) if (current < 0) goto cleanup; else if (current) { - char *xml; + char *name = NULL; if (!(snapshot = virDomainSnapshotCurrent(dom, 0))) goto cleanup; @@ -12065,13 +12067,34 @@ cmdSnapshotCurrent(vshControl *ctl, const vshCmd *cmd) if (!xml) goto cleanup; - vshPrint(ctl, "%s", xml); - VIR_FREE(xml); + if (vshCommandOptBool(cmd, "name")) { + xmlDocPtr xmldoc = NULL; + xmlXPathContextPtr ctxt = NULL; + + xmldoc = xmlReadDoc((const xmlChar *) xml, "domainsnapshot.xml", + NULL, XML_PARSE_NOENT | XML_PARSE_NONET | + XML_PARSE_NOWARNING); + if (!xmldoc) + goto cleanup; + ctxt = xmlXPathNewContext(xmldoc); + if (!ctxt) { + xmlFreeDoc(xmldoc); + } + + name = virXPathString("string(/domainsnapshot/name)", ctxt); + xmlXPathFreeContext(ctxt); + xmlFreeDoc(xmldoc); + if (!name) + goto cleanup; + } + + vshPrint(ctl, "%s", name ? name : xml); } ret = true; cleanup: + VIR_FREE(xml); if (snapshot) virDomainSnapshotFree(snapshot); if (dom) diff --git a/tools/virsh.pod b/tools/virsh.pod index ec778bf..53376d6 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -1569,9 +1569,11 @@ Create a snapshot for domain I<domain> with the given <name> and value. If I<--print-xml> is specified, then XML appropriate for I<snapshot-create> is output, rather than actually creating a snapshot. -=item B<snapshot-current> I<domain> +=item B<snapshot-current> I<domain> [I<--name>] Output the snapshot XML for the domain's current snapshot (if any). +If I<--name> is specified, just list the snapshot name instead of the +full xml. =item B<snapshot-list> I<domain> -- 1.7.4.4

2 6

[libvirt] [PATCH 0/3] xml simplifications
by Eric Blake 19 Aug '11

19 Aug '11

I noticed this while working on my snapshot stuff. There was lots of copy and paste going on in virsh.c, and it felt awful to not be able to represent the same idea in fewer lines, so I fixed it. This will probably mean I have to rebase the entire snapshot series, since it will introduce several conflicts in at least virsh.c; oh well. Eric Blake (3): maint: treat more libxml2 functions as free-like xml: add another convenience function maint: simplify lots of libxml2 clients cfg.mk | 2 + src/conf/domain_conf.c | 20 +------ src/conf/storage_conf.c | 8 +--- src/cpu/cpu.c | 18 +----- src/esx/esx_vi.c | 17 ++----- src/libvirt_private.syms | 2 - src/qemu/qemu_migration.c | 9 +--- src/security/virt-aa-helper.c | 12 +---- src/test/test_driver.c | 13 +---- src/util/xml.c | 53 +++++------------- src/util/xml.h | 84 +++++++++++++++++++++++++--- tests/cputest.c | 9 +--- tools/virsh.c | 120 +++++++++-------------------------------- 13 files changed, 135 insertions(+), 232 deletions(-) -- 1.7.4.4

2 9

[libvirt] Allow to undefine a running domain
by Osier Yang 19 Aug '11

19 Aug '11

Per API virDomainUndefineFlags' doc says, if a domain is running, it will be converted to transient, but keep running. However, the drivers prohibit one undefining a running domain. This patch series modify the internal domainUndefineFlags function of all drivers (except some driver don't need to check if the domain is running, it's handled by the underlying hypervisor, such as ESX and XEND). The principle is: 1) Set vm->persistent = 0 for a running domain 2) domainDestroy and domainShutdown will take care of remove the domain obj from the hash table. [PATCH 1/8] libxl: Allow to undefine a running domain [PATCH 2/8] lxc: Allow to undefine a running domain [PATCH 3/8] openvz: Allow to undefine a running domain [PATCH 4/8] qemu: Allow to undefine a running domain [PATCH 5/8] test: Allow to undefine a running domain [PATCH 6/8] uml: Allow to undefine a running domain [PATCH 7/8] vmware: Allow to undefine a running domain [PATCH 8/8] xen: Allow to undefine a running domain (xm_internal) Split the patches for easy reviewing, will merge when committing if need. Regards Osier

2 28

[libvirt] [PATCH] qemu: Get memory balloon info correctly for text monitor
by Osier Yang 19 Aug '11

