- Devel - libvirt List Archives

[PATCH RESEND 0/5] Introduce pstore device
by Michal Privoznik 26 Jul '24

26 Jul '24

This is a rebased version of: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/UAUK… which was a rebased version of: https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/VWR6… Let's hope no more resends will be needed. Michal Prívozník (5): qemu_capabilities: Introduce QEMU_CAPS_DEVICE_ACPI_ERST conf: Introduce pstore device qemu: Build cmd line for pstore device security: Set seclabels for pstore device NEWS: Document pstore device addition NEWS.rst | 7 + docs/formatdomain.rst | 32 ++++ src/ch/ch_domain.c | 1 + src/conf/domain_conf.c | 153 ++++++++++++++++++ src/conf/domain_conf.h | 19 +++ src/conf/domain_postparse.c | 1 + src/conf/domain_validate.c | 30 ++++ src/conf/schemas/domaincommon.rng | 25 +++ src/conf/virconftypes.h | 2 + src/hyperv/hyperv_driver.c | 1 + src/libvirt_private.syms | 2 + src/libxl/libxl_driver.c | 6 + src/lxc/lxc_driver.c | 6 + src/qemu/qemu_alias.c | 10 ++ src/qemu/qemu_capabilities.c | 2 + src/qemu/qemu_capabilities.h | 1 + src/qemu/qemu_command.c | 52 ++++++ src/qemu/qemu_domain.c | 3 + src/qemu/qemu_domain_address.c | 11 ++ src/qemu/qemu_driver.c | 3 + src/qemu/qemu_hotplug.c | 5 + src/qemu/qemu_validate.c | 26 +++ src/security/security_dac.c | 10 ++ src/security/security_selinux.c | 9 ++ src/security/virt-aa-helper.c | 4 + src/test/test_driver.c | 1 + .../caps_7.0.0_aarch64+hvf.xml | 1 + .../caps_7.0.0_aarch64.xml | 1 + .../qemucapabilitiesdata/caps_7.0.0_ppc64.xml | 1 + .../caps_7.0.0_x86_64.xml | 1 + .../qemucapabilitiesdata/caps_7.1.0_ppc64.xml | 1 + .../caps_7.1.0_x86_64.xml | 1 + tests/qemucapabilitiesdata/caps_7.2.0_ppc.xml | 1 + .../caps_7.2.0_x86_64+hvf.xml | 1 + .../caps_7.2.0_x86_64.xml | 1 + .../caps_8.0.0_x86_64.xml | 1 + .../caps_8.1.0_x86_64.xml | 1 + .../caps_8.2.0_aarch64.xml | 1 + .../caps_8.2.0_armv7l.xml | 1 + .../caps_8.2.0_loongarch64.xml | 1 + .../caps_8.2.0_x86_64.xml | 1 + .../caps_9.0.0_x86_64.xml | 1 + .../caps_9.1.0_x86_64.xml | 1 + .../pstore-acpi-erst.x86_64-latest.args | 38 +++++ .../pstore-acpi-erst.x86_64-latest.xml | 1 + tests/qemuxmlconfdata/pstore-acpi-erst.xml | 53 ++++++ tests/qemuxmlconftest.c | 1 + 47 files changed, 532 insertions(+) create mode 100644 tests/qemuxmlconfdata/pstore-acpi-erst.x86_64-latest.args create mode 120000 tests/qemuxmlconfdata/pstore-acpi-erst.x86_64-latest.xml create mode 100644 tests/qemuxmlconfdata/pstore-acpi-erst.xml -- 2.44.2

3 9

[PATCH 0/4] Rework qemu internal active snapshots to use QMP
by Nikolai Barybin 26 Jul '24

26 Jul '24

These patches make use of QMP recently added snapshot-save/delete commands and reaps HMP savevm/deletevm. The usage of HMP commands are highly discouraged by QEMU. Nikolai Barybin (4): qemu monitor: add snaphot-save/delete QMP commands qemu blockjob: add snapshot-save/delete job types qemu snapshot: use QMP snapshot-save/delete for internal snapshots qemu monitor: reap qemu_monitor_text po/POTFILES | 1 - po/libvirt.pot | 18 ---- src/qemu/meson.build | 1 - src/qemu/qemu_block.c | 2 + src/qemu/qemu_blockjob.c | 6 +- src/qemu/qemu_blockjob.h | 2 + src/qemu/qemu_domain.c | 4 + src/qemu/qemu_monitor.c | 23 +++-- src/qemu/qemu_monitor.h | 16 +++- src/qemu/qemu_monitor_json.c | 66 +++++++++++++++ src/qemu/qemu_monitor_json.h | 13 +++ src/qemu/qemu_monitor_text.c | 88 ------------------- src/qemu/qemu_monitor_text.h | 29 ------- src/qemu/qemu_snapshot.c | 158 ++++++++++++++++++++++++++++++++--- 14 files changed, 267 insertions(+), 160 deletions(-) delete mode 100644 src/qemu/qemu_monitor_text.c delete mode 100644 src/qemu/qemu_monitor_text.h -- 2.43.5

4 12

[PATCH-for-9.1] target/riscv: Remove the deprecated 'any' CPU type
by Philippe Mathieu-Daudé 26 Jul '24

