2015-08-07 08:23:10

by Vineet Gupta

[permalink] [raw]
Subject: [PATCH] ARCv2: spinlock/rwlock/atomics: reduce 1 instruction in exponential backoff

The increment of delay counter was 2 instructions:
Arithmatic Shfit Left (ASL) + set to 1 on overflow

This can be done in 1 using ROtate Left (ROL)

Suggested-by: Nigel Topham <[email protected]>
Cc: Peter Zijlstra (Intel) <[email protected]>
Cc: [email protected]
Signed-off-by: Vineet Gupta <[email protected]>
---
arch/arc/include/asm/atomic.h | 3 +--
arch/arc/include/asm/spinlock.h | 3 +--
2 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/arch/arc/include/asm/atomic.h b/arch/arc/include/asm/atomic.h
index 629dfd0a0c6b..87d18ae53115 100644
--- a/arch/arc/include/asm/atomic.h
+++ b/arch/arc/include/asm/atomic.h
@@ -34,8 +34,7 @@
" mov %[tmp], %[delay] \n" /* tmp = delay */ \
"2: brne.d %[tmp], 0, 2b \n" /* while (tmp != 0) */ \
" sub %[tmp], %[tmp], 1 \n" /* tmp-- */ \
- " asl.f %[delay], %[delay], 1 \n" /* delay *= 2 */ \
- " mov.z %[delay], 1 \n" /* handle overflow */ \
+ " rol %[delay], %[delay] \n" /* delay *= 2 */ \
" b 1b \n" /* start over */ \
"4: ; --- success --- \n" \

diff --git a/arch/arc/include/asm/spinlock.h b/arch/arc/include/asm/spinlock.h
index 7071fc0da56a..db8c59d1eaeb 100644
--- a/arch/arc/include/asm/spinlock.h
+++ b/arch/arc/include/asm/spinlock.h
@@ -260,8 +260,7 @@ static inline void arch_write_unlock(arch_rwlock_t *rw)
" mov %[tmp], %[delay] \n" /* tmp = delay */ \
"2: brne.d %[tmp], 0, 2b \n" /* while (tmp != 0) */ \
" sub %[tmp], %[tmp], 1 \n" /* tmp-- */ \
- " asl.f %[delay], %[delay], 1 \n" /* delay *= 2 */ \
- " mov.z %[delay], 1 \n" /* handle overflow */ \
+ " rol %[delay], %[delay] \n" /* delay *= 2 */ \
" b 1b \n" /* start over */ \
" \n" \
"4: ; --- done --- \n" \
--
1.9.1


2015-08-07 11:00:50

by Peter Zijlstra

[permalink] [raw]
Subject: Re: [PATCH] ARCv2: spinlock/rwlock/atomics: reduce 1 instruction in exponential backoff

On Fri, Aug 07, 2015 at 01:52:19PM +0530, Vineet Gupta wrote:
> The increment of delay counter was 2 instructions:
> Arithmatic Shfit Left (ASL) + set to 1 on overflow
>
> This can be done in 1 using ROtate Left (ROL)

Cute :-)