February 2022 - Devel - libvirt List Archives

[PATCH v4 2/4] hw/i386: Attach CPUs to machine
by Philippe Mathieu-Daudé 07 Feb '22

07 Feb '22

Previously CPUs were exposed in the QOM tree at a path /machine/unattached/device[nn] where the 'nn' of the first CPU is usually zero, but can vary depending on what devices were already created. With this change the CPUs are now at /machine/cpu[nn] where the 'nn' of the first CPU is always zero. Note: This (intentionally) breaks compatibility with current libvirt code that looks for "/machine/unattached/device[0]" in the assumption it is the first CPU. Cc: libvir-list(a)redhat.com Suggested-by: Daniel P. Berrangé <berrange(a)redhat.com> Reviewed-by: Daniel P. Berrangé <berrange(a)redhat.com> Signed-off-by: Philippe Mathieu-Daudé <f4bug(a)amsat.org> --- hw/i386/x86.c | 1 + 1 file changed, 1 insertion(+) diff --git a/hw/i386/x86.c b/hw/i386/x86.c index b84840a1bb9..50bf249c700 100644 --- a/hw/i386/x86.c +++ b/hw/i386/x86.c @@ -108,6 +108,7 @@ void x86_cpu_new(X86MachineState *x86ms, int64_t apic_id, Error **errp) { Object *cpu = object_new(MACHINE(x86ms)->cpu_type); + object_property_add_child(OBJECT(x86ms), "cpu[*]", OBJECT(cpu)); if (!object_property_set_uint(cpu, "apic-id", apic_id, errp)) { goto out; } -- 2.34.1

6 12

[PULL 31/31] linux-user: Remove the deprecated ppc64abi32 target
by Alex Bennée 07 Feb '22