26 Jul '24

The 'any' CPU is deprecated since commit f57d5f8004b ("target/riscv: deprecate the 'any' CPU type"). Users are better off using the default CPUs or the 'max' CPU. Signed-off-by: Philippe Mathieu-Daudé <philmd(a)linaro.org> --- docs/about/deprecated.rst | 13 ------------- docs/about/removed-features.rst | 8 ++++++++ target/riscv/cpu-qom.h | 1 - target/riscv/cpu.c | 28 ---------------------------- 4 files changed, 8 insertions(+), 42 deletions(-) diff --git a/docs/about/deprecated.rst b/docs/about/deprecated.rst index 88f0f037865..0ac49b15b44 100644 --- a/docs/about/deprecated.rst +++ b/docs/about/deprecated.rst @@ -347,19 +347,6 @@ QEMU's ``vhost`` feature, which would eliminate the high latency costs under which the 9p ``proxy`` backend currently suffers. However as of to date nobody has indicated plans for such kind of reimplementation unfortunately. -RISC-V 'any' CPU type ``-cpu any`` (since 8.2) -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - -The 'any' CPU type was introduced back in 2018 and has been around since the -initial RISC-V QEMU port. Its usage has always been unclear: users don't know -what to expect from a CPU called 'any', and in fact the CPU does not do anything -special that isn't already done by the default CPUs rv32/rv64. - -After the introduction of the 'max' CPU type, RISC-V now has a good coverage -of generic CPUs: rv32 and rv64 as default CPUs and 'max' as a feature complete -CPU for both 32 and 64 bit builds. Users are then discouraged to use the 'any' -CPU type starting in 8.2. - RISC-V CPU properties which start with capital 'Z' (since 8.2) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ diff --git a/docs/about/removed-features.rst b/docs/about/removed-features.rst index fc7b28e6373..f3e9474a73e 100644 --- a/docs/about/removed-features.rst +++ b/docs/about/removed-features.rst @@ -850,6 +850,14 @@ The RISC-V no MMU cpus have been removed. The two CPUs: ``rv32imacu-nommu`` and ``rv64imacu-nommu`` can no longer be used. Instead the MMU status can be specified via the CPU ``mmu`` option when using the ``rv32`` or ``rv64`` CPUs. +RISC-V 'any' CPU type ``-cpu any`` (removed in 9.1) +''''''''''''''''''''''''''''''''''''''''''''''''''' + +The 'any' CPU type was introduced back in 2018 and was around since the +initial RISC-V QEMU port. Its usage was always been unclear: users don't know +what to expect from a CPU called 'any', and in fact the CPU does not do anything +special that isn't already done by the default CPUs rv32/rv64. + ``compat`` property of server class POWER CPUs (removed in 6.0) ''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''' diff --git a/target/riscv/cpu-qom.h b/target/riscv/cpu-qom.h index 3670cfe6d9a..4464c0fd7a3 100644 --- a/target/riscv/cpu-qom.h +++ b/target/riscv/cpu-qom.h @@ -29,7 +29,6 @@ #define RISCV_CPU_TYPE_SUFFIX "-" TYPE_RISCV_CPU #define RISCV_CPU_TYPE_NAME(name) (name RISCV_CPU_TYPE_SUFFIX) -#define TYPE_RISCV_CPU_ANY RISCV_CPU_TYPE_NAME("any") #define TYPE_RISCV_CPU_MAX RISCV_CPU_TYPE_NAME("max") #define TYPE_RISCV_CPU_BASE32 RISCV_CPU_TYPE_NAME("rv32") #define TYPE_RISCV_CPU_BASE64 RISCV_CPU_TYPE_NAME("rv64") diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c index a90808a3bac..4bda754b013 100644 --- a/target/riscv/cpu.c +++ b/target/riscv/cpu.c @@ -438,27 +438,6 @@ static void set_satp_mode_default_map(RISCVCPU *cpu) } #endif -static void riscv_any_cpu_init(Object *obj) -{ - RISCVCPU *cpu = RISCV_CPU(obj); - CPURISCVState *env = &cpu->env; - riscv_cpu_set_misa_ext(env, RVI | RVM | RVA | RVF | RVD | RVC | RVU); - -#ifndef CONFIG_USER_ONLY - set_satp_mode_max_supported(RISCV_CPU(obj), - riscv_cpu_mxl(&RISCV_CPU(obj)->env) == MXL_RV32 ? - VM_1_10_SV32 : VM_1_10_SV57); -#endif - - env->priv_ver = PRIV_VERSION_LATEST; - - /* inherited from parent obj via riscv_cpu_init() */ - cpu->cfg.ext_zifencei = true; - cpu->cfg.ext_zicsr = true; - cpu->cfg.mmu = true; - cpu->cfg.pmp = true; -} - static void riscv_max_cpu_init(Object *obj) { RISCVCPU *cpu = RISCV_CPU(obj); @@ -1161,11 +1140,6 @@ static void riscv_cpu_realize(DeviceState *dev, Error **errp) RISCVCPUClass *mcc = RISCV_CPU_GET_CLASS(dev); Error *local_err = NULL; - if (object_dynamic_cast(OBJECT(dev), TYPE_RISCV_CPU_ANY) != NULL) { - warn_report("The 'any' CPU is deprecated and will be " - "removed in the future."); - } - cpu_exec_realizefn(cs, &local_err); if (local_err != NULL) { error_propagate(errp, local_err); @@ -2952,7 +2926,6 @@ static const TypeInfo riscv_cpu_type_infos[] = { .abstract = true, }, #if defined(TARGET_RISCV32) - DEFINE_DYNAMIC_CPU(TYPE_RISCV_CPU_ANY, MXL_RV32, riscv_any_cpu_init), DEFINE_DYNAMIC_CPU(TYPE_RISCV_CPU_MAX, MXL_RV32, riscv_max_cpu_init), DEFINE_DYNAMIC_CPU(TYPE_RISCV_CPU_BASE32, MXL_RV32, rv32_base_cpu_init), DEFINE_VENDOR_CPU(TYPE_RISCV_CPU_IBEX, MXL_RV32, rv32_ibex_cpu_init), @@ -2962,7 +2935,6 @@ static const TypeInfo riscv_cpu_type_infos[] = { DEFINE_BARE_CPU(TYPE_RISCV_CPU_RV32I, MXL_RV32, rv32i_bare_cpu_init), DEFINE_BARE_CPU(TYPE_RISCV_CPU_RV32E, MXL_RV32, rv32e_bare_cpu_init), #elif defined(TARGET_RISCV64) - DEFINE_DYNAMIC_CPU(TYPE_RISCV_CPU_ANY, MXL_RV64, riscv_any_cpu_init), DEFINE_DYNAMIC_CPU(TYPE_RISCV_CPU_MAX, MXL_RV64, riscv_max_cpu_init), DEFINE_DYNAMIC_CPU(TYPE_RISCV_CPU_BASE64, MXL_RV64, rv64_base_cpu_init), DEFINE_VENDOR_CPU(TYPE_RISCV_CPU_SIFIVE_E51, MXL_RV64, rv64_sifive_e_cpu_init), -- 2.45.2

