April 2024 - Devel - libvirt List Archives

[PATCH v4 0/3] test: fix nodedev mdev XML regression
by Cole Robinson 23 Apr '24

23 Apr '24

See last patch for explanation. First two patches are related bugfixes/improvements v4 changes: - 3 patches pushed - mark parsed devices as persistent too - add GetXML fix patch Cole Robinson (3): test: make parsed nodedevs active and persistent test: Sync GetXML INACTIVE behavior with live driver conf: nodedev: Fill active_config at XML parse time src/conf/node_device_conf.c | 5 ++++- src/test/test_driver.c | 19 ++++++++++++++++++- tests/nodedevxml2xmltest.c | 15 --------------- 3 files changed, 22 insertions(+), 17 deletions(-) -- 2.44.0

2 5

[PATCH] virnetdevopenvswitch: Create OVS ports as transient
by Michal Privoznik 22 Apr '24

22 Apr '24

Since OVS keeps desired state in a DB, upon sudden crash of the host we may leave a port behind. There's no problem on VM shutdown or NIC hotunplug as we call corresponding del-port function (virNetDevOpenvswitchRemovePort()). But if the host suddenly crashes we won't ever do that. What happens next, is when OVS starts it finds desired state in its DB and creates a stale port. OVS added support for transient ports in v2.5.0 (Feb 2016) and since its v2.9.0 it even installs a systemd service (ovs-delete-transient-ports) that automatically deletes transient ports on system startup. If we mark a port as transient then OVS won't restore its state on restart after crash. This change may render "--may-exist" argument redundant, but I'm not sure about all the implications if it was removed. Let's keep it for now. Resolves: https://gitlab.com/libvirt/libvirt/-/issues/615 Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/util/virnetdevopenvswitch.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/util/virnetdevopenvswitch.c b/src/util/virnetdevopenvswitch.c index f1765ae1c8..e23f4c83b6 100644 --- a/src/util/virnetdevopenvswitch.c +++ b/src/util/virnetdevopenvswitch.c @@ -164,7 +164,9 @@ int virNetDevOpenvswitchAddPort(const char *brname, const char *ifname, cmd = virNetDevOpenvswitchCreateCmd(&errbuf); virCommandAddArgList(cmd, "--", "--may-exist", - "add-port", brname, ifname, NULL); + "add-port", brname, ifname, + "--", "set", "Port", ifname, "other_config:transient=true", + NULL); virNetDevOpenvswitchConstructVlans(cmd, virtVlan); -- 2.43.2

2 1

[PATCH 0/3] qemu_validate: Reject virtiofs with bootindex on s390 with ccw bus
by Peter Krempa 22 Apr '24

22 Apr '24

qemu doesn't support booting there and thus doesn't even expose bootindex. Peter Krempa (3): qemuxmlconftest: Add test case for virtiofs on s390 using 'ccw' addresses qemuxmlconftest: Decouple input and output files of 'vhost-user-fs-hugepage' case qemu_validate: Reject virtiofs with bootindex on s390x with CCW src/qemu/qemu_validate.c | 19 ++-- ...ost-user-fs-ccw-bootindex.s390x-latest.err | 1 + .../vhost-user-fs-ccw-bootindex.xml | 38 ++++++++ .../vhost-user-fs-ccw.s390x-latest.args | 37 ++++++++ .../vhost-user-fs-ccw.s390x-latest.xml | 43 +++++++++ tests/qemuxmlconfdata/vhost-user-fs-ccw.xml | 40 +++++++++ ...vhost-user-fs-hugepages.x86_64-latest.args | 2 + .../vhost-user-fs-hugepages.x86_64-latest.xml | 87 ++++++++++++++++++- .../vhost-user-fs-hugepages.xml | 8 ++ tests/qemuxmlconftest.c | 3 + 10 files changed, 272 insertions(+), 6 deletions(-) create mode 100644 tests/qemuxmlconfdata/vhost-user-fs-ccw-bootindex.s390x-latest.err create mode 100644 tests/qemuxmlconfdata/vhost-user-fs-ccw-bootindex.xml create mode 100644 tests/qemuxmlconfdata/vhost-user-fs-ccw.s390x-latest.args create mode 100644 tests/qemuxmlconfdata/vhost-user-fs-ccw.s390x-latest.xml create mode 100644 tests/qemuxmlconfdata/vhost-user-fs-ccw.xml mode change 120000 => 100644 tests/qemuxmlconfdata/vhost-user-fs-hugepages.x86_64-latest.xml -- 2.44.0

2 6

[PATCH 0/2] qemu: Substract isolcpus from all online affinity
by Michal Privoznik 22 Apr '24

