- Devel - libvirt List Archives

[libvirt] cont command failing via JSON monitor on restore
by Jim Fehlig 29 Mar '11

29 Mar '11

libvirt 0.8.7 qemu 0.13 I'm looking into a problem with qemu save/restore via JSON monitor. On restore, the vm is left in a paused state with following error returned for 'cont' command An incoming migration is expected before this command can be executed I was trying to debug the issue in gdb, but stepping through the code introduces enough delay between qemudStartVMDaemon() and doStartCPUs() that the latter succeeds. Any suggestions on how to determine when it is safe to call doStartCPUs() to prevent the above error? I don't see this issue with the text monitor btw. Thanks, Jim

4 11

[libvirt] [RFC] Add persistent XML for cpu tunables
by Osier Yang 29 Mar '11

29 Mar '11

Example of cputune XML: <cputune> <shares>2048</shares> <vcpupin vcpu='0' cpuset='0-4,^1'/> <vcpupin vcpu='1' cpuset='1,3'/> <vcpupin vcpu='2' cpuset='0,2'/> </cputune> "shares" is to define the the proportional weighted cpu share for the domain. "vcpupin" is to define the cpu affinities of vcpus, it will not be displayed if one doesn't specify it explicitly in XML or set the cpu affinites for vcpu via "vcpupin", means there will be no vcpupin element in domain XML by default, and the constraints are: - Error if one specify entries more than the count of maxvcpus. - Error when one specify entries for same vcpu. - Error if value of attribute "vcpu" is more than count of "maxvcpus - 1". Attribute "cpuset" works same as "cpuset" of element "vcpu", reuse the codes for parsing and formating value of "cpuset" of element "vcpu". NB, the idea to add persistent XML for "cpushares" is from "Nikunj A. Dadhania": https://www.redhat.com/archives/libvir-list/2011-January/msg01183.html I rebased it and include it together in this patch series. [PATCH 1/8] cputune: Add xml schema for cputune xml [PATCH 2/8] cputune: Add document for cputune xml [PATCH 3/8] cputune: Add data structures presenting cputune xml [PATCH 4/8] cputune: implementations of parsing and formating cputune xml [PATCH 5/8] cputune: support cputune for qemu driver [PATCH 6/8] cputune: support cputune for lxc driver [PATCH 7/8] cputune: support cputune for xend driver [PATCH 8/8] cputune: new tests for testing cputune xml

4 29

[libvirt] [PATCH] qemu: fix regression with fd labeling on migration
by Eric Blake 29 Mar '11

29 Mar '11

My earlier testing for commit 34fa0de0 was done while starting just-built libvirt from an unconfined_t shell, where the fds happened to work when transferring to qemu. But when installed and run under virtd_t, failure to label the raw file (with no compression) or the pipe (with compression) triggers SELinux failures when passing fds over SCM_RIGHTS to svirt_t qemu. * src/qemu/qemu_migration.c (qemuMigrationToFile): When passing FDs, make sure they are labeled. --- This copies the fd-labeling approach added in commit 34a19dda1. With this patch, I tested both unconfined_t and virtd_t SELinux process labels for libvirtd (shell start vs. init start). virsh managedsave dom/virsh start dom works for both raw and compressed save_image_format, both contexts virsh save dom file/virsh restore file works for raw save_image_format, both contexts works for compressed with virtd_t context save fails for compressed with unconfined_t context, with the failure looking identical to the previously-reported failure for restore in the same settings (https://bugzilla.redhat.com/show_bug.cgi?id=691499) So I'm reasonably confident that this is a good patch. src/qemu/qemu_migration.c | 8 ++++++-- 1 files changed, 6 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index 98b9d01..43741e1 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -1304,8 +1304,12 @@ qemuMigrationToFile(struct qemud_driver *driver, virDomainObjPtr vm, if (qemuCaps && qemuCapsGet(qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD) && (!compressor || pipe(pipeFD) == 0)) { /* All right! We can use fd migration, which means that qemu - * doesn't have to open() the file, so we don't have to futz - * around with granting access or revoking it later. */ + * doesn't have to open() the file, so while we still have to + * grant SELinux access, we can do it on fd and avoid cleanup + * later, as well as skip futzing with cgroup. */ + if (virSecurityManagerSetFDLabel(driver->securityManager, vm, + compressor ? pipeFD[1] : fd) < 0) + goto cleanup; is_reg = true; bypassSecurityDriver = true; } else { -- 1.7.4

2 2

[libvirt] [PATCH] qemu: improve error message on failed fd transfer
by Eric Blake 29 Mar '11

