- Devel - libvirt List Archives

[libvirt] Planning for next release
by Daniel Veillard 28 Mar '19

28 Mar '19

It's time to get a new release out. Simplest would be to enter freeze tomorrow, then we can plan an RC2 by Friday, and if everything looks good have the release on time next Monday. Hope this works for everybody, thanks, Daniel -- Daniel Veillard | Red Hat Developers Tools http://developer.redhat.com/ veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

3 3

[libvirt] [PATCH 00/16] snapshot refactoring (incremental backup saga)
by Eric Blake 28 Mar '19

28 Mar '19

A couple of these have seen the list before, but most of them are new. The bulk of this series is about refactoring snapshot_conf.c into smaller pieces that I can then reuse for implementing checkpoints, without having to open-code the hierarchy algorithms a second time (one of John's complaints against my v4 incremental backup series). I suspect that src/vz/ builds may break on one or more of these patches; I could not get a working vz build environment. Eric Blake (16): test: Avoid use-after-free on virDomainSnapshotDelete snapshot: Use accessors for virDomainSnapshot members snapshot: Create virDomainMoment base class vbox: Clean up some snapshot usage snapshot: Drop virDomainSnapshotDef.current snapshot: Track current snapshot in virDomainSnapshotObjList snapshot: Add accessors for updating snapshot list relations snapshot: Access snapshot def directly when needed snapshot: Refactor list filtering snapshot: Factor out virDomainMomentDef class snapshot: Switch type of virDomainSnapshotObj.def snapshot: Rename virDomainSnapshotObjPtr snapshot: Rename file for virDomainMomentObj snapshot: Move snapshot list code into generic file snapshot: Tweaks to support new bulk dumpxml/import API backup: Introduce virDomainCheckpointPtr include/libvirt/virterror.h | 6 +- src/util/virerror.c | 12 +- include/libvirt/libvirt.h | 6 +- src/conf/domain_conf.h | 1 - src/conf/moment_conf.h | 41 +++ src/conf/snapshot_conf.h | 23 +- src/conf/virconftypes.h | 12 +- src/conf/virdomainmomentobjlist.h | 95 +++++ src/conf/virdomainsnapshotobj.h | 50 --- src/conf/virdomainsnapshotobjlist.h | 33 +- src/datatypes.h | 83 ++++- src/qemu/qemu_command.h | 2 +- src/qemu/qemu_domain.h | 6 +- src/qemu/qemu_process.h | 4 +- src/conf/Makefile.inc.am | 6 +- src/conf/moment_conf.c | 40 +++ src/conf/snapshot_conf.c | 179 +++++----- src/conf/virdomainmomentobjlist.c | 521 ++++++++++++++++++++++++++++ src/conf/virdomainsnapshotobj.c | 123 ------- src/conf/virdomainsnapshotobjlist.c | 367 +++++++------------- src/datatypes.c | 132 ++++--- src/esx/esx_driver.c | 82 ++--- src/libvirt-domain-snapshot.c | 26 +- src/libvirt_private.syms | 21 +- src/qemu/qemu_command.c | 2 +- src/qemu/qemu_domain.c | 36 +- src/qemu/qemu_driver.c | 290 ++++++++-------- src/qemu/qemu_process.c | 4 +- src/remote/remote_daemon_dispatch.c | 4 +- src/remote/remote_driver.c | 4 +- src/rpc/gendispatch.pl | 2 +- src/test/test_driver.c | 174 ++++------ src/vbox/vbox_common.c | 207 +++++------ src/vz/vz_driver.c | 89 ++--- src/vz/vz_sdk.c | 4 +- tests/domainsnapshotxml2xmltest.c | 5 +- 36 files changed, 1613 insertions(+), 1079 deletions(-) create mode 100644 src/conf/moment_conf.h create mode 100644 src/conf/virdomainmomentobjlist.h delete mode 100644 src/conf/virdomainsnapshotobj.h create mode 100644 src/conf/moment_conf.c create mode 100644 src/conf/virdomainmomentobjlist.c delete mode 100644 src/conf/virdomainsnapshotobj.c -- 2.20.1

7 70

[libvirt] [PATCH] tests: Avoid writing into $HOME during virsh-snapshot
by Eric Blake 28 Mar '19