22 Apr '24

*** BLURB HERE *** Michal Prívozník (2): virhostcpu: Introduce virHostCPUGetIsolated() qemu: Substract isolcpus from all online affinity src/libvirt_private.syms | 1 + src/qemu/qemu_process.c | 7 +++++++ src/util/virhostcpu.c | 21 +++++++++++++++++++++ src/util/virhostcpu.h | 1 + 4 files changed, 30 insertions(+) -- 2.43.2

4 7

[PATCH 00/10] qemu: Introduce shared_filesystems configuration option
by Andrea Bolognani 19 Apr '24

19 Apr '24

An alternative take on [1] based on review feedback. The need to have something like this in the first place is driven by KubeVirt (see [2] and [3]). A draft version of this series has been integrated into KubeVirt and it has been confirmed that it was effective in removing the need to use LD_PRELOAD hacks in the storage provider. CC'ing Stefan so he can have a look at the TPM part and shout if I've gotten anything wrong :) [1] https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/MMKV… [2] https://issues.redhat.com/browse/CNV-34322 [3] https://issues.redhat.com/browse/CNV-39370 Andrea Bolognani (10): security: Fix alignment security: Fix name for _virSecurityDACChardevCallbackData security: Drop virSecurity(DAC|SELinux)RestoreImageLabelSingle() security: Drop virSecurity(DAC|SELinux)SetImageLabelRelative() qemu: Tweak augeas schema qemu: Introduce shared_filesystems configuration option qemu: Propagate shared_filesystems utils: Use overrides in virFileIsSharedFS() qemu: Always set labels for TPM state NEWS: Document qemu shared_filesystems option NEWS.rst | 7 +++ src/lxc/lxc_controller.c | 2 +- src/lxc/lxc_driver.c | 2 +- src/lxc/lxc_process.c | 4 +- src/qemu/libvirtd_qemu.aug | 11 ++-- src/qemu/qemu.conf.in | 17 ++++++ src/qemu/qemu_conf.c | 17 ++++++ src/qemu/qemu_conf.h | 2 + src/qemu/qemu_domain.c | 2 +- src/qemu/qemu_extdevice.c | 2 +- src/qemu/qemu_migration.c | 12 ++-- src/qemu/qemu_security.c | 14 ++++- src/qemu/qemu_tpm.c | 36 ++++++------ src/qemu/qemu_tpm.h | 8 ++- src/qemu/test_libvirtd_qemu.aug.in | 5 ++ src/security/security_apparmor.c | 2 + src/security/security_dac.c | 67 +++++++++------------- src/security/security_driver.h | 4 ++ src/security/security_manager.c | 34 +++++++----- src/security/security_manager.h | 20 ++++--- src/security/security_nop.c | 4 ++ src/security/security_selinux.c | 58 ++++++++----------- src/security/security_stack.c | 16 ++++-- src/util/virfile.c | 89 +++++++++++++++++++++++++----- src/util/virfile.h | 3 +- tests/securityselinuxlabeltest.c | 2 +- tests/virfiletest.c | 2 +- 27 files changed, 289 insertions(+), 153 deletions(-) -- 2.44.0

4 44

Plans for 10.3.0 release (freeze on Friday 26 Apr)
by Jiri Denemark 19 Apr '24

19 Apr '24

We are getting close to 10.3.0 release of libvirt. To aim for the release on Thursday 02 May I suggest entering the freeze on Friday 26 Apr and tagging RC2 on Tuesday 30 Apr. I hope this works for everyone. Jirka

1 0

[RFC PATCH v1 00/15] node_dev_udev: use workerpool and improve nodedev events
by Marc Hartmayer 19 Apr '24