19 Aug '11

* src/qemu/qemu_monitor_text.c: BALLOON_PREFIX was defined as "balloon: actual=", which cause "actual=" is stripped early before the real parsing. This patch changes BALLOON_PREFIX into "balloon: ", and modifies related functions, also renames "qemuMonitorParseExtraBalloonInfo" to "qemuMonitorParseBalloonInfo", as after the changing, it parses all the info returned by "info balloon". --- src/qemu/qemu_monitor_text.c | 47 +++++++++++++++++++++++++---------------- 1 files changed, 29 insertions(+), 18 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 7bf733d..d17d863 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -547,8 +547,12 @@ static int parseMemoryStat(char **text, unsigned int tag, return 0; } - /* Convert bytes to kilobytes for libvirt */ switch (tag) { + /* Convert megabytes to kilobytes for libvirt */ + case VIR_DOMAIN_MEMORY_STAT_ACTUAL_BALLOON: + value = value << 10; + break; + /* Convert bytes to kilobytes for libvirt */ case VIR_DOMAIN_MEMORY_STAT_SWAP_IN: case VIR_DOMAIN_MEMORY_STAT_SWAP_OUT: case VIR_DOMAIN_MEMORY_STAT_UNUSED: @@ -565,15 +569,17 @@ static int parseMemoryStat(char **text, unsigned int tag, /* The reply from the 'info balloon' command may contain additional memory * statistics in the form: '[,<tag>=<val>]*' */ -static int qemuMonitorParseExtraBalloonInfo(char *text, - virDomainMemoryStatPtr stats, - unsigned int nr_stats) +static int qemuMonitorParseBalloonInfo(char *text, + virDomainMemoryStatPtr stats, + unsigned int nr_stats) { char *p = text; unsigned int nr_stats_found = 0; while (*p && nr_stats_found < nr_stats) { - if (parseMemoryStat(&p, VIR_DOMAIN_MEMORY_STAT_SWAP_IN, + if (parseMemoryStat(&p, VIR_DOMAIN_MEMORY_STAT_ACTUAL_BALLOON, + "actual=", &stats[nr_stats_found]) || + parseMemoryStat(&p, VIR_DOMAIN_MEMORY_STAT_SWAP_IN, ",mem_swapped_in=", &stats[nr_stats_found]) || parseMemoryStat(&p, VIR_DOMAIN_MEMORY_STAT_SWAP_OUT, ",mem_swapped_out=", &stats[nr_stats_found]) || @@ -584,9 +590,7 @@ static int qemuMonitorParseExtraBalloonInfo(char *text, parseMemoryStat(&p, VIR_DOMAIN_MEMORY_STAT_UNUSED, ",free_mem=", &stats[nr_stats_found]) || parseMemoryStat(&p, VIR_DOMAIN_MEMORY_STAT_AVAILABLE, - ",total_mem=", &stats[nr_stats_found]) || - parseMemoryStat(&p, VIR_DOMAIN_MEMORY_STAT_ACTUAL_BALLOON, - ",actual=", &stats[nr_stats_found])) + ",total_mem=", &stats[nr_stats_found])) nr_stats_found++; /* Skip to the next label. When *p is ',' the last match attempt @@ -602,7 +606,7 @@ static int qemuMonitorParseExtraBalloonInfo(char *text, /* The reply from QEMU contains 'ballon: actual=421' where value is in MB */ -#define BALLOON_PREFIX "balloon: actual=" +#define BALLOON_PREFIX "balloon: " /* * Returns: 0 if balloon not supported, +1 if balloon query worked @@ -625,13 +629,22 @@ int qemuMonitorTextGetBalloonInfo(qemuMonitorPtr mon, unsigned int memMB; char *end; offset += strlen(BALLOON_PREFIX); - if (virStrToLong_ui(offset, &end, 10, &memMB) < 0) { - qemuReportError(VIR_ERR_OPERATION_FAILED, - _("could not parse memory balloon allocation from '%s'"), reply); - goto cleanup; + + if (STRPREFIX(offset, "actual=")) { + offset += strlen("actual="); + + if (virStrToLong_ui(offset, &end, 10, &memMB) < 0) { + qemuReportError(VIR_ERR_OPERATION_FAILED, + _("could not parse memory balloon allocation from '%s'"), reply); + goto cleanup; + } + *currmem = memMB * 1024; + ret = 1; + } else { + qemuReportError(VIR_ERR_INTERNAL_ERROR, + _("unexpected balloon information '%s'"), reply); + ret = -1; } - *currmem = memMB * 1024; - ret = 1; } else { /* We don't raise an error here, since its to be expected that * many QEMU's don't support ballooning @@ -660,9 +673,7 @@ int qemuMonitorTextGetMemoryStats(qemuMonitorPtr mon, if ((offset = strstr(reply, BALLOON_PREFIX)) != NULL) { offset += strlen(BALLOON_PREFIX); - if ((offset = strchr(offset, ',')) != NULL) { - ret = qemuMonitorParseExtraBalloonInfo(offset, stats, nr_stats); - } + ret = qemuMonitorParseBalloonInfo(offset, stats, nr_stats); } VIR_FREE(reply); -- 1.7.6

4 11

[libvirt] [PATCH] build: fix compilation on mingw64
by Eric Blake 19 Aug '11

19 Aug '11

* .gnulib: Update to latest, for getcwd fixes. --- Spurred by a gnulib bug report by Marc-André Lureau. Fedora currently uses the mingw32 cross-compiler, but is aiming to switch to the mingw64 cross-compiler; once that is done, failure to upgrade gnulib would trip up cross-compilation during ./autobuild.sh. * .gnulib 4470580...56d6664 (21): > getcwd: fix test failures on mingw > getcwd-lgpl: fix m4 to match relaxed test for BSD > getcwd: fix compilation on mingw64 > pipe2: silence compiler warning > autoupdate > relocatable-prog: fix link error > getaddrinfo: fix sh typo in gai_strerrorA decl checking > base64: Reference latest RFC. > * build-aux/bootstrap (slurp): Remove obsolescent gettext.m4 patch. > po/Makefile.in.in: fix make -q problem > configmake: fix make -q problem > git-version-gen: correct the advice in a comment > base64: fix off-by-one buffer size bug > closein: correct comments > More tests for 'fseeko'. > ChangeLog: fix wrong attribution in last commit > fseeko: remove unneeded hack > fseeko: fix bug on glibc > unictype/base: Fix interoperability with preinstalled libunistring. > iswblank: Detect declaration correctly. > tcgetsid: Detect declaration correctly. .gnulib | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/.gnulib b/.gnulib index 4470580..56d6664 160000 --- a/.gnulib +++ b/.gnulib @@ -1 +1 @@ -Subproject commit 4470580881a7b821b52fb5635102ef3e27aa5af4 +Subproject commit 56d6664559f449af25f0d331457b014b02324d65 -- 1.7.4.4

2 2

[libvirt] [RFC] Add a new feild to qemud_save_header to tell if the saved image is corrupt
by Osier Yang 19 Aug '11

19 Aug '11

If one tries to restore a domain from a corrupt save image, we blindly goes forward to restore from it, this can cause many different errors, depending on how much the image is saved. E.g. https://bugzilla.redhat.com/show_bug.cgi?id=730750 So I'm thinking if we can introduce a new feild to struct qemud_save_header, such as "bool complete;", and set it true if succeeded to save the image, false if not. So that could do some checking while trying to open the image (qemuDomainSaveImageOpen), and quit early if "complete" is false, with a sensiable error message. Thought? Thanks Osier

3 9

[libvirt] [PATCH] storage: Flush host cache after write
by Michal Privoznik 19 Aug '11

19 Aug '11

Although we are flushing cache after some critical writes (e.g. volume creation), after some others we do not (e.g. volume cloning). This patch fix this issue. That is for volume cloning, writing header of logical volume, and storage wipe. --- src/storage/storage_backend.c | 8 ++++++++ src/storage/storage_backend_logical.c | 7 +++++++ src/storage/storage_driver.c | 8 ++++++++ 3 files changed, 23 insertions(+), 0 deletions(-) diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index 6243d1e..889f530 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -208,6 +208,14 @@ virStorageBackendCopyToFD(virStorageVolDefPtr vol, } while ((amtleft -= interval) > 0); } + if (fdatasync(fd) < 0) { + ret = -errno; + virReportSystemError(errno, _("cannot sync data to file '%s'"), + vol->target.path); + goto cleanup; + } + + if (VIR_CLOSE(inputfd) < 0) { ret = -errno; virReportSystemError(errno, diff --git a/src/storage/storage_backend_logical.c b/src/storage/storage_backend_logical.c index c622d2a..ca4166d 100644 --- a/src/storage/storage_backend_logical.c +++ b/src/storage/storage_backend_logical.c @@ -424,6 +424,13 @@ virStorageBackendLogicalBuildPool(virConnectPtr conn ATTRIBUTE_UNUSED, VIR_FORCE_CLOSE(fd); goto cleanup; } + if (fsync(fd) < 0) { + virReportSystemError(errno, + _("cannot flush header of device'%s'"), + pool->def->source.devices[i].path); + VIR_FORCE_CLOSE(fd); + goto cleanup; + } if (VIR_CLOSE(fd) < 0) { virReportSystemError(errno, _("cannot close device '%s'"), diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c index 6715790..68cac1f 100644 --- a/src/storage/storage_driver.c +++ b/src/storage/storage_driver.c @@ -1777,6 +1777,14 @@ storageWipeExtent(virStorageVolDefPtr vol, remaining -= written; } + if (fdatasync(fd) < 0) { + ret = -errno; + virReportSystemError(errno, + _("cannot sync data to volume with path '%s'"), + vol->target.path); + goto out; + } + VIR_DEBUG("Wrote %zu bytes to volume with path '%s'", *bytes_wiped, vol->target.path); -- 1.7.3.4