07 Feb '22

From: Thomas Huth <thuth(a)redhat.com> It's likely broken, and nobody cared for picking it up again during the deprecation phase, so let's remove this now. Since this is the last entry in deprecated_targets_list, remove the related code in the configure script, too. Signed-off-by: Thomas Huth <thuth(a)redhat.com> Reviewed-by: Richard Henderson <richard.henderson(a)linaro.org> Acked-by: Cédric Le Goater <clg(a)kaod.org> Acked-by: Alex Bennée <alex.bennee(a)linaro.org> Message-Id: <20211215084958.185214-1-thuth(a)redhat.com> Signed-off-by: Alex Bennée <alex.bennee(a)linaro.org> Message-Id: <20220105135009.1584676-35-alex.bennee(a)linaro.org> diff --git a/docs/about/deprecated.rst b/docs/about/deprecated.rst index e21e07478f..6f85afdee4 100644 --- a/docs/about/deprecated.rst +++ b/docs/about/deprecated.rst @@ -390,13 +390,6 @@ The above, converted to the current supported format:: linux-user mode CPUs -------------------- -``ppc64abi32`` CPUs (since 5.2) -''''''''''''''''''''''''''''''' - -The ``ppc64abi32`` architecture has a number of issues which regularly -trip up our CI testing and is suspected to be quite broken. For that -reason the maintainers strongly suspect no one actually uses it. - MIPS ``I7200`` CPU (since 5.2) '''''''''''''''''''''''''''''' diff --git a/docs/about/removed-features.rst b/docs/about/removed-features.rst index 4c4da20d0f..380a1b30ea 100644 --- a/docs/about/removed-features.rst +++ b/docs/about/removed-features.rst @@ -601,6 +601,14 @@ the upstream Linux kernel in 2018, and it has also been dropped from glibc, so there is no new Linux development taking place with this architecture. For running the old binaries, you can use older versions of QEMU. +``ppc64abi32`` CPUs (removed in 7.0) +'''''''''''''''''''''''''''''''''''' + +The ``ppc64abi32`` architecture has a number of issues which regularly +tripped up the CI testing and was suspected to be quite broken. For that +reason the maintainers strongly suspected no one actually used it. + + System emulator devices ----------------------- diff --git a/docs/user/main.rst b/docs/user/main.rst index e08d4be63b..6f2ffa080f 100644 --- a/docs/user/main.rst +++ b/docs/user/main.rst @@ -166,7 +166,6 @@ Other binaries - user mode (PowerPC) - * ``qemu-ppc64abi32`` TODO. * ``qemu-ppc64`` TODO. * ``qemu-ppc`` TODO. diff --git a/configure b/configure index 030728d11e..0c57a063c6 100755 --- a/configure +++ b/configure @@ -1273,8 +1273,6 @@ if [ "$ARCH" = "unknown" ]; then fi default_target_list="" -deprecated_targets_list=ppc64abi32-linux-user -deprecated_features="" mak_wilds="" if [ "$softmmu" = "yes" ]; then @@ -1287,16 +1285,6 @@ if [ "$bsd_user" = "yes" ]; then mak_wilds="${mak_wilds} $source_path/configs/targets/*-bsd-user.mak" fi -# If the user doesn't explicitly specify a deprecated target we will -# skip it. -if test -z "$target_list"; then - if test -z "$target_list_exclude"; then - target_list_exclude="$deprecated_targets_list" - else - target_list_exclude="$target_list_exclude,$deprecated_targets_list" - fi -fi - for config in $mak_wilds; do target="$(basename "$config" .mak)" if echo "$target_list_exclude" | grep -vq "$target"; then @@ -1315,11 +1303,9 @@ Standard options: --prefix=PREFIX install in PREFIX [$prefix] --interp-prefix=PREFIX where to find shared libraries, etc. use %M for cpu name [$interp_prefix] - --target-list=LIST set target list (default: build all non-deprecated) + --target-list=LIST set target list (default: build all) $(echo Available targets: $default_target_list | \ fold -s -w 53 | sed -e 's/^/ /') -$(echo Deprecated targets: $deprecated_targets_list | \ - fold -s -w 53 | sed -e 's/^/ /') --target-list-exclude=LIST exclude a set of targets from the default target-list Advanced options (experts only): @@ -1804,13 +1790,6 @@ else done fi -for target in $target_list; do - # if a deprecated target is enabled we note it here - if echo "$deprecated_targets_list" | grep -q "$target"; then - add_to deprecated_features $target - fi -done - # see if system emulation was really requested case " $target_list " in *"-softmmu "*) softmmu=yes @@ -3950,12 +3929,6 @@ else fi fi -if test -n "${deprecated_features}"; then - echo "Warning, deprecated features enabled." - echo "Please see docs/about/deprecated.rst" - echo " features: ${deprecated_features}" -fi - # Create list of config switches that should be poisoned in common code... # but filter out CONFIG_TCG and CONFIG_USER_ONLY which are special. target_configs_h=$(ls *-config-devices.h *-config-target.h 2>/dev/null) diff --git a/configs/targets/ppc64abi32-linux-user.mak b/configs/targets/ppc64abi32-linux-user.mak deleted file mode 100644 index 0945451081..0000000000 --- a/configs/targets/ppc64abi32-linux-user.mak +++ /dev/null @@ -1,8 +0,0 @@ -TARGET_ARCH=ppc64 -TARGET_ABI32=y -TARGET_BASE_ARCH=ppc -TARGET_ABI_DIR=ppc -TARGET_SYSTBL_ABI=common,nospu,32 -TARGET_SYSTBL=syscall.tbl -TARGET_WORDS_BIGENDIAN=y -TARGET_XML_FILES= gdb-xml/power64-core.xml gdb-xml/power-fpu.xml gdb-xml/power-altivec.xml gdb-xml/power-spe.xml gdb-xml/power-vsx.xml diff --git a/linux-user/ppc/target_syscall.h b/linux-user/ppc/target_syscall.h index 8b364697d4..7df9118937 100644 --- a/linux-user/ppc/target_syscall.h +++ b/linux-user/ppc/target_syscall.h @@ -36,7 +36,7 @@ struct target_pt_regs { abi_ulong link; abi_ulong xer; abi_ulong ccr; -#if defined(TARGET_PPC64) && !defined(TARGET_ABI32) +#if defined(TARGET_PPC64) abi_ulong softe; #else abi_ulong mq; /* 601 only (not used at present) */ @@ -58,7 +58,7 @@ struct target_revectored_struct { * flags masks */ -#if defined(TARGET_PPC64) && !defined(TARGET_ABI32) +#if defined(TARGET_PPC64) #ifdef TARGET_WORDS_BIGENDIAN #define UNAME_MACHINE "ppc64" #else diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h index cca561f622..dc7dcad6f7 100644 --- a/linux-user/syscall_defs.h +++ b/linux-user/syscall_defs.h @@ -1607,7 +1607,7 @@ struct target_stat64 { struct target_stat { abi_ulong st_dev; abi_ulong st_ino; -#if defined(TARGET_PPC64) && !defined(TARGET_ABI32) +#if defined(TARGET_PPC64) abi_ulong st_nlink; unsigned int st_mode; #else @@ -1628,12 +1628,12 @@ struct target_stat { abi_ulong target_st_ctime_nsec; abi_ulong __unused4; abi_ulong __unused5; -#if defined(TARGET_PPC64) && !defined(TARGET_ABI32) +#if defined(TARGET_PPC64) abi_ulong __unused6; #endif }; -#if !defined(TARGET_PPC64) || defined(TARGET_ABI32) +#if !defined(TARGET_PPC64) #define TARGET_HAS_STRUCT_STAT64 struct QEMU_PACKED target_stat64 { unsigned long long st_dev; diff --git a/linux-user/elfload.c b/linux-user/elfload.c index cb2f01ff01..107796c44d 100644 --- a/linux-user/elfload.c +++ b/linux-user/elfload.c @@ -718,7 +718,7 @@ static inline void init_thread(struct target_pt_regs *regs, #define ELF_MACHINE PPC_ELF_MACHINE #define ELF_START_MMAP 0x80000000 -#if defined(TARGET_PPC64) && !defined(TARGET_ABI32) +#if defined(TARGET_PPC64) #define elf_check_arch(x) ( (x) == EM_PPC64 ) @@ -870,7 +870,7 @@ static uint32_t get_elf_hwcap2(void) static inline void init_thread(struct target_pt_regs *_regs, struct image_info *infop) { _regs->gpr[1] = infop->start_stack; -#if defined(TARGET_PPC64) && !defined(TARGET_ABI32) +#if defined(TARGET_PPC64) if (get_ppc64_abi(infop) < 2) { uint64_t val; get_user_u64(val, infop->entry + 8); diff --git a/linux-user/ppc/signal.c b/linux-user/ppc/signal.c index 176c9d8503..ec0b9c0df3 100644 --- a/linux-user/ppc/signal.c +++ b/linux-user/ppc/signal.c @@ -477,9 +477,7 @@ void setup_rt_frame(int sig, struct target_sigaction *ka, int i, err = 0; #if defined(TARGET_PPC64) struct target_sigcontext *sc = 0; -#if !defined(TARGET_ABI32) struct image_info *image = ((TaskState *)thread_cpu->opaque)->info; -#endif #endif rt_sf_addr = get_sigframe(ka, env, sizeof(*rt_sf)); @@ -530,7 +528,7 @@ void setup_rt_frame(int sig, struct target_sigaction *ka, env->gpr[5] = (target_ulong) h2g(&rt_sf->uc); env->gpr[6] = (target_ulong) h2g(rt_sf); -#if defined(TARGET_PPC64) && !defined(TARGET_ABI32) +#if defined(TARGET_PPC64) if (get_ppc64_abi(image) < 2) { /* ELFv1 PPC64 function pointers are pointers to OPD entries. */ struct target_func_ptr *handler = @@ -562,7 +560,7 @@ sigsegv: } -#if !defined(TARGET_PPC64) || defined(TARGET_ABI32) +#if !defined(TARGET_PPC64) long do_sigreturn(CPUPPCState *env) { struct target_sigcontext *sc = NULL; @@ -575,12 +573,9 @@ long do_sigreturn(CPUPPCState *env) if (!lock_user_struct(VERIFY_READ, sc, sc_addr, 1)) goto sigsegv; -#if defined(TARGET_PPC64) - set.sig[0] = sc->oldmask + ((uint64_t)(sc->_unused[3]) << 32); -#else __get_user(set.sig[0], &sc->oldmask); __get_user(set.sig[1], &sc->_unused[3]); -#endif + target_to_host_sigset_internal(&blocked, &set); set_sigmask(&blocked); diff --git a/.gitlab-ci.d/buildtest.yml b/.gitlab-ci.d/buildtest.yml index 8f2a3c8f5b..0aa70213fb 100644 --- a/.gitlab-ci.d/buildtest.yml +++ b/.gitlab-ci.d/buildtest.yml @@ -473,33 +473,6 @@ tsan-build: TARGETS: x86_64-softmmu ppc64-softmmu riscv64-softmmu x86_64-linux-user MAKE_CHECK_ARGS: bench V=1 -# These targets are on the way out -build-deprecated: - extends: .native_build_job_template - needs: - job: amd64-debian-user-cross-container - variables: - IMAGE: debian-all-test-cross - CONFIGURE_ARGS: --disable-tools - MAKE_CHECK_ARGS: build-tcg - TARGETS: ppc64abi32-linux-user - artifacts: - expire_in: 2 days - paths: - - build - -# We split the check-tcg step as test failures are expected but we still -# want to catch the build breaking. -check-deprecated: - extends: .native_test_job_template - needs: - - job: build-deprecated - artifacts: true - variables: - IMAGE: debian-all-test-cross - MAKE_CHECK_ARGS: check-tcg - allow_failure: true - # gprof/gcov are GCC features build-gprof-gcov: extends: .native_build_job_template diff --git a/tests/docker/dockerfiles/debian-ppc64el-cross.docker b/tests/docker/dockerfiles/debian-ppc64el-cross.docker index 1146a06be6..5de12b01cd 100644 --- a/tests/docker/dockerfiles/debian-ppc64el-cross.docker +++ b/tests/docker/dockerfiles/debian-ppc64el-cross.docker @@ -16,7 +16,7 @@ RUN apt update && \ # Specify the cross prefix for this image (see tests/docker/common.rc) ENV QEMU_CONFIGURE_OPTS --cross-prefix=powerpc64le-linux-gnu- -ENV DEF_TARGET_LIST ppc64-softmmu,ppc64-linux-user,ppc64abi32-linux-user +ENV DEF_TARGET_LIST ppc64-softmmu,ppc64-linux-user # Install extra libraries to increase code coverage RUN apt update && \ diff --git a/tests/tcg/configure.sh b/tests/tcg/configure.sh index 309335a2bd..763e9b6ad8 100755 --- a/tests/tcg/configure.sh +++ b/tests/tcg/configure.sh @@ -167,7 +167,7 @@ for target in $target_list; do container_image=debian-nios2-cross container_cross_cc=nios2-linux-gnu-gcc ;; - ppc-*|ppc64abi32-*) + ppc-*) container_hosts=x86_64 container_image=debian-powerpc-test-cross container_cross_cc=powerpc-linux-gnu-gcc-10 -- 2.30.2

