- Devel - libvirt List Archives

[libvirt] Libvirt Auth
by Marco Mornati 05 Mar '10

05 Mar '10

Hello guys, I've created a plugin for Hudson (CI) to manage virtmachines using libvirt java (start vm for a build and stop just after). I've just a problem in authentication, in fact I cannot send uname and password using code, libvirt ask me for the credentials on CLI. someone could explain me How I can use authentican or where I can find an usage guide or example? Thanks a lot Bye Marco

3 8

[libvirt] FW: Libvirt and openvz?
by Thomas Graves 05 Mar '10

05 Mar '10

I just recently started using openvz and I'm trying to use libvirt/virsh with it. I followed the directions on page: http://libvirt.org/drvopenvz.html But I get an error trying to run define with the example config file: virsh define virt.conf error: Failed to define domain from virt.conf error: unknown OS type exe sudo virsh -c openvz:/// list error: could not connect to openvz:/// error: failed to connect to the hypervisor I'm using libvirt 0.7.2 with latest rhel5.4 openvz. Does anyone have more example configuration files or other setup instructions? Does Libvirt allow you to specify other resource limits and controls that openvz has - like vzctl set --devnodes device:rw or vzctl set kmemsize? Thanks, Tom

1 1

[libvirt] [PATCH] fix rawhide/F13 --no-add-needed/dlopen-related build failures
by Jim Meyering 05 Mar '10

05 Mar '10

Diego, I saw no follow-up here, http://thread.gmane.org/gmane.comp.emulators.libvirt/21398/focus=21671 so I've fixed your two patches to address Eric's review feedback, and confirmed that they solve the problem on Rawhide. >From 55c182d89cd6e5417eafe9f4bdf2e0ef7f8721f5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Diego=20Elio=20Petten=C3=B2?= <flameeyes(a)gmail.com> Date: Thu, 4 Mar 2010 16:45:02 +0100 Subject: [PATCH 1/2] build: avoid dlopen-related link failure on rawhide/F13 Instead of using AC_CHECK_LIB and hardcoding -ldl, search for the library needed to get dlopen() and then use the cached value. --- configure.ac | 7 +++++-- 1 files changed, 5 insertions(+), 2 deletions(-) diff --git a/configure.ac b/configure.ac index 682b8b5..d5d62ea 100644 --- a/configure.ac +++ b/configure.ac @@ -1700,7 +1700,7 @@ if test "x$with_driver_modules" = "xyes" ; then old_libs="$LIBS" fail=0 AC_CHECK_HEADER([dlfcn.h],[],[fail=1]) - AC_CHECK_LIB([dl], [dlopen],[],[fail=1]) + AC_SEARCH_LIBS([dlopen], [dl], [], [fail=1]) test $fail = 1 && AC_MSG_ERROR([You must have dlfcn.h / dlopen() support to build driver modules]) @@ -1709,7 +1709,10 @@ if test "x$with_driver_modules" = "xyes" ; then fi if test "$with_driver_modules" = "yes"; then DRIVER_MODULE_CFLAGS="-export-dynamic" - DRIVER_MODULE_LIBS="-ldl" + case $ac_cv_search_dlopen in + no*) DRIVER_MODULE_LIBS= ;; + *) DRIVER_MODULE_LIBS=$ac_cv_search_dlopen ;; + esac AC_DEFINE_UNQUOTED([WITH_DRIVER_MODULES], 1, [whether to build drivers as modules]) fi AM_CONDITIONAL([WITH_DRIVER_MODULES], [test "$with_driver_modules" != "no"]) -- 1.7.0.1.475.gaf89b >From ca7a9f299b0bd415f8434b3c9e6c70528d2f85cc Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Diego=20Elio=20Petten=C3=B2?= <flameeyes(a)gmail.com> Date: Thu, 4 Mar 2010 16:48:10 +0100 Subject: [PATCH 2/2] build: vbox: avoid build failure when linking with --no-add-needed With the recent changes to the linking defaults in Fedora 13 (namely enabling --no-add-needed behaviour by default), we have to pass the dlopen()-providing libraries directly at the link of the module; use the same AC_SEARCH_LIBS function as used before to look for it and add it to the Makefile. --- configure.ac | 6 ++++++ src/Makefile.am | 1 + 2 files changed, 7 insertions(+), 0 deletions(-) diff --git a/configure.ac b/configure.ac index d5d62ea..923d7b8 100644 --- a/configure.ac +++ b/configure.ac @@ -303,6 +303,12 @@ fi AM_CONDITIONAL([WITH_OPENVZ], [test "$with_openvz" = "yes"]) if test "x$with_vbox" = "xyes"; then + AC_SEARCH_LIBS([dlopen], [dl], [], [AC_MSG_ERROR([Unable to find dlopen()])]) + case $ac_cv_search_dlopen in + no*) DLOPEN_LIBS= ;; + *) DLOPEN_LIBS=$ac_cv_search_dlopen ;; + esac + AC_SUBST([DLOPEN_LIBS]) AC_DEFINE_UNQUOTED([WITH_VBOX], 1, [whether VirtualBox driver is enabled]) fi AM_CONDITIONAL([WITH_VBOX], [test "$with_vbox" = "yes"]) diff --git a/src/Makefile.am b/src/Makefile.am index 2051e5d..67f8b6d 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -463,6 +463,7 @@ libvirt_driver_vbox_la_CFLAGS = \ if WITH_DRIVER_MODULES libvirt_driver_vbox_la_LDFLAGS = -module -avoid-version endif +libvirt_driver_vbox_la_LIBADD = $(DLOPEN_LIBS) libvirt_driver_vbox_la_SOURCES = $(VBOX_DRIVER_SOURCES) endif -- 1.7.0.1.475.gaf89b

2 1

[libvirt] FW: Bringing up LXC container from virsh command prompt.
by Kumar L Srikanth-B22348 04 Mar '10

04 Mar '10

Hi Daniel, One more thing is that in my host machine vibr0 inteface is missing. Because of the the 'default' network is in 'inactive' state and if I tried to activate, it's giving the following error: error: Failed to start network default error: internal error '/usr/sbin/dnsmasq --strict-order --bind-interfaces --pid-file=/var/run/libvirt/network/default.pid --conf-file= --listen-address 192.168.122.1 --except-interface lo --dhcp-range 192.168.122.2,192.168.122.254' exited with non-zero status 2 and signal 0: dnsmasq: failed to bind listening socket for 192.168.122.1: Address already in use Can you please guide me in this regard? Regards, Srikanth. -----Original Message----- From: Kumar L Srikanth-B22348 Sent: Thursday, March 04, 2010 4:43 PM To: 'Daniel P. Berrange' Cc: libvir-list(a)redhat.com; Reddy B Veera-B22347 Subject: RE: Bringing up LXC container from virsh command prompt. Hi, I want to create eth0 network interface in the container, which is linked to br0 bridge in the host machine. For that, I have created a bridge 'br0' with the IP '172.16.1.10'. Now, I want to create a network interface eth0 in the container with the IP '172.16.1.17'. I created the following network XML for the above requirement. <network> <name>mynetwork</name> <uuid>3e3fce45-4f53-4fa7-bb32-11f34168b82b</uuid> <bridge name="br0" stp="on" delay="5"/> <forward mode="nat" dev="eth0"/> <ip address="172.16.1.17" netmask="255.255.255.0"/> </network> I am able to define the network, but unable to start the network. The following error is coming when I start the network with the following command: #virsh -c lxc:/// net-start mynetwork error: Failed to start network mynetwork error: cannot create bridge 'br0': File exists Can you please guide me to create a network? My network interfaces in the host machine is shown below: br0 Link encap:Ethernet HWaddr 00:50:BF:27:AA:60 inet addr:172.16.1.10 Bcast:172.16.255.255 Mask:255.255.0.0 inet6 addr: fe80::250:bfff:fe27:aa60/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:80728 errors:0 dropped:0 overruns:0 frame:0 TX packets:170078 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:54037401 (51.5 MiB) TX bytes:123567268 (117.8 MiB) eth0 Link encap:Ethernet HWaddr 00:50:BF:27:AA:60 inet6 addr: fe80::250:bfff:fe27:aa60/64 Scope:Link UP BROADCAST RUNNING PROMISC MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:288 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 b) TX bytes:30108 (29.4 KiB) Interrupt:21 Base address:0x1140 eth1 Link encap:Ethernet HWaddr 00:16:76:4B:91:DC inet addr:192.168.1.10 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) eth2 Link encap:Ethernet HWaddr 00:11:95:FC:8D:62 inet addr:10.232.113.80 Bcast:10.232.113.255 Mask:255.255.254.0 inet6 addr: fe80::211:95ff:fefc:8d62/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:563988 errors:0 dropped:0 overruns:0 frame:0 TX packets:391293 errors:1100 dropped:0 overruns:0 carrier:1100 collisions:188525 txqueuelen:1000 RX bytes:565906426 (539.6 MiB) TX bytes:234291753 (223.4 MiB) Interrupt:22 Base address:0xa000 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:23001 errors:0 dropped:0 overruns:0 frame:0 TX packets:23001 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:3478027 (3.3 MiB) TX bytes:3478027 (3.3 MiB) Regards, Srikanth. -----Original Message----- From: Daniel P. Berrange [mailto:berrange@redhat.com] Sent: Thursday, March 04, 2010 3:41 PM To: Kumar L Srikanth-B22348 Cc: libvir-list(a)redhat.com; Reddy B Veera-B22347 Subject: Re: Bringing up LXC container from virsh command prompt. On Thu, Mar 04, 2010 at 03:38:43PM +0530, Kumar L Srikanth-B22348 wrote: > So, again I need to add one network interface in the libvirt XML?. It > will not take the interface which was created earlier while creating > container using lxc-tools?. > But, in the container which was created with lxc-tools(lxc-create, > lxc-start ..etc), I am not getting any such issues. The lxc-XXX tools are completely unrelated. They have a configuration file where you specify what network interfaces are created. Likewise if you use libvirt you need to specify what network interfaces are to be created. > Why I am getting these issues in virsh command prompt, when I am > mounting the same rootfs that was used while creating container using > lxc-tools. You need to setup $PATH in your container. Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://deltacloud.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] Bringing up LXC container from virsh command prompt.
by Kumar L Srikanth-B22348 04 Mar '10

04 Mar '10

Hi, I am new to libvirt virsh console. Currently I am working on Linux Containers(LXC). After lot of exploration, I am able to create a Linux Container(fedora12) with the help of lxc-tools(lxc-create, lxc-start, ...etc). I am able to start that container and use it's console. I am able to assign an IP address to the container, by creating a bridge in the host machine and assigning it to the container. Here, my goal is to bring up this container(fedora12) through libvirt virsh command prompt. For that I created a domain with the following XML file: [vm1_fedora_container.xml] <domain type='lxc' id='1'> <name>vm1_fedora</name> <memory>500000</memory> <os> <type>exe</type> <init>/bin/bash</init> </os> <vcpu>1</vcpu> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <emulator>/usr/libexec/libvirt_lxc</emulator> <filesystem type='mount'> <source dir='/root/lxc/fedora12'/> /*This is the rootfs directory for my linux container(fedora12)*/ <target dir='/'/> /* I mounted it to '/' directory in the Domain. */ </filesystem> <console type='pty' /> </devices> </domain> I executed the following virsh commands: # virsh --connect lxc:/// define vm1_fedora_container.xml Domain vm1_fedora defined from /root/lxc_devel/fedora.xml # virsh --connect lxc:/// list --all Id Name State ---------------------------------- - vm1_fedora shut off # virsh --connect lxc:/// start vm1_fedora Domain vm1_fedora started # virsh --connect lxc:/// list --all Id Name State ---------------------------------- 7805 vm1_fedora running # virsh --connect lxc:/// console vm1_fedora Connected to domain vm1_fedora Escape character is ^] [root@localhost /]# I am able to view the console of my container(fedora12), but when I execute 'ifconfig' in that console, I am getting the following error: bash: ifconfig: command not found When I execute the same 'ifconfig' command through lxc-tools (like lxc_start -n fedora12 /bin/bash), I am able to get the assigned IP address. Can you please suggest me where did I go wrong? Regards, Srikanth.

2 6

[libvirt] [PATCH 0/9] Vol Zeroing API
by David Allan 04 Mar '10

04 Mar '10

This patch set contains a new API for zeroing out volumes. I called it zero out rather than overwrite because it doesn't overwrite anything in the case of sparse files. It does guarantee that any subsequent reads from the volume will return zeroes. The ESX driver change is part of the patch. Without it, the struct is missing its initializer and fails to build. I'm not sure what our position on the style of struct initializers is; there are both forms in the code. If this style is heresy, we should add a statement to that effect in HACKING. Dave David Allan (9): Fix ESX storage driver struct initializer Add public API for volume zeroing Define the internal driver API for vol zeroing Implement the public API for vol zeroing Define wire protocol format for vol zeroing Implement RPC client for vol zeroing Implement the remote dispatch bits of vol zeroing Simplified version of volume zeroing based on feedback from the list. Virsh support for vol zeroing daemon/remote.c | 32 +++++ daemon/remote_dispatch_args.h | 1 + daemon/remote_dispatch_prototypes.h | 8 ++ daemon/remote_dispatch_table.h | 5 + include/libvirt/libvirt.h.in | 2 + src/driver.h | 5 + src/esx/esx_storage_driver.c | 39 +------ src/libvirt.c | 47 ++++++++ src/libvirt_public.syms | 1 + src/remote/remote_driver.c | 27 +++++ src/remote/remote_protocol.c | 11 ++ src/remote/remote_protocol.h | 9 ++ src/remote/remote_protocol.x | 8 +- src/storage/storage_driver.c | 218 +++++++++++++++++++++++++++++++++++ tools/virsh.c | 42 +++++++ 15 files changed, 418 insertions(+), 37 deletions(-)

7 27

[libvirt] [PATCH] Add TCK script for verifying disk based storage pools
by Daniel P. Berrange 04 Mar '10

04 Mar '10

This adds a reasonably comprehensive test script for working with disk based storage pools. It checks - Prevent re-formating an existing partition table (currently missing this check in libvirt) - Formating of a partiton table - Wiping a partition table (currently missing in libvirt) - Creating multiple storage volumes - Deleteing storage volumes * conf/default.cfg: add config entry for specifying host block devices that a test script can use * lib/Sys/Virt/TCK.pm: Add API to get a host block device * lib/Sys/Virt/TCK/StoragePoolBuilder.pm: Fix typo in device XML formatting. Allow pool source format to be specified * scripts/storage/110-disk-pool.t: Test which creates a storage pool, creates some volumes, then deletes it all again. --- conf/default.cfg | 9 ++ lib/Sys/Virt/TCK.pm | 7 ++ lib/Sys/Virt/TCK/StoragePoolBuilder.pm | 13 +++- scripts/storage/110-disk-pool.t | 127 ++++++++++++++++++++++++++++++++ 4 files changed, 155 insertions(+), 1 deletions(-) create mode 100644 scripts/storage/110-disk-pool.t diff --git a/conf/default.cfg b/conf/default.cfg index 5b85cbc..288540f 100644 --- a/conf/default.cfg +++ b/conf/default.cfg @@ -127,3 +127,12 @@ host_pci_devices = ( # } # Can list more than one PCI device if many are available ) + +# List of host block devices that the test suite can trash all +# data on, for purposes of testing. You have been warned ! +# Only full block devices allowed, no single partitions, since +# the test suite itself needs to create partitions. +# The disks should be at *least* 512 MB in size +host_block_devices = ( +# /dev/vdb +) diff --git a/lib/Sys/Virt/TCK.pm b/lib/Sys/Virt/TCK.pm index f101937..d32b03d 100644 --- a/lib/Sys/Virt/TCK.pm +++ b/lib/Sys/Virt/TCK.pm @@ -805,4 +805,11 @@ sub get_host_pci_device { return ($domain, $bus, $slot, $function); } +sub get_host_block_device { + my $self = shift; + my $devindex = @_ ? shift : 0; + + return $self->config("host_block_devices/[$devindex]", undef); +} + 1; diff --git a/lib/Sys/Virt/TCK/StoragePoolBuilder.pm b/lib/Sys/Virt/TCK/StoragePoolBuilder.pm index f92b21a..c9ef2dd 100644 --- a/lib/Sys/Virt/TCK/StoragePoolBuilder.pm +++ b/lib/Sys/Virt/TCK/StoragePoolBuilder.pm @@ -77,6 +77,14 @@ sub target { return $self; } +sub format { + my $self = shift; + + $self->{format} = shift; + + return $self; +} + sub as_xml { my $self = shift; @@ -98,7 +106,7 @@ sub as_xml { } if ($self->{source}->{device}) { foreach my $dev (@{$self->{source}->{device}}) { - $w->emptyTag("dev", path => $dev); + $w->emptyTag("device", path => $dev); } } if ($self->{source}->{adapter}) { @@ -107,6 +115,9 @@ sub as_xml { if ($self->{source}->{name}) { $w->dataElement("name", $self->{source}->{name}); } + if ($self->{format}) { + $w->emptyTag("format", type => $self->{format}); + } $w->endTag("source"); $w->startTag("target"); diff --git a/scripts/storage/110-disk-pool.t b/scripts/storage/110-disk-pool.t new file mode 100644 index 0000000..faa13c5 --- /dev/null +++ b/scripts/storage/110-disk-pool.t @@ -0,0 +1,127 @@ +# -*- perl -*- +# +# Copyright (C) 2009-2010 Red Hat +# Copyright (C) 2009-2010 Daniel P. Berrange +# +# This program is free software; You can redistribute it and/or modify +# it under the GNU General Public License as published by the Free +# Software Foundation; either version 2, or (at your option) any +# later version +# +# The file "LICENSE" distributed along with this file provides full +# details of the terms and conditions +# + +=pod + +=head1 NAME + +storage/110-disk-pool.t - test disk storage pools + +=head1 DESCRIPTION + +The test case validates that it is possible to use all core +functions of the disk pool type. + +=cut + +use strict; +use warnings; + +use Test::More tests => 20; + +use Sys::Virt::TCK; +use Test::Exception; +use File::stat; + +my $tck = Sys::Virt::TCK->new(); +my $conn = eval { $tck->setup(); }; +BAIL_OUT "failed to setup test harness: $@" if $@; +END { + $tck->cleanup if $tck; +} + + +SKIP: { + my $dev = $tck->get_host_block_device(); + + skip "no host block device available", 20 unless defined $dev; + + # Blow away partition table (if any) + open DEV, ">$dev" or die "cannot write to $dev: $!"; + my $data = " "x 512; + + print DEV $data; + close DEV or die "cannot save $dev: $!"; + + my $poolxml = $tck->generic_pool("disk", "tck") + ->source_device($dev) + ->format("dos") + ->target("/dev/") + ->as_xml; + + diag "Defining persistent storage pool $poolxml"; + my $pool; + ok_pool(sub { $pool = $conn->define_storage_pool($poolxml) }, "define persistent storage pool"); + + # Since we blew away the partition table we should not be able to + # start the pool yet + ok_error(sub { $pool->create }, "unable to start un-built storage pool"); + + lives_ok(sub { $pool->build(0) }, "built storage pool"); + + # We should get an error if trying to build a pool which already + # has a partition table written. + ok_error(sub { $pool->build(0) }, "prevent rebuilding existing storage pool"); + + lives_ok(sub { $pool->create }, "started storage pool"); + + my @vols = $pool->list_volumes(); + + is($#vols, -1, "no storage volumes in new pool"); + + my $poolinfo = $pool->get_info(); + + ok($poolinfo->{available} > 1000, "there is some space available in the pool"); + + my $volbase = $dev; + $volbase =~ s,/dev/,,; + + my $vol1xml = $tck->generic_volume($volbase . "1", undef, 1024*1024*256)->as_xml; + my $vol2xml = $tck->generic_volume($volbase . "2", undef, 1024*1024*64)->as_xml; + + diag "Vol $vol1xml $vol2xml"; + my $vol1; + my $vol2; + ok_volume(sub { $vol1 = $pool->create_volume($vol1xml) }, "create disk partition"); + ok_volume(sub { $vol2 = $pool->create_volume($vol2xml) }, "create disk partition"); + + for my $vol (($vol1, $vol2)) { + my $path = xpath($vol, "string(/volume/target/path)"); + my $st = stat($path); + + ok($st, "path $path exists"); + } + + @vols = $pool->list_volumes(); + is($#vols, 1, "two storage volumes in pool"); + + lives_ok(sub { $vol1->delete(0) }, "deleted volume"); + + @vols = $pool->list_volumes(); + is($#vols, 0, "one storage volume in pool"); + + lives_ok(sub { $vol2->delete(0) }, "deleted volume"); + + @vols = $pool->list_volumes(); + is($#vols, -1, "zero storage volume in pool"); + + lives_ok(sub { $pool->destroy() }, "destroyed pool"); + lives_ok(sub { $pool->delete(0) }, "deleted pool"); + + # Since we blew away the partition table we should not be able to + # start the pool anymore + ok_error(sub { $pool->create }, "unable to start un-built storage pool"); + + lives_ok(sub { $pool->undefine() }, "undefined pool"); +} -- 1.6.5.2

3 3

[libvirt] Problem building on Fedora Rawhide
by Laine Stump 04 Mar '10

04 Mar '10

I just checked out a clean tree on a Fedora Rawhide system (up to date as of today), and ran "./autogen.sh && make", and ended up with the following error. Am I missing something, or is configure missing something? make[3]: Entering directory `/home/laine/devel/libvirt/daemon' CC libvirtd-event.o CC libvirtd-libvirtd.o CC libvirtd-remote.o CC libvirtd-dispatch.o CC libvirtd-stream.o CC libvirtd-remote_protocol.o CCLD libvirtd ../src/.libs/libvirt.so: undefined reference to `g_pfnGetFunctions' ../src/.libs/libvirt.so: undefined reference to `g_pVBoxFuncs' ../src/.libs/libvirt.so: undefined reference to `VBoxCGlueInit'

1 0

[libvirt] [PATCH] Fix parser checking of storage pool device
by Daniel P. Berrange 04 Mar '10

04 Mar '10

The storage backend implementations all presume that the XML parser is validating correctness of the source specification. The check for a source device was lost at some point. This allowed for a potential crash in the disk backend. Re-introduce the sanity check * src/conf/storage_conf.c: Re-add check for source device --- src/conf/storage_conf.c | 9 +++++++++ 1 files changed, 9 insertions(+), 0 deletions(-) diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c index 19a1db9..b0f326f 100644 --- a/src/conf/storage_conf.c +++ b/src/conf/storage_conf.c @@ -689,6 +689,15 @@ virStoragePoolDefParseXML(xmlXPathContextPtr ctxt) { } } + /* If DEVICE is the only source type, then its required */ + if (options->flags == VIR_STORAGE_POOL_SOURCE_DEVICE) { + if (!ret->source.ndevice) { + virStorageReportError(VIR_ERR_XML_ERROR, + "%s", _("missing storage pool source device name")); + goto cleanup; + } + } + if ((ret->target.path = virXPathString("string(./target/path)", ctxt)) == NULL) { virStorageReportError(VIR_ERR_XML_ERROR, "%s", _("missing storage pool target path")); -- 1.6.5.2

3 2

[libvirt] [PATCH] Fix mis-leading error message in pool delete API
by Daniel P. Berrange 04 Mar '10

04 Mar '10

When trying to delete a pool the error message claimed the volume could not be deleted. * src/storage/storage_driver.c: Error message referred to volumes instead of pools --- src/storage/storage_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c index 6b1045a..f3be6de 100644 --- a/src/storage/storage_driver.c +++ b/src/storage/storage_driver.c @@ -822,7 +822,7 @@ storagePoolDelete(virStoragePoolPtr obj, if (!backend->deletePool) { virStorageReportError(VIR_ERR_NO_SUPPORT, - "%s", _("pool does not support volume delete")); + "%s", _("pool does not support pool deletion")); goto cleanup; } if (backend->deletePool(obj->conn, pool, flags) < 0) -- 1.6.5.2

3 2

[libvirt] [PATCH 0/2] virsh: consistent N_() usage
by Eric Blake 04 Mar '10

04 Mar '10

As discussed here[1], we aren't being very smart about the use of gettext_noop(). In particular, most packages use N_() to mean gettext_noop rather than a synonym to _(); furthermore, our current definition of N_() meant that we were effectively calling gettext(gettext("string")), which has undefined results. Patch 1/2 also fixes a potential crash due to a translation bug, by telling xgettext to enforce that the translation of a string later handed to printf does not violate printf constraints. [1] https://www.redhat.com/archives/libvir-list/2010-March/msg00197.html [PATCH 1/2] virsh: fix existing N_ uses [PATCH 2/2] virsh: use N_ rather than gettext_noop .../0008-Step-8-of-8-Add-virsh-support.patch | 21 +- src/internal.h | 2 +- tools/virsh.c | 895 ++++++++++---------- 3 files changed, 459 insertions(+), 459 deletions(-)

1 2

[libvirt] [PATCH] Add a define for NFS_SUPER_MAGIC
by Chris Lalancette 04 Mar '10

04 Mar '10

Commit 3c12a67b766cce51b47861ccde2be41de369f832 added a dependency on the NFS_SUPER_MAGIC macro, which is defined in linux/magic.h. Unfortunately linux/magic.h is not available in RHEL-5, and causes a compile error. Just define it locally, since this is something that can't change. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/qemu/qemu_driver.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index ccb9ef9..efc3be1 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -49,8 +49,10 @@ #ifdef __linux__ #include <sys/vfs.h> -#include <linux/magic.h> -#endif +#ifndef NFS_SUPER_MAGIC +#define NFS_SUPER_MAGIC 0x6969 +#endif /* NFS_SUPER_MAGIC */ +#endif /* __linux__ */ #include "virterror_internal.h" #include "logging.h" -- 1.6.6.1

4 5

[libvirt] [PATCH] util: ensure safe{read, write, zero} return is checked
by Eric Blake 04 Mar '10

