Bug 28077

Summary: [4.3 regression] pass39-frag.c produces mudflap violation on alpha
Product: gcc Reporter: Martin Michlmayr <tbm>
Component: libmudflapAssignee: Not yet assigned to anyone <unassigned>
Status: RESOLVED FIXED    
Severity: normal CC: debian-gcc, fche, gcc-bugs, rth
Priority: P5    
Version: 4.1.1   
Target Milestone: 4.3.5   
Host: Target: alpha-linux-gnu
Build: alpha-linux-gnu Known to work:
Known to fail: Last reconfirmed:
Attachments: testsuite logs
more detailed log

Description Martin Michlmayr 2006-06-18 17:16:00 UTC 
libmudflap.cth/pass39-frag.c (and I think pass37-frag.c) produce mudflap violations on Alpha.  In dmesg, you get messages like these:

 pass39-frag.exe(7421): unaligned trap at 0000000120001ff8: b029001c2022000f 29 1 
 pass39-frag.exe(7552): unaligned trap at 0000000120001ff8: b029001c2022000f 29 1
 pass39-frag.exe(7421): unaligned trap at 0000000120001ff8: b029001c2022000f 29 1

I'll attach the relevant portions of libmudflap.log.
Comment 1 Martin Michlmayr 2006-06-18 17:16:30 UTC 
Created attachment 11691 [details]
testsuite logs
Comment 2 Frank Ch. Eigler 2006-06-19 14:01:33 UTC 
It looks like only the statically linked multithreding test cases trigger the problem.  Would you mind trying ot hand-build one of those executables, but adding -rdynamic to LDFLAGS, and run with -backtrace=99 in MUDFLAP_OPTIONS?  That way, the backtraces should have more symbolic information.
Comment 3 Martin Michlmayr 2006-06-19 19:34:42 UTC 
Created attachment 11705 [details]
more detailed log

This is with the options you specified but it seems it doesn't contain so much more information.  Did I do something wrong or is that what you were looking for?
Comment 4 Joseph S. Myers 2008-07-04 21:25:00 UTC 
Closing 4.1 branch.
Comment 5 Uroš Bizjak 2008-12-18 15:31:50 UTC 
This does not fail on 4.4 [1] branch.

[1] http://gcc.gnu.org/ml/gcc-testresults/2008-12/msg01564.html
Comment 6 Uroš Bizjak 2009-02-03 11:18:23 UTC 
Can someone check if this still fails on latest 4.3 branch?

4.4 works OK.
Comment 7 Joseph S. Myers 2009-03-31 19:39:14 UTC 
Closing 4.2 branch.
Comment 8 Richard Biener 2009-08-04 12:27:49 UTC 
GCC 4.3.4 is being released, adjusting target milestone.
Comment 9 Uroš Bizjak 2009-09-29 21:41:16 UTC 
(In reply to comment #6)
> Can someone check if this still fails on latest 4.3 branch?

It doesn't. See [1].

[1] http://gcc.gnu.org/ml/gcc-testresults/2009-09/msg01497.html