Bug 110334

Summary: SPECint2000 255.vortex crashes with -funroll-loops
Product: Red Hat Enterprise Linux 3 Reporter: Mario Held <held>
Component: gcc3Assignee: Jakub Jelinek <jakub>
Status: CLOSED CURRENTRELEASE QA Contact:
Severity: medium Docs Contact:
Priority: medium    
Version: 3.0CC: epasch
Target Milestone: ---   
Target Release: ---   
Hardware: s390   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2004-02-05 16:09:22 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:

Description Mario Held 2003-11-18 16:21:43 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030624

Description of problem:
Run 255.vortex without '-funroll-loops' successfully, afterwards with
the parameter in question which results in an error. Used same
environment for the test and reproduced the behavior twice.




Version-Release number of selected component (if applicable):
RHEL3 gcc-3.2.3-20, glibc-2.3.2-95.3

How reproducible:
Always

Steps to Reproduce:
1.Run SPECint2000 255.vortex with compile option '-funroll-loops' 
2.
3.
    

Actual Results:  specmake  options 2> options.err | tee options.out
COMP: gcc -c -o options.o              -O3 -mno-backchain -funroll-loops
LINK: gcc       -O3 -mno-backchain -funroll-loops      -lm  -o options
Setting Up Run Directories
  Setting up 255.vortex ref base s390_linux default: existing
Running Benchmarks
  Running 255.vortex ref base s390_linux default
Child returned with invalid return code(0)
Child returned with invalid return code(0)
Invalid run; unable to continue.  If you wish to ignore errors please
use '-I' or ignore_errors

Expected Results:  COMP: gcc -c -o options.o              -O3
-mno-backchain
LINK: gcc       -O3 -mno-backchain      -lm  -o options
Setting Up Run Directories
  Setting up 255.vortex ref base s390_linux default: existing
Running Benchmarks
  Running 255.vortex ref base s390_linux default
Success: 1x255.vortex
Producing Reports
mach: default
  ext: s390_linux
    size: ref
      set: int
        format: raw -> /home/bench/cpu2000/result/CINT2000.108.raw
        format: ASCII -> /home/bench/cpu2000/result/CINT2000.108.asc
      set: fp
runspec finished


Additional info:

Can offer binaries and dissasemblies (objdump --disassemble-all) of
the two flavours of the binary.

Comment 1 Mario Held 2004-02-05 16:09:22 UTC
Hello,

I encountered the problem in the recent days in internal drivers with 
gcc-3.4 too. Using ltrace we found the culprit in the SPECint2000 
255.vortex application. 

I tested the one line code change today with RHEL 3.0 and the problem 
is gone. Reported the bug now to Spec.org.