Valgrind results Dec 11: new quartz/filtergraph crash (oh, Lei...)

[Dan Kegel]

New results at
http://kegel.com/wine/valgrind/logs-2007-12-11-summary.txt
The changes from yesterday are probably noise except:

The tests crashed and valgrind hung while running the
quartz/filtergraph test; see
http://kegel.com/wine/valgrind/logs-2007-12-11/vg-quartz_filtergraph-diff.tx...

+Unhandled exception: page fault on read access to 0xffffffff in
32-bit code (0x045e1290).
+Backtrace:
+=>1 0x045e1290 func_filtergraph+0x110()
[/home/dank/wine-git/dlls/quartz/tests/filtergraph.c:161] in
quartz_test (0x7f22fe58)
+  2 0x045e20e8 run_test+0x128(name=0x7f0003f7)
[/home/dank/wine-git/dlls/quartz/tests/../../../include/wine/test.h:389]
in quartz_test (0x7f22fea8)
+  3 0x045e277d main+0x14d(argc=<register ECX not in topmost frame>,
argv=<register ECX not in topmost frame>)
[/home/dank/wine-git/dlls/quartz/tests/../../../include/wine/test.h:437]
in quartz_test (0x7f22fed8)
+  4 0x045e284b __wine_spec_exe_entry+0x5b(peb=0x7ffdf000)
[/home/dank/wine-git/dlls/winecrt0/exe_entry.c:36] in quartz_test
(0x7f22ff08)
+  5 0x045198be start_process+0xee(arg=0x0)
[/home/dank/wine-git/dlls/kernel32/process.c:839] in kernel32
(0x7f22ffe8)
+  6 0x0402c9e7 wine_switch_to_stack+0x17() in libwine.so.1 (0x00000000)
+0x045e1290 func_filtergraph+0x110
[/home/dank/wine-git/dlls/quartz/tests/filtergraph.c:161] in
quartz_test: movl	0x0(%eax),%edx
+161	    hr = IFilterGraph2_Release(pF);

Lei, could you have a look?  You were just in here...
- Dan

[Dmitry Timoshkov]

"Dan Kegel"  wrote:
> New results at
> http://kegel.com/wine/valgrind/logs-2007-12-11-summary.txt
> The changes from yesterday are probably noise except:

I wonder why 
http://kegel.com/wine/valgrind/logs-2007-12-11/vg-kernel32_virtual.txt
still shows NTDLL_queue_process_apc as an offender since Alexandre committed
my patch which supposedly should fix that.

[Lei Zhang]

On Dec 11, 2007 8:11 PM, Dan Kegel  wrote:
> New results at
> http://kegel.com/wine/valgrind/logs-2007-12-11-summary.txt
> The changes from yesterday are probably noise except:
>
> The tests crashed and valgrind hung while running the
> quartz/filtergraph test; see
> http://kegel.com/wine/valgrind/logs-2007-12-11/vg-quartz_filtergraph-diff.tx...
>
> +Unhandled exception: page fault on read access to 0xffffffff in
> 32-bit code (0x045e1290).
> +Backtrace:
> +=>1 0x045e1290 func_filtergraph+0x110()
> [/home/dank/wine-git/dlls/quartz/tests/filtergraph.c:161] in
> quartz_test (0x7f22fe58)
> - 2 0x045e20e8 run_test+0x128(name=0x7f0003f7)
>
> [/home/dank/wine-git/dlls/quartz/tests/../../../include/wine/test.h:389]
> in quartz_test (0x7f22fea8)
> - 3 0x045e277d main+0x14d(argc=<register ECX not in topmost frame>,
>
> argv=<register ECX not in topmost frame>)
> [/home/dank/wine-git/dlls/quartz/tests/../../../include/wine/test.h:437]
> in quartz_test (0x7f22fed8)
> - 4 0x045e284b __wine_spec_exe_entry+0x5b(peb=0x7ffdf000)
>
> [/home/dank/wine-git/dlls/winecrt0/exe_entry.c:36] in quartz_test
> (0x7f22ff08)
> - 5 0x045198be start_process+0xee(arg=0x0)
>
> [/home/dank/wine-git/dlls/kernel32/process.c:839] in kernel32
> (0x7f22ffe8)
> - 6 0x0402c9e7 wine_switch_to_stack+0x17() in libwine.so.1 (0x00000000)
>
> +0x045e1290 func_filtergraph+0x110
> [/home/dank/wine-git/dlls/quartz/tests/filtergraph.c:161] in
> quartz_test: movl       0x0(%eax),%edx
> +161        hr = IFilterGraph2_Release(pF);
>
> Lei, could you have a look?  You were just in here...
> - Dan

It looks as though you're running the current test with an older quartz.dll.so.

http://kegel.com/wine/valgrind/logs-2007-12-11/vg-quartz_filtergraph.txt says:

fixme:quartz:Filtergraph_QueryInterface unknown interface
{36b73882-c2c8-11cf-8b46-00805f6cef60}

the patch that fixed that also added the test that caused your crash.

[Dan Kegel]

On Dec 11, 2007 11:34 PM, Lei Zhang  wrote:
> It looks as though you're running the current test with an older quartz.dll.so.

I'd better add a 'make clean' to my automated script.
And for that matter, a "rm -rf ~/.wine".
- Dan