Bug 29927

Summary: [i915] glean pixelFormats fails with GLES2 enabled
Product: Mesa Reporter: jiajia <jiajia.zheng>
Component: Drivers/DRI/i915Assignee: Eric Anholt <eric>
Status: VERIFIED FIXED QA Contact:
Severity: normal    
Priority: medium    
Version: git   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments: pixelFormat output

Description jiajia 2010-09-01 01:05:30 UTC
Test environment:
-----------------
Platform: Pineview
Mesa:     (master)e3205b05407473e462f2c3668f56062da00ee0ea
          (Mon May 10 08:30:28 2010)

Problem:
If we build Mesa with "--enable-gles2", glean/pixelFormats fails.  

If we build Mesa using the below options without enabling gles2, glean/pixelFormats passes.
"./autogen.sh --prefix=/opt/X11R7 --with-driver=dri --with-dri-drivers="i915 i965 swrast" --with-dri-driverdir=/opt/X11R7/lib/dri --disable-glut --without-demos --disable-gallium"
Comment 1 jiajia 2010-09-01 01:28:05 UTC
May 19 2010 is the first day that we built Mesa with "--enable-gles2" in nightly testing. Before that time, we built Mesa without enabling gles2 and glean/pixelFormats always passed. So we are not sure which commit introduced this issue. 
But since commit 61aa9c86c20b1bc9957aa9fe9976703a93cc4d3c(May 13 10:13:52 2010), with/without "--enable-gles2", glean/pixelFormats fails. Please refer to bug 29925.
Comment 2 jiajia 2010-09-01 01:40:43 UTC
Created attachment 38351 [details]
pixelFormat output
Comment 3 Eric Anholt 2010-09-03 11:23:12 UTC
Did you do a full rebuild of Mesa with a clean tree (git clean -f -d -x is the most reliable way to do so, note that this will remove everything in your tree that isn't in git) when switching configure options?  You have to do that when changing configure options or you'll get strange behavior.
Comment 4 jiajia 2010-09-05 19:36:50 UTC
I did "git clean -f -d -x" before rebuild mesa
Comment 5 Eric Anholt 2011-07-12 16:09:11 UTC
commit 6e6b38860488a0b2b282866f095cea9860503a14
Author: Eric Anholt <eric@anholt.net>
Date:   Mon Jul 11 11:32:04 2011 -0700

    i915: Fix map/unmap mismatches from leaving INTEL_FALLBACK during TNL.
    
    The first rendering after context create didn't know of the color
    buffer yet, triggering a sw fallback.  The intel_prepare_render() from
    intelSpanRenderStart then found the buffer and turned off fallbacks,
    but intelSpanRenderFinish was never called and things were left
    mapped.  By checking buffers before making the call on whether to do
    the fallback pipeline or not, we avoid the fallback change inside of
    the rendering pipeline.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31561
    Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Comment 6 fangxun 2011-07-29 00:41:14 UTC
Verified with Mesa master 0aed27ee37860ba332df776425d89d97ca1168b2.

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.