19 Apr '24

When an udev event occurs for a mediated device (mdev) the mdev config data requires an update via mdevctl as the udev event does not contain all config data. This update needs to occur immediate and to be finished before the libvirt nodedev event is issued to keep the API usage reliable. This patch series already contains the patches from a previous patch series "[RFC PATCH v1 0/5] node_device_udev: small improvements" and has still some TODO's included and is sent therefore as a RFC. Boris Fiuczynski (3): nodedev: fix mdev add udev event data handling nodedev: immediate update of active config on udev add nodedev: reset active config data on udev remove event Marc Hartmayer (12): node_device_udev: Set @def to NULL node_device_udev: Remove the timeout if the data is disposed node_device_udev: Test for mdevctlTimeout != -1 node_device_udev: Add comments about locking node_device_udev: Take lock if `driver->privateData` is modified node_device_udev: Add prefix `udev` for udev related data node_device_udev: Inline `udevRemoveOneDevice` node_device_udev: Use `stateShutdownPrepare` and `stateShutdownWait` node_device_udev: Use a worker pool for processing the udev events node_device_udev: Call `nodeDeviceUpdateMediatedDevices` directly node_device_udev: Don't take `mdevctl` lock for querying `mdevctl list` node_device_udev: Make the code easier to read src/node_device/node_device_driver.h | 5 +- src/util/virmdev.h | 4 + src/conf/node_device_conf.c | 10 +- src/node_device/node_device_driver.c | 19 +- src/node_device/node_device_udev.c | 435 ++++++++++++++++++--------- src/test/test_driver.c | 3 +- src/util/virmdev.c | 20 ++ src/libvirt_private.syms | 2 + 8 files changed, 339 insertions(+), 159 deletions(-) base-commit: 4b5cc57ed35dc24d11673dd3f04bfb8073c0340d -- 2.34.1

3 39

[PATCH v3 0/5] test: fix nodedev mdev XML regression
by Cole Robinson 18 Apr '24

18 Apr '24

The virt-manager test suite is busted with libvirt 10.1.0+ after this commit: commit e67bca23e4fe38a3491749f724b9edf743d0e916 Author: Boris Fiuczynski <fiuczy(a)linux.ibm.com> Date: Thu Feb 22 14:02:01 2024 +0100 nodedev: add an active config to mdev See patch #5 for the full explanation. First 4 patches are nodedev test driver improvements I hit when debugging this v3: really truly send to the correct list v2: Send to the correct mailing list Fix version strings in test driver table Cole Robinson (5): test: Fix `virsh nodedev-list` test: Implement virNodeDeviceIsActive test: Implement virNodeDeviceIsPersistent test: make nodedevs active by default conf: nodedev: Fill active_config at XML parse time src/conf/node_device_conf.c | 5 ++++- src/test/test_driver.c | 35 ++++++++++++++++++++++++++++++++++- tests/nodedevxml2xmltest.c | 15 --------------- 3 files changed, 38 insertions(+), 17 deletions(-) -- 2.44.0

5 21

[PATCH] libvirt_nss: Fix ERROR() macro
by Michal Privoznik 18 Apr '24

18 Apr '24

The purpose of ERROR() macro in our NSS module is to print error message provided as arguments followed by error string corresponding to errno. Historically, we've used strerror_r() for that (please note, we want our NSS module to be free of libvirt internal functions, or glib even - hence, g_strerror() is off the table). Now strerror_r() is documented as: Returns ... a pointer to a string that the function stores in buf, or a pointer to some (immutable) static string (in which case buf is unused). Therefore, we can't rely the string being stored in the buf and really need to store the retval and print that instead. While touching this area, decrease the ebuf size, since its current size (1KiB) is triggering our stack limit (2KiB) in some cases. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- tools/nss/libvirt_nss.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/nss/libvirt_nss.h b/tools/nss/libvirt_nss.h index 2bb313f329..5f356618f3 100644 --- a/tools/nss/libvirt_nss.h +++ b/tools/nss/libvirt_nss.h @@ -37,11 +37,11 @@ # define NULLSTR(s) ((s) ? (s) : "<null>") # define ERROR(...) \ do { \ - char ebuf[1024]; \ - strerror_r(errno, ebuf, sizeof(ebuf)); \ + char ebuf[512]; \ + const char *errmsg = strerror_r(errno, ebuf, sizeof(ebuf)); \ fprintf(stderr, "ERROR %s:%d : ", __FUNCTION__, __LINE__); \ fprintf(stderr, __VA_ARGS__); \ - fprintf(stderr, " : %s\n", ebuf); \ + fprintf(stderr, " : %s\n", errmsg); \ fprintf(stderr, "\n"); \ } while (0) -- 2.43.2

2 1

[PATCH v2 0/4] qemu: Fix migration with custom XML
by Jiri Denemark 17 Apr '24

17 Apr '24

Version 2: - see patch 1/4 for more details - added two cleanups Jiri Denemark (4): qemu: Fix migration with custom XML NEWS: Mention migration bug with custom XML qemu: Change return type of qemuDomainUpdateCPU to void qemu: Change return type of qemuDomainFixupCPUs to void NEWS.rst | 8 +++++++ src/qemu/qemu_domain.c | 51 ++++++++++++++++++++--------------------- src/qemu/qemu_domain.h | 4 ++-- src/qemu/qemu_process.c | 29 ++++++++--------------- 4 files changed, 45 insertions(+), 47 deletions(-) -- 2.44.0

2 7