04 Mar '10

Based on a warning from coverity. The safe* functions guarantee complete transactions on success, but don't guarantee freedom from failure. * src/util/util.h (saferead, safewrite, safezero): Add ATTRIBUTE_RETURN_CHECK. * src/remote/remote_driver.c (remoteIO, remoteIOEventLoop): Ignore some failures. (remoteIOReadBuffer): Adjust error messages on read failure. * daemon/event.c (virEventHandleWakeup): Ignore read failure. --- Coverity claimed that safewrite was checked 37 out of 46 times, but only flagged one instance of a CHECKED_RETURN report, against remote_driver.c. I didn't spot the 9 unchecked calls that coverity's numbers would imply. Meanwhile, adding the attribute ensures that we don't forget this in the future. Please double-check if my use of ignore_value on some of the previously unchecked calls makes sense, or if we need a more invasive patch to deal with real failure. daemon/event.c | 5 +++-- src/remote/remote_driver.c | 13 +++++++++---- src/util/util.h | 9 ++++++--- 3 files changed, 18 insertions(+), 9 deletions(-) diff --git a/daemon/event.c b/daemon/event.c index 2218a3e..36e9dd3 100644 --- a/daemon/event.c +++ b/daemon/event.c @@ -1,8 +1,8 @@ /* * event.c: event loop for monitoring file handles * + * Copyright (C) 2007, 2010 Red Hat, Inc. * Copyright (C) 2007 Daniel P. Berrange - * Copyright (C) 2007 Red Hat, Inc. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -35,6 +35,7 @@ #include "event.h" #include "memory.h" #include "util.h" +#include "ignore-value.h" #define EVENT_DEBUG(fmt, ...) DEBUG(fmt, __VA_ARGS__) @@ -630,7 +631,7 @@ static void virEventHandleWakeup(int watch ATTRIBUTE_UNUSED, { char c; virEventLock(); - saferead(fd, &c, sizeof(c)); + ignore_value (saferead(fd, &c, sizeof(c))); virEventUnlock(); } diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index def4617..47d8c56 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -7809,7 +7809,9 @@ remoteIOReadBuffer(virConnectPtr conn, char errout[1024] = "\0"; if (priv->errfd != -1) { - saferead(priv->errfd, errout, sizeof(errout)); + if (saferead(priv->errfd, errout, sizeof(errout)) < 0) { + strncpy (errout, _("unknown reason"), sizeof errout); + } } virReportSystemError(errno, @@ -7818,7 +7820,9 @@ remoteIOReadBuffer(virConnectPtr conn, } else { char errout[1024] = "\0"; if (priv->errfd != -1) { - saferead(priv->errfd, errout, sizeof(errout)); + if (saferead(priv->errfd, errout, sizeof(errout)) < 0) { + strncpy (errout, _("unknown reason"), sizeof errout); + } } errorf (in_open ? NULL : conn, @@ -8431,7 +8435,8 @@ remoteIOEventLoop(virConnectPtr conn, if (fds[1].revents) { DEBUG0("Woken up from poll by other thread"); - saferead(priv->wakeupReadFD, &ignore, sizeof(ignore)); + ignore_value (saferead(priv->wakeupReadFD, &ignore, + sizeof(ignore))); } if (ret < 0) { @@ -8575,7 +8580,7 @@ remoteIO(virConnectPtr conn, priv->waitDispatch = thiscall; /* Force other thread to wakup from poll */ - safewrite(priv->wakeupSendFD, &ignore, sizeof(ignore)); + ignore_value (safewrite(priv->wakeupSendFD, &ignore, sizeof(ignore))); DEBUG("Going to sleep %d %p %p", thiscall->proc_nr, priv->waitDispatch, thiscall); /* Go to sleep while other thread is working... */ diff --git a/src/util/util.h b/src/util/util.h index cc05abe..34b77fa 100644 --- a/src/util/util.h +++ b/src/util/util.h @@ -2,6 +2,7 @@ /* * utils.h: common, generic utility functions * + * Copyright (C) 2010 Red Hat, Inc. * Copyright (C) 2006, 2007 Binary Karma * Copyright (C) 2006 Shuveb Hussain * @@ -31,9 +32,11 @@ #include <sys/select.h> #include <sys/types.h> -int saferead(int fd, void *buf, size_t count); -ssize_t safewrite(int fd, const void *buf, size_t count); -int safezero(int fd, int flags, off_t offset, off_t len); +int saferead(int fd, void *buf, size_t count) ATTRIBUTE_RETURN_CHECK; +ssize_t safewrite(int fd, const void *buf, size_t count) + ATTRIBUTE_RETURN_CHECK; +int safezero(int fd, int flags, off_t offset, off_t len) + ATTRIBUTE_RETURN_CHECK; enum { VIR_EXEC_NONE = 0, -- 1.6.6.1

1 1

[libvirt] [PATCH] Fix typo in QEMU migration command name
by Daniel P. Berrange 04 Mar '10

04 Mar '10

The QMP code was running query-migration instead of query-migrate. This doesn't work so well * src/qemu/qemu_monitor_json.c: s/query-migration/query-migrate/ --- src/qemu/qemu_monitor_json.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 3a94dd0..9c0245c 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -1108,7 +1108,7 @@ int qemuMonitorJSONGetMigrationStatus(qemuMonitorPtr mon, unsigned long long *total) { int ret; - virJSONValuePtr cmd = qemuMonitorJSONMakeCommand("query-migration", + virJSONValuePtr cmd = qemuMonitorJSONMakeCommand("query-migrate", NULL); virJSONValuePtr reply = NULL; -- 1.6.5.2

2 1

[libvirt] [PATCH] Add TCK script for host PCI device hotplug/unplug
by Daniel P. Berrange 04 Mar '10

04 Mar '10

This adds support for testing host PCI device hotplug/unplug in libvirt drivers. This requires that the host have a usable IOMMU & hardware virt, and that the device in question can be reset safely. Since the TCK can't assume there are any host PCI devices that can be safely messed around with, the person running the test suite must first list one or more devices in the config file. If no devices are listed, the test will automatically skip all parts, rather than failing. * conf/default.cfg: Config entry to specify PCI devices that the TCK can mess around with * lib/Sys/Virt/TCK.pm: API for getting a host PCI device from the config * scripts/domain/250-pci-host-hotplug.t: Test case for PCI device hotplug/unplug. --- conf/default.cfg | 15 ++++ lib/Sys/Virt/TCK.pm | 18 +++++ scripts/domain/250-pci-host-hotplug.t | 117 +++++++++++++++++++++++++++++++++ 3 files changed, 150 insertions(+), 0 deletions(-) create mode 100644 scripts/domain/250-pci-host-hotplug.t diff --git a/conf/default.cfg b/conf/default.cfg index f9c2b5c..5b85cbc 100644 --- a/conf/default.cfg +++ b/conf/default.cfg @@ -112,3 +112,18 @@ host_usb_devices = ( # } # Can list more than one USB device if many are available ) + + +# Host PCI devices that the test suite can safely mess around with +# without risk of breaking the host OS. Using a NIC device is the +# most reliable option. Definitely don't try listing a VGA device. +host_pci_devices = ( +# Slot is only required entry, all others default to 0 +# { +# domain = 0000 +# bus = 00 +# slot = 07 +# function = 0 +# } +# Can list more than one PCI device if many are available +) diff --git a/lib/Sys/Virt/TCK.pm b/lib/Sys/Virt/TCK.pm index ce03935..f101937 100644 --- a/lib/Sys/Virt/TCK.pm +++ b/lib/Sys/Virt/TCK.pm @@ -787,4 +787,22 @@ sub get_host_usb_device { return ($bus, $device, $vendor, $product); } +sub get_host_pci_device { + my $self = shift; + my $devindex = @_ ? shift : 0; + + my $devs = $self->config("host_pci_devices", []); + + if ($devindex > $#{$devs}) { + return (); + } + + my $domain = $self->config("host_pci_devices/[$devindex]/domain", 0); + my $bus = $self->config("host_pci_devices/[$devindex]/bus", 0); + my $slot = $self->config("host_pci_devices/[$devindex]/slot"); + my $function = $self->config("host_pci_devices/[$devindex]/function", 0); + + return ($domain, $bus, $slot, $function); +} + 1; diff --git a/scripts/domain/250-pci-host-hotplug.t b/scripts/domain/250-pci-host-hotplug.t new file mode 100644 index 0000000..9ccd036 --- /dev/null +++ b/scripts/domain/250-pci-host-hotplug.t @@ -0,0 +1,117 @@ +# -*- perl -*- +# +# Copyright (C) 2009 Red Hat +# Copyright (C) 2009 Daniel P. Berrange +# +# This program is free software; You can redistribute it and/or modify +# it under the GNU General Public License as published by the Free +# Software Foundation; either version 2, or (at your option) any +# later version +# +# The file "LICENSE" distributed along with this file provides full +# details of the terms and conditions +# + +=pod + +=head1 NAME + +domain/250-pci-host-hotplug.t - verify hot plug & unplug of a host PCI device + +=head1 DESCRIPTION + +The test case validates that it is possible to hotplug a pci +host device to a running domain, and then unplug it again. +This requires that the TCK configuration file have at least +one host PCI device listed. + +This first searches for the node device, then detachs it from +the host OS. Next it performs a reset. Then does the hotplug +and unplug, before finally reattaching to the host. + +=cut + +use strict; +use warnings; + +use Test::More tests => 10; + +use Sys::Virt::TCK; +use Test::Exception; + +my $tck = Sys::Virt::TCK->new(); +my $conn = eval { $tck->setup(); }; +BAIL_OUT "failed to setup test harness: $@" if $@; +END { + $tck->cleanup if $tck; +} + + +my $xml = $tck->generic_domain("tck")->as_xml; + +diag "Creating a new transient domain"; +my $dom; +ok_domain(sub { $dom = $conn->create_domain($xml) }, "created transient domain object"); + + +my ($domain, $bus, $slot, $function) = $tck->get_host_pci_device(); + +SKIP: { + # Must have one item be non-zero + unless ($domain || $bus || $slot || $function) { + skip "No host pci device available in configuration file", 9; + } + + my $nodedev; + my @devs = $conn->list_node_devices("pci"); + foreach my $dev (@devs) { + my $thisxml = $dev->get_xml_description(); + + my $xp = XML::XPath->new(xml => $thisxml); + + #warn $thisxml; + my $ndomain = $xp->find("string(/device/capability[\@type='pci']/domain[text()])")->value(); + my $nbus = $xp->find("string(/device/capability[\@type='pci']/bus[text()])")->value(); + my $nslot = $xp->find("string(/device/capability[\@type='pci']/slot[text()])")->value(); + my $nfunction = $xp->find("string(/device/capability[\@type='pci']/function[text()])")->value(); + + if ($ndomain == $domain && + $nbus == $bus && + $nslot == $slot && + $nfunction == $function) { + $nodedev = $dev; + last; + } + } + + ok(defined $nodedev, "found PCI device $domain:$bus:$slot.$function on host"); + + lives_ok(sub { $nodedev->dettach() }, "detached device from host OS"); + lives_ok(sub { $nodedev->reset() }, "reset the host PCI device"); + + my $devxml = + "<hostdev mode='subsystem' type='pci' managed='no'>\n" . + " <source>\n" . + " <address domain='$domain' bus='$bus' slot='$slot' function='$function'/>\n" . + " </source>\n" . + "</hostdev>\n"; + + my $initialxml = $dom->get_xml_description; + + diag "Attaching the new dev $devxml"; + lives_ok(sub { $dom->attach_device($devxml); }, "PCI dev has been attached"); + + my $newxml = $dom->get_xml_description; + ok($newxml =~ m|<hostdev|, "new XML has extra PCI dev present"); + + diag "Detaching the new dev $devxml"; + lives_ok(sub { $dom->detach_device($devxml); }, "PCI dev has been detached"); + + lives_ok(sub { $nodedev->reset() }, "reset the host PCI device"); + lives_ok(sub { $nodedev->reattach() }, "reattached device to host OS"); + + my $finalxml = $dom->get_xml_description; + + is($initialxml, $finalxml, "final XML has removed the disk") +} + -- 1.6.5.2

2 1

[libvirt] [PATCH] uml: avoid crash on partial read
by Eric Blake 04 Mar '10

04 Mar '10

Coverity detected a potential dereference of uninitialized memory if recvfrom got cut short. * src/uml/uml_driver.c (umlMonitorCommand): Validate complete read prior to dereferencing res. --- The patch borrows ideas from macvtap.c, the only other file in libvirt that currently uses recvfrom. I did not analyze whether this is a security hole, where a malicious UDP packet could intentionally force the dereferencing of uninitialized memory to misbehave in a controlled manner. src/uml/uml_driver.c | 14 ++++++++++++-- 1 files changed, 12 insertions(+), 2 deletions(-) diff --git a/src/uml/uml_driver.c b/src/uml/uml_driver.c index bbea429..eec239f 100644 --- a/src/uml/uml_driver.c +++ b/src/uml/uml_driver.c @@ -733,14 +733,24 @@ static int umlMonitorCommand(virConnectPtr conn, } do { + ssize_t nbytes; addrlen = sizeof(addr); - if (recvfrom(priv->monitor, &res, sizeof res, 0, - (struct sockaddr *)&addr, &addrlen) < 0) { + nbytes = recvfrom(priv->monitor, &res, sizeof res, 0, + (struct sockaddr *)&addr, &addrlen) < 0; + if (nbytes < 0) { + if (errno == EAGAIN || errno == EINTR) + continue; virReportSystemError(errno, _("cannot read reply %s"), cmd); goto error; } + if (nbytes < sizeof res) { + virReportSystemError(errno, + _("incomplete reply %s"), + cmd); + goto error; + } if (VIR_REALLOC_N(retdata, retlen + res.length) < 0) { virReportOOMError(); -- 1.6.6.1

3 10

[libvirt] [PATCH] don't let a bogus packet trigger over-allocation and segfault
by Jim Meyering 04 Mar '10

04 Mar '10

Another not-really-urgent fix: >From 4b56f03dee82657be3af5c79c826ae3091fbf522 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 3 Mar 2010 16:50:02 +0100 Subject: [PATCH] don't let a bogus packet trigger over-allocation and segfault * src/xen/proxy_internal.c (xenProxyDomainDumpXML): An invalid packet could include a too-large "ans.len" value, which would make us allocate too much memory and then copy data from beyond the end of "ans", possibly evoking a segfault. Ensure that the value we use is no larger than the remaining portion of "ans". Also, change unnecessary memmove to memcpy (src and dest obviously do not overlap, so no need to use memmove). --- src/xen/proxy_internal.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) diff --git a/src/xen/proxy_internal.c b/src/xen/proxy_internal.c index 8e1c226..bd234ec 100644 --- a/src/xen/proxy_internal.c +++ b/src/xen/proxy_internal.c @@ -978,26 +978,27 @@ xenProxyDomainDumpXML(virDomainPtr domain, int flags ATTRIBUTE_UNUSED) memset(&req, 0, sizeof(req)); req.command = VIR_PROXY_DOMAIN_XML; req.data.arg = domain->id; req.len = sizeof(req); ret = xenProxyCommand(domain->conn, &req, &ans, 0); if (ret < 0) { return(NULL); } - if (ans.len <= sizeof(virProxyPacket)) { + if (ans.len <= sizeof(virProxyPacket) + || ans.len > sizeof (ans) - sizeof(virProxyPacket)) { virProxyError(domain->conn, VIR_ERR_OPERATION_FAILED, __FUNCTION__); return (NULL); } xmllen = ans.len - sizeof(virProxyPacket); if (VIR_ALLOC_N(xml, xmllen+1) < 0) { virReportOOMError(); return NULL; } - memmove(xml, &ans.extra.dinfo, xmllen); + memcpy(xml, &ans.extra.dinfo, xmllen); xml[xmllen] = '\0'; return(xml); } /** * xenProxyDomainGetOSType: * @domain: a domain object -- 1.7.0.1.464.g0adc7

2 3

[libvirt] [PATCH] virFileReadLimFD: diagnose maxlen <= 0, rather than passing it on...
by Jim Meyering 04 Mar '10

04 Mar '10

Not urgent. >From 35a24209934cb983ca52e74f580b3ea963d92f57 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 3 Mar 2010 11:42:11 +0100 Subject: [PATCH] virFileReadLimFD: diagnose maxlen <= 0, rather than passing it on... to saferead_lim, which interprets it as a size_t. * src/util/util.c (virFileReadLimFD): Do not malfunction when maxlen < -1. Return -1,EINVAL in that case. Handle maxlen==0 in the same manner. --- src/util/util.c | 11 +++++++++-- 1 files changed, 9 insertions(+), 2 deletions(-) diff --git a/src/util/util.c b/src/util/util.c index 34c585d..7a3a3c4 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -1030,10 +1030,17 @@ saferead_lim (int fd, size_t max_len, size_t *length) /* A wrapper around saferead_lim that maps a failure due to exceeding the maximum size limitation to EOVERFLOW. */ -int virFileReadLimFD(int fd, int maxlen, char **buf) +int +virFileReadLimFD(int fd, int maxlen, char **buf) { size_t len; - char *s = saferead_lim (fd, maxlen+1, &len); + char *s; + + if (maxlen <= 0) { + errno = EINVAL; + return -1; + } + s = saferead_lim (fd, maxlen+1, &len); if (s == NULL) return -1; if (len > maxlen || (int)len != len) { -- 1.7.0.1.464.g0adc7

2 1

[libvirt] [PATCH] qemu restore: don't let corrupt input provoke unwarranted OOM
by Jim Meyering 04 Mar '10

04 Mar '10

As the log says, we must not use a just-read (and unverified) header.xml_len as the size of buffer to allocate. While looking at this, I noticed a small problem with virFileReadLimFD. Passing it a limit < -1 would cause int->size_t trouble with saferead_lim. Separate patch coming up. >From a4906ba24b576f3219234c519c88d102df2173b2 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 3 Mar 2010 11:27:16 +0100 Subject: [PATCH] qemu restore: don't let corrupt input provoke unwarranted OOM * src/qemu/qemu_driver.c (qemudDomainRestore): A corrupt save file (in particular, a too-large header.xml_len value) would cause an unwarranted out-of-memory error. Do not trust the just-read header.xml_len. Instead, merely use that as a hint, and read/allocate up to that number of bytes from the file. Also verify that header.xml_len is positive; if it were negative, passing it to virFileReadLimFD could cause trouble. --- src/qemu/qemu_driver.c | 7 ++++--- 1 files changed, 4 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 7f7c459..5c9d003 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -4725,44 +4725,45 @@ static int qemudDomainRestore(virConnectPtr conn, if (saferead(fd, &header, sizeof(header)) != sizeof(header)) { qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", _("failed to read qemu header")); goto cleanup; } if (memcmp(header.magic, QEMUD_SAVE_MAGIC, sizeof(header.magic)) != 0) { qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", _("image magic is incorrect")); goto cleanup; } if (header.version > QEMUD_SAVE_VERSION) { qemuReportError(VIR_ERR_OPERATION_FAILED, _("image version is not supported (%d > %d)"), header.version, QEMUD_SAVE_VERSION); goto cleanup; } - if (VIR_ALLOC_N(xml, header.xml_len) < 0) { - virReportOOMError(); + if (header.xml_len <= 0) { + qemuReportError(VIR_ERR_OPERATION_FAILED, + _("invalid XML length: %d"), header.xml_len); goto cleanup; } - if (saferead(fd, xml, header.xml_len) != header.xml_len) { + if (virFileReadLimFD(fd, header.xml_len, &xml) != header.xml_len) { qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", _("failed to read XML")); goto cleanup; } /* Create a domain from this XML */ if (!(def = virDomainDefParseString(driver->caps, xml, VIR_DOMAIN_XML_INACTIVE))) { qemuReportError(VIR_ERR_OPERATION_FAILED, "%s", _("failed to parse XML")); goto cleanup; } if (virDomainObjIsDuplicate(&driver->domains, def, 1) < 0) goto cleanup; if (!(vm = virDomainAssignDef(driver->caps, &driver->domains, def))) { -- 1.7.0.1.464.g0adc7

2 1

[libvirt] [PATCH 0/1] - for comments/help only, nearly working root-squash-domain-save
by Laine Stump 04 Mar '10

04 Mar '10

(This patch requires that you first apply the virFileOperation patches that I sent to the list yesterday.) I'm hoping someone else's keen eye can see what I'm failing to see here. The patch following this message makes domain save "work" on root-squash NFS (both the libvirt header and the QEMU state are saved to the file), but the files that it saves don't restore properly, even when simply saving to a local directory w/o forking - the guest screen comes up black, or reboots immediately, or some other such thing. I've looked at the data files and the code, and can't see what difference is causing this - at least the header written by libvirt seems to be the same, and the part written by qemu starts in the same place and looks similar. If anyone's got an idea, please let me know! I'd like to get this figured out and a working patch submitted before the day is done.

3 6

[libvirt] [PATCH 0/4] Misc hotplug fixes for QEMU
by Daniel P. Berrange 04 Mar '10

04 Mar '10

This is a series of fixes for hotplug in QEMU

1 4

[libvirt] [PATCH 0/1] Support domain restore from root-squash NFS
by Laine Stump 04 Mar '10

04 Mar '10

Most of the explanation is in the commit message with the patch proper. When reveiwing please take pains to point out any places where I may not be diligent enough cleaning up loose ends, especially in terms of assuring that the child process always terminates properly no matter what happens. The (unrelated) race condition in qemudDomainRestore() remains - I solved it locally by putting a sleep(10) just before qemuMonitorStartCPUs(), but it needs something with a bit higher comfort factor ;-)

5 11

[libvirt] [PATCH] virsh: silence compiler warning
by Eric Blake 04 Mar '10

04 Mar '10

gcc warns: virsh.c:1879: warning: '0' flag ignored with '-' flag in gnu_printf format * tools/virsh.c (cmdDomjobinfo): Delete useless flag. --- tools/virsh.c | 18 +++++++++--------- 1 files changed, 9 insertions(+), 9 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index a6a637d..65487ed 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -1876,27 +1876,27 @@ cmdDomjobinfo(vshControl *ctl, const vshCmd *cmd) vshPrint(ctl, "%-17s %-12llu ms\n", _("Time remaining:"), info.timeRemaining); if (info.dataTotal || info.dataRemaining || info.dataProcessed) { val = prettyCapacity(info.dataProcessed, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("Data processed:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("Data processed:"), val, unit); val = prettyCapacity(info.dataRemaining, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("Data remaining:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("Data remaining:"), val, unit); val = prettyCapacity(info.dataTotal, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("Data total:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("Data total:"), val, unit); } if (info.memTotal || info.memRemaining || info.memProcessed) { val = prettyCapacity(info.memProcessed, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("Memory processed:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("Memory processed:"), val, unit); val = prettyCapacity(info.memRemaining, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("Memory remaining:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("Memory remaining:"), val, unit); val = prettyCapacity(info.memTotal, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("Memory total:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("Memory total:"), val, unit); } if (info.fileTotal || info.fileRemaining || info.fileProcessed) { val = prettyCapacity(info.fileProcessed, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("File processed:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("File processed:"), val, unit); val = prettyCapacity(info.fileRemaining, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("File remaining:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("File remaining:"), val, unit); val = prettyCapacity(info.fileTotal, &unit); - vshPrint(ctl, "%-17s %-0.3lf %s\n", _("File total:"), val, unit); + vshPrint(ctl, "%-17s %-.3lf %s\n", _("File total:"), val, unit); } } else { ret = FALSE; -- 1.6.6.1

3 4

[libvirt] [PATCH] Add TCK script for host USB device hotplug/unplug
by Daniel P. Berrange 03 Mar '10

03 Mar '10

This adds support for testing host USB device hotplug/unplug in libvirt drivers. This currently fails in all known libvirt drivers :-) The QEMU driver only support the hotplug action, not unplug, hence the failure. Since the TCK can't assume there are any host USB devices that can be safely messed around with, the person running the test suite must first list one or more devices in the config file. If no devices are listed, the test will automatically skip all parts, rather than failing. * conf/default.cfg: Config entry to specify USB devices that the TCK can mess around with * lib/Sys/Virt/TCK.pm: API for getting a host USB device from the config * scripts/domain/240-usb-host-hotplug.t: Test case for USB device hotplug/unplug. --- conf/default.cfg | 13 +++++ lib/Sys/Virt/TCK.pm | 18 +++++++ scripts/domain/240-usb-host-hotplug.t | 89 +++++++++++++++++++++++++++++++++ 3 files changed, 120 insertions(+), 0 deletions(-) create mode 100644 scripts/domain/240-usb-host-hotplug.t diff --git a/conf/default.cfg b/conf/default.cfg index 556225e..f9c2b5c 100644 --- a/conf/default.cfg +++ b/conf/default.cfg @@ -99,3 +99,16 @@ kernels = ( # ostype = exe # } ) + +# Host USB devices that the test suite can safely mess around with +# without risk of breaking the host OS +host_usb_devices = ( +# Must list either vendor+product, or bus+dev, or both +# { +# vendor = 0x0627 +# product = 0x0001 +# bus = 001 +# device = 002 +# } +# Can list more than one USB device if many are available +) diff --git a/lib/Sys/Virt/TCK.pm b/lib/Sys/Virt/TCK.pm index 9e6b554..ce03935 100644 --- a/lib/Sys/Virt/TCK.pm +++ b/lib/Sys/Virt/TCK.pm @@ -769,4 +769,22 @@ sub xpath { return $xp->find($path); } +sub get_host_usb_device { + my $self = shift; + my $devindex = @_ ? shift : 0; + + my $devs = $self->config("host_usb_devices", []); + + if ($devindex > $#{$devs}) { + return (); + } + + my $bus = $self->config("host_usb_devices/[$devindex]/bus", undef); + my $device = $self->config("host_usb_devices/[$devindex]/device", undef); + my $vendor = $self->config("host_usb_devices/[$devindex]/vendor", undef); + my $product = $self->config("host_usb_devices/[$devindex]/product", undef); + + return ($bus, $device, $vendor, $product); +} + 1; diff --git a/scripts/domain/240-usb-host-hotplug.t b/scripts/domain/240-usb-host-hotplug.t new file mode 100644 index 0000000..4729300 --- /dev/null +++ b/scripts/domain/240-usb-host-hotplug.t @@ -0,0 +1,89 @@ +# -*- perl -*- +# +# Copyright (C) 2009 Red Hat +# Copyright (C) 2009 Daniel P. Berrange +# +# This program is free software; You can redistribute it and/or modify +# it under the GNU General Public License as published by the Free +# Software Foundation; either version 2, or (at your option) any +# later version +# +# The file "LICENSE" distributed along with this file provides full +# details of the terms and conditions +# + +=pod + +=head1 NAME + +domain/240-usb-host-hotplug.t - verify hot plug & unplug of a host USB device + +=head1 DESCRIPTION + +The test case validates that it is possible to hotplug a usb +host device to a running domain, and then unplug it again. +This requires that the TCK configuration file have at least +one host USB device listed. + +=cut + +use strict; +use warnings; + +use Test::More tests => 5; + +use Sys::Virt::TCK; +use Test::Exception; + +my $tck = Sys::Virt::TCK->new(); +my $conn = eval { $tck->setup(); }; +BAIL_OUT "failed to setup test harness: $@" if $@; +END { + $tck->cleanup if $tck; +} + + +my $xml = $tck->generic_domain("tck")->as_xml; + +diag "Creating a new transient domain"; +my $dom; +ok_domain(sub { $dom = $conn->create_domain($xml) }, "created transient domain object"); + + +my ($bus, $device, $vendor, $product) = $tck->get_host_usb_device(); + +SKIP: { + # Must have one of the pairs at least + unless (($bus && $device) || ($vendor && $product)) { + skip "No host usb device available in configuration file", 4; + } + + my $devxml = "<hostdev mode='subsystem' type='usb'>\n" . + "<source>\n"; + if ($bus && $device) { + $devxml .= "<address bus='$bus' device='$device'/>\n" + } + if ($vendor && $product) { + $devxml .= "<vendor id='$vendor'/>\n"; + $devxml .= "<product id='$product'/>\n"; + } + $devxml .= "</source>\n" . + "</hostdev>\n"; + + my $initialxml = $dom->get_xml_description; + + diag "Attaching the new dev $devxml"; + lives_ok(sub { $dom->attach_device($devxml); }, "USB dev has been attached"); + + my $newxml = $dom->get_xml_description; + ok($newxml =~ m|<hostdev|, "new XML has extra USB dev present"); + + diag "Detaching the new dev $devxml"; + lives_ok(sub { $dom->detach_device($devxml); }, "USB dev has been detached"); + + + my $finalxml = $dom->get_xml_description; + + is($initialxml, $finalxml, "final XML has removed the disk") +} + -- 1.6.5.2

