FEDORA-2018-bcdb732491 created by ignatenkobrain a year ago for Fedora 29
stable
  • Corrupted frame contents with Vulkan version of DOTA2, Talos Principle and Sascha Willems' demos when they're run Vsynched in fullscreen
  • GPU hang - GS_EMIT without shader outputs
  • [anv] Rise of the Tomb Raider always misrendering, segfault and gpu hang.
  • Compiler crashes on access of non-existent member incremental operations

How to install

sudo dnf upgrade --advisory=FEDORA-2018-bcdb732491

This update has been submitted for testing by ignatenkobrain.

a year ago

This update has been pushed to testing.

a year ago
User Icon cserpentis commented & provided feedback a year ago
karma

works for me

User Icon renault commented & provided feedback a year ago
karma

Works fine, no regressions found

User Icon kparal commented & provided feedback a year ago
karma

gnome desktop works fine, basic opengl and vulkan demos work fine. Intel 620 gpu.

This update has been submitted for batched by bodhi.

a year ago

This update has been submitted for stable by bodhi.

a year ago
User Icon ozeszty commented & provided feedback a year ago
karma

No regressions spotted.

User Icon carlwgeorge commented & provided feedback a year ago
karma

Gnome works fine, but this update wrecks the graphics in Starcraft 2 under wine 3.19. They become extremely choppy and many object fail to render correctly.

User Icon eclipseo commented & provided feedback a year ago
karma

I have a regression with mesa-dri-drivers 12.2.3-1

Using mpv with OpenGL (--vo=gpu) I get the following error:

