[Bug 50277] New: Haven intro video does not play
https://bugs.winehq.org/show_bug.cgi?id=50277 Bug ID: 50277 Summary: Haven intro video does not play Product: Wine Version: 6.0-rc1 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: mfplat Assignee: wine-bugs(a)winehq.org Reporter: kai.sterker+wine(a)gmail.com Distribution: --- Created attachment 68825 --> https://bugs.winehq.org/attachment.cgi?id=68825 Debug logs with +mfplat,+gstreamer Upon testing Haven with the builtin media foundation implementation, the following error is logged, the intro movie does not play and the game is stuck at launch: 020c:err:mfplat:media_source_constructor Failed to play source, error 0. 0218:err:mfplat:bus_watch qtdemux0: Internal data stream error. 0218:err:mfplat:bus_watch qtdemux.c(6605): gst_qtdemux_loop (): /GstBin:bin0/GstDecodeBin:decodebin0/GstQTDemux:qtdemux0: streaming stopped, reason error (-5) Debug logs with +mfplat,+gstreamer attached. I'm on Ubuntu 20.04. If I can provide any other useful information please let me know. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 Nikolay Sivov <bunglehead(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Component|mfplat |winegstreamer --- Comment #1 from Nikolay Sivov <bunglehead(a)gmail.com> --- Thank you. I'll change component to winegstreamer for now. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #2 from Kai Sterker <kai.sterker+wine(a)gmail.com> --- Retested with wine 6.15 and the error above is gone. Intro video does still not play, however. Now the failure reads 0150:fixme:mfplat:media_source_get_service_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714}. Although this might just be a friendlier (and potentially more helpful) log of the same underlying problem. If needs be I can provide debug logs again. Or should I rather file a separate report for this? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 vv221 <vv221+winehq(a)dotslashplay.it> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |vv221+winehq(a)dotslashplay.i | |t --- Comment #3 from vv221 <vv221+winehq(a)dotslashplay.it> --- I get something similar with True Fear: Forsaken Souls Part 2, a game that has no entry in WineHQ AppDB yet. Using WINE 7.0-rc3 on a Debian Bookworm/Sid, with Mesa 21.2.6. The intro video audio is played and subtitles are shown, but the screen stays black instead of showing a video stream. The following debug output is shown when the video issues occur: winegstreamer warning: qtdemux0: Invalid atom size. winegstreamer warning: qtdemux0: ../gst/isomp4/qtdemux.c(4380): gst_qtdemux_loop_state_header (): /GstBin:bin0/GstDecodeBin:decodebin0/GstQTDemux:qtdemux0: Header atom '....' has empty length 0148:fixme:mfplat:media_source_get_service_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714}. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #4 from vv221 <vv221+winehq(a)dotslashplay.it> --- Unsurprisingly, the error reported with True Fear: Forsaken Souls Part 2 happens in the exact same way with the first game of the series, True Fear: Forsaken Souls Part 1. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #5 from vv221 <vv221+winehq(a)dotslashplay.it> --- I found another game affected by what seems to be the same bug, Summer in Mara: https://appdb.winehq.org/objectManager.php?sClass=application&iId=19893 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #6 from vv221 <vv221+winehq(a)dotslashplay.it> --- This is still a current issue with WINE 7.22. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 joaopa <jeremielapuree(a)yahoo.fr> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jeremielapuree(a)yahoo.fr --- Comment #7 from joaopa <jeremielapuree(a)yahoo.fr> --- For me, True fears works like a charm with wine-8.16 Can anyone confirm? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #8 from vv221 <accounts.winehq.org(a)vv221.fr> --- I tried True Fear: Forsaken Souls Part 2 with WINE 8.16, the introduction video playback still does not work, I only get closed captions on a black screen. But there is a change, I now get the following debug output: (wine:1843052): GStreamer-Video-CRITICAL **: 00:54:58.441: gst_video_info_from_caps: assertion 'gst_caps_is_fixed (caps)' failed 01c4:fixme:mfplat:media_source_get_service_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714}. The second line (Unsupported service) is not new, but the first one (assertion 'gst_caps_is_fixed (caps)' failed) I don’t remember seeing with WINE 8.0. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 Maxime <berillions(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |berillions(a)gmail.com --- Comment #9 from Maxime <berillions(a)gmail.com> --- Someone (in Wine's dev) have an idea to fix this issue ? I have the problem with "A Plague Tale : Requiem", the intro video is black and i have only sound. Error message in the output console : #### 00cc:fixme:mfplat:media_engine_SetLoop (00007F3896A0D7E0, 0): stub. 00cc:fixme:thread:NtSetInformationThread ThreadIdealProcessor stub! 021c:fixme:mfplat:stream_handler_BeginCreateObject Unimplemented flags 0x11 (wine:17641): GStreamer-Video-CRITICAL **: 08:52:39.514: gst_video_info_from_caps: assertion 'gst_caps_is_fixed (caps)' failed 01ac:fixme:mfplat:topology_loader_Load iface 00007F3893FC2820, input_topology 00007F3A2F9BED80, ret_topology 00007F3A16CDFCF8, current_topology 0000000000000000 stub! 01ac:fixme:mfplat:video_processor_ProcessMessage iface 00007F3A0404CE80, message 0x10000000, param 0 stub! #### -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #10 from joaopa <jeremielapuree(a)yahoo.fr> --- Did you install the version all the requested gstreamer plugins: base good bad -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #11 from Maxime <berillions(a)gmail.com> --- (In reply to joaopa from comment #10)
Did you install the version all the requested gstreamer plugins: base good bad
Yes of course. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #12 from joaopa <jeremielapuree(a)yahoo.fr> --- I meant the 32-bit version. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 Zeb Figura <z.figura12(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |z.figura12(a)gmail.com --- Comment #13 from Zeb Figura <z.figura12(a)gmail.com> --- (In reply to joaopa from comment #12)
I meant the 32-bit version.
If GStreamer plugins were missing we'd specifically see an error along those lines. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #14 from vv221 <accounts.winehq.org(a)vv221.fr> --- Iron Harvest is another game affected by this bug. GStreamer tools (gst-typefind-1.0) identify the video file triggering the crash as "video/quicktime, variant=(string)iso", the required plugins to handle this format are available in the correct architecture (base + good + libav). The main symptom is the game getting stuck on a black screen at the end of the first mission, when it should display the video instead. This prevents any further progress in the campaign. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #15 from vv221 <accounts.winehq.org(a)vv221.fr> --- Chicken Police is yet another game (based on the Unity engine) affected by the same problem. This time the game is stuck on a black screen when it should play the intro video when starting a new game. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #16 from Ken Sharp <imwellcushtymelike(a)gmail.com> --- If you find an app with the exact same bug just link it in the AppDB. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #17 from vv221 <accounts.winehq.org(a)vv221.fr> ---
If you find an app with the exact same bug just link it in the AppDB.