3 4

[libvirt] [PATCH v2] daemon: initialize GnuTLS
by Michal Privoznik 19 Aug '11

19 Aug '11

When spice_tls is set but listen_tls is not, we don't initialize GnuTLS library. So any later gnutls call (e.g. during migration, where we initialize a certificate) will access uninitialized GnuTLS internal structs and throws an error. Although, we might now initialize GnuTLS twice, it is safe according to the documentation: This function can be called many times, but will only do something the first time. This patch creates 2 functions: virNetTLSInit and virNetTLSDeinit with respect to written above. --- diff to v1: - moved from qemu to daemon - created special init function daemon/libvirtd.c | 2 ++ src/rpc/virnettlscontext.c | 25 ++++++++++++++++++++++--- src/rpc/virnettlscontext.h | 3 +++ 3 files changed, 27 insertions(+), 3 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index b99c637..0530ba5 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -1516,6 +1516,7 @@ int main(int argc, char **argv) { virHookCall(VIR_HOOK_DRIVER_DAEMON, "-", VIR_HOOK_DAEMON_OP_START, 0, "start", NULL); + virNetTLSInit(); if (daemonSetupNetworking(srv, config, sock_file, sock_file_ro, ipsock, privileged) < 0) { @@ -1554,6 +1555,7 @@ cleanup: virNetServerProgramFree(qemuProgram); virNetServerClose(srv); virNetServerFree(srv); + virNetTLSDeinit(); if (statuswrite != -1) { if (ret != 0) { /* Tell parent of daemon what failed */ diff --git a/src/rpc/virnettlscontext.c b/src/rpc/virnettlscontext.c index 19a9b25..8482eaf 100644 --- a/src/rpc/virnettlscontext.c +++ b/src/rpc/virnettlscontext.c @@ -679,9 +679,6 @@ static virNetTLSContextPtr virNetTLSContextNew(const char *cacert, ctxt->refs = 1; - /* Initialise GnuTLS. */ - gnutls_global_init(); - if ((gnutlsdebug = getenv("LIBVIRT_GNUTLS_DEBUG")) != NULL) { int val; if (virStrToLong_i(gnutlsdebug, NULL, 10, &val) < 0) @@ -1399,3 +1396,25 @@ void virNetTLSSessionFree(virNetTLSSessionPtr sess) virMutexDestroy(&sess->lock); VIR_FREE(sess); } + +/* + * This function MUST be called before any + * virNetTLS* because it initializes + * underlying GnuTLS library. According to + * it's documentation, it's safe to be called + * many times, but is not thread safe. Each + * call SHOULD be later followed by + * virNetTLSContextDeinit. + */ +void virNetTLSInit(void) +{ + gnutls_global_init(); +} + +/* + * See virNetTLSInit + */ +void virNetTLSDeinit(void) +{ + gnutls_global_deinit(); +} diff --git a/src/rpc/virnettlscontext.h b/src/rpc/virnettlscontext.h index 641d67e..99f31b9 100644 --- a/src/rpc/virnettlscontext.h +++ b/src/rpc/virnettlscontext.h @@ -30,6 +30,9 @@ typedef struct _virNetTLSSession virNetTLSSession; typedef virNetTLSSession *virNetTLSSessionPtr; +void virNetTLSInit(void); +void virNetTLSDeinit(void); + virNetTLSContextPtr virNetTLSContextNewServerPath(const char *pkipath, bool tryUserPkiPath, const char *const*x509dnWhitelist, -- 1.7.3.4

2 2

[libvirt] [PATCH] schedinfo: add missing documentation
by Taku Izumi 19 Aug '11

19 Aug '11

This patch adds the missing documentation about the scheduler parameter "vcpu_period" and "vcpu_quota". Signed-off-by: Taku Izumi <izumi.taku(a)jp.fujitsu.com> --- tools/virsh.pod | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) Index: libvirt/tools/virsh.pod =================================================================== --- libvirt.orig/tools/virsh.pod +++ libvirt/tools/virsh.pod @@ -712,7 +712,9 @@ I<domain-id> Allows you to show (and set) the domain scheduler parameters. The parameters available for each hypervisor are: -LXC, QEMU/KVM (posix scheduler): cpu_shares +LXC (posix scheduler) : cpu_shares + +QEMU/KVM (posix scheduler): cpu_shares, vcpu_period, vcpu_quota Xen (credit scheduler): weight, cap @@ -729,6 +731,10 @@ Therefore, -1 is a useful shorthand for B<Note>: The weight and cap parameters are defined only for the XEN_CREDIT scheduler and are now I<DEPRECATED>. +B<Note>: The vcpu_period parameter has a valid value range of 1000-1000000 or +0, and the vcpu_quota parameter has a valid value range of 1000-1844674407370955 +or less than 0. + =item B<screenshot> I<domain-id> [I<imagefilepath>] [I<--screen> B<screenID>] Takes a screenshot of a current domain console and stores it into a file.

2 1

[libvirt] [PATCH] maint: ignore generated file
by Eric Blake 19 Aug '11

19 Aug '11

I did 'git add .' while in the middle of 'make syntax-check', and it picked up a temporary file that should not be committed. * .gitignore: Ignore sc_* from syntax check. --- Pushing under the trivial rule. .gitignore | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/.gitignore b/.gitignore index e8b2dbf..39ecd6d 100644 --- a/.gitignore +++ b/.gitignore @@ -53,6 +53,7 @@ /po/* /proxy/ /python/generator.py.stamp +/sc_* /src/libvirt_iohelper /src/locking/qemu-sanlock.conf /src/remote/*_client_bodies.h -- 1.7.4.4

1 0

[libvirt] [RFC] Should libvirt be a proxy for migration data
by Jiri Denemark 18 Aug '11

18 Aug '11

Hi all, Currently when we start a non-tunneled migration, data go straight from source qemu to destination qemu. This is nice in that there is no additional overhead but it also has several disadvantages. If the communication between source and destination qemu breaks, we only get unexpected error message from qemu with no glue about what happened. Another issue is that if qemu cannot send migration data, we cannot cancel the migration because migrate_cancel blocks until all buffers with migration data queued up for transmission are written into the socket. That said, I think we should act as a proxy between source and destination qemu so that we can detect and report normal errors (such as connection reset by peer) and cancel migration at any time. Since we have virNetSocket and we already use that for connecting to destination qemu, we should use it for proxying migration data as well. This approach also has some disadvantages, e.g., a single libvirt thread instead of several qemu processes will now send migration data from all domains that are being migrated. However, I feel like the gain is bigger than the downside. And we already do the same for tunneled migration anyway. Any objections? Jirka

