Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-11-24 05:02:12 +00:00
Code Issues Packages Projects Releases Wiki Activity

parisc: do not inline pa_memcpy() internal functions

Browse Source 
gcc (4.8.x) creates wrong code when the pa_memcpy() functions are
inlined.  Especially in 32bit builds it calculates wrong return values
if we encounter a fault during execution of the memcpy.

Signed-off-by: Helge Deller <deller@gmx.de>
This commit is contained in:
Helge Deller 2013-11-17 22:03:11 +01:00
parent a0ffa8f06f
commit 9af63aedb8
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
arch/parisc/lib/memcpy.c
View File

@ -161,7 +161,7 @@ static inline void prefetch_dst(const void *addr)
/* Copy from a not-aligned src to an aligned dst, using shifts. Handles 4 words
* per loop. This code is derived from glibc.
*/
static inline unsigned long copy_dstaligned(unsigned long dst,
static noinline unsigned long copy_dstaligned(unsigned long dst,
unsigned long src, unsigned long len)
{
/* gcc complains that a2 and a3 may be uninitialized, but actually
@ -276,7 +276,7 @@ handle_store_error:
/* Returns PA_MEMCPY_OK, PA_MEMCPY_LOAD_ERROR or PA_MEMCPY_STORE_ERROR.
* In case of an access fault the faulty address can be read from the per_cpu
* exception data struct. */
static unsigned long pa_memcpy_internal(void *dstp, const void *srcp,
static noinline unsigned long pa_memcpy_internal(void *dstp, const void *srcp,
unsigned long len)
{
register unsigned long src, dst, t1, t2, t3;