[POWERPC] Enable interrupts if we are doing fp math emulation
Anytime we are emulating an instruction we are going to be doing some form of
get_user() to get the instruction image to decode. Since get_user() might
sleep we need to ensure we have interrupts enabled or we might see something
like:
Debug: sleeping function called from invalid context at arch/powerpc/kernel/traps.c:697
in_atomic():0, irqs_disabled():1
Call Trace:
[D6023EB0] [C0007F84] show_stack+0x58/0x174 (unreliable)
[D6023EE0] [C0022C34] __might_sleep+0xbc/0xd0
[D6023EF0] [C000D158] program_check_exception+0x1d8/0x4fc
[D6023F40] [C000E744] ret_from_except_full+0x0/0x4c
--- Exception: 700 at 0x102a7100
LR = 0xdb9ef04
However, we want to ensure that interrupts are disabled when handling a trap
exception that might be used for a kernel breakpoint. This is why ProgramCheck
is marked as EXC_XFER_STD instead of EXC_XFER_EE.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index 6915b91..f038caa 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c
@@ -739,20 +739,7 @@
extern int do_mathemu(struct pt_regs *regs);
/* We can now get here via a FP Unavailable exception if the core
- * has no FPU, in that case no reason flags will be set */
-#ifdef CONFIG_MATH_EMULATION
- /* (reason & REASON_ILLEGAL) would be the obvious thing here,
- * but there seems to be a hardware bug on the 405GP (RevD)
- * that means ESR is sometimes set incorrectly - either to
- * ESR_DST (!?) or 0. In the process of chasing this with the
- * hardware people - not sure if it can happen on any illegal
- * instruction or only on FP instructions, whether there is a
- * pattern to occurences etc. -dgibson 31/Mar/2003 */
- if (!(reason & REASON_TRAP) && do_mathemu(regs) == 0) {
- emulate_single_step(regs);
- return;
- }
-#endif /* CONFIG_MATH_EMULATION */
+ * has no FPU, in that case the reason flags will be 0 */
if (reason & REASON_FP) {
/* IEEE FP exception */
@@ -778,6 +765,20 @@
local_irq_enable();
+#ifdef CONFIG_MATH_EMULATION
+ /* (reason & REASON_ILLEGAL) would be the obvious thing here,
+ * but there seems to be a hardware bug on the 405GP (RevD)
+ * that means ESR is sometimes set incorrectly - either to
+ * ESR_DST (!?) or 0. In the process of chasing this with the
+ * hardware people - not sure if it can happen on any illegal
+ * instruction or only on FP instructions, whether there is a
+ * pattern to occurences etc. -dgibson 31/Mar/2003 */
+ if (do_mathemu(regs) == 0) {
+ emulate_single_step(regs);
+ return;
+ }
+#endif /* CONFIG_MATH_EMULATION */
+
/* Try to emulate it if we should. */
if (reason & (REASON_ILLEGAL | REASON_PRIVILEGED)) {
switch (emulate_instruction(regs)) {