- Devel - libvirt List Archives

[libvirt] [PATCH V3] Add proxy FS support to libvirt
by M. Mohan Kumar 07 Sep '12

07 Sep '12

From: "M. Mohan Kumar" <mohan(a)in.ibm.com> A new FS driver type 'proxy' is added to QEMU 9p server. This patch adds support for using proxy FS driver from libvirt. QEMU proxy FS driver uses socket for communicating between helper and qemu proxy FS driver. Proxy helper (a stand alone binary part of qemu) is invoked with one of the descriptors created using socketpair call and the share path. Similarly QEMU is invoked with another descriptor created using the same socketpair system call and with other required FS driver parameters. Need for proxy FS driver ======================== Pass through security model in QEMU 9p server has following issues: 1) TOCTTOU vulnerability: Following symbolic links in the server could provide access to files beyond 9p export path. 2) Running QEMU with root privilege could be a security issue (pass through security model needs root privilege). Proxy FS driver is implemented to solve these issues. Proxy FS uses chroot + socket combination for securing the vulnerability known with following symbolic links. Intention of adding a new filesystem type is to allow qemu to run in non-root mode, but doing privileged operations in a chroot environment using socket IO. Proxy helper is invoked with root privileges and chroots into 9p export path. QEMU proxy fs driver sends filesystem request to proxy helper and receives the response from it. Proxy helper is designed such a way that it needs only few capabilities related to filesystem operations (such as CAP_DAC_OVERRIDE, CAP_FOWNER, etc) and all other capabilities are dropped (CAP_SYS_CHROOT, etc) Proxy patches http://permalink.gmane.org/gmane.comp.emulators.qemu/128735 Signed-off-by: M. Mohan Kumar <mohan(a)in.ibm.com> --- Changes from previous version V3 * Addressed Daniel's review comments Changes from previous version V2 * Rebased on top of current libvirt git level Changes from previous version * Remove the xml node for specifying the virtfs-proxy-helper path, now it is determined from qemu binary path. docs/formatdomain.html.in | 2 + src/conf/domain_conf.c | 3 +- src/conf/domain_conf.h | 2 +- src/qemu/qemu_capabilities.c | 5 ++- src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 83 ++++++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_command.h | 3 +- tests/qemuhelptest.c | 7 +++- 8 files changed, 96 insertions(+), 10 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 2c5c456..bf74ed8 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -1621,6 +1621,8 @@ while the value <code>immediate</code> means that a host writeback is immediately triggered for all pages touched during a guest file write operation <span class="since">(since 0.9.10)</span>. + or <code>type='proxy'</code> <span class="since">(since + 1.0)</span>. </dd> <dt><code>type='template'</code></dt> <dd> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 851284a..79ca2b7 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -268,7 +268,8 @@ VIR_ENUM_IMPL(virDomainFS, VIR_DOMAIN_FS_TYPE_LAST, VIR_ENUM_IMPL(virDomainFSDriverType, VIR_DOMAIN_FS_DRIVER_TYPE_LAST, "default", "path", - "handle") + "handle", + "proxy") VIR_ENUM_IMPL(virDomainFSAccessMode, VIR_DOMAIN_FS_ACCESSMODE_LAST, "passthrough", diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index fd0e89e..a296f49 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -660,6 +660,7 @@ enum virDomainFSDriverType { VIR_DOMAIN_FS_DRIVER_TYPE_DEFAULT = 0, VIR_DOMAIN_FS_DRIVER_TYPE_PATH, VIR_DOMAIN_FS_DRIVER_TYPE_HANDLE, + VIR_DOMAIN_FS_DRIVER_TYPE_PROXY, VIR_DOMAIN_FS_DRIVER_TYPE_LAST }; @@ -698,7 +699,6 @@ struct _virDomainFSDef { unsigned long long space_soft_limit; /* in bytes */ }; - /* 5 different types of networking config */ enum virDomainNetType { VIR_DOMAIN_NET_TYPE_USER, diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 5472267..88b7e87 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -144,7 +144,6 @@ VIR_ENUM_IMPL(qemuCaps, QEMU_CAPS_LAST, "ich9-ahci", "no-acpi", "fsdev-readonly", - "virtio-blk-pci.scsi", /* 80 */ "blk-sg-io", "drive-copy-on-read", @@ -172,7 +171,7 @@ VIR_ENUM_IMPL(qemuCaps, QEMU_CAPS_LAST, "bridge", /* 100 */ "lsi", "virtio-scsi-pci", - + "fsdev-proxy", ); struct qemu_feature_flags { @@ -1133,6 +1132,8 @@ qemuCapsComputeCmdFlags(const char *help, qemuCapsSet(flags, QEMU_CAPS_FSDEV_READONLY); if (strstr(fsdev, "writeout")) qemuCapsSet(flags, QEMU_CAPS_FSDEV_WRITEOUT); + if (strstr(fsdev, "sock_fd")) + qemuCapsSet(flags, QEMU_CAPS_FSDEV_PROXY); } if (strstr(help, "-smbios type")) qemuCapsSet(flags, QEMU_CAPS_SMBIOS_TYPE); diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index d606890..073a1cd 