[libvirt] [PATCH 35/36] cpu_x86: Read CPU features from IA32_ARCH_CAPABILITIES MSR

Jiri Denemark posted 36 patches 6 years, 10 months ago
[libvirt] [PATCH 35/36] cpu_x86: Read CPU features from IA32_ARCH_CAPABILITIES MSR
Posted by Jiri Denemark 6 years, 10 months ago
This is used by the host capabilities code to construct host CPU
definition.

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
---
 src/cpu/cpu_x86.c | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c
index 1fe6770d57..74385f7992 100644
--- a/src/cpu/cpu_x86.c
+++ b/src/cpu/cpu_x86.c
@@ -2759,6 +2759,30 @@ virCPUx86GetHost(virCPUDefPtr cpu,
         cpuidSet(CPUX86_EXTENDED, cpuData) < 0)
         goto cleanup;
 
+    /* Read the IA32_ARCH_CAPABILITIES MSR (0x10a) if supported.
+     * This is best effort since there might be no way to read the MSR. */
+    if (virCPUx86DataCheckFeature(cpuData, "arch-capabilities") == 1) {
+        uint64_t msr;
+        unsigned long index = 0x10a;
+        int rc = virHostCPUGetMSR(index, &msr);
+
+        if (rc < 0) {
+            virResetLastError();
+        } else if (rc == 0) {
+            virCPUx86DataItem item = {
+                .type = VIR_CPU_X86_DATA_MSR,
+                .data.msr = {
+                    .index = index,
+                    .eax = msr & 0xffffffff,
+                    .edx = msr >> 32,
+                },
+            };
+
+            if (virCPUx86DataAdd(cpuData, &item) < 0)
+                return -1;
+        }
+    }
+
     ret = x86DecodeCPUData(cpu, cpuData, models);
     cpu->microcodeVersion = virHostCPUGetMicrocodeVersion();
 
-- 
2.21.0

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
Re: [libvirt] [PATCH 35/36] cpu_x86: Read CPU features from IA32_ARCH_CAPABILITIES MSR
Posted by Ján Tomko 6 years, 10 months ago
On Mon, Apr 08, 2019 at 10:42:39AM +0200, Jiri Denemark wrote:
>This is used by the host capabilities code to construct host CPU
>definition.
>
>Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
>---
> src/cpu/cpu_x86.c | 24 ++++++++++++++++++++++++
> 1 file changed, 24 insertions(+)
>
>diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c
>index 1fe6770d57..74385f7992 100644
>--- a/src/cpu/cpu_x86.c
>+++ b/src/cpu/cpu_x86.c
>@@ -2759,6 +2759,30 @@ virCPUx86GetHost(virCPUDefPtr cpu,
>         cpuidSet(CPUX86_EXTENDED, cpuData) < 0)
>         goto cleanup;
>
>+    /* Read the IA32_ARCH_CAPABILITIES MSR (0x10a) if supported.
>+     * This is best effort since there might be no way to read the MSR. */
>+    if (virCPUx86DataCheckFeature(cpuData, "arch-capabilities") == 1) {
>+        uint64_t msr;
>+        unsigned long index = 0x10a;
>+        int rc = virHostCPUGetMSR(index, &msr);
>+
>+        if (rc < 0) {
>+            virResetLastError();

Using virResetLastError in a place other than starting a public API
seems fishy.

If we're ignoring the error, we should not have logged it in the first
place.

>+        } else if (rc == 0) {
>+            virCPUx86DataItem item = {
>+                .type = VIR_CPU_X86_DATA_MSR,
>+                .data.msr = {
>+                    .index = index,
>+                    .eax = msr & 0xffffffff,
>+                    .edx = msr >> 32,
>+                },
>+            };
>+
>+            if (virCPUx86DataAdd(cpuData, &item) < 0)
>+                return -1;
>+        }
>+    }
>+
>     ret = x86DecodeCPUData(cpu, cpuData, models);
>     cpu->microcodeVersion = virHostCPUGetMicrocodeVersion();

Reviewed-by: Ján Tomko <jtomko@redhat.com>

Jano
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list