- Devel - libvirt List Archives

[libvirt] Fix cygwin build
by Daniel Veillard 09 Mar '11

09 Mar '11

In virLogEmergencyDumpAll I was using SIGBUS and SIGUSR2, and apparently they arent; defined on cygwin signal.h header. I decided to make the code conditional for all of them and not just those two for portability sake. I resisted doing this on the daemon side because as far as I know we don't compile it on WIN32. So I just pushed the following under the build breaking rule, Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

2 2

[libvirt] [PATCH 1/2] virsh: Remove indirect link against libxml2
by Guido Günther 08 Mar '11

08 Mar '11

as described at http://wiki.debian.org/ToolChain/DSOLinking https://fedoraproject.org/wiki/UnderstandingDSOLinkChange otherwise the build fails on current Debian unstable with: CCLD virsh /usr/bin/ld: virsh-virsh.o: undefined reference to symbol 'xmlSaveTree@@LIBXML2_2.6.8' /usr/bin/ld: note: 'xmlSaveTree@@LIBXML2_2.6.8' is defined in DSO //usr/lib/libxml2.so.2 so try adding it to the linker command line //usr/lib/libxml2.so.2: could not read symbols: Invalid operation O.k. to apply? -- Guido --- tools/Makefile.am | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/tools/Makefile.am b/tools/Makefile.am index 68471ea..3e31dbd 100644 --- a/tools/Makefile.am +++ b/tools/Makefile.am @@ -47,6 +47,7 @@ virsh_LDADD = \ ../src/libvirt.la \ ../src/libvirt-qemu.la \ ../gnulib/lib/libgnu.la \ + $(LIBXML_LIBS) \ $(VIRSH_LIBS) virsh_CFLAGS = \ -I$(top_srcdir)/gnulib/lib -I../gnulib/lib \ -- 1.7.4.1

2 2

[libvirt] [PATCH 0/2] virsh: Change command option parsing
by Michal Privoznik 08 Mar '11

08 Mar '11

This series changes some of vshCommandOpt* functions to return tri-state result according to cases if argument was found and valid and permutations. In addition, we ensure const-correctness of vshCommandOptString function and related code. Michal Privoznik (2): virsh: change vshCommandOptString return type and fix const-correctness virsh: Change option parsing functions to return tri-state information. tools/virsh.c | 717 +++++++++++++++++++++++++++------------------------------ 1 files changed, 335 insertions(+), 382 deletions(-) -- 1.7.4

2 4

[libvirt] [RFC PATCH 0/5] Allow data upload/download to/from storage volumes
by Daniel P. Berrange 08 Mar '11

08 Mar '11

This patch series introduces two new APIs to allow data upload/download to/from storage volumes, using the streams APIs. Mostly I wrote this so I had an easier way to test the streams handling in my RPC patch series than using QEMU tunnelled migration code. Cole has expressed an interest in this kind of functionality in the past to allow virt-manager to upload kernel/initrd to remote hosts for provisioning guests. Most of the code here is pretty boring & trivial. The first patch is the one that is remotely interesting. One might argue we should just spawn 'dd' instead of this, but dd has some annoying aspects (you can't control its write offset in units smaller than the block size) and this could be useful more broadly.

2 9

[libvirt] [PATCH] docs/formatdomain: Add release info for disk <driver> attributes
by Cole Robinson 08 Mar '11

08 Mar '11

--- docs/formatdomain.html.in | 7 ++++--- 1 files changed, 4 insertions(+), 3 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 84b1cab..9b9ab29 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -785,8 +785,7 @@ <dd> The optional driver element allows specifying further details related to the hypervisor driver used to provide the disk. - Since 0.1.8; <code>io</code> attribute - since 0.8.8 + Since 0.1.8 <ul> <li> If the hypervisor supports multiple backend drivers, then @@ -802,16 +801,18 @@ The optional <code>cache</code> attribute controls the cache mechanism, possible values are "default", "none", "writethrough" and "writeback". + Since 0.6.0 </li> <li> The optional <code>error_policy</code> attribute controls how the hypervisor will behave on an error, possible values are "stop", "ignore", and "enospace". + Since 0.8.0 </li> <li> The optional <code>io</code> attribute controls specific policies on I/O; qemu guests support "threads" and - "native". + "native". Since 0.8.8 </li> </ul> </dd> -- 1.7.4