2 2

Re: [libvirt] [Qemu-devel] Killing block migration in qemu?
by Anthony Liguori 18 Aug '11

18 Aug '11

On 08/17/2011 11:51 AM, Paolo Bonzini wrote: > Hi all, > > following discussions yesterday with Juan Quintela and Marcelo Tosatti, > here is my humble proposal: remove block migration from qemu master. +1 I was going to propose the same thing after the conference actually. Regards, Anthony Liguori

1 0

[libvirt] [PATCH 0/5 v2] Some rearrangement & additions to pid file handling/locking
by Daniel P. Berrange 18 Aug '11

18 Aug '11

An update to: https://www.redhat.com/archives/libvir-list/2011-August/msg00349.html Changes in v2: - Fix all docs & code bugs from previous review

4 13

[libvirt] RFC (V2) New virDomainBlockPull API family to libvirt
by Adam Litke 18 Aug '11

18 Aug '11

Here are the patches to implement the BlockPull/BlockJob API as discussed and agreed to. I am testing with a python script (included for completeness as the final patch). The qemu monitor interface is not expected to change in the future. Stefan is planning to submit placeholder commands for upstream qemu until the generic streaming support is implemented. Changes since V1: - Make virDomainBlockPullAbort() and virDomainGetBlockPullInfo() into a generic BlockJob interface. - Added virDomainBlockJobSetSpeed() - Rename VIR_DOMAIN_EVENT_ID_BLOCK_PULL event to fit into block job API - Add bandwidth argument to virDomainBlockPull() Summary of changes since first generation patch series: - Qemu dropped incremental streaming so remove libvirt incremental BlockPull() API - Rename virDomainBlockPullAll() to virDomainBlockPull() - Changes required to qemu monitor handlers for changed command names -- To help speed the provisioning process for large domains, new QED disks are created with backing to a template image. These disks are configured with copy on read such that blocks that are read from the backing file are copied to the new disk. This reduces I/O over a potentially costly path to the backing image. In such a configuration, there is a desire to remove the dependency on the backing image as the domain runs. To accomplish this, qemu will provide an interface to perform sequential copy on read operations during normal VM operation. Once all data has been copied, the disk image's link to the backing file is removed. The virDomainBlockPull API family brings this functionality to libvirt. virDomainBlockPull() instructs the hypervisor to stream the entire device in the background. Progress of this operation can be checked with the function virDomainBlockJobInfo(). An ongoing stream can be cancelled with virDomainBlockJobAbort(). virDomainBlockJobSetSpeed() allows you to limit the bandwidth that the operation may consume. An event (VIR_DOMAIN_EVENT_ID_BLOCK_JOB) will be emitted when a disk has been fully populated or if a BlockPull() operation was terminated due to an error. This event is useful to avoid polling on virDomainBlockJobInfo() for completion and could also be used by the security driver to revoke access to the backing file when it is no longer needed. make check: PASS make syntax-check: PASS make -C tests valgrind: PASS [PATCH 1/8] Add new API virDomainBlockPull* to headers [PATCH 2/8] virDomainBlockPull: Implement the main entry points [PATCH 3/8] Add virDomainBlockPull support to the remote driver [PATCH 4/8] Implement virDomainBlockPull for the qemu driver [PATCH 5/8] Enable the virDomainBlockPull API in virsh [PATCH 6/8] Enable virDomainBlockPull in the python API. [PATCH 7/8] Asynchronous event for BlockJob completion [PATCH 8/8] Test the blockJob/BlockPull API

6 22

Re: [libvirt] [Qemu-devel] Killing block migration in qemu?
by Stefan Hajnoczi 18 Aug '11

18 Aug '11

On Wed, Aug 17, 2011 at 5:51 PM, Paolo Bonzini <pbonzini(a)redhat.com> wrote: > following discussions yesterday with Juan Quintela and Marcelo Tosatti, here > is my humble proposal: remove block migration from qemu master. It seems to > me that keeping block migration is going to slow down further improvements > on migration. The main problems are: > > 1) there are very good reasons to move migration to a separate thread. Only > a limited amount of extra locking, perhaps none is needed in order to do so > for RAM and devices. But the block drivers pretty much need to run under > the I/O thread lock, and coroutines will not help if the I/O thread is taken > by another thread. It's hard/unreliable/pointless to ping-pong migration > between threads. The image streaming approach will also run in the I/O thread for the mid-term future. Is the problem that the block migration code today is too tied into the actual migration code path and therefore stops from using it when migration happens in a separate thread? > > 2) there already are plans to reimplement block migration... it's called > streaming :) and not coincidentially it reuses some of the block migration > code. What are the concrete issues with the existing block migration code? I'm not disagreeing that we should move to a streaming approach but I simply don't know the details of the existing block migration code. > Here is how it would go: This sounds reasonable. In fact we can do both pre-copy and post-copy block migration using streaming (+mirroring). Stefan

2 1

[libvirt] [PATCH] qemu: Initialize GnuTLS
by Michal Privoznik 18 Aug '11

18 Aug '11

When spice_tls is set but listen_tls is not, we don't initialize GnuTLS library. So any later gnutls call (e.g. during migration, where we initialize a certificate) will access uninitialized GnuTLS internal structs and throws an error. Although, we might now initialize GnuTLS twice, it is safe according to the documentation: This function can be called many times, but will only do something the first time. --- src/qemu/qemu_driver.c | 13 +++++++++++++ 1 files changed, 13 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 421a98e..5fe20b6 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -44,6 +44,7 @@ #include <sys/ioctl.h> #include <sys/un.h> #include <byteswap.h> +#include <gnutls/gnutls.h> #include "qemu_driver.h" @@ -537,6 +538,15 @@ qemudStartup(int privileged) { } VIR_FREE(driverConf); + if (qemu_driver->spiceTLS) { + /* Initialize GnuTLS. If it was initialized before, + * it doesn't hurt. From GnuTLS documentation: + * This function can be called many times, + * but will only do something the first time. + */ + gnutls_global_init(); + } + /* We should always at least have the 'nop' manager, so * NULLs here are a fatal error */ @@ -754,6 +764,9 @@ qemudShutdown(void) { qemuProcessAutoDestroyShutdown(qemu_driver); + if (qemu_driver->spiceTLS) + gnutls_global_deinit(); + VIR_FREE(qemu_driver->configDir); VIR_FREE(qemu_driver->autostartDir); VIR_FREE(qemu_driver->logDir); -- 1.7.3.4

