Bug 107022 - [RADV] The Witcher 3: Trembling of trees
Summary: [RADV] The Witcher 3: Trembling of trees
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/Vulkan/radeon (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: mesa-dev
QA Contact: mesa-dev
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-06-25 04:33 UTC by zefkerrigan
Modified: 2019-04-25 21:35 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:


Attachments
The Witcher 3 savegame file in the location where easy to reproduce this bug. (2.54 MB, application/x-xz)
2018-11-07 21:38 UTC, zefkerrigan
Details
All minimal settings (5.72 KB, application/x-xz)
2018-11-08 08:01 UTC, zefkerrigan
Details
The output of R600_DEBUG=info glxgears (3.36 KB, text/plain)
2018-11-08 09:06 UTC, zefkerrigan
Details
RADV_DEBUG=info vulkaninfo (1.92 KB, text/plain)
2018-11-08 09:29 UTC, zefkerrigan
Details
R600_DEBUG without the game starting (996 bytes, application/x-xz)
2018-11-08 09:39 UTC, zefkerrigan
Details

Description zefkerrigan 2018-06-25 04:33:53 UTC
Hi! With Wine Staging+DXVK some trees trembling instead of smoothly swaying from the wind. This does not happen with WineD3D or Windows.

Please, see here for more info:
https://github.com/doitsujin/dxvk/issues/450
Comment 1 ximik 2018-07-08 13:38:35 UTC
on the video on the embedded link is visible. same problem. card r290x, the driver of the message 18-18.2 git


link https://mega.nz/#!Pr5nmYTQ!uGrPyzSW32-Ln60x2jUxrvtW3VH9rG2b2uTgC1iwe18
Comment 2 zefkerrigan 2018-07-09 13:05:54 UTC
(In reply to ximik from comment #1)
> on the video on the embedded link is visible. same problem. card r290x, the
> driver of the message 18-18.2 git
> 
> 
> link https://mega.nz/#!Pr5nmYTQ!uGrPyzSW32-Ln60x2jUxrvtW3VH9rG2b2uTgC1iwe18

It's sad, but Mesa developers ignore this bug report.
Comment 3 Samuel Pitoiset 2018-07-09 13:26:34 UTC
We are not ignoring, we are just busy with something else. I will have a look as soon as possible, don't worry.
Comment 4 zefkerrigan 2018-07-09 17:33:00 UTC
(In reply to Samuel Pitoiset from comment #3)
> We are not ignoring, we are just busy with something else. I will have a
> look as soon as possible, don't worry.

Samuel, thank you!
Comment 5 ximik 2018-07-12 12:47:54 UTC
After installing a stable driver version, this bug is gone. maybe you help too
mesa 18.1.3-1 archlinux
Comment 6 Samuel Pitoiset 2018-07-12 12:51:59 UTC
Really nice, thanks! Closing.
Comment 7 zefkerrigan 2018-07-12 13:28:41 UTC
(In reply to Samuel Pitoiset from comment #6)
> Really nice, thanks! Closing.

Samuel, I do not agree with your decision to close this bug report.
If this bug was not in Mesa 18.1, but it appeared in Mesa-git 18.2, it means that some patch in Mesa-git 18.2 was added, which led to this regression. Therefore, if this regression will not fixed, it will later become part of the Mesa 18.2 stable version.
And so I think that it would be better to find the real cause of this regression instead of closing the bug report without fixing this bug.
Regards. Thank you.
Comment 8 Samuel Pitoiset 2018-07-12 14:19:41 UTC
Yeah sorry, I took the decision too quickly... Can you upload the savegame where you recorded the video please?
Comment 9 zefkerrigan 2018-07-12 14:27:54 UTC
(In reply to Samuel Pitoiset from comment #8)
> Yeah sorry, I took the decision too quickly... Can you upload the savegame
> where you recorded the video please?

I already did this before I created this bug report. Please just follow the link, there are all these files.
Comment 10 Samuel Pitoiset 2018-07-12 14:56:27 UTC
Definitely not enough of caffeine today. Apparently, the issue is SI-specific which means I can't do anything because I don't have the hardware. It works fine on Polaris/Vega.
Comment 11 zefkerrigan 2018-07-12 15:12:14 UTC
(In reply to Samuel Pitoiset from comment #10)
> Definitely not enough of caffeine today. Apparently, the issue is
> SI-specific which means I can't do anything because I don't have the
> hardware. It works fine on Polaris/Vega.

But, can I help you in any way? Or are we finally lost any hope?
Comment 12 Samuel Pitoiset 2018-07-12 15:29:12 UTC
Bisecting the issue can help.
Comment 13 zefkerrigan 2018-07-12 15:41:46 UTC
(In reply to Samuel Pitoiset from comment #12)
> Bisecting the issue can help.

