The Memory Tagging Extensions are hardware acceleration present in some ARM processors that allow memory error detection [1]. Introduce a domain XML knob that turns them on or off. 1: https://www.arm.com/blogs/blueprint/memory-safety-arm-memory-tagging-exte... Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com&gt; --- docs/formatdomain.rst | 7 +++++++ src/conf/domain_conf.c | 6 +++++- src/conf/domain_conf.h | 1 + src/conf/schemas/domaincommon.rng | 5 +++++ src/qemu/qemu_validate.c | 1 + tests/qemuxml2argvdata/aarch64-gic-v3.xml | 1 + tests/qemuxml2xmloutdata/aarch64-gic-v3.aarch64-latest.xml | 1 + 7 files changed, 21 insertions(+), 1 deletion(-) diff --git a/docs/formatdomain.rst b/docs/formatdomain.rst index 99383e725c..1f52f58d37 100644 --- a/docs/formatdomain.rst +++ b/docs/formatdomain.rst @@ -2000,6 +2000,7 @@ Hypervisors may allow certain CPU / machine features to be toggled on/off. <tcg> <tb-cache unit='MiB'>128</tb-cache> </tcg> + <mte state='on'/> </features> ... @@ -2230,6 +2231,12 @@ are: tb-cache The size of translation block cache size an integer (a multiple of MiB) :since:`8.0.0` =========== ============================================== =================================================== ============== +``mte`` + Configure Memory Tagging Extensions for ARM guests. Possible values for the + ``state`` attribute are ``on`` and ``off``. If the attribute is not + defined, the hypervisor default will be used. :since:`Since 9.4.0` (QEMU/KVM + only) + Time keeping ------------ diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 6a864a8db9..047a4c97bf 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -181,6 +181,7 @@ VIR_ENUM_IMPL(virDomainFeature, "sbbc", "ibs", "tcg", + "mte", ); VIR_ENUM_IMPL(virDomainCapabilitiesPolicy, @@ -16645,7 +16646,8 @@ virDomainFeaturesDefParse(virDomainDef *def, case VIR_DOMAIN_FEATURE_HTM: case VIR_DOMAIN_FEATURE_NESTED_HV: - case VIR_DOMAIN_FEATURE_CCF_ASSIST: { + case VIR_DOMAIN_FEATURE_CCF_ASSIST: + case VIR_DOMAIN_FEATURE_MTE: { virTristateSwitch state; if (virXMLPropTristateSwitch(nodes[i], "state", @@ -20486,6 +20488,7 @@ virDomainDefFeaturesCheckABIStability(virDomainDef *src, case VIR_DOMAIN_FEATURE_HTM: case VIR_DOMAIN_FEATURE_NESTED_HV: case VIR_DOMAIN_FEATURE_CCF_ASSIST: + case VIR_DOMAIN_FEATURE_MTE: if (src->features[i] != dst->features[i]) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, _("State of feature '%1$s' differs: source: '%2$s', destination: '%3$s'"), @@ -27005,6 +27008,7 @@ virDomainDefFormatFeatures(virBuffer *buf, case VIR_DOMAIN_FEATURE_HTM: case VIR_DOMAIN_FEATURE_NESTED_HV: case VIR_DOMAIN_FEATURE_CCF_ASSIST: + case VIR_DOMAIN_FEATURE_MTE: switch ((virTristateSwitch) def->features[i]) { case VIR_TRISTATE_SWITCH_LAST: case VIR_TRISTATE_SWITCH_ABSENT: diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index c1cb2ed69d..3f8d6e81c0 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -2169,6 +2169,7 @@ typedef enum { VIR_DOMAIN_FEATURE_SBBC, VIR_DOMAIN_FEATURE_IBS, VIR_DOMAIN_FEATURE_TCG, + VIR_DOMAIN_FEATURE_MTE, VIR_DOMAIN_FEATURE_LAST } virDomainFeature; diff --git a/src/conf/schemas/domaincommon.rng b/src/conf/schemas/domaincommon.rng index f8c7b6a648..37e350ac2c 100644 --- a/src/conf/schemas/domaincommon.rng +++ b/src/conf/schemas/domaincommon.rng @@ -6653,6 +6653,11 @@ <optional> <ref name="tcgfeatures"/> </optional> + <optional> + <element name="mte"> + <ref name="featurestate"/> + </element> + </optional> </interleave> </element> </optional> diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c index da4b9a3b35..99c7775e9b 100644 --- a/src/qemu/qemu_validate.c +++ b/src/qemu/qemu_validate.c @@ -123,6 +123,7 @@ qemuValidateDomainDefFeatures(const virDomainDef *def, break; case VIR_DOMAIN_FEATURE_GIC: + case VIR_DOMAIN_FEATURE_MTE: if (def->features[i] == VIR_TRISTATE_SWITCH_ON && !qemuDomainIsARMVirt(def)) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, diff --git a/tests/qemuxml2argvdata/aarch64-gic-v3.xml b/tests/qemuxml2argvdata/aarch64-gic-v3.xml index 247d6025f7..b9317284b0 100644 --- a/tests/qemuxml2argvdata/aarch64-gic-v3.xml +++ b/tests/qemuxml2argvdata/aarch64-gic-v3.xml @@ -10,6 +10,7 @@ </os> <features> <gic version='3'/> + <mte state='on'/> </features> <cpu mode='host-passthrough' check='none'/> <clock offset='utc'/> diff --git a/tests/qemuxml2xmloutdata/aarch64-gic-v3.aarch64-latest.xml b/tests/qemuxml2xmloutdata/aarch64-gic-v3.aarch64-latest.xml index 5b2fb7df75..1a74903aaa 100644 --- a/tests/qemuxml2xmloutdata/aarch64-gic-v3.aarch64-latest.xml +++ b/tests/qemuxml2xmloutdata/aarch64-gic-v3.aarch64-latest.xml @@ -10,6 +10,7 @@ </os> <features> <gic version='3'/> + <mte state='on'/> </features> <cpu mode='host-passthrough' check='none'/> <clock offset='utc'/> -- 2.39.3