3 3

[PATCH 0/5] virsysinfo: Be more forgiving when decoding OEM strings
by Michal Privoznik 25 Jul '24

25 Jul '24

See 4/5 for explanation. Michal Prívozník (5): vircommand: Initialize dryRunStatus to portable EXIT_SUCCESS instead of 0 virsysinfo: Trim newline when decoding OEM strings tests: Add HPE Apollo test case to sysinfotest virsysinfo: Be more forgiving when decoding OEM strings virsysinfo: Calculate OEM string index better src/util/vircommand.c | 2 +- src/util/virsysinfo.c | 36 +++- .../aarch64-hpe-apollosysinfo.data | 162 ++++++++++++++++++ .../aarch64-hpe-apollosysinfo.expect | 88 ++++++++++ tests/sysinfotest.c | 9 +- 5 files changed, 285 insertions(+), 12 deletions(-) create mode 100644 tests/sysinfodata/aarch64-hpe-apollosysinfo.data create mode 100644 tests/sysinfodata/aarch64-hpe-apollosysinfo.expect -- 2.44.2

2 7

[PATCH] vmx: Ensure unique disk targets when parsing
by Adam Julis 24 Jul '24

24 Jul '24

Disk targets were generated in virVMXParseConfig() with virVMXGenerateDiskTarget(). It works on combination of controller, fix offset, unit and prefix. While SCSI and SATA have same prefix "sd", function virVMXGenerateDiskTarget() could returned in some cases same targets. In this patch, after loaded SCSI and SATA disks to the def, it checks if in array exists any SATA disks after SCSI. If not, nothing happened. If yes, targets of all SATA disks are changed. Disk target is calculated as: last SCSI target value + n, where n is position of disk in array after last SCSI (1,2,..). Because assigned addresses of disks are generated from their indexes, for every changed SATA disk is called virDomainDiskDefAssignAddress() with the updated value. The corresponding tests have been modified to match the index changes. Signed-off-by: Adam Julis <ajulis(a)redhat.com> --- src/vmx/vmx.c | 32 ++++++++++++++++++++++++ tests/vmx2xmldata/esx-in-the-wild-12.xml | 4 +-- tests/vmx2xmldata/esx-in-the-wild-8.xml | 4 +-- 3 files changed, 36 insertions(+), 4 deletions(-) diff --git a/src/vmx/vmx.c b/src/vmx/vmx.c index 227744d062..9fdf0f1cd4 100644 --- a/src/vmx/vmx.c +++ b/src/vmx/vmx.c @@ -1389,6 +1389,9 @@ virVMXParseConfig(virVMXContext *ctx, bool smbios_reflecthost = false; int controller; int bus; + int ndisk; + int last_scsi; + int offset = -1; int port; bool present; int scsi_virtualDev[4] = { -1, -1, -1, -1 }; @@ -1805,6 +1808,35 @@ virVMXParseConfig(virVMXContext *ctx, } } + /* now disks contain only SCSI and SATA, SATA could have same index (dst) as SCSI + * find last SCSI index in array and use it as offset for all SATA indexes + * (overwrite old values) + * finally, regenerate correct addresses, while it depends on the index */ + for (ndisk = 0; ndisk < def->ndisks; ndisk++) { + virDomainDiskDef *dsc = def->disks[ndisk]; + + if (dsc->bus == VIR_DOMAIN_DISK_BUS_SCSI) { + offset = virDiskNameToIndex(dsc->dst); + last_scsi = ndisk; + continue; + } + + if (offset > -1) { + VIR_FREE(def->disks[ndisk]->dst); + def->disks[ndisk]->dst = virIndexToDiskName(offset + ndisk - last_scsi, "sd"); + + if (virDomainDiskDefAssignAddress(NULL, def->disks[ndisk], def) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Could not assign address to disk '%1$s'"), + virDomainDiskGetSource(dsc)); + goto cleanup; + } + } + + else + break; + } + /* def:disks (ide) */ for (bus = 0; bus < 2; ++bus) { for (unit = 0; unit < 2; ++unit) { diff --git a/tests/vmx2xmldata/esx-in-the-wild-12.xml b/tests/vmx2xmldata/esx-in-the-wild-12.xml index 42184501d0..a7730845ee 100644 --- a/tests/vmx2xmldata/esx-in-the-wild-12.xml +++ b/tests/vmx2xmldata/esx-in-the-wild-12.xml @@ -21,9 +21,9 @@ <address type='drive' controller='0' bus='0' target='0' unit='0'/> </disk> <disk type='file' device='cdrom'> - <target dev='sda' bus='sata'/> + <target dev='sdb' bus='sata'/> <readonly/> - <address type='drive' controller='0' bus='0' target='0' unit='0'/> + <address type='drive' controller='0' bus='0' target='0' unit='1'/> </disk> <controller type='scsi' index='0' model='vmpvscsi'/> <controller type='sata' index='0'/> diff --git a/tests/vmx2xmldata/esx-in-the-wild-8.xml b/tests/vmx2xmldata/esx-in-the-wild-8.xml index 47d22ced2a..d5356bda34 100644 --- a/tests/vmx2xmldata/esx-in-the-wild-8.xml +++ b/tests/vmx2xmldata/esx-in-the-wild-8.xml @@ -36,9 +36,9 @@ </disk> <disk type='file' device='cdrom'> <source file='[692eb778-2d4937fe] CentOS-4.7.ServerCD-x86_64.iso'/> - <target dev='sda' bus='sata'/> + <target dev='sdd' bus='sata'/> <readonly/> - <address type='drive' controller='0' bus='0' target='0' unit='0'/> + <address type='drive' controller='0' bus='0' target='0' unit='3'/> </disk> <controller type='scsi' index='0' model='vmpvscsi'/> <controller type='sata' index='0'/> -- 2.45.2

2 1

[libvirt PATCH] qemu: virtiofs: cache: use 'never' instead of 'none'
by Ján Tomko 24 Jul '24

24 Jul '24

The new option style renamed one of the cache modes. Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- src/qemu/qemu_virtiofs.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_virtiofs.c b/src/qemu/qemu_virtiofs.c index 703f1226a2..db987ddfe1 100644 --- a/src/qemu/qemu_virtiofs.c +++ b/src/qemu/qemu_virtiofs.c @@ -147,10 +147,19 @@ qemuVirtioFSBuildCommandLine(virQEMUDriverConfig *cfg, virCommandAddArg(cmd, "--shared-dir"); virCommandAddArg(cmd, fs->src->path); - if (fs->cache) { + switch (fs->cache) { + case VIR_DOMAIN_FS_CACHE_MODE_DEFAULT: + break; + case VIR_DOMAIN_FS_CACHE_MODE_NONE: + virCommandAddArg(cmd, "--cache"); + virCommandAddArg(cmd, "never"); + break; + case VIR_DOMAIN_FS_CACHE_MODE_ALWAYS: virCommandAddArg(cmd, "--cache"); virCommandAddArg(cmd, virDomainFSCacheModeTypeToString(fs->cache)); + break; } + if (fs->sandbox) { virCommandAddArg(cmd, "--sandbox"); virCommandAddArg(cmd, virDomainFSSandboxModeTypeToString(fs->sandbox)); -- 2.45.2

2 1

[PATCH] meson: correct git detection
by John Levon 24 Jul '24

24 Jul '24

The current "building from git" test uses "test -d .git"; however, that doesn't work when libvirt is used as a submodule, as in that case .git is a normal file. Use "test -e .git" instead. Signed-off-by: John Levon <john.levon(a)nutanix.com> --- meson.build | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/meson.build b/meson.build index 3ab8e57e4a..231470e2ee 100644 --- a/meson.build +++ b/meson.build @@ -17,7 +17,7 @@ endif # figure out if we are building from git -git = run_command('test', '-d', '.git', check: false).returncode() == 0 +git = run_command('test', '-e', '.git', check: false).returncode() == 0 if git and not get_option('no_git') run_command('git', 'submodule', 'update', '--init', check: true) -- 2.34.1

2 1

[PATCH] virtiofs: rename member to 'openfiles' for clarity
by Adam Julis 24 Jul '24

