| Message ID | 1542023835-21446-22-git-send-email-julien.thierry@arm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | arm64: provide pseudo NMI with GICv3 | expand |
On Mon, Nov 12, 2018 at 11:57:12AM +0000, Julien Thierry wrote: > diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c > index 5f4d9ac..66344cd 100644 > --- a/arch/arm64/kernel/traps.c > +++ b/arch/arm64/kernel/traps.c > @@ -897,13 +897,17 @@ bool arm64_is_fatal_ras_serror(struct pt_regs *regs, unsigned int esr) > > asmlinkage void do_serror(struct pt_regs *regs, unsigned int esr) > { > - nmi_enter(); > + const bool was_in_nmi = in_nmi(); > + > + if (!was_in_nmi) > + nmi_enter(); > > /* non-RAS errors are not containable */ > if (!arm64_is_ras_serror(esr) || arm64_is_fatal_ras_serror(regs, esr)) > arm64_serror_panic(regs, esr); > > - nmi_exit(); > + if (!was_in_nmi) > + nmi_exit(); > } Do we actually need nmi_enter/exit in the outer do_serror() function? Could we just move it to arm64_serror_panic()?
Hi Catalin, On 04/12/2018 18:09, Catalin Marinas wrote: > On Mon, Nov 12, 2018 at 11:57:12AM +0000, Julien Thierry wrote: >> diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c >> index 5f4d9ac..66344cd 100644 >> --- a/arch/arm64/kernel/traps.c >> +++ b/arch/arm64/kernel/traps.c >> @@ -897,13 +897,17 @@ bool arm64_is_fatal_ras_serror(struct pt_regs *regs, unsigned int esr) >> >> asmlinkage void do_serror(struct pt_regs *regs, unsigned int esr) >> { >> - nmi_enter(); >> + const bool was_in_nmi = in_nmi(); >> + >> + if (!was_in_nmi) >> + nmi_enter(); >> >> /* non-RAS errors are not containable */ >> if (!arm64_is_ras_serror(esr) || arm64_is_fatal_ras_serror(regs, esr)) >> arm64_serror_panic(regs, esr); >> >> - nmi_exit(); >> + if (!was_in_nmi) >> + nmi_exit(); >> } > > Do we actually need nmi_enter/exit in the outer do_serror() function? > Could we just move it to arm64_serror_panic()? They might need to be here in the future: if we support kernel-first we would have extra calls in here that need to be in_nmi(), the same if we call out to APEI to support APCI's NOTIFY_SEI. Thanks, James
diff --git a/arch/arm64/kernel/traps.c b/arch/arm64/kernel/traps.c index 5f4d9ac..66344cd 100644 --- a/arch/arm64/kernel/traps.c +++ b/arch/arm64/kernel/traps.c @@ -897,13 +897,17 @@ bool arm64_is_fatal_ras_serror(struct pt_regs *regs, unsigned int esr) asmlinkage void do_serror(struct pt_regs *regs, unsigned int esr) { - nmi_enter(); + const bool was_in_nmi = in_nmi(); + + if (!was_in_nmi) + nmi_enter(); /* non-RAS errors are not containable */ if (!arm64_is_ras_serror(esr) || arm64_is_fatal_ras_serror(regs, esr)) arm64_serror_panic(regs, esr); - nmi_exit(); + if (!was_in_nmi) + nmi_exit(); } void __pte_error(const char *file, int line, unsigned long val)
Per definition of the daifflags, Serrors can occur during any interrupt context, that includes NMI contexts. Trying to nmi_enter in an nmi context will crash. Skip nmi_enter/nmi_exit when serror occurred during an NMI. Suggested-by: James Morse <james.morse@arm.com> Signed-off-by: Julien Thierry <julien.thierry@arm.com> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Will Deacon <will.deacon@arm.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Dave Martin <dave.martin@arm.com> Cc: James Morse <james.morse@arm.com> --- arch/arm64/kernel/traps.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)