Skip to contentSkip to conversation

ARM: Fix find_next_zero_bit and related assembly

PATCH In review James Jones 4 participants 8 messages Nov 12, 2010 – Nov 24, 2010 Arm (32-bit)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ARM: Fix find_next_zero_bit and related assembly

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

---
 arch/arm/lib/findbit.S |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)
arch/arm/lib/findbit.S +4 -2
index 1e4cbd4..64f6bc1 100644
@@ -174,8 +174,8 @@ ENDPROC(_find_next_bit_be)
  */
 .L_found:
 #if __LINUX_ARM_ARCH__ >= 5
-		rsb	r1, r3, #0
-		and	r3, r3, r1
+		rsb	r0, r3, #0
+		and	r3, r3, r0
 		clz	r3, r3
 		rsb	r3, r3, #31
 		add	r0, r2, r3
@@ -190,5 +190,7 @@ ENDPROC(_find_next_bit_be)
 		addeq	r2, r2, #1
 		mov	r0, r2
 #endif
+		cmp	r1, r0			@ Clamp to maxbit
+		movlo	r0, r1
 		mov	pc, lr
 
Signature · 1 line
1.7.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 11 lines · James Jones

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

Please send to RMK's patch system.

Quoted 36 lines
---
 arch/arm/lib/findbit.S |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)
arch/arm/lib/findbit.S +4 -2
index 1e4cbd4..64f6bc1 100644
@@ -174,8 +174,8 @@ ENDPROC(_find_next_bit_be)
  */
 .L_found:
 #if __LINUX_ARM_ARCH__ >= 5
-		rsb	r1, r3, #0
-		and	r3, r3, r1
+		rsb	r0, r3, #0
+		and	r3, r3, r0
 		clz	r3, r3
 		rsb	r3, r3, #31
 		add	r0, r2, r3
@@ -190,5 +190,7 @@ ENDPROC(_find_next_bit_be)
 		addeq	r2, r2, #1
 		mov	r0, r2
 #endif
+		cmp	r1, r0			@ Clamp to maxbit
+		movlo	r0, r1
 		mov	pc, lr
 
Signature · 7 lines
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
#2 lore raw
Quoted 10 lines · On Tuesday 23 November 2010 13:26:47 Nicolas Pitre wrote:
Quoted 9 lines · James Jones

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

Quoted 6 lines
Quoted 2 lines

Please send to RMK's patch system.

Thanks for the review. It's already in the patch system, but I updated the entry to include your reviewed-by line.

-James

Quoted 36 lines
Quoted 36 lines

---

 arch/arm/lib/findbit.S |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)
arch/arm/lib/findbit.S +2 -2
index 1e4cbd4..64f6bc1 100644
@@ -174,8 +174,8 @@ ENDPROC(_find_next_bit_be)
 
  */
 
 .L_found:
 #if __LINUX_ARM_ARCH__ >= 5
 
-		rsb	r1, r3, #0
-		and	r3, r3, r1
+		rsb	r0, r3, #0
+		and	r3, r3, r0
arch/arm/lib/findbit.S +2 -0
 		clz	r3, r3
 		rsb	r3, r3, #31
 		add	r0, r2, r3
@@ -190,5 +190,7 @@ ENDPROC(_find_next_bit_be)
 
 		addeq	r2, r2, #1
 		mov	r0, r2
 
 #endif
 
+		cmp	r1, r0			@ Clamp to maxbit
+		movlo	r0, r1
 
 		mov	pc, lr
#3 lore raw
Quoted 11 lines · James Jones
Quoted 10 lines · On Tuesday 23 November 2010 13:26:47 Nicolas Pitre wrote:
Quoted 9 lines · James Jones

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

Please send to RMK's patch system.

Thanks for the review. It's already in the patch system, but I updated the entry to include your reviewed-by line.

Should this be sent to the stable tree too?

#4 lore raw
Quoted 14 lines · Stephen Boyd
Quoted 13 lines · James Jones
Quoted 12 lines · On Tuesday 23 November 2010 13:26:47 Nicolas Pitre wrote:
Quoted 11 lines · James Jones

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

Please send to RMK's patch system.

Thanks for the review. It's already in the patch system, but I updated the entry to include your reviewed-by line.

Should this be sent to the stable tree too?

It could, yes. This is hardly an urgent fix though, as the bug has been there virtually forever.

Nicolas

#5 lore raw
Quoted 23 lines · Nicolas Pitre
Quoted 22 lines · Stephen Boyd
Quoted 20 lines · James Jones
Quoted 10 lines · On Tuesday 23 November 2010 13:26:47 Nicolas Pitre wrote:
Quoted 9 lines · James Jones

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

Quoted 6 lines
Quoted 2 lines

Please send to RMK's patch system.

Thanks for the review. It's already in the patch system, but I updated the entry to include your reviewed-by line.

Quoted 7 lines
Quoted 1 line

Should this be sent to the stable tree too?

It could, yes. This is hardly an urgent fix though, as the bug has been there virtually forever.

Nicolas

While ancient, it does cause per-cpu allocations to fail in some situations, which generally causes panics.

Thanks,
-James

#6 lore raw
Quoted 26 lines · James Jones
Quoted 25 lines · Nicolas Pitre
Quoted 23 lines · Stephen Boyd
Quoted 14 lines · James Jones
Quoted 13 lines · On Tuesday 23 November 2010 13:26:47 Nicolas Pitre wrote:
Quoted 11 lines · James Jones

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

Quoted 6 lines
Quoted 3 lines

Please send to RMK's patch system.

Thanks for the review. It's already in the patch system, but I updated the entry to include your reviewed-by line.

Should this be sent to the stable tree too?

Quoted 8 lines
Quoted 5 lines

It could, yes. This is hardly an urgent fix though, as the bug has been there virtually forever.

Nicolas

While ancient, it does cause per-cpu allocations to fail in some situations, which generally causes panics.

That would be a good justification for the stable tree then.

Nicolas

#7 lore raw
Quoted 32 lines · Nicolas Pitre
Quoted 30 lines · James Jones
Quoted 29 lines · Nicolas Pitre
Quoted 15 lines · Stephen Boyd
Quoted 14 lines · James Jones
Quoted 13 lines · On Tuesday 23 November 2010 13:26:47 Nicolas Pitre wrote:
Quoted 11 lines · James Jones

The find_next_bit, find_first_bit, find_next_zero_bit and find_first_zero_bit functions were not properly clamping to the maxbit argument at the bit level. They were instead only checking maxbit at the byte level. To fix this, add a compare and a conditional move instruction to the end of the common bit-within-the- byte code used by all the functions and be sure not to clobber the maxbit argument before it is used.

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

Please send to RMK's patch system.

Thanks for the review. It's already in the patch system, but I updated the entry to include your reviewed-by line.

Should this be sent to the stable tree too?

It could, yes. This is hardly an urgent fix though, as the bug has been there virtually forever.

Quoted 6 lines
Quoted 4 lines
Quoted 1 line

Nicolas

While ancient, it does cause per-cpu allocations to fail in some situations, which generally causes panics.

That would be a good justification for the stable tree then.

And the best way to do that is _not_ to send it to stable at kernel.org, but when you submit the patch to the patch system, add a line in the sign-off area:

and when it goes into mainline (which is a requirement for stable patches) the stable team will automatically pick it up - without anyone needing to do any additional work.