29 Mar '11

First fallout of fd: migration - it looks like SELinux enforcing _does_ require fd labeling (running uninstalled libvirtd from an unconstrained shell had no problems, but once faked out by doing chcon `stat -c %C /usr/sbin/libvirtd` daemon/libvirtd run_init $PWD/daemon/libvirtd to run it with the same context as an init script service, and with SELinux enforcing, I got a rather confusing failure: error: Failed to save domain fedora_12 to fed12.img error: internal error unable to send TAP file handle: No file descriptor supplied via SCM_RIGHTS This fixes the error message, then I need to figure out a subsequent patch that does the fsetfilecon() necessary to keep things happy. It also appears that libvirtd hangs on a failed fd transfer; I don't know if that needs an independent fix. * src/qemu/qemu_monitor_text.c (qemuMonitorTextSendFileHandle): Improve message, since TAP is no longer only client. --- src/qemu/qemu_monitor_text.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 75b2995..5b1290b 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -1679,0 +1679,0 @@ int qemuMonitorTextSendFileHandle(qemuMonitorPtr mon, if (STRNEQ(reply, "")) { qemuReportError(VIR_ERR_INTERNAL_ERROR, - _("unable to send TAP file handle: %s"), - reply); + _("unable to send file handle '%s': %s"), + fdname, reply); goto cleanup; } -- 1.7.4

2 2

[libvirt] [PATCHv2 0/8] Additional functionality for libxl driver
by Markus Groß 29 Mar '11

29 Mar '11

This series of patches adds new functionality to the libxl driver. V2: - Incorporated review from Jim Fehlig - Allow domainGetSchedulerType only on active domains - Fix warnings about virDomainObjUnref Markus Groß (8): Ignore return value of virDomainObjUnref Add event callbacks to libxl driver List authors in copyright headers Add vcpu functions to libxl driver Add domainXMLFromNative/domainXMLToNative to libxl driver Add domainGetSchedulerType to libxl driver Add domainGetOSType to libxl driver Add domainSuspend/Resume to libxl driver configure.ac | 2 + daemon/Makefile.am | 3 + src/Makefile.am | 8 +- src/libxl/libxl_conf.c | 3 + src/libxl/libxl_conf.h | 11 +- src/libxl/libxl_driver.c | 823 +++++++++++++++++++++++++++++++++++++++++++++- src/libxl/libxl_driver.h | 3 + 7 files changed, 832 insertions(+), 21 deletions(-) -- 1.7.4.1

3 19

[libvirt] [PATCH] maint: ignore new built file
by Eric Blake 29 Mar '11

29 Mar '11

* .gitignore: Exclude libvirt_iohelper. --- Pushing under the trivial rule. .gitignore | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/.gitignore b/.gitignore index e8c907e..ba4d351 100644 --- a/.gitignore +++ b/.gitignore @@ -52,6 +52,7 @@ /mkinstalldirs /po/ /proxy/ +/src/libvirt_iohelper /tests/*.log /tests/cputest /tests/nwfilterxml2xmltest -- 1.7.4

1 0

Re: [libvirt] Entering 0.9.0 freeze week, RC1 version to test
by Justin Clift 29 Mar '11

29 Mar '11

Hey all, virsh in 0.9.0 RC1 doesn't work properly. There's a hang at start up: $ virsh <waits forever> Control-C breaks out ok, dropping back to bash prompt. The pieces compiled in were: $ virsh -V Virsh command line tool of libvirt 0.9.0 See web site at http://libvirt.org/ Compiled with support for: Hypervisors: OpenVZ VirtualBox ESX Test Networking: Remote Daemon Storage: Dir SCSI Miscellaneous: Secrets Debug Readline This is definitely something new. With 0.8.8, virsh starts ok: $ virsh Welcome to virsh, the virtualization interactive terminal. Type: 'help' for help with commands 'quit' to quit virsh # ^D $ Compiled with the same options: $ virsh -V Virsh command line tool of libvirt 0.8.8 See web site at http://libvirt.org/ Compiled with support for: Hypervisors: OpenVZ VirtualBox ESX Test Networking: Remote Daemon Storage: Dir SCSI Miscellaneous: Secrets Debug Readline Haven't looked into this in more detail yet, but might get time to today/tomorrow. Kind of wondering if it's related to the win32 prob Matthias mentioned. Regards and best wishes, Justin Clift

1 1

[libvirt] [PATCH] Add missing checks for QEMU domain state in tunables APIs
by Daniel P. Berrange 29 Mar '11

