From mboxrd@z Thu Jan 1 00:00:00 1970
From: bugzilla-daemon-CC+yJ3UmIYqDUpFQwHEjaQ@public.gmane.org
Subject: [Bug 111218] New: Segmentation fault in
nv50_ir::NVC0LegalizeSSA::handleDIV when dividing result of textureSize
Date: Thu, 25 Jul 2019 18:38:46 +0000
Message-ID:
Bug ID
111218
Summary
Segmentation fault in nv50_ir::NVC0LegalizeSSA::handleDIV whe=
n dividing result of textureSize
Product
Mesa
Version
19.0
Hardware
x86-64 (AMD64)
OS
Linux (All)
Status
NEW
Severity
major
Priority
medium
Component
Drivers/DRI/nouveau
Assignee
nouveau@lists.freedesktop.org
Reporter
mmgrqnv@jadamspam.pl
QA Contact
nouveau@lists.freedesktop.org
Created attachment 144869 [details]
Source of small program and somewhat minimized shaders
Hi!
I was looking into a crash[1] in latest version of Blender, which turned ou=
t to
be a crash in the Nouveau shader code generator.
The crash occurs when linking shaders. I traced the crash to the following
lines in the shader code:
ivec2 cell_co =3D ivec2(3, 2);
int cell_per_row =3D textureSize(irradianceGrid, 0).x / cell_co.x;
cell_co.x *=3D cell % cell_per_row;
cell_co.y *=3D cell / cell_per_row;
The crash seems to occur when the result of dividing textureSize() value is
later used in division or modulo operations. Replacing the textureSize() ca=
ll
with a constant avoids the crash. Replacing the division and modulo operati=
ons
with simple assignment (but keeping the first division / cell_co.x) also av=
oids
the crash.
This is the relevant stack trace:
(gdb) bt
#0 nv50_ir::NVC0LegalizeSSA::handleDIV (this=3Dthis@entry=3D0x7fffffff=
ba50,
i=3Di@entry=3D0x555555e592c0) at
../src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp:54
#1 0x00007ffff2e6b38b in nv50_ir::NVC0LegalizeSSA::visit (this=3D0x7ffffff=
fba50,
bb=3D<optimized out>) at
../src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp:334
#2 0x00007ffff2dc40d8 in nv50_ir::Pass::doRun (this=3Dthis@entry=3D0x7=
fffffffba50,
func=3D<optimized out>, ordered=3Dordered@entry=3Dfalse, skipPhi=
=3DskipPhi@entry=3Dtrue)
at ../src/gallium/drivers/nouveau/codegen/nv50_ir_bb.cpp:495
#3 0x00007ffff2dc41b4 in nv50_ir::Pass::doRun (this=3Dthis@entry=3D0x7=
fffffffba50,
prog=3Dprog@entry=3D0x555555e6e9f0, ordered=3Dordered@entry=3Dfalse,
skipPhi=3DskipPhi@entry=3Dtrue)
at ../src/gallium/drivers/nouveau/codegen/nv50_ir_bb.cpp:466
#4 0x00007ffff2dc4273 in nv50_ir::Pass::run (this=3Dthis@entry=3D0x7ff=
fffffba50,
prog=3Dprog@entry=3D0x555555e6e9f0, ordered=3Dordered@entry=3Dfalse,
skipPhi=3DskipPhi@entry=3Dtrue)
at ../src/gallium/drivers/nouveau/codegen/nv50_ir_bb.cpp:457
#5 0x00007ffff2e66dd4 in nv50_ir::TargetNVC0::runLegalizePass (this=3D<=
optimized
out>, prog=3D0x555555e6e9f0, stage=3D<optimized out>) at
../src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp:3145
#6 0x00007ffff2dc150f in nv50_ir_generate_code
(info=3Dinfo@entry=3D0x555555dcd9f0) at
../src/gallium/drivers/nouveau/codegen/nv50_ir.cpp:1265
#7 0x00007ffff2e0988a in nvc0_program_translate
(prog=3Dprog@entry=3D0x555555d79fc0, chipset=3D<optimized out>,
debug=3Ddebug@entry=3D0x5555557a74c8) at
../src/gallium/drivers/nouveau/nvc0/nvc0_program.c:624
#8 0x00007ffff2e1121d in nvc0_sp_state_create (pipe=3D0x5555557a7100,
cso=3D0x7fffffffc490, type=3D1) at
../src/gallium/drivers/nouveau/nvc0/nvc0_state.c:605
#9 0x00007ffff3042963 in st_create_fp_variant (st=3D<optimized out>,
stfp=3Dstfp@entry=3D0x555555d74db0, key=3Dkey@entry=3D0x7fffffffc63=
0) at
../src/mesa/state_tracker/st_program.c:1231
#10 0x00007ffff3045253 in st_get_fp_variant (st=3D<optimized out>,
stfp=3D0x555555d74db0, key=3D0x7fffffffc630) at
../src/mesa/state_tracker/st_program.c:1258
#11 0x00007ffff3045a7c in st_precompile_shader_variant
(st=3Dst@entry=3D0x5555557a4c10, prog=3Dprog@entry=3D0x555555d74db0=
) at
../src/mesa/state_tracker/st_program.c:1965
#12 0x00007ffff30ece0b in st_program_string_notify (ctx=3D<optimized out=
>,
target=3D<optimized out>, prog=3D0x555555d74db0) at
../src/mesa/state_tracker/st_cb_program.c:250
#13 0x00007ffff3112f85 in st_link_shader (ctx=3D0x55555578aed0,
prog=3D0x5555557b6fd0) at ../src/mesa/state_tracker/st_glsl_to_tgsi.cpp:7461
#14 0x00007ffff30b5729 in _mesa_glsl_link_shader (ctx=3Dctx@entry=3D0x5=
5555578aed0,
prog=3Dprog@entry=3D0x5555557b6fd0) at ../src/mesa/program/ir_to_mesa.c=
pp:3174
#15 0x00007ffff3000f8d in link_program (no_error=3D<optimized out>,
shProg=3D<optimized out>, ctx=3D<optimized out>) at
../src/mesa/main/shaderapi.c:1206
#16 link_program_error (ctx=3D0x55555578aed0, shProg=3D0x5555557b6fd0) at
../src/mesa/main/shaderapi.c:1286
#17 0x00005555555553f2 in test_compile ()
#18 0x000055555555586b in main ()
I attach a small program along with preprocessed and *somewhat* minimized
shader code that reproduces the problem on my computer. This code crashes 1=
00%
of the time.
To reproduce the crash:
$ gcc -Wall -o shad shad.c -lX11 -lGL -lGLU -lGLEW
$ ./shad s0vert.i s1frag.i s2geom.i
Compilation successful!
Segmentation fault
System information:
Ubuntu 18.04.2 LTS
Linux-4.15.0-55-generic-x86_64
Graphics card: NVIDIA Corporation GF108 [GeForce GT 730] (rev a1)
Graphics card driver: NVC1 nouveau 4.3 (Core Profile) Mesa 19.0.2
Using GNOME under X
libgl1-mesa-dri: 19.0.2-1ubuntu1.1~18.04.2
The original Blender crash report along with full shader source dump can be
found here:
[1] https://developer.blen=
der.org/T67534