2 1

[libvirt] [PATCH 0/2] Libvirt-java event handling
by js＠alien8.de 17 Aug '11

17 Aug '11

From: Julian Stecklina <jsteckli(a)os.inf.tu-dresden.de> I noticed that event handling in the libvirt Java bindings is incomplete. The first patch adds required data types and constants. The second relaxes the visibility of the Domain constructor, otherwise Callbacks are kind of useless. Regards, Julian Julian Stecklina (2): Allow different types of callbacks (just as the C API). Allow to construct Domain objects from inside callbacks. src/main/java/org/libvirt/Connect.java | 3 ++- src/main/java/org/libvirt/Domain.java | 2 +- src/main/java/org/libvirt/jna/Libvirt.java | 22 ++++++++++++++++++++-- 3 files changed, 23 insertions(+), 4 deletions(-) -- 1.7.3.4

3 4

Re: [libvirt] patch for disk migration verbose progress
by Eric Blake 17 Aug '11

17 Aug '11

[re-adding the list] On 08/17/2011 12:30 AM, Tom Vijlbrief wrote: > Hi Eric > > This is the complete patch including json (not tested) and a typo correction > in an error message... > > Tom > ---------- Doorgestuurd bericht ---------- > Van: "Tom Vijlbrief"<tom(a)nomadbsd.v7f.eu> > Datum: 17 aug. 2011 08:23 > Onderwerp: mon.patch > Aan:<tvijlbrief(a)gmail.com> > > > mon.patch > Thanks again for taking the time to contribute. It is easier to work with patches generated by 'git format-patch' (and 'git send-email' is a nice wrapper that both formats the patch and mails it to the list); as written, your patch had no commit message, so I had to invent one. I also added you to AUTHORS; let me know if I need to update any preferred spellings. > > # NB, must pass the --listen flag to the libvirtd process for this to Is this a stray line in your email, or is it pointing out a further issue? > diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c > index 2a9a078..618a45c 100644 > --- a/src/qemu/qemu_monitor_json.c > +++ b/src/qemu/qemu_monitor_json.c > @@ -1894,6 +1894,31 @@ qemuMonitorJSONGetMigrationStatusReply(virJSONValuePtr reply, > _("migration was active, but RAM 'total' data was missing")); > return -1; > } > + > + virJSONValuePtr disk = virJSONValueObjectGet(ret, "disk"); > + if (!disk) { > + return 0; > + } > + > + unsigned long long t; Our style tends to float declarations to the top, C89 style, even though this is valid as written (we require C99 for other reasons). > + if (virJSONValueObjectGetNumberUlong(disk, "transferred",&t)< 0) { > + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", > + _("disk migration was active, but DISK 'transferred' data was missing")); > + return -1; > + } > + (*transferred)+= t; > + if (virJSONValueObjectGetNumberUlong(disk, "remaining",&t)< 0) { > + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", > + _("disk migration was active, but DISK 'remaining' data was missing")); > + return -1; > + } > + (*remaining)+= t; I tweaked spacing between operators (yeah, it doesn't help that thunderbird corrupted spacing even worse in my reply to your mail). And the () are redundant; this can safely be: *remaining += t; > + if (virJSONValueObjectGetNumberUlong(disk, "total",&t)< 0) { > + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", > + _("disk migration was active, but DISK 'total' data was missing")); > + return -1; > + } > + (*total)+= t; > } > > return 0; > diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c > index 7bf733d..fef6f36 100644 > --- a/src/qemu/qemu_monitor_text.c > +++ b/src/qemu/qemu_monitor_text.c > @@ -149,7 +149,7 @@ int qemuMonitorTextIOProcess(qemuMonitorPtr mon ATTRIBUTE_UNUSED, > passwd = strstr(start, PASSWORD_PROMPT); > if (passwd) { > #if DEBUG_IO > - VIR_DEBUG("Seen a passwowrd prompt [%s]", data + used); > + VIR_DEBUG("Seen a password prompt [%s]", data + used); > #endif > if (msg->passwordHandler) { > int i; > @@ -1183,6 +1183,9 @@ cleanup: > #define MIGRATION_TRANSFER_PREFIX "transferred ram: " > #define MIGRATION_REMAINING_PREFIX "remaining ram: " > #define MIGRATION_TOTAL_PREFIX "total ram: " > +#define MIGRATION_DISK_TRANSFER_PREFIX "transferred disk: " > +#define MIGRATION_DISK_REMAINING_PREFIX "remaining disk: " > +#define MIGRATION_DISK_TOTAL_PREFIX "total disk: " > > int qemuMonitorTextGetMigrationStatus(qemuMonitorPtr mon, > int *status, > @@ -1246,6 +1249,7 @@ int qemuMonitorTextGetMigrationStatus(qemuMonitorPtr mon, > goto cleanup; > } > *remaining *= 1024; > + tmp = end; > > if (!(tmp = strstr(tmp, MIGRATION_TOTAL_PREFIX))) > goto done; > @@ -1257,7 +1261,53 @@ int qemuMonitorTextGetMigrationStatus(qemuMonitorPtr mon, > goto cleanup; > } > *total *= 1024; > + tmp = end; > + > + /* > + * Check for Optional Disk Migration status > + */ > + > + unsigned long long disk_transferred= 0; > + unsigned long long disk_remaining= 0; > + unsigned long long disk_total= 0; > + > + if (!(tmp = strstr(tmp, MIGRATION_DISK_TRANSFER_PREFIX))) > + goto done; > + tmp += strlen(MIGRATION_DISK_TRANSFER_PREFIX); > > + if (virStrToLong_ull(tmp,&end, 10,&disk_transferred)< 0) { > + qemuReportError(VIR_ERR_INTERNAL_ERROR, > + _("cannot parse disk migration data transferred statistic %s"), tmp); > + goto cleanup; > + } > + disk_transferred *= 1024; > + (*transferred)+= disk_transferred; For one less assignment, I merged these two lines into: *transferred += disk_transferred * 1024; > + tmp = end; > + > + if (!(tmp = strstr(tmp, MIGRATION_DISK_REMAINING_PREFIX))) > + goto done; > + tmp += strlen(MIGRATION_DISK_REMAINING_PREFIX); > + > + if (virStrToLong_ull(tmp,&end, 10,&disk_remaining)< 0) { > + qemuReportError(VIR_ERR_INTERNAL_ERROR, > + _("cannot parse disk migration data remaining statistic %s"), tmp); > + goto cleanup; > + } > + disk_remaining *= 1024; > + (*remaining)+= disk_remaining; > + tmp = end; > + > + if (!(tmp = strstr(tmp, MIGRATION_DISK_TOTAL_PREFIX))) > + goto done; > + tmp += strlen(MIGRATION_DISK_TOTAL_PREFIX); > + > + if (virStrToLong_ull(tmp,&end, 10,&disk_total)< 0) { > + qemuReportError(VIR_ERR_INTERNAL_ERROR, > + _("cannot parse disk migration data total statistic %s"), tmp); > + goto cleanup; > + } > + disk_total *= 1024; > + (*total)+= disk_total; > } > } ACK and pushed with those tweaks. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 0

