October 2012 - Devel - libvirt List Archives

[libvirt] [PATCH 0/4] command line fd passing using fd sets
by Corey Bryant 06 Oct '12

06 Oct '12

This series adds command line file descriptor passing support to the -drive option. This is a follow-on to the existing QMP fd passing support provided in the following patch series: comments.gmane.org/gmane.comp.emulators.qemu/165463 An example of using the new -drive fd and opaque options: qemu-kvm -drive fd=24,opaque="rdwr:/path/file",index=0,media=disk When an fd, and optionally opaque, are passed with -drive, they are added to an automatically generated fd set. The file name is also generated in the form of "/dev/fdset/nnn", where nnn is the fd set ID. qemu_open() already knows how to handle a filename of this format. qemu_open() searches the corresponding fd set for an fd and when it finds a match, QEMU goes on to use a dup of that fd just like it would have used an fd that it opened itself. Corey Bryant (4): monitor: Less restrictive fd matching for fd sets monitor: Enable adding an inherited fd to an fd set qemu-config: Add -drive fd and opaque options blockdev: Process -drive fd and opaque options blockdev.c | 40 +++++++++++++++++- monitor.c | 128 ++++++++++++++++++++++++++++++++------------------------ monitor.h | 3 ++ qemu-config.c | 8 ++++ qemu-options.hx | 15 ++++++- 5 files changed, 135 insertions(+), 59 deletions(-) -- 1.7.11.4

3 14

[libvirt] with latest libvirt(from git) on F18, time to take internal snapshot increases as snapshot count rises
by Kashyap Chamarthy 06 Oct '12

06 Oct '12

Hi, the below three internal snapshots were taken when the guest is running 'live'. Any hints on why the time to take snapshots keep increasing ? If you notice, the 2nd and 3rd internal snaphtots took more than 4 minutes each (note: there is no noticeable content change from snap1 ) #==================================================# [root@localhost vmimages]# time virsh snapshot-create-as f17vm3-with-b snap1 snap1-f17vm3-desc Domain snapshot snap1 created real 0m57.734s user 0m0.009s sys 0m0.024s [root@localhost vmimages]# [root@localhost vmimages]# virsh snapshot-list f17vm3-with-b Name Creation Time State ------------------------------------------------------------ snap1 2012-10-04 06:55:38 -0400 running [root@localhost vmimages]# #==================================================# [root@localhost vmimages]# time virsh snapshot-create-as f17vm3-with-b snap2 snap2-f17vm3-desc Domain snapshot snap2 created real 4m19.126s user 0m0.020s sys 0m0.030s [root@localhost vmimages]# #==================================================# [root@localhost vmimages]# time virsh snapshot-create-as f17vm3-with-b snap3 snap3-f17vm3-desc Domain snapshot snap3 created real 4m45.053s user 0m0.013s sys 0m0.032s [root@localhost vmimages]# #==================================================# [root@localhost qemu]# qemu-img info /home/kashyap/vmimages/f17vm3-with-b.qcow2 image: /home/kashyap/vmimages/f17vm3-with-b.qcow2 file format: qcow2 virtual size: 5.0G (5368709120 bytes) disk size: 755M cluster_size: 65536 backing file: /home/kashyap/vmimages/base-f17.qcow2 Snapshot list: ID TAG VM SIZE DATE VM CLOCK 1 snap1 232M 2012-10-04 06:55:38 19:58:19.804 2 snap2 247M 2012-10-04 07:03:09 20:04:53.360 3 snap3 244M 2012-10-04 07:08:23 20:05:48.889 4 snap4 0 2012-10-04 08:51:16 00:00:00.000 5 snap5 0 2012-10-04 08:53:04 00:00:00.000 [root@localhost qemu]# #==================================================# [root@localhost qemu]# virsh snapshot-list f17vm3-with-b Name Creation Time State ------------------------------------------------------------ snap1 2012-10-04 06:55:38 -0400 running snap2 2012-10-04 07:03:08 -0400 running snap3 2012-10-04 07:08:23 -0400 running snap4 2012-10-04 08:51:16 -0400 shutoff snap5 2012-10-04 08:53:04 -0400 shutoff [root@localhost qemu]# #==================================================# -- /kashyap