1 0

[libvirt] [PATCH] web docs -- macvtap mode explanation
by Stefan Berger 03 Mar '10

03 Mar '10

This adds more information about the different macvtap device modes. Regards, Stefan

1 0

[libvirt] Xen Daemon error : internal error domain information incomplete, missing cpu_cap
by "J. Oriol Fitó" 03 Mar '10

03 Mar '10

Hi all, I'm using VirtMonitor on top of Xen and I'm wondering when the 'cpu_cap' feature will be available. Best regards, Oriol

1 0

[libvirt] [PATCH] esx: don't ignore failure on close
by Eric Blake 03 Mar '10

03 Mar '10

Another warning caught by coverity. Continue to perform best-effort closing and resource release, but warn the caller about the failure. * src/esx/esx_driver.c (esxClose): Return an error on failure to close. --- src/esx/esx_driver.c | 18 +++++++++++------- 1 files changed, 11 insertions(+), 7 deletions(-) diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index e125a09..45b389f 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -2,6 +2,7 @@ /* * esx_driver.c: core driver methods for managing VMware ESX hosts * + * Copyright (C) 2010 Red Hat, Inc. * Copyright (C) 2009, 2010 Matthias Bolte <matthias.bolte(a)googlemail.com> * Copyright (C) 2009 Maximilian Wilhelm <max(a)rfc2324.org> * @@ -559,16 +560,19 @@ static int esxClose(virConnectPtr conn) { esxPrivate *priv = conn->privateData; + int result = 0; - esxVI_EnsureSession(priv->host); - - esxVI_Logout(priv->host); + if (esxVI_EnsureSession(priv->host) < 0 || + esxVI_Logout(priv->host) < 0) { + result = -1; + } esxVI_Context_Free(&priv->host); if (priv->vCenter != NULL) { - esxVI_EnsureSession(priv->vCenter); - - esxVI_Logout(priv->vCenter); + if (esxVI_EnsureSession(priv->vCenter) < 0 || + esxVI_Logout(priv->vCenter) < 0) { + result = -1; + } esxVI_Context_Free(&priv->vCenter); } @@ -579,7 +583,7 @@ esxClose(virConnectPtr conn) conn->privateData = NULL; - return 0; + return result; } -- 1.6.6.1

3 2

[libvirt] [PATCH 1/2] Use AC_SEARCH_LIBS to find the library to use for dlopen().
by Diego Elio Pettenò 03 Mar '10

03 Mar '10

Instead of using AC_CHECK_LIB and hardcoding -ldl, search for the library needed to get dlopen() and then use the cached value. --- configure.ac | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/configure.ac b/configure.ac index 5b1eb5f..a3f28c4 100644 --- a/configure.ac +++ b/configure.ac @@ -1705,7 +1705,7 @@ if test "x$with_driver_modules" = "xyes" ; then old_libs="$LIBS" fail=0 AC_CHECK_HEADER([dlfcn.h],[],[fail=1]) - AC_CHECK_LIB([dl], [dlopen],[],[fail=1]) + AC_SEARCH_LIBS([dlopen], [dl], [], [fail=1]) test $fail = 1 && AC_MSG_ERROR([You must have dlfcn.h / dlopen() support to build driver modules]) @@ -1714,7 +1714,7 @@ if test "x$with_driver_modules" = "xyes" ; then fi if test "$with_driver_modules" = "yes"; then DRIVER_MODULES_CFLAGS="-export-dynamic" - DRIVER_MODULES_LIBS="-ldl" + DRIVER_MODULES_LIBS="$ac_cv_search_dlopen" AC_DEFINE_UNQUOTED([WITH_DRIVER_MODULES], 1, [whether to build drivers as modules]) fi AM_CONDITIONAL([WITH_DRIVER_MODULES], [test "$with_driver_modules" != "no"]) -- 1.7.0

4 5

[libvirt] [PATCH] HACKING: add a section on preprocessor conventions
by Eric Blake 03 Mar '10

03 Mar '10

* HACKING: Document recently-discussed style issues. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- Turn recent discussions: https://www.redhat.com/archives/libvir-list/2010-March/msg00058.html https://www.redhat.com/archives/libvir-list/2010-March/msg00120.html into policy, so I don't keep debating style points ;) Can wait till after 0.7.7 (particularly since the vararg macro series is also waiting). HACKING | 11 +++++++++++ 1 files changed, 11 insertions(+), 0 deletions(-) diff --git a/HACKING b/HACKING index be8725d..50b24ee 100644 --- a/HACKING +++ b/HACKING @@ -102,6 +102,17 @@ Usually they're in macro definitions or strings, and should be converted anyhow. +Preprocessor +============ +For variadic macros, stick with C99 syntax: + +#define vshPrint(_ctl, ...) fprintf(stdout, __VA_ARGS__) + +Use parenthesis when checking if a macro is defined: + +#if defined(HAVE_POSIX_FALLOCATE) && !defined(HAVE_FALLOCATE) + + C types ======= Use the right type. -- 1.6.6.1

2 1

[libvirt] Update on 0.7.7 release schedule
by Daniel Veillard 03 Mar '10

03 Mar '10

As you can have noticed we still have patches queued up and not yet pushed, plus added a new interface yesterday. I'm travelling all of this week, so my bandwidth is limited. Plus February is a short month and we skipped a bit for 0.7.6, so time to validate and push patches has been shorter than usual. So I suggest to still push patches today and tomorrow for features, and start the freeze on thusday evening, and delaying the release to mid of next week. But still try to keep the expected cycle for 0.7.8 with a freeze on March 19 and a release on the 26 :-) Hope this doesn't disrupt other people schedules ! 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 1

[libvirt] [PATCH] Fix detection of errors in QEMU device_add command
by Daniel P. Berrange 03 Mar '10

03 Mar '10

The code assumed that 'device_add' returned an empty string upon success. This is not true, it sometimes prints random debug info. THus we need to check for an explicit fail string * src/qemu/qemu_monitor_text.c: Fix error checking of the device_add monitor command --- src/qemu/qemu_monitor_text.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 7604ae8..c60e840 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -2122,9 +2122,11 @@ int qemuMonitorTextAddDevice(qemuMonitorPtr mon, goto cleanup; } - if (STRNEQ(reply, "")) { + /* If the command failed qemu prints: + * Could not add ... */ + if (strstr(reply, "Could not add ")) { qemuReportError(VIR_ERR_OPERATION_FAILED, - _("adding %s device failed: %s"), devicestr, reply); + _("adding %s device failed"), devicestr); goto cleanup; } -- 1.6.5.2

2 1

[libvirt] [PATCH] Fix safezero()
by Jiri Denemark 03 Mar '10

03 Mar '10

Various safezero() implementations used either -1, errno or -errno return values. This patch fixes them all to return -1 and set errno appropriately. There was also a bug in size parameter passed to safewrite() which could result in an attempt to write gigabytes out of a megabyte buffer. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/storage/storage_backend.c | 4 ++-- src/util/util.c | 16 +++++++++------- 2 files changed, 11 insertions(+), 9 deletions(-) diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index 3742493..849f01b 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -316,7 +316,7 @@ static int createRawFileOpHook(int fd, void *data) { if ((r = safezero(fd, 0, hdata->vol->allocation - remain, bytes)) != 0) { ret = errno; - virReportSystemError(r, _("cannot fill file '%s'"), + virReportSystemError(errno, _("cannot fill file '%s'"), hdata->vol->target.path); goto cleanup; } @@ -327,7 +327,7 @@ static int createRawFileOpHook(int fd, void *data) { if ((r = safezero(fd, 0, 0, remain)) != 0) { ret = errno; - virReportSystemError(r, _("cannot fill file '%s'"), + virReportSystemError(errno, _("cannot fill file '%s'"), hdata->vol->target.path); goto cleanup; } diff --git a/src/util/util.c b/src/util/util.c index cf7bba5..34c585d 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -146,11 +146,11 @@ int safezero(int fd, int flags ATTRIBUTE_UNUSED, off_t offset, off_t len) */ r = ftruncate(fd, offset + len); if (r < 0) - return -errno; + return -1; buf = mmap(NULL, len, PROT_READ | PROT_WRITE, MAP_SHARED, fd, offset); if (buf == MAP_FAILED) - return -errno; + return -1; memset(buf, 0, len); munmap(buf, len); @@ -167,24 +167,26 @@ int safezero(int fd, int flags ATTRIBUTE_UNUSED, off_t offset, off_t len) unsigned long long remain, bytes; if (lseek(fd, offset, SEEK_SET) < 0) - return errno; + return -1; /* Split up the write in small chunks so as not to allocate lots of RAM */ remain = len; bytes = 1024 * 1024; r = VIR_ALLOC_N(buf, bytes); - if (r < 0) - return -ENOMEM; + if (r < 0) { + errno = ENOMEM; + return -1; + } while (remain) { if (bytes > remain) bytes = remain; - r = safewrite(fd, buf, len); + r = safewrite(fd, buf, bytes); if (r < 0) { VIR_FREE(buf); - return r; + return -1; } /* safewrite() guarantees all data will be written */ -- 1.7.0

3 3

[libvirt] [PATCH] lxc: detect socket failure
by Eric Blake 03 Mar '10

03 Mar '10

Coverity warned about a potential close(-1). * src/lxc/lxc_controller.c (lxcControllerMain): React to failure to accept. --- src/lxc/lxc_controller.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-) diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index 95970cc..318106e 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -1,5 +1,6 @@ /* + * Copyright (C) 2010 Red Hat, Inc. * Copyright IBM Corp. 2008 * * lxc_controller.c: linux container process controller * @@ -348,8 +349,13 @@ static int lxcControllerMain(int monitor, numEvents = epoll_wait(epollFd, &epollEvent, 1, timeout); if (numEvents > 0) { if (epollEvent.data.fd == monitor) { int fd = accept(monitor, NULL, 0); + if (fd < 0) { + virReportSystemError(errno, "%s", + _("epoll_ctl(client) failed")); + goto cleanup; + } if (client != -1) { /* Already connected, so kick new one out */ close(fd); continue; } -- 1.6.6.1

2 1

[libvirt] [PATCH] build: silence coverity warning in node_device
by Eric Blake 03 Mar '10

03 Mar '10

All other uses of get_str_prop in this file that ignored failure explicitly cast to void. * src/node_device/node_device_hal.c (dev_create): Silence coverity warning. --- src/node_device/node_device_hal.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/node_device/node_device_hal.c b/src/node_device/node_device_hal.c index 6cc2864..4bf445d 100644 --- a/src/node_device/node_device_hal.c +++ b/src/node_device/node_device_hal.c @@ -463,7 +463,7 @@ static void dev_create(const char *udi) goto cleanup; /* Some devices don't have a path in sysfs, so ignore failure */ - get_str_prop(ctx, udi, "linux.sysfs_path", &devicePath); + (void)get_str_prop(ctx, udi, "linux.sysfs_path", &devicePath); dev = virNodeDeviceAssignDef(&driverState->devs, def); -- 1.6.6.1

3 2

[libvirt] [PATCH] Fix USB/PCI device address aliases in QEMU hotplug driver
by Daniel P. Berrange 02 Mar '10

02 Mar '10

The USB/PCI device hotplug code for the QEMU driver was forgetting to allocate a unique device alias. * src/qemu/qemu_driver.c: Fill in device alias for USB/PCI devices --- src/qemu/qemu_driver.c | 11 ++++++++--- 1 files changed, 8 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index c2bd46d..5ac3316 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -5878,6 +5878,8 @@ static int qemudDomainAttachHostPciDevice(struct qemud_driver *driver, } if (qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) { + if (qemuAssignDeviceHostdevAlias(vm->def, hostdev, -1) < 0) + goto error; if (qemuDomainPCIAddressEnsureAddr(priv->pciaddrs, &hostdev->info) < 0) goto error; @@ -5926,9 +5928,12 @@ static int qemudDomainAttachHostUsbDevice(struct qemud_driver *driver, qemuDomainObjPrivatePtr priv = vm->privateData; char *devstr = NULL; - if ((qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) && - !(devstr = qemuBuildUSBHostdevDevStr(hostdev))) - goto error; + if (qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) { + if (qemuAssignDeviceHostdevAlias(vm->def, hostdev, -1) < 0) + goto error; + if (!(devstr = qemuBuildUSBHostdevDevStr(hostdev))) + goto error; + } if (VIR_REALLOC_N(vm->def->hostdevs, vm->def->nhostdevs+1) < 0) { virReportOOMError(); -- 1.6.6.1

2 1

[libvirt] [PATCH 0/6] Add more guest clock mods
by Daniel P. Berrange 02 Mar '10

02 Mar '10

An update of http://www.redhat.com/archives/libvir-list/2010-February/msg00104.html The two current options are: <clock offset='utc'/> <clock offset='localtime'/> This introduces a way to set a specific timezone <clock offset='timezone' timezone='Europe/Paris'/> And a way to set a completely arbitrary time, by giving the number of seconds offset from UTC. <clock offset='variable' adjustment='123456'/> Since last time, I removed the 'timezone' attribute from the existing offset='localtime' option, and introduced the new explicit option offset='timezone'. That way we don't change semantics of the existing options at all.

4 22

[libvirt] [PATCH 0/6] SCSI-Disk-Hotremove
by Wolfgang Mauerer 02 Mar '10

02 Mar '10

[PATCH 0/6] SCSI-Disk-Hotremove This series adds support for hot-removing SCSI disks when qemu supports the drive/device model. It also contains fixes for SCSI controller hotremoval and the build process. Cheers, Wolfgang --- Siemens AG, Corporate Technology, CT T TC 4 Corporate Competence Centre Embedded Linux Makefile.am | 7 +-- qemu/qemu_driver.c | 97 +++++++++++++++++++++++++++++++++++++++++++---- qemu/qemu_monitor.c | 13 ++++++ qemu/qemu_monitor.h | 3 + qemu/qemu_monitor_json.c | 24 +++++++++++ qemu/qemu_monitor_json.h | 3 + qemu/qemu_monitor_text.c | 40 +++++++++++++++++++ qemu/qemu_monitor_text.h | 3 + 8 files changed, 180 insertions(+), 10 deletions(-)

4 19

[libvirt] libvirt/dnsmasq integration.
by Simon Kelley 02 Mar '10

02 Mar '10

As the principal maintainer of dnsmasq, I'm seeing increasing reports of problems on systems which run both dnsmasq and libvirt. I'm fairly sure I understand what's going on in these cases, and I have a few proposals for changes in libvir and dnsmasq that should fix things. The problem is that libvirt runs a private instance of dnsmasq: on machines which are also running a "system" dnsmasq daemon, this can cause problems. Some background: dnsmasq can run in two modes. Default mode: dnsmasq binds the wildcard address and does network magic to determine which interface request packets actually come from, so that the results can be sent back with the correct source address. This has the advantage that network interfaces can come and go and change IP address and dnsmasq will keep working. It's possible to restrict dnsmasq to only reply to requests on some interfaces; requests from other interfaces will be read by dnsmasq and then silently dropped. Telling dnsmasq to use an interface which doesn't exist but might in the future will result in a logged warning, but dnsmasq will still start and when the interface comes up it will work. Bind-interfaces mode: This is the traditional way to do UDP servers. At startup dnsmasq enumerates all the extant interfaces and then opens a socket for each one, listening on the interfaces's IP address. Interfaces may be skipped if excluded by the --interface and --except-interface flags, and any interface specified in --interface which doesn't exist at start-up will generate a fatal error. In almost all cases, default mode is better: --bind-interfaces is only there to cope with old platforms which don't support enough socket options to do default mode. The only time when --bind-interfaces works better is when it's desirable to run more than one instance of dnsmasq or have dnsmasq co-exist with another DNS server. This is not possible in default mode, but it does work in bind-interfaces mode, providing than _all_ instances of dnsmasq are in bind-interfaces mode, and that they listen on a disjoint set of interfaces. Therefore, to allow multiple dnsmasq instances libvirt's private dnsmasq instance is started in bind-interfaces mode: that forces one of the dnsmasq instances to do bind-interfaces. Many of the Linux distibution dnsmasq packages have now implemented an /etc/dnsmasq.d directory where configuration fragments can be dropped. Their libvirt packages are putting a file there which contains a bind-interfaces command, so that the "system" dnsmasq is automatically forced into the same mode, and the two can co-exist. This works, sort-of, but there some disadvantages. Installing libvirt drops the configuration change for the system dnsmasq, but the packages frequently don't restart the system daemon, so that things transiently fail until everything has rebooted. Much worse, the system dnsmasq is forced into bind-interfaces mode and then service to transient interfaces (usb, ad-hoc wifi) no longer works, or, because those interfaces are mentioned in the dnsmasq configuration, dnsmasq now fails at start-up when the interfaces don't exist. I don't think there is a solution to this problem wholly within dnsmasq: the nature of the BSD-sockets universe means that there never be two instances of dnsmasq without bind-interfaces, and it's, at best, difficult to get round the limitations of that WRT transient network interfaces. There is, an alternative solution, which would involve changes to both dnsmasq, libvirt and their packaging. Hence this mail. My proposal is to get rid of the necessity for two dnsmasq instances. Libvirt should check for the existance of a "system" dnsmasq and, if the system daemon exists, libvirt should drop the required configuration into /etc/dnsmasq.d and then restart it. If the system daemon is not installed or enabled, libvirt can start a private instance as now. The difficulty with this scheme is that libvirt needs to create some configuration which enables the services it needs on the virtual network without disturbing, or being disturbed by, whatever configuration exists for the system daemon. That's not currently possible, but it can be made possible. I'm assuming that libvirt needs to provide a set of IP address / MAC address mappings, and range of IP addresses on a virtual network. It needs DHCP and DNS service on the virtual network. The dhcp-host IP/MAC mappings are a non-problem: they will be ignored for any other subnet where the IP addresses don't fit, and any other dhcp-hosts in the system configuration will be similarly ignored for DHCP on the virtual network subnet. The dhcp-range is more of a problem. Service to particular networks in dnsmasq is controlled by interface=<interface name"> lines in the configuration. If there are none of these, service is provided to all interfaces. If they exist, service is limited to the interfaces specified. The existence of any dhcp-range line in dnsmasq's configuration enables the DHCP server for any subnet unless explicitly limited to particular interfaces. So a default dnsmasq installation, (with no interface=<interface>) which provides DNS everywhere but DHCP nowhere would be turned into one which provided DHCP on every interface by libvirt adding a dhcp-range. Since there wouldn't be a suitable DHCP range for most subnets, this would only result in logged errors, but it is still not good. Worse, there's no good answer to the question 'should libvirt include interface=virt0"' in the configuration it supplies? If it does, then the "enable DHCP on all interfaces" problem is solved, but a default system configuration with no interface declaration is transformed from one which provides DNS everywhere to one which provides DNS only to the virtual interface. If libvirt doesn't provide "interface=virt0" and the system configuration includes interface declarations, then there will be no DNS or DHCP service to the virtual network. To solve this, I propose to add an optional interface name to the dhcp-range declaration. The semantics of this would be rather odd, but solve the problem perfectly. 1) for DHCP, if any other dhcp-range exists _without_ an interface name, them the interface name is ignored and and things behave as before, otherwise DHCP is only provided to interfaces mentioned in dhcp-range declarations. 2) for DNS, if there are no interface declarations, things work as before. If there are interface declarations, the interfaces mentioned in dhcp-ranges are added to the set which get DNS service. With these rules, it should be possible for libvirt to drop eg dhcp-range=interface:virt0,192.168.0.1,192.168.0.240 into the configuration of the system dnsmasq and get DHCP and DNS service for virt0, irrespective of any other configuration in the system dnsmasq, and doing so shouldn't affect the services supplied elsewhere. The code in libvirt to make this work looks like this: echo dhcp-range=interface:virt0,<ip range> >>/etc/dnsmasq.d/libvirt if <system dnsmasq is not installed or not enabled> dnsmasq --interface=virt0\ --bind-interfaces --conf-file=/etc/dnsmasq.d/libvirt else /etc/init.d/dnsmasq restart (The --bind-interfaces in the private-dnsmasq instance keeps dnsmasq from clashing with other nameservers eg BIND which may be running.) The system dnsmasq package has to ensure that /etc/dnsmasq.d is read for configuration fragments, and the dnsmasq package and the libvirt package will have to co-operate to manage transitions between private and system dnsmasq mode caused by package installation or removal. Does that make sense? It's a long and involved explanation to come to cold. I fear I may have over-simplified what libvirt is doing with dnsmasq, in which case please enlighten me and I'll modify my scheme to take that into account. If this looks good I can easily have the necessary dnsmasq changes in the next release. Cheers, Simon.

3 8

[libvirt] [PATCH] x86Decode: avoid NULL-dereference upon questionable input
by Jim Meyering 02 Mar '10

02 Mar '10

Clang warned about the potential NULL-dereference via the STREQ/strcmp below. models[i] could be NULL. Even "models" could be NULL, and the "allowed = ..." test makes that appear to be deliberately allowed. The change below is the least invasive and cleanest I could come up with, assuming there is no need to diagnose (e.g., by returning -1) the condition of a NULL models[i] pointer. >From 623ac546a93f3e5b554647b05524b5041a642530 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Sun, 28 Feb 2010 13:34:06 +0100 Subject: [PATCH] x86Decode: avoid NULL-dereference upon questionable input * src/cpu/cpu_x86.c (x86Decode): Don't dereference NULL when passed a NULL "models" pointer, or when passed a nonzero "nmodels" value and a corresponding NULL models[i]. --- src/cpu/cpu_x86.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c index 2194c32..b263629 100644 --- a/src/cpu/cpu_x86.c +++ b/src/cpu/cpu_x86.c @@ -957,89 +957,89 @@ x86Compare(virCPUDefPtr host, virCPUDefPtr cpu) { return x86Compute(host, cpu, NULL); } static virCPUCompareResult x86GuestData(virCPUDefPtr host, virCPUDefPtr guest, union cpuData **data) { return x86Compute(host, guest, data); } static int x86Decode(virCPUDefPtr cpu, const union cpuData *data, const char **models, unsigned int nmodels) { int ret = -1; struct x86_map *map; const struct x86_model *candidate; virCPUDefPtr cpuCandidate; virCPUDefPtr cpuModel = NULL; struct cpuX86cpuid *cpuid; int i; if (data == NULL || (map = x86LoadMap()) == NULL) return -1; candidate = map->models; while (candidate != NULL) { bool allowed = (models == NULL); for (i = 0; i < candidate->ncpuid; i++) { cpuid = x86DataCpuid(data, candidate->cpuid[i].function); if (cpuid == NULL || !x86cpuidMatchMasked(cpuid, candidate->cpuid + i)) goto next; } for (i = 0; i < nmodels; i++) { - if (STREQ(models[i], candidate->name)) { + if (models && models[i] && STREQ(models[i], candidate->name)) { allowed = true; break; } } if (!allowed) { VIR_DEBUG("CPU model %s not allowed by hypervisor; ignoring", candidate->name); goto next; } if (!(cpuCandidate = x86DataToCPU(data, candidate, map))) goto out; if (cpuModel == NULL || cpuModel->nfeatures > cpuCandidate->nfeatures) { virCPUDefFree(cpuModel); cpuModel = cpuCandidate; } else virCPUDefFree(cpuCandidate); next: candidate = candidate->next; } if (cpuModel == NULL) { virCPUReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Cannot find suitable CPU model for given data")); goto out; } cpu->model = cpuModel->model; cpu->nfeatures = cpuModel->nfeatures; cpu->features = cpuModel->features; VIR_FREE(cpuModel); ret = 0; out: x86MapFree(map); virCPUDefFree(cpuModel); return ret; } -- 1.7.0.1.414.g89213d

2 2

[libvirt] Problems with attach-device/detach-device using libvirt 0.7.6
by Rolf Eike Beer 02 Mar '10

02 Mar '10

Hi, first a short summary of what the situation is: I'm running qemu 0.12.2, libvirt 0.7.6 on a E5520 based machine with kernel 2.6.33. I also tried 2.6.32.7 as I was using that before and still had the old kernel around, but kernel version doesn't matter AFAICT. The host is a gentoo AMD64 installation, the client is SuSE 11.0 (both 32 and 64 bit tested). This system is our machine to run automated PCIe device tests on, so the idea is to move PCIe devices from the host to different vms, run some testcases there and then move them away again. This has worked before (I'm not absolutely sure, but I would bet a bit of money that it was libvirt 0.7.5). So, what I'm going to do is simply virsh # attach-device suse11.0-AMD64 dev-pci-AD1-CL.xml Device attached successfully virsh # detach-device suse11.0-AMD64 dev-pci-AD1-CL.xml Device detached successfully This is just was I expect and what I now get with the two patches attached. The original situation was somewhat different (i.e. broken): -first I got complains in the log stating: tried to create id "(null)" twice for "device" Well, adding two devices with the same id is bad, and if that id is a NULL pointer anyway IMHO it's better to not pass the parameter at all. That's what the "null-pci-id" patch does. -next, detaching failed. I found out that the command sent to qemu was " pci_del pci_addr=fc67add0:7f41:fe38b65a" which is obviously bogus. After some hours of digging through the command stack up and down I found that it was the memcpy addressed in the "pci_memcpy" patch. When the device is attached qemu just replies with an empty string. Afterwards from the guestAddr the address of the PCI device in the client is copied back to the info structure. Too bad that this structure was never initialized to anything useful (at least not for me). When I kick out the memcpy() everything works fine. Configuration files of one device and one guest are attached, the others look similar. Note: when the client PCI id is passed to qemu anyway it would be cool if that could be specified in the device XML. For USB devices, too. Eike Please keep me CC'ed as I'm not on the list.