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -138,6 +138,7 @@ enum qemuCapsFlags { QEMU_CAPS_NETDEV_BRIDGE = 100, /* bridge helper support */ QEMU_CAPS_SCSI_LSI = 101, /* -device lsi */ QEMU_CAPS_VIRTIO_SCSI_PCI = 102, /* -device virtio-scsi-pci */ + QEMU_CAPS_FSDEV_PROXY = 103, /* -fsdev proxy supported */ QEMU_CAPS_LAST, /* this must always be the last item */ }; diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 4ca3047..3f78635 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -46,6 +46,7 @@ #include <sys/utsname.h> #include <sys/stat.h> #include <fcntl.h> +#include <libgen.h> #define VIR_FROM_THIS VIR_FROM_QEMU @@ -116,7 +117,8 @@ VIR_ENUM_DECL(qemuDomainFSDriver) VIR_ENUM_IMPL(qemuDomainFSDriver, VIR_DOMAIN_FS_DRIVER_TYPE_LAST, "local", "local", - "handle"); + "handle", + "proxy"); static void @@ -2632,9 +2634,59 @@ error: return NULL; } +/* + * Invokes the Proxy Helper with one of the socketpair as its parameter + * + */ +static int qemuInvokeProxyHelper(const char *emulator, int sock, const char *path) +{ +#define HELPER "virtfs-proxy-helper" + int ret_val, status; + virCommandPtr cmd; + char *helper, *dname, *dp; + + dp = strdup(emulator); + if (!dp) { + virReportOOMError(); + return -1; + } + dname = strrchr(dp, '/'); + if (!dname) { + VIR_FREE(dp); + VIR_FORCE_CLOSE(sock); + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Unable to get path details from %s"), dp); + return -1; + } + *dname = '\0'; + if (virAsprintf(&helper, "%s/%s", dp, HELPER) < 0) { + VIR_FREE(dp); + VIR_FORCE_CLOSE(sock); + virReportOOMError(); + return -1; + } + + cmd = virCommandNewArgList(helper, NULL); + virCommandAddArg(cmd, "-f"); + virCommandAddArgFormat(cmd, "%d", sock); + virCommandAddArg(cmd, "-p"); + virCommandAddArgFormat(cmd, "%s", path); + virCommandTransferFD(cmd, sock); + virCommandDaemonize(cmd); + ret_val = virCommandRun(cmd, &status); + if (ret_val < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("%s can't execute"), helper); + VIR_FORCE_CLOSE(sock); + } + virCommandFree(cmd); + VIR_FREE(helper); + VIR_FREE(dp); + return ret_val; +} char *qemuBuildFSStr(virDomainFSDefPtr fs, - virBitmapPtr qemuCaps ATTRIBUTE_UNUSED) + virBitmapPtr qemuCaps ATTRIBUTE_UNUSED, int qemuSocket) { virBuffer opt = VIR_BUFFER_INITIALIZER; const char *driver = qemuDomainFSDriverTypeToString(fs->fsdriver); @@ -2683,6 +2735,10 @@ char *qemuBuildFSStr(virDomainFSDefPtr fs, } virBufferAsprintf(&opt, ",id=%s%s", QEMU_FSDEV_HOST_PREFIX, fs->info.alias); + + if (fs->fsdriver == VIR_DOMAIN_FS_DRIVER_TYPE_PROXY) + virBufferAsprintf(&opt, ",sock_fd=%d", qemuSocket); + virBufferAsprintf(&opt, ",path=%s", fs->src); if (fs->readonly) { @@ -5153,10 +5209,31 @@ qemuBuildCommandLine(virConnectPtr conn, if (qemuCapsGet(qemuCaps, QEMU_CAPS_FSDEV)) { for (i = 0 ; i < def->nfss ; i++) { char *optstr; + int sockets[2] = {-1, -1}; virDomainFSDefPtr fs = def->fss[i]; + /* + * If its a proxy FS, we need to create a socket pair + * and invoke proxy_helper + */ + if (fs->fsdriver == VIR_DOMAIN_FS_DRIVER_TYPE_PROXY) { + if (qemuCapsGet(qemuCaps, QEMU_CAPS_FSDEV_PROXY) < 0) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("proxy helper not supported")); + goto error; + } + /* create a socket pair */ + if (socketpair(PF_UNIX, SOCK_STREAM, 0, sockets) < 0) { + virReportSystemError(errno, _("socketpair %s"), "failed"); + goto error; + } + virCommandTransferFD(cmd, sockets[1]); + if (qemuInvokeProxyHelper(def->emulator, sockets[0], + fs->src) < 0) + goto error; + } virCommandAddArg(cmd, "-fsdev"); - if (!(optstr = qemuBuildFSStr(fs, qemuCaps))) + if (!(optstr = qemuBuildFSStr(fs, qemuCaps, sockets[1]))) goto error; virCommandAddArg(cmd, optstr); VIR_FREE(optstr); diff --git a/src/qemu/qemu_command.h b/src/qemu/qemu_command.h index e999bc7..2da5ad0 100644 --- a/src/qemu/qemu_command.h +++ b/src/qemu/qemu_command.h @@ -89,7 +89,8 @@ char *qemuBuildDriveStr(virConnectPtr conn, bool bootable, virBitmapPtr qemuCaps); char *qemuBuildFSStr(virDomainFSDefPtr fs, - virBitmapPtr qemuCaps); + virBitmapPtr qemuCaps, + int qemuSocket); /* Current, best practice */ char * qemuBuildDriveDevStr(virDomainDefPtr def, diff --git a/tests/qemuhelptest.c b/tests/qemuhelptest.c index 2d15c94..6af2e68 100644 --- a/tests/qemuhelptest.c +++ b/tests/qemuhelptest.c @@ -679,7 +679,9 @@ mymain(void) QEMU_CAPS_SCSI_BLOCK, QEMU_CAPS_SCSI_CD, QEMU_CAPS_IDE_CD, - QEMU_CAPS_SCSI_LSI); + QEMU_CAPS_SCSI_LSI, + QEMU_CAPS_FSDEV_PROXY); + DO_TEST("qemu-1.1.0", 1001000, 0, 0, QEMU_CAPS_VNC_COLON, QEMU_CAPS_NO_REBOOT, @@ -759,7 +761,8 @@ mymain(void) QEMU_CAPS_NEC_USB_XHCI, QEMU_CAPS_NETDEV_BRIDGE, QEMU_CAPS_SCSI_LSI, - QEMU_CAPS_VIRTIO_SCSI_PCI); + QEMU_CAPS_VIRTIO_SCSI_PCI, + QEMU_CAPS_FSDEV_PROXY); return ret == 0 ? EXIT_SUCCESS : EXIT_FAILURE; } -- 1.7.7.6