Re: [libvirt] Patch for verbose Disk transfer
by Eric Blake 17 Aug '11

17 Aug '11

[re-adding the list] On 08/16/2011 11:52 PM, Tom Vijlbrief wrote: > Hi Eric, > > Yes json has similar info and I also patched it, but I don't know how to > configure vir-sh so that it uses the json channel for at least testing the > patch. I can send you the patch anyway, it's a simple change... Use of JSON depends on which version of qemu you are running; if you are using Fedora 15 or newer, then JSON should automatically be in use. It's not something you can configure at the virsh command line. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 0

[libvirt] [PATCH] qemu: Init reattaching related members of pciDevice before reattach
by Osier Yang 17 Aug '11

17 Aug '11

Otherwise the device will still be bound to pci-stub driver even it's set as "managed=yes" when do detaching. Of course, it won't triger any driver reprobing too. --- src/qemu/qemu_hotplug.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 5f449fb..b7fdfa0 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -1761,6 +1761,7 @@ int qemuDomainDetachHostPciDevice(struct qemud_driver *driver, pciDeviceListDel(driver->activePciHostdevs, pci); if (pciResetDevice(pci, driver->activePciHostdevs, NULL) < 0) ret = -1; + pciDeviceReAttachInit(pci); qemuReattachPciDevice(pci, driver); pciFreeDevice(pci); } -- 1.7.6

2 2

[libvirt] [PATCH] conf: Do not load domain if the domain with same already exists
by Osier Yang 17 Aug '11

17 Aug '11

We don't allow to define domain with same name and different UUID, or with same UUID, so it's reasonable to not load the domain config if there is domain with same name already exists. Otherwise it can cause problem like: 1) % cp /etc/libvirt/qemu/dom.xml /etc/libvirt/qemu/dom_diffuuid.xml 2) remove the line with "uuid" in the "dom_diffuuid.xml" 3) % for i in {1..10}; do kill -SIGHUP $(pidof libvirtd); done 4) % virsh list --all There will be 11 domains listed with the same name, as if there is no UUID specified in domain XML, libvirt will generate one for it, which will be definitely different with the original one. --- src/conf/domain_conf.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index ce1f3c5..0bed929 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -10535,7 +10535,8 @@ static virDomainObjPtr virDomainLoadConfig(virCapsPtr caps, /* if the domain is already in our hashtable, we only need to * update the autostart flag */ - if ((dom = virDomainFindByUUID(doms, def->uuid))) { + if ((dom = virDomainFindByUUID(doms, def->uuid)) || + (dom = virDomainFindByName(doms, def->name))) { dom->autostart = autostart; VIR_FREE(configFile); -- 1.7.6

2 2

[libvirt] [PATCH] Ensure stream is aborted when exiting console
by Daniel P. Berrange 17 Aug '11

17 Aug '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> After running 'virsh console' in interactive mode, there was a missing call to virStreamAbort, which meant the server kept the stream resources open * tools/console.c: Abort stream when exiting --- tools/console.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/tools/console.c b/tools/console.c index 171ebc9..82ef537 100644 --- a/tools/console.c +++ b/tools/console.c @@ -90,9 +90,11 @@ static void virConsoleShutdown(virConsolePtr con) { con->quit = true; - virStreamEventRemoveCallback(con->st); - if (con->st) + if (con->st) { + virStreamEventRemoveCallback(con->st); + virStreamAbort(con->st); virStreamFree(con->st); + } if (con->stdinWatch != -1) virEventRemoveHandle(con->stdinWatch); if (con->stdoutWatch != -1) -- 1.7.6

3 6

[libvirt] [PATCH] Ensure async packets never get marked for sync replies
by Daniel P. Berrange 17 Aug '11

17 Aug '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> If a client had initiated a stream abort, it will have a call waiting for a reply in the queue. If more data continues to arrive on the stream, the abort command could mistakenly get signalled as complete. Remove the code from async data processing that looked for waiting calls. Add a sanity check to ensure no async call can ever be marked as needing a reply * src/rpc/virnetclient.c: Ensure async data packets can't trigger a reply --- src/rpc/virnetclient.c | 12 +++++++----- 1 files changed, 7 insertions(+), 5 deletions(-) diff --git a/src/rpc/virnetclient.c b/src/rpc/virnetclient.c index efaac88..055361d 100644 --- a/src/rpc/virnetclient.c +++ b/src/rpc/virnetclient.c @@ -627,11 +627,6 @@ static int virNetClientCallDispatchStream(virNetClientPtr client) case VIR_NET_CONTINUE: { if (virNetClientStreamQueuePacket(st, &client->msg) < 0) return -1; - - if (thecall && thecall->expectReply) { - VIR_DEBUG("Got sync data packet completion"); - thecall->mode = VIR_NET_CLIENT_MODE_COMPLETE; - } return 0; } @@ -1193,6 +1188,13 @@ int virNetClientSend(virNetClientPtr client, virNetClientCallPtr call; int ret = -1; + if (expectReply && + (msg->header.status == VIR_NET_CONTINUE)) { + virNetError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Attempt to send an asynchronous message with a synchronous reply")); + return -1; + } + if (VIR_ALLOC(call) < 0) { virReportOOMError(); return -1; -- 1.7.6

2 2

[libvirt] [PATCH] Don't attempt to read from a stream if it is closed
by Daniel P. Berrange 17 Aug '11

17 Aug '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> The I/O event callback processes incoming packets first, and then does outgoing packets. If the incoming packet caused the stream to close, then the attempt to process outgoing data resulted in an error. This caused libvirt to then send an error back to the client, but the stream had already been stopped. This confused the client since it sees 2 error events. * daemon/stream.c: Don't attempt read if stream is closed --- daemon/stream.c | 10 +++++++++- 1 files changed, 9 insertions(+), 1 deletions(-) diff --git a/daemon/stream.c b/daemon/stream.c index 7dd9ae7..7d2b367 100644 --- a/daemon/stream.c +++ b/daemon/stream.c @@ -678,7 +678,15 @@ daemonStreamHandleRead(virNetServerClientPtr client, size_t bufferLen = VIR_NET_MESSAGE_PAYLOAD_MAX; int ret; - VIR_DEBUG("client=%p, stream=%p", client, stream); + VIR_DEBUG("client=%p, stream=%p tx=%d closed=%d", + client, stream, stream->tx, stream->closed); + + /* We might have had an event pending before we shut + * down the stream, so if we're marked as closed, + * then do nothing + */ + if (stream->closed) + return 0; /* Shouldn't ever be called unless we're marked able to * transmit, but doesn't hurt to check */ -- 1.7.6