2 3

[libvirt] [PATCH] openvzDomainDefineCmd: remove useless increment
by Jim Meyering 02 Mar '10

02 Mar '10

Another dead store. Refreshingly shallow. The function below starts like this: static int openvzDomainDefineCmd(virConnectPtr conn, const char *args[], int maxarg, virDomainDefPtr vmdef) { int narg; int veid; int max_veid; and there are no intervening uses of "max_veid". >From 7333ff9e0cf83d243dba752d293d9007ded5e9f8 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 1 Mar 2010 21:38:06 +0100 Subject: [PATCH] openvzDomainDefineCmd: remove useless increment * src/openvz/openvz_driver.c (openvzDomainDefineCmd): Remove useless increment of "max_veid". --- src/openvz/openvz_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/openvz/openvz_driver.c b/src/openvz/openvz_driver.c index 5057b81..4673b6c 100644 --- a/src/openvz/openvz_driver.c +++ b/src/openvz/openvz_driver.c @@ -137,65 +137,65 @@ openvzDomainDefineCmd(virConnectPtr conn, narg = 0; ADD_ARG_LIT(VZCTL); ADD_ARG_LIT("--quiet"); ADD_ARG_LIT("create"); if ((fp = popen(VZLIST " -a -ovpsid -H 2>/dev/null", "r")) == NULL) { openvzError(NULL, VIR_ERR_INTERNAL_ERROR, "%s", _("popen failed")); return -1; } max_veid = 0; while (!feof(fp)) { if (fscanf(fp, "%d\n", &veid) != 1) { if (feof(fp)) break; openvzError(NULL, VIR_ERR_INTERNAL_ERROR, "%s", _("Failed to parse vzlist output")); goto cleanup; } if (veid > max_veid) { max_veid = veid; } } fclose(fp); if (max_veid == 0) { max_veid = 100; } else { max_veid++; } - sprintf(str_id, "%d", max_veid++); + sprintf(str_id, "%d", max_veid); ADD_ARG_LIT(str_id); ADD_ARG_LIT("--name"); ADD_ARG_LIT(vmdef->name); if (vmdef->nfss == 1 && vmdef->fss[0]->type == VIR_DOMAIN_FS_TYPE_TEMPLATE) { ADD_ARG_LIT("--ostemplate"); ADD_ARG_LIT(vmdef->fss[0]->src); } #if 0 if ((vmdef->profile && *(vmdef->profile))) { ADD_ARG_LIT("--config"); ADD_ARG_LIT(vmdef->profile); } #endif ADD_ARG(NULL); return 0; no_memory: openvzError(conn, VIR_ERR_INTERNAL_ERROR, _("Could not put argument to %s"), VZCTL); return -1; cleanup: fclose(fp); return -1; #undef ADD_ARG #undef ADD_ARG_LIT } -- 1.7.0.1.414.g89213d

2 1

[libvirt] [PATCH] phypUUIDTable_Push: do not corrupt output stream
by Jim Meyering 02 Mar '10

02 Mar '10

This started with clang's report of the two dead stores. With the increment and decrement by "sent" being performed outside of the intended loop, the two assignments had no net effect. BTW, I presume these hard-coded names, from the same function, are somehow temporary and/or debugging-related: char remote_file[] = "/home/hscroot/libvirt_uuid_table"; char local_file[] = "./uuid_table"; >From a1104c4d1a95b10cab19bacf91ed073769f1d2d7 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 1 Mar 2010 21:26:59 +0100 Subject: [PATCH] phypUUIDTable_Push: do not corrupt output stream upon partial write * src/phyp/phyp_driver.c (phypUUIDTable_Push): Move incr/decr of ptr/nread into the loop where those variables are used. Also, remove "exit" label and just-preceding "goto". --- src/phyp/phyp_driver.c | 7 ++----- 1 files changed, 2 insertions(+), 5 deletions(-) diff --git a/src/phyp/phyp_driver.c b/src/phyp/phyp_driver.c index 1e8ed30..3fc5a0c 100644 --- a/src/phyp/phyp_driver.c +++ b/src/phyp/phyp_driver.c @@ -1982,54 +1982,51 @@ phypUUIDTable_Push(virConnectPtr conn) do { nread = fread(buffer, 1, sizeof(buffer), fd); if (nread <= 0) { if (feof(fd)) { /* end of file */ break; } else { VIR_ERROR("Failed to read from '%s'", local_file); goto err; } } ptr = buffer; sent = 0; do { /* write the same data over and over, until error or completion */ rc = libssh2_channel_write(channel, ptr, nread); if (LIBSSH2_ERROR_EAGAIN == rc) { /* must loop around */ continue; } else if (rc > 0) { /* rc indicates how many bytes were written this time */ sent += rc; } + ptr += sent; + nread -= sent; } while (rc > 0 && sent < nread); - ptr += sent; - nread -= sent; } while (1); - goto exit; - - exit: if (channel) { libssh2_channel_send_eof(channel); libssh2_channel_wait_eof(channel); libssh2_channel_wait_closed(channel); libssh2_channel_free(channel); channel = NULL; } return 0; err: if (channel) { libssh2_channel_send_eof(channel); libssh2_channel_wait_eof(channel); libssh2_channel_wait_closed(channel); libssh2_channel_free(channel); channel = NULL; } return -1; } int phypUUIDTable_Pull(virConnectPtr conn) { -- 1.7.0.1.414.g89213d

2 1

[libvirt] [PATCH] add an assert, to avoid a false-positive NULL-deref warning from clang
by Jim Meyering 02 Mar '10

02 Mar '10

Here's a case in which using an assertion appears to be the only way to tell clang that "client" really is non-NULL at that point. I'm sure clang's analyzers will eventually improve, and hence avoid this sort of false positive, so have marked this with a FIXME comment, to help ensure we eventually remove this otherwise unnecessary assertion. >From ba26bad7aec8713ded0fd3300e951eac3673cc72 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 1 Mar 2010 15:19:48 +0100 Subject: [PATCH] add an assert, to avoid a false-positive NULL-deref warning from clang * daemon/libvirtd.c (qemudWorker): Assert. --- daemon/libvirtd.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index 9bdbecb..a357914 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -24,34 +24,35 @@ #include <config.h> #include <sys/types.h> #include <sys/wait.h> #include <sys/stat.h> #include <unistd.h> #include <fcntl.h> #include <limits.h> #include <sys/socket.h> #include <sys/un.h> #include <sys/poll.h> #include <netinet/in.h> #include <netinet/tcp.h> #include <netdb.h> #include <stdlib.h> #include <pwd.h> #include <stdio.h> +#include <assert.h> #include <stdarg.h> #include <syslog.h> #include <string.h> #include <errno.h> #include <getopt.h> #include <fnmatch.h> #include <grp.h> #include <signal.h> #include <netdb.h> #include "libvirt_internal.h" #include "virterror_internal.h" #define VIR_FROM_THIS VIR_FROM_QEMU #include "libvirtd.h" #include "dispatch.h" @@ -1504,34 +1505,38 @@ static void *qemudWorker(void *data) virMutexLock(&server->lock); while (((client = qemudPendingJob(server)) == NULL) && !worker->quitRequest) { if (virCondWait(&server->job, &server->lock) < 0) { virMutexUnlock(&server->lock); return NULL; } } if (worker->quitRequest) { if (client) virMutexUnlock(&client->lock); virMutexUnlock(&server->lock); return NULL; } worker->processingCall = 1; virMutexUnlock(&server->lock); + /* Tell clang we know what we're doing. + FIXME: remove when clang improves. */ + assert (client); + /* We own a locked client now... */ client->refs++; /* Remove our message from dispatch queue while we use it */ msg = qemudClientMessageQueueServe(&client->dx); /* This function drops the lock during dispatch, * and re-acquires it before returning */ if (remoteDispatchClientRequest (server, client, msg) < 0) { VIR_FREE(msg); qemudDispatchClientFailure(client); client->refs--; virMutexUnlock(&client->lock); continue; } client->refs--; -- 1.7.0.1.414.g89213d

5 7

[libvirt] [RFC PATCH] build: consistently use C99 varargs macros
by Eric Blake 02 Mar '10

02 Mar '10

Prior to this patch, there was an inconsistent mix between GNU and C99. For consistency, and potential portability to other compilers, stick with the C99 vararg macro syntax. * src/conf/cpu_conf.c (virCPUReportError): Use C99 rather than GNU vararg macro syntax. * src/conf/domain_conf.c (virDomainReportError): Likewise. * src/conf/domain_event.c (eventReportError): Likewise. * src/conf/interface_conf.c (virInterfaceReportError): Likewise. * src/conf/network_conf.c (virNetworkReportError): Likewise. * src/conf/node_device_conf.h (virNodeDeviceReportError): Likewise. * src/conf/secret_conf.h (virSecretReportError): Likewise. * src/conf/storage_conf.h (virStorageReportError): Likewise. --- Any objections to expanding this into a full-blown patch series? This touches 8 files; I counted 43 more files that could be altered, using: $ git grep -l 'define.*[a-zA-Z]\.\.\.' src/conf/cpu_conf.c | 4 ++-- src/conf/domain_conf.c | 4 ++-- src/conf/domain_event.c | 5 +++-- src/conf/interface_conf.c | 4 ++-- src/conf/network_conf.c | 6 +++--- src/conf/node_device_conf.h | 5 +++-- src/conf/secret_conf.h | 6 +++--- src/conf/storage_conf.h | 6 +++--- 8 files changed, 21 insertions(+), 19 deletions(-) diff --git a/src/conf/cpu_conf.c b/src/conf/cpu_conf.c index ed83188..612e376 100644 --- a/src/conf/cpu_conf.c +++ b/src/conf/cpu_conf.c @@ -31,9 +31,9 @@ #define VIR_FROM_THIS VIR_FROM_CPU -#define virCPUReportError(code, fmt...) \ +#define virCPUReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_CPU, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) VIR_ENUM_IMPL(virCPUMatch, VIR_CPU_MATCH_LAST, "minimum", diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 4420445..ea84863 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -232,9 +232,9 @@ VIR_ENUM_IMPL(virDomainNetdevMacvtap, VIR_DOMAIN_NETDEV_MACVTAP_MODE_LAST, "private", "bridge") -#define virDomainReportError(code, fmt...) \ +#define virDomainReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_DOMAIN, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) #ifndef PROXY diff --git a/src/conf/domain_event.c b/src/conf/domain_event.c index b474b1c..e791cab 100644 --- a/src/conf/domain_event.c +++ b/src/conf/domain_event.c @@ -1,6 +1,7 @@ /* * domain_event.c: domain event queue processing helpers * + * Copyright (C) 2010 Red Hat, Inc. * Copyright (C) 2008 VirtualIron * * This library is free software; you can redistribute it and/or @@ -30,9 +31,9 @@ #define VIR_FROM_THIS VIR_FROM_NONE -#define eventReportError(conn, code, fmt...) \ +#define eventReportError(conn, code, ...) \ virReportErrorHelper(conn, VIR_FROM_THIS, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) /** diff --git a/src/conf/interface_conf.c b/src/conf/interface_conf.c index a0d2dfa..bc9e958 100644 --- a/src/conf/interface_conf.c +++ b/src/conf/interface_conf.c @@ -45,9 +45,9 @@ static int virInterfaceDefDevFormat(virBufferPtr buf, const virInterfaceDefPtr def, int level); -#define virInterfaceReportError(code, fmt...) \ +#define virInterfaceReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_INTERFACE, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) static void virInterfaceIpDefFree(virInterfaceIpDefPtr def) { diff --git a/src/conf/network_conf.c b/src/conf/network_conf.c index 6d3c3c0..39ebd62 100644 --- a/src/conf/network_conf.c +++ b/src/conf/network_conf.c @@ -1,7 +1,7 @@ /* * network_conf.c: network XML handling * - * Copyright (C) 2006-2009 Red Hat, Inc. + * Copyright (C) 2006-2010 Red Hat, Inc. * Copyright (C) 2006-2008 Daniel P. Berrange * * This library is free software; you can redistribute it and/or @@ -53,9 +53,9 @@ VIR_ENUM_IMPL(virNetworkForward, VIR_NETWORK_FORWARD_LAST, "none", "nat", "route" ) -#define virNetworkReportError(code, fmt...) \ +#define virNetworkReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_NETWORK, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) virNetworkObjPtr virNetworkFindByUUID(const virNetworkObjListPtr nets, const unsigned char *uuid) diff --git a/src/conf/node_device_conf.h b/src/conf/node_device_conf.h index 3a5432c..cbaad9b 100644 --- a/src/conf/node_device_conf.h +++ b/src/conf/node_device_conf.h @@ -1,6 +1,7 @@ /* * node_device_conf.h: config handling for node devices * + * Copyright (C) 2010 Red Hat, Inc. * Copyright (C) 2008 Virtual Iron Software, Inc. * Copyright (C) 2008 David F. Lively * @@ -217,9 +218,9 @@ struct _virDeviceMonitorState { void *privateData; /* driver-specific private data */ }; -#define virNodeDeviceReportError(code, fmt...) \ +#define virNodeDeviceReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_NODEDEV, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) int virNodeDeviceHasCap(const virNodeDeviceObjPtr dev, const char *cap); diff --git a/src/conf/secret_conf.h b/src/conf/secret_conf.h index da00920..afcf8f1 100644 --- a/src/conf/secret_conf.h +++ b/src/conf/secret_conf.h @@ -1,7 +1,7 @@ /* * secret_conf.h: internal <secret> XML handling API * - * Copyright (C) 2009 Red Hat, Inc. + * Copyright (C) 2009-2010 Red Hat, Inc. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -26,9 +26,9 @@ #include "internal.h" #include "util.h" -#define virSecretReportError(code, fmt...) \ +#define virSecretReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_SECRET, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) VIR_ENUM_DECL(virSecretUsageType) diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h index c643984..5441c8e 100644 --- a/src/conf/storage_conf.h +++ b/src/conf/storage_conf.h @@ -1,7 +1,7 @@ /* * storage_conf.h: config handling for storage driver * - * Copyright (C) 2006-2008 Red Hat, Inc. + * Copyright (C) 2006-2008, 2010 Red Hat, Inc. * Copyright (C) 2006-2008 Daniel P. Berrange * * This library is free software; you can redistribute it and/or @@ -318,9 +318,9 @@ static inline int virStoragePoolObjIsActive(virStoragePoolObjPtr pool) { return pool->active; } -#define virStorageReportError(code, fmt...) \ +#define virStorageReportError(code, ...) \ virReportErrorHelper(NULL, VIR_FROM_STORAGE, code, __FILE__, \ - __FUNCTION__, __LINE__, fmt) + __FUNCTION__, __LINE__, __VA_ARGS__) int virStoragePoolLoadAllConfigs(virStoragePoolObjListPtr pools, const char *configDir, -- 1.6.6.1

4 4

[libvirt] [RFC] Proposal for introduction of network traffic filtering capabilities for filtering of network traffic from and to VMs
by Stefan Berger 02 Mar '10

02 Mar '10

Hello! The following is a proposal to introduce network traffic filtering capabilities for the network traffic originating from and destined to virtual machines. Libvirt's capabilities will be extended to allow users to describe what traffic filtering rules are to be applied on a virtual machine (using XML) and libvirt then creates the appropriate ebtables and iptables rules and applies those on the host when the virtual machine starts up, resumes after suspension or resumes on a new host after migration. libvirt tears down the traffic filtering rules when the virtual machine shuts down, suspends, or a VM is migrated to another host. It will also be possible to modify the filtering rules while a virtual machine is running. In this architecture we apply the firewall rules on the outside of the virtual machines on the Linux host and make use of the fact that virtual machines can be configured by libvirt to have their network traffic pass through the host and the host exposes (tap) 'backend' interfaces on which the firewall rules can be applied. The application of the firewall rules is optional and those who do not want to introduce a raw network throughput performance hit on their system due to the evaluation of every packet passing through the filtering chains, do not have to use these capabilities. An initial implementation would be done for libvirt's Qemu support. As stated above, the application of firewall rules on virtual machines will require some form of XML description that lets the user choose what type of filtering is to be performed. In effect the user needs to be able to tell libvirt which ebtables and iptables rules to generate so that the appropriate filtering can be done. We realize that ebtables and iptables have lots of capabilities but think that we need to restrict which capabilities can actually be 'reached' through this XML. The following proposal is based on an XML as defined by the DMTF ( http://www.dmtf.org/standards/cim/cim_schema_v2230/CIM_Network.pdf, slide 10) with extensions for processing of ARP packets.It gives control over the evaluation of Ethernet (MAC) frames, ARP packet data and IP header information. A (draft) XML skeleton in this case could look as follows: <FilterEntry> <Action>DROP|ACCEPT</Action>  <TrafficType>incoming [to VM]|outgoing [from VM]</TrafficType> <Hdr8021Filter> <HdrSrcMACAddr8021> </HdrSrcMACAddr8021> <HdrSrcMACMask8021> </HdrSrcMACMask8021> <HdrDestMACAddr8021> </HdrDestMACAddr8021> <HdrDestMACMask8021> </HdrDestMACMask8021> <HdrProtocolID8021> numeric and/or string? </HdrProtocolID8021> <HdrPriorityValue8021></HdrPriorityValue8021> <HdrVLANID8021> </HdrVLANID8021> </Hdr8021Filter> <ARPFilter> <HWType> </HWType> <ProtocolType> </ProtocolType> <OPCode> </OPCode> <SrcMACAddr> </SrcMACAddr> <SrcIPAddr> </SrcIPAddr> <DestMACAddr> </DestMACAddr> <DestIPAddr> </DestIPAddr> </ARPFilter> <IPHeadersFilter> <HdrIPVersion> </HdrIPVersion> <HdrSrcAddress> </HdrSrcAddress> <HdrSrcMask> </HdrSrcMask> <HdrDestAddress> </HdrDestAddress> <HdrDestMask> </HdrDestMask> <HdrProtocolID> numeric and/or string? </HdrProtocolID> <HdrSrcPortStart> </HdrSrcPortStart> <HdrSrcPortEnd> </HdrSrcPortEnd> <HdrDestPortStart> </HdrDestPortStart> <HdrDestPortEnd> </HdrDestPortEnd> <HdrDSCP> </HdrDSCP> </IPHeadersFilter> </FilterEntry> Since the ebtables and iptables command cannot accept all possible parameters at the same time, only a certain subset of the above parameters may be set for any given filter command. Higher level application writers will likely use a library that lets them choose which features they would want to have enforced, such as no-broadcast or no-multicast, enforcement of MAC spoofing prevention or ARP poisoning prevention, which then generates this lower level XML rules in the appropriate order of the rules. Further, we will introduce filter pools where a collection of the above filter rules can be stored and referenced to by virtual machines' individual interface. A reference to such a filter pool entry will be given in the interface description and may look as in the following draft XML: <interface type='bridge'> <source bridge='virbr0'/> <script path='vif-bridge'/> <firewall> <reference profile='generic-layer2' ip_address='10.0.0.1'/> <reference profile='VM-specific-layer3'/> </firewall> </interface> The above XML has one reference to a generic layer2 filter template XML that should be usable by multiple virtual machines but would need to be customized with interface-specific parameters. In this case, the IP address of the interface is explicitly provided in order to make the filter XML template a concrete XML for the particular interface. The MAC address of the interface is not explicitly provided since it will already have been randomly generated by libvirt at the point when this layer2 filter XML needs to be converted into concrete ebtables commands/rules. We still need to determine how the format of a template should look like, though an idea would be to indicate a placeholder for a VM's MAC address using THIS_MAC and similarly for the IP address with THIS_IP as placeholder. Further, we would introduce the management of filter 'pools'. Considering existing functionality in libvirt and CLI commands for similar type of management functionality, the following CLI commands would be added: virsh nwfilter-create <file> virsh nwfilter-destroy <profile name> virsh nwfilter-list <options> virsh nwfilter-dumpxml <profile name> virsh nwfilter-update <filename> (performs an update on an existing profile replacing all rules) possibly also: virsh nwfilter-edit <profile name> virsh nwfilter-create-from <profile name> Please let us know what you think of this proposal. Regards Stefan, Gerhard and Vivek

8 25

[libvirt] [PATCH] qemudNetworkIfaceConnect: remove dead store
by Jim Meyering 02 Mar '10

02 Mar '10

clang reported this "dead store". As you can see, just a few lines before there's an identical assignment, so this may have been an editing error (didn't dig). As clang reports, that final assignment is a dead store, since there is no following use of "errobj". >From 92edbcf2ffa0dad25251723cd0749be2397a2321 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 1 Mar 2010 15:59:42 +0100 Subject: [PATCH] qemudNetworkIfaceConnect: remove dead store * src/qemu/qemu_conf.c (qemudNetworkIfaceConnect): Remove extraneous virSaveLastError call, whose result was unused. --- src/qemu/qemu_conf.c | 1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 1f8a9c3..eee6222 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -1520,11 +1520,10 @@ qemudNetworkIfaceConnect(virConnectPtr conn, errobj = virSaveLastError(); virNetworkFree(network); virSetError(errobj); virFreeError(errobj); - errobj = virSaveLastError(); if (fail) return -1; } else if (net->type == VIR_DOMAIN_NET_TYPE_BRIDGE) { if (!(brname = strdup(net->data.bridge.brname))) { -- 1.7.0.1.414.g89213d

2 2

[libvirt] [PATCH] udevEnumerateDevices: remove dead code
by Jim Meyering 02 Mar '10

02 Mar '10

Clang reported the dead store. >From 0557e23ca2f5eb78ed9f260040895011785f670f Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 1 Mar 2010 16:14:35 +0100 Subject: [PATCH] udevEnumerateDevices: remove dead code * src/node_device/node_device_udev.c (udevEnumerateDevices): Remove unnecessary call to udev_list_entry_get_name. --- src/node_device/node_device_udev.c | 4 +--- 1 files changed, 1 insertions(+), 3 deletions(-) diff --git a/src/node_device/node_device_udev.c b/src/node_device/node_device_udev.c index 68509f4..eee44c4 100644 --- a/src/node_device/node_device_udev.c +++ b/src/node_device/node_device_udev.c @@ -1,11 +1,11 @@ /* * node_device_udev.c: node device enumeration - libudev implementation * - * Copyright (C) 2009 Red Hat + * Copyright (C) 2009-2010 Red Hat * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * * This library is distributed in the hope that it will be useful, @@ -1333,30 +1333,28 @@ static int udevProcessDeviceListEntry(struct udev *udev, } static int udevEnumerateDevices(struct udev *udev) { struct udev_enumerate *udev_enumerate = NULL; struct udev_list_entry *list_entry = NULL; - const char *name = NULL; int ret = 0; udev_enumerate = udev_enumerate_new(udev); ret = udev_enumerate_scan_devices(udev_enumerate); if (0 != ret) { VIR_ERROR("udev scan devices returned %d", ret); goto out; } udev_list_entry_foreach(list_entry, udev_enumerate_get_list_entry(udev_enumerate)) { udevProcessDeviceListEntry(udev, list_entry); - name = udev_list_entry_get_name(list_entry); } out: udev_enumerate_unref(udev_enumerate); return ret; } -- 1.7.0.1.414.g89213d

2 1

[libvirt] [PATCH] cmdPoolDiscoverSources: initialize earlier to avoid FP from clang
by Jim Meyering 02 Mar '10

02 Mar '10

This change hoists the initialization of srcSpec to the top, to keep clang from reporting a false-positive used-uninitialized problem. >From f487655124c887310be3bc82d7beae3cd47e7d21 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 1 Mar 2010 15:41:15 +0100 Subject: [PATCH] cmdPoolDiscoverSources: initialize earlier to avoid FP from clang * tools/virsh.c (cmdPoolDiscoverSources): Always initialize srcSpec. Otherwise, clang would report that srcSpec could be used uninitialized in the call to virConnectFindStoragePoolSources. --- tools/virsh.c | 7 +++---- 1 files changed, 3 insertions(+), 4 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 3b1011c..191e7ba 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -4431,17 +4431,16 @@ static const vshCmdOptDef opts_find_storage_pool_sources[] = { static int cmdPoolDiscoverSources(vshControl * ctl, const vshCmd * cmd ATTRIBUTE_UNUSED) { - char *type, *srcSpec, *srcSpecFile, *srcList; + char *type, *srcSpecFile, *srcList; + char *srcSpec = NULL; int found; type = vshCommandOptString(cmd, "type", &found); if (!found) return FALSE; srcSpecFile = vshCommandOptString(cmd, "srcSpec", &found); - if (!found) { + if (!found) srcSpecFile = NULL; - srcSpec = NULL; - } if (!vshConnectionUsability(ctl, ctl->conn, TRUE)) return FALSE; -- 1.7.0.1.414.g89213d

2 1

[libvirt] [PATCH] qemu: avoid null dereference on failed migration
by Eric Blake 02 Mar '10

02 Mar '10

From: Eric Blake <eblake(a)redhat.com> * src/qemu/qemu_monitor_text.c (qemuMonitorTextGetMigrationStatus): Check for failed strchr, to silence a coverity warning. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- src/qemu/qemu_monitor_text.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_monitor_text.c b/src/qemu/qemu_monitor_text.c index 62ffcc6..e7b4b1f 100644 --- a/src/qemu/qemu_monitor_text.c +++ b/src/qemu/qemu_monitor_text.c @@ -989,6 +989,11 @@ int qemuMonitorTextGetMigrationStatus(qemuMonitorPtr mon, if ((tmp = strstr(reply, MIGRATION_PREFIX)) != NULL) { tmp += strlen(MIGRATION_PREFIX); end = strchr(tmp, '\r'); + if (end == NULL) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, + _("unexpected migration status in %s"), reply); + goto cleanup; + } *end = '\0'; if ((*status = qemuMonitorMigrationStatusTypeFromString(tmp)) < 0) { -- 1.6.6.1

3 2

[libvirt] [PATCH] Do not search xenstore for disk device IDs
by Jim Fehlig 02 Mar '10