And what do we need to do for this?
Comment 14 Samuel Pitoiset 2018-07-12 16:22:08 UTC
See https://bugs.freedesktop.org/show_bug.cgi?id=107156#c5
Comment 15 zefkerrigan 2018-08-26 12:09:45 UTC
I found out that this bug is only apparent when I compiled Mesa-git with llvm-svn installed.
But this bug is completely missing when I compiled Mesa-git with installed llvm 6.0.1 (stable version) .
And so what do I do? Should I open a bug report on llvm bugzilla?
Comment 16 Samuel Pitoiset 2018-08-26 13:13:20 UTC
No, I should be able to have a look in the next few weeks.
Comment 17 zefkerrigan 2018-09-30 00:50:27 UTC
(In reply to Samuel Pitoiset from comment #16)
> No, I should be able to have a look in the next few weeks.

I guess this may not be the issue that is SI-specific which means this is apparently due to the fact that only those people who use Mesa compiled using LLVM 7 have this issue, but this issue is not present in those people who use Mesa compiled using LLVM 6.
Do you have any news about some any progress on the way to fix this bug?
Thank you.
Comment 18 Samuel Pitoiset 2018-11-07 19:44:49 UTC
Are you guys still able to reproduce that problem? I'm trying to reproduce it.
Comment 19 zefkerrigan 2018-11-07 21:38:12 UTC
Created attachment 142402 [details]
The Witcher 3 savegame file in the location where easy to reproduce this bug.
Comment 20 zefkerrigan 2018-11-07 21:38:48 UTC
(In reply to Samuel Pitoiset from comment #18)
> Are you guys still able to reproduce that problem? I'm trying to reproduce
> it.

With commit 16f10230 this bug unfortunately is still present.
Comment 21 Samuel Pitoiset 2018-11-08 07:54:01 UTC
What settings are you using?
Comment 22 zefkerrigan 2018-11-08 08:01:37 UTC
Created attachment 142404 [details]
All minimal settings
Comment 23 Samuel Pitoiset 2018-11-08 08:22:39 UTC
I'm having troubles to reproduce the problem on my Pitcairn card. Can you show me the output of R600_DEBUG=info glxgears ?
Comment 24 zefkerrigan 2018-11-08 09:06:23 UTC
Created attachment 142405 [details]
The output of R600_DEBUG=info glxgears

(In reply to Samuel Pitoiset from comment #23)
> I'm having troubles to reproduce the problem on my Pitcairn card. Can you
> show me the output of R600_DEBUG=info glxgears ?
Comment 25 Samuel Pitoiset 2018-11-08 09:11:13 UTC
Well, you don't have the OpenGL driver installed? (ie. radeonsi)
Comment 26 Samuel Pitoiset 2018-11-08 09:12:48 UTC
RADV_DEBUG=info vulkaninfo please.
Comment 27 zefkerrigan 2018-11-08 09:15:45 UTC
(In reply to Samuel Pitoiset from comment #25)
> Well, you don't have the OpenGL driver installed? (ie. radeonsi)

I use Mesa-git, Wine Staging+DXVK, so RadeonSI not in use.
Comment 28 zefkerrigan 2018-11-08 09:20:23 UTC
(In reply to Samuel Pitoiset from comment #25)
> Well, you don't have the OpenGL driver installed? (ie. radeonsi)

But I haven't this issue with WineD3D/RadeonSI, only with DXVK/RADV.
Comment 29 zefkerrigan 2018-11-08 09:29:30 UTC
Created attachment 142406 [details]
RADV_DEBUG=info vulkaninfo

(In reply to Samuel Pitoiset from comment #26)
> RADV_DEBUG=info vulkaninfo please.
Comment 30 Samuel Pitoiset 2018-11-08 09:31:36 UTC
Well, the output doesn't help... It should look like https://hastebin.com/ujuyujobon
Comment 31 zefkerrigan 2018-11-08 09:39:23 UTC
Created attachment 142407 [details]
R600_DEBUG without the game starting

A try #2
Comment 32 Samuel Pitoiset 2018-11-08 09:42:43 UTC
You have to run the native version of vulkaninfo, not via wine.

"RADV_DEBUG=info vulkaninfo" in a new terminal should just work.
Comment 33 zefkerrigan 2018-11-08 09:46:10 UTC
(In reply to Samuel Pitoiset from comment #32)
> You have to run the native version of vulkaninfo, not via wine.
> 
> "RADV_DEBUG=info vulkaninfo" in a new terminal should just work.

This is exactly what I did this time. The fact was that I had misunderstood in the past times how it should have been done correctly. I believe that now I did everything right this time.
See here: https://bugs.freedesktop.org/attachment.cgi?id=142407
Comment 34 zefkerrigan 2019-04-25 21:35:26 UTC
Hi! I just tested it again, but this time using Mesa 19.0.3 with LLVM 8.0.0 and now it is fully fixed and works well, just like on Windows. This is similar to the fact that it was a LLVM regression that fixed now in LLVM 8, and not a Mesa bug at all.
Thank you very much. GOOD LUCK!


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.