[Bug 105333] [gallium-nine] missing geometry after commit ac: replace ac_build_kill with ac_build_kill_if_false

List overview All Threads
Download

newer

older

[PATCH 0/3] drm/omapdrm: Replace...

[PATCH v4 0/4] drm/omap:...

bugzilla-daemon＠freedesktop.org

3 Mar 2018 3 Mar '18

2:05 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

Bug ID: 105333 Summary: [gallium-nine] missing geometry after commit ac: replace ac_build_kill with ac_build_kill_if_false Product: Mesa Version: git Hardware: x86-64 (AMD64) OS: Linux (All) Status: NEW Severity: normal Priority: medium Component: Drivers/Gallium/radeonsi Assignee: dri-devel@lists.freedesktop.org Reporter: hadack@gmx.de QA Contact: dri-devel@lists.freedesktop.org

Created attachment 137765 --> https://bugs.freedesktop.org/attachment.cgi?id=137765&action=edit Xorg.0.log

Since commit 1ff9e27cbdd1fe770f18e5316c6f32a867e39095 "ac: replace ac_build_kill with ac_build_kill_if_false" the train engines in the game Mashinky under wine with nine are missing. On branch 17.3 everything is fine. I tried different llvm(5.0.1,release_60 and master), wine and kernel versions, makes no difference. This is on a RX 570.

-- You are receiving this mail because: You are the assignee for the bug.

Attachments:

attachment.html (text/html — 2.5 KB)

Show replies by date

bugzilla-daemon＠freedesktop.org

3 Mar 3 Mar

2:06 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #1 from hadack@gmx.de --- Created attachment 137766 --> https://bugs.freedesktop.org/attachment.cgi?id=137766&action=edit screenshot with mesa 17.3

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

2:07 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #2 from hadack@gmx.de --- Created attachment 137767 --> https://bugs.freedesktop.org/attachment.cgi?id=137767&action=edit screenshot with mesa git

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

10 Apr 10 Apr

7:41 a.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

iive@yahoo.com changed:

What |Removed |Added ---------------------------------------------------------------------------- Keywords| |bisected, regression

--- Comment #3 from iive@yahoo.com --- Would you make and upload an apitrace of the game demonstrating the issue?

There is an ftp for traces of Gallium Nine, you can ask in Freenode #d3d9 channel for the password, or simply use google drive or other better file sharing. The trace might be kept for future regression testing. (Unless you explicitly oppose.)

It would be nice if you also open issue at github ixit.Mesa-3D, where most of Gallium Nine bugs are tracked. Since the bug itself is outside of Nine, it is good that you have it filled here.

BTW, since the bug is regression and you've located it, be sure to put the regression and bisected tags/keywords. I just added them. It help developers if they know that the mundane and time consuming work has already been done. (I hope.)

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

3:43 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #4 from hadack@gmx.de --- Thanks for the infos! I will make a apitrace and upload it, but it will take me some days. I will then also open a report on the github page.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

21 Apr 21 Apr

11:13 a.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #5 from hadack@gmx.de --- I tried to make an apitrace, but it crashes for me when replaying. The apitrace can be downloaded here: https://www.dropbox.com/s/ynr9v7ls561qgdw/Mashinky.6.trace?dl=0 I opened a bug at the ixit github side here: https://github.com/iXit/Mesa-3D/issues/314

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

31 Aug 31 Aug

9:35 a.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

Timothy Arceri t_arceri@yahoo.com.au changed:

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

9 Sep 9 Sep

12:19 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

Axel Davy davyaxel0@gmail.com changed:

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

10:20 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #7 from Timothy Arceri t_arceri@yahoo.com.au --- (In reply to Axel Davy from comment #6)

...

Given the regression bisect found the issue came with a radeonsi patch, I reassign to radeonsi. Maybe the patch introduces a small difference in the driver behaviour that regresses nine ?

Well you can do that but it's unlikely to be investigated by anyone not interested in gallium-nine which is why I moved it there.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

11 Sep 11 Sep

6:01 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #8 from Marek Olšák maraeo@gmail.com --- ac_build_kill used an intrinsic that is no longer in LLVM. ac_build_kill_if_false replaced it. The behavior should be identical.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

6:04 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #9 from Marek Olšák maraeo@gmail.com --- You can try to set glsl_correct_derivatives_after_discard=true, but I don't know if that works with nine.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

13 Sep 13 Sep

4:48 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #10 from hadack@gmx.de ---

(In reply to Marek Olšák from comment #9)

...

You can try to set glsl_correct_derivatives_after_discard=true, but I don't know if that works with nine.

I tried with the env variable set with latest git checkout of llvm and mesa and the bug is still there. Is there anything else i could try, like getting some logs or so?

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

6:24 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #11 from Axel Davy davyaxel0@gmail.com --- (In reply to Marek Olšák from comment #9)

...

You can try to set glsl_correct_derivatives_after_discard=true, but I don't know if that works with nine.

