csky: optimize memcpy_{from,to}io() and memset_io()
authorGuo Ren <guoren@linux.alibaba.com>
Wed, 6 Apr 2022 13:32:22 +0000 (21:32 +0800)
committerGuo Ren <guoren@linux.alibaba.com>
Mon, 18 Apr 2022 13:23:55 +0000 (21:23 +0800)
commit8318f7c231d5be09e47410c5ab387b9bef6fe19e
tree0072d121bd5f7b4d5615e72a34d0b6ebb9841a98
parente4df2d5e852a7d24df3672ae9951eb79e179be08
csky: optimize memcpy_{from,to}io() and memset_io()

Optimize memcpy_{from,to}io() and memset_io() by transferring in
64 bit as much as possible with minimized barrier usage.  This
simplest optimization brings faster throughput compare to current
byte-by-byte read and write with barrier in the loop. Code's
skeleton is taken from the powerpc & arm64.

Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@kernel.org>
arch/csky/include/asm/io.h
arch/csky/kernel/Makefile
arch/csky/kernel/io.c [new file with mode: 0644]