2 4

[libvirt] [PATCH] S390: Buffer too small for large CPU numbers.
by Viktor Mihajlovski 06 Oct '12

06 Oct '12

The output buffer for virFileReadAll was too small for systems with more than 30 CPUs which leads to a log entry and incorrect behavior. The new size will be sufficient for the current architectural limits. Signed-off-by: Viktor Mihajlovski <mihajlov(a)linux.vnet.ibm.com> --- src/util/sysinfo.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/util/sysinfo.c b/src/util/sysinfo.c index d38b4ff..8add936 100644 --- a/src/util/sysinfo.c +++ b/src/util/sysinfo.c @@ -361,7 +361,7 @@ virSysinfoRead(void) { goto no_memory; /* Gather info from /proc/cpuinfo */ - if (virFileReadAll(CPUINFO, 2048, &outbuf) < 0) { + if (virFileReadAll(CPUINFO, 8192, &outbuf) < 0) { virReportError(VIR_ERR_INTERNAL_ERROR, _("Failed to open %s"), CPUINFO); return NULL; @@ -376,7 +376,7 @@ virSysinfoRead(void) { VIR_FREE(outbuf); /* Gather info from /proc/sysinfo */ - if (virFileReadAll(SYSINFO, 4096, &outbuf) < 0) { + if (virFileReadAll(SYSINFO, 8192, &outbuf) < 0) { virReportError(VIR_ERR_INTERNAL_ERROR, _("Failed to open %s"), SYSINFO); return NULL; -- 1.7.0.4

2 2

[libvirt] [PATCH] Check for private symbols presence as well
by Michal Privoznik 05 Oct '12

05 Oct '12

Currently, we are checking if libvirt.so contains public symbols. However, sometimes we rename an internal symbol and forget to change libvirt_private.syms accordingly. Hence, it's safer to check for internal symbols as well. --- src/Makefile.am | 7 ++++++- src/check-symfile.pl | 2 +- 2 files changed, 7 insertions(+), 2 deletions(-) diff --git a/src/Makefile.am b/src/Makefile.am index ae3d491..c5840c0 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -318,8 +318,13 @@ if WITH_LINUX check-symfile: libvirt.syms libvirt.la $(AM_V_GEN)$(PERL) $(srcdir)/check-symfile.pl libvirt.syms \ .libs/libvirt.so + +check-private-symfile: libvirt_private.syms libvirt.la + $(AM_V_GEN)$(PERL) $(srcdir)/check-symfile.pl libvirt_private.syms \ + .libs/libvirt.so else check-symfile: +check-private-symfile: endif PROTOCOL_STRUCTS = \ @@ -344,7 +349,7 @@ else !WITH_REMOTE check-protocol: endif EXTRA_DIST += $(PROTOCOL_STRUCTS) check-symfile.pl -check-local: check-protocol check-symfile +check-local: check-protocol check-symfile check-private-symfile .PHONY: check-protocol $(PROTOCOL_STRUCTS:structs=struct) # Mock driver, covering domains, storage, networks, etc diff --git a/src/check-symfile.pl b/src/check-symfile.pl index ac37b46..435e045 100755 --- a/src/check-symfile.pl +++ b/src/check-symfile.pl @@ -44,7 +44,7 @@ foreach my $elflib (@elflibs) { close NM; } -foreach my $sym (@wantsyms) { +foreach my $sym (keys(%wantsyms)) { next if exists $gotsyms{$sym}; print STDERR "Expected symbol $sym is not in ELF library\n"; -- 1.7.8.6

2 5

[libvirt] [libvirt-php] Function calls blocking on Ubuntu 12.04
by Rory Slegtenhorst 05 Oct '12

05 Oct '12

Dear List, Please note I am not subscribed to this list. I just wanted to state my problem, and will wait for a fix when it comes. I have sucessfully compiled libvirt-php-0.4.5 on Ubuntu 12.04. I ran into some problems when using the library in simple php code, specifically trying to retrieve a screenshot. 1. libvirt_domain_get_screenshot_api returns a PHP fatal about it not being defined ?!? I have not looked into this at all. It was the first of the two I tried, and went on for the other :) 2. libvirt_domain_get_screenshot blocks forever I investigated this one, as from my debug logging, I could see things were going right. My browser just never finishes loading. I traced it down vnc.c on line x where a loop is started. The loop never seems to finish (unsure why though), and commenting out the loop fixes my problem. With kind regards, Rory Slegtenhorst rory dot slegtenhorst at gmail dot com