mpv drama_115846.mp4-720.mp4 21:31:21 Playing: drama_115846.mp4-720.mp4 (+) Video --vid=1 () (h264 1280x720 25.000fps) (+) Audio --aid=1 --alang=eng () (aac 2ch 48000Hz) Cannot load libcuda.so.1 Using hardware decoding (vaapi). VO: [gpu] 1280x720 vaapi[nv12] [vo/gpu/opengl] fragment shader source: [vo/gpu/opengl] [ 1] #version 440 [vo/gpu/opengl] [ 2] #define tex1D texture [vo/gpu/opengl] [ 3] #define tex3D texture [vo/gpu/opengl] [ 4] #define LUT_POS(x, lut_size) mix(0.5 / (lut_size), 1.0 - 0.5 / (lut_size), (x)) [vo/gpu/opengl] [ 5] out vec4 out_color; [vo/gpu/opengl] [ 6] in vec2 texcoord0; [vo/gpu/opengl] [ 7] layout(std140, binding=0) uniform UBO { [vo/gpu/opengl] [ 8] layout(offset=0) vec2 texture_size0; [vo/gpu/opengl] [ 9] layout(offset=16) mat2 texture_rot0; [vo/gpu/opengl] [ 10] layout(offset=48) vec2 texture_off0; [vo/gpu/opengl] [ 11] layout(offset=56) vec2 pixel_size0; [vo/gpu/opengl] [ 12] }; [vo/gpu/opengl] [ 13] uniform float random;
[vo/gpu/opengl] [ 14] uniform sampler2D texture0;
[vo/gpu/opengl] [ 15] #define HOOKED_raw texture0
[vo/gpu/opengl] [ 16] #define HOOKED_pos texcoord0
[vo/gpu/opengl] [ 17] #define HOOKED_size texture_size0
[vo/gpu/opengl] [ 18] #define HOOKED_rot texture_rot0
[vo/gpu/opengl] [ 19] #define HOOKED_pt pixel_size0
[vo/gpu/opengl] [ 20] #define HOOKED_map texmap0
[vo/gpu/opengl] [ 21] #define HOOKED_mul 1.000000
[vo/gpu/opengl] [ 22] #define HOOKED_tex(pos) (HOOKED_mul * vec4(texture(HOOKED_raw, pos)).rgba)
[vo/gpu/opengl] [ 23] #define HOOKED_texOff(off) HOOKED_tex(HOOKED_pos + HOOKED_pt * vec2(off))
[vo/gpu/opengl] [ 24] #define LUMA_raw texture0
[vo/gpu/opengl] [ 25] #define LUMA_pos texcoord0
[vo/gpu/opengl] [ 26] #define LUMA_size texture_size0
[vo/gpu/opengl] [ 27] #define LUMA_rot texture_rot0
[vo/gpu/opengl] [ 28] #define LUMA_pt pixel_size0
[vo/gpu/opengl] [ 29] #define LUMA_map texmap0
[vo/gpu/opengl] [ 30] #define LUMA_mul 1.000000
[vo/gpu/opengl] [ 31] #define LUMA_tex(pos) (LUMA_mul * vec4(texture(LUMA_raw, pos)).rgba)
[vo/gpu/opengl] [ 32] #define LUMA_texOff(off) LUMA_tex(LUMA_pos + LUMA_pt * vec2(off))
[vo/gpu/opengl] [ 33] float mod289(float x) { return x - floor(x * 1.0/289.0) * 289.0; }
[vo/gpu/opengl] [ 34] float permute(float x) { return mod289((34.0x + 1.0) * x); }
[vo/gpu/opengl] [ 35] float rand(float x) { return fract(x * 1.0/41.0); }
[vo/gpu/opengl] [ 36] vec4 average(float range, inout float h) {
[vo/gpu/opengl] [ 37] float dist = rand(h) * range; h = permute(h);
[vo/gpu/opengl] [ 38] float dir = rand(h) * 6.2831853; h = permute(h);
[vo/gpu/opengl] [ 39] vec2 o = dist * vec2(cos(dir), sin(dir));
[vo/gpu/opengl] [ 40] vec4 ref[4];
[vo/gpu/opengl] [ 41] ref[0] = HOOKED_texOff(vec2( o.x, o.y));
[vo/gpu/opengl] [ 42] ref[1] = HOOKED_texOff(vec2(-o.y, o.x));
[vo/gpu/opengl] [ 43] ref[2] = HOOKED_texOff(vec2(-o.x, -o.y));
[vo/gpu/opengl] [ 44] ref[3] = HOOKED_texOff(vec2( o.y, -o.x));
[vo/gpu/opengl] [ 45] return (ref[0] + ref[1] + ref[2] + ref[3])
0.25;
[vo/gpu/opengl] [ 46] }
[vo/gpu/opengl] [ 47] void main() {
[vo/gpu/opengl] [ 48] vec4 color = vec4(0.0, 0.0, 0.0, 1.0);
[vo/gpu/opengl] [ 49] {
[vo/gpu/opengl] [ 50] vec3 _m = vec3(HOOKED_pos, random) + vec3(1.0);
[vo/gpu/opengl] [ 51] float h = permute(permute(permute(_m.x)+_m.y)+_m.z);
[vo/gpu/opengl] [ 52] color = HOOKED_tex(HOOKED_pos);
[vo/gpu/opengl] [ 53] vec4 avg, diff;
[vo/gpu/opengl] [ 54] avg = average(16.000000, h);
[vo/gpu/opengl] [ 55] diff = abs(color - avg);
[vo/gpu/opengl] [ 56] color = mix(avg, color, greaterThan(diff, vec4(0.003906)));
[vo/gpu/opengl] [ 57] vec3 noise;
[vo/gpu/opengl] [ 58] noise.x = rand(h); h = permute(h);
[vo/gpu/opengl] [ 59] noise.y = rand(h); h = permute(h);
[vo/gpu/opengl] [ 60] noise.z = rand(h); h = permute(h);
[vo/gpu/opengl] [ 61] color.xyz += 0.005859 * (noise - vec3(0.5));
[vo/gpu/opengl] [ 62] }
[vo/gpu/opengl] [ 63] color.g = 0.000000;
[vo/gpu/opengl] [ 64] color.b = 0.000000;
[vo/gpu/opengl] [ 65] color.a = 1.000000;
[vo/gpu/opengl] [ 66] out_color = color;
[vo/gpu/opengl] [ 67] }
[vo/gpu/opengl] fragment shader compile log (status=0):
[vo/gpu/opengl] 0:36(27): error: invalid input layout qualifier used