2 2

[PATCH] virhostcpu: fix getting CPU freq for Apple Silicon
by Menci 04 Feb '22

04 Feb '22

The current implementation of virHostCPUGetInfo for macOS (__APPLE__) reads "hw.cpufrequency" from sysctl, which is available only on x86_64 architecture. On Apple Silicon ARM Macs, it's not available: $ sysctl hw.cpufrequency # No output $ arch -x86_64 sysctl hw.cpufrequency # Run with Rosetta 2 hw.cpufrequency: 2400000000 When running libvirtd on Apple Silicon, I got the error: cannot obtain CPU freq: No such file or directory. To fix it, we can calculate it with "hw.tbfrequency" and "kern.clockrate" instead: $ sysctl hw.tbfrequency hw.tbfrequency: 24000000 $ sysctl kern.clockrate kern.clockrate: { hz = 100, tick = 10000, tickadj = 0, profhz = 100, stathz = 100 } The result value would be hw.tbfrequency * kern.clockrate.hz. Signed-off-by: Menci <huanghaorui301(a)gmail.com> --- src/util/virhostcpu.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/src/util/virhostcpu.c b/src/util/virhostcpu.c index a07c00a0e9..72983c91f3 100644 --- a/src/util/virhostcpu.c +++ b/src/util/virhostcpu.c @@ -927,9 +927,18 @@ virHostCPUGetInfo(virArch hostarch G_GNUC_UNUSED, *mhz = cpu_freq; # else + /* This works for Intel Macs */ if (sysctlbyname("hw.cpufrequency", &cpu_freq, &cpu_freq_len, NULL, 0) < 0) { - virReportSystemError(errno, "%s", _("cannot obtain CPU freq")); - return -1; + /* On Apple Silicon fallback to hw.tbfrequency and kern.clockrate.hz */ + struct clockinfo clockrate; + size_t clockrate_len = sizeof(clockrate); + if (sysctlbyname("hw.tbfrequency", &cpu_freq, &cpu_freq_len, NULL, 0) < 0 || + sysctlbyname("kern.clockrate", &clockrate, &clockrate_len, NULL, 0) < 0) { + virReportSystemError(errno, "%s", _("cannot obtain CPU freq")); + return -1; + } + + cpu_freq *= clockrate.hz; } *mhz = cpu_freq / 1000000; -- 2.34.1

1 0

[libvirt PATCH 00/14] qemu: Improve swtpm internal API
by Andrea Bolognani 04 Feb '22

04 Feb '22

Clean up inconsistencies to make navigating the module easier. Andrea Bolognani (14): qemu: Rename qemuExtTPMStartEmulator() qemu: Fully document qemuTPMEmulatorStart() qemu: Rename qemuTPM{Create,Delete}EmulatorStorage() qemu: Document qemuTPMEmulatorDeleteStorage() qemu: Drop qemuTPMEmulatorInitStorage() qemu: Make qemuTPMEmulatorCreateStorage() take a virDomainTPMDef* qemu: Introduce qemuExtTPMEmulatorSetupCgroup() qemu: Introduce qemuTPMEmulatorCleanupHost() qemu: Rename path-building functions qemu: Call virDomainDefGetShortName() less frequently qemu: Fix description of swtpmStateDir qemu: Move utility functions close together qemu: Move entry points close together qemu: Move high-level actions close together src/qemu/qemu_tpm.c | 606 +++++++++++++++++++++++--------------------- 1 file changed, 323 insertions(+), 283 deletions(-) -- 2.34.1

2 15

[PATCH 0/2] qemu_process: Start QEMU for caps probing more robustly
by Michal Privoznik 04 Feb '22

04 Feb '22

I'm not going to lie, my true motive is to be able to have a very small bash wrapper that spawns QEMU under valgrind, e.g. like this: #!/bin/bash exec /usr/bin/valgrind --tool=memcheck --leak-check=full \ --track-fds=yes --log-file=/tmp/valgrind_qemu.log \ qemu-system-x86_64 "$@" But if a malicious user provides slightly different script (see 2/2) then they can starve libvirt worker pool. Hence, we have two reasons for merging these. Michal Prívozník (2): qemu_process: Be nicer to killing QEMU when probing caps qemu_process: Start QEMU for caps probing more robustly src/qemu/qemu_process.c | 64 +++++++++++++++++++++++++++-------------- src/qemu/qemu_process.h | 1 - 2 files changed, 43 insertions(+), 22 deletions(-) -- 2.34.1

1 2

[PATCH] qemu: Validate domain definition even on migration
by Michal Privoznik 04 Feb '22