02 Mar '10

Disk devices can be referenced by name in Xen, e.g. when modifying their configuration or remvoving them. As such, don't search xenstore for a device ID corresponding to the disk device. Instead, search the disks contained in the domain definition and use the disk's target name if found. This approach allows removing a disk when domain is inactive. We obviously can't search xenstore when the domain is inactive. --- src/xen/xend_internal.c | 38 +++++++++++++++++++++++--------------- 1 files changed, 23 insertions(+), 15 deletions(-) diff --git a/src/xen/xend_internal.c b/src/xen/xend_internal.c index 9d95291..d0d32e2 100644 --- a/src/xen/xend_internal.c +++ b/src/xen/xend_internal.c @@ -93,6 +93,7 @@ xenDaemonFormatSxprOnePCI(virConnectPtr conn, static int virDomainXMLDevID(virDomainPtr domain, + virDomainDefPtr domDef, virDomainDeviceDefPtr dev, char *class, char *ref, @@ -4212,7 +4213,7 @@ xenDaemonAttachDeviceFlags(virDomainPtr domain, const char *xml, sexpr = virBufferContentAndReset(&buf); - if (virDomainXMLDevID(domain, dev, class, ref, sizeof(ref))) { + if (virDomainXMLDevID(domain, def, dev, class, ref, sizeof(ref))) { /* device doesn't exist, define it */ ret = xend_op(domain->conn, domain->name, "op", "device_create", "config", sexpr, NULL); @@ -4306,7 +4307,7 @@ xenDaemonDetachDeviceFlags(virDomainPtr domain, const char *xml, def, xml, VIR_DOMAIN_XML_INACTIVE))) goto cleanup; - if (virDomainXMLDevID(domain, dev, class, ref, sizeof(ref))) + if (virDomainXMLDevID(domain, def, dev, class, ref, sizeof(ref))) goto cleanup; if (dev->type == VIR_DOMAIN_DEVICE_HOSTDEV) { @@ -6048,6 +6049,7 @@ error: */ static int virDomainXMLDevID(virDomainPtr domain, + virDomainDefPtr domDef, virDomainDeviceDefPtr dev, char *class, char *ref, @@ -6056,27 +6058,33 @@ virDomainXMLDevID(virDomainPtr domain, xenUnifiedPrivatePtr priv = domain->conn->privateData; char *xref; char *tmp; + unsigned int i; + virDomainDiskDefPtr disk; if (dev->type == VIR_DOMAIN_DEVICE_DISK) { + if (dev->data.disk->dst == NULL) + return -1; if (dev->data.disk->driverName && STREQ(dev->data.disk->driverName, "tap")) strcpy(class, "tap"); else strcpy(class, "vbd"); - if (dev->data.disk->dst == NULL) - return -1; - xenUnifiedLock(priv); - xref = xenStoreDomainGetDiskID(domain->conn, domain->id, - dev->data.disk->dst); - xenUnifiedUnlock(priv); - if (xref == NULL) - return -1; - - tmp = virStrcpy(ref, xref, ref_len); - VIR_FREE(xref); - if (tmp == NULL) - return -1; + /* For disks, the device name can be used directly. + * If disk device exists in domain definintion, + * copy it to ref and return success. + */ + for (i = 0; i < domDef->ndisks; i++) { + disk = domDef->disks[i]; + if (STREQ(dev->data.disk->dst, disk->dst)) { + tmp = virStrcpy(ref, disk->dst, ref_len); + if (tmp == NULL) + return -1; + else + return 0; + } + } + return -1; } else if (dev->type == VIR_DOMAIN_DEVICE_NET) { char mac[30]; virDomainNetDefPtr def = dev->data.net; -- 1.6.0.2

3 4

[libvirt] running libvirt on Xen Cloud Platform 0.7.1
by Gadi Naor 02 Mar '10

02 Mar '10

Hi, I am trying to run libvirt daemon in Dom0 of Xen Cloud Platform 0.7.1. For that purpose I downloaded the libvirt source files 0.7.6 and trying to build the libvirt rpm with XEN driver using the XCP DDK. I am running into to many problems getting it done - so I was wondering whether this is supported or if anyone is using it in such way. Thank You, Gadi _________________________________________________________________ Hotmail: Free, trusted and rich email service. https://signup.live.com/signup.aspx?id=60969

1 0

[libvirt] and for LXC?
by Mihamina Rakotomandimby 01 Mar '10

01 Mar '10

Manao ahoana, Hello, Bonjour, Is there some plan/schedule to make libvirt LXC friendly? Or is it already? Misaotra, Thanks, Merci. -- Architecte Informatique chez Blueline/Gulfsat: Administration Systeme, Recherche & Developpement +261 34 29 155 34 / +261 33 11 207 36

3 2

[libvirt] (no subject)
by Daniel P. Berrange 01 Mar '10

01 Mar '10

Repost of http://www.redhat.com/archives/libvir-list/2010-February/msg00177.html The virsh output is now scaled to show KB/MB as needed, instead of always bytes. The public API gained extra status VIR_DOMAIN_JOB_COMPLETED = 3, /* Job has finished, but isn't cleaned up */ VIR_DOMAIN_JOB_FAILED = 4, /* Job hit error, but isn't cleaned up */ VIR_DOMAIN_JOB_CANCELLED = 6, /* Job was aborted, but isn't cleaned up */ to allow to more easily detecting end of a job, once we allow for truely async usage

2 21

[libvirt] [PATCH] [macvtap] interface type direct web docs addition
by Stefan Berger 01 Mar '10

01 Mar '10

This adds a description about the 'direct' type of interface recently added for macvtap device type support on the host.

3 2

[libvirt] [PATCH] Free the macvtap mode string
by Stefan Berger 01 Mar '10

01 Mar '10

I forgot to free the macvtap mode string. This fixes it. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com>

3 2

[libvirt] Hypervisor raw options
by Daniel Dehennin 01 Mar '10

01 Mar '10

Hello, I wonder if it's possible to add hypervisor pass-through options, I add some usage of kvm -startdate and need to use a script which call kvm with the desired option. Something like: <emulator>/usr/bin/kvm</emulator> <emulator-options> <option>-option1</option> <option>-option2</option> </emulator-options> The several <option/> will be passed to the <emulator/> command. Regards. -- Daniel Dehennin Récupérer ma clef GPG: gpg --keyserver pgp.mit.edu --recv-keys 0x6A2540D1

2 1

[libvirt] reload libvirtd.conf has a little problem.
by Osier Yang 01 Mar '10

01 Mar '10

hi When libvirtd is running, I changed "unix_sock_ro_perms=0777" into ""unix_sock_ro_perms=0755", and then reload the configuration with: kill -SIGHUP `pidof libvirtd` or "service libvirtd reload" the permission of /var/lib/libvirt/libvirt-sock-ro still is 0777, but not 0755. and after "service libvirtd restart", the permission is changed into 0755. think it's a little problem. libvirt version: libvirt-0.7.6-1.fc12.x86_64 Thanks and Regards osier

2 2

[libvirt] [PATCH] build: avoid warning about unused variables
by Jim Meyering 01 Mar '10

01 Mar '10

>From 8ddff3e6cdccc2b4289509c6941b43f2ddf8e643 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 14:19:33 +0100 Subject: [PATCH] build: avoid warning about unused variables * tools/virsh.c (cmdCPUBaseline): Remove declarations of unused variables, p and cur. --- tools/virsh.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 5fdbbe5..89eefcf 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -7048,12 +7048,11 @@ cmdCPUBaseline(vshControl *ctl, const vshCmd *cmd) int found; int ret = TRUE; char *buffer; - char *p; char *result = NULL; const char **list = NULL; unsigned int count = 0; xmlDocPtr doc = NULL; - xmlNodePtr node_list, cur; + xmlNodePtr node_list; xmlXPathContextPtr ctxt = NULL; xmlSaveCtxtPtr sctxt = NULL; xmlBufferPtr buf = NULL; -- 1.7.0.401.g84adb

2 1

[libvirt] [PATCH] libvirtd: do not ignore failure to set group ID in privileged mode
by Jim Meyering 27 Feb '10

27 Feb '10

Coverity spotted these unchecked setgid calls: >From 7eddaf457d54974859a8c22cced6f43e2d267746 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 26 Feb 2010 10:42:14 +0100 Subject: [PATCH] libvirtd: do not ignore failure to set group ID in privileged mode * daemon/libvirtd.c (qemudListenUnix): Diagnose and fail upon failure to set or restore group-ID. --- daemon/libvirtd.c | 12 ++++++++---- 1 files changed, 8 insertions(+), 4 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index cee6f13..9bdbecb 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -560,8 +560,10 @@ static int qemudListenUnix(struct qemud_server *server, oldgrp = getgid(); oldmask = umask(readonly ? ~unix_sock_ro_mask : ~unix_sock_rw_mask); - if (server->privileged) - setgid(unix_sock_gid); + if (server->privileged && setgid(unix_sock_gid)) { + VIR_ERROR(_("Failed to set group ID to %d"), unix_sock_gid); + goto cleanup; + } if (bind(sock->fd, (struct sockaddr *)&addr, sizeof(addr)) < 0) { VIR_ERROR(_("Failed to bind socket to '%s': %s"), @@ -569,8 +571,10 @@ static int qemudListenUnix(struct qemud_server *server, goto cleanup; } umask(oldmask); - if (server->privileged) - setgid(oldgrp); + if (server->privileged && setgid(oldgrp)) { + VIR_ERROR(_("Failed to restore group ID to %d"), oldgrp); + goto cleanup; + } if (listen(sock->fd, 30) < 0) { VIR_ERROR(_("Failed to listen for connections on '%s': %s"), -- 1.7.0.442.g55ad1

2 2

[libvirt] [PATCH] build: avoid warning about return-with-value in void function
by Jim Meyering 27 Feb '10

27 Feb '10

I saw this: virsh.c: In function 'virshReportError': virsh.c:363: warning: 'return' with a value, in function returning void >From fc83f08ac204cec5fba474a23bd19386d1a415d7 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 26 Feb 2010 11:50:34 +0100 Subject: [PATCH] build: avoid warning about return-with-value in void function * tools/virsh.c: Just "return;", not "return NULL;". --- tools/virsh.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index f7fd7d4..3b1011c 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -360,7 +360,7 @@ virshReportError(vshControl *ctl) * no error was ever raised, so just ignore */ last_error = virSaveLastError(); if (!last_error || last_error->code == VIR_ERR_OK) - return NULL; + return; } if (last_error->code == VIR_ERR_OK) { -- 1.7.0.442.g55ad1

2 1

[libvirt] build: update gnulib submodule to latest
by Jim Meyering 27 Feb '10

27 Feb '10

Thanks, Eric. >From bd3ff6b34892fe75cfcc67f1aeab43e8a273c1b6 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 26 Feb 2010 17:14:01 +0100 Subject: [PATCH] build: update gnulib submodule to latest * .gnulib: Update to latest. Commit 89bdf84bcd9c6032e37 inadvertently rewound the .gnulib submodule by 51 commits. This corrects it. Spotted by Eric Blake. --- .gnulib | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/.gnulib b/.gnulib index 11fbc57..80cd995 160000 --- a/.gnulib +++ b/.gnulib @@ -1 +1 @@ -Subproject commit 11fbc57405a118e6ec9a3ebc19bbf5ececdae4d6 +Subproject commit 80cd995cdcbf4b9ded895a43621a11f11806ad8d -- 1.7.0.442.g55ad1

2 1

[libvirt] [PATCH] Only build virDomainObjFormat if not building proxy.
by Chris Lalancette 26 Feb '10

26 Feb '10

While building under RHEL-5, I got a compile warning because virDomainObjFormat was defined but not used. That came about because in RHEL-5 we build with "#define PROXY", and virDomainObjFormat is only used with !PROXY. Move the define. Signed-off-by: Chris Lalancette <clalance(a)redhat.com> --- src/conf/domain_conf.c | 5 ++--- 1 files changed, 2 insertions(+), 3 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index b6d2c61..1d7f458 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -5668,6 +5668,8 @@ char *virDomainDefFormat(virDomainDefPtr def, return NULL; } +#ifndef PROXY + static char *virDomainObjFormat(virCapsPtr caps, virDomainObjPtr obj, int flags) @@ -5703,9 +5705,6 @@ error: return NULL; } - -#ifndef PROXY - int virDomainSaveXML(const char *configDir, virDomainDefPtr def, const char *xml) -- 1.6.6.1

3 3

[libvirt] Odd implementation of libvirt qemudDomainSetMaxMemory
by Chris Lalancette 26 Feb '10

26 Feb '10

Hey Cole, In commit 09a33fd8a9cc85e5e1064168472bff7aecff15fb, you made it so that qemudDomainSetMaxMemory() was hotplug only. However, with the changes you pushed, this function no longer makes a lot of sense. It looks up the domain, checks to make sure it is active, and then if it is active, it fails. Assuming that qemu doesn't support changing the maximum amount of memory on the fly, wouldn't it be better to just remove this function entirely and let the higher libvirt layers report "function not supported"? -- Chris Lalancette

2 1

[libvirt] [PATCH] openvzGetVEID: don't leak (memory + file descriptor)
by Jim Meyering 26 Feb '10

26 Feb '10

Coverity spotted the leaks. Here's a proposed fix. I didn't bother with a separate diagnostic for the unlikely event that we read a negative number from the pipe. Seems far-fetched enough not to bother, but it's easy to add, if anyone cares. >From f3439c7eae46681eacf5b469a6f0e22cb8fec1b4 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 19:24:50 +0100 Subject: [PATCH] openvzGetVEID: don't leak (memory + file descriptor) * src/openvz/openvz_conf.c (openvzGetVEID): Always call fclose. Diagnose parse failure also when vzlist output is empty. If somehow we read a -1, diagnose that (albeit as a parse failure). --- src/openvz/openvz_conf.c | 19 +++++++------------ 1 files changed, 7 insertions(+), 12 deletions(-) diff --git a/src/openvz/openvz_conf.c b/src/openvz/openvz_conf.c index ce0c4d3..3713a45 100644 --- a/src/openvz/openvz_conf.c +++ b/src/openvz/openvz_conf.c @@ -964,6 +964,7 @@ int openvzGetVEID(const char *name) { char *cmd; int veid; FILE *fp; + bool ok; if (virAsprintf(&cmd, "%s %s -ovpsid -H", VZLIST, name) < 0) { openvzError(NULL, VIR_ERR_INTERNAL_ERROR, "%s", @@ -979,18 +980,12 @@ int openvzGetVEID(const char *name) { return -1; } - if (fscanf(fp, "%d\n", &veid ) != 1) { - if (feof(fp)) - return -1; - - openvzError(NULL, VIR_ERR_INTERNAL_ERROR, - "%s", _("Failed to parse vzlist output")); - goto cleanup; - } - - return veid; - - cleanup: + ok = fscanf(fp, "%d\n", &veid ) == 1; fclose(fp); + if (ok && veid >= 0) + return veid; + + openvzError(NULL, VIR_ERR_INTERNAL_ERROR, + "%s", _("Failed to parse vzlist output")); return -1; } -- 1.7.0.401.g84adb

2 3

[libvirt] avoid "make rpm" failure in doc/
by Jim Meyering 26 Feb '10

26 Feb '10

"make rpm" fails like this: Rebuilding the HTML pages from the XML API make[3]: *** No rule to make target `html/libvirt-libvirt.html', needed by `distdir'. Stop. make[3]: *** Waiting for unfinished jobs.... Validating the resulting XHTML pages These two patches fix that: (though "make rpm" still fails for me: extracting debug info from /j/rpmbuild/BUILDROOT/libvirt-0.7.6-1.fc12.x86_6h.. *** ERROR: No build ID note found in /j/rpmbuild/BUILDROOT/libvirt-0.7.6-1.h.. The first (1/2) change is not officially required to solve this problem, but it overlapped, and does fix a non-distdir build failure. >From ace4dca5dfbc022b057ce1d9f3d2770ad3a34716 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 22:51:47 +0100 Subject: [PATCH 1/2] build: teach apibuild.py to work in a non-srcdir build * docs/Makefile.am (libvirt-api.xml libvirt-refs.xml): Generalize apibuild.py to work in a non-srcdir build. Pass "srcdir" to it. * docs/apibuild.py (rebuild): Honor the $srcdir envvar. --- docs/Makefile.am | 2 +- docs/apibuild.py | 10 +++++++--- 2 files changed, 8 insertions(+), 4 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index e47b668..dbbd1ef 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -132,7 +132,7 @@ devhelp/index.html devhelp/libvirt.devhelp: libvirt-api.xml $(devhelpxsl) libvirt-api.xml libvirt-refs.xml: apibuild.py \ $(srcdir)/../include/libvirt/*.h \ $(srcdir)/../src/libvirt.c $(srcdir)/../src/util/virterror.c - -$(srcdir)/apibuild.py + -srcdir=$(srcdir) $(srcdir)/apibuild.py clean-local: rm -f *~ *.bak *.hierarchy *.signals *-unused.txt diff --git a/docs/apibuild.py b/docs/apibuild.py index e233c8b..0ab5db2 100755 --- a/docs/apibuild.py +++ b/docs/apibuild.py @@ -2102,10 +2102,14 @@ class docBuilder: def rebuild(): builder = None - if glob.glob("../src/libvirt.c") != [] : + srcdir = os.environ["srcdir"] + if glob.glob(srcdir + "/../src/libvirt.c") != [] : print "Rebuilding API description for libvirt" - builder = docBuilder("libvirt", ["../src", "../src/util", "../include/libvirt"], - []) + builder = docBuilder("libvirt", + [srcdir + "/../src", + srcdir + "/../src/util", + srcdir + "/../include/libvirt"], + []) elif glob.glob("src/libvirt.c") != [] : print "Rebuilding API description for libvir" builder = docBuilder("libvirt", ["src", "src/util", "include/libvirt"], -- 1.7.0.401.g84adb >From a79fb1c491197a91094cd2345c3d7bbba40da2fd Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 10:35:20 +0100 Subject: [PATCH 2/2] build: avoid "make rpm" failure in docs/ Add missing rule to build html/libvirt-libvirt.html. Use a GNU Make pattern rule to avoid running apibuild.py once for each out-of-date target, in a parallel build. * docs/Makefile.am --- docs/Makefile.am | 15 +++++++++------ 1 files changed, 9 insertions(+), 6 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index dbbd1ef..edf6fc8 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -123,17 +123,20 @@ html/index.html: libvirt-api.xml newapi.xsl page.xsl sitemap.html.in $(XMLLINT) --nonet --valid --noout html/*.html ; \ else echo "missing XHTML1 DTD" ; fi ; fi ); -devhelp/index.html devhelp/libvirt.devhelp: libvirt-api.xml $(devhelpxsl) +$(devhelphtml): libvirt-api.xml $(devhelpxsl) -@(echo Rebuilding devhelp files) -@(if [ -x $(XSLTPROC) ] ; then \ - $(XSLTPROC) --nonet -o devhelp/libvirt.devhelp $(top_srcdir)/docs/devhelp/devhelp.xsl $(top_srcdir)/docs/libvirt-api.xml ; fi ); + $(XSLTPROC) --nonet -o devhelp/libvirt.devhelp \ + $(top_srcdir)/docs/devhelp/devhelp.xsl libvirt-api.xml ; fi ); - -libvirt-api.xml libvirt-refs.xml: apibuild.py \ - $(srcdir)/../include/libvirt/*.h \ - $(srcdir)/../src/libvirt.c $(srcdir)/../src/util/virterror.c +html/%-%.html html/%-virterror.html %-api.xml %-refs.xml: $(srcdir)/apibuild.py -srcdir=$(srcdir) $(srcdir)/apibuild.py +html/%-%.html html/%-virterror.html %-api.xml %-refs.xml: \ + $(srcdir)/../include/%/*.h \ + $(srcdir)/../src/%.c \ + $(srcdir)/../src/util/virterror.c + clean-local: rm -f *~ *.bak *.hierarchy *.signals *-unused.txt -- 1.7.0.401.g84adb

2 3

[libvirt] [PATCH] build: avoid non-srcdir "make distcheck" failures (CLEANFILES)
by Jim Meyering 26 Feb '10

26 Feb '10

A quick review won't hurt, but this doesn't really need one. >From 9ac893a1342b31ef7df6fc57c03767c49ef98d5b Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Thu, 25 Feb 2010 09:28:51 +0100 Subject: [PATCH 1/4] build: avoid non-srcdir "make distcheck" failures (CLEANFILES) * docs/Makefile.am (MAINTAINERCLEANFILES): Use this variable for generated-and-distributed files, not "CLEANFILES". Besides, "make clean" and "make distclean" should not delete distributed files. --- docs/Makefile.am | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index 07ea9bf..dbbd1ef 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -17,7 +17,7 @@ apipng = \ html/home.png \ html/right.png -devhelphtml = \ +devhelphtml = \ devhelp/libvirt.devhelp \ devhelp/index.html \ devhelp/general.html \ @@ -79,7 +79,7 @@ EXTRA_DIST= \ sitemap.html.in \ ChangeLog.awk -CLEANFILES = $(dot_html) $(apihtml) $(devhelphtml) +MAINTAINERCLEANFILES = $(dot_html) $(apihtml) $(devhelphtml) all: web -- 1.7.0.401.g84adb

2 1

[libvirt] [PATCH] qemu: Report binary path if error parsing -help
by Cole Robinson 26 Feb '10

26 Feb '10

Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/qemu/qemu_conf.c | 14 ++++++++------ src/qemu/qemu_conf.h | 3 ++- tests/qemuhelptest.c | 3 ++- 3 files changed, 12 insertions(+), 8 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index e3da08a..1f8a9c3 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -1237,7 +1237,8 @@ static unsigned long long qemudComputeCmdFlags(const char *help, #define SKIP_BLANKS(p) do { while ((*(p) == ' ') || (*(p) == '\t')) (p)++; } while (0) -int qemudParseHelpStr(const char *help, +int qemudParseHelpStr(const char *qemu, + const char *help, unsigned long long *flags, unsigned int *version, unsigned int *is_kvm, @@ -1308,8 +1309,8 @@ fail: p = strndup(help, p - help); qemuReportError(VIR_ERR_INTERNAL_ERROR, - _("cannot parse QEMU version number in '%s'"), - p ? p : help); + _("cannot parse %s version number in '%s'"), + qemu, p ? p : help); VIR_FREE(p); @@ -1340,12 +1341,13 @@ int qemudExtractVersionInfo(const char *qemu, enum { MAX_HELP_OUTPUT_SIZE = 1024*64 }; int len = virFileReadLimFD(newstdout, MAX_HELP_OUTPUT_SIZE, &help); if (len < 0) { - virReportSystemError(errno, "%s", - _("Unable to read QEMU help output")); + virReportSystemError(errno, + _("Unable to read %s help output"), qemu); goto cleanup2; } - if (qemudParseHelpStr(help, &flags, &version, &is_kvm, &kvm_version) == -1) + if (qemudParseHelpStr(qemu, help, &flags, + &version, &is_kvm, &kvm_version) == -1) goto cleanup2; if (retversion) diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h index ae187dc..c8757c5 100644 --- a/src/qemu/qemu_conf.h +++ b/src/qemu/qemu_conf.h @@ -175,7 +175,8 @@ int qemudExtractVersionInfo (const char *qemu, unsigned int *version, unsigned long long *qemuCmdFlags); -int qemudParseHelpStr (const char *str, +int qemudParseHelpStr (const char *qemu, + const char *str, unsigned long long *qemuCmdFlags, unsigned int *version, unsigned int *is_kvm, diff --git a/tests/qemuhelptest.c b/tests/qemuhelptest.c index 0f2b509..ad355d7 100644 --- a/tests/qemuhelptest.c +++ b/tests/qemuhelptest.c @@ -49,7 +49,8 @@ static int testHelpStrParsing(const void *data) if (virtTestLoadFile(path, &help, MAX_HELP_OUTPUT_SIZE) < 0) return -1; - if (qemudParseHelpStr(help, &flags, &version, &is_kvm, &kvm_version) == -1) + if (qemudParseHelpStr("QEMU", help, &flags, + &version, &is_kvm, &kvm_version) == -1) return -1; if (flags != info->flags) { -- 1.6.5.2

2 2

[libvirt] [PATCH] remote: Improve daemon startup error reporting
by Cole Robinson 25 Feb '10

25 Feb '10

If I toggle enable_tcp in libvirtd.conf and add --listen in /etc/init.d/libvirtd, I get the unhelpful error: Starting libvirtd daemon: error: Unable to initialize network sockets. Running without --daemon provides much more useful info: sudo libvirtd --listen 11:29:26.117: error : remoteCheckCertFile:270 : Cannot access CA certificate '/etc/pki/CA/cacert.pem': No such file or directory The daemon architecture makes it difficult to report this useful info if daemonized, so point users to /var/log/messages and dropping the --daemon flag if they want more info. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- daemon/libvirtd.c | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/daemon/libvirtd.c b/daemon/libvirtd.c index fefa7aa..cee6f13 100644 --- a/daemon/libvirtd.c +++ b/daemon/libvirtd.c @@ -474,7 +474,10 @@ static int daemonForkIntoBackground(void) { goto again; if (ret == 1 && status != 0) { - fprintf(stderr, "error: %s\n", virDaemonErrTypeToString(status)); + fprintf(stderr, + "error: %s. Check /var/log/messages or run without " + "--daemon for more info.\n", + virDaemonErrTypeToString(status)); } _exit(ret == 1 && status == 0 ? 0 : 1); } -- 1.6.5.2

2 1

[libvirt] [PATCH] virsh: Show errors reported by nonAPI functions
by Cole Robinson 25 Feb '10

25 Feb '10

Only API calls trigger the error callback, which is required for proper virsh error reporting. Since we use non API functions from util/, make sure we properly report these errors. Fixes lack of error message from 'virsh create idontexit.xml' Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- tools/virsh.c | 12 ++++++++++-- 1 files changed, 10 insertions(+), 2 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 5fdbbe5..74e3fad 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -352,8 +352,16 @@ virshErrorHandler(void *unused ATTRIBUTE_UNUSED, virErrorPtr error) static void virshReportError(vshControl *ctl) { - if (last_error == NULL) - return; + if (last_error == NULL) { + /* Calling directly into libvirt util functions won't trigger the + * error callback (which sets last_error), so check it ourselves. + * + * If the returned error has CODE_OK, this most likely means that + * no error was ever raised, so just ignore */ + last_error = virSaveLastError(); + if (!last_error || last_error->code == VIR_ERR_OK) + return NULL; + } if (last_error->code == VIR_ERR_OK) { vshError(ctl, "%s", _("unknown error")); -- 1.6.5.2