24 Jul '24

New element 'openfiles' had confusing name. Since the patch with this new element wasn't propagate yet, old name ('rlimit_nofile') was changed. ... <binary> <openfiles max='122333'/> </binary> ... Signed-off-by: Adam Julis <ajulis(a)redhat.com> --- docs/formatdomain.rst | 6 +++--- src/conf/domain_conf.c | 14 +++++++------- src/conf/domain_conf.h | 2 +- src/conf/schemas/domaincommon.rng | 4 ++-- src/qemu/qemu_virtiofs.c | 4 ++-- ... vhost-user-fs-fd-openfiles.x86_64-latest.args} | 0 .../vhost-user-fs-fd-openfiles.x86_64-latest.xml | 1 + ...d-rlimit.xml => vhost-user-fs-fd-openfiles.xml} | 2 +- .../vhost-user-fs-fd-rlimit.x86_64-latest.xml | 1 - tests/qemuxmlconftest.c | 2 +- 10 files changed, 18 insertions(+), 18 deletions(-) rename tests/qemuxmlconfdata/{vhost-user-fs-fd-rlimit.x86_64-latest.args => vhost-user-fs-fd-openfiles.x86_64-latest.args} (100%) create mode 120000 tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.x86_64-latest.xml rename tests/qemuxmlconfdata/{vhost-user-fs-fd-rlimit.xml => vhost-user-fs-fd-openfiles.xml} (97%) delete mode 120000 tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.x86_64-latest.xml diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 10584dfe83..c2a6d0b910 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -3734,10 +3734,10 @@ A directory on the host that can be accessed directly from the guest. The thread pool helps increase the number of requests in flight when used with storage that has a higher latency. However, it has an overhead, and so for fast, low latency filesystems, it may be best to turn it off. ( :since:`Since 8.5.0` ) - Element ``rlimit_profile`` accepts one attribute ``size`` which defines the + Element ``openfiles`` accepts one attribute ``max`` which defines the maximum number of file descriptors. Non-positive values are forbidden. - Although numbers greater than 1M are allowed, the virtiofsd documentation - states that in this case its set by virtiofsd to the 1M. ( :since:`Since 10.6.0` ) + The upper bound on the number of open files is implementation defined. + ( :since:`Since 10.6.0` ) ``source`` The resource on the host that is being accessed in the guest. The ``name`` attribute must be used with ``type='template'``, and the ``dir`` attribute diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 6733857a3a..d1c59f7a91 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -8868,7 +8868,7 @@ virDomainFSDefParseXML(virDomainXMLOption *xmlopt, g_autofree char *queue_size = virXPathString("string(./driver/@queue)", ctxt); g_autofree char *binary = virXPathString("string(./binary/@path)", ctxt); g_autofree char *thread_pool_size = virXPathString("string(./binary/thread_pool/@size)", ctxt); - g_autofree char *rlimit_nofile = virXPathString("string(./binary/rlimit_nofile/@size)", ctxt); + g_autofree char *openfiles = virXPathString("string(./binary/openfiles/@max)", ctxt); xmlNodePtr binary_node = virXPathNode("./binary", ctxt); xmlNodePtr binary_lock_node = virXPathNode("./binary/lock", ctxt); xmlNodePtr binary_cache_node = virXPathNode("./binary/cache", ctxt); @@ -8892,11 +8892,11 @@ virDomainFSDefParseXML(virDomainXMLOption *xmlopt, goto error; } - if (rlimit_nofile && - virStrToLong_ull(rlimit_nofile, NULL, 10, &def->rlimit_nofile) < 0) { + if (openfiles && + virStrToLong_ull(openfiles, NULL, 10, &def->openfiles) < 0) { virReportError(VIR_ERR_XML_ERROR, - _("cannot parse rlimit_nofile '%1$s' for virtiofs"), - rlimit_nofile); + _("cannot parse openfiles '%1$s' for virtiofs"), + openfiles); goto error; } @@ -23424,8 +23424,8 @@ virDomainFSDefFormat(virBuffer *buf, if (def->thread_pool_size >= 0) virBufferAsprintf(&binaryBuf, "<thread_pool size='%d'/>\n", def->thread_pool_size); - if (def->rlimit_nofile > 0) - virBufferAsprintf(&binaryBuf, "<rlimit_nofile size='%lld'/>\n", def->rlimit_nofile); + if (def->openfiles > 0) + virBufferAsprintf(&binaryBuf, "<openfiles max='%lld'/>\n", def->openfiles); } diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 8283493dfc..6972f6ae9b 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -890,7 +890,7 @@ struct _virDomainFSDef { bool symlinksResolved; char *binary; unsigned long long queue_size; - unsigned long long rlimit_nofile; + unsigned long long openfiles; virTristateSwitch xattr; virDomainFSCacheMode cache; virTristateSwitch posix_lock; diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index ab5374d5f0..8dae6416e9 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -3381,9 +3381,9 @@ </element> </optional> <optional> - <element name="rlimit_nofile"> + <element name="openfiles"> <optional> - <attribute name="size"> + <attribute name="max"> <data type="positiveInteger"/> </attribute> </optional> diff --git a/src/qemu/qemu_virtiofs.c b/src/qemu/qemu_virtiofs.c index 703f1226a2..a8f2416273 100644 --- a/src/qemu/qemu_virtiofs.c +++ b/src/qemu/qemu_virtiofs.c @@ -194,8 +194,8 @@ qemuVirtioFSBuildCommandLine(virQEMUDriverConfig *cfg, if (fs->thread_pool_size >= 0) virCommandAddArgFormat(cmd, "--thread-pool-size=%i", fs->thread_pool_size); - if (fs->rlimit_nofile > 0) - virCommandAddArgFormat(cmd, "--rlimit-nofile=%llu", fs->rlimit_nofile); + if (fs->openfiles > 0) + virCommandAddArgFormat(cmd, "--rlimit-nofile=%llu", fs->openfiles); if (cfg->virtiofsdDebug) { if (virBitmapIsBitSet(fs->caps, QEMU_VHOST_USER_FS_FEATURE_SEPARATE_OPTIONS)) diff --git a/tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.x86_64-latest.args b/tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.x86_64-latest.args similarity index 100% rename from tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.x86_64-latest.args rename to tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.x86_64-latest.args diff --git a/tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.x86_64-latest.xml b/tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.x86_64-latest.xml new file mode 120000 index 0000000000..74761b0d0f --- /dev/null +++ b/tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.x86_64-latest.xml @@ -0,0 +1 @@ +vhost-user-fs-fd-openfiles.xml \ No newline at end of file diff --git a/tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.xml b/tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.xml similarity index 97% rename from tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.xml rename to tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.xml index 2983d3f275..3e925a8c8b 100644 --- a/tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.xml +++ b/tests/qemuxmlconfdata/vhost-user-fs-fd-openfiles.xml @@ -32,7 +32,7 @@ <cache mode='always'/> <sandbox mode='chroot'/> <thread_pool size='16'/> - <rlimit_nofile size='122333'/> + <openfiles max='122333'/> </binary> <idmap> <uid start='0' target='100000' count='65535'/> diff --git a/tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.x86_64-latest.xml b/tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.x86_64-latest.xml deleted file mode 120000 index 68c4e8a482..0000000000 --- a/tests/qemuxmlconfdata/vhost-user-fs-fd-rlimit.x86_64-latest.xml +++ /dev/null @@ -1 +0,0 @@ -vhost-user-fs-fd-rlimit.xml \ No newline at end of file diff --git a/tests/qemuxmlconftest.c b/tests/qemuxmlconftest.c index cc984440ea..bf88bd2f8d 100644 --- a/tests/qemuxmlconftest.c +++ b/tests/qemuxmlconftest.c @@ -2816,7 +2816,7 @@ mymain(void) DO_TEST_CAPS_ARCH_LATEST("launch-security-s390-pv", "s390x"); DO_TEST_CAPS_LATEST("vhost-user-fs-fd-memory"); - DO_TEST_CAPS_LATEST("vhost-user-fs-fd-rlimit"); + DO_TEST_CAPS_LATEST("vhost-user-fs-fd-openfiles"); DO_TEST_CAPS_LATEST("vhost-user-fs-hugepages"); DO_TEST_CAPS_LATEST_PARSE_ERROR("vhost-user-fs-readonly"); -- 2.45.2

