- Devel - libvirt List Archives

[libvirt] [PATCH 3/5] macvtap support for libvirt -- qemu support
by Stefan Berger 11 Feb '10

11 Feb '10

This part adds support for qemu making a macvtap tap device available via file descriptor passed to qemu command line. This also attempts to tear down the macvtap device when a VM terminates. This includes support for attachment and detachment to/from running VM. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

3 4

[libvirt] [PATCH 2/5] macvtap support for libvirt -- parse new interface XML
by Stefan Berger 11 Feb '10

11 Feb '10

This part adds support to domain_conf.{c|h} for parsing the new interface XML of type 'direct'. The parsed mode is now stored as an int. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

3 3

[libvirt] [PATCH 0/5] macvtap support for Qemu/KVM VMs via libvirt
by Stefan Berger 11 Feb '10

11 Feb '10

Hello! This is a re-post of previously posted patches following Daniel Berrange's request for changes along with other fixes. The following patches provide support for making the macvtap networking device type available to Qemu/KVM VMs. The patches rely on the macvtap driver that just became available through the Linux net-next tree (fixes still may be necessary) and make the device available to Qemu/KVM via a tap file descriptor similar to a 'regular' tap device. Following up on previous discussions, the libvirt patches allow using the following XML in the domain description to enable qemu network connectivity via this type of device: <interface type='direct'> <source dev='eth1' mode='vepa'/> <model type='virtio'/> </interface> The above XML indicates that eth1 is the Ethernet interface to link the macvtap device to and communicate to the network. As a consequence, libvirt will create an instance of a macvtap device, assign it the same MAC address as the VM's interface has and open a file descriptor of the associated character device /dev/tap%d and pass it via command line to Qemu/kvm. In the above XML the mode can be chosen as 'vepa', 'private' or 'bridge' and is by default set to 'vepa'(by the driver) if omitted. Attachment and detachment of macvtap to/from a running VM also works. Regards, Stefan

3 3

[libvirt] [PATCH 5/5] macvtap support for libvirt -- schema extensions
by Stefan Berger 11 Feb '10

11 Feb '10

This part extends the domain xml schema to support the new interface type 'direct'. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

2 1

[libvirt] virStoragePoolSourceFree(S) does not free S
by Jim Meyering 11 Feb '10

11 Feb '10

I was surprised to see that virStoragePoolSourceFree(S) does not free S. The other three vir*Free functions in storage_conf *do* free S. There are at least three places where this causes leaks. Here's one: virStoragePoolSourcePtr virStoragePoolDefParseSourceString(virConnectPtr conn, const char *srcSpec, int pool_type) { xmlDocPtr doc = NULL; xmlNodePtr node = NULL; xmlXPathContextPtr xpath_ctxt = NULL; virStoragePoolSourcePtr def = NULL, ret = NULL; doc = xmlReadDoc((const xmlChar *)srcSpec, "srcSpec.xml", NULL, XML_PARSE_NOENT | XML_PARSE_NONET | XML_PARSE_NOERROR | XML_PARSE_NOWARNING); if (doc == NULL) { virStorageReportError(conn, VIR_ERR_XML_ERROR, "%s", _("bad <source> spec")); goto cleanup; } xpath_ctxt = xmlXPathNewContext(doc); if (xpath_ctxt == NULL) { virReportOOMError(conn); goto cleanup; } if (VIR_ALLOC(def) < 0) { <<<====== def is alloc'd virReportOOMError(conn); goto cleanup; } node = virXPathNode(conn, "/source", xpath_ctxt); if (!node) { virStorageReportError(conn, VIR_ERR_XML_ERROR, "%s", _("root element was not source")); goto cleanup; } if (virStoragePoolDefParseSource(conn, xpath_ctxt, def, pool_type, node) < 0) goto cleanup; ret = def; def = NULL; cleanup: if (def) virStoragePoolSourceFree(def); <<<<======== yet not freed xmlFreeDoc(doc); xmlXPathFreeContext(xpath_ctxt); return ret; } One fix might be to call VIR_FREE(def) in the "if (def)..." block, but that seems short-sighted, since the name virStoragePoolSourceFree makes me think *it* should be doing the whole job. However, if I make the logical change to virStoragePoolSourceFree, diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c index 62b8394..ffe38cc 100644 --- a/src/conf/storage_conf.c +++ b/src/conf/storage_conf.c @@ -291,6 +291,7 @@ virStoragePoolSourceFree(virStoragePoolSourcePtr source) { VIR_FREE(source->auth.chap.login); VIR_FREE(source->auth.chap.passwd); } + VIR_FREE(source); } that causes "make check" to fail miserably due to heap corruption, as reported by valgrind: ==30311== Invalid free() / delete / delete[] ==30311== at 0x4A04D72: free (vg_replace_malloc.c:325) ==30311== by 0x42950C: virFree (memory.c:177) ==30311== by 0x4A2687: virStoragePoolSourceFree (storage_conf.c:294) ==30311== by 0x4A26BD: virStoragePoolDefFree (storage_conf.c:304) ==30311== by 0x4A2723: virStoragePoolObjFree (storage_conf.c:319) ==30311== by 0x4A27A2: virStoragePoolObjListFree (storage_conf.c:334) ==30311== by 0x4757AD: storageDriverShutdown (storage_driver.c:246) ==30311== by 0x4C1E3F: virStateCleanup (libvirt.c:909) ==30311== by 0x4103C5: qemudCleanup (libvirtd.c:2402) ==30311== by 0x41273B: main (libvirtd.c:3196) ==30311== Address 0x5c48c38 is 56 bytes inside a block of size 184 alloc'd ==30311== at 0x4A04481: calloc (vg_replace_malloc.c:418) ==30311== by 0x4293F5: virAlloc (memory.c:100) ==30311== by 0x4A330E: virStoragePoolDefParseXML (storage_conf.c:615) ==30311== by 0x4A3A21: virStoragePoolDefParseNode (storage_conf.c:762) ==30311== by 0x4A3BE6: virStoragePoolDefParse (storage_conf.c:810) ==30311== by 0x4A3C7F: virStoragePoolDefParseFile (storage_conf.c:834) ==30311== by 0x4A5816: virStoragePoolObjLoad (storage_conf.c:1495) ==30311== by 0x4A5BDB: virStoragePoolLoadAllConfigs (storage_conf.c:1575) ==30311== by 0x475598: storageDriverStartup (storage_driver.c:164) ==30311== by 0x4C1D9D: virStateInitialize (libvirt.c:888) ==30311== by 0x4125E9: main (libvirtd.c:3153) I tracked the problem down to this definition in src/conf/storage_conf.h: typedef struct _virStoragePoolDef virStoragePoolDef; typedef virStoragePoolDef *virStoragePoolDefPtr; struct _virStoragePoolDef { /* General metadata */ char *name; unsigned char uuid[VIR_UUID_BUFLEN]; int type; /* virStoragePoolType */ unsigned long long allocation; unsigned long long capacity; unsigned long long available; virStoragePoolSource source; <== this is a *STRUCT*, not a ptr virStoragePoolTarget target; <== Likewise }; Hence, when calling virStoragePoolDefFree (defined like this) void virStoragePoolDefFree(virStoragePoolDefPtr def) { if (!def) return; VIR_FREE(def->name); virStoragePoolSourceFree(&def->source); That function must *not* call VIR_FREE on &def->source, because as valgrind tells us, above, that is in the middle of a virStoragePoolDef buffer. Thus, the naive patch seems to be the way to go, unless you want to change the struct members to be pointers (and adjust all uses -- much more invasive). There are other leaks associated with similar uses of virStoragePoolSourceFree just like this one. Here's a patch that fixes them all at once. If you agree in principle, I'll write a longer commit log than the one below: [note that there's no need to guard uses of virStoragePoolSourceFree against NULL args, so I removed those. ] >From a53efcee7ec74948a2cf4c1c02902419856b0154 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 5 Feb 2010 17:09:43 +0100 Subject: [PATCH] plug four virStoragePoolSourceFree-related leaks * src/conf/storage_conf.c (virStoragePoolDefParseSourceString): * src/storage/storage_backend_fs.c: (virStorageBackendFileSystemNetFindPoolSourcesFunc): (virStorageBackendFileSystemNetFindPoolSources): * src/test/test_driver.c (testStorageFindPoolSources): --- src/conf/storage_conf.c | 4 ++-- src/storage/storage_backend_fs.c | 8 ++++---- src/test/test_driver.c | 3 ++- 3 files changed, 8 insertions(+), 7 deletions(-) diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c index 62b8394..b98b8e9 100644 --- a/src/conf/storage_conf.c +++ b/src/conf/storage_conf.c @@ -524,8 +524,8 @@ virStoragePoolDefParseSourceString(virConnectPtr conn, ret = def; def = NULL; cleanup: - if (def) - virStoragePoolSourceFree(def); + virStoragePoolSourceFree(def); + VIR_FREE(def); xmlFreeDoc(doc); xmlXPathFreeContext(xpath_ctxt); diff --git a/src/storage/storage_backend_fs.c b/src/storage/storage_backend_fs.c index 0d1c7a7..70e91a0 100644 --- a/src/storage/storage_backend_fs.c +++ b/src/storage/storage_backend_fs.c @@ -174,8 +174,8 @@ virStorageBackendFileSystemNetFindPoolSourcesFunc(virConnectPtr conn, src = NULL; ret = 0; cleanup: - if (src) - virStoragePoolSourceFree(src); + virStoragePoolSourceFree(src); + VIR_FREE(src); return ret; } @@ -236,8 +236,8 @@ virStorageBackendFileSystemNetFindPoolSources(virConnectPtr conn, for (i = 0; i < state.list.nsources; i++) virStoragePoolSourceFree(&state.list.sources[i]); - if (source) - virStoragePoolSourceFree(source); + virStoragePoolSourceFree(source); + VIR_FREE(source); VIR_FREE(state.list.sources); diff --git a/src/test/test_driver.c b/src/test/test_driver.c index 2122a1b..8e5ecf2 100644 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -1,7 +1,7 @@ /* * test.c: A "mock" hypervisor for use by application unit tests * - * Copyright (C) 2006-2009 Red Hat, Inc. + * Copyright (C) 2006-2010 Red Hat, Inc. * Copyright (C) 2006 Daniel P. Berrange * * This library is free software; you can redistribute it and/or @@ -3763,6 +3763,7 @@ testStorageFindPoolSources(virConnectPtr conn, cleanup: virStoragePoolSourceFree(source); + VIR_FREE(source); return ret; } -- 1.7.0.rc1.204.gb96e

