https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65459
Bug ID: 65459
Summary: munaligned-access still produce split access codes
Product: gcc
Version: 5.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: ma.jiang at zte dot com.cn
Hi all,
The munaligned-access option on ARM can not stop gcc from producing split
access codes. A very simple test
char mt[20];
void main()
{
void *mm=&(mt[1]);
*(( int *)mm)=4;
}
compiled with -O2 -munaligned-access could reproduce this problem.As we can
see,"*(( int *)mm)=4;" cause gcc to emit 4 strb instructions not a unaligned
stw .
In my opinion, this bug is caused by the SLOW_UNALIGNED_ACCESS macro,which is
defined to be 1 in arm.h. It is very strange that the munaligned-access option
can affect neither SLOW_UNALIGNED_ACCESS nor STRICT_ALIGNMENT. On PPC,the
mstrict-align can affect STRICT_ALIGNMENT， and SLOW_UNALIGNED_ACCESS is
determined by STRICT_ALIGNMENT, ALIGN, and MODE(args of the
SLOW_UNALIGNED_ACCESS macro). This is a good example for ARM, I think.
