| Summary: | 3D textures don't work on R300/R400/R500 DRI | ||
|---|---|---|---|
| Product: | Mesa | Reporter: | Aditya Kadambi <rakadambi> |
| Component: | Drivers/DRI/r300 | Assignee: | Default DRI bug account <dri-devel> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | medium | CC: | MostAwesomedude, nhaehnle |
| Version: | 7.2 | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: |
Simple 3D texture code that I found on the web URL I mentioned in the report
The screenshot of the sample 3D texture code. Shows just red blob. Screenshot of output from OpenRM sample application |
||
|
Description
Aditya Kadambi
2008-08-12 13:47:07 UTC
Created attachment 18244 [details]
The screenshot of the sample 3D texture code. Shows just red blob.
Created attachment 18245 [details]
Screenshot of output from OpenRM sample application
Thanks for the report. I'm not really surprised that 3D textures don't work, we should add an automatic test to Piglit for this. Can I help with any testing? debugging? this should be fixed in mesa git: 57acb0fad9e6299eb2e1e2c06d043b2a418e3732 (In reply to comment #5) > this should be fixed in mesa git: > 57acb0fad9e6299eb2e1e2c06d043b2a418e3732 Right, that fixed mesa/progs/demos/stex3d on my RV350. R5xx may need more work. (In reply to comment #6) > > Right, that fixed mesa/progs/demos/stex3d on my RV350. R5xx may need more work. > Yeah, doesn't fix it on my X1700. One sec, lemme try some stuff. Just tested my code on R300 (X800 PCIE) card. The example programs work great. I will test this on R500 (X1650 AGP) and report back. Thanks for the fixes. The demo programs work fine on X1650 too. I don't think they are very different from 1700, are they? (In reply to comment #9) > The demo programs work fine on X1650 too. > I don't think they are very different from 1700, are they? > Should be the same. Looking at the docs the 3d texture stuff is the same across r3xx-r5xx, so I think we are all set. They're quite similar. I've added a testcase for Piglit in which I've found that there are still problems that must be due to texture memory alignment. The most common cases should work for the time being. I'll take a look at it in the r300-bufmgr branch since I'm reworking texture handling anyway. Great! do you suggest I resolve this bug? or leave it open for your fixes? I get a SW fallback message sometimes: *********************************WARN_ONCE********************************* File r300_render.c function r300Fallback line 451 Software fallback:ctx->RenderMode != GL_RENDER *************************************************************************** This happens when I have multiple viewports in a single window. I suspect this is probably an artifact. It doesn't seem to fallback to SW rendering after this message. Because SW rendering of these volumes would be unusable. The transformations after this message just seem fine. That warning is not a bug (maybe we should just remove it or move it to a _mesa_debug to stop getting those reports). It indicates that the application sets RenderMode to SELECT or FEEDBACK, which we can't do in hardware. It is therefore going to be slower than other things, but it should work. In any case, that message itself is not an indication of a problem other than a potential performance problem. Well, I just tested this on my new ATI X1950Pro (RV580) card and it doesn't work. So ATI x800 PCIE (R480) -> Works ATI x1650Pro AGP -> Works ATI x1950Pro (RV580) -> Does _not_ work. I tried with the same programs as before. Just tested with 7.2.0.14.fc10 on Fedora 10 X86_84 system. Still doesn't work on X1950Pro card. I ran Piglit: The tex3d test failed: @test: texturing/tex3d info: @@@Returncode: 1 Errors: Render 3D texture: Mismatch at 0x0x1 Expected: 87,205,222,213 Readback: 120,152,34,125 Failure with texture size 1x1x2, format = GL_RGBA Output: errors! Render 3D texture: Mismatch at 0x0x1 Expected: 87,205,222,213 Readback: 120,152,34,125 Failure with texture size 1x1x2, format = GL_RGBA note: Returncode was 1 returncode: 1 result: fail time: 0.088436126709 I confirm it. Looks like only 3d textures with rowstride < 64 bytes are affected (found with mesa/progs/demos/stex3d). Fixed in F11 Mesa 7.5-Devel The small 3D textures are fixed now (commit e0eafde746fee546b16205e4a605e28f78e120f9). Wow! thanks! |
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.