diff mbox

sparc64 mcount misaligned stack

Message ID 20091205.201802.71100445.davem@davemloft.net
State Accepted
Delegated to: David Miller
Headers show

Commit Message

David Miller Dec. 6, 2009, 4:18 a.m. UTC 
From: Igor Kovalenko <igor.v.kovalenko@gmail.com>
Date: Sun, 6 Dec 2009 01:27:08 +0300

> I wonder if mcount ever works on sparc64, if it does work indeed I'd
> like to know what I'm missing there.

Yep, two distinct bugs there.  Please try this patch, thanks
for the report.

sparc64: Fix stack debugging IRQ stack regression.

Commit 4f70f7a91bffdcc39f088748dc678953eb9a3fbd
(sparc64: Implement IRQ stacks.) has two bugs.

First, the softirq range check forgets to subtract STACK_BIAS
before comparing with %sp.  Next, on failure the wrong label
is jumped to, resulting in a bogus stack being loaded.

Reported-by: Igor Kovalenko <igor.v.kovalenko@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

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

Comments

Igor V. Kovalenko Dec. 6, 2009, 5:37 p.m. UTC | #1 
On Sun, Dec 6, 2009 at 7:18 AM, David Miller <davem@davemloft.net> wrote:
> From: Igor Kovalenko <igor.v.kovalenko@gmail.com>
> Date: Sun, 6 Dec 2009 01:27:08 +0300
>
>> I wonder if mcount ever works on sparc64, if it does work indeed I'd
>> like to know what I'm missing there.
>
> Yep, two distinct bugs there.  Please try this patch, thanks
> for the report.

Works here, thanks!

- does not trap too early
- prints "Stack overflow" and exits to prom without fault (I tried
removing STACK_BIAS adjustment to make it trap)
David Miller Dec. 6, 2009, 7:21 p.m. UTC | #2 
From: Igor Kovalenko <igor.v.kovalenko@gmail.com>
Date: Sun, 6 Dec 2009 20:37:03 +0300

> On Sun, Dec 6, 2009 at 7:18 AM, David Miller <davem@davemloft.net> wrote:
>> From: Igor Kovalenko <igor.v.kovalenko@gmail.com>
>> Date: Sun, 6 Dec 2009 01:27:08 +0300
>>
>>> I wonder if mcount ever works on sparc64, if it does work indeed I'd
>>> like to know what I'm missing there.
>>
>> Yep, two distinct bugs there. ,A Please try this patch, thanks
>> for the report.
> 
> Works here, thanks!
> 
> - does not trap too early
> - prints "Stack overflow" and exits to prom without fault (I tried
> removing STACK_BIAS adjustment to make it trap)

Great, thanks for testing.  I'll push the fix around.
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/arch/sparc/lib/mcount.S b/arch/sparc/lib/mcount.S
index 7ce9c65..24b8b12 100644
--- a/arch/sparc/lib/mcount.S
+++ b/arch/sparc/lib/mcount.S
@@ -64,8 +64,9 @@  mcount:
 2:	 sethi		%hi(softirq_stack), %g3
 	or		%g3, %lo(softirq_stack), %g3
 	ldx		[%g3 + %g1], %g7
+	sub		%g7, STACK_BIAS, %g7
 	cmp		%sp, %g7
-	bleu,pt		%xcc, 2f
+	bleu,pt		%xcc, 3f
 	 sethi		%hi(THREAD_SIZE), %g3
 	add		%g7, %g3, %g7
 	cmp		%sp, %g7
@@ -75,7 +76,7 @@  mcount:
 	 * again, we are already trying to output the stack overflow
 	 * message.
 	 */
-	sethi		%hi(ovstack), %g7		! cant move to panic stack fast enough
+3:	sethi		%hi(ovstack), %g7		! cant move to panic stack fast enough
 	 or		%g7, %lo(ovstack), %g7
 	add		%g7, OVSTACKSIZE, %g3
 	sub		%g3, STACK_BIAS + 192, %g3