2 1

[libvirt] [PATCH] Use standard spacing for user/pass prompt
by Cole Robinson 25 Feb '10

25 Feb '10

Kind of minor, but it annoys me that the default auth callback doesn't put a space between the prompt and the input, like a typical terminal, ssh, etc. This patch changes the current prompt: Please enter your authentication name:myuser to Please enter your authentication name: myuser Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/libvirt.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index be17668..9d50c92 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -159,7 +159,7 @@ static int virConnectAuthCallbackDefault(virConnectCredentialPtr cred, case VIR_CRED_AUTHNAME: case VIR_CRED_ECHOPROMPT: case VIR_CRED_REALM: - if (printf("%s:", cred[i].prompt) < 0) + if (printf("%s: ", cred[i].prompt) < 0) return -1; if (fflush(stdout) != 0) return -1; @@ -178,7 +178,7 @@ static int virConnectAuthCallbackDefault(virConnectCredentialPtr cred, case VIR_CRED_PASSPHRASE: case VIR_CRED_NOECHOPROMPT: - if (printf("%s:", cred[i].prompt) < 0) + if (printf("%s: ", cred[i].prompt) < 0) return -1; if (fflush(stdout) != 0) return -1; -- 1.6.5.2

2 1

[libvirt] the price we pay for using gettext-0.14
by Jim Meyering 25 Feb '10

25 Feb '10

Make distcheck currently fails like this: make[2]: Entering directory `/t/libvirt-0.7.6/_build/po' /bin/sh @MKINSTALLDIRS@ /t/libvirt-0.7.6/_inst/share /bin/sh: @MKINSTALLDIRS@: No such file or directory make[2]: *** [install-data-yes] Error 127 That is due to our use of a version of gettext that is way too old: $ grep GETT configure.ac AM_GNU_GETTEXT_VERSION([0.14.1]) The work-around is to define (via AC_SUBST) the now-obsolete symbol, MKINSTALLDIRS, so that @MKINSTALLDIRS@ in the autopoint-provided Makefile.in.in (from the old version of gettext) ends up being substituted. As I recall, people insist on using this old version of gettext, because nothing newer is supplied by RHEL5.x. IMHO, that is RHEL5's problem, and should not be allowed to cause continuing problems in upstream projects. It is easy and safe to install a newer version of gettext, especially since any project using autopoint will automatically continue to get the same bit-rotting old files: a gettext installation includes an archive of all older versions of its files, to provide just this type of backwards compatibility. We had a similar problem in libguestfs, and it was not trivial to diagnose there, either, even though I'd seen it at least once or twice before then. Here's the work-around: [just realized this requires a comment. adding that now] >From 16b9418427101a495b7fdd6917f518f254d22720 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 10:53:44 +0100 Subject: [PATCH] build: ensure that MKINSTALLDIRS is AC_SUBST-defined since we're using gettext-0.14.1, which uses that now-obsolete automake symbol. Otherwise, make distcheck would fails like this: make[2]: Entering directory `/t/libvirt-0.7.6/_build/po' /bin/sh @MKINSTALLDIRS@ /t/libvirt-0.7.6/_inst/share /bin/sh: @MKINSTALLDIRS@: No such file or directory make[2]: *** [install-data-yes] Error 127 * configure.ac (MKINSTALLDIRS): Define. --- configure.ac | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/configure.ac b/configure.ac index 117cb20..2f79970 100644 --- a/configure.ac +++ b/configure.ac @@ -49,6 +49,8 @@ AC_PROG_CPP AC_OBJEXT +AC_SUBST([MKINSTALLDIRS], [$ac_aux_dir/mkinstalldirs]) + dnl gl_INIT uses m4_foreach_w, yet that is not defined in autoconf-2.59. dnl In order to accommodate developers with such old tools, here's a dnl replacement definition. -- 1.7.0.367.g566c3

3 4

[libvirt] [PATCH] maint: relax git minimum version
by Eric Blake 25 Feb '10

25 Feb '10

From: Eric Blake <eblake(a)redhat.com> Requiring git 1.6.4, just for the optional GNULIB_SRCDIR support, was too harsh. Resynchronize from gnulib. * .gnulib: Import from latest gnulib. * bootstrap: Re-synchronize from .gnulib/build-aux. * bootstrap.conf: Drop git to 1.5.5. * README-hacking: Document use of GNULIB_SRCDIR. --- Chris is currently testing that RHEL 5 with git 1.5.5 can actually grok this patch. We don't have a README-prereq (but README-hacking was borrowed from coreutils, which does). Should we add one? The README-prereq in coreutils is not quite relevant for a straight-across copy. .gnulib | 2 +- README-hacking | 9 +++++++-- bootstrap | 41 +++++++++++++++++++++++++++++++++++------ bootstrap.conf | 2 +- 4 files changed, 44 insertions(+), 10 deletions(-) diff --git a/.gnulib b/.gnulib index 11fbc57..108cfb8 160000 --- a/.gnulib +++ b/.gnulib @@ -1 +1 @@ -Subproject commit 11fbc57405a118e6ec9a3ebc19bbf5ececdae4d6 +Subproject commit 108cfb803174abf17eb101e9999b9d1a642ce1ba diff --git a/README-hacking b/README-hacking index 4105a3e..80b022c 100644 --- a/README-hacking +++ b/README-hacking @@ -12,8 +12,7 @@ requirements on anyone wishing to build from the just-checked-out sources. Note the requirements to build the released archive are much less and are just the requirements of the standard ./configure && make procedure. Specific development tools and versions will be checked for and listed by -the bootstrap script. See README-prereq for specific notes on obtaining -these prerequisite tools. +the bootstrap script. Valgrind <http://valgrind.org/> is also highly recommended, if Valgrind supports your architecture. See also README-valgrind. @@ -28,6 +27,12 @@ You can get a copy of the source repository like this: $ git clone git://libvirt.org/libvirt $ cd libvirt +As an optional step, if you already have a copy of the gnulib git +repository on your hard drive, then you can use it as a reference to +reduce download time and disk space requirements: + + $ export GNULIB_SRCDIR=/path/to/gnulib + The next step is to get all required pieces from gnulib, to run autoreconf, and to invoke ./configure: diff --git a/bootstrap b/bootstrap index be96444..365a3d9 100755 --- a/bootstrap +++ b/bootstrap @@ -1,4 +1,6 @@ #! /bin/sh +# Print a version string. +scriptversion=2010-02-24.17; # UTC # Bootstrap this package from checked-out sources. @@ -17,7 +19,15 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. -# Written by Paul Eggert. +# Originally written by Paul Eggert. The canonical version of this +# script is maintained as build-aux/bootstrap in gnulib, however, to +# be useful to your project, you should place a copy of it under +# version control in the top-level directory of your project. The +# intent is that all customization can be done with a bootstrap.conf +# file also maintained in your version control; gnulib comes with a +# template build-aux/bootstrap.conf to get you started. + +# Please report bugs or propose patches to bug-gnulib(a)gnu.org. nl=' ' @@ -161,7 +171,6 @@ vc_ignore=auto # die otherwise. find_tool () { - # Find sha1sum, named gsha1sum on MacPorts. find_tool_envvar=$1 shift find_tool_names=$@ @@ -424,12 +433,28 @@ case ${GNULIB_SRCDIR--} in GNULIB_SRCDIR=$gnulib_path ;; *) - # Use GNULIB_SRCDIR as a reference. Assumes git 1.6.4 or newer. + # Use GNULIB_SRCDIR as a reference. if test -d "$GNULIB_SRCDIR"/.git && \ git_modules_config submodule.gnulib.url >/dev/null; then echo "$0: getting gnulib files..." - git submodule update --init --reference "$GNULIB_SRCDIR" \ + if git submodule --help|grep reference > /dev/null; then + # Prefer the one-liner available in git 1.6.4 or newer. + git submodule update --init --reference "$GNULIB_SRCDIR" \ "$gnulib_path" || exit $? + else + # This fallback allows at least git 1.5.5. + if test -f "$gnulib_path"/gnulib-tool; then + # Since file already exists, assume submodule init already complete. + git submodule update || exit $? + else + # Older git can't clone into an empty directory. + rmdir "$gnulib_path" 2>/dev/null + git clone --reference "$GNULIB_SRCDIR" \ + "$(git_modules_config submodule.gnulib.url)" "$gnulib_path" \ + && git submodule init && git submodule update \ + || exit $? + fi + fi GNULIB_SRCDIR=$gnulib_path fi ;; @@ -819,6 +844,10 @@ bootstrap_epilogue echo "$0: done. Now you can run './configure'." -# Local Variables: -# indent-tabs-mode: nil +# Local variables: +# eval: (add-hook 'write-file-hooks 'time-stamp) +# time-stamp-start: "scriptversion=" +# time-stamp-format: "%:y-%02m-%02d.%02H" +# time-stamp-time-zone: "UTC" +# time-stamp-end: "; # UTC" # End: diff --git a/bootstrap.conf b/bootstrap.conf index d315e04..58ef2ab 100644 --- a/bootstrap.conf +++ b/bootstrap.conf @@ -124,7 +124,7 @@ autoconf 2.59 automake 1.9.6 autopoint - gettext - -git 1.6.4 +git 1.5.5 gzip - libtool - perl 5.5 -- 1.6.6.1

2 2

[libvirt] [PATCH v2] libvirtd: Better initscript error reporting
by Cole Robinson 25 Feb '10

25 Feb '10

>From time to time I bork my install, and hate it when the initscript returns no info. This patch removes the sanity check, which lets the shell give us 'command not found' or 'permission denied' errors. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- daemon/libvirtd.init.in | 3 --- 1 files changed, 0 insertions(+), 3 deletions(-) diff --git a/daemon/libvirtd.init.in b/daemon/libvirtd.init.in index 071fe2c..b808ab3 100644 --- a/daemon/libvirtd.init.in +++ b/daemon/libvirtd.init.in @@ -32,9 +32,6 @@ # pidfile: @localstatedir@/run/libvirtd.pid # -# Sanity checks. -[ -x @sbindir@/libvirtd ] || exit 0 - # Source function library. . @sysconfdir@/rc.d/init.d/functions -- 1.6.5.2

1 0

[libvirt] [PATCH] remote: Improve error message when libvirtd isn't running
by Cole Robinson 25 Feb '10

25 Feb '10

While this could lead people in the wrong direction, people most commonly hit this error when libvirtd isn't running, so let's mention a possible fix. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/remote/remote_driver.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index 8914c39..e10312c 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -723,8 +723,8 @@ doRemoteOpen (virConnectPtr conn, } } virReportSystemError(errno, - _("unable to connect to '%s'"), - sockname); + _("unable to connect to '%s', libvirtd may need to be started"), + sockname); goto failed; } -- 1.6.5.2

2 1

[libvirt] [PATCH] libvirtd: Better initscript error reporting
by Cole Robinson 25 Feb '10

25 Feb '10

>From time to time I bork my install, and hate it when the initscript returns no info. This patch makes things a bit more clear. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- daemon/libvirtd.init.in | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/daemon/libvirtd.init.in b/daemon/libvirtd.init.in index 071fe2c..c5d4eee 100644 --- a/daemon/libvirtd.init.in +++ b/daemon/libvirtd.init.in @@ -32,8 +32,10 @@ # pidfile: @localstatedir@/run/libvirtd.pid # +LIBVIRTD_BIN="@sbindir@/libvirtd" + # Sanity checks. -[ -x @sbindir@/libvirtd ] || exit 0 +[ ! -e $LIBVIRTD_BIN ] && echo "$LIBVIRTD_BIN not found." && exit 1 # Source function library. . @sysconfdir@/rc.d/init.d/functions -- 1.6.5.2

2 2

[libvirt] [PATCH] Ignore SIGWINCH in remote client call to poll(2) (RHBZ#567931).
by Richard W.M. Jones 25 Feb '10

25 Feb '10

-- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into Xen guests. http://et.redhat.com/~rjones/virt-p2v

3 6

[libvirt] [PATCH] conf: storage: Remove iSCSI <auth> parsing
by Cole Robinson 25 Feb '10

25 Feb '10

This was never wired up, and even generated broken XML until 0.7.2, so clearly no one was trying to use it. Dan recommended its removal, so lets drop it. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- docs/schemas/storagepool.rng | 19 --------- src/conf/storage_conf.c | 48 ----------------------- src/conf/storage_conf.h | 18 -------- tests/storagepoolxml2xmlin/pool-iscsi-auth.xml | 17 -------- tests/storagepoolxml2xmlout/pool-iscsi-auth.xml | 20 --------- tests/storagepoolxml2xmltest.c | 1 - 6 files changed, 0 insertions(+), 123 deletions(-) delete mode 100644 tests/storagepoolxml2xmlin/pool-iscsi-auth.xml delete mode 100644 tests/storagepoolxml2xmlout/pool-iscsi-auth.xml diff --git a/docs/schemas/storagepool.rng b/docs/schemas/storagepool.rng index 247664e..bcdca62 100644 --- a/docs/schemas/storagepool.rng +++ b/docs/schemas/storagepool.rng @@ -234,22 +234,6 @@ </element> </define> - <define name='sourceinfoauth'> - <element name='auth'> - <attribute name='type'> - <choice> - <value>chap</value> - </choice> - </attribute> - <attribute name='login'> - <text/> - </attribute> - <attribute name='passwd'> - <text/> - </attribute> - </element> - </define> - <define name='sourcefmtfs'> <optional> <element name='format'> @@ -374,9 +358,6 @@ <optional> <ref name='initiatorinfoiqn'/> </optional> - <optional> - <ref name='sourceinfoauth'/> - </optional> </element> </define> diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c index 19a1db9..dd375b9 100644 --- a/src/conf/storage_conf.c +++ b/src/conf/storage_conf.c @@ -283,11 +283,6 @@ virStoragePoolSourceFree(virStoragePoolSourcePtr source) { VIR_FREE(source->name); VIR_FREE(source->adapter); VIR_FREE(source->initiator.iqn); - - if (source->authType == VIR_STORAGE_POOL_AUTH_CHAP) { - VIR_FREE(source->auth.chap.login); - VIR_FREE(source->auth.chap.passwd); - } } void @@ -363,26 +358,6 @@ virStoragePoolObjRemove(virStoragePoolObjListPtr pools, static int -virStoragePoolDefParseAuthChap(xmlXPathContextPtr ctxt, - virStoragePoolAuthChapPtr auth) { - auth->login = virXPathString("string(./auth/@login)", ctxt); - if (auth->login == NULL) { - virStorageReportError(VIR_ERR_XML_ERROR, - "%s", _("missing auth host attribute")); - return -1; - } - - auth->passwd = virXPathString("string(./auth/@passwd)", ctxt); - if (auth->passwd == NULL) { - virStorageReportError(VIR_ERR_XML_ERROR, - "%s", _("missing auth passwd attribute")); - return -1; - } - - return 0; -} - -static int virStoragePoolDefParseSource(xmlXPathContextPtr ctxt, virStoragePoolSourcePtr source, int pool_type, @@ -445,25 +420,6 @@ virStoragePoolDefParseSource(xmlXPathContextPtr ctxt, source->dir = virXPathString("string(./dir/@path)", ctxt); source->adapter = virXPathString("string(./adapter/@name)", ctxt); - authType = virXPathString("string(./auth/@type)", ctxt); - if (authType == NULL) { - source->authType = VIR_STORAGE_POOL_AUTH_NONE; - } else { - if (STREQ(authType, "chap")) { - source->authType = VIR_STORAGE_POOL_AUTH_CHAP; - } else { - virStorageReportError(VIR_ERR_XML_ERROR, - _("unknown auth type '%s'"), - (const char *)authType); - goto cleanup; - } - } - - if (source->authType == VIR_STORAGE_POOL_AUTH_CHAP) { - if (virStoragePoolDefParseAuthChap(ctxt, &source->auth.chap) < 0) - goto cleanup; - } - ret = 0; cleanup: ctxt->node = relnode; @@ -867,10 +823,6 @@ virStoragePoolSourceFormat(virBufferPtr buf, } - if (src->authType == VIR_STORAGE_POOL_AUTH_CHAP) - virBufferVSprintf(buf," <auth type='chap' login='%s' passwd='%s'/>\n", - src->auth.chap.login, - src->auth.chap.passwd); virBufferAddLit(buf," </source>\n"); return 0; diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h index c643984..1408128 100644 --- a/src/conf/storage_conf.h +++ b/src/conf/storage_conf.h @@ -134,19 +134,6 @@ enum virStoragePoolDeviceType { }; -enum virStoragePoolAuthType { - VIR_STORAGE_POOL_AUTH_NONE, - VIR_STORAGE_POOL_AUTH_CHAP, -}; - -typedef struct _virStoragePoolAuthChap virStoragePoolAuthChap; -typedef virStoragePoolAuthChap *virStoragePoolAuthChapPtr; -struct _virStoragePoolAuthChap { - char *login; - char *passwd; -}; - - /* * For remote pools, info on how to reach the host */ @@ -232,11 +219,6 @@ struct _virStoragePoolSource { /* Initiator IQN */ virStoragePoolSourceInitiatorAttr initiator; - int authType; /* virStoragePoolAuthType */ - union { - virStoragePoolAuthChap chap; - } auth; - int format; /* Pool type specific format such as filesystem type, or lvm version, etc */ }; diff --git a/tests/storagepoolxml2xmlin/pool-iscsi-auth.xml b/tests/storagepoolxml2xmlin/pool-iscsi-auth.xml deleted file mode 100644 index f7d4d52..0000000 --- a/tests/storagepoolxml2xmlin/pool-iscsi-auth.xml +++ /dev/null @@ -1,17 +0,0 @@ -<pool type='iscsi'> - <name>virtimages</name> - <uuid>e9392370-2917-565e-692b-d057f46512d6</uuid> - <source> - <host name="iscsi.example.com"/> - <device path="demo-target"/> - <auth type='chap' login='foobar' passwd='frobbar'/> - </source> - <target> - <path>/dev/disk/by-path</path> - <permissions> - <mode>0700</mode> - <owner>0</owner> - <group>0</group> - </permissions> - </target> -</pool> diff --git a/tests/storagepoolxml2xmlout/pool-iscsi-auth.xml b/tests/storagepoolxml2xmlout/pool-iscsi-auth.xml deleted file mode 100644 index 557295d..0000000 --- a/tests/storagepoolxml2xmlout/pool-iscsi-auth.xml +++ /dev/null @@ -1,20 +0,0 @@ -<pool type='iscsi'> - <name>virtimages</name> - <uuid>e9392370-2917-565e-692b-d057f46512d6</uuid> - <capacity>0</capacity> - <allocation>0</allocation> - <available>0</available> - <source> - <host name='iscsi.example.com'/> - <device path='demo-target'/> - <auth type='chap' login='foobar' passwd='frobbar'/> - </source> - <target> - <path>/dev/disk/by-path</path> - <permissions> - <mode>0700</mode> - <owner>0</owner> - <group>0</group> - </permissions> - </target> -</pool> diff --git a/tests/storagepoolxml2xmltest.c b/tests/storagepoolxml2xmltest.c index 4550407..1d7094b 100644 --- a/tests/storagepoolxml2xmltest.c +++ b/tests/storagepoolxml2xmltest.c @@ -91,7 +91,6 @@ mymain(int argc, char **argv) DO_TEST("pool-logical-create"); DO_TEST("pool-disk"); DO_TEST("pool-iscsi"); - DO_TEST("pool-iscsi-auth"); DO_TEST("pool-netfs"); DO_TEST("pool-scsi"); DO_TEST("pool-mpath"); -- 1.6.6

3 3

[libvirt] [PATCH 1/2] build: avoid non-srcdir installation failure (apibuild.py)
by Jim Meyering 25 Feb '10

25 Feb '10

Running "make distcheck" from a non-srcdir build was failing like this: /bin/sh: ./apibuild.py: No such file or directory Fixed that with the first patch below (the apibuild.py script is not in ".", but in $(srcdir)). Then retrying, it got a little further but failed like this: (../../docs/apibuild.py) make[3]: *** No rule to make target `sitemap.html.in', needed by `html/index.html'. Stop. That was because sitemap.html.in was not being distributed. Adding it to EXTRA_DIST solved that: >From f0fb87ed1838a9a67d03985e35a4c7625da9d95e Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 14:18:40 +0100 Subject: [PATCH 1/2] build: avoid non-srcdir installation failure (apibuild.py) * docs/Makefile.am (libvirt-api.xml): Insert missing "$(srcdir)/". --- docs/Makefile.am | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index cce7a48..26b1d29 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -131,7 +131,7 @@ devhelp/index.html devhelp/libvirt.devhelp: libvirt-api.xml $(devhelpxsl) libvirt-api.xml libvirt-refs.xml: apibuild.py \ $(srcdir)/../include/libvirt/*.h \ $(srcdir)/../src/libvirt.c $(srcdir)/../src/util/virterror.c - -(./apibuild.py) + -($(srcdir)/apibuild.py) clean-local: rm -f *~ *.bak *.hierarchy *.signals *-unused.txt -- 1.7.0.367.g566c3 >From 8e948091c4c3b2346be8ede0309ce768136621e9 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 14:15:26 +0100 Subject: [PATCH 2/2] build: avoid non-srcdir installation failure (sitemap.html.in) * docs/Makefile.am (EXTRA_DIST): Add sitemap.html.in. --- docs/Makefile.am | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index 26b1d29..4828752 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -76,6 +76,7 @@ EXTRA_DIST= \ $(devhelphtml) $(devhelppng) $(devhelpcss) $(devhelpxsl) \ $(xml) $(fig) $(png) \ $(patches) \ + sitemap.html.in \ ChangeLog.awk CLEANFILES = $(dot_html) $(apihtml) $(devhelphtml) -- 1.7.0.367.g566c3

2 2

[libvirt] [PATCH] build: make git submodule checking more reliable
by Jim Meyering 25 Feb '10

25 Feb '10

After investigating some autogen.sh/bootstrap-vs-new-gnulib problems today, I realized that libvirt's SHA1-comparing code lacked improvements I'd made in libguestfs, which does the same thing. AFAIK, this is not related to the problems we've seen today, but it might well help someone down the road. Here's the result of that synchronization: >From fada075bcaf46d1bd0c51165475594544e18ecfd Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 12:01:09 +0100 Subject: [PATCH] build: make git submodule checking more reliable * cfg.mk (_curr_status): Accommodate leading "+" or "-" in the output of git submodule status. * autogen.sh: Likewise. --- autogen.sh | 2 +- cfg.mk | 4 +++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/autogen.sh b/autogen.sh index 833b1d1..9ae8d67 100755 --- a/autogen.sh +++ b/autogen.sh @@ -65,7 +65,7 @@ fi # Ensure that whenever we pull in a gnulib update or otherwise change to a # different version (i.e., when switching branches), we also rerun ./bootstrap. curr_status=.git-module-status -t=$(git submodule status) +t=$(git submodule status|sed 's/^[ +-]//;s/ .*//') if test "$t" = "$(cat $curr_status 2>/dev/null)"; then : # good, it's up to date, all we need is autoreconf autoreconf -if diff --git a/cfg.mk b/cfg.mk index 17bb341..8317ae2 100644 --- a/cfg.mk +++ b/cfg.mk @@ -262,7 +262,9 @@ ifeq (0,$(MAKELEVEL)) # submodule name. Contrast these: # -b653eda3ac4864de205419d9f41eec267cb89eeb .gnulib # b653eda3ac4864de205419d9f41eec267cb89eeb .gnulib (v0.0-2286-gb653eda) - _submodule_hash = sed 's/.//;s/ .*//' + # $ cat .git-module-status + # b653eda3ac4864de205419d9f41eec267cb89eeb + _submodule_hash = sed 's/^[ +-]//;s/ .*//' _update_required := $(shell \ actual=$$(git submodule status | $(_submodule_hash)); \ stamp="$$($(_submodule_hash) $(_curr_status) 2>/dev/null)"; \ -- 1.7.0.377.g90adb

2 1

[libvirt] [PATCH] build: avoid non-srcdir "make distcheck" failures (XML_EXAMPLES)
by Jim Meyering 25 Feb '10

25 Feb '10

This avoids two of the 3 remaining failures, -error: loading host definition file '/t/libvirt-0.7.6/examples/xml/test/testnode.xml': No such file or directory -error: failed to connect to the hypervisor +error: failed to get domain '4294967298' >From b2318b64abc8ac23b4fa43dc52d3c2a32226b8cb Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 15:52:44 +0100 Subject: [PATCH] build: avoid non-srcdir "make distcheck" failures (XML_EXAMPLES) * Makefile.am (XML_EXAMPLES): Use $(wildcard in $(srcdir)-aware manner. --- Makefile.am | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/Makefile.am b/Makefile.am index a02cd26..09439c8 100644 --- a/Makefile.am +++ b/Makefile.am @@ -9,7 +9,9 @@ SUBDIRS = gnulib/lib include src daemon tools proxy docs gnulib/tests \ ACLOCAL_AMFLAGS = -I m4 -I gnulib/m4 -XML_EXAMPLES = $(wildcard examples/xml/test/*.xml) $(wildcard examples/xml/storage/*.xml) +XML_EXAMPLES = \ + $(patsubst $(srcdir)/%,%,$(wildcard $(addprefix $(srcdir)/examples/xml/, \ + test/*.xml storage/*.xml))) EXTRA_DIST = \ ChangeLog-old \ -- 1.7.0.367.g566c3

2 2

[libvirt] [PATCH 1/4] Avoid searching for windres when not building for cygwin or mingw.
by Diego Elio Pettenò 25 Feb '10

25 Feb '10

Just checking for a windres tool might hit even on Linux systems when building for Linux (e.g.: when using Gentoo and having built binutils with multitarget support), and will then fail to link properly at the end of the build. Check the host string before deciding whether to look for windres or not. --- configure.ac | 9 +++++++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff --git a/configure.ac b/configure.ac index 117cb20..5b1eb5f 100644 --- a/configure.ac +++ b/configure.ac @@ -1682,8 +1682,13 @@ AC_SUBST([CYGWIN_EXTRA_PYTHON_LIBADD]) AC_SUBST([MINGW_EXTRA_LDFLAGS]) dnl Look for windres to build a Windows icon resource. -AC_CHECK_TOOL([WINDRES], [windres], [no]) -AM_CONDITIONAL([WITH_WIN_ICON], [test "$WINDRES" != "no"]) +case "$host" in + *cygwin* | *mingw* ) + AC_CHECK_TOOL([WINDRES], [windres], []) + ;; +esac + +AM_CONDITIONAL([WITH_WIN_ICON], [test "$WINDRES" != ""]) -- 1.7.0

