Bug 32006 - [4.3 Regression] ICE (segv) in tree-vrp (use of uninitialised values)
Summary: [4.3 Regression] ICE (segv) in tree-vrp (use of uninitialised values)
Status: RESOLVED DUPLICATE of bug 32018
Alias: None
Product: gcc
Classification: Unclassified
Component: middle-end (show other bugs)
Version: 4.3.0
: P3 normal
Target Milestone: 4.3.0
Assignee: Not yet assigned to anyone
URL:
Keywords: ice-on-valid-code
Depends on:
Blocks:
 
Reported: 2007-05-20 09:58 UTC by tobias.burnus
Modified: 2007-05-24 12:47 UTC (History)
5 users (show)

See Also:
Host:
Target:
Build:
Known to work:
Known to fail:
Last reconfirmed:

Attachments
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 tobias.burnus 2007-05-20 09:58:36 UTC 
Using test_fpu.f90 of the Polyhedron testsuite, http://www.polyhedron.co.uk/pb05/polyhedron_benchmark_suite.html, fails since 2007-05-15 (r124736) on x86_64-unknown-linux-gnu, 2007-05-14 (r124708) was still ok. Last tested: 2007-05-20 (r124860).

The ICE occures with:
gfortran -O3 -ftree-vectorize -ffast-math -funroll-loops test_fpu.f90
With -m32 the ICE does not occure.

Running GCC through valgrind shows:

==30881== Conditional jump or move depends on uninitialised value(s)
==30881==    at 0x7A2E38: vrp_evaluate_conditional_warnv (tree-vrp.c:4803)
==30881==    by 0x7A34D9: vrp_evaluate_conditional (tree-vrp.c:4935)
==30881==    by 0x76488B: substitute_and_fold (tree-ssa-propagate.c:1131)
==30881==    by 0x79E418: vrp_finalize (tree-vrp.c:5914)
==30881==    by 0x7A006A: execute_vrp (tree-vrp.c:6000)
==30881==    by 0x609540: execute_one_pass (passes.c:1065)
==30881==    by 0x6096FB: execute_pass_list (passes.c:1117)
==30881==    by 0x60970D: execute_pass_list (passes.c:1118)
==30881==    by 0x6D3337: tree_rest_of_compilation (tree-optimize.c:406)
==30881==    by 0x810D9F: cgraph_expand_function (cgraphunit.c:1086)
==30881==    by 0x813241: cgraph_optimize (cgraphunit.c:1155)
==30881==    by 0x466F5C: gfc_be_parse_file (f95-lang.c:307)
Comment 1 Tobias Burnus 2007-05-22 23:15:54 UTC 
The patch causing this regression is
r124727 | rakdver | 2007-05-14 23:58:42 +0200 (Mo, 14 Mai 2007) | 46 lines
http://gcc.gnu.org/ml/gcc-cvs/2007-05/msg00425.html

Zdenek, could you have a look?
Comment 2 Tobias Burnus 2007-05-22 23:35:22 UTC 
Seems to be fixed by the patch for PR 32018. I will bootstrap and retest tomorrow.
Comment 3 Tobias Burnus 2007-05-24 12:47:25 UTC 

*** This bug has been marked as a duplicate of 32018 ***