http://bugs.winehq.org/show_bug.cgi?id=20917
Summary: tests crash in init_format_filter_info() under valgrind Product: Wine Version: 1.1.33 Platform: PC OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: dank@kegel.com
The wine test suite reliably crashes under valgrind in the function init_format_filter_info() in the tests d3d8/device d3d9/device ddrawex/surface See e.g. http://kegel.com/wine/valgrind/logs/2009-12-03-17.17/vg-d3d8_device.txt http://kegel.com/wine/valgrind/logs/2009-12-03-17.17/vg-ddrawex_surface.txt It looks like a wild pointer bug in the nvidia driver, but it could also be a problem with valgrind or wine.
If it's a driver bug, we should report it upstream after testing with the latest driver. If it's not a valgrind bug, we should evaluate whether we should be calling the driver more carefully somehow.
I'll attach an nvidia bug report log for completeness.
http://bugs.winehq.org/show_bug.cgi?id=20917
--- Comment #1 from Dan Kegel dank@kegel.com 2009-12-05 02:28:10 --- I'm going to stop running these three tests in valgrind for now by touching the corresponding 'ok' files in the script; I'll leave a comment pointing to this bug in the test script.
http://bugs.winehq.org/show_bug.cgi?id=20917
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, source, testcase
http://bugs.winehq.org/show_bug.cgi?id=20917
--- Comment #2 from Dan Kegel dank@kegel.com 2009-12-05 12:15:54 --- Henri says it's probably a disagreement between valgrind and the driver: http://www.winehq.org/pipermail/wine-devel/2009-December/080292.html
http://bugs.winehq.org/show_bug.cgi?id=20917
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |UPSTREAM
--- Comment #3 from Austin English austinenglish@gmail.com 2012-04-01 23:07:22 CDT --- No longer in 295.33 (current nvidia drivers).
There are two other warnings now: ==13781== Conditional jump or move depends on uninitialised value(s) ==13781== at 0x61EB666: ??? (in /usr/lib/libnvidia-glcore.so.295.33) ==13781== Uninitialised value was created by a stack allocation ==13781== at 0x61EADC4: ??? (in /usr/lib/libnvidia-glcore.so.295.33) ==13781== { <insert_a_suppression_name_here> Memcheck:Cond obj:/usr/lib/libnvidia-glcore.so.295.33 }
and
==13781== 120 bytes in 1 blocks are definitely lost in loss record 588 of 925 ==13781== at 0x4024F20: malloc (vg_replace_malloc.c:236) ==13781== by 0x41F5B6D: getdelim (iogetdelim.c:68) ==13781== by 0x6C792AA: ??? (in /dev/nvidia0) ==13781== by 0x6C8276C: ??? (in /dev/nvidia0) ==13781== by 0x6C710F3: ??? (in /dev/nvidia0) ==13781== by 0x400DC0B: call_init (dl-init.c:70) ==13781== by 0x400DD28: _dl_init (dl-init.c:134) ==13781== by 0x4011DE8: dl_open_worker (dl-open.c:463) ==13781== by 0x400D835: _dl_catch_error (dl-error.c:178) ==13781== by 0x4011635: _dl_open (dl-open.c:554) ==13781== by 0x42F2C0A: dlopen_doit (dlopen.c:67) ==13781== by 0x400D835: _dl_catch_error (dl-error.c:178) ==13781== by 0x42F309B: _dlerror_run (dlerror.c:164) ==13781== by 0x42F2B40: dlopen@@GLIBC_2.1 (dlopen.c:88) ==13781== by 0x4FE8DDD: ??? (in /usr/lib/libGL.so.295.33) ==13781== { <insert_a_suppression_name_here> Memcheck:Leak fun:malloc fun:getdelim obj:/dev/nvidia0 obj:/dev/nvidia0 obj:/dev/nvidia0 fun:call_init fun:_dl_init fun:dl_open_worker fun:_dl_catch_error fun:_dl_open fun:dlopen_doit fun:_dl_catch_error fun:_dlerror_run fun:dlopen@@GLIBC_2.1 obj:/usr/lib/libGL.so.295.33 }
http://bugs.winehq.org/show_bug.cgi?id=20917
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #4 from Austin English austinenglish@gmail.com 2012-04-01 23:07:41 CDT --- This was in 1.5.1, btw.
-
wine-bugs＠winehq.org