x86: msr: Don't test bits 63:32 of SYSENTER MSRs on 32-bit builds
PATCH 1/1 x86: msr: Don't test bits 63:32 of SYSENTER MSRs on 32-bit builds
Squish the "address" stuffed into SYSENTER_EIP/ESP into an unsigned long so as to drop bits 63:32 on 32-bit builds. VMX diverges from bare metal in the sense that the associated VMCS fields are natural width fields, whereas the actual MSRs hold 64-bit values, even on CPUs that don't support 64-bit mode. This causes the tests to fail if bits 63:32 are non-zero and a VM-Exit/VM-Enter occurs on and/or between WRMSR/RDMSR, e.g. when running the tests in L1 or deeper.
Don't bother trying to actually test that bits 63:32 are dropped, the behavior depends on the (virtual) CPU capabilities, not the build, and the behavior is specific to VMX as both SVM and bare metal preserve the full 64-bit values. And because practically no one cares about 32-bit KVM, let alone an obscure architectural quirk that doesn't affect real world kernels.
--- x86/msr.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
x86/msr.c +3 -2
@@ -16,6 +16,7 @@ struct msr_info {#define addr_64 0x0000123456789abcULL +#define addr_ul (unsigned long)addr_64struct msr_info msr_info[] = {
@@ -23,10 +24,10 @@ struct msr_info msr_info[] = .val_pairs = {{ .valid = 1, .value = 0x1234, .expected = 0x1234}} }, { .index = 0x00000175, .name = "MSR_IA32_SYSENTER_ESP", - .val_pairs = {{ .valid = 1, .value = addr_64, .expected = addr_64}} + .val_pairs = {{ .valid = 1, .value = addr_ul, .expected = addr_ul}} }, { .index = 0x00000176, .name = "IA32_SYSENTER_EIP", - .val_pairs = {{ .valid = 1, .value = addr_64, .expected = addr_64}} + .val_pairs = {{ .valid = 1, .value = addr_ul, .expected = addr_ul}} }, { .index = 0x000001a0, .name = "MSR_IA32_MISC_ENABLE", // reserved: 1:2, 4:6, 8:10, 13:15, 17, 19:21, 24:33, 35:63
Signature · 1 line
2.26.0
Quoted 36 lines · Sean Christopherson
Squish the "address" stuffed into SYSENTER_EIP/ESP into an unsigned long so as to drop bits 63:32 on 32-bit builds. VMX diverges from bare metal in the sense that the associated VMCS fields are natural width fields, whereas the actual MSRs hold 64-bit values, even on CPUs that don't support 64-bit mode. This causes the tests to fail if bits 63:32 are non-zero and a VM-Exit/VM-Enter occurs on and/or between WRMSR/RDMSR, e.g. when running the tests in L1 or deeper.
Don't bother trying to actually test that bits 63:32 are dropped, the behavior depends on the (virtual) CPU capabilities, not the build, and the behavior is specific to VMX as both SVM and bare metal preserve the full 64-bit values. And because practically no one cares about 32-bit KVM, let alone an obscure architectural quirk that doesn't affect real world kernels.
Signed-off-by Sean Christopherson <[email protected]>--- x86/msr.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)x86/msr.c +1
-0@@ -16,6 +16,7 @@ struct msr_info {#define addr_64 0x0000123456789abcULL +#define addr_ul (unsigned long)addr_64struct msr_info msr_info[] = {@@ -23,10 +24,10 @@ struct msr_info msr_info[] = .val_pairs = {{ .valid = 1, .value = 0x1234, .expected = 0x1234}} },
x86/msr.c struct msr_info msr_info[] = · line 33
Quoted 10 lines
x86/msr.c +2
-2{ .index = 0x00000175, .name = "MSR_IA32_SYSENTER_ESP", - .val_pairs = {{ .valid = 1, .value = addr_64, .expected = addr_64}} + .val_pairs = {{ .valid = 1, .value = addr_ul, .expected = addr_ul}} }, { .index = 0x00000176, .name = "IA32_SYSENTER_EIP", - .val_pairs = {{ .valid = 1, .value = addr_64, .expected = addr_64}} + .val_pairs = {{ .valid = 1, .value = addr_ul, .expected = addr_ul}} }, { .index = 0x000001a0, .name = "MSR_IA32_MISC_ENABLE", // reserved: 1:2, 4:6, 8:10, 13:15, 17, 19:21, 24:33, 35:63
Queued, thanks.
Paolo