powerpc: memcpy optimization for 64bit LE

Unaligned stores take alignment exceptions on POWER7 running in little-endian.
This is a dumb little-endian base memcpy that prevents unaligned stores.
Once booted the feature fixup code switches over to the VMX copy loops
(which are already endian safe).

The question is what we do before that switch over. The base 64bit
memcpy takes alignment exceptions on POWER7 so we can't use it as is.
Fixing the causes of alignment exception would slow it down, because
we'd need to ensure all loads and stores are aligned either through
rotate tricks or bytewise loads and stores. Either would be bad for
all other 64bit platforms.

[ I simplified the loop a bit - Anton ]

Signed-off-by: Philippe Bergheaud <felix@linux.vnet.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

2 files changed, 16 insertions, 2 deletions

diff --git a/arch/powerpc/lib/Makefile b/arch/powerpc/lib/Makefile
index 95a20e17dbff..59fa2de9546d 100644
--- a/arch/powerpc/lib/Makefile
+++ b/arch/powerpc/lib/Makefile
@@ -23,9 +23,7 @@ obj-y			+= checksum_$(CONFIG_WORD_SIZE).o
 obj-$(CONFIG_PPC64)     += checksum_wrappers_64.o
 endif
 
-ifeq ($(CONFIG_CPU_LITTLE_ENDIAN),)
 obj-$(CONFIG_PPC64)             += memcpy_power7.o memcpy_64.o 
-endif
 
 obj-$(CONFIG_PPC_EMULATE_SSTEP) += sstep.o ldstfp.o
 
diff --git a/arch/powerpc/lib/memcpy_64.S b/arch/powerpc/lib/memcpy_64.S
index 72ad055168a3..dc4ba7953b92 100644
--- a/arch/powerpc/lib/memcpy_64.S
+++ b/arch/powerpc/lib/memcpy_64.S
@@ -12,12 +12,27 @@
         .align  7
 _GLOBAL(memcpy)
 BEGIN_FTR_SECTION
+#ifdef __LITTLE_ENDIAN__
+        cmpdi   cr7,r5,0
+#else
         std     r3,48(r1)       /* save destination pointer for return value */
+#endif
 FTR_SECTION_ELSE
 #ifndef SELFTEST
         b       memcpy_power7
 #endif
 ALT_FTR_SECTION_END_IFCLR(CPU_FTR_VMX_COPY)
+#ifdef __LITTLE_ENDIAN__
+        /* dumb little-endian memcpy that will get replaced at runtime */
+        addi r9,r3,-1
+        addi r4,r4,-1
+        beqlr cr7
+        mtctr r5
+1:      lbzu r10,1(r4)
+        stbu r10,1(r9)
+        bdnz 1b
+        blr
+#else
         PPC_MTOCRF(0x01,r5)
         cmpldi  cr1,r5,16
         neg     r6,r3           # LS 3 bits = # bytes to 8-byte dest bdry
@@ -203,3 +218,4 @@ END_FTR_SECTION_IFCLR(CPU_FTR_UNALIGNED_LD_STD)
         stb     r0,0(r3)
 4:      ld      r3,48(r1)       /* return dest pointer */
         blr
+#endif