- Devel - libvirt List Archives

[PATCH 0/4] ui: add support for 'secret' object to provide VNC/SPICE passwords
by Daniel P. Berrangé 11 Mar '21

11 Mar '21

This fixes a long standing limitation of the VNC/SPICE code which was unable to securely accept passswords on the CLI, instead requiring use of separate monitor commands after startup. This takes the opportunity to also remove previously deprecated ACL functionality from VNC. Daniel P. Berrangé (4): ui: introduce "password-secret" option for VNC servers ui: introduce "password-secret" option for SPICE server ui: deprecate "password" option for SPICE server ui, monitor: remove deprecated VNC ACL option and HMP commands docs/system/deprecated.rst | 24 ++-- docs/system/removed-features.rst | 13 +++ hmp-commands.hx | 76 ------------- monitor/misc.c | 187 ------------------------------- qemu-options.hx | 17 ++- ui/spice-core.c | 32 +++++- ui/vnc.c | 61 ++++------ 7 files changed, 88 insertions(+), 322 deletions(-) -- 2.29.2

3 8

[PATCH 00/23] RFC: get rid of macros when dealing with block io tunes
by Nikolay Shirokovskiy 11 Mar '21

11 Mar '21

Hi, all. I started this work as adding missing parts/fixing issues/etc in block iotune code but then turned to refactoring code. We use a lot of macros in this place and if we get rid of them I belive we will have much more readable/reusable/ extendable code. Most of macros usage is for iterating over unsigned long long values. I'm talking about parsing/formating xml, converting from/to virTypedParameterPtr list etc. These places do not care about tune semantics and thus we can add tools for the said iteration. See patch [1] for the first such conversion. Patches before it partially prepare for this conversion, partially just improve code reuse and add missing parts. The work on removing macros in code handling iotunes is not finished as I wanted to get an approvement that I have taken a right direction. At the same time series shows the potential of the approach (take a look on how qemuDomainSetBlockIoTune and testDomainSetBlockIoTune are looking now!). Some places like qemuDomainSetBlockIoTuneDefaults deal with fields semantics. So we need another approach to remove macros there but it is a matter of a different RFC. Nikolay Shirokovskiy (23): qemu: pass supportGroupNameOption as expected qemu: factor out qemuValidateDomainBlkdeviotune qemu: reuse validation in qemuDomainSetBlockIoTune qemu: remove extra check for QEMU_BLOCK_IOTUNE_MAX conf: factor out virDomainBlockIoTuneValidate qemu: reuse virDomainBlockIoTuneValidate test driver: reuse virDomainBlockIoTuneValidate qemu: reset max iotune setting when needed qemu: add max iotune settings check to virDomainBlockIoTuneValidate qemu: remove iotune max checks test driver: remove iotune max checks qemu: prepare for removing qemuBlockIoTuneSetFlags qemu: use group_name instead of QEMU_BLOCK_IOTUNE_SET_GROUP_NAME qemu: remove qemuBlockIoTuneSetFlags usage in qemuDomainSetBlockIoTune qemu: remove qemuBlockIoTuneSetFlags enum completly conf: get rid of macros in virDomainDiskDefIotuneParse [1] conf: get rid of macros in virDomainDiskDefFormatIotune conf: add virDomainBlockIoTuneFromParams conf: add virDomainBlockIoTuneToEventParams qemu: qemuDomainSetBlockIoTune use functions to convert to/from params test driver: remove TEST_BLOCK_IOTUNE_MAX checks test driver: prepare to delete macros in testDomainSetBlockIoTune test driver: testDomainSetBlockIoTune: use functions to convert to/from params src/conf/domain_conf.c | 303 +++++++++++++++++++++++++++++++++-------------- src/conf/domain_conf.h | 16 +++ src/libvirt_private.syms | 3 + src/qemu/qemu_driver.c | 281 ++++++++++++------------------------------- src/qemu/qemu_validate.c | 100 +++++++++------- src/qemu/qemu_validate.h | 4 + src/test/test_driver.c | 156 ++---------------------- 7 files changed, 380 insertions(+), 483 deletions(-) -- 1.8.3.1

2 42

[PATCH v3 0/4] Drop deprecated floppy config & bogus -drive if=T
by Markus Armbruster 11 Mar '21

11 Mar '21

v3: * PATCH 1: New [Daniel] v2: * Rebased, straightforward conflict with commit f5d33dd51f "hw/block/fdc: Remove the check_media_rate property" resolved * PATCH 2: Commit message fixed [Kevin] Markus Armbruster (3): fdc: Drop deprecated floppy configuration fdc: Inline fdctrl_connect_drives() into fdctrl_realize_common() blockdev: Drop deprecated bogus -drive interface type docs/system/deprecated.rst | 33 -- docs/system/removed-features.rst | 33 ++ include/sysemu/blockdev.h | 1 - blockdev.c | 37 +- hw/block/fdc.c | 73 +--- softmmu/vl.c | 8 +- tests/qemu-iotests/172 | 31 +- tests/qemu-iotests/172.out | 562 +------------------------------ 8 files changed, 59 insertions(+), 719 deletions(-) -- 2.26.2