I did that for Iron Harvest, but it seems that Chicken Police is not included in the AppDB yet. I plan to populate the AppDB with new reports soon, so I will link the new Chicken Police entry to this bug report when I add it. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 Shmerl <shtetldik(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |shtetldik(a)gmail.com --- Comment #18 from Shmerl <shtetldik(a)gmail.com> --- Could be a similar issue. I get these errors when videos don't play in Blood West: 0:00:00.094873436 823464 0x7f06a8000db0 ERROR audio-info audio-info.c:302:gst_audio_info_from_caps: no rate property given (wine:823464): GStreamer-Video-CRITICAL **: 19:53:46.923: gst_video_info_from_caps: assertion 'gst_caps_is_fixed (caps)' failed 0:00:00.107035676 823464 0x7f06a8001000 ERROR audio-info audio-info.c:302:gst_audio_info_from_caps: no rate property given 01e0:fixme:mfplat:media_source_get_service_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714}. 01e0:fixme:mfplat:MFPluginControl_IsDisabled (0 {88753b26-5b24-49bd-b2e7-0c445c78c982}) 01e0:fixme:mfplat:MFPluginControl_GetPreferredClsidByIndex (0 0 00007F099DA1EEC8 00007F099DA1EEF0) 0314:fixme:mfplat:video_processor_ProcessMessage Ignoring message 0. According to this collection https://github.com/HonkingGoose/proton_codecs_media_support_tracker/blob/mai... dd0ac3d8-40e3-4128-ac48-c0add067b714 refers to H.264+AAC. Is anything missing in Wine to make it work over gstreamer? I have all gstreamer plugins installed. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 mrdeathjr28(a)yahoo.es changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mrdeathjr28(a)yahoo.es -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #19 from mrdeathjr28(a)yahoo.es --- elderborn (same devs of blood west) show a similar error, audio works without video -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #20 from joaopa <jeremielapuree(a)yahoo.fr> --- Created attachment 76764 --> https://bugs.winehq.org/attachment.cgi?id=76764 console output for elderborn I tested elderborn. Videos are black there is sound, but no mfplat error messages. Surely not the same bug. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #21 from joaopa <jeremielapuree(a)yahoo.fr> --- Speak too fast. I have the mfplat errors (wine:5289): GStreamer-Video-CRITICAL **: 15:31:06.141: gst_video_info_from_caps: assertion 'gst_caps_is_fixed (caps)' failed 0150:fixme:mfplat:media_source_get_service_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714}. 0150:fixme:mfplat:MFPluginControl_IsDisabled (0 {88753b26-5b24-49bd-b2e7-0c445c78c982}) 0150:fixme:mfplat:MFPluginControl_GetPreferredClsidByIndex (0 0 00007D7BDF5EF198 00007D7BDF5EF1C0) 0230:fixme:mfplat:video_processor_ProcessMessage Ignoring message 0. (wine:5289): GStreamer-Video-CRITICAL **: 15:31:06.594: gst_video_info_from_caps: assertion 'gst_caps_is_fixed (caps)' failed 0154:fixme:mfplat:media_source_get_service_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714}. 0154:fixme:mfplat:MFPluginControl_IsDisabled (0 {88753b26-5b24-49bd-b2e7-0c445c78c982}) 0154:fixme:mfplat:MFPluginControl_GetPreferredClsidByIndex (0 0 00007D7BDF4DF198 00007D7BDF4DF1C0) -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #22 from Shmerl <shtetldik(a)gmail.com> --- Some report that Proton handles Blood West correct. What is the exact difference and can this be upstreamed or for some reason not? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #23 from Shmerl <shtetldik(a)gmail.com> --- FYI, using new ffmpeg backend for MF doesn't yet fix this. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #24 from Shmerl <shtetldik(a)gmail.com> --- Btw, what exactly is dd0ac3d8-40e3-4128-ac48-c0add067b714? Is it something explicitly missing in Wine now that's not enabling it? I'm sure ffmpeg can support H.264 + AAC if anything. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #25 from Shmerl <shtetldik(a)gmail.com> --- Actually my bad, I built Wine without libav*-dev packages. I rebuilt it now and winedmo is not a dummy anymore. Here are some more detailed errors when starting Blood West: ``` 01fc:fixme:mfplat:media_source_IMFGetService_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714} / riid {26afea53-d9ed-42b5-ab80-e64f9ee34779} 01fc:fixme:mfplat:MFPluginControl_IsDisabled (0 {88753b26-5b24-49bd-b2e7-0c445c78c982}) 01fc:fixme:mfplat:MFPluginControl_IsDisabled (0 {62ce7e72-4c71-4d20-b15d-452831a87d9d}) 01fc:fixme:mfplat:MFPluginControl_IsDisabled (0 {82d353df-90bd-4382-8bc2-3f6192b76e34}) 01fc:fixme:mfplat:MFPluginControl_GetPreferredClsidByIndex (0 0 00007FBE03CDEED8 00007FBE03CDEF00) 01fc:fixme:mfplat:transform_GetAttributes iface 00007FBCE4D4DD68, attributes 00007FBE03CDF078 semi-stub! 01fc:fixme:mfplat:MFPluginControl_IsDisabled (0 {88753b26-5b24-49bd-b2e7-0c445c78c982}) 01fc:fixme:mfplat:MFPluginControl_GetPreferredClsidByIndex (0 0 00007FBE03CDEDB8 00007FBE03CDEDE0) 01fc:fixme:mfplat:MFPluginControl_IsDisabled (0 {f447b69e-1884-4a7e-8055-346f74d6edb3}) 01fc:fixme:mfplat:MFPluginControl_IsDisabled (0 {2eeb4adf-4578-4d10-bca7-bb955f56320a}) 01fc:fixme:mfplat:MFPluginControl_IsDisabled (0 {32d186a7-218f-4c75-8876-dd77273a8999}) 01fc:fixme:mfplat:MFPluginControl_GetPreferredClsidByIndex (0 0 00007FBE03CDEED8 00007FBE03CDEF00) 0:00:00.070471028 12403 0x7fbb68ea2cf0 ERROR aacparse gstaacparse.c:358:gst_aac_parse_sink_setcaps:<aacparse0> Need codec_data for raw AAC 0:00:00.070479268 12403 0x7fbb68ea2cf0 ERROR aacparse gstaacparse.c:358:gst_aac_parse_sink_setcaps:<aacparse0> Need codec_data for raw AAC 0:00:00.070904783 12403 0x7fbb68ea2cf0 ERROR WINE unixlib.c:226:link_element_to_sink: Failed to find src pad compatible to <'':sink> on <audioresample1>. 0:00:00.070923863 12403 0x7fbb68ea2cf0 ERROR WINE wg_transform.c:694:wg_transform_create: Failed to create winegstreamer transform. 031c:fixme:mfplat:transform_ProcessMessage iface 00007FBCE4E2F4E0, message 0x10000003, param 0000000000000000 stub! 031c:err:dmo:buffer_unlock Failed to update buffer length, hr 0x80070057 031c:err:dmo:winedmo_demuxer_read Failed to read sample, status 0xc0000023 031c:err:dmo:buffer_unlock Failed to update buffer length, hr 0x80070057 031c:err:dmo:winedmo_demuxer_read Failed to read sample, status 0xc0000023 031c:fixme:mfplat:video_processor_ProcessMessage Ignoring message 0x10000003. 031c:err:dmo:buffer_unlock Failed to update buffer length, hr 0x80070057 031c:err:dmo:winedmo_demuxer_read Failed to read sample, status 0xc0000023 ... 