On 03/09/2012 08:38 AM, Vic Lee wrote:
On 03/09/2012 11:19 PM, Brian Paul wrote:
Vic, it looks like your code is slightly out of date. In current
git/master, line 148 reads:
unsigned generic_inputs =
svga->curr.fs ? svga->curr.fs->generic_inputs : 0;
Zack added the null ptr check in January wi
On 03/09/2012 11:19 PM, Brian Paul wrote:
Vic, it looks like your code is slightly out of date. In current
git/master, line 148 reads:
unsigned generic_inputs =
svga->curr.fs ? svga->curr.fs->generic_inputs : 0;
Zack added the null ptr check in January with commit cff0eac702
On 03/08/2012 08:16 AM, Vic Lee wrote:
Hi Brian,
Please the the gdb backtrace if I only add vs check. Maybe you have a
better idea on this issue. I am using the git 8.0 branch.
vic@debian:~/mesa/src/gallium$ git diff
diff --git a/src/gallium/drivers/svga/svga_st
Hi Brian,
Please the the gdb backtrace if I only add vs check. Maybe you have a
better idea on this issue. I am using the git 8.0 branch.
vic@debian:~/mesa/src/gallium$ git diff
diff --git a/src/gallium/drivers/svga/svga_state_need_swtnl.c
b/src/gallium/driver
On 03/07/2012 10:01 AM, Vic Lee wrote:
Hi Brian,
On 03/07/2012 11:44 PM, Brian Paul wrote:
On 03/07/2012 01:08 AM, Vic Lee wrote:
Hi Brian,
I created a smallest possible test case to reproduce the segfault, as
attached.
OK, I was assuming that you found this with an OpenGL program, not with
Hi Brian,
On 03/07/2012 11:44 PM, Brian Paul wrote:
On 03/07/2012 01:08 AM, Vic Lee wrote:
Hi Brian,
I created a smallest possible test case to reproduce the segfault, as
attached.
OK, I was assuming that you found this with an OpenGL program, not with
a custom direct-to-gallium program.
Rig
On 03/07/2012 01:08 AM, Vic Lee wrote:
Hi Brian,
I created a smallest possible test case to reproduce the segfault, as
attached.
OK, I was assuming that you found this with an OpenGL program, not
with a custom direct-to-gallium program.
Please see file header on how I compile it. vs and f
Hi Brian,
I created a smallest possible test case to reproduce the segfault, as
attached. Please see file header on how I compile it. vs and fs both
needs to be checked to avoid the segfault.
If I replace the screen creation part to use swrast or r600, they both
work fine.
Thanks.
Vic
On
On 03/03/2012 04:17 AM, Vic Lee wrote:
This patch fix segfault if an application clears a render target
before setting any vertex or fragment shaders, which is a legal
situation.
I can't seem to trigger this problem with a test program. Can you
post a test case?
The first part of your patch
This patch fix segfault if an application clears a render target before
setting any vertex or fragment shaders, which is a legal situation.
Vic
>From 1de6aa23d3e472db7deb819e52b1f75ac09fca55 Mon Sep 17 00:00:00 2001
From: Vic Lee
Date: Sat, 3 Mar 2012 05:57:16 -0500
Subject: [PATCH] svga: Check
10 matches
Mail list logo