29 Mar '11

The methods qemuDomain{Get,Set}{Memory,Blkio,Scheduler}Parameters all forgot to do a check on virDomainIsActive(), resulting in bogus error messages from later parts of their impl * src/qemu/qemu_driver.c: Add missing checks on virDomainIsActive() --- src/qemu/qemu_driver.c | 38 +++++++++++++++++++++++++++++++++++++- 1 files changed, 37 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index e97c1d9..f54e41c 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -4245,6 +4245,12 @@ static int qemuDomainSetBlkioParameters(virDomainPtr dom, goto cleanup; } + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto cleanup; + } + if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) { qemuReportError(VIR_ERR_INTERNAL_ERROR, _("cannot find cgroup for domain %s"), vm->def->name); @@ -4321,6 +4327,12 @@ static int qemuDomainGetBlkioParameters(virDomainPtr dom, goto cleanup; } + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto cleanup; + } + if ((*nparams) == 0) { /* Current number of blkio parameters supported by cgroups */ *nparams = QEMU_NB_BLKIO_PARAM; @@ -4405,6 +4417,12 @@ static int qemuDomainSetMemoryParameters(virDomainPtr dom, goto cleanup; } + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto cleanup; + } + if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) { qemuReportError(VIR_ERR_INTERNAL_ERROR, _("cannot find cgroup for domain %s"), vm->def->name); @@ -4508,6 +4526,12 @@ static int qemuDomainGetMemoryParameters(virDomainPtr dom, goto cleanup; } + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto cleanup; + } + if ((*nparams) == 0) { /* Current number of memory parameters supported by cgroups */ *nparams = QEMU_NB_MEM_PARAM; @@ -4621,6 +4645,12 @@ static int qemuSetSchedulerParameters(virDomainPtr dom, goto cleanup; } + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto cleanup; + } + if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) { qemuReportError(VIR_ERR_INTERNAL_ERROR, _("cannot find cgroup for domain %s"), vm->def->name); @@ -4692,6 +4722,12 @@ static int qemuGetSchedulerParameters(virDomainPtr dom, goto cleanup; } + if (!virDomainObjIsActive(vm)) { + qemuReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("domain is not running")); + goto cleanup; + } + if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) { qemuReportError(VIR_ERR_INTERNAL_ERROR, _("cannot find cgroup for domain %s"), vm->def->name); @@ -4751,7 +4787,7 @@ qemudDomainBlockStats (virDomainPtr dom, if (qemuDomainObjBeginJob(vm) < 0) goto cleanup; - if (!virDomainObjIsActive (vm)) { + if (!virDomainObjIsActive(vm)) { qemuReportError(VIR_ERR_OPERATION_INVALID, "%s", _("domain is not running")); goto endjob; -- 1.7.4

2 1

[libvirt] [PATCH 0/8] pending patch queue
by Eric Blake 29 Mar '11

29 Mar '11

All of these have been posted before (sometimes buried as replies in another thread), but I'm reposting to make it easier to get feedback in one place. My upcoming fd: migration v5 series was tested with all of these applied, and while I can probably rebase things to skip any or all of these patches, it would be nicer to get them in to 0.9.0. Eric Blake (8): remote: fix memory leak qemu: don't restore state label twice docs: document recent hook additions tests: don't alter state in $HOME virsh: optimize creation of default connection maint: use space, not tab, in remote_protocol-structs logging: always NUL-terminate circular buffer build: fix driver module build docs/hooks.html.in | 23 +- src/Makefile.am | 13 +- src/libvirt_private.syms | 20 + src/libvirt_xenxs.syms | 22 + src/qemu/qemu_driver.c | 1 + src/qemu/qemu_process.c | 4 +- src/remote/remote_driver.c | 120 ++-- src/remote_protocol-structs | 1489 ++++++++++++++++++++++--------------------- src/util/logging.c | 8 +- tests/Makefile.am | 30 +- tests/int-overflow | 4 +- tools/virsh.c | 423 +++++++----- 12 files changed, 1152 insertions(+), 1005 deletions(-) create mode 100644 src/libvirt_xenxs.syms -- 1.7.4

4 21

[libvirt] [PATCH] Update event loop example programs to demonstrate best practice
by Daniel P. Berrange 29 Mar '11

