| Message ID | 1231757365-15717-3-git-send-email-amit.shah@redhat.com (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Headers | show |
On Mon, Jan 12, 2009 at 10:49:25AM +0000, Amit Shah wrote: > CPUID functions 4, 0xb and 0xd behave differently for different values of ECX. > Store these values if userspace passes them. > > Signed-off-by: Amit Shah <amit.shah@redhat.com> > --- > arch/x86/kvm/x86.c | 18 +++++++++++++++--- > 1 files changed, 15 insertions(+), 3 deletions(-) > > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index 06b44fb..8dc4b29 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -1126,7 +1126,7 @@ static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu, > struct kvm_cpuid *cpuid, > struct kvm_cpuid_entry __user *entries) > { > - int r, i; > + int r, i, count; > struct kvm_cpuid_entry *cpuid_entries; > > r = -E2BIG; > @@ -1146,8 +1146,20 @@ static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu, > vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx; > vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx; > vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx; > - vcpu->arch.cpuid_entries[i].index = 0; > - vcpu->arch.cpuid_entries[i].flags = 0; > + switch (cpuid_entries[i].function) { > + case 4: > + case 0xb: > + case 0xd: > + vcpu->arch.cpuid_entries[i].index = count++; Isn't this using count uninitialized? Cheers, Muli
Amit Shah wrote: > CPUID functions 4, 0xb and 0xd behave differently for different values of ECX. > Store these values if userspace passes them. > Userspace has no way of knowing whether the kernel contains this fix or not. Use the SET_CPUID2 interface instead.
On Tue, Jan 13, 2009 at 11:27:07AM +0200, Avi Kivity wrote: > Amit Shah wrote: >> CPUID functions 4, 0xb and 0xd behave differently for different values of ECX. >> Store these values if userspace passes them. >> > > Userspace has no way of knowing whether the kernel contains this fix or not. > > Use the SET_CPUID2 interface instead. OK; So first attempt to use the SET_CPUID2 interface. If that fails (new userspace on older kernel), what should we do? Amit -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Amit Shah wrote: >> Use the SET_CPUID2 interface instead. >> > > OK; So first attempt to use the SET_CPUID2 interface. If that fails (new > userspace on older kernel), what should we do? > SET_CPUID2 was introduced on 2.6.25; that is now the minimum requirement for qemu and kvm-userspace anyway.
On Tue, Jan 13, 2009 at 11:45:55AM +0200, Avi Kivity wrote: > Amit Shah wrote: >>> Use the SET_CPUID2 interface instead. >>> >> >> OK; So first attempt to use the SET_CPUID2 interface. If that fails (new >> userspace on older kernel), what should we do? >> > > SET_CPUID2 was introduced on 2.6.25; that is now the minimum requirement > for qemu and kvm-userspace anyway. OK; that's great. I'll respin the patches. Amit. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 06b44fb..8dc4b29 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -1126,7 +1126,7 @@ static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu, struct kvm_cpuid *cpuid, struct kvm_cpuid_entry __user *entries) { - int r, i; + int r, i, count; struct kvm_cpuid_entry *cpuid_entries; r = -E2BIG; @@ -1146,8 +1146,20 @@ static int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu, vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx; vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx; vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx; - vcpu->arch.cpuid_entries[i].index = 0; - vcpu->arch.cpuid_entries[i].flags = 0; + switch (cpuid_entries[i].function) { + case 4: + case 0xb: + case 0xd: + vcpu->arch.cpuid_entries[i].index = count++; + vcpu->arch.cpuid_entries[i].flags = + KVM_CPUID_FLAG_SIGNIFICANT_INDEX; + break; + default: + vcpu->arch.cpuid_entries[i].index = 0; + vcpu->arch.cpuid_entries[i].flags = 0; + } + if (cpuid_entries[i].function != cpuid_entries[i+1].function) + count = 0; vcpu->arch.cpuid_entries[i].padding[0] = 0; vcpu->arch.cpuid_entries[i].padding[1] = 0; vcpu->arch.cpuid_entries[i].padding[2] = 0;
CPUID functions 4, 0xb and 0xd behave differently for different values of ECX. Store these values if userspace passes them. Signed-off-by: Amit Shah <amit.shah@redhat.com> --- arch/x86/kvm/x86.c | 18 +++++++++++++++--- 1 files changed, 15 insertions(+), 3 deletions(-)