![David S. Miller](/assets/img/avatar_default.png)
For the case where the source is not aligned modulo 8 we don't use load-twins to suck the data in and this kills performance since normal loads allocate in the L1 cache (unlike load-twin) and thus big memcpys swipe the entire L1 D-cache. We need to allocate a register window to implement this properly, but that actually simplifies a lot of things as a nice side-effect. Signed-off-by: David S. Miller <davem@davemloft.net>
38 lines
770 B
ArmAsm
38 lines
770 B
ArmAsm
/* NGcopy_from_user.S: Niagara optimized copy from userspace.
|
|
*
|
|
* Copyright (C) 2006, 2007 David S. Miller (davem@davemloft.net)
|
|
*/
|
|
|
|
#define EX_LD(x) \
|
|
98: x; \
|
|
.section .fixup; \
|
|
.align 4; \
|
|
99: wr %g0, ASI_AIUS, %asi;\
|
|
ret; \
|
|
restore %g0, 1, %o0; \
|
|
.section __ex_table,"a";\
|
|
.align 4; \
|
|
.word 98b, 99b; \
|
|
.text; \
|
|
.align 4;
|
|
|
|
#ifndef ASI_AIUS
|
|
#define ASI_AIUS 0x11
|
|
#endif
|
|
|
|
#define FUNC_NAME NGcopy_from_user
|
|
#define LOAD(type,addr,dest) type##a [addr] ASI_AIUS, dest
|
|
#define LOAD_TWIN(addr_reg,dest0,dest1) \
|
|
ldda [addr_reg] ASI_BLK_INIT_QUAD_LDD_AIUS, dest0
|
|
#define EX_RETVAL(x) %g0
|
|
|
|
#ifdef __KERNEL__
|
|
#define PREAMBLE \
|
|
rd %asi, %g1; \
|
|
cmp %g1, ASI_AIUS; \
|
|
bne,pn %icc, memcpy_user_stub; \
|
|
nop
|
|
#endif
|
|
|
|
#include "NGmemcpy.S"
|