031c:err:dmo:buffer_unlock Failed to update buffer length, hr 0x80070057 031c:err:dmo:winedmo_demuxer_read Failed to read sample, status 0xc0000023 031c:fixme:mfplat:transform_GetAttributes iface 00007FBCE4D4DD68, attributes 00007FBCE427FC30 semi-stub! 031c:fixme:mfplat:transform_GetOutputStreamAttributes iface 00007FBCE4D4DD68, id 0, attributes 00007FBCE427FC30 semi-stub! 031c:err:dmo:buffer_unlock Failed to update buffer length, hr 0x80070057 031c:err:dmo:winedmo_demuxer_read Failed to read sample, status 0xc0000023 031c:err:dmo:buffer_unlock Failed to update buffer length, hr 0x80070057 ... ``` -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #26 from Shmerl <shtetldik(a)gmail.com> --- Not sure if it's the game update that changed something or something changed in Wine, but videos in Blood West started working! -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #27 from Kai Sterker <kai.sterker+wine(a)gmail.com> --- Retried Haven with wine 9.21, but it's still not working fully. The intro normally looks like this: https://youtu.be/apDarietN0M With wine it looks like on the attached screenshot. Music is playing fine, and the game reacts to inputs, allowing to skip to the main menu. Waiting long enough will also bring up the main menu eventually, so it seems the intro is playing as expected, just not rendering correctly. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #28 from Kai Sterker <kai.sterker+wine(a)gmail.com> --- Created attachment 77417 --> https://bugs.winehq.org/attachment.cgi?id=77417 Haven intro movie renders like this -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 Kai Sterker <kai.sterker+wine(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #68825|0 |1 is obsolete| | --- Comment #29 from Kai Sterker <kai.sterker+wine(a)gmail.com> --- Created attachment 77418 --> https://bugs.winehq.org/attachment.cgi?id=77418 New logs with +mfplat,+gstreamer -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #30 from vv221 <accounts.winehq.org(a)vv221.fr> --- I ran tests with WINE 9.21, using several games affected by this bug. Haven The video can now be skipped, allowing to reach the game menu and then the actual gameplay parts. While the introduction video itself can still not been seen in game, the game is now playable. Some short videos showing logos before the longer introduction video are shown without any rendering problems, that was not the case with WINE 9.0. Chicken Police The intro video (played after starting a new game) fails to render anything but a black screen and the sound is messed up. But there is sound, unlike with WINE 9.0, and the video can now be skipped. The actual gameplay can now be reached. Atone: Heart of the Elder Tree The failure to render the intro video with WINE 9.0 prevented the game menu from rendering. With WINE 9.21 the intro video prior to reaching the menu can be skipped, a new game can be started, and the following video can be skipped, finally allowing to reach the actual gameplay. None of the skippable videos are rendered correctly, I only get a black screen with audio. Iron Harvest The video at the end of the first campaign map fails to render anything and its audio is messed up, but it can be skipped when WINE 9.21 is used. This allows progressing further in the campaign, while we used to be stuck there with WINE 9.0. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #31 from vv221 <accounts.winehq.org(a)vv221.fr> --- PS: I don’t remember having tested any version between WINE 9.0 and 9.21, so I can’t say for sure the improvements are specific to 9.21 or introduced with an earlier version. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=50277 login(a)b-vo.fr changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |login(a)b-vo.fr -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 Cedric Heintz <cedric(a)ced117.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |cedric(a)ced117.net -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 Stian Low <wineryyyyy(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |wineryyyyy(a)gmail.com --- Comment #32 from Stian Low <wineryyyyy(a)gmail.com> --- (In reply to Kai Sterker from comment #28)
Created attachment 77417 [details] Haven intro movie renders like this
Confirming for 10.13 commit d74d97cbf87 using shared wow64 which reproduces snapshot of the background image over which the video is supposed to play. GE-Proton10-11 plays the video outside Steam. Video also works via Steam Proton experimental. Running steam via this command may be needed to enable additional video features: steam steam://unlockh264 unlockh264 is another workaround suggested for The Signifier start menu background videos reported broken: https://bugs.winehq.org/show_bug.cgi?id=56614#c23 https://github.com/ValveSoftware/Proton/issues/8989#issuecomment-3258020085 GE-Proton seems to enable unlockh264 features by default. GE-Proton-10-15 gets never loads past the initial loading screen to even play the video but that may be related to a separate issue for wine 10.15 that I need to test. Recently GE-Proton has saved a lot of time otherwise wasted trying to fix bugs that are already technically fixed. This begs the question. Are these video fixes considered valid? If so then why are fixes not merged back to winehq? Is it accidental? If it's intentional then documenting the reason why may help newcomers from confusing video bugs as high priority bugs when actually very low and instead focusing efforts on much higher priority bugs. What is needed to close these bugs? Extract diffs that fix videos from proton and apply to vanilla wine? Please advise. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #33 from Stian Low <wineryyyyy(a)gmail.com> --- (In reply to Stian Low from comment #32)
This begs the question. Are these video fixes considered valid? If so then why are fixes not merged back to winehq? Is it accidental? If it's intentional then documenting the reason why may help newcomers from confusing video bugs as high priority bugs when actually very low and instead focusing efforts on much higher priority bugs.
What is needed to close these bugs? Extract diffs that fix videos from proton and apply to vanilla wine?
Please advise.
Same question for video bugs applies for vulkan bugs fixed by dxvk: https://bugs.winehq.org/show_bug.cgi?id=58495 Why are dxvk fixes not merged back to winehq? Intentional? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #34 from Stian Low <wineryyyyy(a)gmail.com> --- (In reply to Shmerl from comment #22)
Some report that Proton handles Blood West correct. What is the exact difference and can this be upstreamed or for some reason not?
This seems to be an open-ended question. Maybe by design? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #35 from Stian Low <wineryyyyy(a)gmail.com> --- (In reply to Stian Low from comment #34)
(In reply to Shmerl from comment #22)
Some report that Proton handles Blood West correct. What is the exact difference and can this be upstreamed or for some reason not?
This seems to be an open-ended question. Maybe by design?
Partly answering my own question: https://forum.winehq.org/viewtopic.php?p=148593#p148593 Much less ambiguity but still some regarding solution for WINE. DXVK C++ fixes translated to C considered valid? Anything else blocking in theory? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #36 from Shmerl <shtetldik(a)gmail.com> --- Blood West videos work in Wine now, as I mentioned above. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #37 from Stian Low <wineryyyyy(a)gmail.com> --- (In reply to Shmerl from comment #36)
Blood West videos work in Wine now, as I mentioned above.
