Bug #95772 for ifdef: Tests may segfault

Sun May 18 16:03:41 2014 SREZIC [...] cpan.org - Ticket created

Subject:

Tests may segfault

Some of the tests in the test suite may segfault. See http://www.cpantesters.org/cpan/report/981000d0-dc74-11e3-b0f2-7117e1bfc7aa for a sample report. Here's a backtrace of such a failure (sorry, I currently have no perl with debugging symbols available...) Core was generated by `/opt/perl-5.20.0-RC1/bin/perl t/ifdef05.t'. Program terminated with signal 11, Segmentation fault. #0 0x00000000004b4448 in Perl_sv_upgrade () (gdb) bt #0 0x00000000004b4448 in Perl_sv_upgrade () #1 0x00000000004b27e8 in Perl_sv_setsv_flags () #2 0x00000000004a7bdc in Perl_pp_aassign () #3 0x00000000004a4e03 in Perl_runops_standard () #4 0x0000000000440024 in perl_run () #5 0x0000000000422285 in main () Regards, Slaven

Thu May 22 22:02:43 2014 ANDK [...] cpan.org - Correspondence added

I've tried to bisect this one. First condition to get the segv is threaded perl, second condition is Test::More >= 1.99. Bisect leads us to v5.19.4-224-g411e93c which was the "Upgrade Test-Simple from version 0.98 to 0.99". A debugging stack trace is: (gdb) bt #0 0x00000000004b7b99 in Perl_sv_upgrade (my_perl=my_perl@entry=0x1b7d010, sv=sv@entry=0x1dbe2e8, new_type=SVt_PV) at sv.c:1355 #1 0x00000000004b611b in Perl_sv_setsv_flags ( my_perl=my_perl@entry=0x1b7d010, dstr=dstr@entry=0x1dbe2e8, sstr=0x1ee7ab8, flags=flags@entry=1536) at sv.c:4134 #2 0x00000000004aaf99 in Perl_pp_aassign (my_perl=0x1b7d010) at pp_hot.c:1065 #3 0x00000000004a7d76 in Perl_runops_standard (my_perl=0x1b7d010) at run.c:42 #4 0x000000000043dea4 in S_run_body (oldscope=1, my_perl=0x1b7d010) at perl.c:2500 #5 perl_run (my_perl=0x1b7d010) at perl.c:2416 #6 0x000000000041e27b in main (argc=2, argv=0x7fff01986758, env=0x7fff01986770) at perlmain.c:112

Thu May 22 22:02:43 2014 The RT System itself - Status changed from 'new' to 'open'

Thu May 22 22:15:52 2014 ANDK [...] cpan.org - Correspondence added

Stacktrace redone with 5.20.0-RC1 (the above was 5.19.4): (gdb) bt #0 0x00000000004b9d19 in Perl_sv_upgrade (my_perl=my_perl@entry=0x1ac8010, sv=sv@entry=0x1e51660, new_type=SVt_PV) at sv.c:1427 #1 0x00000000004b821b in Perl_sv_setsv_flags ( my_perl=my_perl@entry=0x1ac8010, dstr=dstr@entry=0x1e51660, sstr=0x1d12b48, flags=flags@entry=1536) at sv.c:4203 #2 0x00000000004accef in Perl_pp_aassign (my_perl=0x1ac8010) at pp_hot.c:1071 #3 0x00000000004a9c96 in Perl_runops_standard (my_perl=0x1ac8010) at run.c:42 #4 0x000000000043cf94 in S_run_body (oldscope=1, my_perl=0x1ac8010) at perl.c:2456 #5 perl_run (my_perl=0x1ac8010) at perl.c:2372 #6 0x000000000041be1b in main (argc=2, argv=0x7ffff3a20548, env=0x7ffff3a20560) at perlmain.c:114 (gdb)