2 2

[libvirt] [PATCH 1/5] macvtap support for libvirt -- build support
by Stefan Berger 11 Feb '10

11 Feb '10

This patch adds build support for libvirt checking for certain contents of /usr/include/linux/if_link.h to see whether macvtap support is compilable on that system. One can disable macvtap support in libvirt via --without-macvtap passed to configure. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

3 2

[libvirt] Node devices?
by Ed Swierk 11 Feb '10

11 Feb '10

After staring at the code implementing node devices in libvirt, I have decided that they're really neat, but I have no idea what they are for or how to use them. Reading the API docs and searching libvir-list didn't help much, and the user documentation (http://libvirt.org/archnode.html and http://libvirt.org/formatnode.html) is a bit, err, high-level. Can someone enlighten me? --Ed

5 5

[libvirt] [PATCH] Fix disk stats retrieval with QEMU >= 0.12
by Daniel P. Berrange 10 Feb '10

10 Feb '10

With QEMU >= 0.12 the host and guest side of disks no longer have the same naming convention. Specifically the host side will now get a 'drive-' prefix added to its name. The 'info blockstats' monitor command returns the host side name, so it is neccessary to strip this off when looking up stats since libvirt stores the guest side name ! * src/qemu/qemu_conf.c, src/qemu/qemu_conf.h: Move 'drive-' prefix string to a defined constant * src/qemu/qemu_monitor_json.c, src/qemu/qemu_monitor_text.c: Strip off 'drive-' prefix (if found) when looking up disk stats --- src/qemu/qemu_conf.c | 4 ++-- src/qemu/qemu_conf.h | 2 ++ src/qemu/qemu_monitor_json.c | 7 +++++++ src/qemu/qemu_monitor_text.c | 7 +++++++ 4 files changed, 18 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 3988582..c1d03cd 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2301,7 +2301,7 @@ qemuBuildDriveStr(virDomainDiskDefPtr disk, virBufferAddLit(&opt, ",media=cdrom"); if (qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) { - virBufferVSprintf(&opt, ",id=drive-%s", disk->info.alias); + virBufferVSprintf(&opt, ",id=%s%s", QEMU_DRIVE_HOST_PREFIX, disk->info.alias); } else { if (busid == -1 && unitid == -1) { if (idx != -1) @@ -2390,7 +2390,7 @@ qemuBuildDriveDevStr(virDomainDiskDefPtr disk) _("unsupported disk bus '%s' with device setup"), bus); goto error; } - virBufferVSprintf(&opt, ",drive=drive-%s", disk->info.alias); + virBufferVSprintf(&opt, ",drive=%s%s", QEMU_DRIVE_HOST_PREFIX, disk->info.alias); virBufferVSprintf(&opt, ",id=%s", disk->info.alias); if (virBufferError(&opt)) { diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h index d26bb90..498f1d1 100644 --- a/src/qemu/qemu_conf.h +++ b/src/qemu/qemu_conf.h @@ -157,6 +157,8 @@ typedef qemuDomainPCIAddressSet *qemuDomainPCIAddressSetPtr; /* Config type for XML import/export conversions */ #define QEMU_CONFIG_FORMAT_ARGV "qemu-argv" +#define QEMU_DRIVE_HOST_PREFIX "drive-" + #define qemuReportError(code, fmt...) \ virReportErrorHelper(NULL, VIR_FROM_QEMU, code, __FILE__, \ __FUNCTION__, __LINE__, fmt) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index e5288ce..032afef 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -754,6 +754,13 @@ int qemuMonitorJSONGetBlockStatsInfo(qemuMonitorPtr mon, goto cleanup; } + /* New QEMU has separate names for host & guest side of the disk + * and libvirt gives the host side a 'drive-' prefix. The passed + * in devname is the guest side though + */ + if (STRPREFIX(thisdev, QEMU_DRIVE_HOST_PREFIX)) + thisdev += strlen(QEMU_DRIVE_HOST_PREFIX); + if (STRNEQ(thisdev, devname)) continue; diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 23fc4cf..a6a4598 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -649,6 +649,13 @@ int qemuMonitorTextGetBlockStatsInfo(qemuMonitorPtr mon, p = info; while (*p) { + /* New QEMU has separate names for host & guest side of the disk + * and libvirt gives the host side a 'drive-' prefix. The passed + * in devname is the guest side though + */ + if (STRPREFIX(p, QEMU_DRIVE_HOST_PREFIX)) + p += strlen(QEMU_DRIVE_HOST_PREFIX); + if (STREQLEN (p, devname, devnamelen) && p[devnamelen] == ':' && p[devnamelen+1] == ' ') { -- 1.6.5.2

3 2

[libvirt] domblkstat not working for qemu-kvm > 0.11
by Thomas Treutner 10 Feb '10

10 Feb '10

Hi, with all qemu-kvm > 0.11, virsh domblkstat is not working: virsh -c qemu://node03/system domblkstat wp01 vdb : invalid argument in no stats found for device virtio-disk1 (vda, vdb etc. works with 0.11) "virtio-disk1" seems to be a red herring: virsh -c qemu://node03/system domblkstat wp01 virtio-disk1 : invalid argument in invalid path: virtio-disk1 Most probably there have been made some changes with 0.12 libvirt does not resemble yet. If you need more debugging output, just tell. kr,t

2 2

[libvirt] [PATCH] qemu: Properly report a startup timeout error
by Cole Robinson 10 Feb '10

10 Feb '10

The timeout errors were unconditionally being overwritten by the less helpful 'unable to start guest' error. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/qemu/qemu_driver.c | 17 +++++++---------- 1 files changed, 7 insertions(+), 10 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 4374291..2172c99 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -1358,15 +1358,15 @@ qemudReadLogOutput(virConnectPtr conn, buf[got] = '\0'; if (got == buflen-1) { qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, - _("Out of space while reading %s log output"), - what); + _("Out of space while reading %s log output: %s"), + what, buf); return -1; } if (isdead) { qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, - _("Process exited while reading %s log output"), - what); + _("Process exited while reading %s log output: %s"), + what, buf); return -1; } @@ -1378,7 +1378,8 @@ qemudReadLogOutput(virConnectPtr conn, } qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, - _("Timed out while reading %s log output"), what); + _("Timed out while reading %s log output: %s"), + what, buf); return -1; } @@ -1557,12 +1558,8 @@ qemudWaitForMonitor(virConnectPtr conn, virStrerror(errno, ebuf, sizeof ebuf)); } - if (ret < 0) { - /* Unexpected end of file - inform user of QEMU log data */ - qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, - _("unable to start guest: %s"), buf); + if (ret < 0) return -1; - } VIR_DEBUG("Connect monitor to %p '%s'", vm, vm->def->name); if (qemuConnectMonitor(vm) < 0) -- 1.6.5.2

3 6

[libvirt] [PATCH] test: Fake security driver support in capabilities
by Cole Robinson 10 Feb '10

10 Feb '10

Having some value in capabilities helps testing this stuff in virt-manager. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/test/test_driver.c | 8 ++++++++ 1 files changed, 8 insertions(+), 0 deletions(-) diff --git a/src/test/test_driver.c b/src/test/test_driver.c index 411c5cd..3f95a22 100644 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -205,6 +205,14 @@ testBuildCapabilities(virConnectPtr conn) { caps->privateDataAllocFunc = testDomainObjPrivateAlloc; caps->privateDataFreeFunc = testDomainObjPrivateFree; + caps->host.secModel.model = strdup("testSecurity"); + if (!caps->host.secModel.model) + goto no_memory; + + caps->host.secModel.doi = strdup(""); + if (!caps->host.secModel.doi) + goto no_memory; + return caps; no_memory: -- 1.6.5.2

2 2

[libvirt] [PATCH] (absolutePathFromBaseFile): fix up preceding commit
by Jim Meyering 10 Feb '10

10 Feb '10

To my chagrin, I saw that my most recent commit introduced compilation errors. Sorry about that. Here's how I propose to fix it. >From 2d948a373ecebec6c06274f61b31d1ae9c40ae41 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 5 Feb 2010 14:57:35 +0100 Subject: [PATCH] (absolutePathFromBaseFile): fix up preceding commit * src/util/storage_file.c: Include <assert.h>. (absolutePathFromBaseFile): Assert that converting size_t to int is valid. Reverse length/string args to match "%.*s". Explicitly ignore the return value of virAsprintf. --- src/util/storage_file.c | 7 ++++++- 1 files changed, 6 insertions(+), 1 deletions(-) diff --git a/src/util/storage_file.c b/src/util/storage_file.c index 2c79fa9..135acec 100644 --- a/src/util/storage_file.c +++ b/src/util/storage_file.c @@ -26,7 +26,9 @@ #include <unistd.h> #include <fcntl.h> +#include <assert.h> #include "dirname.h" +#include "ignore-value.h" #include "memory.h" #include "virterror_internal.h" @@ -255,7 +257,10 @@ absolutePathFromBaseFile(const char *base_file, const char *path) if (*path == '/' || d_len == 0) return strdup(path); - virAsprintf(&res, "%.*s/%s", base_file, d_len, path); + /* Ensure that the following cast-to-int is valid. */ + assert (d_len <= INT_MAX); + + ignore_value(virAsprintf(&res, "%.*s/%s", (int) d_len, base_file, path)); return res; } -- 1.7.0.rc1.204.gb96e

