powerpc/64/kuap: Conditionally restore AMR in interrupt exit

The AMR update is made conditional on AMR actually changing, which
should be the less common case on most workloads (though kernel page
faults on uaccess could be frequent, this doesn't significantly slow
down that case).

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200429065654.1677541-4-npiggin@gmail.com
此提交包含在:
Nicholas Piggin
2020-04-29 16:56:51 +10:00
提交者 Michael Ellerman
父節點 cb2b53cbff
當前提交 579940bb45
共有 2 個檔案被更改,包括 29 行新增7 行删除

查看文件

@@ -242,6 +242,10 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
BUG_ON(!FULL_REGS(regs));
BUG_ON(regs->softe != IRQS_ENABLED);
/*
* We don't need to restore AMR on the way back to userspace for KUAP.
* AMR can only have been unlocked if we interrupted the kernel.
*/
kuap_check_amr();
local_irq_save(flags);
@@ -313,13 +317,14 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign
unsigned long *ti_flagsp = &current_thread_info()->flags;
unsigned long flags;
unsigned long ret = 0;
unsigned long amr;
if (IS_ENABLED(CONFIG_PPC_BOOK3S) && unlikely(!(regs->msr & MSR_RI)))
unrecoverable_exception(regs);
BUG_ON(regs->msr & MSR_PR);
BUG_ON(!FULL_REGS(regs));
kuap_check_amr();
amr = kuap_get_and_check_amr();
if (unlikely(*ti_flagsp & _TIF_EMULATE_STACK_STORE)) {
clear_bits(_TIF_EMULATE_STACK_STORE, ti_flagsp);
@@ -367,10 +372,11 @@ again:
#endif
/*
* We don't need to restore AMR on the way back to userspace for KUAP.
* The value of AMR only matters while we're in the kernel.
* Don't want to mfspr(SPRN_AMR) here, because this comes after mtmsr,
* which would cause Read-After-Write stalls. Hence, we take the AMR
* value from the check above.
*/
kuap_restore_amr(regs);
kuap_restore_amr(regs, amr);
return ret;
}