28 Mar '19

In a constrained CI environment, where it is intentional that attempts to write outside the current directory will fail, virsh-snapshot was failing: @@ -1,2 +1,3 @@ error: invalid argument: parent s3 for snapshot s2 not found error: marker +error: Failed to create '/home/travis/.cache/libvirt/virsh': Permission denied FAIL virsh-snapshot (exit status: 1) But we've already solved the problem in virsh-uriprecedence: tell virsh to use XDG locations pointing to somewhere we can write rather than its default of falling back to $HOME with the test being at risk of breaking due to the user's environment and/or unacceptably altering the user's normal cache. Hoist that solution into test-lib.sh, so that all scripts can use it as needed. Fixes: 280a2b41e Reported-by: Daniel P. Berrangé <berrange(a)redhat.com> Signed-off-by: Eric Blake <eblake(a)redhat.com> --- tests/test-lib.sh | 13 +++++++++++++ tests/virsh-snapshot | 2 ++ tests/virsh-uriprecedence | 12 +----------- 3 files changed, 16 insertions(+), 11 deletions(-) diff --git a/tests/test-lib.sh b/tests/test-lib.sh index 49e8d22095..64f0b0d401 100644 --- a/tests/test-lib.sh +++ b/tests/test-lib.sh @@ -222,6 +222,19 @@ mkfifo_or_skip_() fi } +# Create mock XDG files/directories to avoid permission problems. +# As it points inside $test_dir_, it is automatically cleaned. +mock_xdg_() +{ + export XDG_CONFIG_HOME="$test_dir_/.config" + export XDG_CACHE_HOME="$test_dir_/.cache" + export XDG_RUNTIME_HOME="XDG_CACHE_HOME" + + mkdir -p "$XDG_CONFIG_HOME/libvirt" "$XDG_CONFIG_HOME/virsh" + mkdir -p "$XDG_CACHE_HOME/libvirt" "$XDG_CACHE_HOME/virsh" + mkdir -p "$XDG_RUNTIME_HOME/libvirt" "$XDG_RUNTIME_HOME/virsh" +} + test_dir_=$(pwd) this_test_() { echo "./$0" | sed 's,.*/,,'; } diff --git a/tests/virsh-snapshot b/tests/virsh-snapshot index fb8a99dd43..cb498cf54e 100755 --- a/tests/virsh-snapshot +++ b/tests/virsh-snapshot @@ -26,6 +26,8 @@ fi fail=0 +mock_xdg_ || framework_failure + # The test driver loses states between restarts, so we perform a script # with some convenient markers for later post-processing of output. $abs_top_builddir/tools/virsh --connect test:///default >out 2>err ' diff --git a/tests/virsh-uriprecedence b/tests/virsh-uriprecedence index 564e3dc42c..fd6ce108c0 100755 --- a/tests/virsh-uriprecedence +++ b/tests/virsh-uriprecedence @@ -11,17 +11,7 @@ virsh_cmd="$virsh_bin" counter=0 ret=0 -cleanup_() { rm -rf "$tmphome"; } - -# Create all mock files/directories to avoid permission problems -tmphome="$PWD/tmp_home" -export XDG_CONFIG_HOME="$tmphome/.config" -export XDG_CACHE_HOME="$tmphome/.cache" -export XDG_RUNTIME_HOME="XDG_CACHE_HOME" - -mkdir -p "$XDG_CONFIG_HOME/libvirt" "$XDG_CONFIG_HOME/virsh" -mkdir -p "$XDG_CACHE_HOME/libvirt" "$XDG_CACHE_HOME/virsh" -mkdir -p "$XDG_RUNTIME_HOME/libvirt" "$XDG_RUNTIME_HOME/virsh" +mock_xdg_ || framework_failure is_uri_good() { -- 2.20.1

3 4

[libvirt] [PATCH 0/3] qemu: clean up USB address releasing
by Peter Krempa 28 Mar '19

28 Mar '19

Peter Krempa (3): qemu: hotplug: Don't release USB address twice when removing disk qemu: Always use 'alias' in warning message when removing USB address qemu: address: Stop reporting warning when USB address can't be released src/qemu/qemu_domain_address.c | 9 ++---- src/qemu/qemu_domain_address.h | 3 +- src/qemu/qemu_hotplug.c | 53 +++++++++++++++++----------------- 3 files changed, 29 insertions(+), 36 deletions(-) -- 2.20.1

2 4

Re: [libvirt] Can jobs suck like qemu-pr-helper does be transfered to libvirtd?
by Zhangbo (Oscar) 28 Mar '19