3 4

[libvirt] [PATCH 0/8] Remove virConnectPtr from yet more places
by Daniel P. Berrange 10 Feb '10

10 Feb '10

Following on from previous two postings http://www.redhat.com/archives/libvir-list/2010-February/msg00198.html http://www.redhat.com/archives/libvir-list/2010-February/msg00279.html With this series, none of the shared driver code uses virConnectPtr anymore, with the exception of src/conf/domain_events.h which has a genuine need for it - events are dispatched back to a specifici connection handle.

2 17

[libvirt] [PATCH 0/4] Remove use of virConnectPtr from more internal APIs
by Daniel P. Berrange 10 Feb '10

10 Feb '10

Following on from Matthias' work http://www.redhat.com/archives/libvir-list/2010-February/msg00198.html This series kills the virConnectPtr from QEMU, and LXC drivers, the security driver support code and the domain XML handling code.

3 11

[libvirt] [PATCH] Convert qemu command line flags to 64-bit int
by Daniel P. Berrange 10 Feb '10

10 Feb '10

The QEMU flags are commonly stored as a signed or unsigned int, allowing only 31 flags. This limit is rather close, so to aid future patches, change it to a 64-bit int * src/qemu/qemu_conf.c, src/qemu/qemu_conf.h, src/qemu/qemu_driver.c, tests/qemuargv2xmltest.c, tests/qemuhelptest.c, tests/qemuxml2argvtest.c: Use 'unsigned long long' for QEMU flags --- src/qemu/qemu_conf.c | 28 ++++++++++++++-------------- src/qemu/qemu_conf.h | 14 +++++++------- src/qemu/qemu_driver.c | 12 ++++++------ tests/qemuargv2xmltest.c | 2 +- tests/qemuhelptest.c | 13 +++++++------ tests/qemuxml2argvtest.c | 7 ++++--- 6 files changed, 39 insertions(+), 37 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 257b4a4..e3769c8 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -1108,12 +1108,12 @@ virCapsPtr qemudCapsInit(virCapsPtr old_caps) { return NULL; } -static unsigned int qemudComputeCmdFlags(const char *help, - unsigned int version, - unsigned int is_kvm, - unsigned int kvm_version) +static unsigned long long qemudComputeCmdFlags(const char *help, + unsigned int version, + unsigned int is_kvm, + unsigned int kvm_version) { - unsigned int flags = 0; + unsigned long long flags = 0; if (strstr(help, "-no-kqemu")) flags |= QEMUD_CMD_FLAG_KQEMU; @@ -1236,7 +1236,7 @@ static unsigned int qemudComputeCmdFlags(const char *help, #define SKIP_BLANKS(p) do { while ((*(p) == ' ') || (*(p) == '\t')) (p)++; } while (0) int qemudParseHelpStr(const char *help, - unsigned int *flags, + unsigned long long *flags, unsigned int *version, unsigned int *is_kvm, unsigned int *kvm_version) @@ -1316,14 +1316,14 @@ fail: int qemudExtractVersionInfo(const char *qemu, unsigned int *retversion, - unsigned int *retflags) { + unsigned long long *retflags) { const char *const qemuarg[] = { qemu, "-help", NULL }; const char *const qemuenv[] = { "LC_ALL=C", NULL }; pid_t child; int newstdout = -1; int ret = -1, status; unsigned int version, is_kvm, kvm_version; - unsigned int flags = 0; + unsigned long long flags = 0; if (retflags) *retflags = 0; @@ -1425,7 +1425,7 @@ int qemudNetworkIfaceConnect(virConnectPtr conn, struct qemud_driver *driver, virDomainNetDefPtr net, - int qemuCmdFlags) + unsigned long long qemuCmdFlags) { char *brname = NULL; int err; @@ -1637,7 +1637,7 @@ no_memory: int -qemuAssignDeviceDiskAlias(virDomainDiskDefPtr def, int qemuCmdFlags) +qemuAssignDeviceDiskAlias(virDomainDiskDefPtr def, unsigned long long qemuCmdFlags) { if (qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE) { if (qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) @@ -1720,7 +1720,7 @@ qemuAssignDeviceControllerAlias(virDomainControllerDefPtr controller) static int -qemuAssignDeviceAliases(virDomainDefPtr def, int qemuCmdFlags) +qemuAssignDeviceAliases(virDomainDefPtr def, unsigned long long qemuCmdFlags) { int i; @@ -2189,7 +2189,7 @@ qemuSafeSerialParamValue(virConnectPtr conn, char * qemuBuildDriveStr(virDomainDiskDefPtr disk, int bootable, - int qemuCmdFlags) + unsigned long long qemuCmdFlags) { virBuffer opt = VIR_BUFFER_INITIALIZER; const char *bus = virDomainDiskQEMUBusTypeToString(disk->bus); @@ -3017,7 +3017,7 @@ no_memory: static char * qemuBuildSmpArgStr(const virDomainDefPtr def, - int qemuCmdFlags) + unsigned long long qemuCmdFlags) { virBuffer buf = VIR_BUFFER_INITIALIZER; @@ -3057,7 +3057,7 @@ int qemudBuildCommandLine(virConnectPtr conn, virDomainDefPtr def, virDomainChrDefPtr monitor_chr, int monitor_json, - unsigned int qemuCmdFlags, + unsigned long long qemuCmdFlags, const char ***retargv, const char ***retenv, int **tapfds, diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h index 5525b6c..f6397b7 100644 --- a/src/qemu/qemu_conf.h +++ b/src/qemu/qemu_conf.h @@ -170,10 +170,10 @@ virCapsPtr qemudCapsInit (virCapsPtr old_caps); int qemudExtractVersion (struct qemud_driver *driver); int qemudExtractVersionInfo (const char *qemu, unsigned int *version, - unsigned int *flags); + unsigned long long *qemuCmdFlags); int qemudParseHelpStr (const char *str, - unsigned int *flags, + unsigned long long *qemuCmdFlags, unsigned int *version, unsigned int *is_kvm, unsigned int *kvm_version); @@ -183,7 +183,7 @@ int qemudBuildCommandLine (virConnectPtr conn, virDomainDefPtr def, virDomainChrDefPtr monitor_chr, int monitor_json, - unsigned int qemuCmdFlags, + unsigned long long qemuCmdFlags, const char ***retargv, const