5 18

[libvirt] [PATCH] build: fix typos in configure-set makefile variable names
by Jim Meyering 25 Feb '10

25 Feb '10

Without this change, the ..._MODULES_... variables are set in configure.ac, but never used. The variables that are actually used in src/Makefile.am are named _MODULE_... >From a3361a6ce032f71273d6f7a6b4b4f54438b1dd73 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 12:04:27 +0100 Subject: [PATCH] build: fix typos in makefile variable names * configure.ac: Fix typos: s/DRIVER_MODULES_CFLAGS/DRIVER_MODULE_CFLAGS/ s/DRIVER_MODULES_LIBS/DRIVER_MODULE_LIBS/ --- configure.ac | 14 +++++++------- 1 files changed, 7 insertions(+), 7 deletions(-) diff --git a/configure.ac b/configure.ac index 2f79970..43ea771 100644 --- a/configure.ac +++ b/configure.ac @@ -1695,8 +1695,8 @@ AC_ARG_WITH([driver-modules], [], [with_driver_modules=no]) -DRIVER_MODULES_CFLAGS= -DRIVER_MODULES_LIBS= +DRIVER_MODULE_CFLAGS= +DRIVER_MODULE_LIBS= if test "x$with_driver_modules" = "xyes" ; then old_cflags="$CFLAGS" old_libs="$LIBS" @@ -1710,13 +1710,13 @@ if test "x$with_driver_modules" = "xyes" ; then LIBS="$old_libs" fi if test "$with_driver_modules" = "yes"; then - DRIVER_MODULES_CFLAGS="-export-dynamic" - DRIVER_MODULES_LIBS="-ldl" + DRIVER_MODULE_CFLAGS="-export-dynamic" + DRIVER_MODULE_LIBS="-ldl" AC_DEFINE_UNQUOTED([WITH_DRIVER_MODULES], 1, [whether to build drivers as modules]) fi AM_CONDITIONAL([WITH_DRIVER_MODULES], [test "$with_driver_modules" != "no"]) -AC_SUBST([DRIVER_MODULES_CFLAGS]) -AC_SUBST([DRIVER_MODULES_LIBS]) +AC_SUBST([DRIVER_MODULE_CFLAGS]) +AC_SUBST([DRIVER_MODULE_LIBS]) # Set LV_LIBTOOL_OBJDIR to "." or $lt_cv_objdir, depending on whether @@ -1931,7 +1931,7 @@ AC_MSG_NOTICE([]) AC_MSG_NOTICE([Driver Loadable Modules]) AC_MSG_NOTICE([]) if test "$with_driver_modules" != "no" ; then -AC_MSG_NOTICE([ dlopen: $DRIVER_MODULES_CFLAGS $DRIVER_MODULES_LIBS]) +AC_MSG_NOTICE([ dlopen: $DRIVER_MODULE_CFLAGS $DRIVER_MODULE_LIBS]) else AC_MSG_NOTICE([ dlopen: no]) fi -- 1.7.0.367.g566c3

3 6

[libvirt] [PATCH] build: avoid non-srcdir installation failure
by Jim Meyering 25 Feb '10

25 Feb '10

>From 28193acd8f1cab9b91cde77afef68781e5c740b2 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 14:01:58 +0100 Subject: [PATCH] build: avoid non-srcdir installation failure * docs/Makefile.am (html/index.html): Insert missing $(srcdir)/ on argument to $(XSLTPROC). This would make "make distcheck" fail when run from a non-srcdir build. Reported by Amit Shah --- docs/Makefile.am | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index 4df5154..cce7a48 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -115,7 +115,7 @@ ChangeLog.html.in: ChangeLog.xml ChangeLog.xsl html/index.html: libvirt-api.xml newapi.xsl page.xsl sitemap.html.in -@(if [ -x $(XSLTPROC) ] ; then \ echo "Rebuilding the HTML pages from the XML API" ; \ - $(XSLTPROC) --nonet $(srcdir)/newapi.xsl libvirt-api.xml ; fi ) + $(XSLTPROC) --nonet $(srcdir)/newapi.xsl $(srcdir)/libvirt-api.xml; fi) -@(if [ -x $(XMLLINT) -a -x $(XMLCATALOG) ] ; then \ if $(XMLCATALOG) /etc/xml/catalog "-//W3C//DTD XHTML 1.0 Strict//EN" > /dev/null ; then \ echo "Validating the resulting XHTML pages" ; \ -- 1.7.0.367.g566c3

2 1

[libvirt] libvirt/dnsmasq integration (formatting fixed).
by Simon Kelley 25 Feb '10

25 Feb '10

(apologies for first attempt: I was using a strange mailer.) As the principal maintainer of dnsmasq, I'm seeing increasing reports of problems on systems which run both dnsmasq and libvirt. I'm fairly sure I understand what's going on in these cases, and I have a few proposals for changes in libvir and dnsmasq that should fix things. The problem is that libvirt runs a private instance of dnsmasq: on machines which are also running a "system" dnsmasq daemon, this can cause problems. Some background: dnsmasq can run in two modes. Default mode: dnsmasq binds the wildcard address and does network magic to determine which interface request packets actually come from, so that the results can be sent back with the correct source address. This has the advantage that network interfaces can come and go and change IP address and dnsmasq will keep working. It's possible to restrict dnsmasq to only reply to requests on some interfaces; requests from other interfaces will be read by dnsmasq and then silently dropped. Telling dnsmasq to use an interface which doesn't exist but might in the future will result in a logged warning, but dnsmasq will still start and when the interface comes up it will work. Bind-interfaces mode: This is the traditional way to do UDP servers. At startup dnsmasq enumerates all the extant interfaces and then opens a socket for each one, listening on the interfaces's IP address. Interfaces may be skipped if excluded by the --interface and --except-interface flags, and any interface specified in --interface which doesn't exist at start-up will generate a fatal error. In almost all cases, default mode is better: --bind-interfaces is only there to cope with old platforms which don't support enough socket options to do default mode. The only time when --bind-interfaces works better is when it's desirable to run more than one instance of dnsmasq or have dnsmasq co-exist with another DNS server. This is not possible in default mode, but it does work in bind-interfaces mode, providing than _all_ instances of dnsmasq are in bind-interfaces mode, and that they listen on a disjoint set of interfaces. Therefore, to allow multiple dnsmasq instances libvirt's private dnsmasq instance is started in bind-interfaces mode: that forces one of the dnsmasq instances to do bind-interfaces. Many of the Linux distibution dnsmasq packages have now implemented an /etc/dnsmasq.d directory where configuration fragments can be dropped. Their libvirt packages are putting a file there which contains a bind-interfaces command, so that the "system" dnsmasq is automatically forced into the same mode, and the two can co-exist. This works, sort-of, but there some disadvantages. Installing libvirt drops the configuration change for the system dnsmasq, but the packages frequently don't restart the system daemon, so that things transiently fail until everything has rebooted. Much worse, the system dnsmasq is forced into bind-interfaces mode and then service to transient interfaces (usb, ad-hoc wifi) no longer works, or, because those interfaces are mentioned in the dnsmasq configuration, dnsmasq now fails at start-up when the interfaces don't exist. I don't think there is a solution to this problem wholly within dnsmasq: the nature of the BSD-sockets universe means that there never be two instances of dnsmasq without bind-interfaces, and it's, at best, difficult to get round the limitations of that WRT transient network interfaces. There is, an alternative solution, which would involve changes to both dnsmasq, libvirt and their packaging. Hence this mail. My proposal is to get rid of the necessity for two dnsmasq instances. Libvirt should check for the existance of a "system" dnsmasq and, if the system daemon exists, libvirt should drop the required configuration into /etc/dnsmasq.d and then restart it. If the system daemon is not installed or enabled, libvirt can start a private instance as now. The difficulty with this scheme is that libvirt needs to create some configuration which enables the services it needs on the virtual network without disturbing, or being disturbed by, whatever configuration exists for the system daemon. That's not currently possible, but it can be made possible. I'm assuming that libvirt needs to provide a set of IP address / MAC address mappings, and range of IP addresses on a virtual network. It needs DHCP and DNS service on the virtual network. The dhcp-host IP/MAC mappings are a non-problem: they will be ignored for any other subnet where the IP addresses don't fit, and any other dhcp-hosts in the system configuration will be similarly ignored for DHCP on the virtual network subnet. The dhcp-range is more of a problem. Service to particular networks in dnsmasq is controlled by interface=<interface name"> lines in the configuration. If there are none of these, service is provided to all interfaces. If they exist, service is limited to the interfaces specified. The existence of any dhcp-range line in dnsmasq's configuration enables the DHCP server for any subnet unless explicitly limited to particular interfaces. So a default dnsmasq installation, (with no interface=<interface>) which provides DNS everywhere but DHCP nowhere would be turned into one which provided DHCP on every interface by libvirt adding a dhcp-range. Since there wouldn't be a suitable DHCP range for most subnets, this would only result in logged errors, but it is still not good. Worse, there's no good answer to the question 'should libvirt include interface=virt0"' in the configuration it supplies? If it does, then the "enable DHCP on all interfaces" problem is solved, but a default system configuration with no interface declaration is transformed from one which provides DNS everywhere to one which provides DNS only to the virtual interface. If libvirt doesn't provide "interface=virt0" and the system configuration includes interface declarations, then there will be no DNS or DHCP service to the virtual network. To solve this, I propose to add an optional interface name to the dhcp-range declaration. The semantics of this would be rather odd, but solve the problem perfectly. 1) for DHCP, if any other dhcp-range exists _without_ an interface name, them the interface name is ignored and and things behave as before, otherwise DHCP is only provided to interfaces mentioned in dhcp-range declarations. 2) for DNS, if there are no interface declarations, things work as before. If there are interface declarations, the interfaces mentioned in dhcp-ranges are added to the set which get DNS service. With these rules, it should be possible for libvirt to drop eg dhcp-range=interface:virt0,192.168.0.1,192.168.0.240 into the configuration of the system dnsmasq and get DHCP and DNS service for virt0, irrespective of any other configuration in the system dnsmasq, and doing so shouldn't affect the services supplied elsewhere. The code in libvirt to make this work looks like this: echo dhcp-range=interface:virt0,<ip range> >>/etc/dnsmasq.d/libvirt if <system dnsmasq is not installed or not enabled> dnsmasq --interface=virt0\ --bind-interfaces --conf-file=/etc/dnsmasq.d/libvirt else /etc/init.d/dnsmasq restart (The --bind-interfaces in the private-dnsmasq instance keeps dnsmasq from clashing with other nameservers eg BIND which may be running.) The system dnsmasq package has to ensure that /etc/dnsmasq.d is read for configuration fragments, and the dnsmasq package and the libvirt package will have to co-operate to manage transitions between private and system dnsmasq mode caused by package installation or removal. Does that make sense? It's a long and involved explanation to come to cold. I fear I may have over-simplified what libvirt is doing with dnsmasq, in which case please enlighten me and I'll modify my scheme to take that into account. If this looks good I can easily have the necessary dnsmasq changes in the next release. Cheers, Simon.

1 0

[libvirt] [PATCH] build: avoid non-srcdir "make distcheck" failures (srcdir vs wildcard)
by Jim Meyering 25 Feb '10

25 Feb '10

With this, I'm down to 3 of 43 test failures. >From dd6f73de8afe2bcac80cc0a5b7f60bc83e0e30fc Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 15:18:51 +0100 Subject: [PATCH] build: avoid non-srcdir "make distcheck" failures (srcdir vs wildcard) * tests/xencapsdata/Makefile.am: Use $(wildcard in $(srcdir)-aware manner * tests/xmconfigdata/Makefile.am: Likewise. * tests/xml2sexprdata/Makefile.am: Likewise. --- tests/xencapsdata/Makefile.am | 5 +++-- tests/xmconfigdata/Makefile.am | 4 ++-- tests/xml2sexprdata/Makefile.am | 4 ++-- 3 files changed, 7 insertions(+), 6 deletions(-) diff --git a/tests/xencapsdata/Makefile.am b/tests/xencapsdata/Makefile.am index 7d77575..ed08bc4 100644 --- a/tests/xencapsdata/Makefile.am +++ b/tests/xencapsdata/Makefile.am @@ -1,2 +1,3 @@ - -EXTRA_DIST = $(wildcard *.xml) $(wildcard *.cpuinfo) $(wildcard *.caps) +EXTRA_DIST = \ + $(patsubst $(srcdir)/%,%,$(wildcard \ + $(srcdir)/*.xml $(srcdir)/*.cpuinfo $(srcdir)/*.caps)) diff --git a/tests/xmconfigdata/Makefile.am b/tests/xmconfigdata/Makefile.am index 05201a0..af0cb09 100644 --- a/tests/xmconfigdata/Makefile.am +++ b/tests/xmconfigdata/Makefile.am @@ -1,2 +1,2 @@ - -EXTRA_DIST = $(wildcard *.xml) $(wildcard *.cfg) +EXTRA_DIST = \ + $(patsubst $(srcdir)/%,%,$(wildcard $(srcdir)/*.xml $(srcdir)/*.cfg)) diff --git a/tests/xml2sexprdata/Makefile.am b/tests/xml2sexprdata/Makefile.am index 43117d5..20028bb 100644 --- a/tests/xml2sexprdata/Makefile.am +++ b/tests/xml2sexprdata/Makefile.am @@ -1,2 +1,2 @@ - -EXTRA_DIST = $(wildcard *.xml) $(wildcard *.sexpr) +EXTRA_DIST = \ + $(patsubst $(srcdir)/%,%,$(wildcard $(srcdir)/*.xml $(srcdir)/*.sexpr)) -- 1.7.0.367.g566c3

1 0

[libvirt] [PATCH] build: avoid non-srcdir "make distcheck" failure (test_conf.sh)
by Jim Meyering 25 Feb '10

25 Feb '10

"make distcheck" from a non-srcdir build dir was failing the "test_conf.sh" test TEST: test_conf.sh Failed to process /t/libvirt/.j/libvirt-0.7.6/tests/confdata/*.conf diff: /t/libvirt/.j/libvirt-0.7.6/tests/confdata/*.out: No such file or directory! 1 FAILED due to the fact that tests/confdata/Makefile.am was specifying its files via this: EXTRA_DIST = $(wildcard *.conf) $(wildcard *.out) That led to a distribution tarball that lacked those files, which are required by the test_conf.sh test. Automake normally warns us not to use $(wildcard ...), but we've suppressed the warning. This is one case in which using $(wildcard...) really should be avoided. In a non-srcdir build, each of those expands to the empty string, since the actual files are not in that directory, but in $(srcdir). My first reflex was to just hard-code the 4 file names: -EXTRA_DIST = $(wildcard *.conf) $(wildcard *.out) +EXTRA_DIST = \ + fc4.conf \ + fc4.out \ + libvirtd.conf \ + libvirtd.out But that'd make for more maintenance overhead if/when we add more files. Here it doesn't really matter, but in the next patch you'll see that it really does, in general. So here's a better way: Account for the fact that the files are available in $(srcdir): EXTRA_DIST = \ $(patsubst $(srcdir)/%,%,$(wildcard $(srcdir)/*.out $(srcdir)/*.conf)) >From ecce58c9432e83386f2a9ee275c873775577152b Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Wed, 24 Feb 2010 15:01:11 +0100 Subject: [PATCH] build: avoid non-srcdir "make distcheck" failure (test_conf.sh) * tests/confdata/Makefile.am (EXTRA_DIST): Apply $(wildcard... to $(srcdir)/..., and then remove the prefix. --- tests/confdata/Makefile.am | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/confdata/Makefile.am b/tests/confdata/Makefile.am index 5e97605..eaaadbc 100644 --- a/tests/confdata/Makefile.am +++ b/tests/confdata/Makefile.am @@ -1,2 +1,2 @@ - -EXTRA_DIST = $(wildcard *.conf) $(wildcard *.out) +EXTRA_DIST = \ + $(patsubst $(srcdir)/%,%,$(wildcard $(srcdir)/*.out $(srcdir)/*.conf)) -- 1.7.0.367.g566c3

1 0

[libvirt] [PATCH 0/4] maint: improve bootstrap
by Eric Blake 24 Feb '10

24 Feb '10

My biggest motivation in writing this patch series is the fact that I already have the complete gnulib.git on my hard drive, and repeatedly doing shallow clones of gnulib into multiple projects gets expensive in both network traffic and disk space. With this series, it is now possible to do: GNULIB_SRCDIR=/path/to/reference/gnulib.git ./autogen.sh so that the .gnulib subdirectory will exploit git's --reference capabilities rather than re-download all the files from scratch; it also results in less disk usage. The only caveat is that you should make sure that your reference gnulib.git is never rewound beyond contents referred to by a submodule, so that you never risk creating delta objects in the client repositories that would break if the reference module eventually prunes objects rendered unused by the rewind (in practice, this is only a problem if you are prone to pointing the .gnulib submodule to a non-published commit, in order to test out pending changes to gnulib.git, and can be avoided by doing your gnulib development in a different directory than in the reference gnulib.git that only tracks upstream). .gitignore | 20 +- autogen.sh | 21 +- bootstrap | 852 +++++++++++++++++++++++++++++++++++++++++++++----- bootstrap.conf | 145 +++++++++ build-aux/.gitignore | 12 +- po/.gitignore | 26 +- po/Makevars | 41 --- 7 files changed, 967 insertions(+), 150 deletions(-) Eric Blake (4): maint: fix quoting in autogen.sh maint: start factoring bootstrap maint: import modern bootstrap maint: sort .gitignore -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

4 19

[libvirt] RFC: Libvirt FAQ
by Cole Robinson 24 Feb '10

24 Feb '10

Hi all, If you spend enough time on #virt and various virt mailing lists, there are quite a few libvirt questions that come up regularly. So, largely in an effort to save myself and everyone from repeating ourselves, I've put together a libvirt FAQ, found here: http://wiki.libvirt.org/page/FAQ Some of the really common questions that I tried to answer (paraphrased): Q) Where does libvirt store the VM XML? A) You shouldn't really care, use virsh edit/dumpxml/define/undefine Q) How do I access the qemu monitor with libvirt? A) You don't, monitor commands should be provided by libvirt APIs, take a look at this complete list... Q) Why is libvirt messing with iptables? A) Providing NAT based network access for your VMs. If you don't want it, here's how to get rid of it. I've also tried to answer some questions that I don't feel have good straightforward answers in a highly visible place, like: - What is libvirt? - What is some of the major functionality provided by libvirt? - What does it mean when someone says 'I use libvirt'? - Difference between qemu:///system and qemu:///session Feel free to make any additions or mention any feedback. At the very least, please take a glance, so when someone inevitably asks one of these questions, you know to point them to the FAQ, rather than type out an answer. If people think this is useful, we can link to it from the main libvirt.org page, and merge the current static FAQ content linked from the sidebar into the wiki page. Thanks, Cole

2 1

[libvirt] [PATCH 00/10] Various documentation updates
by Cole Robinson 24 Feb '10

24 Feb '10

Over the months I've noticed a few XML pieces that aren't documented, so I finally sat down to hammer these out. Patch #1 is actually a bug I noticed in the process of trying to document the affected attribute. Thanks, Cole Cole Robinson (10): storage: conf: Correctly calculate exabyte unit docs: storage: Document capacity/alloc 'unit' docs: storage: <volume><key> is always generated. docs: storage: Fix backingStore <format> docs docs: storage: Document <auth> element docs: storage: Document SCSI pools docs: storage: Document multipath pools docs: domain: Document <description> element docs: network: Document STP and delay attributes docs: network: Document <domain> element docs/formatdomain.html.in | 7 ++++ docs/formatnetwork.html.in | 12 ++++++- docs/formatstorage.html.in | 34 ++++++++++++++++--- docs/storage.html.in | 76 ++++++++++++++++++++++++++++++++++++++++++++ src/conf/storage_conf.c | 4 +- 5 files changed, 124 insertions(+), 9 deletions(-)

4 18

[libvirt] PATCH [0/1] Filesystem pool formatting
by David Allan 24 Feb '10

24 Feb '10

Here's an updated patch based on the feedback from DV and danpb. I haven't added support for any additional filesystems beyond what mke2fs supports. The addition of additional filesystems is easily done and better as a separate patch. Dave

4 7

[libvirt] [PATCH] virsh.c: avoid leak on OOM error path
by Jim Meyering 23 Feb '10

23 Feb '10

No one really cares if we leak memory while dying, but who knows... freeing that buffer may let us go down more gracefully. FYI, the leak is triggered when virFileReadAll succeeds (it allocates "buffer"), yet xmlNewDoc fails: if (virFileReadAll(from, VIRSH_MAX_XML_FILE, &buffer) < 0) return FALSE; doc = xmlNewDoc(NULL); if (doc == NULL) goto no_memory; >From 03c7a44e3d5b2e7c992bebc98fc8c6a7bf63881e Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Fri, 19 Feb 2010 18:03:41 +0100 Subject: [PATCH] virsh.c: avoid leak on OOM error path * tools/virsh.c (cmdCPUBaseline): Also free "buffer" upon OOM. --- tools/virsh.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index dd916f3..8756a7a 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -7139,6 +7139,7 @@ cleanup: return ret; no_memory: + VIR_FREE(buffer); vshError(ctl, "%s", _("Out of memory")); ret = FALSE; return ret; -- 1.7.0.233.g05e1a

3 6

[libvirt] [PATCH] Create raw storage files with O_DSYNC (again)
by Jiri Denemark 23 Feb '10

23 Feb '10

Recently we introduced O_DYSNC flag when creating raw storage files to avoid filling all disk cache with dirty pages. However, the patch got lost when virStorageBackendCreateRaw was reworked using virFileOperation. Let's use O_DYSNC again. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/storage/storage_backend.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/storage/storage_backend.c b/src/storage/storage_backend.c index 8b9ed5d..3742493 100644 --- a/src/storage/storage_backend.c +++ b/src/storage/storage_backend.c @@ -356,7 +356,8 @@ virStorageBackendCreateRaw(virConnectPtr conn ATTRIBUTE_UNUSED, goto cleanup; } - if ((createstat = virFileOperation(vol->target.path, O_RDWR | O_CREAT | O_EXCL, + if ((createstat = virFileOperation(vol->target.path, + O_RDWR | O_CREAT | O_EXCL | O_DSYNC, vol->target.perms.mode, vol->target.perms.uid, vol->target.perms.gid, createRawFileOpHook, &hdata, -- 1.7.0

3 3

[libvirt] domain restore race condition
by Laine Stump 23 Feb '10

23 Feb '10

As noted in another message, the problem I was seeing is a race condition in qemudDomainRestore(), not with my modifications to qemudDmainSave(). Here's some discussion about that problem from IRC, with a question at the bottom: > <laine> Does anyone else see a failure of domain restore (immediately > after domain save? I'm very definitely seeing it on my machine with > F12+updates testing and libvirt built from unpatched sources. > <laine> It's very reproduceable - with virsh I do "save domain > filename", then "restore filename" and it pretty much always gives me > a black screen. Then I force shutdown the guest (with virt-manager) > and do "restore filename" again. Tada! It's restored and running! > [...] > <danpb> laine: possible race condition > <danpb> laine: try putting a sleep(10) before the qemuMonitorStartCPUs > in qemuDomainRestore() Dan's suggestion *did* eliminate the failures. > [...] > <danpb> laine: this sounds like the issue with libvirt prematurely > starting execution of the CPUs before QEMU has even started restoring > (or soemthing like that) > <danpb> laine: search the archives for a mail from Charles Duffy on > this subject some time ago > Here's the BZ filed by Charles Duffy https://bugzilla.redhat.com/show_bug.cgi?id=537938 It looks like he's dealing with a race condition earlier in the restore, since his solution was to wait for the migration process to terminate somewhere inside qemudStartVMDaemon(), rather than waiting until qemudStartVMDaemon() was finished (which is what it does now). Since this wait has already been done anyway by the time of Dan's sleep(10) in my test, I don't think Charles' patch would help this situation. So is there something that libvirt can wait on here to ensure proper start? Or is there a problem in qemu? (I'm still running 0.11. I'll also try upgrading to 0.12 and see if there are changes in behavior.)

