[prev in list] [next in list] [prev in thread] [next in thread] 

List:       wine-devel
Subject:    Re: Lots of failures in the new WoW 32-bit tests
From:       Bernhard_Kölbl <besentv () gmail ! com>
Date:       2023-06-18 11:07:24
Message-ID: CAAWc_oKxgXU1j7BV+hUXNxkw5vb-tgq3uC=9GvEaXaXrbTZnGg () mail ! gmail ! com
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]


Hi,

i think the failures in mf, quartz and msvfw are related to wow64 tunks not
existing in winegstreamer.

--
Bernhard

Le dim. 18 juin 2023, 11:21, Francois Gouget <fgouget@codeweavers.com> a
écrit :

>
> So as far as I understand all that's needed to build Wine with the
> new Windows-on-Windows support is to do:
>
>   ./configure --enable-archs=i386,x86_64
>   make
>
> But then when I run WineTest I get 60+ failures instead of the usual ~16
> on my machine.
>
> https://test.winehq.org/queue/err5z_5H/report
> https://test.winehq.org/queue/errloGNj/report
>
> That does not entirely surprise me since this mode has never been tested
> (neither GitLab nor the TestBot support it). But did I miss something?
>
>
> For instance one failure I get is with devenum:devenum:
>
> devenum:devenum start dlls/devenum/tests/devenum.c
> devenum.c:691: Test failed: Got hr 0x80004005.
> Unhandled exception: page fault on read access to 0x00000000 in 32-bit
> code (0x00405a75).
>
> The reason for the crash is that wg_transform_create() in main.c returns
> NULL:
>
>      if (WINE_UNIX_CALL(unix_wg_transform_create, &params))
>          return NULL;  // <---- we return NULL here
>
> without actually calling the implementation in wg_transform.c as far as
> I can see.
>
>
>
>
> Here's the list of test units that fail in one of the runs. Lots of
> Direct3D (but not all) and multimedia (GStreamer?) tests. Maybe there
> are some quick gains to be had?
> (then there's msvcrt:file which gets -2 as a file descriptor)
>
> advapi32:security
> amstream:amstream
> comctl32:subclass
> d2d1:d2d1
> d3d10core:d3d10core
> d3d11:d3d11
> d3d8:visual
> d3d9:visual
> d3dcompiler_43:hlsl_d3d11
> d3dcompiler_43:hlsl_d3d9
> d3dcompiler_46:hlsl_d3d11
> d3dcompiler_46:hlsl_d3d9
> d3dcompiler_47:hlsl_d3d11
> d3dcompiler_47:hlsl_d3d9
> d3dx10_34:d3dx10
> d3dx10_35:d3dx10
> d3dx10_36:d3dx10
> d3dx10_37:d3dx10
> d3dx10_38:d3dx10
> d3dx10_39:d3dx10
> d3dx10_40:d3dx10
> d3dx10_41:d3dx10
> d3dx10_42:d3dx10
> d3dx10_43:d3dx10
> d3dx9_36:core
> d3dx9_36:math
> d3dx9_36:texture
> d3dx9_36:volume
> dbghelp:dbghelp
> ddraw:d3d
> ddraw:visual
> devenum:devenum
> explorer.exe:explorer
> gdi32:font
> iphlpapi:iphlpapi
> kernel32:debugger
> kernel32:thread
> kernel32:virtual
> mfmediaengine:mfmediaengine
> mfplay:mfplay
> mf:mf
> mf:transform
> msvcrt:file
> msvfw32:mciwnd
> ntdll:info
> ntdll:virtual
> ntdll:wow64
> opengl32:opengl
> qasf:asfreader
> qedit:mediadet
> quartz:filtergraph
> quartz:vmr7
> quartz:vmr9
> riched20:editor
> user32:class
> user32:clipboard
> user32:msg
> user32:win
> win32u:win32u
> wininet:internet
> winmm:mci
> wmp:media
> wmvcore:wmvcore
>
> --
> Francois Gouget <fgouget@codeweavers.com>
>
>
>

[Attachment #5 (text/html)]