char ***retenv, int **tapfds, @@ -205,12 +205,12 @@ char * qemuBuildNicStr(virConnectPtr conn, /* Current, best practice */ char * qemuBuildNicDevStr(virDomainNetDefPtr net, - int qemuCmdFlags); + int vlan); /* Both legacy & current support */ char *qemuBuildDriveStr(virDomainDiskDefPtr disk, int bootable, - int qemuCmdFlags); + unsigned long long qemuCmdFlags); /* Current, best practice */ char * qemuBuildDriveDevStr(virConnectPtr conn, @@ -244,7 +244,7 @@ char * qemuBuildUSBHostdevDevStr(virDomainHostdevDefPtr dev); int qemudNetworkIfaceConnect (virConnectPtr conn, struct qemud_driver *driver, virDomainNetDefPtr net, - int qemuCmdFlags); + unsigned long long qemuCmdFlags); int qemudProbeMachineTypes (const char *binary, virCapsGuestMachinePtr **machines, @@ -283,7 +283,7 @@ int qemuAssignDevicePCISlots(virDomainDefPtr def, qemuDomainPCIAddressSetPtr ad int qemuDomainNetVLAN(virDomainNetDefPtr def); int qemuAssignDeviceNetAlias(virDomainDefPtr def, virDomainNetDefPtr net, int idx); -int qemuAssignDeviceDiskAlias(virDomainDiskDefPtr def, int qemuCmdFlags); +int qemuAssignDeviceDiskAlias(virDomainDiskDefPtr def, unsigned long long qemuCmdFlags); int qemuAssignDeviceHostdevAlias(virDomainDefPtr def, virDomainHostdevDefPtr net, int idx); int qemuAssignDeviceControllerAlias(virDomainControllerDefPtr controller); diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 4374291..54d33cd 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2520,7 +2520,7 @@ static int qemudStartVMDaemon(virConnectPtr conn, struct stat sb; int *tapfds = NULL; int ntapfds = 0; - unsigned int qemuCmdFlags; + unsigned long long qemuCmdFlags; fd_set keepfd; const char *emulator; pid_t child; @@ -4707,7 +4707,7 @@ static char *qemuDomainXMLToNative(virConnectPtr conn, virDomainDefPtr def = NULL; virDomainChrDef monConfig; const char *emulator; - unsigned int qemuCmdFlags; + unsigned long long qemuCmdFlags; struct stat sb; const char **retargv = NULL; const char **retenv = NULL; @@ -5586,7 +5586,7 @@ static int qemudDomainAttachNetDevice(virConnectPtr conn, struct qemud_driver *driver, virDomainObjPtr vm, virDomainNetDefPtr net, - unsigned int qemuCmdFlags) + unsigned long long qemuCmdFlags) { qemuDomainObjPrivatePtr priv = vm->privateData; char *tapfd_name = NULL; @@ -5901,7 +5901,7 @@ static int qemudDomainAttachDevice(virDomainPtr dom, struct qemud_driver *driver = dom->conn->privateData; virDomainObjPtr vm; virDomainDeviceDefPtr dev = NULL; - unsigned int qemuCmdFlags; + unsigned long long qemuCmdFlags; virCgroupPtr cgroup = NULL; int ret = -1; @@ -7437,7 +7437,7 @@ qemudDomainMigratePrepareTunnel(virConnectPtr dconn, int ret = -1; int internalret; char *unixfile = NULL; - unsigned int qemuCmdFlags; + unsigned long long qemuCmdFlags; struct qemuStreamMigFile *qemust = NULL; qemuDriverLock(driver); @@ -7884,7 +7884,7 @@ static int doTunnelMigrate(virDomainPtr dom, virStreamPtr st = NULL; char *unixfile = NULL; int internalret; - unsigned int qemuCmdFlags; + unsigned long long qemuCmdFlags; int status; unsigned long long transferred, remaining, total; diff --git a/tests/qemuargv2xmltest.c b/tests/qemuargv2xmltest.c index 1f1914b..c125a34 100644 --- a/tests/qemuargv2xmltest.c +++ b/tests/qemuargv2xmltest.c @@ -75,7 +75,7 @@ static int testCompareXMLToArgvFiles(const char *xml, struct testInfo { const char *name; - int extraFlags; + unsigned long long extraFlags; const char *migrateFrom; }; diff --git a/tests/qemuhelptest.c b/tests/qemuhelptest.c index 34a6048..0f2b509 100644 --- a/tests/qemuhelptest.c +++ b/tests/qemuhelptest.c @@ -12,10 +12,10 @@ struct testInfo { const char *name; - unsigned flags; - unsigned version; - unsigned is_kvm; - unsigned kvm_version; + unsigned long long flags; + unsigned int version; + unsigned int is_kvm; + unsigned int kvm_version; }; static char *progname; @@ -41,7 +41,8 @@ static int testHelpStrParsing(const void *data) char path[PATH_MAX]; char helpStr[MAX_HELP_OUTPUT_SIZE]; char *help = &(helpStr[0]); - unsigned flags, version, is_kvm, kvm_version; + unsigned int version, is_kvm, kvm_version; + unsigned long long flags; snprintf(path, PATH_MAX, "%s/qemuhelpdata/%s", abs_srcdir, info->name); @@ -52,7 +53,7 @@ static int testHelpStrParsing(const void *data) return -1; if (flags != info->flags) { - fprintf(stderr, "Computed flags do not match: got 0x%x, expected 0x%x\n", + fprintf(stderr, "Computed flags do not match: got 0x%llx, expected 0x%llx\n", flags, info->flags); if (getenv("VIR_TEST_DEBUG")) diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 67dc47e..f4908b8 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -24,7 +24,7 @@ static struct qemud_driver driver; static int testCompareXMLToArgvFiles(const char *xml, const char *cmd, - int extraFlags, + unsigned long long extraFlags, const char *migrateFrom) { char argvData[MAX_FILE]; char *expectargv = &(argvData[0]); @@ -32,7 +32,8 @@ static int testCompareXMLToArgvFiles(const char *xml, const char **argv = NULL; const char **qenv = NULL; const char **tmp = NULL; - int ret = -1, len, flags; + int ret = -1, len; + unsigned long long flags; virDomainDefPtr vmdef = NULL; virDomainChrDef monitor_chr; @@ -142,7 +143,7 @@ static int testCompareXMLToArgvFiles(const char *xml, struct testInfo { const char *name; - int extraFlags; + unsigned long long extraFlags; const char *migrateFrom; }; -- 1.6.6