1 0

[libvirt] PATCH [0/3] Filesystem pool formatting
by David Allan 23 Feb '10

23 Feb '10

I finished these patches this afternoon, and I need to do a bit more testing before I'm comfortable having these patches committed, but I'm sending them along so I can get everybody's feedback overnight. In particular, I'm wondering if my move of the pool unref call is correct. Dave

5 11

[libvirt] Failures in existing qemu domain save/restore.
by Laine Stump 22 Feb '10

22 Feb '10

After running a bunch more tests, I am now noticing failures in domain save/restore *without* my patch (ie sources built straight from git), not 100% in either case (with or without patches), and the same saved domain sometimes fails to restore and sometimes is successful. (Most of the time (but not all), when I try to restore a domain immediately after a save, it fails, but if I then force shutdown of that failed restore and restore again from the same file, it succeeds. (Note that this is all to/from a local disk - it's not related to NFS) Is there a known problem in domain save/restore? Is anyone else seeing any failures? On 02/19/2010 02:30 AM, Laine Stump wrote: > (This patch requires that you first apply the virFileOperation patches > that I sent to the list yesterday.) > > I'm hoping someone else's keen eye can see what I'm failing to see > here. > > The patch following this message makes domain save "work" on > root-squash NFS (both the libvirt header and the QEMU state are saved > to the file), but the files that it saves don't restore properly, even > when simply saving to a local directory w/o forking - the guest screen > comes up black, or reboots immediately, or some other such thing. I've > looked at the data files and the code, and can't see what difference > is causing this - at least the header written by libvirt seems to be > the same, and the part written by qemu starts in the same place and > looks similar. > > If anyone's got an idea, please let me know! I'd like to get this > figured out and a working patch submitted before the day is done. > > -- > libvir-list mailing list > libvir-list(a)redhat.com > https://www.redhat.com/mailman/listinfo/libvir-list > >

1 0

[libvirt] [PATCH] docs: add 3 missing spaces
by Dan Kenigsberg 21 Feb '10

21 Feb '10

From: danken <danken(a)redhat.com> --- docs/architecture.html.in | 2 +- docs/auth.html.in | 2 +- docs/logging.html.in | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/architecture.html.in b/docs/architecture.html.in index 8eb6458..f5cc520 100644 --- a/docs/architecture.html.in +++ b/docs/architecture.html.in @@ -83,7 +83,7 @@ drivers present in driver.h: <li>xen_internal: provide the implementation of the functions possible via direct hypervisor access</li> <li>proxy_internal: provide read-only Xen access via a proxy, the proxy code - is in the <code>proxy/</code>directory.</li> + is in the <code>proxy/</code> directory.</li> <li>xm_internal: provide support for Xen defined but not running domains.</li> <li>qemu_internal: implement the driver functions for QEmu and diff --git a/docs/auth.html.in b/docs/auth.html.in index 3ca40c3..ab6c3e9 100644 --- a/docs/auth.html.in +++ b/docs/auth.html.in @@ -72,7 +72,7 @@ available. The two libvirt daemon actions available are named <code>org.libvirt. for the RO socket, and <code>org.libvirt.unix.manage</code> for the RW socket. -As an example, to allow a user <code>fred</code>full access to the RW socket, +As an example, to allow a user <code>fred</code> full access to the RW socket, while requiring <code>joe</code> to authenticate with the admin password, would require adding the following snippet to <code>PolicyKit.conf</code>. diff --git a/docs/logging.html.in b/docs/logging.html.in index 63a0b05..f4a5878 100644 --- a/docs/logging.html.in +++ b/docs/logging.html.in @@ -96,7 +96,7 @@ <li><code>x:stderr</code> output goes to stderr</li> <li><code>x:syslog:name</code> use syslog for the output and use the given <code>name</code> as the ident</li> - <li><code>x:file:file_path</code>output to a file, with the given + <li><code>x:file:file_path</code> output to a file, with the given filepath</li> </ul> In all cases the x prefix is the minimal level, acting as a filter: -- 1.6.6

2 1

[libvirt] [RFC][PATCH 2/2] Automatically recreate macvtap interface and reconnect to qemu
by Ed Swierk 20 Feb '10

20 Feb '10

This patch implements a thread for each VM that iterates through the list of macvtap interfaces, attempting to recreate each one and reconnect it to qemu. See the PATCH 0 email for discussion. --- Index: libvirt-0.7.6/src/qemu/qemu_driver.c =================================================================== --- libvirt-0.7.6.orig/src/qemu/qemu_driver.c +++ libvirt-0.7.6/src/qemu/qemu_driver.c @@ -76,6 +76,7 @@ #include "xml.h" #include "cpu/cpu.h" #include "macvtap.h" +#include "threads.h" #define VIR_FROM_THIS VIR_FROM_QEMU @@ -96,6 +97,8 @@ struct _qemuDomainObjPrivate { int *vcpupids; qemuDomainPCIAddressSetPtr pciaddrs; + + pthread_t reconnectNetBackendsThread; }; static int qemudShutdown(void); @@ -143,6 +146,8 @@ static int qemudDomainDisconnectNetBacke virDomainObjPtr vm, virDomainNetDefPtr net); +static void *qemudDomainReconnectNetBackends(void *opaque); + static struct qemud_driver *qemu_driver = NULL; @@ -2794,6 +2799,10 @@ static int qemudStartVMDaemon(virConnect if (virDomainSaveStatus(conn, driver->caps, driver->stateDir, vm) < 0) goto abort; + if (pthread_create(&priv->reconnectNetBackendsThread, NULL, + qemudDomainReconnectNetBackends, vm->def->uuid)) + goto abort; + return 0; cleanup: @@ -5627,7 +5636,7 @@ static int qemudDomainConnectNetBackend( unsigned int qemuCmdFlags) { qemuDomainObjPrivatePtr priv = vm->privateData; - char *tapfd_name = NULL; + char *hostnet_name = NULL; int tapfd = -1; char *netstr = NULL; int ret = -1; @@ -5664,27 +5673,30 @@ static int qemudDomainConnectNetBackend( if (tapfd < 0) return -1; - if (virAsprintf(&tapfd_name, "fd-%s", net->info.alias) < 0) { + if (virAsprintf(&hostnet_name, "host%s", net->info.alias) < 0) { virReportOOMError(conn); close(tapfd); return -1; } if (!(netstr = qemuBuildHostNetStr(conn, net, ' ', - vlan, tapfd_name))) { - VIR_FREE(tapfd_name); + vlan, hostnet_name))) { + VIR_FREE(hostnet_name); close(tapfd); return -1; } qemuDomainObjEnterMonitorWithDriver(driver, vm); - if (qemuMonitorSendFileHandle(priv->mon, tapfd_name, tapfd) < 0) + if (qemuMonitorRemoveHostNetwork(priv->mon, vlan, hostnet_name) < 0) + VIR_INFO0("Did not remove existing host network"); + + if (qemuMonitorSendFileHandle(priv->mon, hostnet_name, tapfd) < 0) goto out; if (qemuMonitorAddHostNetwork(priv->mon, netstr) < 0) { - if (qemuMonitorCloseFileHandle(priv->mon, tapfd_name) < 0) - VIR_WARN(_("Failed to close tapfd with '%s'"), tapfd_name); + if (qemuMonitorCloseFileHandle(priv->mon, hostnet_name) < 0) + VIR_WARN(_("Failed to close tapfd with '%s'"), hostnet_name); goto out; } @@ -5693,7 +5705,7 @@ static int qemudDomainConnectNetBackend( out: qemuDomainObjExitMonitorWithDriver(driver, vm); VIR_FREE(netstr); - VIR_FREE(tapfd_name); + VIR_FREE(hostnet_name); close(tapfd); return ret; } @@ -8549,6 +8561,72 @@ out: return ret; } +static void *qemudDomainReconnectNetBackends(void *opaque) +{ + virConnectPtr conn = virConnectOpen("qemu:///system"); + struct qemud_driver *driver = conn->privateData; + const unsigned char *uuid = opaque; + + while (1) { + virDomainObjPtr vm; + unsigned int qemuCmdFlags; + int i; + + usleep(1000000); + + VIR_ERROR(_("qemuDomainReconnectNetBackends (%p %p)"), + conn, driver); + + qemuDriverLock(driver); + vm = virDomainFindByUUID(&driver->domains, uuid); + if (!vm) { + VIR_ERROR0("qemuDomainReconnectNetBackends done"); + qemuDriverUnlock(driver); + break; + } + + VIR_ERROR(_("qemuDomainReconnectNetBackends (%s)"), + vm->def->name); + + if (qemudExtractVersionInfo(vm->def->emulator, + NULL, + &qemuCmdFlags) < 0) + goto cleanup; + + if (qemuDomainObjBeginJobWithDriver(driver, vm) < 0) + goto cleanup; + + if (!virDomainObjIsActive(vm)) + goto endjob; + + for (i = 0 ; i < vm->def->nnets ; i++) { + virDomainNetDefPtr net = vm->def->nets[i]; + + VIR_ERROR(_("qemuDomainReconnectNetBackends (%s)"), net->ifname); + + if (net->type != VIR_DOMAIN_NET_TYPE_DIRECT) + continue; + + if (qemudDomainConnectNetBackend(conn, driver, vm, + net, qemuCmdFlags) < 0) + VIR_ERROR(_("qemudDomainConnectNetBackend failed (%s)"), + net->ifname); + } + + endjob: + if (vm && + qemuDomainObjEndJob(vm) == 0) + vm = NULL; + + cleanup: + if (vm) + virDomainObjUnlock(vm); + qemuDriverUnlock(driver); + } + + return NULL; +} + static int qemuCPUCompare(virConnectPtr conn, const char *xmlDesc,

1 0

[libvirt] [RFC][PATCH 1/2] Separate backend setup from NIC device add/remove
by Ed Swierk 20 Feb '10

20 Feb '10

This patch moves backend setup code from qemudDomainAttachNetDevice and qemudDomainDetachNetDevice into separate functions. While the intent is to allow the new functions to be called separately without affecting VM NIC devices, this is arguably a worthwhile refactoring on its own. Currently it's based on 0.7.6 with some macvtap patches. I can rebase it against git if there's interest. Signed-off-by: Ed Swierk <eswierk(a)aristanetworks.com> --- Move the code that creates and destroys the network backend and connects it to qemu from the code that creates the virtual NIC. Index: libvirt-0.7.6/src/qemu/qemu_driver.c =================================================================== --- libvirt-0.7.6.orig/src/qemu/qemu_driver.c +++ libvirt-0.7.6/src/qemu/qemu_driver.c @@ -132,6 +132,17 @@ static int qemuDetectVcpuPIDs(virConnect static int qemuUpdateActivePciHostdevs(struct qemud_driver *driver, virDomainDefPtr def); +static int qemudDomainConnectNetBackend(virConnectPtr conn, + struct qemud_driver *driver, + virDomainObjPtr vm, + virDomainNetDefPtr net, + unsigned int qemuCmdFlags); + +static int qemudDomainDisconnectNetBackend(virConnectPtr conn, + struct qemud_driver *driver, + virDomainObjPtr vm, + virDomainNetDefPtr net); + static struct qemud_driver *qemu_driver = NULL; @@ -5609,19 +5620,17 @@ error: } -static int qemudDomainAttachNetDevice(virConnectPtr conn, - struct qemud_driver *driver, - virDomainObjPtr vm, - virDomainNetDefPtr net, - unsigned int qemuCmdFlags) +static int qemudDomainConnectNetBackend(virConnectPtr conn, + struct qemud_driver *driver, + virDomainObjPtr vm, + virDomainNetDefPtr net, + unsigned int qemuCmdFlags) { qemuDomainObjPrivatePtr priv = vm->privateData; char *tapfd_name = NULL; int tapfd = -1; - char *nicstr = NULL; char *netstr = NULL; int ret = -1; - virDomainDevicePCIAddress guestAddr; int vlan; if (!(qemuCmdFlags & QEMUD_CMD_FLAG_HOST_NET_ADD)) { @@ -5630,142 +5639,132 @@ static int qemudDomainAttachNetDevice(vi return -1; } + if (priv->monConfig->type != VIR_DOMAIN_CHR_TYPE_UNIX) { + qemudReportError(conn, dom, NULL, VIR_ERR_NO_SUPPORT, + _("network device type '%s' cannot be attached: " + "qemu is not using a unix socket monitor"), + virDomainNetTypeToString(net->type)); + return -1; + } + + if ((vlan = qemuDomainNetVLAN(net)) < 0) { + qemudReportError(conn, NULL, NULL, VIR_ERR_NO_SUPPORT, "%s", + _("Unable to connect network backend without vlan")); + goto out; + } + if (net->type == VIR_DOMAIN_NET_TYPE_BRIDGE || net->type == VIR_DOMAIN_NET_TYPE_NETWORK) { - if (priv->monConfig->type != VIR_DOMAIN_CHR_TYPE_UNIX) { - qemudReportError(conn, dom, NULL, VIR_ERR_NO_SUPPORT, - _("network device type '%s' cannot be attached: " - "qemu is not using a unix socket monitor"), - virDomainNetTypeToString(net->type)); - return -1; - } - - if ((tapfd = qemudNetworkIfaceConnect(conn, driver, net, qemuCmdFlags)) < 0) - return -1; + tapfd = qemudNetworkIfaceConnect(conn, driver, net, qemuCmdFlags); } else if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) { - if (priv->monConfig->type != VIR_DOMAIN_CHR_TYPE_UNIX) { - qemudReportError(conn, dom, NULL, VIR_ERR_NO_SUPPORT, - _("direct device type '%s' cannot be attached: " - "qemu is not using a unix socket monitor"), - virDomainNetTypeToString(net->type)); - return -1; - } + tapfd = qemudPhysIfaceConnect(conn, net, + net->data.direct.linkdev, + net->data.direct.mode); + } + if (tapfd < 0) + return -1; - if ((tapfd = qemudPhysIfaceConnect(conn, net, - net->data.direct.linkdev, - net->data.direct.mode)) < 0) - return -1; + if (virAsprintf(&tapfd_name, "fd-%s", net->info.alias) < 0) { + virReportOOMError(conn); + close(tapfd); + return -1; } - if (VIR_REALLOC_N(vm->def->nets, vm->def->nnets+1) < 0) - goto no_memory; + if (!(netstr = qemuBuildHostNetStr(conn, net, ' ', + vlan, tapfd_name))) { + VIR_FREE(tapfd_name); + close(tapfd); + return -1; + } + + qemuDomainObjEnterMonitorWithDriver(driver, vm); + + if (qemuMonitorSendFileHandle(priv->mon, tapfd_name, tapfd) < 0) + goto out; + + if (qemuMonitorAddHostNetwork(priv->mon, netstr) < 0) { + if (qemuMonitorCloseFileHandle(priv->mon, tapfd_name) < 0) + VIR_WARN(_("Failed to close tapfd with '%s'"), tapfd_name); + goto out; + } + + ret = 0; + +out: + qemuDomainObjExitMonitorWithDriver(driver, vm); + VIR_FREE(netstr); + VIR_FREE(tapfd_name); + close(tapfd); + return ret; +} + +static int qemudDomainAttachNetDevice(virConnectPtr conn, + struct qemud_driver *driver, + virDomainObjPtr vm, + virDomainNetDefPtr net, + unsigned int qemuCmdFlags) +{ + qemuDomainObjPrivatePtr priv = vm->privateData; + char *nicstr = NULL; + int ret = -1; + virDomainDevicePCIAddress guestAddr; + int vlan; + + if (VIR_REALLOC_N(vm->def->nets, vm->def->nnets+1) < 0) { + virReportOOMError(conn); + return -1; + } if ((qemuCmdFlags & QEMUD_CMD_FLAG_NET_NAME) || (qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE)) { if (qemuAssignDeviceNetAlias(vm->def, net, -1) < 0) - goto cleanup; + goto out; } if ((qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) && qemuDomainPCIAddressEnsureAddr(priv->pciaddrs, &net->info) < 0) - goto cleanup; - - vlan = qemuDomainNetVLAN(net); + goto out; - if (vlan < 0) { + if ((vlan = qemuDomainNetVLAN(net)) < 0) { qemudReportError(conn, NULL, NULL, VIR_ERR_NO_SUPPORT, "%s", _("Unable to attach network devices without vlan")); - goto cleanup; - } - - if (tapfd != -1) { - if (virAsprintf(&tapfd_name, "fd-%s", net->info.alias) < 0) - goto no_memory; - - qemuDomainObjEnterMonitorWithDriver(driver, vm); - if (qemuMonitorSendFileHandle(priv->mon, tapfd_name, tapfd) < 0) { - qemuDomainObjExitMonitorWithDriver(driver, vm); - goto cleanup; - } - qemuDomainObjExitMonitorWithDriver(driver, vm); + goto out; } - if (!(netstr = qemuBuildHostNetStr(conn, net, ' ', - vlan, tapfd_name))) - goto try_tapfd_close; - - qemuDomainObjEnterMonitorWithDriver(driver, vm); - if (qemuMonitorAddHostNetwork(priv->mon, netstr) < 0) { - qemuDomainObjExitMonitorWithDriver(driver, vm); - goto try_tapfd_close; + if (qemudDomainConnectNetBackend(conn, driver, vm, net, qemuCmdFlags) < 0) { + VIR_WARN0(_("Failed to connect network backend")); + goto out; } - qemuDomainObjExitMonitorWithDriver(driver, vm); - - if (tapfd != -1) - close(tapfd); - tapfd = -1; if (!(nicstr = qemuBuildNicStr(conn, net, NULL, vlan))) - goto try_remove; + goto out_remove; qemuDomainObjEnterMonitorWithDriver(driver, vm); if (qemuMonitorAddPCINetwork(priv->mon, nicstr, &guestAddr) < 0) { qemuDomainObjExitMonitorWithDriver(driver, vm); - goto try_remove; + goto out_remove; } net->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI; memcpy(&net->info.addr.pci, &guestAddr, sizeof(guestAddr)); qemuDomainObjExitMonitorWithDriver(driver, vm); + vm->def->nets[vm->def->nnets++] = net; ret = 0; + goto out; - vm->def->nets[vm->def->nnets++] = net; +out_remove: + if (qemudDomainDisconnectNetBackend(conn, driver, vm, net) < 0) + VIR_WARN0("Unable to disconnect network backend"); -cleanup: +out: if ((ret != 0) && (qemuCmdFlags & QEMUD_CMD_FLAG_DEVICE) && (net->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) && qemuDomainPCIAddressReleaseAddr(priv->pciaddrs, &net->info) < 0) VIR_WARN0("Unable to release PCI address on NIC"); - VIR_FREE(nicstr); - VIR_FREE(netstr); - VIR_FREE(tapfd_name); - if (tapfd != -1) - close(tapfd); - return ret; - -try_remove: - if (vlan < 0) { - VIR_WARN0(_("Unable to remove network backend")); - } else { - char *hostnet_name; - if (virAsprintf(&hostnet_name, "host%s", net->info.alias) < 0) - goto no_memory; - qemuDomainObjEnterMonitorWithDriver(driver, vm); - if (qemuMonitorRemoveHostNetwork(priv->mon, vlan, hostnet_name) < 0) - VIR_WARN(_("Failed to remove network backend for vlan %d, net %s"), - vlan, hostnet_name); - qemuDomainObjExitMonitorWithDriver(driver, vm); - VIR_FREE(hostnet_name); - } - goto cleanup; - -try_tapfd_close: - if (tapfd_name) { - qemuDomainObjEnterMonitorWithDriver(driver, vm); - if (qemuMonitorCloseFileHandle(priv->mon, tapfd_name) < 0) - VIR_WARN(_("Failed to close tapfd with '%s'"), tapfd_name); - qemuDomainObjExitMonitorWithDriver(driver, vm); - } - - goto cleanup; - -no_memory: - virReportOOMError(conn); - goto cleanup; } @@ -6207,82 +6206,95 @@ cleanup: return ret; } -static int -qemudDomainDetachNetDevice(virConnectPtr conn, - struct qemud_driver *driver, - virDomainObjPtr vm, - virDomainDeviceDefPtr dev) +static int qemudDomainDisconnectNetBackend(virConnectPtr conn, + struct qemud_driver *driver, + virDomainObjPtr vm, + virDomainNetDefPtr net) { - int i, ret = -1; - virDomainNetDefPtr detach = NULL; qemuDomainObjPrivatePtr priv = vm->privateData; int vlan; char *hostnet_name = NULL; + if ((vlan = qemuDomainNetVLAN(net)) < 0) { + qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, + "%s", _("unable to determine original VLAN")); + return -1; + } + + if (virAsprintf(&hostnet_name, "host%s", net->info.alias) < 0) { + virReportOOMError(NULL); + return -1; + } + + qemuDomainObjEnterMonitorWithDriver(driver, vm); + if (qemuMonitorRemoveHostNetwork(priv->mon, vlan, hostnet_name) < 0) + VIR_WARN0("Failed to remove host network"); + qemuDomainObjExitMonitorWithDriver(driver, vm); + +#if WITH_MACVTAP + if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) { + if (net->ifname) + delMacvtap(conn, net->ifname); + } +#endif + + VIR_FREE(hostnet_name); + return 0; +} + +static int qemudDomainDetachNetDevice(virConnectPtr conn, + struct qemud_driver *driver, + virDomainObjPtr vm, + virDomainDeviceDefPtr dev) +{ + int i = -1; + virDomainNetDefPtr net = NULL; + qemuDomainObjPrivatePtr priv = vm->privateData; + for (i = 0 ; i < vm->def->nnets ; i++) { - virDomainNetDefPtr net = vm->def->nets[i]; + virDomainNetDefPtr n = vm->def->nets[i]; - if (!memcmp(net->mac, dev->data.net->mac, sizeof(net->mac))) { - detach = net; + if (!memcmp(n->mac, dev->data.net->mac, sizeof(n->mac))) { + net = n; break; } } - if (!detach) { + if (!net) { qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, _("network device %02x:%02x:%02x:%02x:%02x:%02x not found"), dev->data.net->mac[0], dev->data.net->mac[1], dev->data.net->mac[2], dev->data.net->mac[3], dev->data.net->mac[4], dev->data.net->mac[5]); - goto cleanup; + return -1; } - if (!virDomainDeviceAddressIsValid(&detach->info, + if (!virDomainDeviceAddressIsValid(&net->info, VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI)) { qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, "%s", _("device cannot be detached without a PCI address")); - goto cleanup; - } - - if ((vlan = qemuDomainNetVLAN(detach)) < 0) { - qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED, - "%s", _("unable to determine original VLAN")); - goto cleanup; - } - - if (virAsprintf(&hostnet_name, "host%s", detach->info.alias) < 0) { - virReportOOMError(NULL); - goto cleanup; + return -1; } qemuDomainObjEnterMonitorWithDriver(driver, vm); if (qemuMonitorRemovePCIDevice(priv->mon, - &detach->info.addr.pci) < 0) { - qemuDomainObjExitMonitorWithDriver(driver, vm); - goto cleanup; - } - - if (qemuMonitorRemoveHostNetwork(priv->mon, vlan, hostnet_name) < 0) { + &net->info.addr.pci) < 0) { qemuDomainObjExitMonitorWithDriver(driver, vm); - goto cleanup; + return -1; } qemuDomainObjExitMonitorWithDriver(driver, vm); -#if WITH_MACVTAP - if (detach->type == VIR_DOMAIN_NET_TYPE_DIRECT) { - if (detach->ifname) - delMacvtap(conn, detach->ifname); - } -#endif + if (qemudDomainDisconnectNetBackend(conn, driver, vm, net) < 0) + VIR_WARN0("Failed to disconnect network backend"); - if ((driver->macFilter) && (detach->ifname != NULL)) { + if ((driver->macFilter) && (net->ifname != NULL)) { if ((errno = networkDisallowMacOnPort(conn, driver, - detach->ifname, - detach->mac))) { + net->ifname, + net->mac))) { virReportSystemError(conn, errno, _("failed to remove ebtables rule on '%s'"), - detach->ifname); + net->ifname); } } @@ -6299,13 +6311,9 @@ qemudDomainDetachNetDevice(virConnectPtr VIR_FREE(vm->def->nets); vm->def->nnets = 0; } - virDomainNetDefFree(detach); + virDomainNetDefFree(net); - ret = 0; - -cleanup: - VIR_FREE(hostnet_name); - return ret; + return 0; } static int qemudDomainDetachHostPciDevice(virConnectPtr conn,

1 0

[libvirt] [RFC][PATCH 0/2] Dynamic backend setup for macvtap interfaces
by Ed Swierk 20 Feb '10

20 Feb '10

Using a bridge to connect a qemu NIC to a host interface offers a fair amount of flexibility to reconfigure the host without restarting the VM. For example, if the bridge connects host interface eth0 to the qemu tap0 interface, eth0 can be hot-removed and hot-plugged without affecting the VM. Similarly, if the bridge connects host VLAN interface vlan0 to the qemu tap0 interface, the admin can easily replace vlan0 with vlan1 without the VM noticing. Using the macvtap driver instead of a kernel bridge, the host interface is much more tightly tied to the VM. Qemu communicates with the macvtap interface through a file descriptor, and the macvtap interface is bound permanently to a specific host interface when it is first created. What's more, if the underlying host interface disappears, the macvtap interface vanishes along with it, leaving the VM holding a file descriptor for a deleted file. To avoid race conditions during system startup, I would like libvirt to allow starting up the VM with a NIC even if the underlying host interface doesn't yet exist, deferring creation of the macvtap interface (analogous to starting up the VM with a tap interface bound to an orphan bridge). To support adding and removing a host interface without restarting the VM, I would like libvirt to react to the (re)appearance of the underlying host interface, creating a new macvtap interface and passing the new fd to qemu to reconnect to the NIC. (It would also be nice if libvirt allowed the user to change which underlying host interface the qemu NIC is connected to. I'm ignoring this issue for now, except to note that implementing the above features should make this easier.) The libvirt API already supports domainAttachDevice and domainDetachDevice to add or remove an interface while the VM is running. In the qemu implementation, these commands add or remove the VM NIC device as well as reconfiguring the host side. This works only if the OS and application running in the VM can handle PCI hotplug and dynamically reconfigure its network. I would like to isolate the VM from changes to the host network setup, whether you use macvtap or a bridge. The changes I think are needed to implement this include: 1. Refactor qemudDomainAttachNetDevice/qemudDomainDetachNetDevice, which currently handle both backend (host) setup and adding/removing the VM NIC device; move the backend setup code into separate functions that can called separately without affecting VM devices. 2. Implement a thread or task that watches for changes to the underlying host interface for each configured macvtap interface, and reacts by invoking the appropriate backend setup code. 3. Change qemudBuildCommandLine to defer backend setup if qemu supports the necessary features for doing it later (e.g. the host_net_add monitor command). 4. Implement appropriate error handling and reporting, and any necessary changes to the configuration schema. The following patches are a partial implementation of the above as a proof of concept. Patch 1 implements change (1) above, moving the backend setup code to new functions qemudDomainConnectNetBackend/qemudDomainDisconnectNetBackend, and calling these functions from the existing qemudDomainAttachNetDevice/qemudDomainDetachNetDevice. I think this change is useful on its own: it breaks up two monster functions into more manageable pieces, and eliminates some code duplication (e.g. the try_remove clause at the end of qemudDomainAttachNetDevice). Patch 2 is a godawful hack roughly implementing change (2) above (did I mention that this is a proof of concept?). It spawns a thread that simply tries reconnecting the backend of each macvtap interface once a second. As long as the interface is already up, the reconnection fails. If the macvtap interface goes away because the underlying host interface disappears, the reconnection fails until the host interface reappears. I ran into two major issues while implementing change (2): - Can we use the existing virEvent functions to invoke the reconnection process, triggered either by a timer or by an event from the host? It seems like this ought to work, but it appears that communication between libvirt and the qemu monitor relies on an event, and since all events run in the same thread, there's no way for an event to call the monitor. - Should the reconnection process use udev or hal to get notifications, or leverage the node device code which itself uses udev or hal? Currently there doesn't appear to be a way to get notifications of changes to node devices; if there were, we'd still need to address the threading issue. If we use node devices, what changes to the configuration schema would be needed to associate a macvtap interface with the underlying node device? I didn't attempt to implement changes (3) or (4) above but would welcome input on those items as well. --Ed

1 0

[libvirt] [PATCH 0/1] A couple of problems with the daemon-conf test
by David Allan 20 Feb '10

20 Feb '10

Here's a patch fixing two problems I found with the daemon-conf test that prevented the test from passing if the system libvirt was running on my system. The first change only affects the direct running of the daemon-conf script, i.e, if someone executes ./daemon-conf, which will always exit failure because the default config file contains a line that the test believes should not be there. The invocation by make-check supplies a different config file that's not subject to that problem. The second problem is that the corrupt config tests don't supply the pid-file argument to libvirt, so on my system it was attempting to use the same pidfile as the running system daemon and failing all the tests except the valid config test. Supplying the --pid-file argument when running the corrupt config tests as it is supplied for the valid config tests fixes the problem. Dave

4 4