Thanks. I'll assume upstreaming proton diffs that fix remaining otherwise broken videos like Haven for vanilla wine is considered valid unless told otherwise. Found yet another semi-duplicate video bug also fixed by proton: https://bugs.winehq.org/show_bug.cgi?id=57908 UnityPlayer.dll seems common among all of these video bugs but Blood West also depends on Unity so whatever fixed it seems hit or miss. Still unclear how WINE views DXVK but seems clear enough for now to proceed. Remaining ambiguity: DXVK considered a workaround or solution? I assume WINE ideally rathers DXVK implemented in C and packaged with WINE. But C++ projects fulfilling gaps like graphics may be considered valid enough that priority to port to C isn't as high as other efforts. Perhaps this is an ongoing consideration not fully decided and still evolving. I'll assume that if C++ fixes are easy enough to port to C then its worth the effort. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #38 from Stian Low <wineryyyyy(a)gmail.com> --- (In reply to Shmerl from comment #36)
Blood West videos work in Wine now, as I mentioned above.
Confirming Blood West videos still work for Wine 10.14 commit e6df334f178 with shared wow64. So probably not a duplicate of this bug so Blood West video bug can probably be closed: https://bugs.winehq.org/show_bug.cgi?id=56936 However: Blood West gameplay responds to no keyboard/mouse inputs: 0294:err:winediag:ntlm_check_version ntlm_auth was not found. Make sure that ntlm_auth >= 3.0.25 is in your path. Usually, you can find it in the winbind package of your distribution. All inputs worked fine leading up to actual start of gameplay. I'll create a separate bug to handle broken input if determined legit after more testing. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #39 from Stian Low <wineryyyyy(a)gmail.com> --- (In reply to Stian Low from comment #38)
(In reply to Shmerl from comment #36)
Blood West videos work in Wine now, as I mentioned above.
Confirming Blood West videos still work for Wine 10.14 commit e6df334f178 with shared wow64.
Correction: Blood West videos played for wine-10.0 (Debian 10.0~repack-6) provided by Debian apt pkg. But retesting for Wine 10.14 confirms that Blood West videos still work or latest so indeed not a duplicate of Haven video bugs.
Blood West gameplay responds to no keyboard/mouse inputs:
Wine 10.14 also fixes input bugs encountered for Debian wine-10.0. Blood West has very laggy gameplay but that's outside the scope of video bugs. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 Hugo Osvaldo Barrera <hugo@whynothugo.nl> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |hugo@whynothugo.nl -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #40 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81434 --> http://bugs.winehq.org/attachment.cgi?id=81434 Screenshot intro video playing via hack Hack that allows Haven videos to play same as signifier: https://bugs.winehq.org/show_bug.cgi?id=56614#c26 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #41 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Shmerl from comment #24)
Btw, what exactly is dd0ac3d8-40e3-4128-ac48-c0add067b714? Is it something explicitly missing in Wine now that's not enabling it? I'm sure ffmpeg can support H.264 + AAC if anything.
dd0ac3d8-40e3-4128-ac48-c0add067b714 is MF_SCRUBBING_SERVICE and may be a red herring for these video playback errors: https://www.magnumdb.com/search?q=dd0ac3d8-40e3-4128-ac48-c0add067b714 dxgi_resource_GetSharedHandle logs may indicate the problem which are skipped via hack that forces video playback natively as MFVideoFormat_NV12 rather than converting to MFVideoFormat_ABGR32 which may be implemented in a multthreaded manner that requires dxgi_resource_GetSharedHandle. 01a8:fixme:mfplat:media_source_get_service_GetService Unsupported service {dd0ac3d8-40e3-4128-ac48-c0add067b714}. 01a8:fixme:mfplat:MFPluginControl_IsDisabled (0 {88753b26-5b24-49bd-b2e7-0c445c78c982}) 01a8:fixme:mfplat:MFPluginControl_GetPreferredClsidByIndex (0 0 00007F3B8998F258 00007F3B8998F280) 0234:fixme:dmo:video_processor_ProcessMessage Ignoring message 0x10000003. 0244:fixme:dmo:video_processor_ProcessMessage Ignoring message 0. 0234:fixme:dmo:video_processor_ProcessMessage Ignoring message 0x10000003. 0024:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F39AE9C5688, shared_handle 00007FFFFE1FF330 stub! 0024:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F39AF2109E8, shared_handle 00007FFFFE1FF330 stub! 0024:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F39AE9C5768, shared_handle 00007FFFFE1FF330 stub! 0024:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F39AE9C5688, shared_handle 00007FFFFE1FF330 stub! 0024:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F39AF2109E8, shared_handle 00007FFFFE1FF330 stub! 0024:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F39AE9C5768, shared_handle 00007FFFFE1FF330 stub! -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #42 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81437 --> http://bugs.winehq.org/attachment.cgi?id=81437 Patch that works for WINE_D3D_CONFIG=csmt=0x2,renderer=vulkan Patch provides minimal support for dxgi_resource_GetSharedHandle and d3d11_device_OpenSharedResource which allows intro video to play (choppy) for renderer=vulkan (damvand) via WINE_D3D_CONFIG=csmt=0x2 to force single-threading. WINE_D3D_CONFIG=csmt=0x1 for multithreading crashes probably because wined3d lacks thread safety for some UnityPlayer.dll calls. renderer=gl seem incompatible with WINE_D3D_CONFIG=csmt=0x2 as of wine-11.12 because it crashes more immediately and more severely without backtrace. Unclear why csmt=0x2 cannot be supported for renderer=gl like vulkan. renderer=vulkan renders a few frames for multithreading via WINE_D3D_CONFIG=csmt=0x1 but eventually crashes which seems to clarify that lack of wined3d thread-safety is probably partly a problem. renderer=vulkan may support texture sharing more natively vs renderer=gl which may require a more explicit solution to support texture sharing across devices however UnityPlayer.dll handles it. Choppy video for renderer=vulkan may be a separate issue. Useful UnityPlayer args for further troubleshooting these types of video playback errors: https://docs.unity3d.com/6000.0/Documentation/Manual/PlayerCommandLineArgume... --force-d3d12 forces vkd3d which still displays blackscreen regardless of patch so d3d12 may need more to support texture sharing across devices. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #43 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #42)
Unclear why csmt=0x2 cannot be supported for renderer=gl like vulkan.