3 6

[libvirt] [PATCH] Fix security driver configuration
by Daniel P. Berrange 10 Feb '10

10 Feb '10

The security driver was mistakenly initialized before the QEMU config file was loaded. This prevents it being turned off again. The capabilities XML was also getting the wrong security driver name, due to the stacked driver arrangement. * src/qemu/qemu_driver.c: Fix initialization order and capabilities model name --- src/qemu/qemu_driver.c | 20 ++++++++++---------- 1 files changed, 10 insertions(+), 10 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 54d33cd..cac5b8a 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -933,7 +933,7 @@ qemudSecurityInit(struct qemud_driver *qemud_drv) static virCapsPtr qemuCreateCapabilities(virCapsPtr oldcaps, - virSecurityDriverPtr secDriver) + struct qemud_driver *driver) { virCapsPtr caps; @@ -951,11 +951,11 @@ qemuCreateCapabilities(virCapsPtr oldcaps, /* Security driver data */ - if (secDriver) { + if (driver->securityPrimaryDriver) { const char *doi, *model; - doi = virSecurityDriverGetDOI(secDriver); - model = virSecurityDriverGetModel(secDriver); + doi = virSecurityDriverGetDOI(driver->securityPrimaryDriver); + model = virSecurityDriverGetModel(driver->securityPrimaryDriver); if (!(caps->host.secModel.model = strdup(model))) goto no_memory; @@ -1098,20 +1098,20 @@ qemudStartup(int privileged) { virStrerror(-rc, buf, sizeof(buf))); } + if (qemudLoadDriverConfig(qemu_driver, driverConf) < 0) { + goto error; + } + if (qemudSecurityInit(qemu_driver) < 0) goto error; if ((qemu_driver->caps = qemuCreateCapabilities(NULL, - qemu_driver->securityDriver)) == NULL) + qemu_driver)) == NULL) goto error; if ((qemu_driver->activePciHostdevs = pciDeviceListNew()) == NULL) goto error; - if (qemudLoadDriverConfig(qemu_driver, driverConf) < 0) { - goto error; - } - if (privileged) { if (chown(qemu_driver->libDir, qemu_driver->user, qemu_driver->group) < 0) { virReportSystemError(errno, @@ -3065,7 +3065,7 @@ static char *qemudGetCapabilities(virConnectPtr conn) { qemuDriverLock(driver); if ((caps = qemuCreateCapabilities(qemu_driver->caps, - qemu_driver->securityDriver)) == NULL) { + qemu_driver)) == NULL) { virCapabilitiesFree(caps); goto cleanup; } -- 1.6.6

2 1

[libvirt] PATCH: Depend on libcgroup & turn on cgconfig initscript
by Daniel P. Berrange 10 Feb '10

10 Feb '10

Since libvirt uses cgroups for various aspects of guest tuning, we need to make sure that the host is configured with them enabled out of the box. The libcgroup maintainer has now provided a suitable default config that mounts them all in /cgroup. So libvirt simply needs to depend on libcgroup RPM, and turn on the 'cgconfig' initscript at install time if QEMU or LXC is enabled. diff -u -r1.81 libvirt.spec --- libvirt.spec 3 Feb 2010 17:54:54 -0000 1.81 +++ libvirt.spec 9 Feb 2010 12:49:25 -0000 @@ -61,6 +61,7 @@ %define with_udev 0%{!?_without_udev:0} %define with_hal 0%{!?_without_hal:0} %define with_yajl 0%{!?_without_yajl:0} +%define with_cgconfig 0%{!?_without_cgconfig:0} # Non-server/HV driver defaults which are always enabled %define with_python 0%{!?_without_python:1} @@ -147,6 +148,13 @@ %define with_yajl 0%{!?_without_yajl:%{server_drivers}} %endif +# Pull in cgroups config system +%if 0%{?fedora} >= 12 || 0%{?rhel} >= 6 +%if %{with_qemu} || %{with_lxc} +%define with_cgconfig 0%{!?_without_cgconfig:1} +%endif +%endif + # Force QEMU to run as non-root %if 0%{?fedora} >= 12 || 0%{?rhel} >= 6 %define qemu_user qemu @@ -244,6 +252,9 @@ # For multipath support Requires: device-mapper %endif +%if %{with_cgconfig} +Requires: libcgroup +%endif %if %{with_xen} BuildRequires: xen-devel %endif @@ -634,6 +645,12 @@ fi %endif +%if %{with_cgconfig} +if [ "$1" = "1" ]; then +/sbin/chkconfig cgconfig on +fi +%endif + /sbin/chkconfig --add libvirtd if [ "$1" -ge "1" ]; then /sbin/service libvirtd condrestart > /dev/null 2>&1 -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