1 1

[libvirt] [PATCH] private.syms: Drop virKillProcess
by Michal Privoznik 05 Oct '12

05 Oct '12

After it got renamed to virProcessKill (cf470068a11) it is no longer needed in the private syms file. --- It is trivial, but maybe worth inventing cfg.mk macro that does check if all symbols fro private syms are defined somewhere. However, I am not that familiar with cfg.mk, so if anybody has any idea ... src/libvirt_private.syms | 1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index eebc52a..a8c81e7 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1246,7 +1246,6 @@ virGetUserName; virHexToBin; virIndexToDiskName; virIsDevMapperDevice; -virKillProcess; virParseNumber; virParseVersionString; virPipeReadUntilEOF; -- 1.7.8.6

2 2

[libvirt] [PATCH v3 0/2] security: support for names on DAC labels
by Marcelo Cerri 05 Oct '12

05 Oct '12

This patch series extends label parsing for DAC security driver and updates the related documentation. This version basically uses virReportError instead of VIR_DEBUG for error messages. Marcelo Cerri (2): security: also parse user/group names instead of just IDs for DAC labels doc: update description about security labels on formatdomain.html docs/formatdomain.html.in | 14 ++++++- src/security/security_dac.c | 92 ++++++++++++++++++++++++++++++++++----------- 2 files changed, 82 insertions(+), 24 deletions(-) -- 1.7.12

3 8

[libvirt] [PATCH] spec: prefer canonical name of util-linux
by Eric Blake 05 Oct '12

05 Oct '12

I noticed that in two places, we require util-linux, and in a third, we require util-linux-ng. On Fedora (I tested F15 through rawhide), util-linux-ng is obsoleted by util-linux; on RHEL 6, util-linux is obsoleted by util-linux-ng. That is, on either platform, either name will get you the correct package installed (where the preferred name on fedora is util-linux, and on RHEL 6 is util-linux-ng). But on RHEL 5, there is no util-linux-ng * libvirt.spec.in (Requires): Use util-linux, not util-linux-ng. --- I debated about pushing this under the trivial rule, but decided not to. Also, if you look at Fedora's git repository for libvirt, you'll notice that Peter Robinson attempted this same patch for F17 in commit f9de2f6, but with no explanation why. libvirt.spec.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 11e3199..e3d0a2d 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -614,7 +614,7 @@ Requires: PolicyKit >= 0.6 %if %{with_storage_fs} Requires: nfs-utils # For mkfs -Requires: util-linux-ng +Requires: util-linux # For pool-build probing for existing pools BuildRequires: libblkid-devel >= 2.17 # For glusterfs -- 1.7.11.4

2 2

[libvirt] migration.html page issues
by Eric Blake 05 Oct '12

05 Oct '12

Compare the table of contents on http://libvirt.org/migration.html to most other pages on libvirt.org. For some reason, the build process of converting migration.html.in into live html is rendering all links to internal anchors as '#' instead of full '#name' links. I'm throwing this out there in case someone else knows more quickly how to fix the problem (I'm not very fluent with all the xsltproc conversion stuff in the docs directory). -- Eric Blake eblake(a)redhat.com +1-919-301-3266 Libvirt virtualization library http://libvirt.org

2 2

[libvirt] feature or bug
by Gene Czarcinski 05 Oct '12

05 Oct '12

I notice that radvd is started for any ISOLATED (non-forwarding) network with an IPv6 definition. Why? In fact, I believe that there should be a way of defining a virtual network where neither dnsmasq nor radvd are started. My assumption is that virtual servers will be defined on the network as needed. I assume that the network addresses, etc. are needed by software somewhere deep in the bowels of the kernel and related software. Thus, it should not be possible to have guests use IPv4 or IPv6 addresses on a network other than those defined. Comments? Gene

1 0