4 13

[PATCH v2 00/12] qemu: Prepare for QAPIfied object-add
by Peter Krempa 10 Mar '21

10 Mar '21

QEMU plans to QAPIfy object add. This series prepares for the API change (drop of 'props' wrapper for the object) and adds testing based on our qemuxml2argv test data which forces the output to JSON and validates it agains the schema. Based on Kevin's qemu patches: https://listman.redhat.com/archives/libvir-list/2021-February/msg01168.html Last patch forces more files to be processed and as the summary suggest will not be pushed. The patch updating qemucapabilitiesdata will be updated after Kevin's patches hit upstream repo. Peter Krempa (12): qemu: command: Generate commandline of 'masterKey0' secret via JSON qemu: command: Generate commandline of 'sev0' sev-guest object via JSON qemu: command: Generate commandline of iothread objects JSON qemu: capabilities: Introduce QEMU_CAPS_OBJECT_QAPIFIED tests: qemuxml2argv: Validate generation of JSON props for object-add qemu: command: Introduce raw JSON passthrough for '-object' for testing qemu: monitor: Make wrapping of 'props' of 'object-add' optional qemuMonitorCreateObjectPropsWrap: Open-code in qemuBuildMemoryBackendProps qemu: monitor: Don't add 'props' wrapper if qemu has QEMU_CAPS_OBJECT_QAPIFIED qemumonitorjsontest: Remove bomb guarding object-add tests: qemucapabilities: Update qemu caps for object-add qapification [DON'T PUSH] Force-check all configs with latest capabilities src/qemu/qemu_capabilities.c | 4 + src/qemu/qemu_capabilities.h | 3 + src/qemu/qemu_command.c | 326 +- src/qemu/qemu_monitor.c | 78 +- src/qemu/qemu_monitor.h | 4 - src/util/virqemu.c | 48 +- src/util/virqemu.h | 3 +- .../caps_6.0.0.x86_64.replies | 3238 ++++++++++++----- .../caps_6.0.0.x86_64.xml | 83 +- tests/qemumonitorjsontest.c | 14 - ...v-missing-platform-info.x86_64-2.12.0.args | 2 +- .../launch-security-sev.x86_64-2.12.0.args | 2 +- tests/qemuxml2argvtest.c | 24 +- 13 files changed, 2775 insertions(+), 1054 deletions(-) -- 2.29.2

2 18

[PATCH] XML validate that 'ramfb' has no address
by Kristina Hanicova 10 Mar '21

10 Mar '21

With this, XML fails if config video type 'ramfb' contains address, since address is not supported for 'ramfb' video devices. Previously it didn't raise error. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1891416 Signed-off-by: Kristina Hanicova <khanicov(a)redhat.com> --- src/conf/domain_validate.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/src/conf/domain_validate.c b/src/conf/domain_validate.c index b4e09e21fe..c9c677ddf5 100644 --- a/src/conf/domain_validate.c +++ b/src/conf/domain_validate.c @@ -151,6 +151,14 @@ virDomainVideoDefValidate(const virDomainVideoDef *video, } } + if (video->type == VIR_DOMAIN_VIDEO_TYPE_RAMFB) { + if (video->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("address not supported for video type ramfb")); + return -1; + } + } + return 0; } -- 2.29.2

2 1

[PATCH] virQEMUCapsInitQMPArch: Refactor cleanup
by Yi Li 10 Mar '21

10 Mar '21

Switch to using the 'g_auto*' helpers. Signed-off-by: Yi Li <yili(a)winhong.com> --- src/qemu/qemu_capabilities.c | 13 ++++--------- 1 file changed, 4 insertions(+), 9 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 790375ac38..447cf77875 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -5044,23 +5044,18 @@ static int virQEMUCapsInitQMPArch(virQEMUCapsPtr qemuCaps, qemuMonitorPtr mon) { - char *archstr = NULL; - int ret = -1; + g_autofree char *archstr = NULL; if (!(archstr = qemuMonitorGetTargetArch(mon))) - goto cleanup; + return -1; if ((qemuCaps->arch = virQEMUCapsArchFromString(archstr)) == VIR_ARCH_NONE) { virReportError(VIR_ERR_INTERNAL_ERROR, _("Unknown QEMU arch %s"), archstr); - goto cleanup; + return -1; } - ret = 0; - - cleanup: - VIR_FREE(archstr); - return ret; + return 0; } -- 2.25.3