04 Feb '22

When we are about to spawn QEMU, we validate the domain definition against qemuCaps. Except when domain is/was already running before (i.e. on incoming migration, snapshots, resume from a file). However, especially on incoming migration it may happen that the destination QEMU is different to the source QEMU, e.g. the destination QEMU may have some devices disabled. And we have a function that validates devices/features requested in domain XML against the desired QEMU capabilities (aka qemuCaps) - it's virDomainDefValidate() which calls qemuValidateDomainDef() and qemuValidateDomainDeviceDef() subsequently. But the problem here is that the validation function is explicitly skipped over in specific scenarios (like incoming migration, restore from a snapshot or previously saved file). This in turn means that we may spawn QEMU and request device/features it doesn't support. When that happens QEMU fails to load migration stream: qemu-kvm: ... 'virtio-mem-pci' is not a valid device model name (NB, while the example shows one particular device, the problem is paramount) This problem is easier to run into since we are slowly moving validation from qemu_command.c into said validation functions. The solution is simple: do the validation in all cases. And while it may happen that users would be unable to migrate/restore a guest due to a bug in our validator, spawning QEMU without validation is worse (especially when you consider that users can supply their own XMLs for migrate/restore operations - these were never validated). Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2048435 Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Technically, this is v2 of: https://listman.redhat.com/archives/libvir-list/2022-January/msg01307.html but since it implements completely different approach I've reset the counter. src/qemu/qemu_process.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index c13280c8f3..ea586e54c1 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -5399,11 +5399,7 @@ qemuProcessStartValidate(virQEMUDriver *driver, } - /* Checks below should not be executed when starting a qemu process for a - * VM that was running before (migration, snapshots, save). It's more - * important to start such VM than keep the configuration clean */ - if ((flags & VIR_QEMU_PROCESS_START_NEW) && - virDomainDefValidate(vm->def, 0, driver->xmlopt, qemuCaps) < 0) + if (virDomainDefValidate(vm->def, 0, driver->xmlopt, qemuCaps) < 0) return -1; if (qemuProcessStartValidateGraphics(vm) < 0) -- 2.34.1

2 1

[PATCH v3 0/3] qemu_tpm: Get swtpm pid without binary validation
by Vasiliy Ulyanov 04 Feb '22

04 Feb '22

[v1] https://listman.redhat.com/archives/libvir-list/2022-January/msg00008.html [v2] https://listman.redhat.com/archives/libvir-list/2022-January/msg00582.html As suggesed in the review comments: - dropped virFileGetLockOwner; - simplified lock validation by using VIR_AUTOCLOSE and just trying to lock the file; - introduced virPidFileReadPathIfLocked to preserve the existing behaviour of virPidFileReadPathIfAlive. Vasiliy Ulyanov (3): virpidfile: Add virPidFileReadPathIfLocked func qemu: tpm: Get swtpm pid without binary validation qemu: gpu: Get pid without binary validation src/libvirt_private.syms | 1 + src/qemu/qemu_tpm.c | 40 +++++++++++++++++++++------------- src/qemu/qemu_vhost_user_gpu.c | 11 +++++----- src/util/virpidfile.c | 34 +++++++++++++++++++++++++++++ src/util/virpidfile.h | 2 ++ 5 files changed, 67 insertions(+), 21 deletions(-) -- 2.34.1

3 9

[libvirt PATCH] conf: Fix documentation for virDomainDefGetShortName()
by Andrea Bolognani 03 Feb '22

03 Feb '22

The @unique argument didn't exist even when the function was introduced in a042275a396e, and the @vm argument was not renamed when the function was changed to take a virDomainDef* instead of a virDomainObj* in 7ed6934f3b92. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- src/conf/domain_conf.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 467f708350..d66496c9b7 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -30086,8 +30086,7 @@ virDomainDefHasMemballoon(const virDomainDef *def) /** * virDomainDefGetShortName: - * @vm: Machine for which to get a name - * @unique: Make sure the name is unique (use id as well) + * @def: domain definition * * Shorten domain name to avoid possible path length limitations. */ -- 2.34.1

2 1

[libvirt PATCH v8 0/3] Ignore EPERM on implicit clearing of VF VLAN ID
by Dmitrii Shcherbakov 03 Feb '22