28 Mar '19

>> Hi all: >> qemu-pr-helper exits to help qemu do the high-privileged scsi related jobs. >LIBVIRTD is responsible to launch qemu-pr-helper and qemu, and set >selinux/DAC labels for them and their socket. >> >> # >> # >> # >> # >> # ___________ >> # ___________|libvirtd |__________ >> # | |___________| | >> # | | | >> # | | | >> # _________|________ _____|______ ___|____ >> # | qemu-pr-helper |__| vm1.sock |_____| qemu1 | >> # |__________________| |____________| |________| >> >> There may be other jobs quite like qemu-pr-helper, shall we make them as >"complex" as the qemu-pr-helper scheme? >> Will it be OK to just let qemu send an EVENT to libvirtd, and let libvirtd do the >SCSI/other jobs? Will it be OK if we remove qemu-pr-helper and similar >processes? >> What's the disadvantage if we let libvirtd do there high privileged jobs instead >qemu-pr-helper-like processes? > >No what you describe is not desired. Any process which is vital for the >runtime of the VM needs to stay as simple and minimal as possible and >separate from libvirtd. > >Libvirtd is a complex big application dealing with multiple things at >once and also the architecture allows for libvirtd to be stopped and >upgraded. > >This does not play well with the fact that processes as the PR helper >are always required. > >Merging them into libvirtd would make the VM stop until libvirtd is >running again. Additionally if any of the operations require persistent >kernel state as e.g. file descriptors, this would be impossible as >stopping libvirtd process would close the FDs which may be then >impossible to reopen properly e.g. due to state. Thanks! Besides these reasons above, will it weaken security if we let libvirtd do these jobs? For example, Such sayings, like, "libvirtd would become the focus suffered from attacking forces", make sense?

1 0

[libvirt] [PATCH] qemu: don't fail on destroy if domain is inactive
by Nikolay Shirokovskiy 28 Mar '19

28 Mar '19

Mgmt can not track if domain is already inactive before calling destroy because domain can become inactive because of crash/shutdown from guest. Thus it is make sense to report success in this case. Another option is to return special error code but this is a bit more complicated. Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy(a)virtuozzo.com> --- src/qemu/qemu_driver.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 62d8d97..0789efc 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -2172,8 +2172,10 @@ qemuDomainDestroyFlags(virDomainPtr dom, if (virDomainDestroyFlagsEnsureACL(dom->conn, vm->def) < 0) goto cleanup; - if (virDomainObjCheckActive(vm) < 0) + if (!virDomainObjIsActive(vm)) { + ret = 0; goto cleanup; + } state = virDomainObjGetState(vm, &reason); starting = (state == VIR_DOMAIN_PAUSED && -- 1.8.3.1

3 7

[libvirt] [PATCH] util: suppress unimportant ovs-vsctl errors when getting interface stats
by Laine Stump 28 Mar '19

28 Mar '19

commit edaf13565 modified the stats retrieval for OVS interfaces to not fail when one of the fields was unrecognized by the ovs-vsctl command, but ovs-vsctl was still returning an error, and libvirt was cluttering the logs with these inconsequential error messages. This patch modifies the GET_STAT macro to add "--if-exists" to the ovs-vsctl command, which causes it to return an empty string (and exit with success) if the requested statistic isn't in its database, thus eliminating the ugly error messages from the log. Resolves: https://bugzilla.redhat.com/1683175 Signed-off-by: Laine Stump <laine(a)laine.org> --- src/util/virnetdevopenvswitch.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/util/virnetdevopenvswitch.c b/src/util/virnetdevopenvswitch.c index 4fa3a5742a..e5a45f59ec 100644 --- a/src/util/virnetdevopenvswitch.c +++ b/src/util/virnetdevopenvswitch.c @@ -335,10 +335,10 @@ virNetDevOpenvswitchInterfaceStats(const char *ifname, virCommandFree(cmd); \ cmd = virCommandNew(OVSVSCTL); \ virNetDevOpenvswitchAddTimeout(cmd); \ - virCommandAddArgList(cmd, "get", "Interface", ifname, \ - "statistics:" name, NULL); \ + virCommandAddArgList(cmd, "--if-exists", "get", "Interface", \ + ifname, "statistics:" name, NULL); \ virCommandSetOutputBuffer(cmd, &output); \ - if (virCommandRun(cmd, NULL) < 0) { \ + if (virCommandRun(cmd, NULL) < 0 || !*output || *output == '\n') { \ stats->member = -1; \ } else { \ if (virStrToLong_ll(output, &tmp, 10, &stats->member) < 0 || \ -- 2.20.1

2 1

[libvirt] [PATCH] tools: Tweak workding for iothreadset
by John Ferlan 28 Mar '19

28 Mar '19

Update the wording to note the values for polling are purely dynamic and won't be saved across domain stop/(re)start or save/restore. Signed-off-by: John Ferlan <jferlan(a)redhat.com> --- tools/virsh.pod | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/tools/virsh.pod b/tools/virsh.pod index a0fe949c55..f00564fa5f 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -1772,7 +1772,10 @@ reach the maximum polling time. If a 0 (zero) is provided, then the default factor will be used. The I<--poll-shrink> is the quotient by which the current polling time will be reduced in order to get below the maximum polling interval. If a 0 (zero) is provided, then -the default quotient will be used. +the default quotient will be used. The polling values are purely dynamic +for a running guest. Saving, destroying, stopping, etc. the guest will +result in the polling values returning to hypervisor defaults at the +next start, restore, etc. If I<--live> is specified, affect a running guest. If the guest is not running an error is returned. -- 2.20.1