Setting glsl_correct_derivatives_after_discard should work with nine.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

6:36 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #12 from Marek Olšák maraeo@gmail.com --- It's possible that nine ignores the env var.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

7:37 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #13 from Axel Davy davyaxel0@gmail.com --- I inserted a printf inside the if condition that calls driQueryOptionb on this option in si_pipe.c and ran a nine app, and it did print, so the env var is not ignored.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

22 Sep 22 Sep

1:10 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #14 from Axel Davy davyaxel0@gmail.com --- The app uses both alpha test and kill_if (with some kill_if inside an if). Maybe an issue combining ifs ? I see old code was doing a min between the kill conditions, not sure how it is handled now.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

2:30 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

Axel Davy davyaxel0@gmail.com changed:

What |Removed |Added ---------------------------------------------------------------------------- CC| |davyaxel0@gmail.com

--- Comment #15 from Axel Davy davyaxel0@gmail.com --- Created attachment 141686 --> https://bugs.freedesktop.org/attachment.cgi?id=141686&action=edit potentially the buggy shader

Here is a shader that may be the buggy one ?

Thanks with iive who was running a trace on his side (r600 doesn't have the bug), we identified a drawcall where the engine would appear on his side and not mine (radeonsi).

The draw calls has alpha test disabled, but has stencil test enabled. I used monolithic shader to ease reading.

It may be a false alarms, as the conditions of the test cannot guarantee this is really the first moment the two cards have a difference in their rendering (which may affect the stencil test for example).

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

2:56 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

Axel Davy davyaxel0@gmail.com changed:

--- Comment #16 from Axel Davy davyaxel0@gmail.com --- Created attachment 141687 --> https://bugs.freedesktop.org/attachment.cgi?id=141687&action=edit probably the buggy shader

Ok, so we figured out the shader I mentionned in my last message was supposed to render on top of a "white" engine which didn't appear on radeonsi.

So here is now the shader that draws (earlier) the "white" engine.

This one has alpha test enabled and stencil test enabled. There are two version of the shader in the log that get compiled (I guess two alpha test set of parameters), I don't know which is used, thus I put both.

It features on kill_if inside an if. Perhaps something to look at.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

4:56 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #17 from Axel Davy davyaxel0@gmail.com --- My investigation found that the alpha test for this shader always passes, so we can discard that (I also tried disabling it).

As for the stencil test, it is set to always pass, so it's not that either.

To play with the shader, I disabled some texkill instructions, and found that if I disabled all third texkill (texkill == kill_if) instructions in all the ps shaders, it would work. Hopefully, this is not due to some interference with some previous shader.

The third kill_if instruction in this shader is: DCL IN[4], COLOR[1], COLOR, CENTROID

31: MOV TEMP[1], IN[4].wwww 32: KILL_IF TEMP[1]

and if I can extract the ISA correctly:

v_interp_p1_f32 v4, v0, attr4.w v_interp_p2_f32 v4, v1, attr4.w v_cmpx_le_f32_e32 vcc, 0, v4

(note v_cmpx_le_f32_e32 is v_cmpx_lt_f32_e32 in the attached shader, due to a test I had done around that, but that doesn't change anything)

It doesn't look wrong, I don't understand why it causes this behaviour (it kills too much).

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

5:44 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #18 from Axel Davy davyaxel0@gmail.com --- PixWin enables to use ProcessVertices which outputs the vs shader outputs.

I looked at the output which would give the IN[4].w

If found that for llvmpipe (used for ProcessVertices) the vs output is between 0 and 1. However I tried with radeonsi, and... it displays a NaN for that output...

I guess one of the input is NaN and llvmpipe clamps it (the vs output is the subtraction between two vs inputs).

If that output is really NaN, I guess it can explain the regression. A small change (x < 0 vs 0 > x) will give different result with NaN.

I tried to change the order for the test in radeonsi's kil_emit, but the generated isa doesn't change the test order between 0 and x.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

9:05 p.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #19 from Axel Davy davyaxel0@gmail.com --- Allowing NaN to not be rejected by the kill test by replacing in radeonsi's kil_emit LLVMRealOGE by LLVMRealUGE works.

Is that ok relative to the gl spec ?

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

25 Sep 25 Sep

1:36 a.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #20 from Marek Olšák maraeo@gmail.com --- It's fine with me.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

7:39 a.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

--- Comment #21 from hadack@gmx.de --- Hi, I can confirm that the patch on the mailing list fixes the bug, thanks a lot! I will close this when the patch lands.

-- You are receiving this mail because: You are the assignee for the bug.

bugzilla-daemon＠freedesktop.org

26 Sep 26 Sep

11:17 a.m.

https://bugs.freedesktop.org/show_bug.cgi?id=105333

hadack@gmx.de changed:

-- You are receiving this mail because: You are the assignee for the bug.

2423

Age (days ago)

2630

Last active (days ago)

dri-devel@lists.freedesktop.org

23 comments

1 participants

tags (0)

participants (1)

bugzilla-daemon＠freedesktop.org