2 1

[libvirt] [PATCH] escape strings serialized in XML
by Daniel Veillard 10 Feb '10

10 Feb '10

A large part of the serialization code in src/conf/node_device_conf.c forgot that rule, this led to https://bugzilla.redhat.com/show_bug.cgi?id=547045 Escape strings serialized in XML * src/conf/node_device_conf.c: string saved into XML should be escaped with the use of virBufferEscapeString() diff --git a/src/conf/node_device_conf.c b/src/conf/node_device_conf.c index 4e8a492..69f42a9 100644 --- a/src/conf/node_device_conf.c +++ b/src/conf/node_device_conf.c @@ -372,19 +372,21 @@ char *virNodeDeviceDefFormat(virConnectPtr conn, virBufferVSprintf(&buf, " <protocol>%d</protocol>\n", data->usb_if.protocol); if (data->usb_if.description) - virBufferVSprintf(&buf, " <description>%s</description>\n", + virBufferEscapeString(&buf, + " <description>%s</description>\n", data->usb_if.description); break; case VIR_NODE_DEV_CAP_NET: - virBufferVSprintf(&buf, " <interface>%s</interface>\n", + virBufferEscapeString(&buf, " <interface>%s</interface>\n", data->net.ifname); if (data->net.address) - virBufferVSprintf(&buf, " <address>%s</address>\n", + virBufferEscapeString(&buf, " <address>%s</address>\n", data->net.address); if (data->net.subtype != VIR_NODE_DEV_CAP_NET_LAST) { const char *subtyp = virNodeDevNetCapTypeToString(data->net.subtype); - virBufferVSprintf(&buf, " <capability type='%s'/>\n", subtyp); + virBufferEscapeString(&buf, " <capability type='%s'/>\n", + subtyp); } break; case VIR_NODE_DEV_CAP_SCSI_HOST: @@ -392,10 +394,10 @@ char *virNodeDeviceDefFormat(virConnectPtr conn, data->scsi_host.host); if (data->scsi_host.flags & VIR_NODE_DEV_CAP_FLAG_HBA_FC_HOST) { virBufferAddLit(&buf, " <capability type='fc_host'>\n"); - virBufferVSprintf(&buf, - " <wwnn>%s</wwnn>\n", data->scsi_host.wwnn); - virBufferVSprintf(&buf, - " <wwpn>%s</wwpn>\n", data->scsi_host.wwpn); + virBufferEscapeString(&buf, " <wwnn>%s</wwnn>\n", + data->scsi_host.wwnn); + virBufferEscapeString(&buf, " <wwpn>%s</wwpn>\n", + data->scsi_host.wwpn); virBufferAddLit(&buf, " </capability>\n"); } if (data->scsi_host.flags & VIR_NODE_DEV_CAP_FLAG_HBA_VPORT_OPS) { @@ -405,8 +407,8 @@ char *virNodeDeviceDefFormat(virConnectPtr conn, break; case VIR_NODE_DEV_CAP_SCSI_TARGET: - virBufferVSprintf(&buf, " <target>%s</target>\n", - data->scsi_target.name); + virBufferEscapeString(&buf, " <target>%s</target>\n", + data->scsi_target.name); break; case VIR_NODE_DEV_CAP_SCSI: @@ -416,23 +418,23 @@ char *virNodeDeviceDefFormat(virConnectPtr conn, data->scsi.target); virBufferVSprintf(&buf, " <lun>%d</lun>\n", data->scsi.lun); if (data->scsi.type) - virBufferVSprintf(&buf, " <type>%s</type>\n", - data->scsi.type); + virBufferEscapeString(&buf, " <type>%s</type>\n", + data->scsi.type); break; case VIR_NODE_DEV_CAP_STORAGE: - virBufferVSprintf(&buf, " <block>%s</block>\n", + virBufferEscapeString(&buf, " <block>%s</block>\n", data->storage.block); if (data->storage.bus) - virBufferVSprintf(&buf, " <bus>%s</bus>\n", + virBufferEscapeString(&buf, " <bus>%s</bus>\n", data->storage.bus); if (data->storage.drive_type) - virBufferVSprintf(&buf, " <drive_type>%s</drive_type>\n", + virBufferEscapeString(&buf, " <drive_type>%s</drive_type>\n", data->storage.drive_type); if (data->storage.model) - virBufferVSprintf(&buf, " <model>%s</model>\n", + virBufferEscapeString(&buf, " <model>%s</model>\n", data->storage.model); if (data->storage.vendor) - virBufferVSprintf(&buf, " <vendor>%s</vendor>\n", + virBufferEscapeString(&buf, " <vendor>%s</vendor>\n", data->storage.vendor); if (data->storage.serial) virBufferVSprintf(&buf, " <serial>%s</serial>\n", -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

2 2

[libvirt] [0/3] Rebased patches for SPICE graphics support
by Daniel Veillard 10 Feb '10

10 Feb '10

This is a rebase of the 4 RHEL specific patches posted last sptember by Dan Berrange (but with patch 3 and 4 merged). It's actually a rebase from the gitorious spice set from Dan http://www.redhat.com/archives/libvir-list/2009-September/msg00865.html -- 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 7

[libvirt] [PATCH] fix AppArmor security driver when libvirt is compiled with libcap-ng
by Jamie Strandboge 10 Feb '10

10 Feb '10

The calls to virExec() in security_apparmor.c when invoking virt-aa-helper use VIR_EXEC_CLEAR_CAPS. When compiled without libcap-ng, this is not a problem (it's effectively a no-op) but with libcap-ng this causes MAC_ADMIN to be cleared. MAC_ADMIN is needed by virt-aa-helper to manipulate apparmor profiles and without it VMs will not start[1]. This patch calls virExec with the default VIR_EXEC_NONE instead. [1] https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/517714 -- Jamie Strandboge | http://www.canonical.com

2 1

[libvirt] [PATCH 0/9] virDomain{Attach,Detach}DeviceFlags patches
by Jim Fehlig 09 Feb '10

09 Feb '10

This set implements two new APIs, virDomainAttachDeviceFlags and virDomainDetachDeviceFlags as discussed here https://www.redhat.com/archives/libvir-list/2009-December/msg00124.html Introduce two new APIs virDomainAttachDeviceFlags(virDomainPtr dom, const char *xml, unsigned flags) virDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, unsigned flags) with flags being one or more from VIR_DOMAIN_DEVICE_MODIFY_CURRENT, VIR_DOMAIN_DEVICE_MODIFY_LIVE, VIR_DOMAIN_DEVICE_MODIFY_CONFIG. If caller specifies CURRENT (default), add or remove the device depending on the current state of domain. E.g. if domain is active add or remove the device to/from live domain, if it is inactive change the persistent config. If caller specifies LIVE, only change the active domain. If caller specifies CONFIG, only change persistent config - even if the domain is active. If caller specifies both LIVE and CONFIG, then change both. If a driver can not satisfy the exact requested flags it must return an error. E.g if user specified LIVE but the driver can only change live and persisted config, the driver must fail the request. The existing virDomain{Attach,Detach}Device is now explicitly restricted to active domains and is equivalent to virDomain{Attach,Detach}DeviceFlags(LIVE). Finally, virsh {attach,detach}-{disk,interface,device} has been modified to add a --persistent flag in order to set the appropriate flags when calling the new APIs. Jim Fehlig (9): Restrict virDomain{Attach,Detach}Device to active domains Public API Internal API Public API Implementation Wire protocol format Remote driver Server side dispatcher domain{Attach,Detach}DeviceFlags handler for drivers Modify virsh commands daemon/remote.c | 53 ++++++++++++++++++ include/libvirt/libvirt.h.in | 13 +++++ src/driver.h | 10 ++++ src/esx/esx_driver.c | 2 + src/libvirt.c | 120 ++++++++++++++++++++++++++++++++++++++---- src/libvirt_public.syms | 6 ++ src/lxc/lxc_driver.c | 2 + src/opennebula/one_driver.c | 2 + src/openvz/openvz_driver.c | 2 + src/phyp/phyp_driver.c | 2 + src/qemu/qemu_driver.c | 26 +++++++++ src/remote/remote_driver.c | 54 +++++++++++++++++++ src/remote/remote_protocol.x | 17 ++++++- src/test/test_driver.c | 2 + src/uml/uml_driver.c | 2 + src/vbox/vbox_tmpl.c | 24 ++++++++ src/xen/proxy_internal.c | 4 +- src/xen/xen_driver.c | 42 +++++++++++++-- src/xen/xen_driver.h | 4 +- src/xen/xen_hypervisor.c | 4 +- src/xen/xen_inotify.c | 4 +- src/xen/xend_internal.c | 98 ++++++++++++++++++++++++++++------ src/xen/xm_internal.c | 30 +++++++---- src/xen/xs_internal.c | 4 +- tools/virsh.c | 55 +++++++++++++++++-- 25 files changed, 523 insertions(+), 59 deletions(-)