2 3

[libvirt] [PATCH] events: Fix domain event race on client disconnect
by Christophe Fergeau 07 Sep '12

07 Sep '12

GNOME Boxes sometimes stops getting domain events from libvirtd, even after restarting it. Further investigation in libvirtd shows that events are properly queued with virDomainEventStateQueue, but the timer virDomainEventTimer which flushes the events and sends them to the clients never gets called. Looking at the event queue in gdb shows that it's non-empty and that its size increases with each new events. virDomainEventTimer is set up in virDomainEventStateRegister[ID] when going from 0 client connecte to 1 client connected, but is initially disabled. The timer is removed in virDomainEventStateRegister[ID] when the last client is disconnected (going from 1 client connected to 0). This timer (which handles sending the events to the clients) is enabled in virDomainEventStateQueue when queueing an event on an empty queue (queue containing 0 events). It's disabled in virDomainEventStateFlush after flushing the queue (ie removing all the elements from it). This way, no extra work is done when the queue is empty, and when the next event comes up, the timer will get reenabled because the queue will go from 0 event to 1 event, which triggers enabling the timer. However, with this Boxes bug, we have a client connected (Boxes), a non-empty queue (there are events waiting to be sent), but a disabled timer, so something went wrong. When Boxes connects (it's the only client connecting to the libvirtd instance I used for debugging), the event timer is not set as expected (state->timer == -1 when virDomainEventStateRegisterID is called), but at the same time the event queue is not empty. In other words, we had no clients connected, but pending events. This also explains why the timer never gets enabled as this is only done when an event is queued on an empty queue. I think this can happen if an event gets queued using virDomainEventStateQueue and the client disconnection happens before the event timer virDomainEventTimer gets a chance to run and flush the event. In this situation, virDomainEventStateDeregister[ID] will get called with a non-empty event queue, the timer will be destroyed if this was the only client connected. Then, when other clients connect at a later time, they will never get notified about domain events as the event timer will never get enabled because the timer is only enabled if the event queue is empty when virDomainEventStateRegister[ID] gets called, which will is no longer the case. To avoid this issue, this commit makes sure to remove all events from the event queue when the last client in unregistered. As there is no longer anyone interested in receiving these events, these events are stale so there is no need to keep them around. A client connecting later will have no interest in getting events that happened before it got connected. --- src/conf/domain_event.c | 24 +++++++++++++++++++++--- 1 file changed, 21 insertions(+), 3 deletions(-) diff --git a/src/conf/domain_event.c b/src/conf/domain_event.c index 43ecdcf..7ab973b 100644 --- a/src/conf/domain_event.c +++ b/src/conf/domain_event.c @@ -525,13 +525,13 @@ void virDomainEventFree(virDomainEventPtr event) } /** - * virDomainEventQueueFree: + * virDomainEventQueueClear: * @queue: pointer to the queue * - * Free the memory in the queue. We process this like a list here + * Removes all elements from the queue */ static void -virDomainEventQueueFree(virDomainEventQueuePtr queue) +virDomainEventQueueClear(virDomainEventQueuePtr queue) { int i; if (!queue) @@ -541,6 +541,22 @@ virDomainEventQueueFree(virDomainEventQueuePtr queue) virDomainEventFree(queue->events[i]); } VIR_FREE(queue->events); + queue->count = 0; +} + +/** + * virDomainEventQueueFree: + * @queue: pointer to the queue + * + * Free the memory in the queue. We process this like a list here + */ +static void +virDomainEventQueueFree(virDomainEventQueuePtr queue) +{ + if (!queue) + return; + + virDomainEventQueueClear(queue); VIR_FREE(queue); } @@ -1559,6 +1575,7 @@ virDomainEventStateDeregister(virConnectPtr conn, state->timer != -1) { virEventRemoveTimeout(state->timer); state->timer = -1; + virDomainEventQueueClear(state->queue); } virDomainEventStateUnlock(state); @@ -1596,6 +1613,7 @@ virDomainEventStateDeregisterID(virConnectPtr conn, state->timer != -1) { virEventRemoveTimeout(state->timer); state->timer = -1; + virDomainEventQueueClear(state->queue); } virDomainEventStateUnlock(state); -- 1.7.11.4