WINE_D3D_CONFIG=csmt=0x2 works with renderer=gl with these Unity args: -force-gfx-direct -force-d3d11-singlethreaded Single-threaded gl shows a distorted menu screen and eventually crashes unlike single-threaded renderer=vulkan. The screen distortion may indicate that cross device sharing is attempted but maybe not accounted by wglShareLists (or equiv) which leads to crashes for gl. dxgi_resource_GetSharedHandle and d3d11_device_OpenSharedResource are still called even for single-thread runs so cross-device sharing may still a problem for gl vs vulkan which handles more natively. Remaining work after minimal support for dxgi_resource_GetSharedHandle and d3d11_device_OpenSharedResource: 1. Support thread-safety for wined3d calls by UnityPlayer.dll 2. Support gl cross-device sharing via wglShareLists mechanisms 3. Test if renderer=gl has choppy video like renderer=vulkan 4. Fix choppy video as necessary to resolve this bug -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #44 from Stian Low <wineryyyyy@gmail.com> --- renderer=vulkan renders dark like many Unity games for default `winecfg -v win10` until this merge: https://gitlab.winehq.org/wine/wine/-/merge_requests/10567 https://bugs.winehq.org/show_bug.cgi?id=45364 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #45 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #43)
Remaining work after minimal support for dxgi_resource_GetSharedHandle and d3d11_device_OpenSharedResource: 1. Support thread-safety for wined3d calls by UnityPlayer.dll 2. Support gl cross-device sharing via wglShareLists mechanisms 3. Test if renderer=gl has choppy video like renderer=vulkan 4. Fix choppy video as necessary to resolve this bug
5. Support vkd3d video playback via -force-d3d12 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #46 from Stian Low <wineryyyyy@gmail.com> --- This patch alone forces UnityPlayer.dll to use software video decoding which seems to skip cross-device sharing and plays videos for renderer=gl and vulkan even with multi-threading enabled via WINE_D3D_CONFIG=csmt=0x1. diff --git a/dlls/dxgi/dxgi_main.c b/dlls/dxgi/dxgi_main.c index ca1fb525535..34719b1c5ef 100644 --- a/dlls/dxgi/dxgi_main.c +++ b/dlls/dxgi/dxgi_main.c @@ -147,6 +147,11 @@ HRESULT WINAPI DXGID3D10CreateDevice(HMODULE d3d10core, IDXGIFactory *factory, I if (flags) FIXME("Ignoring flags %#x.\n", flags); + if (flags & 0x800) + { + FIXME("failing for D3D11_CREATE_DEVICE_VIDEO_SUPPORT\n"); + return E_FAIL; + } if (TRACE_ON(dxgi)) dump_feature_levels(feature_levels, level_count); dxgi_resource_GetSharedHandle and d3d11_device_OpenSharedResource are no longer logged which may be specific to D3D11_CREATE_DEVICE_VIDEO_SUPPORT skipped. Vulkan still plays video choppy vs gl which plays without any obvious issues. -force-d3d12 via vkd3d also plays the video choppy similar to damavand. Maybe some combo of Unity flags supports forcing software rendering without need for the hack. D3D11_CREATE_DEVICE_VIDEO_SUPPORT seems to be lacking support which probably requires all multi-thread safety work. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #47 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #46)
This patch alone forces UnityPlayer.dll to use software video decoding which seems to skip cross-device sharing and plays videos for renderer=gl and vulkan even with multi-threading enabled via WINE_D3D_CONFIG=csmt=0x1.
MR created to force UnityPlayer.dll to use software rendering until D3D11_CREATE_DEVICE_VIDEO_SUPPORT is better supported via the cross-device multi-thread safety needed: https://gitlab.winehq.org/wine/wine/-/merge_requests/11398 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 Afif An Nafie <afifsworld132@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |afifsworld132@gmail.com --- Comment #48 from Afif An Nafie <afifsworld132@gmail.com> --- I found the same isdue with Hollow Knight: Silksong and In Falsus game intros as well. https://github.com/ValveSoftware/wine/commit/164af86dd770f975cdff3e09884f14e... fixed it, though it had memory leaks (tested on macOS through DXMT). -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #49 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Afif An Nafie from comment #48)
I found the same isdue with Hollow Knight: Silksong and In Falsus game intros as well. https://github.com/ValveSoftware/wine/commit/ 164af86dd770f975cdff3e09884f14ebc14b856b fixed it, though it had memory leaks (tested on macOS through DXMT).
Thanks for more games to test. I reviewed proton's way of handling these videos a while back but concluded it deviated too far from vanilla wine to be useful for fixing these bugs. Besides DXVK, proton wine also contains significant deviations such as support for GetSharedHandle and OpenSharedResource which may not be easily portable for latest wine-11.12. Unclear if proton somehow falls back to software video decoding or attempts support for cross-device multi-thread-safety mechanisms. Proton seems not to have any special handling for D3D11_CREATE_DEVICE_VIDEO_SUPPORT which may imply it supports the more advanced video decoding. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #50 from Afif An Nafie <afifsworld132@gmail.com> --- Fwiw, I just apply that patch over current wine and video then works fine. Data from fences was being emitted by game, but got clobbered before being read. Then when renderer reads from fence, it gets wrong value causing issues. The miscflags carried data which would have fence info but it gets set to 0. Also DXMT did a similar hack as the MR (https://github.com/3Shain/dxmt/pull/105/commits/14038e4441fbfc396238eeef2d6d...), so that videos would use software rendering, but was found to not be sufficient to fix the bug. I didn't need to do any other changes or any special changes to my setup (just using latest gstreamer-development) Maybe there are other ways to make sure the fence values don't get flushed before reading the value. But then again, i am just testing on macOS and didn't test Linux... -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #51 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Afif An Nafie from comment #50)
Also DXMT did a similar hack as the MR (https://github.com/3Shain/dxmt/pull/105/commits/ 14038e4441fbfc396238eeef2d6dc8800ce415e3), so that videos would use software rendering, but was found to not be sufficient to fix the bug.
Are you able to test if my patch fixes the bug that DXMT does not fix? DXMT seems to be failing at a deeper level than my patch near top. Would be brilliant if UnityPlayer just supported args to force software video decoding. Maybe there is a way to force it by failing in a less severe manner. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #52 from Afif An Nafie <afifsworld132@gmail.com> --- Sure I can give the patch a try tomorrow. I'll test wined3d, dxvk, and dxmt. For reference, DXMT does have an open issue about this: https://github.com/3Shain/dxmt/issues/135 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #53 from Stian Low <wineryyyyy@gmail.com> --- Work-in-process merge-request to support cross-device multi-thread involving Get/OpenSharedResource(): https://gitlab.winehq.org/wine/wine/-/merge_requests/11404 MR-11404 should override MR-11398 which in the meantime forces UnityPlayer.dll software video decoding by failing for D3D11_CREATE_DEVICE_VIDEO_SUPPORT: https://gitlab.winehq.org/wine/wine/-/merge_requests/11398 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #54 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #44)
renderer=vulkan renders dark like many Unity games for default `winecfg -v win10` until this merge:
Correction: Only signifier is fixed by MR-10567 so Haven is only affected by this one bug vs signifier which is affected by both bugs. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #55 from Stian Low <wineryyyyy@gmail.com> --- Latest patch supports `WINE_D3D_CONFIG=no_create_flags=0x800` to control when to fail for UnityPlayer.dll secondary device creation for D3D11_CREATE_DEVICE_VIDEO_SUPPORT to fallback to software video decoding: https://gitlab.winehq.org/wine/wine/-/merge_requests/11398#note_146227 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #56 from Stian Low <wineryyyyy@gmail.com> --- Duplicate: https://bugs.winehq.org/show_bug.cgi?id=59503#c3 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #57 from Stian Low <wineryyyyy@gmail.com> --- Greak is duplicate but not Cult of Lamb which plays videos despite same logs that cause Haven videos to fail as blackscreen: 01ec:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007FDE80DFC848, shared_handle 00007FDF0015DD40 stub! https://bugs.winehq.org/show_bug.cgi?id=53559#c10 Maybe Unity/Player.dll version diffs cause these inconsistencies and quirks. I will debug to determine if software video decoding may be triggered for Haven similar to Cult of Lamb which does not need MR-11398 to force fail for D3D11_CREATE_DEVICE_VIDEO_SUPPORT as workaround. MR-11404 should enable Cult of Lamb to use GetSharedHandle method of rendering vs however its currently able to play videos which seem likely to be UnityPlayer.dll software video decoding which may be triggered by GetSharedHandle failing vs Haven which so far seems to need to fail for DXGID3D10CreateDevice flags 0x800 instead to trigger fallback video playback. If Unity version diffs are the cause then MR-11398 may still be a valid workaround required until MR-11404 is ready. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #58 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #57)
Maybe Unity/Player.dll version diffs cause these inconsistencies and quirks. I will debug to determine if software video decoding may be triggered for Haven similar to Cult of Lamb which does not need MR-11398 to force fail for D3D11_CREATE_DEVICE_VIDEO_SUPPORT as workaround.
