Bug 38584

I've posted two patches to the mesa-dev mailing list.  Could you verify whether or not they fix this issue for you?  The second patch should fix the segfault that you hit, and the first patch fixes a similar issue that I hit.

http://marc.info/?l=mesa3d-dev&m=130896214323707&w=2
http://marc.info/?l=mesa3d-dev&m=130896218123735&w=2

(In reply to comment #1)
> I've posted two patches to the mesa-dev mailing list.  Could you verify whether
> or not they fix this issue for you?  The second patch should fix the segfault
> that you hit, and the first patch fixes a similar issue that I hit.
> 
> http://marc.info/?l=mesa3d-dev&m=130896214323707&w=2
> http://marc.info/?l=mesa3d-dev&m=130896218123735&w=2

no such luck, it's var->name that's NULL. sorry for not looking into it deeper earlier; i got it work-arounded with

diff --git a/src/glsl/ir_print_visitor.cpp b/src/glsl/ir_print_visitor.cpp
index 5248330..20db268 100644
--- a/src/glsl/ir_print_visitor.cpp
+++ b/src/glsl/ir_print_visitor.cpp
@@ -108,6 +108,9 @@ ir_print_visitor::unique_name(ir_variable *var)
    if (name != NULL)
       return name;
 
+   if (!var->name) {
+      return "<<var->name == NULL>>";
+   }
    /* If there's no conflict, just use the original name */
    if (_mesa_symbol_table_find_symbol(this->symbols, -1, var->name) == NULL) {
       name = var->name;


... and it gets triggered like this:
(function order_ps_input
  (signature void
    (parameters
      (declare (in ) (array vec4 12) <<var->name == NULL>>)
    )
    (
    ))

(it's always order_ps_input, the GLSL source says:
#version 120
uniform vec4 VC[256];
uniform vec4 posFixup;
void order_ps_input(in vec4[12]);
ivec4 A0;
vec4 OUT[12];
[...]

(In reply to comment #2)
> (In reply to comment #1)
> > I've posted two patches to the mesa-dev mailing list.  Could you verify whether
> > or not they fix this issue for you?  The second patch should fix the segfault
> > that you hit, and the first patch fixes a similar issue that I hit.
> > 
> > http://marc.info/?l=mesa3d-dev&m=130896214323707&w=2
> > http://marc.info/?l=mesa3d-dev&m=130896218123735&w=2
> 
> no such luck, it's var->name that's NULL. sorry for not looking into it deeper
> earlier; i got it work-arounded with

Nope, I should have been paying more attention.  I should have also run it on the provided shader after making my fix (instead of just running it before).  How about:

http://marc.info/?l=mesa3d-dev&m=130920996000346&w=2
http://marc.info/?l=mesa3d-dev&m=130920999500426&w=2

Fixed on master by the commits below.  These have been cherry-picked to 7.11 (f28cf18 and b8972db) and 7.10 (7424f9c and cd73c06).

commit 174cef7fee7d400fc89a3ce68b7791d2aa3eb90f
Author: Ian Romanick <ian.d.romanick@intel.com>
Date:   Fri Jun 24 17:30:41 2011 -0700

    glsl: Don't choke when printing an anonymous function parameter
    
    NOTE: This is a candidate for the 7.10 and 7.11 branches.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38584
    Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>

commit dbda466fc05a6262ba857a7887e16347cf3d3e96
Author: Ian Romanick <ian.d.romanick@intel.com>
Date:   Fri Jun 24 17:12:31 2011 -0700

    ir_to_mesa: Allocate temporary instructions on the visitor's ralloc context
    
    And don't delete them.  Let ralloc clean them up.  Deleting the
    temporary IR leaves dangling references in the prog_instruction.  That
    results in a bad dereference when printing the IR with MESA_GLSL=dump.
    
    NOTE: This is a candidate for the 7.10 and 7.11 branches.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=38584
    Reviewed-by: Eric Anholt <eric@anholt.net>
    Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>