3 5

[libvirt] [RFC PATCH v1 0/2] Qemu/Gluster support in Libvirt
by Harsh Prateek Bora 07 Sep '12

07 Sep '12

This patchset provides support for Gluster protocol based network disks. It is based on the proposed gluster support in Qemu on qemu-devel: http://lists.gnu.org/archive/html/qemu-devel/2012-08/msg01539.html TODO: - Add support for IPv6 format based server addr - Support for transport types other than socket. Harsh Prateek Bora (2): Qemu/Gluster: Add Gluster protocol as supported network disk formats. tests: Add tests for gluster protocol based network disks support docs/schemas/domaincommon.rng | 8 ++ src/conf/domain_conf.c | 14 ++- src/conf/domain_conf.h | 3 +- src/qemu/qemu_command.c | 123 +++++++++++++++++++++ tests/qemuargv2xmltest.c | 1 + .../qemuxml2argv-disk-drive-network-gluster.args | 1 + .../qemuxml2argv-disk-drive-network-gluster.xml | 33 ++++++ tests/qemuxml2argvtest.c | 2 + 8 files changed, 182 insertions(+), 3 deletions(-) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-network-gluster.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-disk-drive-network-gluster.xml -- 1.7.11.2

9 17

[libvirt] [PATCH V4] implement offline migration
by liguang 07 Sep '12

07 Sep '12

allow migration even domain isn't active by inserting some stubs to tunnel migration path. Signed-off-by: liguang <lig.fnst(a)cn.fujitsu.com> --- src/qemu/qemu_driver.c | 2 +- src/qemu/qemu_migration.c | 181 +++++++++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_migration.h | 3 +- 3 files changed, 178 insertions(+), 8 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index d74bf52..00ca211 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -9779,7 +9779,7 @@ qemuDomainMigratePrepareTunnel3(virConnectPtr dconn, virCheckFlags(QEMU_MIGRATION_FLAGS, -1); - if (!dom_xml) { + if (!dom_xml && !(flags & VIR_MIGRATE_OFFLINE)) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("no domain XML passed")); goto cleanup; diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index 1b21ef6..991bcc5 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -70,6 +70,7 @@ enum qemuMigrationCookieFlags { QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS, QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE, QEMU_MIGRATION_COOKIE_FLAG_PERSISTENT, + QEMU_MIGRATION_COOKIE_FLAG_OFFLINE, QEMU_MIGRATION_COOKIE_FLAG_LAST }; @@ -77,12 +78,13 @@ enum qemuMigrationCookieFlags { VIR_ENUM_DECL(qemuMigrationCookieFlag); VIR_ENUM_IMPL(qemuMigrationCookieFlag, QEMU_MIGRATION_COOKIE_FLAG_LAST, - "graphics", "lockstate", "persistent"); + "graphics", "lockstate", "persistent", "offline"); enum qemuMigrationCookieFeatures { QEMU_MIGRATION_COOKIE_GRAPHICS = (1 << QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS), QEMU_MIGRATION_COOKIE_LOCKSTATE = (1 << QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE), QEMU_MIGRATION_COOKIE_PERSISTENT = (1 << QEMU_MIGRATION_COOKIE_FLAG_PERSISTENT), + QEMU_MIGRATION_COOKIE_OFFLINE = (1 << QEMU_MIGRATION_COOKIE_FLAG_OFFLINE), }; typedef struct _qemuMigrationCookieGraphics qemuMigrationCookieGraphics; @@ -101,6 +103,10 @@ struct _qemuMigrationCookie { unsigned int flags; unsigned int flagsMandatory; + /*offline migration flag*/ + int offline; + char *mig_file; + /* Host properties */ unsigned char localHostuuid[VIR_UUID_BUFLEN]; unsigned char remoteHostuuid[VIR_UUID_BUFLEN]; @@ -139,6 +145,8 @@ static void qemuMigrationCookieFree(qemuMigrationCookiePtr mig) if (mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) qemuMigrationCookieGraphicsFree(mig->graphics); + if (mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) + VIR_FREE(mig->mig_file); VIR_FREE(mig->localHostname); VIR_FREE(mig->remoteHostname); @@ -439,6 +447,12 @@ qemuMigrationCookieXMLFormat(struct qemud_driver *driver, virBufferAdjustIndent(buf, -2); } + if (mig->flags & QEMU_MIGRATION_COOKIE_OFFLINE) { + virBufferAsprintf(buf, " <offline mig_ol='%d' mig_file='% s'>\n", + mig->offline, mig->mig_file); + virBufferAddLit(buf, " </offline>\n"); + } + virBufferAddLit(buf, "</qemu-migration>\n"); return 0; } @@ -662,6 +676,18 @@ qemuMigrationCookieXMLParse(qemuMigrationCookiePtr mig, VIR_FREE(nodes); } + if ((flags & QEMU_MIGRATION_COOKIE_OFFLINE) && + virXPathBoolean("count(./offline) > 0", ctxt)) { + if (virXPathInt("string(./offline/@mig_ol)", ctxt, &mig->offline) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("missing mig_ol attribute in migration data")); + goto error; + } + mig->mig_file = virXPathString("string(./offline/@mig_file)", ctxt); + if (mig->mig_file && STREQ(mig->mig_file, "")) + VIR_FREE(mig->mig_file); + } + return 0; error: @@ -721,6 +747,12 @@ qemuMigrationBakeCookie(qemuMigrationCookiePtr mig, qemuMigrationCookieAddPersistent(mig, dom) < 0) return -1; + if (flags & QEMU_MIGRATION_COOKIE_OFFLINE) { + mig->flags |= QEMU_MIGRATION_COOKIE_OFFLINE; + mig->offline = 1; + } + + if (!(*cookieout = qemuMigrationCookieXMLFormatStr(driver, mig))) return -1; @@ -1307,6 +1339,27 @@ qemuMigrationPrepareAny(struct qemud_driver *driver, /* Domain starts inactive, even if the domain XML had an id field. */ vm->def->id = -1; + if (tunnel) { + if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen, + QEMU_MIGRATION_COOKIE_OFFLINE))) + return ret; + else if (mig->offline) { + char *file, *str, *tmp = NULL; + ret = 0; + for (str = mig->mig_file; ; str = NULL) { + file = strtok_r(str, " ", &tmp); + if (file == NULL) + break; + if (virFDStreamCreateFile(st, file, 0, 0, O_WRONLY, 0) < 0) { + virReportSystemError(errno, "%s", + _("cannot setup stream for tunnelled migration\n")); + ret = -1; + } + } + goto endjob; + } + } + if (tunnel && (pipe(dataFD) < 0 || virSetCloseExec(dataFD[1]) < 0)) { virReportSystemError(errno, "%s", @@ -2303,6 +2356,117 @@ cleanup: return ret; } +static int +doReadFile(virStreamPtr st ATTRIBUTE_UNUSED, + char *buf, size_t nbytes, void *opaque) +{ + int *fd = opaque; + + return saferead(*fd, buf, nbytes); +} + + +/* + * do offline migration + */ +static int doMigrateOffline(struct qemud_driver *driver, + virConnectPtr dconn, + virDomainObjPtr vm, + const char *cookiein, + int cookieinlen, + char **cookieout, + int *cookieoutlen, + unsigned long flags, + const char *dom_xml, + const char *dname, + virStreamPtr st, + unsigned long resource) +{ + xmlDocPtr xml = NULL; + xmlXPathContextPtr ctxt = NULL; + xmlNodePtr *disks = NULL; + qemuMigrationCookiePtr mig = NULL; + virBuffer buf = VIR_BUFFER_INITIALIZER; + int i = 0, cn = 0, fd = -1, ret = -1; + char *src[] = {NULL}, *files = NULL; + + VIR_DEBUG("driver=%p, vm=%p, st=%p, cookiein=%s, cookieinlen=%d, " + "cookieout=%p, cookieoutlen=%p, dom_xml=%s," + "dname=%s, flags=%lx, resource=%lu", + driver, vm, st, NULLSTR(cookiein), cookieinlen, + cookieout, cookieoutlen, dom_xml, dname, flags, resource); + + xml = virXMLParseStringCtxt(dom_xml, _("(domain_definition)"), &ctxt); + if (!xml) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("can't parse dom_xml for offline migration \n")); + goto cleanup; + } + cn = virXPathNodeSet("./devices/disk", ctxt, &disks); + if (cn < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Fail to get disk node\n")); + goto cleanup; + } + cn = 1; + + for (i = 0 ; i < cn ; i++) { + ctxt->node = disks[i]; + src[i] = virXPathString("string(./source/@file" + "|./source/@dir" + "|./source/@name)", ctxt); + virBufferAsprintf(&buf, "%s ", src[i]); + } + + files = virBufferContentAndReset(&buf); + + if (!