[vo/gpu/opengl]
[vo/gpu/opengl] shader link log (status=0): error: linking with uncompiled/unspecialized shader
[vo/gpu/opengl] fragment shader source:
[vo/gpu/opengl] [ 1] #version 440
[vo/gpu/opengl] [ 2] #define tex1D texture
[vo/gpu/opengl] [ 3] #define tex3D texture
[vo/gpu/opengl] [ 4] #define LUT_POS(x, lut_size) mix(0.5 / (lut_size), 1.0 - 0.5 / (lut_size), (x))
[vo/gpu/opengl] [ 5] out vec4 out_color;
[vo/gpu/opengl] [ 6] in vec2 texcoord0;
[vo/gpu/opengl] [ 7] layout(std140, binding=0) uniform UBO {
[vo/gpu/opengl] [ 8] layout(offset=0) vec2 texture_size0;
[vo/gpu/opengl] [ 9] layout(offset=16) mat2 texture_rot0;
[vo/gpu/opengl] [ 10] layout(offset=48) vec2 texture_off0;
[vo/gpu/opengl] [ 11] layout(offset=56) vec2 pixel_size0;
[vo/gpu/opengl] [ 12] };
[vo/gpu/opengl] [ 13] uniform float random;
[vo/gpu/opengl] [ 14] uniform sampler2D texture0;
[vo/gpu/opengl] [ 15] #define HOOKED_raw texture0
[vo/gpu/opengl] [ 16] #define HOOKED_pos texcoord0
[vo/gpu/opengl] [ 17] #define HOOKED_size texture_size0
[vo/gpu/opengl] [ 18] #define HOOKED_rot texture_rot0
[vo/gpu/opengl] [ 19] #define HOOKED_pt pixel_size0
[vo/gpu/opengl] [ 20] #define HOOKED_map texmap0
[vo/gpu/opengl] [ 21] #define HOOKED_mul 1.000000
[vo/gpu/opengl] [ 22] #define HOOKED_tex(pos) (HOOKED_mul * vec4(texture(HOOKED_raw, pos)).rgba)
[vo/gpu/opengl] [ 23] #define HOOKED_texOff(off) HOOKED_tex(HOOKED_pos + HOOKED_pt * vec2(off))
[vo/gpu/opengl] [ 24] #define CHROMA_raw texture0
[vo/gpu/opengl] [ 25] #define CHROMA_pos texcoord0
[vo/gpu/opengl] [ 26] #define CHROMA_size texture_size0
[vo/gpu/opengl] [ 27] #define CHROMA_rot texture_rot0
[vo/gpu/opengl] [ 28] #define CHROMA_pt pixel_size0
[vo/gpu/opengl] [ 29] #define CHROMA_map texmap0
[vo/gpu/opengl] [ 30] #define CHROMA_mul 1.000000
[vo/gpu/opengl] [ 31] #define CHROMA_tex(pos) (CHROMA_mul * vec4(texture(CHROMA_raw, pos)).rgba)
[vo/gpu/opengl] [ 32] #define CHROMA_texOff(off) CHROMA_tex(CHROMA_pos + CHROMA_pt * vec2(off))
[vo/gpu/opengl] [ 33] float mod289(float x) { return x - floor(x * 1.0/289.0) * 289.0; }
[vo/gpu/opengl] [ 34] float permute(float x) { return mod289((34.0x + 1.0) * x); }
[vo/gpu/opengl] [ 35] float rand(float x) { return fract(x * 1.0/41.0); }
[vo/gpu/opengl] [ 36] vec4 average(float range, inout float h) {
[vo/gpu/opengl] [ 37] float dist = rand(h) * range; h = permute(h);
[vo/gpu/opengl] [ 38] float dir = rand(h) * 6.2831853; h = permute(h);
[vo/gpu/opengl] [ 39] vec2 o = dist * vec2(cos(dir), sin(dir));
[vo/gpu/opengl] [ 40] vec4 ref[4]; [vo/gpu/opengl] [ 41] ref[0] = HOOKED_texOff(vec2( o.x, o.y)); [vo/gpu/opengl] [ 42] ref[1] = HOOKED_texOff(vec2(-o.y, o.x)); [vo/gpu/opengl] [ 43] ref[2] = HOOKED_texOff(vec2(-o.x, -o.y)); [vo/gpu/opengl] [ 44] ref[3] = HOOKED_texOff(vec2( o.y, -o.x)); [vo/gpu/opengl] [ 45] return (ref[0] + ref[1] + ref[2] + ref[3])
0.25; [vo/gpu/opengl] [ 46] } [vo/gpu/opengl] [ 47] void main() { [vo/gpu/opengl] [ 48] vec4 color = vec4(0.0, 0.0, 0.0, 1.0); [vo/gpu/opengl] [ 49] { [vo/gpu/opengl] [ 50] vec3 _m = vec3(HOOKED_pos, random) + vec3(1.0); [vo/gpu/opengl] [ 51] float h = permute(permute(permute(_m.x)+_m.y)+_m.z); [vo/gpu/opengl] [ 52] color = HOOKED_tex(HOOKED_pos); [vo/gpu/opengl] [ 53] vec4 avg, diff; [vo/gpu/opengl] [ 54] avg = average(16.000000, h); [vo/gpu/opengl] [ 55] diff = abs(color - avg); [vo/gpu/opengl] [ 56] color = mix(avg, color, greaterThan(diff, vec4(0.003906))); [vo/gpu/opengl] [ 57] vec3 noise; [vo/gpu/opengl] [ 58] noise.x = rand(h); h = permute(h); [vo/gpu/opengl] [ 59] noise.y = rand(h); h = permute(h); [vo/gpu/opengl] [ 60] noise.z = rand(h); h = permute(h); [vo/gpu/opengl] [ 61] color.xyz += 0.005859 * (noise - vec3(0.5)); [vo/gpu/opengl] [ 62] } [vo/gpu/opengl] [ 63] color.b = 0.000000; [vo/gpu/opengl] [ 64] color.a = 1.000000; [vo/gpu/opengl] [ 65] out_color = color; [vo/gpu/opengl] [ 66] } [vo/gpu/opengl] fragment shader compile log (status=0): [vo/gpu/opengl] 0:36(27): error: invalid input layout qualifier used [vo/gpu/opengl] [vo/gpu/opengl] shader link log (status=0): error: linking with uncompiled/unspecialized shader

