diff options
author | Jeffrey A Law <law@cygnus.com> | 1999-05-25 13:14:40 +0000 |
---|---|---|
committer | Jeff Law <law@gcc.gnu.org> | 1999-05-25 07:14:40 -0600 |
commit | 3ce7c5a21da0c2e0d865a9457a8fda2529abb959 (patch) | |
tree | c856716ededa6050c016dd3bc3dba3d046e5dd6a /gcc/flow.c | |
parent | c2c08b29b98521bc05f483fc06e3759b37e6f023 (diff) | |
download | gcc-3ce7c5a21da0c2e0d865a9457a8fda2529abb959.zip gcc-3ce7c5a21da0c2e0d865a9457a8fda2529abb959.tar.gz gcc-3ce7c5a21da0c2e0d865a9457a8fda2529abb959.tar.bz2 |
flow.c (mark_set_1): Do not record BLKmode stores as dead store elimination candidates.
�
* flow.c (mark_set_1): Do not record BLKmode stores as dead
store elimination candidates.
From-SVN: r27140
Diffstat (limited to 'gcc/flow.c')
-rw-r--r-- | gcc/flow.c | 3 |
1 files changed, 3 insertions, 0 deletions
@@ -3258,6 +3258,9 @@ mark_set_1 (needed, dead, x, insn, significant) invalidate_mems_from_autoinc (insn); if (GET_CODE (reg) == MEM && ! side_effects_p (reg) + /* We do not know the size of a BLKmode store, so we do not track + them for redundant store elimination. */ + && GET_MODE (reg) != BLKmode /* There are no REG_INC notes for SP, so we can't assume we'll see everything that invalidates it. To be safe, don't eliminate any stores though SP; none of them should be redundant anyway. */ |