2 1

[PATCH] qemu_hotplug: Do not allow absent values in rom settings
by Kristina Hanicova 24 Jul '24

24 Jul '24

If there are absent values in an already existing element specifying rom settings, we simply use the old ones. This behaviour is not desired, as users might think that deleting the element from XML would delete the setting (because the hotplug succeeds) - which does not happen. Because of that, we should not accept an interface without elements that cannot be changed. Therefore, we should not allow absent values for already existing rom setting during hotplug. Resolves: https://issues.redhat.com/browse/RHEL-7109 Signed-off-by: Kristina Hanicova <khanicov(a)redhat.com> --- src/qemu/qemu_hotplug.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 1f4620d833..7f158ddcd5 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -3838,16 +3838,12 @@ qemuDomainChangeNet(virQEMUDriver *driver, /* device alias is checked already in virDomainDefCompatibleDevice */ - if (newdev->info.rombar == VIR_TRISTATE_SWITCH_ABSENT) - newdev->info.rombar = olddev->info.rombar; if (olddev->info.rombar != newdev->info.rombar) { virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", _("cannot modify network device rom bar setting")); goto cleanup; } - if (!newdev->info.romfile) - newdev->info.romfile = g_strdup(olddev->info.romfile); if (STRNEQ_NULLABLE(olddev->info.romfile, newdev->info.romfile)) { virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", _("cannot modify network rom file")); @@ -3860,8 +3856,6 @@ qemuDomainChangeNet(virQEMUDriver *driver, goto cleanup; } - if (newdev->info.romenabled == VIR_TRISTATE_BOOL_ABSENT) - newdev->info.romenabled = olddev->info.romenabled; if (olddev->info.romenabled != newdev->info.romenabled) { virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", _("cannot modify network device rom enabled setting")); -- 2.45.1

2 1

[PATCH v3] qemu: add a monitor to /proc/$pid when killing times out
by Boris Fiuczynski 24 Jul '24

