If the previous context had interrupts disabled, we should better keep them disabled. This was noticed in the unwinding code where a copy_from_kernel_nofault() triggered a page fault, and after the fixup by the page fault handler interrupts where suddenly enabled. Signed-off-by: Sven Schnelle <svens@xxxxxxxxxxxxxx> --- arch/parisc/kernel/traps.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/arch/parisc/kernel/traps.c b/arch/parisc/kernel/traps.c index 690e6abcaf22..3c5d968da415 100644 --- a/arch/parisc/kernel/traps.c +++ b/arch/parisc/kernel/traps.c @@ -480,9 +480,9 @@ void notrace handle_interruption(int code, struct pt_regs *regs) int si_code; if (code == 1) - pdc_console_restart(); /* switch back to pdc if HPMC */ - else - local_irq_enable(); + pdc_console_restart(); /* switch back to pdc if HPMC */ + else if (!irqs_disabled_flags(regs->gr[0])) + local_irq_enable(); /* Security check: * If the priority level is still user, and the -- 2.33.0