2 2

[libvirt] [PATCH 0/3] device-detaching with generated mac
by Michal Privoznik 08 Mar '11

08 Mar '11

When detaching interface via detach-device or virDomainDetachDevice() we parse input XML and (probably) generate a random MAC. This leads then into not finding interface and thus error. When a mac wasn't specified and domain has exactly one interface, semantic is clear. Otherwise we require <mac> in input XML. Michal Privoznik (3): Introduce flag representing if MAC address of interface was generated or not. qemu: Check for generated MACs while detaching interface xen: Check for generated MACs while detaching interface src/conf/domain_conf.c | 2 ++ src/conf/domain_conf.h | 1 + src/qemu/qemu_hotplug.c | 26 ++++++++++++++++++++------ src/xen/xm_internal.c | 3 ++- 4 files changed, 25 insertions(+), 7 deletions(-) -- 1.7.4

5 8

[libvirt] [PATCH v2 0/4] Support domain snapshots with current QMP
by Jiri Denemark 08 Mar '11

08 Mar '11

This version 2 of this series implements the third alternative (suggested by Dan Berrange) for falling back to HMP if appropriate snapshot command does not exist in QMP. Version 2: - refactor text monitor commands into parts which can be called directly from json monitor code - implement HMP fallback for loadvm and delvm in addition to savevm Jiri Denemark (4): qemu: Refactor text monitor snapshot commands qemu: Fallback to HMP when savevm QMP command is not found qemu: Refactor qemuDomainSnapshotCreateXML qemu: Stop guest CPUs before creating a snapshot src/qemu/qemu_driver.c | 156 ++++++++++++++++++++++---------- src/qemu/qemu_monitor_json.c | 204 +++++++++++++++++++++++++++++++++++------- src/qemu/qemu_monitor_text.c | 179 ++++++++++++++++++++++++++----------- src/qemu/qemu_monitor_text.h | 8 ++ 4 files changed, 411 insertions(+), 136 deletions(-) -- 1.7.4.1

1 4

[libvirt] Add persistent XML for cpu tunables
by Osier Yang 08 Mar '11

08 Mar '11

These are pretty rough patches, something like a draft, I beleive there must be many problems, please review it heavily, :-) New 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 1/6] cputune: new tests for testing cputune xml [PATCH 2/6] cputune: support cputune for xend driver [PATCH 3/6] cputune: support cputune for qemu driver [PATCH 4/6] cputune: support cputune xml for LXC driver [PATCH 5/6] cputune: support cputune xml config [PATCH 6/6] cputune: Add docs xml schema for cputune xml

2 11

[libvirt] libvirt-php: various fixes
by Tiziano Mueller 08 Mar '11

08 Mar '11

While building a libvirt-php package for Gentoo Linux I had to workaround a couple of build issues. The first two patches are the results. The last patch is a fix for the case libvirt has been built without Xen support. In this case libvirt_connect (and phpinfo()) fails because virGetVersion is called with a non-NULL typeVer argument and libvirt therefore checks for Xen version. I guess the documentation for virGetVersion should be updated since it also returns 0 if both type and typeVer are NULL. [PATCH 1/3] Fix building with threaded php. [PATCH 2/3] add --with-php-config flag [PATCH 3/3] fix libvirt_connect failure without Xen

2 5

[libvirt] [PATCH] Fix performance problem of virStorageVolCreateXMLFrom()
by Minoru Usui 08 Mar '11