(mig = qemuMigrationEatCookie(driver, vm, NULL, 0, 0))) + goto cleanup; + + mig->mig_file = files; + + if (qemuMigrationBakeCookie(mig, driver, vm, + cookieout, cookieoutlen, + QEMU_MIGRATION_COOKIE_OFFLINE) < 0) + goto cleanup; + + cookiein = *cookieout; + cookieinlen = *cookieoutlen; + cookieout = NULL; + cookieoutlen = 0; + + qemuDomainObjEnterRemoteWithDriver(driver, vm); + ret = dconn->driver->domainMigratePrepareTunnel3 + (dconn, st, cookiein, cookieinlen, + cookieout, cookieoutlen, + flags, dname, resource, dom_xml); + qemuDomainObjExitRemoteWithDriver(driver, vm); + if (ret == -1) + goto cleanup; + + for (i = 0; i < cn; i++) { + if ((fd = open(src[i], O_RDONLY)) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s %s", + _("Fail to open file for offline migration \n"), src[i]); + goto cleanup; + } + if (virStreamSendAll(st, doReadFile, &fd) < 0) + goto cleanup; + if (VIR_CLOSE(fd) < 0) + goto cleanup; + } + + if (virStreamFinish(st) < 0) + goto cleanup; + + ret = 0; + +cleanup: + VIR_FREE(files); + xmlXPathFreeContext(ctxt); + xmlFreeDoc(xml); + return ret; +} /* This is essentially a re-impl of virDomainMigrateVersion3 * from libvirt.c, but running in source libvirtd context, @@ -2357,7 +2521,13 @@ static int doPeer2PeerMigrate3(struct qemud_driver *driver, if (flags & VIR_MIGRATE_TUNNELLED) { if (!(st = virStreamNew(dconn, 0))) goto cleanup; - + if (flags & VIR_MIGRATE_OFFLINE) { + if (doMigrateOffline(driver, dconn, vm, cookiein, cookieinlen, + &cookieout, &cookieoutlen, flags, dom_xml, + dname, st, resource) != -1) + ret = 0; + goto cleanup; + } qemuDomainObjEnterRemoteWithDriver(driver, vm); ret = dconn->driver->domainMigratePrepareTunnel3 (dconn, st, cookiein, cookieinlen, @@ -2477,7 +2647,7 @@ finish: vm->def->name); cleanup: - if (ddomain) { + if (ddomain || (flags & VIR_MIGRATE_OFFLINE)) { virObjectUnref(ddomain); ret = 0; } else { @@ -2557,7 +2727,7 @@ static int doPeer2PeerMigrate(struct qemud_driver *driver, if (!virDomainObjIsActive(vm)) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("guest unexpectedly quit")); - goto cleanup; + flags |= VIR_MIGRATE_OFFLINE; } /* Change protection is only required on the source side (us), and @@ -2573,7 +2743,6 @@ static int doPeer2PeerMigrate(struct qemud_driver *driver, else ret = doPeer2PeerMigrate2(driver, sconn, dconn, vm, dconnuri, flags, dname, resource); - cleanup: orig_err = virSaveLastError(); qemuDomainObjEnterRemoteWithDriver(driver, vm); @@ -2620,7 +2789,7 @@ qemuMigrationPerformJob(struct qemud_driver *driver, if (!virDomainObjIsActive(vm)) { virReportError(VIR_ERR_OPERATION_INVALID, "%s", _("domain is not running")); - goto endjob; + flags |= VIR_MIGRATE_OFFLINE; } if (!qemuMigrationIsAllowed(driver, vm, NULL)) diff --git a/src/qemu/qemu_migration.h b/src/qemu/qemu_migration.h index 1740204..2bcaea0 100644 --- a/src/qemu/qemu_migration.h +++ b/src/qemu/qemu_migration.h @@ -36,7 +36,8 @@ VIR_MIGRATE_NON_SHARED_DISK | \ VIR_MIGRATE_NON_SHARED_INC | \ VIR_MIGRATE_CHANGE_PROTECTION | \ - VIR_MIGRATE_UNSAFE) + VIR_MIGRATE_UNSAFE | \ + VIR_MIGRATE_OFFLINE) enum qemuMigrationJobPhase { QEMU_MIGRATION_PHASE_NONE = 0, -- 1.7.2.5