2 2

[libvirt] [PATCH v2 0/2] qemu_capabilities; Drop virQEMUCapsSetVAList
by Michal Privoznik 27 Mar '19

27 Mar '19

Technically, this is v2 of: https://www.redhat.com/archives/libvir-list/2019-March/msg01778.html but is this one implements different approach. Plus it breaks dependency between two unrelated enums that I've noticed when debugging broken build on a 32bit arch. Michal Prívozník (2): qemu_capabilities; Drop virQEMUCapsSetVAList qemuxml2argvtest: Drop dependency between testInfoArgName and virQEMUCapsFlags enums src/qemu/qemu_capabilities.c | 15 +++---------- src/qemu/qemu_capabilities.h | 2 -- tests/qemuxml2argvtest.c | 41 +++++++++++++++++++++++------------- 3 files changed, 29 insertions(+), 29 deletions(-) -- 2.19.2

2 4

[libvirt] [PATCH] qemu_capabilities: Pass pointer to va_list in virQEMUCapsSetVAList
by Michal Privoznik 27 Mar '19

27 Mar '19

There is one specific caller (testInfoSetArgs() in qemuxml2argvtest.c) which expect the va_list argument to change after returning from the virQEMUCapsSetVAList() function. However, since we are passing plain va_list this is not guaranteed. The man page of stdarg(3) says: If ap is passed to a function that uses va_arg(ap,type), then the value of ap is undefined after the return of that function. (ap is a variable of type va_list) I've seen this in action in fact: on i686 the qemuxml2argvtest fails on the second test case because testInfoSetArgs() sees ARG_QEMU_CAPS and callse virQEMUCapsSetVAList to process the capabilities (in this case there's just one QEMU_CAPS_SECCOMP_BLACKLIST). But since the changes are not reflected in the caller, in the next iteration testInfoSetArgs() sees the qemu capability and not ARG_END. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- This passed successfully on x86_64 and i686 in my testing. src/qemu/qemu_capabilities.c | 6 +++--- src/qemu/qemu_capabilities.h | 2 +- tests/qemuxml2argvtest.c | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index c5954edaf0..e182a2c2e5 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -1665,11 +1665,11 @@ virQEMUCapsSet(virQEMUCapsPtr qemuCaps, void virQEMUCapsSetVAList(virQEMUCapsPtr qemuCaps, - va_list list) + va_list *list) { int flag; - while ((flag = va_arg(list, int)) < QEMU_CAPS_LAST) + while ((flag = va_arg(*list, int)) < QEMU_CAPS_LAST) ignore_value(virBitmapSetBit(qemuCaps->flags, flag)); } @@ -1680,7 +1680,7 @@ virQEMUCapsSetList(virQEMUCapsPtr qemuCaps, ...) va_list list; va_start(list, qemuCaps); - virQEMUCapsSetVAList(qemuCaps, list); + virQEMUCapsSetVAList(qemuCaps, &list); va_end(list); } diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h index 7625d754a3..ceab8c9154 100644 --- a/src/qemu/qemu_capabilities.h +++ b/src/qemu/qemu_capabilities.h @@ -519,7 +519,7 @@ void virQEMUCapsSet(virQEMUCapsPtr qemuCaps, virQEMUCapsFlags flag) ATTRIBUTE_NONNULL(1); void virQEMUCapsSetVAList(virQEMUCapsPtr qemuCaps, - va_list list) ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2); + va_list *list) ATTRIBUTE_NONNULL(1) ATTRIBUTE_NONNULL(2); void virQEMUCapsSetList(virQEMUCapsPtr qemuCaps, ...) ATTRIBUTE_NONNULL(1); void virQEMUCapsClear(virQEMUCapsPtr qemuCaps, diff --git a/tests/qemuxml2argvtest.c b/tests/qemuxml2argvtest.c index 6e7d1b0b9a..a7a8d77e1d 100644 --- a/tests/qemuxml2argvtest.c +++ b/tests/qemuxml2argvtest.c @@ -650,7 +650,7 @@ testInfoSetArgs(struct testInfo *info, case ARG_QEMU_CAPS: if (qemuCaps || !(qemuCaps = virQEMUCapsNew())) goto cleanup; - virQEMUCapsSetVAList(qemuCaps, argptr); + virQEMUCapsSetVAList(qemuCaps, &argptr); break; case ARG_GIC: -- 2.19.2

4 6