08 Mar '11

virStorageVolCreateXMLFrom() is slow if destination Pool is LVM. Because write block size is not appropriate. On linux environment, block size of LVM Pool which is made by virStoragePoolCreateXML() is 4096 bytes. On the other hand, write block size of virStorageVolCreateXMLFrom() is 512 bytes. This value are hard corded in virStorageBackendCopyToFD(). This patch fix this block size problem on linux. [lvm block size] # virsh vol-dumpxml vol.img --pool lvm-pool |grep '<path>' <path>/dev/lvm-pool/vol.img</path> # blockdev --getbsz /dev/lvm-pool/vol.img 4096 [before patch] # echo 1 > /proc/sys/vm/drop_caches # time virsh vol-create-from lvm-pool vol.xml --inputpool dir-pool vol.img real 1m17.736s user 0m0.007s sys 0m0.021s [after patch] # echo 1 > /proc/sys/vm/drop_caches # time virsh vol-create-from lvm-pool vol.xml --inputpool dir-pool vol.img real 0m36.883s user 0m0.011s sys 0m0.031s Signed-off-by: Minoru Usui <usui(a)mxm.nes.nec.co.jp> --- src/storage/storage_backend.c | 23 +++++++++++++++++++---- 1 files changed, 19 insertions(+), 4 deletions(-) diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index 174155d..a1a930c 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -36,6 +36,10 @@ #include <sys/stat.h> #include <sys/param.h> #include <dirent.h> +#ifdef __linux__ +#include <sys/ioctl.h> +#include <linux/fs.h> +#endif #if HAVE_SELINUX # include <selinux/selinux.h> @@ -108,6 +112,9 @@ enum { TOOL_QCOW_CREATE, }; +#define READ_BLOCK_SIZE_DEFAULT (1024 * 1024) +#define WRITE_BLOCK_SIZE_DEFAULT (512) + static int ATTRIBUTE_NONNULL (2) virStorageBackendCopyToFD(virStorageVolDefPtr vol, virStorageVolDefPtr inputvol, @@ -119,8 +126,10 @@ virStorageBackendCopyToFD(virStorageVolDefPtr vol, int amtread = -1; int ret = 0; unsigned long long remain; - size_t rbytes = 1024 * 1024; /* For Read */ - char zerobuf[512]; + size_t rbytes = READ_BLOCK_SIZE_DEFAULT; + size_t wbytes = WRITE_BLOCK_SIZE_DEFAULT; + int interval; + char zerobuf[WRITE_BLOCK_SIZE_DEFAULT]; char *buf = NULL; if ((inputfd = open(inputvol->target.path, O_RDONLY)) < 0) { @@ -131,6 +140,12 @@ virStorageBackendCopyToFD(virStorageVolDefPtr vol, goto cleanup; } +#ifdef __linux__ + if (ioctl(fd, BLKBSZGET, &wbytes) < 0) { + wbytes = WRITE_BLOCK_SIZE_DEFAULT; + } +#endif + bzero(&zerobuf, sizeof(zerobuf)); if (VIR_ALLOC_N(buf, rbytes) < 0) { @@ -160,7 +175,7 @@ virStorageBackendCopyToFD(virStorageVolDefPtr vol, * blocks */ amtleft = amtread; do { - int interval = ((512 > amtleft) ? amtleft : 512); + interval = ((wbytes > amtleft) ? amtleft : wbytes); int offset = amtread - amtleft; if (is_dest_file && memcmp(buf+offset, zerobuf, interval) == 0) { @@ -179,7 +194,7 @@ virStorageBackendCopyToFD(virStorageVolDefPtr vol, goto cleanup; } - } while ((amtleft -= 512) > 0); + } while ((amtleft -= interval) > 0); } if (VIR_CLOSE(inputfd) < 0) { -- Minoru Usui <usui(a)mxm.nes.nec.co.jp>

2 2