2 1

[libvirt] [RFC] Host CPU passthrough
by Jiri Denemark 17 Aug '11

17 Aug '11

Hi, AFAIK this topic is not new but I think we still do not have a good solution for it. Libvirt already supports specifying what CPU and its features a guest should see but imagine one wants to run a guest on the best possible CPU. The current way is to copy the <cpu> element from capabilities XML into domain XML. This approach is fine since it provides reproducible environment and such domain can even be migrated to a different host. But the CPU shown provided to a guest is not the same as the real host CPU. It's based on a model which doesn't reflect all aspects of real CPUs. Ideally, we would model everything but that's quite complicated and may need updating anytime a new CPU is introduced. I think we should add a possibility to passthrough host CPU into a domain no matter what the CPU looks like. Domains configured in such a way will not be allowed to be migrated since we cannot guarantee the same CPU on destination host but they would be able to see the real CPU. Another benefit is that a single XML configuration would work on any CPU without modifications. Since possible issues with such domains will be harder to debug we can mark them as tainted just like we do if someone calls qemu monitor commands directly. I'm not sure what XML representation makes the most sense so I'll mention several options which I'm thinking about: - "host" ("native", "passthrough", whatever) special model; same name space as real CPU models, I don't like this too much: <cpu match='exact'> <model>host</model> </cpu> - "host" match policy; has nice first-look semantics that the guest CPU matches host's but doesn't play very nice with the defined semantics of "match" attribute: <cpu match='host'/> - new "mode" attribute; IMHO the best one, similar to what we have for SMBIOS: <cpu mode='host'/> vs. <cpu mode='custom' match='...'> <model>...</model. ... </cpu> What do you think? Jirka

3 6

[libvirt] [PATCH] snapshot: tweak misleading wording
by Eric Blake 17 Aug '11

17 Aug '11

Fixes confusion introduced in commit 98369d3. * tools/virsh.c (cmdSnapshotParent): Operates on named snapshot, not current. --- virsh.pod got it right. I'm pushing this under the trivial rule. tools/virsh.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 51ba0a8..49917f3 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -12282,7 +12282,7 @@ cleanup: * "snapshot-parent" command */ static const vshCmdInfo info_snapshot_parent[] = { - {"help", N_("Get the name of the parent of the current snapshot")}, + {"help", N_("Get the name of the parent of a snapshot")}, {"desc", N_("Extract the snapshot's parent, if any")}, {NULL, NULL} }; -- 1.7.4.4

1 0

[libvirt] main: fix some compilation issues on non-linux platforms
by Stefan Berger 17 Aug '11

17 Aug '11

This patch fixes *some* compilation issues on non-Linux platforms (cygwin). Signed-off-by: Stefan Berger <stefanb(a)linux.vnet.ibm.com> diff --git a/src/util/interface.c b/src/util/interface.c index aec12f5..5d473b7 100644 --- a/src/util/interface.c +++ b/src/util/interface.c @@ -1317,7 +1317,7 @@ ifaceGetPhysicalFunction(const char *ifname, char **pfname) } #else int -ifaceIsVirtualFunction(const char *ifname) +ifaceIsVirtualFunction(const char *ifname ATTRIBUTE_UNUSED) { ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", _("ifaceIsVirtualFunction is not supported on non-linux " @@ -1326,8 +1326,9 @@ ifaceIsVirtualFunction(const char *ifname) } int -ifaceGetVirtualFunctionIndex(const char *pfname, const char *vfname, - int *vf_index) +ifaceGetVirtualFunctionIndex(const char *pfname ATTRIBUTE_UNUSED, + const char *vfname ATTRIBUTE_UNUSED, + int *vf_index ATTRIBUTE_UNUSED) { ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", _("ifaceGetVirtualFunctionIndex is not supported on non-linux " @@ -1336,7 +1337,8 @@ ifaceGetVirtualFunctionIndex(const char *pfname, const char *vfname, } int -ifaceGetPhysicalFunction(const char *ifname, char **pfname) +ifaceGetPhysicalFunction(const char *ifname ATTRIBUTE_UNUSED, + char **pfname ATTRIBUTE_UNUSED) { ifaceError(VIR_ERR_INTERNAL_ERROR, "%s", _("ifaceGetPhysicalFunction is not supported on non-linux " diff --git a/src/util/pci.c b/src/util/pci.c index 7040e55..9873f33 100644 --- a/src/util/pci.c +++ b/src/util/pci.c @@ -1685,6 +1685,8 @@ int pciDeviceIsAssignable(pciDevice *dev, return 1; } +#ifdef __linux__ + /* * returns 1 if equal and 0 if not */ @@ -1804,7 +1806,6 @@ out: return ret; } -#ifdef __linux__ /* * Returns Physical function given a virtual function */ @@ -2010,8 +2011,8 @@ out: } #else int -pciGetPhysicalFunction(const char *vf_sysfs_path, - struct pci_config_address **physical_function) +pciGetPhysicalFunction(const char *vf_sysfs_path ATTRIBUTE_UNUSED, + struct pci_config_address **physical_function ATTRIBUTE_UNUSED) { pciReportError(VIR_ERR_INTERNAL_ERROR, _("pciGetPhysicalFunction is not " "supported on non-linux platforms")); @@ -2019,9 +2020,9 @@ pciGetPhysicalFunction(const char *vf_sysfs_path, } int -pciGetVirtualFunctions(const char *sysfs_path, - struct pci_config_address ***virtual_functions, - unsigned int *num_virtual_functions) +pciGetVirtualFunctions(const char *sysfs_path ATTRIBUTE_UNUSED, + struct pci_config_address ***virtual_functions ATTRIBUTE_UNUSED, + unsigned int *num_virtual_functions ATTRIBUTE_UNUSED) { pciReportError(VIR_ERR_INTERNAL_ERROR, _("pciGetVirtualFunctions is not " "supported on non-linux platforms")); @@ -2029,7 +2030,7 @@ pciGetVirtualFunctions(const char *sysfs_path, } int -pciDeviceIsVirtualFunction(const char *vf_sysfs_device_link) +pciDeviceIsVirtualFunction(const char *vf_sysfs_device_link ATTRIBUTE_UNUSED) { pciReportError(VIR_ERR_INTERNAL_ERROR, _("pciDeviceIsVirtualFunction is " "not supported on non-linux platforms")); @@ -2037,9 +2038,9 @@ pciDeviceIsVirtualFunction(const char *vf_sysfs_device_link) } int -pciGetVirtualFunctionIndex(const char *pf_sysfs_device_link, - const char *vf_sysfs_device_link, - int *vf_index) +pciGetVirtualFunctionIndex(const char *pf_sysfs_device_link ATTRIBUTE_UNUSED, + const char *vf_sysfs_device_link ATTRIBUTE_UNUSED, + int *vf_index ATTRIBUTE_UNUSED) { pciReportError(VIR_ERR_INTERNAL_ERROR, _("pciGetVirtualFunctionIndex is " "not supported on non-linux platforms")); @@ -2048,7 +2049,8 @@ pciGetVirtualFunctionIndex(const char *pf_sysfs_device_link, } int -pciDeviceNetName(char *device_link_sysfs_path, char **netname) +pciDeviceNetName(char *device_link_sysfs_path ATTRIBUTE_UNUSED, + char **netname ATTRIBUTE_UNUSED) { pciReportError(VIR_ERR_INTERNAL_ERROR, _("pciDeviceNetName is not " "supported on non-linux platforms")); diff --git a/src/util/virfile.c b/src/util/virfile.c index 0edf058..1158998 100644 --- a/src/util/virfile.c +++ b/src/util/virfile.c @@ -44,7 +44,7 @@ int virFileClose(int *fdptr, bool preserve_errno) { - int saved_errno; + int saved_errno = 0; int rc = 0; if (*fdptr >= 0) { @@ -62,7 +62,7 @@ int virFileClose(int *fdptr, bool preserve_errno) int virFileFclose(FILE **file, bool preserve_errno) { - int saved_errno; + int saved_errno = 0; int rc = 0; if (*file) { diff --git a/src/util/virpidfile.c b/src/util/virpidfile.c index 7dd3c51..1927051 100644 --- a/src/util/virpidfile.c +++ b/src/util/virpidfile.c @@ -213,6 +213,8 @@ int virPidFileReadPathIfAlive(const char *path, #ifdef __linux__ if (virFileLinkPointsTo(procpath, binpath) == 0) *pid = -1; +#else + (void)binpath; #endif VIR_FREE(procpath);

