[PATCH 1/2] alpha: irq: run all handlers with interrupts disabled

From: Will Deacon
Date: Sun Mar 17 2013 - 17:50:14 EST


Linux has expected that interrupt handlers are executed with local
interrupts disabled for a while now, so ensure that this is the case on
Alpha even for non-device interrupts such as IPIs.

Without this patch, secondary boot results in the following backtrace:

------------[ cut here ]------------
warning: at kernel/softirq.c:139 __local_bh_enable+0xb8/0xd0()
modules linked in:
fffffc0000867c70 fffffc0000876618 fffffc0000330628 0000000000000200
fffffc000088e1b0 fffffc0000868000 0000000000000000 0000000000000000
fffffc0000331594 fffffc00008c79a0 fffffc000085fac0 0000000000000001
fffffc00003597e0 0000000000000008 fffffc000031b3b4 0000000000000001
0000000000000001 0000000000000000 fffffc0000315c4c fffffc00008b88f8
fffffc0000856e20 fffffc0000856b00 fffffc0001a060c0 0000000000000007
trace:
[<fffffc0000330628>] __local_bh_enable+0xb8/0xd0
[<fffffc0000331594>] irq_enter+0x74/0xa0
[<fffffc00003597e0>] scheduler_ipi+0x50/0x100
[<fffffc000031b3b4>] handle_ipi+0x84/0x260
[<fffffc0000315c4c>] do_entint+0x1ac/0x2e0
[<fffffc0000331620>] irq_exit+0x60/0xa0
[<fffffc00003159e8>] handle_irq+0x98/0x100
[<fffffc0000315d68>] do_entint+0x2c8/0x2e0
[<fffffc0000310d80>] ret_from_sys_call+0x0/0x10
[<fffffc0000365124>] load_balance+0x3e4/0x870
[<fffffc0000312bd4>] cpu_idle+0x24/0x80
[<fffffc0000391bd0>] rcu_eqs_enter_common.isra.38+0x0/0x120
[<fffffc0000312bf0>] cpu_idle+0x40/0x80
[<fffffc0000715030>] rest_init+0xc0/0xe0
[<fffffc000031001c>] _stext+0x1c/0x20

---[ end trace 5bbac1dc3a0ec32e ]---

A similar dump occurs if you try to reboot using magic-sysrq.

Cc: Richard Henderson <rth@xxxxxxxxxxx>
Cc: Ivan Kokshaysky <ink@xxxxxxxxxxxxxxxxxxxx>
Cc: Matt Turner <mattst88@xxxxxxxxx>
Signed-off-by: Will Deacon <will.deacon@xxxxxxx>
---
arch/alpha/kernel/irq.c | 7 -------
arch/alpha/kernel/irq_alpha.c | 9 ++++++++-
2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/arch/alpha/kernel/irq.c b/arch/alpha/kernel/irq.c
index 2872acc..7b2be25 100644
--- a/arch/alpha/kernel/irq.c
+++ b/arch/alpha/kernel/irq.c
@@ -117,13 +117,6 @@ handle_irq(int irq)
return;
}

- /*
- * From here we must proceed with IPL_MAX. Note that we do not
- * explicitly enable interrupts afterwards - some MILO PALcode
- * (namely LX164 one) seems to have severe problems with RTI
- * at IPL 0.
- */
- local_irq_disable();
irq_enter();
generic_handle_irq_desc(irq, desc);
irq_exit();
diff --git a/arch/alpha/kernel/irq_alpha.c b/arch/alpha/kernel/irq_alpha.c
index 772ddfdb..1216dfb 100644
--- a/arch/alpha/kernel/irq_alpha.c
+++ b/arch/alpha/kernel/irq_alpha.c
@@ -45,6 +45,14 @@ do_entInt(unsigned long type, unsigned long vector,
unsigned long la_ptr, struct pt_regs *regs)
{
struct pt_regs *old_regs;
+
+ /*
+ * Disable interrupts during IRQ handling.
+ * Note that there is no matching local_irq_enable() due to
+ * severe problems with RTI at IPL0 and some MILO PALcode
+ * (namely LX164).
+ */
+ local_irq_disable();
switch (type) {
case 0:
#ifdef CONFIG_SMP
@@ -62,7 +70,6 @@ do_entInt(unsigned long type, unsigned long vector,
{
long cpu;

- local_irq_disable();
smp_percpu_timer_interrupt(regs);
cpu = smp_processor_id();
if (cpu != boot_cpuid) {
--
1.8.0

--
To unsubscribe from this list: send the line "unsubscribe linux-alpha" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html