https://bugs.kde.org/show_bug.cgi?id=433500
--- Comment #4 from Mark Wielaard <m...@klomp.org> --- (In reply to Paul Floyd from comment #3) > With the suggested changes, for DRD I get, when running > > /usr/bin/time -p perl tests/vg_regtest drd/tests > > == 130 tests, 1 stderr failure, 0 stdout failures, 0 stderrB failures, 0 > stdoutB failures, 0 post failures == > drd/tests/pth_cleanup_handler (stderr) > > The above testcase that fails generates > > +Process terminating with default action of signal 14 (SIGALRM) > + at 0x........: (within libpthread-?.?.so) > + by 0x........: pthread_once (drd_pthread_intercepts.c:?) > + by 0x........: UnknownInlinedFun (gthr-default.h:700) > + by 0x........: uw_init_context_1 (unwind-dw2.c:1598) > + by 0x........: _Unwind_ForcedUnwind (unwind.inc:211) > + by 0x........: (within libpthread-?.?.so) > + by 0x........: (within libpthread-?.?.so) > + by 0x........: f (pth_cleanup_handler.c:?) > + by 0x........: vgDrd_thread_wrapper (drd_pthread_intercepts.c:?) > + by 0x........: (within libpthread-?.?.so) > + by 0x........: clone (in /...libc...) > +Destroying condition variable that is being waited upon: cond 0x........, > mutex 0x........ locked by thread 0 > + at 0x........: pthread_cond_wait@@GLIBC_2.3.2(within libpthread-?.?.so) > + by 0x........: pthread_cond_wait (drd_pthread_intercepts.c:?) > + by 0x........: pthread_cond_wait (drd_pthread_intercepts.c:?) > + by 0x........: vgDrd_sema_down (drd_pthread_intercepts.c:?) > + by 0x........: pthread_create (drd_pthread_intercepts.c:?) > + by 0x........: main (pth_cleanup_handler.c:?) > + > + > +ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0) I cannot tell whether that is good or bad. It seems the test doesn't expect this failure mode. But it seems the dwarf reader worked as expected doesn't it? > I also noticed a substantial slowdown Maybe that explains the above test failure? A timeout because valgrind is reading all that debuginfo? > Runtime with change and debuginfo installed > real 2200.95 > user 2085.74 > sys 23.89 > > Without the changes and with debuginfo removed [0 failures] > real 244.29 > user 149.81 > sys 13.18 > > With the changes and still withoug debuginfo [0 failures] > real 245.38 > user 151.01 > sys 13.17 Yes, this is also expected sadly. The dwarf reader is really slow and it is basically reading all the gcc runtime library debuginfo eagerly as soon as libstdc++.so is loaded. That takes a lot of time. -- You are receiving this mail because: You are watching all bug changes.
