diff mbox

[RESEND,v2,2/2] arm: KVM: keep arm vfp/simd exit handling consistent with arm64

Message ID 1434739306-23417-1-git-send-email-m.smarduch@samsung.com (mailing list archive)
State New, archived
Headers show

Commit Message

Mario Smarduch June 19, 2015, 6:41 p.m. UTC
After enhancing arm64 FP/SIMD exit handling, 32-bit VFP exit branch is moved
to guest trap handling, keeping exit handling flow for both architectures 
consistent.

Also replaced an unused label with a comment.

Signed-off-by: Mario Smarduch <m.smarduch@samsung.com>
---
 arch/arm/kvm/interrupts.S |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

Comments

Mario Smarduch June 22, 2015, 3:17 a.m. UTC | #1
On 06/19/2015 11:41 AM, Mario Smarduch wrote:
[...]

>
>  
> -host_switch_to_hyp:
> +	/* Guest is calling a hyp function, save minimum context branch to it */ 

Hi Marc,
    please correct Guest to Host (or word this to your preference).
Somehow wondered off on Friday :)

Thanks,
- Mario

>  	pop	{r0, r1, r2}
>  
>  	/* Check for __hyp_get_vectors */
> @@ -411,6 +407,10 @@ guest_trap:
>  
>  	@ Check if we need the fault information
>  	lsr	r1, r1, #HSR_EC_SHIFT
> +#ifdef CONFIG_VFPv3
> +	cmp	r1, #HSR_EC_CP_0_13
> +	beq	switch_to_guest_vfp
> +#endif
>  	cmp	r1, #HSR_EC_IABT
>  	mrceq	p15, 4, r2, c6, c0, 2	@ HIFAR
>  	beq	2f
>
Marc Zyngier June 24, 2015, 3:44 p.m. UTC | #2
Hi Mario,

On 19/06/15 19:41, Mario Smarduch wrote:
> After enhancing arm64 FP/SIMD exit handling, 32-bit VFP exit branch is moved
> to guest trap handling, keeping exit handling flow for both architectures 
> consistent.
> 
> Also replaced an unused label with a comment.
> 
> Signed-off-by: Mario Smarduch <m.smarduch@samsung.com>
> ---
>  arch/arm/kvm/interrupts.S |   10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/arch/arm/kvm/interrupts.S b/arch/arm/kvm/interrupts.S
> index 79caf79..16f00f1 100644
> --- a/arch/arm/kvm/interrupts.S
> +++ b/arch/arm/kvm/interrupts.S
> @@ -363,10 +363,6 @@ hyp_hvc:
>  	@ Check syndrome register
>  	mrc	p15, 4, r1, c5, c2, 0	@ HSR
>  	lsr	r0, r1, #HSR_EC_SHIFT
> -#ifdef CONFIG_VFPv3
> -	cmp	r0, #HSR_EC_CP_0_13
> -	beq	switch_to_guest_vfp
> -#endif
>  	cmp	r0, #HSR_EC_HVC
>  	bne	guest_trap		@ Not HVC instr.
>  
> @@ -380,7 +376,7 @@ hyp_hvc:
>  	cmp     r2, #0
>  	bne	guest_trap		@ Guest called HVC
>  
> -host_switch_to_hyp:
> +	/* Guest is calling a hyp function, save minimum context branch to it */
>  	pop	{r0, r1, r2}
>  
>  	/* Check for __hyp_get_vectors */
> @@ -411,6 +407,10 @@ guest_trap:
>  
>  	@ Check if we need the fault information
>  	lsr	r1, r1, #HSR_EC_SHIFT
> +#ifdef CONFIG_VFPv3
> +	cmp	r1, #HSR_EC_CP_0_13
> +	beq	switch_to_guest_vfp
> +#endif
>  	cmp	r1, #HSR_EC_IABT
>  	mrceq	p15, 4, r2, c6, c0, 2	@ HIFAR
>  	beq	2f
> 

Two things: I think you can now drop the load_vcpu that is in
switch_to_guest_vfp (you just executed it in guest_trap). You also have
an extra store to VCPU_HSR, but I don't think that's a big deal.

Thanks,

	M.
Mario Smarduch June 25, 2015, 3:33 a.m. UTC | #3
On 06/24/2015 08:44 AM, Marc Zyngier wrote:
> Hi Mario,
> 
> On 19/06/15 19:41, Mario Smarduch wrote:
>> After enhancing arm64 FP/SIMD exit handling, 32-bit VFP exit branch is moved
>> to guest trap handling, keeping exit handling flow for both architectures 
>> consistent.
>>
>> Also replaced an unused label with a comment.
>>
>> Signed-off-by: Mario Smarduch <m.smarduch@samsung.com>
>> ---
>>  arch/arm/kvm/interrupts.S |   10 +++++-----
>>  1 file changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/arch/arm/kvm/interrupts.S b/arch/arm/kvm/interrupts.S
>> index 79caf79..16f00f1 100644
>> --- a/arch/arm/kvm/interrupts.S
>> +++ b/arch/arm/kvm/interrupts.S
>> @@ -363,10 +363,6 @@ hyp_hvc:
>>  	@ Check syndrome register
>>  	mrc	p15, 4, r1, c5, c2, 0	@ HSR
>>  	lsr	r0, r1, #HSR_EC_SHIFT
>> -#ifdef CONFIG_VFPv3
>> -	cmp	r0, #HSR_EC_CP_0_13
>> -	beq	switch_to_guest_vfp
>> -#endif
>>  	cmp	r0, #HSR_EC_HVC
>>  	bne	guest_trap		@ Not HVC instr.
>>  
>> @@ -380,7 +376,7 @@ hyp_hvc:
>>  	cmp     r2, #0
>>  	bne	guest_trap		@ Guest called HVC
>>  
>> -host_switch_to_hyp:
>> +	/* Guest is calling a hyp function, save minimum context branch to it */
>>  	pop	{r0, r1, r2}
>>  
>>  	/* Check for __hyp_get_vectors */
>> @@ -411,6 +407,10 @@ guest_trap:
>>  
>>  	@ Check if we need the fault information
>>  	lsr	r1, r1, #HSR_EC_SHIFT
>> +#ifdef CONFIG_VFPv3
>> +	cmp	r1, #HSR_EC_CP_0_13
>> +	beq	switch_to_guest_vfp
>> +#endif
>>  	cmp	r1, #HSR_EC_IABT
>>  	mrceq	p15, 4, r2, c6, c0, 2	@ HIFAR
>>  	beq	2f
>>
> 
> Two things: I think you can now drop the load_vcpu that is in
> switch_to_guest_vfp (you just executed it in guest_trap). You also have
> an extra store to VCPU_HSR, but I don't think that's a big deal.

I left VCPU_HSR as is.

- Mario

> 
> Thanks,
> 
> 	M.
>
diff mbox

Patch

diff --git a/arch/arm/kvm/interrupts.S b/arch/arm/kvm/interrupts.S
index 79caf79..16f00f1 100644
--- a/arch/arm/kvm/interrupts.S
+++ b/arch/arm/kvm/interrupts.S
@@ -363,10 +363,6 @@  hyp_hvc:
 	@ Check syndrome register
 	mrc	p15, 4, r1, c5, c2, 0	@ HSR
 	lsr	r0, r1, #HSR_EC_SHIFT
-#ifdef CONFIG_VFPv3
-	cmp	r0, #HSR_EC_CP_0_13
-	beq	switch_to_guest_vfp
-#endif
 	cmp	r0, #HSR_EC_HVC
 	bne	guest_trap		@ Not HVC instr.
 
@@ -380,7 +376,7 @@  hyp_hvc:
 	cmp     r2, #0
 	bne	guest_trap		@ Guest called HVC
 
-host_switch_to_hyp:
+	/* Guest is calling a hyp function, save minimum context branch to it */
 	pop	{r0, r1, r2}
 
 	/* Check for __hyp_get_vectors */
@@ -411,6 +407,10 @@  guest_trap:
 
 	@ Check if we need the fault information
 	lsr	r1, r1, #HSR_EC_SHIFT
+#ifdef CONFIG_VFPv3
+	cmp	r1, #HSR_EC_CP_0_13
+	beq	switch_to_guest_vfp
+#endif
 	cmp	r1, #HSR_EC_IABT
 	mrceq	p15, 4, r2, c6, c0, 2	@ HIFAR
 	beq	2f