Bug 47988 - ICE: in reload_cse_simplify_operands, at postreload.c:403: insn does not satisfy its constraints
Summary: ICE: in reload_cse_simplify_operands, at postreload.c:403: insn does not sati...
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: target (show other bugs)
Version: 4.6.0
: P3 normal
Target Milestone: ---
Assignee: Not yet assigned to anyone
URL:
Keywords: ice-on-valid-code
Depends on:
Blocks: 45689
  Show dependency tree / graph
 
Reported: 2011-03-04 13:27 UTC by Zdenek Sojka
Modified: 2018-01-04 23:38 UTC (History)
0 users

See Also:
Host: x86_64-pc-linux-gnu
Target: x86_64-pc-linux-gnu
Build:
Known to work:
Known to fail: 4.6.0
Last reconfirmed: 2017-12-04 00:00:00

Attachments
reduced testcase (292 bytes, text/plain)
2011-03-04 13:27 UTC, Zdenek Sojka 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Zdenek Sojka 2011-03-04 13:27:56 UTC 
Created attachment 23542 [details]
reduced testcase

Compiler output:
$ gcc -O2 -fira-algorithm=priority -fpeel-loops -fno-split-wide-types -fno-tree-ter testcase.f90
testcase.f90: In function 'test':
testcase.f90:20:0: error: insn does not satisfy its constraints:
(insn 614 613 615 38 (set (reg:TI 44 r15 [orig:248 D.1840 ] [248])
        (mem/c:TI (reg/f:DI 7 sp) [0 S16 A128])) testcase.f90:16 60 {*movti_internal_rex64}
     (nil))
testcase.f90:20:0: internal compiler error: in reload_cse_simplify_operands, at postreload.c:403
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Two files from the testsuite fail the same way, with the same flags:
gfortran.dg/cshift_large_1.f90
gfortran.dg/eoshift_large_1.f90

Tested revisions:
r170622 - crash
4.5 r170013 - OK
Comment 1 Dominique d'Humieres 2017-12-04 09:01:09 UTC 
Are sure that the expected values after cshift are right?

Is the problem still present?
Comment 2 Zdenek Sojka 2017-12-13 14:39:18 UTC 
(In reply to Dominique d'Humieres from comment #1)
> Are sure that the expected values after cshift are right?
> 
> Is the problem still present?

The testcase was reduced from testsuite/gfortran.dg/cshift_large_1.f90

I can not reproduce it anymore with trunk.
Comment 3 Dominique d'Humieres 2018-01-04 23:38:06 UTC 
I have tested gfortran.dg/cshift_large_1.f90 with/without the options in comment 0 without any problem from 4.8 up to trunk. Closing as FIXED.

BTW I think expected values after cshift in the attached test are wrong.