5 10

[libvirt] [libvirt-snmp][ANNOUNCE] libvirt-snmp-0.0.3
by Michal Privoznik 07 Sep '12

07 Sep '12

I am pleased to announce that a new release of the libvirt-snmp package, version 0.0.3, is now available from ftp://libvirt.org/libvirt/snmp/ The new package contains several bug fixes and enhancements. Thanks to all the people involved in contributing to this release. Michal

1 0

Re: [libvirt] Memory free in libvirt JNA
by Benjamin Wang (gendwang) 07 Sep '12

07 Sep '12

Hi, Overview Part of JNA API describes as following: 1. Description1: If the native method returns char* and actually allocates memory, a return type of Pointer should be used to avoid leaking the memory. It is then up to you to take the necessary steps to free the allocated memory. 2. Description2: Declare the method as returning a Structure of the appropriate type, then invoke Structure.toArray(int) to convert to an array of initialized structures of the appropriate size. Note that your Structure class must have a no-args constructor, and you are responsible for freeing the returned memory if applicable in whatever way is appropriate for the called function. And the example code shows as following: // Original C code struct Display* get_displays(int* pcount); void free_displays(struct Display* displays); // Equivalent JNA mapping Display get_displays(IntByReference pcount); void free_displays(Display[] displays); ... IntByReference pcount = new IntByReference(); Display d = lib.get_displays(pcount); Display[] displays = (Display[])d.toArray(pcount.getValue()); ... lib.free_displays(displays); That's to say. All the memory allocated by native code must be freed explicitly in JNA part. We must add some free memory methods to support the memory-freeing. Any comments? B.R. Benjamin Wang -----Original Message----- From: Daniel Veillard [mailto:veillard@redhat.com] Sent: 2012年8月20日 14:25 To: Benjamin Wang (gendwang) Cc: stoty(a)tvnet.hu; Daniel.Schwager(a)dtnet.de Subject: Re: Memory free in libvirt JNA On Mon, Aug 20, 2012 at 05:15:45AM +0000, Benjamin Wang (gendwang) wrote: > Hi Veillard, > Thanks for your reply. I checked the current Libvirt-JNA > implementation. I find that a method named "free" defined in Domain > class which is used to free the domain object. If this is mandatory, that's to say, we should a lot of methods into the current Libvirt-jna implementation to free the memory which is allocated by libvirt API. Please correct me! As far as I understat free() is aliased as finalize() on that object so the java runtime will call free() automatically on garbage collection. I'm not a java expert, check some Java litterature for more details about how this is done and the cases where free() might be better called directly. 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/

1 0

[libvirt] [PATCH] Fix location of SELinux mount during RPM builds
by Daniel P. Berrange 07 Sep '12

07 Sep '12

From: "Daniel P. Berrange" <berrange(a)redhat.com> When building RPMs the host kernel cannot be assumed to match the target OS kernel. Thus auto-detecting /selinux vs /sys/fs/selinux based on the host kernel can result in the wrong choice (eg F18 builds on a RHEL6 host kernel) --- libvirt.spec.in | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/libvirt.spec.in b/libvirt.spec.in index 044b00f..fb35934 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1222,6 +1222,15 @@ of recent versions of Linux (and other OSes). %if 0%{?enable_autotools} autoreconf -if %endif + +%if %{with_selinux} +%if %{?fedora} >= 17 || %{?rhel} >= 7 +%define with_selinux_mount --with-selinux-mount="/sys/fs/cgroup" +%else +%define with_selinux_mount --with-selinux-mount="/selinux" +%endif +%endif + %configure %{?_without_xen} \ %{?_without_qemu} \ %{?_without_openvz} \ @@ -1254,6 +1263,7 @@ autoreconf -if %{?_without_capng} \ %{?_without_netcf} \ %{?_without_selinux} \ + %{?_with_selinux_mount} \ %{?_without_hal} \ %{?_without_udev} \ %{?_without_yajl} \ -- 1.7.11.2

3 2

[libvirt] [PATCHv3 0/4] Add range checking for scheduler tunables
by Peter Krempa 07 Sep '12

07 Sep '12

This version improves virsh so that it doesn't update all tunables but only those that changed (and now correctly). This series contains a new patch that adds a helper to deal with typed parameters. Regarding typed parameters: Shouldn't we make the helpers to deal with them public? (Patches 1 and 2 didn't change to v2) Peter Krempa (4): qemu: clean up qemuSetSchedulerParametersFlags() qemu: Add range checking for scheduler tunables when changed by API util: Add helper to assign typed params from string virsh: Update only changed scheduler tunables src/libvirt_private.syms | 1 + src/qemu/qemu_driver.c | 96 ++++++++++++++++----------- src/util/virtypedparam.c | 97 ++++++++++++++++++++++++++++ src/util/virtypedparam.h | 6 ++ tools/virsh-domain.c | 164 +++++++++++++++++++++-------------------------- 5 files changed, 236 insertions(+), 128 deletions(-) -- 1.7.12