Fri May 23 01:30:21 2014 smueller [...] cpan.org - Correspondence added

On Thu May 22 22:15:52 2014, ANDK wrote: Show quoted text

> Stacktrace redone with 5.20.0-RC1 (the above was 5.19.4): > > (gdb) bt > #0 0x00000000004b9d19 in Perl_sv_upgrade > (my_perl=my_perl@entry=0x1ac8010, > sv=sv@entry=0x1e51660, new_type=SVt_PV) at sv.c:1427 > #1 0x00000000004b821b in Perl_sv_setsv_flags ( > my_perl=my_perl@entry=0x1ac8010, dstr=dstr@entry=0x1e51660, > sstr=0x1d12b48, flags=flags@entry=1536) at sv.c:4203 > #2 0x00000000004accef in Perl_pp_aassign (my_perl=0x1ac8010) at > pp_hot.c:1071 > #3 0x00000000004a9c96 in Perl_runops_standard (my_perl=0x1ac8010) at > run.c:42 > #4 0x000000000043cf94 in S_run_body (oldscope=1, my_perl=0x1ac8010) > at perl.c:2456 > #5 perl_run (my_perl=0x1ac8010) at perl.c:2372 > #6 0x000000000041be1b in main (argc=2, argv=0x7ffff3a20548, > env=0x7ffff3a20560) at perlmain.c:114 > (gdb)

I briefly tried to reproduce this just now but while I did see failing tests, I could not get a crash with a bleadperl from last night. --Steffen

Fri May 23 01:43:27 2014 SREZIC [...] cpan.org - Correspondence added

On 2014-05-23 01:30:21, SMUELLER wrote: Show quoted text

> On Thu May 22 22:15:52 2014, ANDK wrote:

> > Stacktrace redone with 5.20.0-RC1 (the above was 5.19.4): > > > > (gdb) bt > > #0 0x00000000004b9d19 in Perl_sv_upgrade > > (my_perl=my_perl@entry=0x1ac8010, > > sv=sv@entry=0x1e51660, new_type=SVt_PV) at sv.c:1427 > > #1 0x00000000004b821b in Perl_sv_setsv_flags ( > > my_perl=my_perl@entry=0x1ac8010, dstr=dstr@entry=0x1e51660, > > sstr=0x1d12b48, flags=flags@entry=1536) at sv.c:4203 > > #2 0x00000000004accef in Perl_pp_aassign (my_perl=0x1ac8010) at > > pp_hot.c:1071 > > #3 0x00000000004a9c96 in Perl_runops_standard (my_perl=0x1ac8010) at > > run.c:42 > > #4 0x000000000043cf94 in S_run_body (oldscope=1, my_perl=0x1ac8010) > > at perl.c:2456 > > #5 perl_run (my_perl=0x1ac8010) at perl.c:2372 > > #6 0x000000000041be1b in main (argc=2, argv=0x7ffff3a20548, > > env=0x7ffff3a20560) at perlmain.c:114 > > (gdb)

> > I briefly tried to reproduce this just now but while I did see failing > tests, I could not get a crash with a bleadperl from last night.

Actually I don't think it's a bleadperl issue. Here's a report on a 5.14.2 box with a similar pattern: http://www.cpantesters.org/cpan/report/c45cbc06-7ff0-11e1-9319-76eb9a551a3f Regards, Slaven

Fri May 23 03:15:45 2014 ANDK [...] cpan.org - Correspondence added

I did another bisect with the precondition uptodate Test::More met and it ended with v5.19.2-126-g706a6eb. Quite possible that this is a red herring if nobody else can reproduce it.