UnityPlayer.dll 2022 may handle software video decoding fallback better than 2019: - Cult of Lamb released 2022 - UnityPlayer.dll 2022.3.62 - Greak released 2021 - UnityPlayer.dll 2019.4.14 https://bugs.winehq.org/show_bug.cgi?id=53559#c11 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #59 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #37)
Found yet another semi-duplicate video bug also fixed by proton: https://bugs.winehq.org/show_bug.cgi?id=57908
Duplicate confirmed. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #60 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #59)
Duplicate confirmed.
I'm willing to properly mark duplicates if Bugzilla admins elevate my account privileges. Otherwise I'll continue linking the lesser better way. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #61 from Stian Low <wineryyyyy@gmail.com> --- Stories Untold (GOG) bug may have reported Unity video bugs even earlier than Haven and is also fixed by MR-11398: - https://bugs.winehq.org/show_bug.cgi?id=44675#c12 UnityPlayer.dll does not even exist for Stories Untold but version is output to logs: Unity Player [version: Unity 5.6.3p2_b3d7a6428558] Therefore MR-11398 seems to fix videos for UnityPlayer 5.6 up to 2019. Some version after 2019 started triggering software video decoding via some other method that I'm still debugging. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #62 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Maxime from comment #9)
Someone (in Wine's dev) have an idea to fix this issue ? I have the problem with "A Plague Tale : Requiem", the intro video is black and i have only sound.
A Plague Tale : Requiem should have a separate bug report because it unrelated to because it's not a Unity game and MangoHUD reports VKD3D so it uses d3d12 vs d3d11 for Haven and all duplicates. Also seems to have regressed showing blackscreen for everything instead of just for intro videos for wine-11.13-d30dcd75bc9. Spams d3d12_pipeline_state_GetCachedBlob and then vkd3d_create_texture_uav: vkd3d:0168:fixme:vkd3d:d3d12_pipeline_state_GetCachedBlob iface 00007EF9050C5520, blob 00007EFC5D6BF370 stub! vkd3d:016c:fixme:vkd3d:d3d12_pipeline_state_GetCachedBlob iface 00007EF9013E65A0, blob 00007EFC5D59F370 stub! 0024:fixme:dxgi:dxgi_output_GetDisplayModeList1 iface 00007EFA95AEC280, format DXGI_FORMAT_R8G8B8A8_UNORM, flags 0, mode_count 00007FFFFE216B80, modes 0000000000000000 partial stub! 0024:fixme:dxgi:dxgi_output_GetDisplayModeList1 iface 00007EFA95AEC280, format DXGI_FORMAT_R8G8B8A8_UNORM, flags 0, mode_count 00007FFFFE216B80, modes 00007EFACC598A20 partial stub! vkd3d:0208:fixme:vkd3d:vkd3d_create_texture_uav Unhandled depth view 0-1. vkd3d:0208:fixme:vkd3d:vkd3d_create_texture_uav Unhandled depth view 1-1. Some bugs already reported for the game but maybe not for video blackscreen and this regression in which case separate bugs should be created for them. I'll debug A Plague Tale : Requiem after UnityPlayer.dll video bugs are fixed. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #63 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81560 --> http://bugs.winehq.org/attachment.cgi?id=81560 Screenshots videos playing for True Fear: Forsaken Souls Part 2 for wine-11.13-d30dcd75bc9 for renderer=gl and vulkan (In reply to Stian Low from comment #61) Screenshots videos playing for True Fear: Forsaken Souls Part 2 for wine-11.13-d30dcd75bc9 for renderer=gl and vulkan Videos play despite logs that otherwise indicate video bugs for Haven and duplicates: 0024:fixme:dxgi:DXGID3D10CreateDevice Ignoring flags 0x820. 0024:fixme:d3d11:d3d11_device_CreateFence iface 00007F9C03BA9020, initial_value 0, flags 0, iid {affde9d1-1df7-4bb7-8a34-0f46251dab80}, fence 00007FFFFE1FF308 stub! 0024:fixme:d3d11:d3d10_multithread_SetMultithreadProtected iface 00007F9C03BA9030, enable 0x1 stub! 01f0:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F9C03E6C538, shared_handle 00007F9CB837DAB0 stub! (In reply to Stian Low from comment #61)
Therefore MR-11398 seems to fix videos for UnityPlayer 5.6 up to 2019
UnityPlayer.dll seem to be 2018.3 which puts it between 5.6 and 2019 which seems to suggest 2019 and earlier in fact may support a way to trigger software video decoding without needing to be forced via MR-11398. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #64 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #58)
UnityPlayer.dll 2022 may handle software video decoding fallback better than 2019:
- Cult of Lamb released 2022 - UnityPlayer.dll 2022.3.62 - Greak released 2021 - UnityPlayer.dll 2019.4.14
Last Stand: Aftermath plays videos for UnityPlayer.dll 2019.4.30f1 which seems to confirm that these videos bugs are not UnityPlayer version specific afterall: - https://bugs.winehq.org/show_bug.cgi?id=54858#c25 Both renderer=gl and vulkan play intro videos but vulkan seems to render with better quality which may indicate another UnityPlayer quirk is triggered but needs more testing to confirm (screenshots 5-6): - https://bugs.winehq.org/attachment.cgi?id=81562 Still seeking a fix for Haven video bugs without MR-11398 since other Unity games sharing similar versions do not have bugged videos. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #65 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81567 --> http://bugs.winehq.org/attachment.cgi?id=81567 Screenshots `winecfg /v win8` rendering better than win10 in general vs just for videos (In reply to Stian Low from comment #64)
Both renderer=gl and vulkan play intro videos but vulkan seems to render with better quality which may indicate another UnityPlayer quirk is triggered but needs more testing to confirm (screenshots 5-6): - https://bugs.winehq.org/attachment.cgi?id=81562 Seems additional work may be needed to fully support win10 rendering for some Unity games. Will test MR-10567 to see if it improves win10 rendering which seems unlikely because it addresses swapchain RTVs. Unlike win8 fixing dark for renderer=vulkan it seems to make no diff for fixing video bugs.
-- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #66 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #65)
Unlike win8 fixing dark for renderer=vulkan it seems to make no diff for fixing video bugs.
Night Security is the first case of win8 fixing win10 video bugs: https://bugs.winehq.org/show_bug.cgi?id=55775#c30 Night Security is also the first case of MR-11398 not fixing videos for win10 as it does for Haven and all duplicates so far. Will try to improve MR-11398 to include Night Security as duplicate. MR-11404 to support GetSharedHandle method of video playback may be ready first which should also make Night Security a duplicate. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #67 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #66)
Night Security is the first case of win8 fixing win10 video bugs: https://bugs.winehq.org/show_bug.cgi?id=55775#c30
For further clarity: win8 does not fix Haven videos nor duplicates which are instead fixed by MR-11398 unlike Night Security only fixed by win8 so far -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 winehq@b-vo.fr changed: What |Removed |Added ---------------------------------------------------------------------------- CC|winehq@b-vo.fr | -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #68 from Stian Low <wineryyyyy@gmail.com> --- Duplicate: - https://bugs.winehq.org/show_bug.cgi?id=56297 Othercide -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #69 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81638 --> http://bugs.winehq.org/attachment.cgi?id=81638 Screenshots MR-11398 fixing backtrace/videos for wine-11.14-1012f3d9950 for UnityPlayer.dll 2018.3.0a1 (In reply to Afif An Nafie from comment #52)
Sure I can give the patch a try tomorrow. I'll test wined3d, dxvk, and dxmt. For reference, DXMT does have an open issue about this: https://github.com/3Shain/dxmt/issues/135
Silksong crashes with backtrace at launch without MR-11398 `WINE_D3D_CONFIG=no_create_flags=0x800` and with one of the signature logs that typically indicates MR-11398 applies as fix/work around: 0024:fixme:dxgi:DXGID3D10CreateDevice Ignoring flags 0x820. Oddly the other GetSharedHandle log typically paired with 0x820 is not logged which may suggest another Unity/Player.dll quirk. Unclear if GetSharedHandle log missing means MR-11404 does not apply for Silksong which needs to be tested once ready. If MR-11404 does not also fix Silksong then it may indicate D3D11_CREATE_DEVICE_VIDEO_SUPPORT needs more than MR-11404 GetSharedHandle and further raise relevance of MR-11398 as a workaround in the meantime until whatever else is bugged is fixed. renderer=vulkan (damavand) is also not darker than gl despite also having signature log that indicates MR-10567 is needed to fix SWAP_EFFECT_FLIP_SEQUENTIAL for default win10: 0024:fixme:d3d:wined3d_swapchain_init Unimplemented swap effect 0x3. Therefore Silksong is an outlier among the other typical duplicates reported here so far but seems at least semi if not a full duplicate also. Will test orig Hollow Knight also which has an AppDB ratings vs Silksong which is missing: - https://appdb.winehq.org/objectManager.php?sClass=application&iId=18289 Protondb has some ratings for Silksong: - https://www.protondb.com/app/1030300 Unclear which version of Silksong I'm using so I'll try to retest for more clearly common like Steam/GOG. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #70 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81639 --> http://bugs.winehq.org/attachment.cgi?id=81639 Backtrace for both renderer=gl and vulkan without MR-11398 `WINE_D3D_CONFIG=no_create_flags=0x800` (In reply to Stian Low from comment #69)
Silksong crashes with backtrace at launch without MR-11398
-- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #71 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #69)
Silksong crashes with backtrace at launch without MR-11398 `WINE_D3D_CONFIG=no_create_flags=0x800` and with one of the signature logs that typically indicates MR-11398 applies as fix/work around:
Magic the Gathering Arena bug that crashed prior to wine-11.14 mentions mono override as a possible workaround without MR-11398: - https://bugs.winehq.org/show_bug.cgi?id=47753#c4 I will retest older version of MTGA to reproduce the crash and mono override workaround which may help clarify some of the causes of quirks/diffs among the Unity games related to MR-11398. Unity games contain their own mono versions modded for specific compatibility so I figured it was too problematic to override them because my previous attempts caused instability/crashes. I'll reattempt to determine how much of a factor mono may be regarding video bugs related to MR-11398. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #72 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #69)
Unclear which version of Silksong I'm using so I'll try to retest for more clearly common like Steam/GOG.
Silksong GOG-89650 also crashes without MR-11398 which has been updated to support all versions of d3d despite only known to be bugged for d3d11 so far just in case it has wider use: - https://gitlab.winehq.org/wine/wine/-/merge_requests/11398#note_147538 Silksong only logs 0x820 and not GetSharedHandle like the other video bugs/duplicates. MR-11398 fixes Silksong crash at launch which is unique compared to the other bugs that only fix videos. Because GetSharedHandle is not logged it may indicate that Silksong uses software video decoding by default and therefore only requires MR-11398 to fix the crash vs actually fixing the videos which may already been supported. Silksong tests seem to indicate that MR-11398 has wider/different scope than originally thought and is independent of MR-11404 rather than just a temporary workaround. Unity may need wine to better supoort D3D11_CREATE_DEVICE_VIDEO_SUPPORT in order to better handle GetSharedHandle so that forcing Unity fallback methods is not required to workaround discrepancies for now. I'll open a separate merge request to address whatever else may be missing for D3D11_CREATE_DEVICE_VIDEO_SUPPORT so that Silksong does not require MR-11398 eventually. Maybe D3D11_CREATE_DEVICE_VIDEO_SUPPORT should fail in the meantime since it seems not fully supported if at all but seems d3d maintainers are unlikely to accept that change because it may break other unknown apps. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #73 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #72)
Silksong GOG-89650 also crashes without MR-11398 which has been updated to
MR-11404 does not fix Silksong crash so support for GetSharedHandle seems not to apply for Silksong. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #74 from Stian Low <wineryyyyy@gmail.com> --- Rune Factory 5 is duplicate: - https://bugs.winehq.org/show_bug.cgi?id=55775#c34 Night Security is not quite a duplicate but seems related so I'll investigate it further. Other games mentioned by bug-55775 will also be tested soon to determine relation to Haven video bugs. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #75 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Antoine Le Gonidec from comment #5)
I found another game affected by what seems to be the same bug, Summer in Mara: https://appdb.winehq.org/objectManager.php?sClass=application&iId=19893
(In reply to Stian Low from comment #56)
bug-59503 reports Summer in Mara which is indeed a duplicate of Haven video bug. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #76 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Stian Low from comment #73)
MR-11404 does not fix Silksong crash so support for GetSharedHandle seems not to apply for Silksong.