2 7

[libvirt] RFC: APIs to manage KSM
by Osier Yang 07 Sep '12

07 Sep '12

Triggered by the requirement to control the new sys knob /sys/kernel/mm/ksm/merge_nodes for a NUMA aware host by patch: http://permalink.gmane.org/gmane.linux.kernel.mm/81497 It shows there is no methods to manage the KSM in libvirt. This proposes the new APIs like following, thanks for suggestions/idea in advance. == The macros and enums for the params' fields /* virNodeKSMRunMode: * * Represents the modes of /sys/kernel/mm/ksm/run. */ typedef enum { /* Stop ksmd from running but keep merged pages */ VIR_NODE_STOP_KSM_KEEP_MERGED_PAGES = 0, /* Start ksmd */ VIR_NODE_START_KSM = 1, /* Stop ksmd and and unmerge all pages currently merged */ VIR_NODE_STOP_KSM_UNMERGE_PAGES = 2, } virNodeKSMRunMode; /* * VIR_NODE_KSM_PAGES_TO_SCAN: * * Macro for typed parameter that represents how many present pages * to scan before ksmd goes to sleep. */ # define VIR_NODE_KSM_PAGES_TO_SCAN "pages_to_scan" /* * VIR_NODE_KSM_SLEEP_MILLISECS: * * Macro for typed parameter that represents how many milliseconds * ksmd should sleep before next scan. */ # define VIR_NODE_KSM_SLEEP_MILLISECS "sleep_millisecs" /* * VIR_NODE_KSM_RUN: * * Macro for typed parameter that is used to control ksmd's state, as * an int containing a virNodeKSMRunMode value. */ # define VIR_NODE_KSM_RUN "run" /* * VIR_NODE_KSM_PAGES_SHARED: * * Macro for typed parameter that represents how many ksm shared pages * are being used. */ # define VIR_NODE_KSM_PAGES_SHARED "pages_shared" /* * VIR_NODE_KSM_PAGES_SHARING: * * Macro for typed parameter that represents how many sites are * sharing the pages i.e. how much saved. */ # define VIR_NODE_KSM_PAGES_SHARING "pages_sharing" /* VIR_NODE_KSM_PAGES_UNSHARED: * * Macro for typed parameter that represents how many pages unique * but repeatedly checked for merging. */ # define VIR_NODE_KSM_PAGES_UNSHARED "pages_unshared" /* VIR_NODE_KSM_PAGES_VOLATILE: * * Macro for typed parameter that represents how many pages changing * too fast to be placed in a tree. */ # define VIR_NODE_KSM_PAGES_VOLATILE "pages_volatile" /* VIR_NODE_KSM_FULL_SCAN: * * Macro for typed parameter that represents how many times all * mergeable areas have been scanned. */ # define VIR_NODE_KSM_FULL_SCAN "full_scan" == API to Get KSM parameters. /* * virNodeGetKSMParameters: * @conn: pointer to the hypervisor connection * @params: pointer to memory parameter object * (return value, allocated by the caller) * @nparams: pointer to number of memory parameters; input and output * @flags: extra flags; not used yet, so callers should always pass 0 * * ... Detailed comments ommitted ... * * Returns 0 in case of success, and -1 in case of failure. */ int virNodeGetKSMParameters(virConnectPtr conn, virTypedParameterPtr params, int *nparams, unsigned int flags); == API to set KSM parameters. Currently only "run", "pages_to_scan", and "sleep_millisecs" are writable. The new boolearn sys knob "merge_nodes" could be added once it's acceptted by kernel upstream. /* * virNodeSetKSMParameters: * @conn: pointer to the hypervisor connection * @params: pointer to scheduler parameter objects * @naprams: number of scheduler parameter objects * (this value can be the same or less than the returned * value nparams of virDomainGetSchedulerType) * @flags: extra flags; not used yet, so callers should always pass 0 * * ... Detailed comments ommitted ... * * Returns 0 in case of success, -1 in case of failure. */ int virNodeSetKSMParameters(virConnectPtr conn, virTypedParameterPtr params, int nparams, unsigned int flags); BTW, the other thing libvirt misses is to tell ksmtuned to retune when a domain is started or destroyed. Which is outlined in the Red Hat Virtualization Administration Guide doc, but it's not true. So either we have a hole in libvirt, or lacks a fix for the document. <...> The KSM tuning service The ksmtuned service does not have any options. The ksmtuned service loops and adjusts ksm. The ksmtuned service is notified by libvirt when a guest is created or destroyed. </...> https://access.redhat.com/knowledge/docs/en-US/Red_Hat_Enterprise_Linux/6/h… Regards, Osier

3 6

[libvirt] [PATCH] Refactor ESX storage driver and add iSCSI support
by Ata E Husain Bohra 07 Sep '12

07 Sep '12

Resent of http://www.redhat.com/archives/libvir-list/2012-August/msg01353.html. Merging recent branch updates to this patch. Ata E Husain Bohra (1): Refactor ESX storage driver and add iSCSI support src/Makefile.am | 16 +- src/esx/esx_driver.c | 4 +- src/esx/esx_storage_backend_iscsi.c | 794 +++++++++++++++++++ src/esx/esx_storage_backend_iscsi.h | 31 + src/esx/esx_storage_backend_vmfs.c | 1471 +++++++++++++++++++++++++++++++++++ src/esx/esx_storage_backend_vmfs.h | 31 + src/esx/esx_storage_driver.c | 1303 +++++-------------------------- src/esx/esx_vi.c | 341 +++++++- src/esx/esx_vi.h | 21 +- src/esx/esx_vi_generator.input | 308 +++++++- src/esx/esx_vi_generator.py | 19 + 11 files changed, 3218 insertions(+), 1121 deletions(-) create mode 100644 src/esx/esx_storage_backend_iscsi.c create mode 100644 src/esx/esx_storage_backend_iscsi.h create mode 100644 src/esx/esx_storage_backend_vmfs.c create mode 100644 src/esx/esx_storage_backend_vmfs.h -- 1.7.9.5

2 1