| To: | Ralf Baechle <ralf@linux-mips.org>, linux-mips <linux-mips@linux-mips.org> |
|---|---|
| Subject: | [PATCH 0/2] Rearrange MIPS barriers and optimize for Octeon. |
| From: | David Daney <ddaney@caviumnetworks.com> |
| Date: | Fri, 08 Jan 2010 17:16:29 -0800 |
| Sender: | linux-mips-bounce@linux-mips.org |
| User-agent: | Thunderbird 2.0.0.21 (X11/20090320) |
Locking/atomic performance on Octeon can be improved by using optimized barrier instructions. The first patch in this set rearranges and simplifies (at least in my mind) the existing barriers. Then the second patch adds Octeon specific barriers. I will reply with the two patches. David Daney (2): MIPS: New macro smp_mb__before_llsc. MIPS: Octeon: Use optimized memory barrier primitives. arch/mips/Kconfig | 1 - arch/mips/include/asm/atomic.h | 16 ++++++------arch/mips/include/asm/barrier.h | 52 +++++++++++++++++++++++++++----------- arch/mips/include/asm/bitops.h | 8 +++--- arch/mips/include/asm/cmpxchg.h | 10 +++--- arch/mips/include/asm/spinlock.h | 4 +- arch/mips/include/asm/system.h | 4 +++ 7 files changed, 60 insertions(+), 35 deletions(-) |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | [PATCH] MIPS: Octeon: Use non-overflowing arithmetic in sched_clock, David Daney |
|---|---|
| Next by Date: | [PATCH 2/2] MIPS: Octeon: Use optimized memory barrier primitives., David Daney |
| Previous by Thread: | [PATCH] MIPS: Octeon: Use non-overflowing arithmetic in sched_clock, David Daney |
| Next by Thread: | [PATCH 2/2] MIPS: Octeon: Use optimized memory barrier primitives., David Daney |
| Indexes: | [Date] [Thread] [Top] [All Lists] |