Silksong is not a duplicate despite having similar workaround via MR-11398. - https://gitlab.winehq.org/wine/wine/-/merge_requests/11398#note_147677 Silksong crashes for wine-11.14 because `d3d11_device_CreateFence` returns E_NOTIMPL fixed by returning S_OK instead after which game works including videos without MR-11398 workaround to skip d3d create flag 0x800 which is not used for Silksong like other Haven video bug duplicates. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #77 from Afif An Nafie <afifsworld132@gmail.com> --- Hi. I am sorry I could not get the required tests yet. I promise i'll get back to it soon! There seems to be many variables which either make videos work or not on Silksong / other games. One thing I can say though is that when I started Silksong back in roughly wine 10.20, wined3d gave us a software rendered video just fine, and it still does work for me. One thing is though, I have never encountered even once throughout my testing that Silksong crashes on something. If the videos are not working, not on wined3d, DXVK or DXMT. I can usually just go to the Player.log found in AppData/LocalLow to see whether Shared Resources are being used, otherwise there would be repeated errors there about it. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #78 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Afif An Nafie from comment #77)
Hi. I am sorry I could not get the required tests yet. I promise i'll get back to it soon! There seems to be many variables which either make videos work or not on Silksong / other games. One thing I can say though is that when I started Silksong back in roughly wine 10.20, wined3d gave us a software rendered video just fine, and it still does work for me. One thing is though, I have never encountered even once throughout my testing that Silksong crashes on something. If the videos are not working, not on wined3d, DXVK or DXMT. I can usually just go to the Player.log found in AppData/LocalLow to see whether Shared Resources are being used, otherwise there would be repeated errors there about it.
No worries. Thanks for checking back. After hacking `d3d11_device_CreateFence` to return S_OK instead of E_NOTIMPL, Silksong logs GetSharedHandle right after but still seems to somehow fallback to software rendering unlike Haven duplicates: 0024:fixme:dxgi:DXGID3D10CreateDevice Ignoring flags 0x820. 0024:fixme:d3d:wined3d_device_create wined3d 00007F85C8C6B740, adapter 00007F85C20AB660, device_type 0x1, focus_window 0000000000000000, flags 0x820, surface_alignment 4, feature_levels 00006FFFF8977E80, feature_level_count 6, device_parent 00007F84F6FC4F58, device 00007F84F6FC4EB0. 0024:fixme:d3d11:d3d11_device_CreateFence iface 00007F84F6FC4ED0, initial_value 0, flags 0, iid {affde9d1-1df7-4bb7-8a34-0f46251dab80}, fence 00007FFFFE1FF368 stub! 0024:fixme:d3d11:d3d10_multithread_SetMultithreadProtected iface 00007F84F6FC4EE0, enable 0x1 stub! 0220:fixme:dxgi:dxgi_resource_GetSharedHandle iface 00007F85C216F348, shared_handle 00007F856BFEDC50 stub! Silksong seems not to be a Haven duplicate, so whatever remaining issues you have seem better handled via bug-59414: - https://bugs.winehq.org/show_bug.cgi?id=59414 If you have any other games that you think may be duplicates of Haven bugs besides Silksong please report them here. Also, forgive some of my rushed Bugzilla comments. Hopefully they are coherent enough. Please let me know if any need clarification. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #79 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81681 --> http://bugs.winehq.org/attachment.cgi?id=81681 Screenshots IronHarvest as duplicate for wine-11.14-4dbf7bef7a0 for renderer=gl and vulkan worked around via MR-11398 pending MR-11404 for GetSharedHandle; damavand darker fixed via `winecfg /v win8` IronHarvest has other bugs besides Haven video bug which may be regressions (also shown for screenshots): - renderer=gl renders ground without textures but renders trees - renderer=vulkan renders ground but not trees - Starts up fullscreen seemingly every other run despite setting windowed mode Final screenshot shows intro video for 1st campaign rendering ground partially correctly for renderer=gl which is played like a video but seems likely an in-game/gameplay cinematic rather than a true 2D video shown for the first screenshots which are the actual duplicates of Haven video bugs. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #80 from Stian Low <wineryyyyy@gmail.com> --- Also related/duplicate: - https://bugs.winehq.org/show_bug.cgi?id=56780#c3 Saint Kotar confirmed duplicate. The Long Dark and The Room 4: Old Sins are also likely duplicates but unfortunately do not have intros for quick test and I do not have time to wait/find cutscenes for now. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #81 from Stian Low <wineryyyyy@gmail.com> --- Created attachment 81685 --> http://bugs.winehq.org/attachment.cgi?id=81685 Screenshots Elderborn videos fixed via MR-11398 for renderer=gl and vulkan and darker rendering for vulkan fixed via `winecfg /v win8` until MR-10567 for win10 (In reply to Shmerl from comment #18)
Could be a similar issue. I get these errors when videos don't play in Blood West:
(In reply to mrdeathjr28 from comment #19)
elderborn (same devs of blood west) show a similar error, audio works without video
(In reply to Shmerl from comment #26)
Not sure if it's the game update that changed something or something changed in Wine, but videos in Blood West started working!
Despite Elderborn sharing the same developers it seems a duplicate of Haven video bug vs BloodWest which is able to fallback to software video decoding without MR-11398 for still unknown reasons. UnityPlayer/2020.3.10f1 args tested but none seem to support equiv MR-11398 workaround: - https://docs.unity3d.com/2020.3/Documentation/Manual/PlayerCommandLineArgume... Some UnityPlayer flags seem to be respected but not others: - -force-d3d12 never shows vkd3d but instead wined3d for renderer=gl and damavand for vulkan which may be a separate bug - -force-d3d11-bitblt-model does not fix renderer=vulkan darker rendering fixed by `winecfg /v win8` until MR-10567 for win10 - -force-vulkan and -force-glcore cause popup errors "Failed to initialize player" with details "InitializeEngineGraphics failed" which may need a separate bug Other bugs: - Alt+tab for fullscreen results in not responding state requiring force close for wayland test (seen for other games and may be a regression, x11 not tested yet; windowed mode works for wayland) - Fullscreen seems to override windowed mode every other run similar to IronHarvest which may also be a regression and makes Alt+tab bug more obvious MR-11398 was recently reviewed so I'll be working to do what I can to get it merged while I continue to also prepare to push a more complete/correct solution for MR-11404 this week which currently only works via `WINE_D3D_CONFIG=csmt=0x2,renderer=vulkan` as proof of an alternative partial fix for vulkan vs MR-11398 which supports both vulkan and gl and default csmt=0x1. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #82 from Stian Low <wineryyyyy@gmail.com> --- (In reply to Shmerl from comment #22)
Some report that Proton handles Blood West correct. What is the exact difference and can this be upstreamed or for some reason not?
I spent time last year trying to understand proton/dxvk diffs but concluded there were too many deviations for a practical port. Therefore my fixes so far have been irrespective of however proton/dxvk handles these videos. Attempts were made around 2021 to merge proton's changes that support videos for vulkan but were rejected for wine. Recently an alternative foundational architecture has been merged for supporting GetSharedHandle for wine. Still I have been directed to dxvk by winehq as a guide for how to better support GetSharedHandle for MR-11404 so I will take a closer look again for supporting it. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
http://bugs.winehq.org/show_bug.cgi?id=50277 --- Comment #83 from Stian Low <wineryyyyy@gmail.com> --- Ashes of sun confirmed duplicate: - https://bugs.winehq.org/show_bug.cgi?id=60018#c21 -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
participants (2)
-
WineHQ Bugzilla -
WineHQ Bugzilla