<div dir="auto">Hi,<div dir="auto"><br></div><div dir="auto">i think the failures in \
mf, quartz and msvfw are related to wow64 tunks not existing in \
winegstreamer.</div><div dir="auto"><br></div><div dir="auto">--</div><div \
dir="auto">Bernhard  </div></div><br><div class="gmail_quote"><div dir="ltr" \
class="gmail_attr">Le dim. 18 juin 2023, 11:21, Francois Gouget &lt;<a \
href="mailto:fgouget@codeweavers.com">fgouget@codeweavers.com</a>&gt; a écrit  \
:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px \
#ccc solid;padding-left:1ex"><br> So as far as I understand all that&#39;s needed to \
build Wine with the <br> new Windows-on-Windows support is to do:<br>
<br>
   ./configure --enable-archs=i386,x86_64<br>
   make<br>
<br>
But then when I run WineTest I get 60+ failures instead of the usual ~16 <br>
on my machine.<br>
<br>
<a href="https://test.winehq.org/queue/err5z_5H/report" rel="noreferrer noreferrer" \
target="_blank">https://test.winehq.org/queue/err5z_5H/report</a><br> <a \
href="https://test.winehq.org/queue/errloGNj/report" rel="noreferrer noreferrer" \
target="_blank">https://test.winehq.org/queue/errloGNj/report</a><br> <br>
That does not entirely surprise me since this mode has never been tested <br>
(neither GitLab nor the TestBot support it). But did I miss something?<br>
<br>
<br>
For instance one failure I get is with devenum:devenum:<br>
<br>
devenum:devenum start dlls/devenum/tests/devenum.c<br>
devenum.c:691: Test failed: Got hr 0x80004005.<br>
Unhandled exception: page fault on read access to 0x00000000 in 32-bit code \
(0x00405a75).<br> <br>
The reason for the crash is that wg_transform_create() in main.c returns <br>
NULL: <br>
<br>
        if (WINE_UNIX_CALL(unix_wg_transform_create, &amp;params))<br>
              return NULL;   // &lt;---- we return NULL here<br>
<br>
without actually calling the implementation in wg_transform.c as far as <br>
I can see.<br>
<br>
<br>
<br>
<br>
Here&#39;s the list of test units that fail in one of the runs. Lots of <br>
Direct3D (but not all) and multimedia (GStreamer?) tests. Maybe there <br>
are some quick gains to be had?<br>
(then there&#39;s msvcrt:file which gets -2 as a file descriptor)<br>
<br>
advapi32:security<br>
amstream:amstream<br>
comctl32:subclass<br>
d2d1:d2d1<br>
d3d10core:d3d10core<br>
d3d11:d3d11<br>
d3d8:visual<br>
d3d9:visual<br>
d3dcompiler_43:hlsl_d3d11<br>
d3dcompiler_43:hlsl_d3d9<br>
d3dcompiler_46:hlsl_d3d11<br>
d3dcompiler_46:hlsl_d3d9<br>
d3dcompiler_47:hlsl_d3d11<br>
d3dcompiler_47:hlsl_d3d9<br>
d3dx10_34:d3dx10<br>
d3dx10_35:d3dx10<br>
d3dx10_36:d3dx10<br>
d3dx10_37:d3dx10<br>
d3dx10_38:d3dx10<br>
d3dx10_39:d3dx10<br>
d3dx10_40:d3dx10<br>
d3dx10_41:d3dx10<br>
d3dx10_42:d3dx10<br>
d3dx10_43:d3dx10<br>
d3dx9_36:core<br>
d3dx9_36:math<br>
d3dx9_36:texture<br>
d3dx9_36:volume<br>
dbghelp:dbghelp<br>
ddraw:d3d<br>
ddraw:visual<br>
devenum:devenum<br>
explorer.exe:explorer<br>
gdi32:font<br>
iphlpapi:iphlpapi<br>
kernel32:debugger<br>
kernel32:thread<br>
kernel32:virtual<br>
mfmediaengine:mfmediaengine<br>
mfplay:mfplay<br>
mf:mf<br>
mf:transform<br>
msvcrt:file<br>
msvfw32:mciwnd<br>
ntdll:info<br>
ntdll:virtual<br>
ntdll:wow64<br>
opengl32:opengl<br>
qasf:asfreader<br>
qedit:mediadet<br>
quartz:filtergraph<br>
quartz:vmr7<br>
quartz:vmr9<br>
riched20:editor<br>
user32:class<br>
user32:clipboard<br>
user32:msg<br>
user32:win<br>
win32u:win32u<br>
wininet:internet<br>
winmm:mci<br>
wmp:media<br>
wmvcore:wmvcore<br>
<br>
-- <br>
Francois Gouget &lt;<a href="mailto:fgouget@codeweavers.com" target="_blank" \
rel="noreferrer">fgouget@codeweavers.com</a>&gt;<br> <br>
<br>
</blockquote></div>





[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic