* [PATCH] KVM: arm64: Avoid rewinding the PC for trapped SMCs
@ 2020-07-14 9:45 Andrew Scull
2020-07-15 13:16 ` James Morse
0 siblings, 1 reply; 2+ messages in thread
From: Andrew Scull @ 2020-07-14 9:45 UTC (permalink / raw)
To: kvmarm; +Cc: maz, kernel-team
When handling a guest's SError that was identified in __guest_exit, it
is injected back into the guest at the instruction that caused the
switch into hyp. A trapped SMC does not advance the PC so no adjustment
needs to be made to the PC.
Fixes: ddb3d07cfe90 ("arm64: KVM: Inject a Virtual SError if it was pending")
Signed-off-by: Andrew Scull <ascull@google.com>
---
arch/arm64/kvm/handle_exit.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c
index 98ab33139982..7d9e198ebdf2 100644
--- a/arch/arm64/kvm/handle_exit.c
+++ b/arch/arm64/kvm/handle_exit.c
@@ -244,12 +244,11 @@ int handle_exit(struct kvm_vcpu *vcpu, struct kvm_run *run,
u8 esr_ec = ESR_ELx_EC(kvm_vcpu_get_esr(vcpu));
/*
- * HVC/SMC already have an adjusted PC, which we need
+ * HVC already has an adjusted PC, which we need
* to correct in order to return to after having
* injected the SError.
*/
- if (esr_ec == ESR_ELx_EC_HVC32 || esr_ec == ESR_ELx_EC_HVC64 ||
- esr_ec == ESR_ELx_EC_SMC32 || esr_ec == ESR_ELx_EC_SMC64) {
+ if (esr_ec == ESR_ELx_EC_HVC32 || esr_ec == ESR_ELx_EC_HVC64) {
u32 adj = kvm_vcpu_trap_il_is32bit(vcpu) ? 4 : 2;
*vcpu_pc(vcpu) -= adj;
}
--
2.27.0.389.gc38d7665816-goog
_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] KVM: arm64: Avoid rewinding the PC for trapped SMCs
2020-07-14 9:45 [PATCH] KVM: arm64: Avoid rewinding the PC for trapped SMCs Andrew Scull
@ 2020-07-15 13:16 ` James Morse
0 siblings, 0 replies; 2+ messages in thread
From: James Morse @ 2020-07-15 13:16 UTC (permalink / raw)
To: Andrew Scull, maz; +Cc: kernel-team, kvmarm
Hi Andrew,
On 14/07/2020 10:45, Andrew Scull wrote:
> When handling a guest's SError that was identified in __guest_exit, it
> is injected back into the guest at the instruction that caused the
> switch into hyp. A trapped SMC does not advance the PC so no adjustment
> needs to be made to the PC.
Oops!
Reviewed-by: James Morse <james.morse@arm.com>
Thanks,
James
> Fixes: ddb3d07cfe90 ("arm64: KVM: Inject a Virtual SError if it was pending")
> Signed-off-by: Andrew Scull <ascull@google.com>
> ---
> arch/arm64/kvm/handle_exit.c | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c
> index 98ab33139982..7d9e198ebdf2 100644
> --- a/arch/arm64/kvm/handle_exit.c
> +++ b/arch/arm64/kvm/handle_exit.c
> @@ -244,12 +244,11 @@ int handle_exit(struct kvm_vcpu *vcpu, struct kvm_run *run,
> u8 esr_ec = ESR_ELx_EC(kvm_vcpu_get_esr(vcpu));
>
> /*
> - * HVC/SMC already have an adjusted PC, which we need
> + * HVC already has an adjusted PC, which we need
> * to correct in order to return to after having
> * injected the SError.
> */
> - if (esr_ec == ESR_ELx_EC_HVC32 || esr_ec == ESR_ELx_EC_HVC64 ||
> - esr_ec == ESR_ELx_EC_SMC32 || esr_ec == ESR_ELx_EC_SMC64) {
> + if (esr_ec == ESR_ELx_EC_HVC32 || esr_ec == ESR_ELx_EC_HVC64) {
> u32 adj = kvm_vcpu_trap_il_is32bit(vcpu) ? 4 : 2;
> *vcpu_pc(vcpu) -= adj;
> }
>
_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2020-07-15 13:16 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-07-14 9:45 [PATCH] KVM: arm64: Avoid rewinding the PC for trapped SMCs Andrew Scull
2020-07-15 13:16 ` James Morse
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).