Bug 886891

Summary: FTBFS: lots of valgrind failures during self checks
Product: [Fedora] Fedora Reporter: Karsten Hopp <karsten>
Component: curlAssignee: Kamil Dudka <kdudka>
Status: CLOSED CURRENTRELEASE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: high Docs Contact:
Priority: medium    
Version: rawhideCC: kdudka, paul
Target Milestone: ---   
Target Release: ---   
Hardware: powerpc   
OS: Linux   
Whiteboard:
Fixed In Version: curl-7.28.1-3.fc19 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-01-20 20:35:56 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
root.log
none
build.log none

Description Karsten Hopp 2012-12-13 13:53:41 UTC 
Created attachment 662971 [details]
root.log

Description of problem:
curl-7.28.1-1.fc19 fails to build on PPC64, PPC(32) is ok.
All failures look similar to this one:

 ==65518== Memcheck, a memory error detector
 ==65518== Copyright (C) 2002-2012, and GNU GPL'd, by Julian Seward et al.
 ==65518== Using Valgrind-3.8.1 and LibVEX; rerun with -h for copyright info
 ==65518== Command: /builddir/build/BUILD/curl-7.28.1/tests/libtest/.libs/lt-libntlmconnect http://127.0.0.1:6490/2032
 ==65518== Parent PID: 65517
 ==65518== 
 ==65518== Jump to the invalid address stated on the next line
 ==65518==    at 0x0: ???
 ==65518==    by 0x4023827: __assert_fail (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x400AC37: _dl_map_object_from_fd (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x400CCBB: _dl_map_object (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x40043CF: map_doit (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x4017F23: _dl_catch_error (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x40036D7: do_preload (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x4007DE7: dl_main (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x4022A9B: _dl_sysdep_start (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x40041AB: _dl_start_final (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x400830F: _dl_start (in /usr/lib64/ld-2.16.90.so)
 ==65518==    by 0x4003E0F: _start (in /usr/lib64/ld-2.16.90.so)
 ==65518==  Address 0x0 is not stack'd, malloc'd or (recently) free'd
...
TESTDONE: 45 tests out of 682 reported OK: 6%
...


Version-Release number of selected component (if applicable):
curl-7.28.1-1.fc19

How reproducible:


Steps to Reproduce:
1. ppc-koji build --scratch f19 curl-7.28.1-1.fc19.src.rpm
2.
3.
  
Actual results:
http://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=820237

Expected results:


Additional info:
Comment 1 Karsten Hopp 2012-12-13 13:55:23 UTC 
Created attachment 662972 [details]
build.log
Comment 2 Kamil Dudka 2012-12-13 15:46:10 UTC 
I will disable valgrind for ppc64.  Please keep it disabled.  If anybody needs to debug the ppc64 tool chain, it can be done off the build system.  Build system is intended for building packages in the first place.
Comment 3 Karsten Hopp 2012-12-14 12:42:05 UTC 
Further testing points at glibc as a possible cause for the invalid address issues. glibc-2.16.90-25.fc19.ppc64 is ok, glibc-2.16.90-30.fc19.ppc64 and newer show this issue.
With glibc-2.16.90-25 only test 24 fails and that doesn't seem to be curl's fault:
 output (log/stderr24) FAILED:
--- log/check-expected  2012-12-14 13:04:10.872609526 +0100
+++ log/check-generated 2012-12-14 13:04:10.872609526 +0100
@@ -1 +1,2 @@
+ls: relocation error: ls: symbol clock_gettime, version GLIBC_2.17 not defined in file libc.so.6 with link time reference
Comment 4 Kamil Dudka 2013-01-15 14:09:36 UTC 
fixed in git:

http://pkgs.fedoraproject.org/cgit/curl.git/commit/?id=32b2d061