03 Feb '22

SmartNIC DPUs may not expose some privileged eswitch operations to the hypervisor hosts. For example, this happens with Bluefield devices running in the ECPF (default) mode [1] for security reasons. While VF MAC address programming is possible via an RTM_SETLINK operation, trying to set a VLAN ID in the same operation may fail with EPERM. Specifically for the mlx5 driver this behavior was altered in the Linux kernel upstream [2] to avoid getting EPERM when trying to program VLAN 0. This may also get backported to older downstream kernels (e.g. [3]). However, Libvirt could potentially handle this case gracefully without needing a specific kernel version or depending on a specific driver fix. In the kernel a relevant call chain may look like do_setlink -> do_setvfinfo -> dev->netdev_ops->set_vf_vlan which calls a driver-specific function like [4] eventually. The equivalent ip link commands below provide an illustration: 1. This will work without an error: sudo ip link set enp130s0f0 vf 2 mac de:ad:be:ef:ca:fe 2. Setting (or clearing) a VLAN will fail with EPERM: sudo ip link set enp130s0f0 vf 2 vlan 0 RTNETLINK answers: Operation not permitted 3. This is what Libvirt attempts to do today when trying to clear a VF VLAN at the same time as programming a VF MAC: sudo ip link set enp130s0f0 vf 2 vlan 0 mac de:ad:be:ef:ca:fe RTNETLINK answers: Operation not permitted If setting an explicit VLAN ID results in an EPERM, clearing a VLAN (setting a VLAN ID to 0) can be handled gracefully by ignoring the EPERM error with the rationale being that if we cannot set this state in the first place, we cannot clear it either. Thus, virNetDevSetVfConfig is split into two distinct functions. If clearing a VLAN ID fails with EPERM when clearing is implicit, the error is simply ignored. For explicit clearing EPERM is still a fatal error. Both new functions rely virNetDevSendVfSetLinkRequest that implements common functionality related to formatting a request, sending it and handling error conditions and returns 0 or an error since in both cases the payload is either NLMSG_DONE (no error) or NLMSG_ERROR where an error message is needed by the caller to handle known cases appropriately. This function allows the conditional code to be unit tested. An alternative to this could be providing a higher level control plane mechanism that would provide metadata about a device being remotely managed in which case Libvirt would avoid trying to set or clear a VLAN ID. This would be more complicated since other software (like Nova in the OpenStack case) would have to annotate every guest device with an attribute indicating whether a device is remotely managed or not based on operator provided configuration so that Libvirt can act on this and avoid VLAN programming. https://gitlab.com/dmitriis/libvirt/-/pipelines/460293963 v8 change: * Rebased on top of the latest changes to Libvirt; * Added relevant upstream Linux and downstream kernel references to this cover letter. [1] https://docs.mellanox.com/display/BlueFieldSWv35111601/Modes+of+Operation#M… [2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… [3] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1957753 [4] https://github.com/torvalds/linux/blob/v5.15/drivers/net/ethernet/mellanox/… Dmitrii Shcherbakov (3): Set VF MAC and VLAN ID in two different operations Allow VF vlanid to be passed as a pointer Ignore EPERM on implicit clearing of VF VLAN ID NEWS.rst | 14 ++ src/hypervisor/virhostdev.c | 4 +- src/libvirt_private.syms | 7 + src/util/virnetdev.c | 256 +++++++++++++++++++++++++----------- src/util/virnetdevpriv.h | 44 +++++++ tests/virnetdevtest.c | 249 ++++++++++++++++++++++++++++++++++- 6 files changed, 496 insertions(+), 78 deletions(-) create mode 100644 src/util/virnetdevpriv.h -- 2.32.0

3 14

[libvirt PATCH] docs: Add example for isa-debugcon
by Andrea Bolognani 03 Feb '22

03 Feb '22

Most people will want to use isa-debugcon to obtain debug output for SeaBIOS / EDK II, so let's include a ready-made example for that scenario in our documentation. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- docs/formatdomain.rst | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 8fa5940469..583c195568 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -6395,6 +6395,14 @@ Serial port <source path='/dev/pts/3'/> <target port='0'/> </serial> +  + <serial type='pty'> + <target type='isa-debug'> + <model type='isa-debugcon'/> + </target> + <address type='isa' iobase='0x402'/> + </console> + </devices> ... -- 2.34.1

2 1