2 8

[libvirt] Bug - libvirt persists on having dnsmasq
by Zdenek Styblik 17 Aug '11

17 Aug '11

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, I've noticed libvirt is persisting to have dnsmasq present and use it. Version of libvirt in question is 0.9.4. There is no dnsmasq present on the Host during libvirt compilation, yet it seems to me ./configure wrongfully finds one: ~~~ SNIP ~~~ configure:46087: checking for dnsmasq configure:46119: result: dnsmasq ~~~ SNIP ~~~ shouldn't there be "result: no"? As far as I remember - no dnsmasq present during compilation, no usage of it later. However I might have missed something - change in behavior, requirements etc. This is possible. Still, if required "component" is missing, I say compilation resp. configure should fail. And this of course breaks things later on: ~~~ SNIP ~~~ virsh # net-start virtnet01 error: Failed to start network virtnet01 error: Cannot find 'dnsmasq' in path: No such file or directory ~~~ SNIP ~~~ and there is no DHCP defined for 'virtnet01' thus there is no reason to bring dnsmasq in. Neither can be new network created and started via virt-manager. Thanks, Zdenek - -- Zdenek Styblik email: stybla(a)turnovfree.net jabber: stybla(a)jabber.turnovfree.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk5H2OwACgkQ8MreUbSH7ingtgCePYxU4zroZQo+3xrSxT7/hekN Qc8An3U/nDY/C8Z0oLQ76Za1SrR++0ON =TkKh -----END PGP SIGNATURE-----

4 10

[libvirt] [PATCH] Fix race condition in abort of stream
by Daniel P. Berrange 17 Aug '11

17 Aug '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> If a stream gets a server initiated abort, the client may still send an abort request before it receives the server side abort. This causes the server to send back another abort for the stream. Since the protocol defines that abort is the last thing to be sent, the client gets confused by this second abort from the server. If the stream is already shutdown, just drop any client requested abort, rather than sending back another message. This fixes the regression from previous versions. Tested as follows In one virsh session virsh # start foo virsh # console foo In other virsh session virsh # destroy foo The first virsh session should be able to continue issues commands without error. Prior to this patch it saw virsh # list error: Failed to list active domains error: An error occurred, but the cause is unknown virsh # list error: Failed to list active domains error: no call waiting for reply with prog 536903814 vers 1 serial 9 * src/rpc/virnetserverprogram.c: Drop abort requests for streams which no longer exist --- src/rpc/virnetserverprogram.c | 25 ++++++++----------------- 1 files changed, 8 insertions(+), 17 deletions(-) diff --git a/src/rpc/virnetserverprogram.c b/src/rpc/virnetserverprogram.c index 63a6b6d..ca80ae0 100644 --- a/src/rpc/virnetserverprogram.c +++ b/src/rpc/virnetserverprogram.c @@ -257,23 +257,14 @@ int virNetServerProgramDispatch(virNetServerProgramPtr prog, * stream packets after we closed down a stream. Just drop & ignore * these. */ - if (msg->header.status == VIR_NET_CONTINUE) { - VIR_INFO("Ignoring unexpected stream data serial=%d proc=%d status=%d", - msg->header.serial, msg->header.proc, msg->header.status); - /* Send a dummy reply to free up 'msg' & unblock client rx */ - memset(msg, 0, sizeof(*msg)); - msg->header.type = VIR_NET_REPLY; - if (virNetServerClientSendMessage(client, msg) < 0) { - ret = -1; - goto cleanup; - } - } else { - VIR_INFO("Unexpected stream control message serial=%d proc=%d status=%d", - msg->header.serial, msg->header.proc, msg->header.status); - virNetError(VIR_ERR_RPC, - _("Unexpected stream control message serial=%d proc=%d status=%d"), - msg->header.serial, msg->header.proc, msg->header.status); - goto error; + VIR_INFO("Ignoring unexpected stream data serial=%d proc=%d status=%d", + msg->header.serial, msg->header.proc, msg->header.status); + /* Send a dummy reply to free up 'msg' & unblock client rx */ + memset(msg, 0, sizeof(*msg)); + msg->header.type = VIR_NET_REPLY; + if (virNetServerClientSendMessage(client, msg) < 0) { + ret = -1; + goto cleanup; } ret = 0; break; -- 1.7.6

2 1

[libvirt] [PATCH 0/4 v3] macvtap: Support for sending port profile message for a SRIOV VF to its PF
by Roopa Prabhu 17 Aug '11

17 Aug '11

This patch tries to fix getPhysFn in macvtap.c to get the physical function(PF) of the direct attach interface, if the interface is a SR-IOV VF. It moves some of the sriov pci device handling code from node_device_driver to src/util/pci.[ch]. This patch series implements the following 01/4 - pci: Move some pci sriov helper code out of node device driver to util/pci 02/4 - pci: Add helper functions for sriov devices 03/4 - interface: Add functions to get sriov PF/VF relationship of a net interface 04/4 - macvtap: Fix getPhysfn to get the PF of a direct attach network interface Changelog: ---------- v1: RFC patch introduced new functions to get PF/VF relationship of SRIOV net devices by comparing PF/VF sysfs device links. Also mentioned the existance of some related code in node_device_driver. The feedback was to move and use the node_device_driver code v2: Moves sriov get_physical_function and get_virtual_functions from node_device_driver to src/util/pci*. And reworks the rest of the patches around the new code. v3: Incorporated Stefan Berger's comments. Fixed indentation errors shown by 'make syntax-check'. Used pciReportError in src/util/pci Signed-off-by: Roopa Prabhu <roprabhu(a)cisco.com> Signed-off-by: Christian Benvenuti <benve(a)cisco.com> Signed-off-by: David Wang <dwang2(a)cisco.com>

2 10