3 33

[libvirt] [PATCH] Create raw storage files with O_DSYNC
by Jiri Denemark 09 Feb '10

09 Feb '10

When creating preallocated large raw files opening them with O_DSYNC prevents long delays in reading because cache pages can be immediately reused without writing them on a disk first. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/storage/storage_backend.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index e79294a..8aab9ab 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -301,7 +301,7 @@ virStorageBackendCreateRaw(virConnectPtr conn, goto cleanup; } - if ((fd = open(vol->target.path, O_RDWR | O_EXCL)) < 0) { + if ((fd = open(vol->target.path, O_RDWR | O_EXCL | O_DSYNC)) < 0) { virReportSystemError(conn, errno, _("cannot open new path '%s'"), vol->target.path); -- 1.6.6.1

2 2

[libvirt] ANNOUNCE: New releases virt-manager 0.8.3 and virtinst 0.500.2
by Cole Robinson 09 Feb '10

09 Feb '10

I'm happy to announce two new releases: virt-manager 0.8.3: virt-manager is a desktop application for managing KVM and Xen virtual machines via libvirt. virtinst 0.500.2: virtinst is a collection of command line tools for provisioning libvirt virtual machines, including virt-install and virt-clone. The releases can be downloaded from: http://virt-manager.org/download.html The direct download links are: http://virt-manager.org/download/sources/virt-manager/virt-manager-0.8.3.ta… http://virt-manager.org/download/sources/virtinst/virtinst-0.500.2.tar.gz The virt-manager release includes: - New ability to manage network interfaces: start, stop, and view existing interfaces. Provision new bridge, bond, and vlan devices. - New option to 'customize VM before install', which allows adjusting most VM options from the install wizard. The virtinst release includes: - virt-install: --autostart option for setting domain autostart flag - virt-install: Allow specifying --host-device with lsusb and lspci format (0x1234:0x5678 for USB, etc) - virtinst: New 'Interface' module for building libvirt interface XML Thanks to everyone who has contributed to this release through testing, bug reporting, submitting patches, and otherwise sending in feedback! Thanks, Cole

1 0

Re: [libvirt] Live migration woes
by Lance Albertson 09 Feb '10

09 Feb '10

> On Fri, Feb 05, 2010 at 03:21:32PM -0500, Thomas Sjolshagen wrote: >>> is anyone using libvirt-git and qemu-kvm post-0.11 and can live >>> migrate VMs successfully for- and backwards? qemu-kvm-0.11 seems >>> to be the latest release were live migration works, and I'm >>> trying to figure out if its a kvm or libvirt problem, since there >>> were other things (balloon parameter f.e.) that were changed with >>> qemu-kvm-0.12 and caused troubles when using libvirt. >> >> Yes. >> >> I'm using the virt-preview repo (rawhide) and I just noticed, >> today, after re-upgrading to 0.12.2-5 (since -4 has a rather nasty >> IO problem for virtio devices against raw image files) that my >> live-migrations fail with the following error message on the >> destination system (i.e. the system the I'm trying to migrate to): > >> /var/log/libvirt/qemu/<guest>.log: Unknown savevm section >> -2038417646 load of migration failed > > This looks like a new QEMU bug to me I can confirm this bug is happening on Ganeti as well [1]. It appears that if you sync the time on the guest VM before migrating back to the other node it will "succeed" most of the time. Has anyone opened a bug for this on the qemu-kvm bug tracker yet? I was about to do that myself. Thanks- [1] http://groups.google.com/group/ganeti/browse_thread/thread/9b2c556557e749cf -- Lance Albertson lance <at> osuosl.org Systems Administrator / Architect Open Source Lab Network Services Oregon State University Work: 541-737-9975 PGP Key: 0x27F4B742 GPG Fingerprint 0423 92F3 544A 1282 5AB1 4D07 416F A15D 27F4 B742

2 1

[libvirt] Inbound NAT and iptables rules
by Karl Vogel 09 Feb '10

09 Feb '10

This issue has been brought up a few times, but I haven't found any real solution yet. The problem is with the way libvirt adds iptables rules. I have a KVM Host with 1 public IP address, on which a few virtual machines run. I want to be able to direct certain ports on the public IP to certain virtual machine's internal IP. For example: allow conneciton to 'public-ip' tcp port: 80 to 'vm1' Which is possible using following rules: $ iptables -I FORWARD 1 -d vm1 -p tcp --dport 80 -j ACCEPT $ iptables -t nat -A PREROUTING -p tcp -d public-ip --dport 80 -j DNAT --to-destination vm1 However the problem is that libvirt _always_ adds it's rules to the top of the FORWARD rules, hence it blocks any 'custom' rules to the VM machines. ie. Chain FORWARD (policy ACCEPT 0 packets, 0 bytes) pkts bytes target prot opt in out source destination 39 2964 ACCEPT all -- * virbr0 0.0.0.0/0 192.168.122.0/24 state RELATED,ESTABLISHED 39 2964 ACCEPT all -- virbr0 * 192.168.122.0/24 0.0.0.0/0 2 656 ACCEPT all -- virbr0 virbr0 0.0.0.0/0 0.0.0.0/0 2 120 REJECT all -- * virbr0 0.0.0.0/0 0.0.0.0/0 reject-with icmp-port-unreachable 0 0 REJECT all -- virbr0 * 0.0.0.0/0 0.0.0.0/0 reject-with icmp-port-unreachable Only established connections are allowed with the libvirt rules, hence the forwarded connection to port 80 will be rejected by the 4th libvirt rule. The only way to make it work is to insert the rule _after_ libvirt has started the network.. and if the libvirt network ever gets restarted, the rule will be disabled since libvirt inserts it's rules at the top of the chain. Is there a proper way to achieve this with libvirt ?!

1 1