linux/arch/sparc/lib
David S. Miller 2077cef4d5 sparc64: Fix several bugs in memmove().
Firstly, handle zero length calls properly.  Believe it or not there
are a few of these happening during early boot.

Next, we can't just drop to a memcpy() call in the forward copy case
where dst <= src.  The reason is that the cache initializing stores
used in the Niagara memcpy() implementations can end up clearing out
cache lines before we've sourced their original contents completely.

For example, considering NG4memcpy, the main unrolled loop begins like
this:

     load   src + 0x00
     load   src + 0x08
     load   src + 0x10
     load   src + 0x18
     load   src + 0x20
     store  dst + 0x00

Assume dst is 64 byte aligned and let's say that dst is src - 8 for
this memcpy() call.  That store at the end there is the one to the
first line in the cache line, thus clearing the whole line, which thus
clobbers "src + 0x28" before it even gets loaded.

To avoid this, just fall through to a simple copy only mildly
optimized for the case where src and dst are 8 byte aligned and the
length is a multiple of 8 as well.  We could get fancy and call
GENmemcpy() but this is good enough for how this thing is actually
used.

Reported-by: David Ahern <david.ahern@oracle.com>
Reported-by: Bob Picco <bpicco@meloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-03-23 09:22:10 -07:00
..
ashldi3.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
ashrdi3.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
atomic32.c sparc32: Implement xchg and atomic_xchg using ATOMIC_HASH locks 2014-11-07 12:51:44 -08:00
atomic_64.S locking, sparc64: Fix atomics 2014-09-10 11:45:04 +02:00
bitext.c sparc/srmmu: clear trailing edge of bitmap properly 2013-03-31 19:29:12 -04:00
bitops.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
blockops.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
bzero.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
checksum_32.S sparc32: Fixed unaligned memory copying in function __csum_partial_copy_sparc_generic 2011-05-11 21:35:04 -07:00
checksum_64.S
clear_page.S sparc64: Make PAGE_OFFSET variable. 2013-11-12 15:22:34 -08:00
cmpdi2.c
copy_in_user.S
copy_page.S sparc64: Make PAGE_OFFSET variable. 2013-11-12 15:22:34 -08:00
copy_user.S
COPYING.LIB
csum_copy.S
csum_copy_from_user.S
csum_copy_to_user.S
divdi3.S sparc32: Kill off software 32-bit multiply/divide routines. 2012-05-15 11:23:47 -07:00
ffs.S sparc: Use popc when possible for ffs/__ffs/ffz. 2011-08-02 21:28:53 -07:00
GENbzero.S
GENcopy_from_user.S
GENcopy_to_user.S
GENmemcpy.S
GENpage.S
GENpatch.S
hweight.S sparc: Use popc if possible for hweight routines. 2011-08-02 21:28:50 -07:00
iomap.c sparc: switch to GENERIC_PCI_IOMAP 2011-12-04 15:59:49 +02:00
ipcsum.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
ksyms.c locking,arch,sparc: Fold atomic_ops 2014-08-14 12:48:13 +02:00
libgcc.h
locks.S
lshrdi3.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
Makefile sparc32: introduce asm-generic/io.h 2014-05-02 01:30:21 -04:00
mcount.S sparc64,ftrace: Remove check of obsolete variable function_trace_stop 2014-07-18 13:57:03 -04:00
memcmp.S
memcpy.S sparc32: Correct the return value of memcpy. 2011-10-20 15:17:23 -07:00
memmove.S sparc64: Fix several bugs in memmove(). 2015-03-23 09:22:10 -07:00
memscan_32.S
memscan_64.S
memset.S sparc: Let memset return the address argument 2014-09-09 16:38:10 -07:00
muldi3.S sparc32: Kill off software 32-bit multiply/divide routines. 2012-05-15 11:23:47 -07:00
NG2copy_from_user.S
NG2copy_to_user.S
NG2memcpy.S sparc64: Add membar to Niagara2 memcpy code. 2014-05-17 11:28:05 -07:00
NG2patch.S
NG4clear_page.S sparc64: Niagara-4 bzero/memset, plus use MRU stores in page copy. 2012-10-05 13:45:26 -07:00
NG4copy_from_user.S sparc64: Fix comment type in NG4 copy from user. 2012-09-27 14:26:41 -07:00
NG4copy_page.S sparc64: Niagara-4 bzero/memset, plus use MRU stores in page copy. 2012-10-05 13:45:26 -07:00
NG4copy_to_user.S sparc64: Add SPARC-T4 optimized memcpy. 2012-09-27 00:35:11 -07:00
NG4memcpy.S sparc64: Fix FPU register corruption with AES crypto offload. 2014-10-14 19:37:58 -07:00
NG4memset.S sparc64: Niagara-4 bzero/memset, plus use MRU stores in page copy. 2012-10-05 13:45:26 -07:00
NG4patch.S sparc64: Niagara-4 bzero/memset, plus use MRU stores in page copy. 2012-10-05 13:45:26 -07:00
NGbzero.S
NGcopy_from_user.S
NGcopy_to_user.S
NGmemcpy.S
NGpage.S sparc64: Add SPARC-T4 optimized memcpy. 2012-09-27 00:35:11 -07:00
NGpatch.S
PeeCeeI.c sparc64: update IO access functions in PeeCeeI 2014-07-21 21:43:18 -07:00
strlen.S
strncmp_32.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
strncmp_64.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00
U1copy_from_user.S
U1copy_to_user.S
U1memcpy.S sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. 2012-06-27 01:25:23 -07:00
U3copy_from_user.S
U3copy_to_user.S
U3memcpy.S
U3patch.S
ucmpdi2.c sparc32: add ucmpdi2 2012-05-19 15:23:57 -07:00
udivdi3.S sparc32: Kill off software 32-bit multiply/divide routines. 2012-05-15 11:23:47 -07:00
user_fixup.c
VISsave.S
xor.S sparc: Convert some assembler over to linakge.h's ENTRY/ENDPROC 2012-05-11 20:33:22 -07:00