Downgrading mesa-dri-drivers solve the issue.

This update has been pushed to stable.

a year ago

I thought that bodhi is supposed to stop auto-pushing once someone gave -1... It seems to not be the case any longer.

@eclipseo, @carlwgeorge, please open bugs in upstream mesa and give me a link.

@ignatenkobrain https://bugs.freedesktop.org/show_bug.cgi?id=108646

I try to bisect the issue unsuccessfully, I'm not even sure the issue is Mesa… If I build locally i965_dri.so 18.2.2-1, it fails too, whereas the Fedora package works. My conclusion is that a dependency used during the build was working when Fedora Mesa 18.2.2-1 was built on Oct 5 but that dependency was updated and i965_dri.so 18.2.2 when built now with that updated dependency fails.

User Icon equeim commented & provided feedback a year ago
karma

I have a problem with Plasma and Mesa 18.2.3. When compositing is enabled, transparency in Plasma doesn't work, instead it is white background. Downgrading to 18.2.2 solved the issue.

Seems the culprit is gcc: compiled with gcc-8.2.1-2 it works, with gcc-8.2.1-4 it doesn't.

Any chance to revert this update? It's also breaking Godot: https://bugzilla.redhat.com/show_bug.cgi?id=1646888

@ignatenkobrain I believe -1 karma will disable autopush, but that only has an effect before the build is actually submitted stable, either manually or using autopush. In this case, it was submitted already. But I strongly recommend to not use autopush for critical path upgrades in the first place, and always submit them manually after a sufficient time has passed and no negative feedback has been received. Alternatively, if you insist on using autopush, just set the default karma level really high, e.g. to 10 or 15, to make sure this gets sufficient testing. Sending new mesa to stable after just a day with just 3 upvotes is way too fast.


Please login to add feedback.

Metadata
Type
bugfix
Karma
1
Signed
Content Type
RPM
Test Gating
Settings
Unstable by Karma
-3
Stable by Karma
3
Dates
submitted
a year ago
in testing
a year ago
in stable
a year ago

Automated Test Results