- Devel - libvirt List Archives

[libvirt] Update on PHP libvirt
by Radek Hladik 01 Jul '10

01 Jul '10

I would like to post some update about PHP bindings for libvirt I started last year. As usual I do not have enough time to work on it however the last version 0.3 seems to be quite stable. I did get some reports from people using it and it seems that the code is working without problems. I will probably change the status from alpha to beta. The code (available at http://phplibvirt.cybersales.cz/ ) implements basic Libvirt functions for readonly access and machine live-cycle management. I should be able to add any other libvirt function but i do not know all the libvirt functions so well. So if you miss some function and you are willing to test it I can implement it. I would appreciate any feedback and suggestions on this code. Radek

5 10

[libvirt] [PATCH] cgroup: Fix compilation broken on MinGW due to dirent->d_type
by Ryota Ozaki 01 Jul '10

01 Jul '10

As pointed out by Eric Blake, using dirent->d_type breaks compilation on MinGW. This patch addresses this by using '#if defined' as same as doing for virCgroupForDriver. --- src/util/cgroup.c | 8 ++++++++ 1 files changed, 8 insertions(+), 0 deletions(-) diff --git a/src/util/cgroup.c b/src/util/cgroup.c index 9fa64dc..024036a 100644 --- a/src/util/cgroup.c +++ b/src/util/cgroup.c @@ -608,6 +608,7 @@ cleanup: } #endif +#if defined _DIRENT_HAVE_D_TYPE static int virCgroupRemoveRecursively(char *grppath) { DIR *grpdir; @@ -656,6 +657,13 @@ static int virCgroupRemoveRecursively(char *grppath) return rc; } +#else +static int virCgroupRemoveRecursively(char *grppath ATTRIBUTE_UNUSED) +{ + /* Claim no support */ + return -ENXIO; +} +#endif /** * virCgroupRemove: -- 1.6.5.2

2 3

[libvirt] [PATCH] virsh: tweak help output for VSH_OT_DATA
by Eric Blake 30 Jun '10

30 Jun '10

https://bugzilla.redhat.com/show_bug.cgi?id=609044 complained that 'virsh help pool-create-as' didn't document the shortcut that you can do 'virsh pool-create-as $name $type --target $target' rather than having to supply the four optional source- arguments in order to fill out the necessary positional arguments. This one-liner changes the help output to hopefully make this more obvious: NAME pool-create-as - create a pool from a set of args SYNOPSIS pool-create-as <name> [--print-xml] <type> [<source-host>] [<source-path>] [<source-dev>] [<source-name>] [<target>] [--source-format <string>] DESCRIPTION Create a pool. OPTIONS [--name] <string> name of the pool --print-xml print XML document, but don't define/create [--type] <string> type of the pool [--source-host] <string> source-host for underlying storage [--source-path] <string> source path for underlying storage [--source-dev] <string> source device for underlying storage [--source-name] <string> source name for underlying storage [--target] <string> target for underlying storage --source-format <string> format for underlying storage * tools/virsh.c (vshCmddefHelp): Make it more obvious that data arguments may, but not must, be specified by option leaders. --- Wow - the commit message is 30 times bigger than the commit itself! If additionally wanted, I could make the help output wrap if the left column is long, so that the right column lines up, looking more like: OPTIONS [--name] <string> name of the pool --print-xml print XML document, but don't define/create [--type] <string> type of the pool [--source-host] <string> source-host for underlying storage [--source-path] <string> source path for underlying storage [--source-dev] <string> source device for underlying storage [--source-name] <string> source name for underlying storage [--target] <string> target for underlying storage --source-format <string> format for underlying storage tools/virsh.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index f1bb295..cdac9c0 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -9558,7 +9558,7 @@ vshCmddefHelp(vshControl *ctl, const char *cmdname) else if (opt->type == VSH_OT_STRING) snprintf(buf, sizeof(buf), _("--%s <string>"), opt->name); else if (opt->type == VSH_OT_DATA) - snprintf(buf, sizeof(buf), "<%s>", opt->name); + snprintf(buf, sizeof(buf), "[--%s] <string>", opt->name); fprintf(stdout, " %-15s %s\n", buf, _(opt->help)); } -- 1.7.0.1

2 4

[libvirt] [PATCH] Avoid invoking the qemu monitor destroy callback if the constructor fails
by Daniel P. Berrange 30 Jun '10