Sat Mar 26 03:59:24 2016 SREZIC [...] cpan.org - Correspondence added

On 2014-05-18 16:03:41, SREZIC wrote: Show quoted text

> Some of the tests in the test suite may segfault. See > http://www.cpantesters.org/cpan/report/981000d0-dc74-11e3-b0f2- > 7117e1bfc7aa for a sample report. > > Here's a backtrace of such a failure (sorry, I currently have no perl > with debugging symbols available...) > > Core was generated by `/opt/perl-5.20.0-RC1/bin/perl t/ifdef05.t'. > Program terminated with signal 11, Segmentation fault. > #0 0x00000000004b4448 in Perl_sv_upgrade () > (gdb) bt > #0 0x00000000004b4448 in Perl_sv_upgrade () > #1 0x00000000004b27e8 in Perl_sv_setsv_flags () > #2 0x00000000004a7bdc in Perl_pp_aassign () > #3 0x00000000004a4e03 in Perl_runops_standard () > #4 0x0000000000440024 in perl_run () > #5 0x0000000000422285 in main () >

Just stumbled over this again while smoking for perl 5.23.9: Test Summary Report ------------------- t/ifdef05.t (Wstat: 11 Tests: 2 Failed: 0) Non-zero wait status: 11 t/ifdef05a.t (Wstat: 11 Tests: 2 Failed: 0) Non-zero wait status: 11 t/ifdef05b.t (Wstat: 11 Tests: 2 Failed: 0) Non-zero wait status: 11 t/ifdef05c.t (Wstat: 11 Tests: 2 Failed: 0) Non-zero wait status: 11 Files=9, Tests=132, 2 wallclock secs ( 0.02 usr 0.07 sys + 0.61 cusr 0.95 csys = 1.65 CPU) Result: FAIL Failed 4/9 test programs. 0/132 subtests failed. make: *** [test_dynamic] Fehler 255

Sat Mar 26 07:03:11 2016 liz [...] dijkmat.nl - Correspondence added

Subject:	Re: [rt.cpan.org #95772] Tests may segfault
Date:	Sat, 26 Mar 2016 12:02:50 +0100
To:	bug-ifdef [...] rt.cpan.org
From:	Elizabeth Mattijsen <liz [...] dijkmat.nl>

Sorry, I’m no longer supporting any of my Perl 5 modules. They are on ADOPT-ME status. Show quoted text

> On 26 Mar 2016, at 08:59, Slaven_Rezic via RT <bug-ifdef@rt.cpan.org> wrote: > > Queue: ifdef > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=95772 > > > On 2014-05-18 16:03:41, SREZIC wrote:

>> Some of the tests in the test suite may segfault. See >> http://www.cpantesters.org/cpan/report/981000d0-dc74-11e3-b0f2- >> 7117e1bfc7aa for a sample report. >> >> Here's a backtrace of such a failure (sorry, I currently have no perl >> with debugging symbols available...) >> >> Core was generated by `/opt/perl-5.20.0-RC1/bin/perl t/ifdef05.t'. >> Program terminated with signal 11, Segmentation fault. >> #0 0x00000000004b4448 in Perl_sv_upgrade () >> (gdb) bt >> #0 0x00000000004b4448 in Perl_sv_upgrade () >> #1 0x00000000004b27e8 in Perl_sv_setsv_flags () >> #2 0x00000000004a7bdc in Perl_pp_aassign () >> #3 0x00000000004a4e03 in Perl_runops_standard () >> #4 0x0000000000440024 in perl_run () >> #5 0x0000000000422285 in main () >>

> > Just stumbled over this again while smoking for perl 5.23.9: > > Test Summary Report > ------------------- > t/ifdef05.t (Wstat: 11 Tests: 2 Failed: 0) > Non-zero wait status: 11 > t/ifdef05a.t (Wstat: 11 Tests: 2 Failed: 0) > Non-zero wait status: 11 > t/ifdef05b.t (Wstat: 11 Tests: 2 Failed: 0) > Non-zero wait status: 11 > t/ifdef05c.t (Wstat: 11 Tests: 2 Failed: 0) > Non-zero wait status: 11 > Files=9, Tests=132, 2 wallclock secs ( 0.02 usr 0.07 sys + 0.61 cusr 0.95 csys = 1.65 CPU) > Result: FAIL > Failed 4/9 test programs. 0/132 subtests failed. > make: *** [test_dynamic] Fehler 255 > >

Sun Nov 10 04:25:30 2019 SREZIC [...] cpan.org - Broken in 0.10 added