The MTE feature is not supported by all QEMUs, only those with QEMU_CAPS_MACHINE_VIRT_MTE capability. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com&gt; --- src/qemu/qemu_validate.c | 25 +++++++++++++++++-------- 1 file changed, 17 insertions(+), 8 deletions(-) diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c index 99c7775e9b..aec1801672 100644 --- a/src/qemu/qemu_validate.c +++ b/src/qemu/qemu_validate.c @@ -124,14 +124,23 @@ qemuValidateDomainDefFeatures(const virDomainDef *def, case VIR_DOMAIN_FEATURE_GIC: case VIR_DOMAIN_FEATURE_MTE: - if (def->features[i] == VIR_TRISTATE_SWITCH_ON && - !qemuDomainIsARMVirt(def)) { - virReportError(VIR_ERR_CONFIG_UNSUPPORTED, - _("The '%1$s' feature is not supported for architecture '%2$s' or machine type '%3$s'"), - featureName, - virArchToString(def->os.arch), - def->os.machine); - return -1; + if (def->features[i] == VIR_TRISTATE_SWITCH_ON) { + if (!qemuDomainIsARMVirt(def)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("The '%1$s' feature is not supported for architecture '%2$s' or machine type '%3$s'"), + featureName, + virArchToString(def->os.arch), + def->os.machine); + return -1; + } + + if (i == VIR_DOMAIN_FEATURE_MTE && + !virQEMUCapsGet(qemuCaps, QEMU_CAPS_MACHINE_VIRT_MTE)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("The '%1$s' feature is not supported with this QEMU binary"), + featureName); + return -1; + } } break; -- 2.39.3

This is pretty trivia, just append "mte=on/off" to -machine arguments. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com&gt; --- src/qemu/qemu_command.c | 6 ++++++ tests/qemuxml2argvdata/aarch64-gic-v3.args | 2 +- 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 2a6d9408f6..9b993c3aad 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -6797,6 +6797,12 @@ qemuAppendDomainFeaturesMachineParam(virBuffer *buf, virBufferAsprintf(buf, ",cap-ibs=%s", str); } + if (def->features[VIR_DOMAIN_FEATURE_MTE] != VIR_TRISTATE_SWITCH_ABSENT) { + const char *str; + str = virTristateSwitchTypeToString(def->features[VIR_DOMAIN_FEATURE_MTE]); + virBufferAsprintf(buf, ",mte=%s", str); + } + return 0; } diff --git a/tests/qemuxml2argvdata/aarch64-gic-v3.args b/tests/qemuxml2argvdata/aarch64-gic-v3.args index 0d7a1c259a..0244951d87 100644 --- a/tests/qemuxml2argvdata/aarch64-gic-v3.args +++ b/tests/qemuxml2argvdata/aarch64-gic-v3.args @@ -10,7 +10,7 @@ XDG_CONFIG_HOME=/var/lib/libvirt/qemu/domain--1-aarch64test/.config \ -name guest=aarch64test,debug-threads=on \ -S \ -object '{"qom-type":"secret","id":"masterKey0","format":"raw","file":"/var/lib/libvirt/qemu/domain--1-aarch64test/master-key.aes"}' \ --machine virt,usb=off,gic-version=3,dump-guest-core=off,memory-backend=mach-virt.ram \ +-machine virt,usb=off,gic-version=3,mte=on,dump-guest-core=off,memory-backend=mach-virt.ram \ -accel kvm \ -cpu host \ -m 1024 \ -- 2.39.3

On 5/16/23 18:32, Andrea Bolognani wrote: We can still revert the patches, if needed. I admit that my QEMU code base understanding is limited, but the patch you've linked doesn't really expose any CPU feature that libvirt could set. Instead, it enables MTE for KVM under the same API, i.e. -machine virt,mte=on. Fair enough, feel free to revert my patches. Then it's a matter of: diff --git i/hw/arm/virt.c w/hw/arm/virt.c index b99ae18501..ead3555dfa 100644 --- i/hw/arm/virt.c +++ w/hw/arm/virt.c @@ -3111,11 +3111,13 @@ static void virt_machine_class_init(ObjectClass *oc, void *data) "Set on/off to enable/disable reporting host memory errors " "to a KVM guest using ACPI and guest external abort exceptions"); - object_class_property_add_bool(oc, "mte", virt_get_mte, virt_set_mte); - object_class_property_set_description(oc, "mte", - "Set on/off to enable/disable emulating a " - "guest CPU which implements the ARM " - "Memory Tagging Extension"); + if (kvm_arm_mte_supported()) { + object_class_property_add_bool(oc, "mte", virt_get_mte, virt_set_mte); + object_class_property_set_description(oc, "mte", + "Set on/off to enable/disable emulating a " + "guest CPU which implements the ARM " + "Memory Tagging Extension"); + } object_class_property_add_bool(oc, "its", virt_get_its, virt_set_its); Or querying KVM extensions in libvirt (we already do that for some features). Michal