Hi,
I noticed Marc had applied this patch in his kvmarm tree, with the
updated commit message. Thanks Marc :)
On 2019/7/25 20:48, James Morse wrote:
Hi,
On 13/07/2019 05:40, Zenghui Yu wrote:
We've added two ESR exception classes for new ARM hardware extensions:
ESR_ELx_EC_PAC and ESR_ELx_EC_SVE.
This patch updates "kvm_arm_exception_class" for these two EC, which the
new EC will be parsed in kvm_exit trace events (for guest's usage of
new EC will be visible to user-space via kvm_exit() trace events... ?
Pointer Authentication and Scalable Vector Extension). The same updates
to "esr_class_str" for ESR_ELx_EC_PAC, by which we can get more accurate
debug info.
Its accurate either way round. This stops the trace-point print from printing
the EC in
hex, instead giving it a name, like we do for all the others.
So this should be "readable".
Trivial changes, update them in one go.
(I don't think this bit needs to go in the git-log!)
diff --git a/arch/arm64/include/asm/kvm_arm.h b/arch/arm64/include/asm/kvm_arm.h
index a8b205e..ddf9d76 100644
--- a/arch/arm64/include/asm/kvm_arm.h
+++ b/arch/arm64/include/asm/kvm_arm.h
@@ -316,9 +316,10 @@
#define kvm_arm_exception_class \
ECN(UNKNOWN), ECN(WFx), ECN(CP15_32), ECN(CP15_64), ECN(CP14_MR), \
- ECN(CP14_LS), ECN(FP_ASIMD), ECN(CP10_ID), ECN(CP14_64), ECN(SVC64), \
- ECN(HVC64), ECN(SMC64), ECN(SYS64), ECN(IMP_DEF), ECN(IABT_LOW), \
- ECN(IABT_CUR), ECN(PC_ALIGN), ECN(DABT_LOW), ECN(DABT_CUR), \
+ ECN(CP14_LS), ECN(FP_ASIMD), ECN(CP10_ID), ECN(PAC), ECN(CP14_64), \
+ ECN(SVC64), ECN(HVC64), ECN(SMC64), ECN(SYS64), ECN(SVE), \
+ ECN(IMP_DEF), ECN(IABT_LOW), ECN(IABT_CUR), \
+ ECN(PC_ALIGN), ECN(DABT_LOW), ECN(DABT_CUR), \
ECN(SP_ALIGN), ECN(FP_EXC32), ECN(FP_EXC64), ECN(SERROR), \
ECN(BREAKPT_LOW), ECN(BREAKPT_CUR), ECN(SOFTSTP_LOW), \
ECN(SOFTSTP_CUR), ECN(WATCHPT_LOW), ECN(WATCHPT_CUR), \
Do we need to add:
ESR_ELx_EC_ILL
ESR_ELx_EC_HVC32
ESR_ELx_EC_SMC32
I think they're needed. But these three EC are not included into
kvm_arm_exception_class at the time it was introduced, I don't know
the reason.
(I don't see how any of the _CUR entries could ever happen as these EC
originally come
from ESR_EL2, but they're harmless)
(I agree with you.)
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c
index 8c03456..969e156 100644
--- a/arch/arm64/kernel/traps.c
+++ b/arch/arm64/kernel/traps.c
@@ -734,6 +734,7 @@ asmlinkage void __exception do_sysinstr(unsigned int esr,
struct pt_regs *regs)
[ESR_ELx_EC_CP14_LS] = "CP14 LDC/STC",
[ESR_ELx_EC_FP_ASIMD] = "ASIMD",
[ESR_ELx_EC_CP10_ID] = "CP10 MRC/VMRS",
+ [ESR_ELx_EC_PAC] = "PAC",
[ESR_ELx_EC_CP14_64] = "CP14 MCRR/MRRC",
[ESR_ELx_EC_ILL] = "PSTATE.IL",
[ESR_ELx_EC_SVC32] = "SVC (AArch32)",
Reviewed-by: James Morse <[email protected]>
Thanks for your reviewing!
Zenghui