30 Jun '10

Some, but not all, codepaths in the qemuMonitorOpen() method would trigger the destroy callback. The caller does not expect this to be invoked if construction fails, only during normal release of the monitor. This resulted in a possible double-unref of the virDomainObjPtr, because the caller explicitly unrefs the virDomainObjPtr if qemuMonitorOpen() fails * src/qemu/qemu_monitor.c: Don't invoke destroy callback from qemuMonitorOpen() failure paths --- src/qemu/qemu_monitor.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index f428665..ff613a0 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -671,6 +671,12 @@ qemuMonitorOpen(virDomainObjPtr vm, return mon; cleanup: + /* We don't want the 'destroy' callback invoked during + * cleanup from construction failure, because that can + * give a double-unref on virDomainObjPtr in the caller, + * so kill the callbacks now. + */ + mon->cb = NULL; qemuMonitorUnlock(mon); qemuMonitorClose(mon); return NULL; -- 1.6.6.1

3 3

[libvirt] [PATCH] cgroup: Add missing errno == ENOENT check in virCgroupRemoveRecursively
by Ryota Ozaki 30 Jun '10

30 Jun '10

ENOENT happens normally when a subsystem is enabled with any other subsystems and the directory of the target group has already removed in a prior loop. In that case, the function should just return without leaving an error message. NB this is the same behavior as before introducing virCgroupRemoveRecursively. --- src/util/cgroup.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/util/cgroup.c b/src/util/cgroup.c index 62b1446..9fa64dc 100644 --- a/src/util/cgroup.c +++ b/src/util/cgroup.c @@ -616,6 +616,8 @@ static int virCgroupRemoveRecursively(char *grppath) grpdir = opendir(grppath); if (grpdir == NULL) { + if (errno == ENOENT) + return 0; VIR_ERROR(_("Unable to open %s (%d)"), grppath, errno); rc = -errno; return rc; -- 1.6.5.2

2 4

[libvirt] [PATCH] vbox: Let configure detect/set the XPCOMC directory
by Matthias Bolte 30 Jun '10