29 Mar '11

The example C event loop code is a nasty hack and not compliant with the require API semantics. Delete this, so that developers don't mistakenly copy it. Instead call the new public event loop APIs. Update the python event loop example, so that it can optionally use the public event APIs, as an alternative to the pure python code. The pure python event code is a good working example, so don't delete it. * examples/domain-events/events-c/event-test.c: Replace event loop code with use of public APIs * examples/domain-events/events-python/event-test.py: Allow optional use of new public event APIs --- examples/domain-events/events-c/event-test.c | 163 +------------------ examples/domain-events/events-python/event-test.py | 29 ++++- 2 files changed, 36 insertions(+), 156 deletions(-) diff --git a/examples/domain-events/events-c/event-test.c b/examples/domain-events/events-c/event-test.c index 6e3a160..ee68bb1 100644 --- a/examples/domain-events/events-c/event-test.c +++ b/examples/domain-events/events-c/event-test.c @@ -5,10 +5,8 @@ #include <string.h> #include <signal.h> -#if HAVE_SYS_POLL_H -# include <sys/types.h> -# include <sys/poll.h> # include <libvirt/libvirt.h> +# include <libvirt/virterror.h> # define VIR_DEBUG0(fmt) printf("%s:%d :: " fmt "\n", \ __func__, __LINE__) @@ -20,22 +18,6 @@ # define ATTRIBUTE_UNUSED __attribute__((__unused__)) # endif -/* handle globals */ -int h_fd = 0; -virEventHandleType h_event = 0; -virEventHandleCallback h_cb = NULL; -virFreeCallback h_ff = NULL; -void *h_opaque = NULL; - -/* timeout globals */ -# define TIMEOUT_MS 1000 -int t_active = 0; -int t_timeout = -1; -virEventTimeoutCallback t_cb = NULL; -virFreeCallback t_ff = NULL; -void *t_opaque = NULL; - - /* Prototypes */ const char *eventToString(int event); int myEventAddHandleFunc (int fd, int event, @@ -266,94 +248,6 @@ static void myFreeFunc(void *opaque) } -/* EventImpl Functions */ -int myEventHandleTypeToPollEvent(virEventHandleType events) -{ - int ret = 0; - if(events & VIR_EVENT_HANDLE_READABLE) - ret |= POLLIN; - if(events & VIR_EVENT_HANDLE_WRITABLE) - ret |= POLLOUT; - if(events & VIR_EVENT_HANDLE_ERROR) - ret |= POLLERR; - if(events & VIR_EVENT_HANDLE_HANGUP) - ret |= POLLHUP; - return ret; -} - -virEventHandleType myPollEventToEventHandleType(int events) -{ - virEventHandleType ret = 0; - if(events & POLLIN) - ret |= VIR_EVENT_HANDLE_READABLE; - if(events & POLLOUT) - ret |= VIR_EVENT_HANDLE_WRITABLE; - if(events & POLLERR) - ret |= VIR_EVENT_HANDLE_ERROR; - if(events & POLLHUP) - ret |= VIR_EVENT_HANDLE_HANGUP; - return ret; -} - -int myEventAddHandleFunc(int fd, int event, - virEventHandleCallback cb, - void *opaque, - virFreeCallback ff) -{ - VIR_DEBUG("Add handle %d %d %p %p", fd, event, cb, opaque); - h_fd = fd; - h_event = myEventHandleTypeToPollEvent(event); - h_cb = cb; - h_ff = ff; - h_opaque = opaque; - return 0; -} - -void myEventUpdateHandleFunc(int fd, int event) -{ - VIR_DEBUG("Updated Handle %d %d", fd, event); - h_event = myEventHandleTypeToPollEvent(event); - return; -} - -int myEventRemoveHandleFunc(int fd) -{ - VIR_DEBUG("Removed Handle %d", fd); - h_fd = 0; - if (h_ff) - (h_ff)(h_opaque); - return 0; -} - -int myEventAddTimeoutFunc(int timeout, - virEventTimeoutCallback cb, - void *opaque, - virFreeCallback ff) -{ - VIR_DEBUG("Adding Timeout %d %p %p", timeout, cb, opaque); - t_active = 1; - t_timeout = timeout; - t_cb = cb; - t_ff = ff; - t_opaque = opaque; - return 0; -} - -void myEventUpdateTimeoutFunc(int timer ATTRIBUTE_UNUSED, int timeout) -{ - /*VIR_DEBUG("Timeout updated %d %d", timer, timeout);*/ - t_timeout = timeout; -} - -int myEventRemoveTimeoutFunc(int timer) -{ - VIR_DEBUG("Timeout removed %d", timer); - t_active = 0; - if (t_ff) - (t_ff)(t_opaque); - return 0; -} - /* main test functions */ void usage(const char *pname) @@ -372,7 +266,6 @@ static void stop(int sig) int main(int argc, char **argv) { - int sts; int callback1ret = -1; int callback2ret = -1; int callback3ret = -1; @@ -386,17 +279,12 @@ int main(int argc, char **argv) action_stop.sa_handler = stop; - if(argc > 1 && STREQ(argv[1],"--help")) { + if (argc > 1 && STREQ(argv[1],"--help")) { usage(argv[0]); return -1; } - virEventRegisterImpl( myEventAddHandleFunc, - myEventUpdateHandleFunc, - myEventRemoveHandleFunc, - myEventAddTimeoutFunc, - myEventUpdateTimeoutFunc, - myEventRemoveTimeoutFunc); + virEventRegisterDefaultImpl(); virConnectPtr dconn = NULL; dconn = virConnectOpenReadOnly (argv[1] ? argv[1] : NULL); @@ -451,38 +339,12 @@ int main(int argc, char **argv) (callback5ret != -1) && (callback6ret != -1) && (callback7ret != -1)) { - while(run) { - struct pollfd pfd = { .fd = h_fd, - .events = h_event, - .revents = 0}; - - sts = poll(&pfd, 1, TIMEOUT_MS); - - /* if t_timeout < 0 then t_cb must not be called */ - if (t_cb && t_active && t_timeout >= 0) { - t_cb(t_timeout,t_opaque); - } - - if (sts == 0) { - /* VIR_DEBUG0("Poll timeout"); */ - continue; - } - if (sts < 0 ) { - VIR_DEBUG0("Poll failed"); - continue; - } - if ( pfd.revents & POLLHUP ) { - VIR_DEBUG0("Reset by peer"); - return -1; + while (run) { + if (virEventRunDefaultImpl() < 0) { + virErrorPtr err = virGetLastError(); + fprintf(stderr, "Failed to run event loop: %s\n", + err && err->message ? err->message : "Unknown error"); } - - if(h_cb) { - h_cb(0, - h_fd, - myPollEventToEventHandleType(pfd.revents & h_event), - h_opaque); - } - } VIR_DEBUG0("Deregistering event handlers"); @@ -496,17 +358,10 @@ int main(int argc, char **argv) } VIR_DEBUG0("Closing connection"); - if( dconn && virConnectClose(dconn)<0 ) { + if (dconn && virConnectClose(dconn) < 0) { printf("error closing\n"); } printf("done\n"); return 0; } - -#else -int main(void) { - printf("event-test program not available without sys/poll.h support\n"); - return 1; -} -#endif diff --git a/examples/domain-events/events-python/event-test.py b/examples/domain-events/events-python/event-test.py index c149ed9..df75dce 100644 --- a/examples/domain-events/events-python/event-test.py +++ b/examples/domain-events/events-python/event-test.py @@ -15,6 +15,17 @@ import errno import time import threading +# For the sake of demonstration, this example program includes +# an implementation of a pure python event loop. Most applications +# would be better off just using the default libvirt event loop +# APIs, instead of implementing this in python. The exception is +# where an application wants to integrate with an existing 3rd +# party event loop impl +# +# Change this to 'False' to make the demo use the native +# libvirt event loop impl +use_pure_python_event_loop = True + do_debug = False def debug(msg): global do_debug @@ -392,6 +403,10 @@ def virEventLoopPureRun(): global eventLoop eventLoop.run_loop() +def virEventLoopNativeRun(): + while True: + libvirt.virEventRunDefaultImpl() + # Spawn a background thread to run the event loop def virEventLoopPureStart(): global eventLoopThread @@ -400,6 +415,13 @@ def virEventLoopPureStart(): eventLoopThread.setDaemon(True) eventLoopThread.start() +def virEventLoopNativeStart(): + global eventLoopThread + libvirt.virEventRegisterDefaultImpl() + eventLoopThread = threading.Thread(target=virEventLoopNativeRun, name="libvirtEventLoop") + eventLoopThread.setDaemon(True) + eventLoopThread.start() + ########################################################################## # Everything that now follows is a simple demo of domain lifecycle events @@ -474,9 +496,12 @@ def main(): print "Using uri:" + uri # Run a background thread with the event loop - virEventLoopPureStart() + if use_pure_python_event_loop: + virEventLoopPureStart() + else: + virEventLoopNativeStart() - vc = libvirt.open(uri) + vc = libvirt.openReadOnly(uri) # Close connection on exit (to test cleanup paths) old_exitfunc = getattr(sys, 'exitfunc', None) -- 1.7.4

2 2