2 1

[PATCH 0/4] Discourage use of some GLib functions
by Michal Privoznik 10 Mar '21

10 Mar '21

See 4/4 for the list and explanation. Michal Prívozník (4): qemu_shim: Replace g_file_get_contents() with virFileReadAll() virutil: Do not use g_get_host_name() to obtain hostname syntax-check: Update list of gethostname exceptions docs: Discourage use of some glib functions build-aux/syntax-check.mk | 2 +- docs/glib-adoption.rst | 17 +++++++++++++++++ src/qemu/qemu_shim.c | 18 +++++++++++------- src/util/virutil.c | 12 +++++++++--- 4 files changed, 38 insertions(+), 11 deletions(-) -- 2.26.2

3 8

[PATCH 00/16] Improve virStrcpy and remove virStrncpy
by Peter Krempa 10 Mar '21

10 Mar '21

Peter Krempa (16): util: virstring: Always copy string in virStrcpy virProcessRunInForkHelper: Use virStrcpyStatic for static buffers sanlock: Use virStrcpy instead of virStrncpy virDevMapperGetTargetsImpl: Use virStrcpy instead of virStrncpy virFileLoopDeviceAssociate: Use virStrcpy instead of virStrncpy xenParseXMDisk: Replace g_new + virStrncpy by g_strndup test_driver: Rewrite testBuildFilename xenParseVif: Refactor parser xenParseSxprSound: Refactor parsing of model list openvzReadNetworkConf: Rework parser xenParseXLChannel: Use g_strndup instead of virStrndup xenParseXLUSBController: Avoid use of virStrndup xenParseXLUSB: Rewrite to avoid virStrncpy virNetLibsshAuthenticatePrivkeyCb: Use g_autofree for 'actual_prompt' virNetLibsshAuthenticatePrivkeyCb: Use virStrcpy instead of virStrncpy util: virstring: Remove virStrncpy docs/coding-style.rst | 13 --- src/libvirt_private.syms | 1 - src/libxl/xen_common.c | 182 +++++++++++------------------- src/libxl/xen_xl.c | 84 ++++---------- src/libxl/xen_xm.c | 9 +- src/locking/lock_driver_sanlock.c | 4 +- src/openvz/openvz_conf.c | 77 +++++-------- src/rpc/virnetlibsshsession.c | 17 +-- src/test/test_driver.c | 35 +----- src/util/virdevmapper.c | 2 +- src/util/virfile.c | 3 +- src/util/virprocess.c | 8 +- src/util/virstring.c | 54 +-------- src/util/virstring.h | 5 +- 14 files changed, 137 insertions(+), 357 deletions(-) -- 2.29.2

2 33

RFC: do we want/need the "Ptr" typedefs for internal code ?
by Daniel P. Berrangé 10 Mar '21

10 Mar '21

One of the conventions we have had since the early days of libvirt is that every struct typedef, has a corresponding "Ptr" typedef too. For example typedef struct _virDomainDef virDomainDef; typedef virDomainDef *virDomainDefPtr; Periodically someone has questioned what the purpose of these Ptr typedefs is, and we've not had an compelling answer, other than that's what we've always done. There are a few things that make me question the status quo - If we want a const pointer, we can't use const virDomainDefPtr def because that expands to "struct _virDomainDef * const", which is not what we need semantically. Instead we must write const virDomainDef *def It is not at all obvious why these two are different, but none the less they are, which is confusing to contributors To me this a compelling reason to consider the "Ptr" typedefs a waste of time, if not actively harmful. Please don't suggest adding virDomainDefConstPtr too ! - Writing 'virDomainDefPtr' is actually two characters more typing than 'virDomainDef *'. IOW these "Ptr" typedefs aren't saving us time when writing code. - This convention of having "Ptr" typedefs is atypical among C projects I've worked on. Anything that is peculiar to libvirt is another item that new contributors need to learn, so has a cost to the project that must be weighed against its benefit. We can't do anything about the use "Ptr" in the include/ files because that is public ABI. We can potentially eliminate "Ptr" types everywhere else in the codebase, even the src/libvirt*.c files corresponding to the public includes. Does anyone have suggestions for how these "Ptr" typedefs are benefiting libvirt ? Would anyone miss them ? Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|

7 10

[PATCH 0/2] fix crash on invalid construction of a gvariant
by Peter Krempa 10 Mar '21

10 Mar '21

Peter Krempa (2): virSystemdCreateMachine: Use proper format string for uint64_t when constructing gvariant virsystemdtest: Call at least one virSystemdCreateMachine with 'maxthreads' > 0 src/util/virsystemd.c | 5 +++-- tests/virsystemdtest.c | 2 +- 2 files changed, 4 insertions(+), 3 deletions(-) -- 2.29.2

2 3