Bug 23857

Created attachment 29405 [details]
incorrect rendering in openarena

*** Bug 23868 has been marked as a duplicate of this bug. ***

I have build latest mesa on my G45 box. It seems everything is ok.
Can you retest it?

The bug is still there, but some maps work better than others. 

To reproduce the bug, try OA_RPG3DM2 and walk close to walls and corners.

(In reply to comment #4)
> The bug is still there, but some maps work better than others. 
> To reproduce the bug, try OA_RPG3DM2 and walk close to walls and corners.

Yes. It is still very easy to reproduce on G45 with the newest code on master branch. 

fixed in 7_6 branch

Branch: mesa_7_6_branch
Commit: 76e836a41ad47237f80195c5dacb8bc57a7f2b69
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=76e836a41ad47237f80195c5dacb8bc57a7f2b69

Author: Zou Nan hai <nanhai.zou@intel.com>
Date:   Wed Sep 16 13:25:46 2009 +0800

i965: do a flush in clear, fix openarena render issue,
      fd.o bug# 23857

---

 src/mesa/drivers/dri/intel/intel_clear.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/src/mesa/drivers/dri/intel/intel_clear.c b/src/mesa/drivers/dri/intel/intel_clear.c
index bce2372..9efe6a2 100644
--- a/src/mesa/drivers/dri/intel/intel_clear.c
+++ b/src/mesa/drivers/dri/intel/intel_clear.c
@@ -170,6 +170,7 @@ intelClear(GLcontext *ctx, GLbitfield mask)
 	 }
 	 DBG("\n");
       }
+      intelFlush(&intel->ctx);
       _mesa_meta_clear(&intel->ctx, tri_mask);
    }
 

(In reply to comment #6)
> fixed in 7_6 branch

Nanhai, with your commit, its incorrect renderring has been fixed successfully, but the performance of ut2004-demo has a downgrade of 5% ~ 10%. 

Applying a hack like this is not a fix for the bug. (worse, the commit message didn't clearly state that it was only a hack around a real bug)

commit cc8084932cc552587e3036dbbf62c77db3b4a08e
Author: Eric Anholt <eric@anholt.net>
Date:   Thu Sep 24 16:15:52 2009 -0700

    intel: Flush the batch when we're about to subdata into a VBO.
    
    This fixes the clears in openarena with the new metaops clear code, and
    the new piglit vbo-subdata-sync test.
    
    Bug #23857.

There is now no rendering issues and only 1%~2% slower but I think it is acceptable. 
Libdrm:         (master)ac71f0849928f4b2fbb69c01304ac6f9df8916a1
Mesa:           (mesa_7_6_branch)cc8084932cc552587e3036dbbf62c77db3b4a08e
Xserver:        (server-1.6-branch)507e57381fea6334f7dc8da6925e53d2c76fddcb
Xf86_video_intel:    (master)a92bbcc94904684e7709b3ddaad82bc04607af26