| Submitter | Aldy Hernandez |
|---|---|
| Date | Jan. 18, 2012, 2 p.m. |
| Message ID | <4F16D07D.6040405@redhat.com> |
| Download | mbox | patch |
| Permalink | /patch/136613/ |
| State | New |
| Headers | show |
Comments
On Wed, Jan 18, 2012 at 3:00 PM, Aldy Hernandez <aldyh@redhat.com> wrote: > >>> So you can do probably the following: >>> >>> Index: trans-mem.c >>> =================================================================== >>> --- trans-mem.c (revision 183253) >>> +++ trans-mem.c (working copy) >>> @@ -1497,8 +1497,6 @@ requires_barrier (basic_block entry_block, tree x, >>> to needs_to_live_in_memory until we eliminate >>> lower_sequence_tm altogether. */ >>> needs_to_live_in_memory (x) >>> - /* X escapes. */ >>> - || is_global_var (x)) >> >> >> True, because needs_to_live_in_memory is true for all global vars already >> (and includes all address-taken automatic vars as well, thus >> may_be_aliased (x), >> too). > > > Sweet. Thanks guys. > > Richard, I assume you are OK with this? Yes. Richard. > Tested on x86-64 Linux.
Patch
Index: trans-mem.c =================================================================== --- trans-mem.c (revision 183272) +++ trans-mem.c (working copy) @@ -1496,9 +1496,7 @@ requires_barrier (basic_block entry_bloc during lower_sequence_tm/gimplification, leave the call to needs_to_live_in_memory until we eliminate lower_sequence_tm altogether. */ - needs_to_live_in_memory (x) - /* X escapes. */ - || is_global_var (x)) + needs_to_live_in_memory (x)) return true; else {