24 Jul '24

In cases when a QEMU process takes longer than the time sigterm and sigkill are issued to kill the process do not simply fail and leave the VM in state VIR_DOMAIN_SHUTDOWN until the daemon stops. Instead set up an fd on /proc/$pid and get notified when the QEMU process finally has terminated to cleanup the VM state. Resolves: https://issues.redhat.com/browse/RHEL-28819 Signed-off-by: Boris Fiuczynski <fiuczy(a)linux.ibm.com> --- src/qemu/qemu_domain.c | 8 +++ src/qemu/qemu_domain.h | 2 + src/qemu/qemu_driver.c | 18 ++++++ src/qemu/qemu_process.c | 124 ++++++++++++++++++++++++++++++++++++++-- src/qemu/qemu_process.h | 1 + 5 files changed, 148 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 2134b11038..8147ff02fd 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -1889,6 +1889,11 @@ qemuDomainObjPrivateFree(void *data) virChrdevFree(priv->devs); + if (priv->pidMonitored >= 0) { + virEventRemoveHandle(priv->pidMonitored); + priv->pidMonitored = -1; + } + /* This should never be non-NULL if we get here, but just in case... */ if (priv->mon) { VIR_ERROR(_("Unexpected QEMU monitor still active during domain deletion")); @@ -1934,6 +1939,8 @@ qemuDomainObjPrivateAlloc(void *opaque) priv->blockjobs = virHashNew(virObjectUnref); priv->fds = virHashNew(g_object_unref); + priv->pidMonitored = -1; + /* agent commands block by default, user can choose different behavior */ priv->agentTimeout = VIR_DOMAIN_AGENT_RESPONSE_TIMEOUT_BLOCK; priv->migMaxBandwidth = QEMU_DOMAIN_MIG_BANDWIDTH_MAX; @@ -11680,6 +11687,7 @@ qemuProcessEventFree(struct qemuProcessEvent *event) case QEMU_PROCESS_EVENT_RESET: case QEMU_PROCESS_EVENT_NBDKIT_EXITED: case QEMU_PROCESS_EVENT_MONITOR_EOF: + case QEMU_PROCESS_EVENT_SHUTDOWN_COMPLETED: case QEMU_PROCESS_EVENT_LAST: break; } diff --git a/src/qemu/qemu_domain.h b/src/qemu/qemu_domain.h index d777559119..a5092dd7f0 100644 --- a/src/qemu/qemu_domain.h +++ b/src/qemu/qemu_domain.h @@ -119,6 +119,7 @@ struct _qemuDomainObjPrivate { bool beingDestroyed; char *pidfile; + int pidMonitored; virDomainPCIAddressSet *pciaddrs; virDomainUSBAddressSet *usbaddrs; @@ -469,6 +470,7 @@ typedef enum { QEMU_PROCESS_EVENT_UNATTENDED_MIGRATION, QEMU_PROCESS_EVENT_RESET, QEMU_PROCESS_EVENT_NBDKIT_EXITED, + QEMU_PROCESS_EVENT_SHUTDOWN_COMPLETED, QEMU_PROCESS_EVENT_LAST } qemuProcessEventType; diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 9f3013e231..6b1e4084f6 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -4041,6 +4041,21 @@ processNbdkitExitedEvent(virDomainObj *vm, } +static void +processShutdownCompletedEvent(virQEMUDriver *driver, + virDomainObj *vm) +{ + if (virDomainObjBeginJob(vm, VIR_JOB_MODIFY) < 0) + return; + + if (virDomainObjIsActive(vm)) + qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_UNKNOWN, + VIR_ASYNC_JOB_NONE, 0); + + virDomainObjEndJob(vm); +} + + static void qemuProcessEventHandler(void *data, void *opaque) { struct qemuProcessEvent *processEvent = data; @@ -4101,6 +4116,9 @@ static void qemuProcessEventHandler(void *data, void *opaque) case QEMU_PROCESS_EVENT_NBDKIT_EXITED: processNbdkitExitedEvent(vm, processEvent->data); break; + case QEMU_PROCESS_EVENT_SHUTDOWN_COMPLETED: + processShutdownCompletedEvent(driver, vm); + break; case QEMU_PROCESS_EVENT_LAST: break; } diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 25dfd04272..c6f7d34168 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -25,6 +25,9 @@ #include <unistd.h> #include <signal.h> #include <sys/stat.h> +#if WITH_SYS_SYSCALL_H +# include <sys/syscall.h> +#endif #if defined(__linux__) # include <linux/capability.h> #elif defined(__FreeBSD__) @@ -8387,9 +8390,114 @@ qemuProcessCreatePretendCmdBuild(virDomainObj *vm, } +#if WITH_SYS_SYSCALL_H && defined(SYS_pidfd_open) +typedef struct { + virDomainObj *vm; + int pidfd; +} qemuProcessInShutdownEventData; + + +static qemuProcessInShutdownEventData* +qemuProcessInShutdownEventDataNew(virDomainObj *vm, int pidfd) +{ + qemuProcessInShutdownEventData *d = g_new(qemuProcessInShutdownEventData, 1); + d->vm = virObjectRef(vm); + d->pidfd = pidfd; + return d; +} + + +static void +qemuProcessInShutdownEventDataFree(qemuProcessInShutdownEventData *d) +{ + virObjectUnref(d->vm); + VIR_FORCE_CLOSE(d->pidfd); + g_free(d); +} + + +static void +qemuProcessInShutdownPidfdCb(int watch, + int fd, + int events G_GNUC_UNUSED, + void *opaque) +{ + qemuProcessInShutdownEventData *data = opaque; + virDomainObj *vm = data->vm; + + VIR_DEBUG("vm=%p name=%s pid=%lld fd=%d", + vm, vm->def->name, (long long)vm->pid, fd); + + virEventRemoveHandle(watch); + + virObjectLock(vm); + + VIR_INFO("QEMU process %lld finally completed termination", + (long long)vm->pid); + + QEMU_DOMAIN_PRIVATE(vm)->pidMonitored = -1; + qemuProcessEventSubmit(vm, QEMU_PROCESS_EVENT_SHUTDOWN_COMPLETED, + 0, 0, NULL); + + virObjectUnlock(vm); +} +#endif /* WITH_SYS_SYSCALL_H && defined(SYS_pidfd_open) */ + + +static int +qemuProcessInShutdownStartMonitor(virDomainObj *vm) +{ +#if WITH_SYS_SYSCALL_H && defined(SYS_pidfd_open) + qemuDomainObjPrivate *priv = vm->privateData; + qemuProcessInShutdownEventData *data; + int pidfd; + int ret = -1; + + VIR_DEBUG("vm=%p name=%s pid=%lld pidMonitored=%d", + vm, vm->def->name, (long long)vm->pid, + priv->pidMonitored); + + if (priv->pidMonitored >= 0) { + VIR_DEBUG("Monitoring qemu in-shutdown process %i already set up", vm->pid); + goto cleanup; + } + + pidfd = syscall(SYS_pidfd_open, vm->pid, 0); + if (pidfd < 0) { + if (errno == ESRCH) /* process has already terminated */ + ret = 1; + goto cleanup; + } + + data = qemuProcessInShutdownEventDataNew(vm, pidfd); + if ((priv->pidMonitored = virEventAddHandle(pidfd, + VIR_EVENT_HANDLE_READABLE, + qemuProcessInShutdownPidfdCb, + data, + (virFreeCallback)qemuProcessInShutdownEventDataFree)) < 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("failed to monitor qemu in-shutdown process %1$i"), + vm->pid); + qemuProcessInShutdownEventDataFree(data); + goto cleanup; + } + VIR_DEBUG("Monitoring qemu in-shutdown process %i for termination", vm->pid); + ret = 0; + + cleanup: + return ret; +#else /* !WITH_SYS_SYSCALL_H || !defined(SYS_pidfd_open) */ + VIR_DEBUG("Monitoring qemu process %i not implemented", vm->pid); + return -1; +#endif /* !WITH_SYS_SYSCALL_H || !defined(SYS_pidfd_open) */ +} + + int qemuProcessKill(virDomainObj *vm, unsigned int flags) { + int ret = -1; + VIR_DEBUG("vm=%p name=%s pid=%lld flags=0x%x", vm, vm->def->name, (long long)vm->pid, flags); @@ -8410,10 +8518,16 @@ qemuProcessKill(virDomainObj *vm, unsigned int flags) /* Request an extra delay of two seconds per current nhostdevs * to be safe against stalls by the kernel freeing up the resources */ - return virProcessKillPainfullyDelay(vm->pid, - !!(flags & VIR_QEMU_PROCESS_KILL_FORCE), - vm->def->nhostdevs * 2, - false); + ret = virProcessKillPainfullyDelay(vm->pid, + !!(flags & VIR_QEMU_PROCESS_KILL_FORCE), + vm->def->nhostdevs * 2, + false); + + if (ret < 0 && (flags & VIR_QEMU_PROCESS_KILL_MONITOR_ON_ERROR)) + if (qemuProcessInShutdownStartMonitor(vm) == 1) + ret = 0; /* process termination detected */ + + return ret; } @@ -8438,7 +8552,7 @@ qemuProcessBeginStopJob(virDomainObj *vm, * cleared inside qemuProcessStop */ priv->beingDestroyed = true; - if (qemuProcessKill(vm, killFlags) < 0) + if (qemuProcessKill(vm, killFlags|VIR_QEMU_PROCESS_KILL_MONITOR_ON_ERROR) < 0) goto error; /* Wake up anything waiting on domain condition */ diff --git a/src/qemu/qemu_process.h b/src/qemu/qemu_process.h index cb67bfcd2d..2324aeb7bd 100644 --- a/src/qemu/qemu_process.h +++ b/src/qemu/qemu_process.h @@ -180,6 +180,7 @@ typedef enum { VIR_QEMU_PROCESS_KILL_FORCE = 1 << 0, VIR_QEMU_PROCESS_KILL_NOWAIT = 1 << 1, VIR_QEMU_PROCESS_KILL_NOCHECK = 1 << 2, /* bypass the running vm check */ + VIR_QEMU_PROCESS_KILL_MONITOR_ON_ERROR = 1 << 3, /* on error enable process monitor */ } virQemuProcessKillMode; int qemuProcessKill(virDomainObj *vm, unsigned int flags); -- 2.45.0

3 5