2019-07-05 12:46:32

by Paolo Bonzini

[permalink] [raw]
Subject: [PATCH] KVM: LAPIC: ARBPRI is a reserved register for x2APIC

kvm-unit-tests were adjusted to match bare metal behavior, but KVM
itself was not doing what bare metal does; fix that.

Signed-off-by: Paolo Bonzini <[email protected]>
---
arch/x86/kvm/lapic.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index d6ca5c4f29f1..2e4470f2685a 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -1318,7 +1318,7 @@ int kvm_lapic_reg_read(struct kvm_lapic *apic, u32 offset, int len,
unsigned char alignment = offset & 0xf;
u32 result;
/* this bitmask has a bit cleared for each reserved register */
- static const u64 rmask = 0x43ff01ffffffe70cULL;
+ u64 rmask = 0x43ff01ffffffe70cULL;

if ((alignment + len) > 4) {
apic_debug("KVM_APIC_READ: alignment error %x %d\n",
@@ -1326,6 +1326,10 @@ int kvm_lapic_reg_read(struct kvm_lapic *apic, u32 offset, int len,
return 1;
}

+ /* ARBPRI is also reserved on x2APIC */
+ if (apic_x2apic_mode(apic))
+ rmask &= ~(1 << (APIC_ARBPRI >> 4));
+
if (offset > 0x3f0 || !(rmask & (1ULL << (offset >> 4)))) {
apic_debug("KVM_APIC_READ: read reserved register %x\n",
offset);
--
1.8.3.1


2019-07-05 13:49:12

by Paolo Bonzini

[permalink] [raw]
Subject: Re: [PATCH] KVM: LAPIC: ARBPRI is a reserved register for x2APIC

On 05/07/19 15:37, Nadav Amit wrote:
>> On Jul 5, 2019, at 5:14 AM, Paolo Bonzini <[email protected]> wrote:
>>
>> kvm-unit-tests were adjusted to match bare metal behavior, but KVM
>> itself was not doing what bare metal does; fix that.
>>
>> Signed-off-by: Paolo Bonzini <[email protected]>
>
> Reported-by ?

I found it myself while running the tests, was there a report too?

Paolo

2019-07-05 13:51:42

by Nadav Amit

[permalink] [raw]
Subject: Re: [PATCH] KVM: LAPIC: ARBPRI is a reserved register for x2APIC

> On Jul 5, 2019, at 6:43 AM, Paolo Bonzini <[email protected]> wrote:
>
> On 05/07/19 15:37, Nadav Amit wrote:
>>> On Jul 5, 2019, at 5:14 AM, Paolo Bonzini <[email protected]> wrote:
>>>
>>> kvm-unit-tests were adjusted to match bare metal behavior, but KVM
>>> itself was not doing what bare metal does; fix that.
>>>
>>> Signed-off-by: Paolo Bonzini <[email protected]>
>>
>> Reported-by ?
>
> I found it myself while running the tests, was there a report too?

Perhaps it is in my mind. I thought that fixing a test is equivalent to a
bug report.

2019-07-05 14:06:05

by Liran Alon

[permalink] [raw]
Subject: Re: [PATCH] KVM: LAPIC: ARBPRI is a reserved register for x2APIC



> On 5 Jul 2019, at 15:14, Paolo Bonzini <[email protected]> wrote:
>
> kvm-unit-tests were adjusted to match bare metal behavior, but KVM
> itself was not doing what bare metal does; fix that.
>
> Signed-off-by: Paolo Bonzini <[email protected]>
> ---
> arch/x86/kvm/lapic.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
> index d6ca5c4f29f1..2e4470f2685a 100644
> --- a/arch/x86/kvm/lapic.c
> +++ b/arch/x86/kvm/lapic.c
> @@ -1318,7 +1318,7 @@ int kvm_lapic_reg_read(struct kvm_lapic *apic, u32 offset, int len,
> unsigned char alignment = offset & 0xf;
> u32 result;
> /* this bitmask has a bit cleared for each reserved register */
> - static const u64 rmask = 0x43ff01ffffffe70cULL;
> + u64 rmask = 0x43ff01ffffffe70cULL;

Why not rename this to “used_bits_mask” and calculate it properly by macros?
It seems a lot nicer than having a pre-calculated magic.

-Liran

>
> if ((alignment + len) > 4) {
> apic_debug("KVM_APIC_READ: alignment error %x %d\n",
> @@ -1326,6 +1326,10 @@ int kvm_lapic_reg_read(struct kvm_lapic *apic, u32 offset, int len,
> return 1;
> }
>
> + /* ARBPRI is also reserved on x2APIC */
> + if (apic_x2apic_mode(apic))
> + rmask &= ~(1 << (APIC_ARBPRI >> 4));
> +
> if (offset > 0x3f0 || !(rmask & (1ULL << (offset >> 4)))) {
> apic_debug("KVM_APIC_READ: read reserved register %x\n",
> offset);
> --
> 1.8.3.1
>

2019-07-05 14:07:54

by Paolo Bonzini

[permalink] [raw]
Subject: Re: [PATCH] KVM: LAPIC: ARBPRI is a reserved register for x2APIC

On 05/07/19 14:37, Liran Alon wrote:
>> + u64 rmask = 0x43ff01ffffffe70cULL;
> Why not rename this to “used_bits_mask” and calculate it properly by macros?
> It seems a lot nicer than having a pre-calculated magic.

Yes, I wanted to do the same but I didn't have time right now. I am
planning to cleanup after the merge window, but if a patch comes soon I
can apply it instead of course. :)

Something like

#define APIC_VALID_REG_MASK(reg) (1ull << ((reg) >> 4))
#define APIC_VALID_REGS_MASK(first, count) \
(APIC_VALID_REG_MASK(first) * (1ull << ((count) - 1)))

followed by

if (offset > 0x3f0 ||
!(valid_regs_mask & APIC_VALID_REG_MASK(offset))

Paolo

> -Liran
>

2019-07-05 14:09:57

by Nadav Amit

[permalink] [raw]
Subject: Re: [PATCH] KVM: LAPIC: ARBPRI is a reserved register for x2APIC

> On Jul 5, 2019, at 5:14 AM, Paolo Bonzini <[email protected]> wrote:
>
> kvm-unit-tests were adjusted to match bare metal behavior, but KVM
> itself was not doing what bare metal does; fix that.
>
> Signed-off-by: Paolo Bonzini <[email protected]>

Reported-by ?