30 Jun '10

This allows the user to give an explicit path to configure ./configure --with-vbox=/path/to/virtualbox instead of having the VirtualBox driver probe a set of possible paths at runtime. If no explicit path is specified then configure probes the set of "known" paths. https://bugzilla.redhat.com/show_bug.cgi?id=609185 --- configure.ac | 73 +++++++++++++++++++++++++++++++++++++++++++- src/Makefile.am | 1 + src/vbox/vbox_XPCOMCGlue.c | 22 ++----------- 3 files changed, 76 insertions(+), 20 deletions(-) diff --git a/configure.ac b/configure.ac index 691ef69..a1ba2f3 100644 --- a/configure.ac +++ b/configure.ac @@ -232,7 +232,7 @@ AC_ARG_WITH([phyp], AC_ARG_WITH([xenapi], AC_HELP_STRING([--with-xenapi], [add XenAPI support @<:@default=check@:>@]),[],[with_xenapi=check]) AC_ARG_WITH([vbox], - AC_HELP_STRING([--with-vbox], [add VirtualBox support @<:@default=yes@:>@]),[],[with_vbox=yes]) + AC_HELP_STRING([--with-vbox=@<:@PFX@:>@], [VirtualBox XPCOMC location @<:@default=check@:>@]),[],[with_vbox=check]) AC_ARG_WITH([lxc], AC_HELP_STRING([--with-lxc], [add Linux Container support @<:@default=check@:>@]),[],[with_lxc=check]) AC_ARG_WITH([one], @@ -315,6 +315,77 @@ if test "$with_openvz" = "yes"; then fi AM_CONDITIONAL([WITH_OPENVZ], [test "$with_openvz" = "yes"]) + +dnl +dnl check for VirtualBox XPCOMC location +dnl + +vbox_xpcomc_dir= + +if test "x$with_vbox" = "xyes" || test "x$with_vbox" = "xcheck"; then + AC_MSG_CHECKING([for VirtualBox XPCOMC location]) + + if test -f /usr/lib/virtualbox/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/usr/lib/virtualbox + else + if test -f /usr/lib/VirtualBox/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/usr/lib/VirtualBox + else + if test -f /opt/virtualbox/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/opt/virtualbox + else + if test -f /opt/virtualbox/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/opt/virtualbox + else + if test -f /opt/VirtualBox/i386/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/opt/VirtualBox/i386 + else + if test -f /opt/VirtualBox/amd64/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/opt/VirtualBox/amd64 + else + if test -f /usr/local/lib/virtualbox/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/usr/local/lib/virtualbox + else + if test -f /usr/lib/local/VirtualBox/VBoxXPCOMC.so; then + vbox_xpcomc_dir=/usr/lib/local/VirtualBox + else + if test -f /Application/VirtualBox.app/Contents/MacOS/VBoxXPCOMC.dylib; then + vbox_xpcomc_dir=/Application/VirtualBox.app/Contents/MacOS + fi + fi + fi + fi + fi + fi + fi + fi + fi + + if test -n "$vbox_xpcomc_dir"; then + AC_MSG_RESULT([$vbox_xpcomc_dir]) + with_vbox=yes + else + if test "x$with_vbox" = "xcheck"; then + AC_MSG_RESULT([not found, disabling VirtualBox driver]) + with_vbox=no + else + AC_MSG_RESULT([not found]) + AC_MSG_ERROR([VirtualBox XPCOMC is required for the VirtualBox driver]) + fi + fi +else + if test "x$with_vbox" != "xno"; then + if test -f ${with_vbox}/VBoxXPCOMC.so || test -f ${with_vbox}/VBoxXPCOMC.dylib; then + vbox_xpcomc_dir=$with_vbox + with_vbox=yes + else + AC_MSG_ERROR([$with_vbox does not contain VirtualBox XPCOMC]) + fi + fi +fi + +AC_SUBST([vbox_xpcomc_dir]) + if test "x$with_vbox" = "xyes"; then AC_SEARCH_LIBS([dlopen], [dl], [], [AC_MSG_ERROR([Unable to find dlopen()])]) case $ac_cv_search_dlopen in diff --git a/src/Makefile.am b/src/Makefile.am index ece18a6..d3c7087 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -19,6 +19,7 @@ INCLUDES = \ -DPKGDATADIR=\""$(pkgdatadir)"\" \ -DLOCAL_STATE_DIR=\""$(localstatedir)"\" \ -DGETTEXT_PACKAGE=\"$(PACKAGE)\" \ + -DVBOX_XPCOMC_DIR=\"$(vbox_xpcomc_dir)\" \ $(WARN_CFLAGS) \ $(LOCK_CHECKING_CFLAGS) \ -DIN_LIBVIRT \ diff --git a/src/vbox/vbox_XPCOMCGlue.c b/src/vbox/vbox_XPCOMCGlue.c index 0987c1b..fcae0cb 100644 --- a/src/vbox/vbox_XPCOMCGlue.c +++ b/src/vbox/vbox_XPCOMCGlue.c @@ -192,28 +192,12 @@ int VBoxCGlueInit(void) return tryLoadOne(pszHome, 0); /* - * Try the known standard locations. + * Try the configured location. */ g_szVBoxErrMsg[0] = '\0'; -#if defined(__gnu__linux__) || defined(__linux__) - if (tryLoadOne("/opt/VirtualBox", 1) == 0) - return 0; - if (tryLoadOne("/usr/lib/virtualbox", 1) == 0) - return 0; -#elif defined(__sun__) - if (tryLoadOne("/opt/VirtualBox/amd64", 1) == 0) - return 0; - if (tryLoadOne("/opt/VirtualBox/i386", 1) == 0) - return 0; -#elif defined(__APPLE__) - if (tryLoadOne("/Application/VirtualBox.app/Contents/MacOS", 1) == 0) - return 0; -#elif defined(__FreeBSD__) - if (tryLoadOne("/usr/local/lib/virtualbox", 1) == 0) + + if (tryLoadOne(VBOX_XPCOMC_DIR, 1) == 0) return 0; -#else -# error "port me" -#endif /* * Finally try the dynamic linker search path. -- 1.7.0.4

2 1

[libvirt] [PATCH 0/2]: Small fixes for PCI device assignment
by Chris Lalancette 29 Jun '10

29 Jun '10

This series fixes a couple of bugs found while testing libvirt PCI device assignment.

3 8

[libvirt] [PATCH] virFileResolveLink: guarantee an absolute path
by Eric Blake 29 Jun '10

29 Jun '10

https://bugzilla.redhat.com/show_bug.cgi?id=608092 * src/util/util.c (virFileResolveLink): Use canonicalize_file_name, rather than areadlink. --- For this patch, I chose to avoid any symlink resolution if an lstat shows the last element doesn't need any; but if we know we need symlink resolution, it is much easier to use existing interfaces to correctly resolve relative paths, even if it ends up doing more resolution work than what we strictly need. src/util/util.c | 26 ++++++++++++++++---------- 1 files changed, 16 insertions(+), 10 deletions(-) diff --git a/src/util/util.c b/src/util/util.c index 445fd4e..d058113 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -63,7 +63,7 @@ # include <mntent.h> #endif -#include "areadlink.h" +#include "dirname.h" #include "virterror_internal.h" #include "logging.h" #include "event.h" @@ -1178,8 +1178,9 @@ int virFileLinkPointsTo(const char *checkLink, /* - * Attempt to resolve a symbolic link, returning the - * real path + * Attempt to resolve a symbolic link, returning an + * absolute path where only the last component is guaranteed + * not to be a symlink. * * Return 0 if path was not a symbolic, or the link was * resolved. Return -1 with errno set upon error @@ -1191,16 +1192,21 @@ int virFileResolveLink(const char *linkpath, *resultpath = NULL; - if (lstat(linkpath, &st) < 0) - return -1; - - if (!S_ISLNK(st.st_mode)) { - if (!(*resultpath = strdup(linkpath))) + /* We don't need the full canonicalization of intermediate + * directories, if linkpath is absolute and the basename is + * already a non-symlink. */ + if (IS_ABSOLUTE_FILE_NAME(linkpath)) { + if (lstat(linkpath, &st) < 0) return -1; - return 0; + + if (!S_ISLNK(st.st_mode)) { + if (!(*resultpath = strdup(linkpath))) + return -1; + return 0; + } } - *resultpath = areadlink (linkpath); + *resultpath = canonicalize_file_name(linkpath); return *resultpath == NULL ? -1 : 0; } -- 1.7.0.1

2 2

[libvirt] Storage segfaults after phyp patches
by Cole Robinson 29 Jun '10

29 Jun '10

Basically: > $ gdb virsh ... > (gdb) run --connect qemu:///session pool-list --all > Starting program: /usr/bin/virsh --connect qemu:///session pool-list --all > [Thread debugging using libthread_db enabled] > Detaching after fork from child process 29579. > > Program received signal SIGSEGV, Segmentation fault. ... > (gdb) bt > #0 _libssh2_channel_open (session=0x0, channel_type=0x7ffff7b82737 "session", > channel_type_len=7, window_size=65536, packet_size=32768, message=0x0, > message_len=0) at channel.c:139 > #1 0x00000030cb008a7c in libssh2_channel_open_ex (session=0x0, > type=0x7ffff7b82737 "session", type_len=7, window_size=65536, > packet_size=32768, msg=<value optimized out>, msg_len=0) at channel.c:338 > #2 0x00007ffff7ac9edb in phypExec (session=0x0, > cmd=0x63ba50 "viosvrcmd -m (null) --id 0 -c 'lsvg'|grep -c '^.*$'", > exit_status=0x7fffffffdbcc, conn=<value optimized out>) > at phyp/phyp_driver.c:124 > #3 0x00007ffff7acae1c in phypNumOfStoragePools (conn=0x6341b0) > at phyp/phyp_driver.c:3032 > #4 0x00007ffff7a771be in virConnectNumOfStoragePools (conn=0x6341b0) > at libvirt.c:7223 > #5 0x0000000000415f9e in cmdPoolList (ctl=0x7fffffffdf80, cmd=0x1) > at virsh.c:4911 > #6 0x0000000000409dcc in vshCommandRun (ctl=0x7fffffffdf80, cmd=0x62d500) > at virsh.c:9647 > #7 0x00000000004186ed in main (argc=5, argv=<value optimized out>) > at virsh.c:10656 Looks like the phyp storage driver is being used instead of the libvirtd storage driver (and then segfaulting). Looked briefly for a fix, but it wasn't obvious to me. Thanks, Cole

5 8

[libvirt] [PATCH] Avoid calling virStorageFileIsSharedFS with NULL
by Laine Stump 29 Jun '10

29 Jun '10

From: Laine Stump <laine(a)redhat.com> This code was just recently added (by me) and didn't account for the fact that stdin_path is sometimes NULL. If it's NULL, and SetSecurityAllLabel fails, a segfault would result. --- src/qemu/qemu_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 2a277a5..e8c5c35 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -3429,7 +3429,7 @@ static int qemudStartVMDaemon(virConnectPtr conn, if (driver->securityDriver && driver->securityDriver->domainSetSecurityAllLabel && driver->securityDriver->domainSetSecurityAllLabel(vm, stdin_path) < 0) { - if (virStorageFileIsSharedFS(stdin_path) != 1) + if (stdin_path && virStorageFileIsSharedFS(stdin_path) != 1) goto cleanup; } -- 1.7.1

3 3