#1515 closed defect (fixed)
[ARM/Thumb] Invalid stack frame usage at -Os
Reported by: | Sebastian Huber | Owned by: | Ralf Corsepius |
---|---|---|---|
Priority: | normal | Milestone: | 4.10 |
Component: | tool/gcc | Version: | 4.9 |
Severity: | normal | Keywords: | |
Cc: | joel.sherrill@… | Blocked By: | |
Blocking: |
Description (last modified by Gedare Bloom)
We should wait with the RTEMS 4.10 release until we have a workaround or fix for this GCC bug:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44091
The bug vanishes if you change the optimization level or use ARM code instead of Thumb code. Who knows if this really prevents such failures?
Change History (4)
comment:1 Changed on 05/12/10 at 11:43:38 by Ralf Corsepius
Cc: | Joel Sherrill added |
---|
comment:2 Changed on 05/18/10 at 10:43:33 by Sebastian Huber
Resolution: | → fixed |
---|---|
Status: | new → closed |
comment:3 Changed on 11/22/14 at 20:07:54 by Gedare Bloom
Description: | modified (diff) |
---|---|
Milestone: | 2 → 4.10 |
Version: | unknown → 4.9 |
comment:4 Changed on 10/10/17 at 05:58:26 by Sebastian Huber
Component: | GCC → tool/gcc |
---|
Note: See
TracTickets for help on using
tickets.
It seems that -fno-schedule-insns2 is a workaround. At least the test case generates valid code after disabling this optimization.