[Mesa-dev] [Bug 29407] llvm 2.8 asserts: "Tried to create an integer operation on a non-integer type!"

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=29407

--- Comment #3 from nobled <nobled at dreamwidth.org> 2010-08-05 09:31:49 PDT ---
Created an attachment (id=37590)
 View: https://bugs.freedesktop.org/attachment.cgi?id=37590
 Review: https://bugs.freedesktop.org/review?bug=29407&attachment=37590

gallivm: Always use floating point operators for floating point values (v1)

(In reply to comment #2)
> We haven't been testing with LLVM 2.8 / SVN much.  You shouldn't have trouble
> with LLVM 2.6 or 2.7.
Or I could just turn asserts off, yeah, but I was hoping to help patch the bug.

Unfortunately fixing this(or at least in the three files I looked at)
introduces regressions; with this patch the asserts are gone as far as I tested
but lp_test_format fails on PIPE_FORMAT_R8G8Bx_SNORM.

Output:
---
Testing PIPE_FORMAT_R10SG10SB10SA2U_NORM (float) ...
Testing PIPE_FORMAT_R10SG10SB10SA2U_NORM (unorm8) ...
Testing PIPE_FORMAT_R8G8Bx_SNORM (float) ...
FAILED
  Packed: 7f 00 00 00
  Unpacked (0,0): 1.000000 0.000000 0.000086 1.000000 obtained
                  1.000000 0.000000 0.000000 1.000000 expected
FAILED
  Packed: 81 00 00 00
  Unpacked (0,0): -1.000000 0.000000 0.000086 1.000000 obtained
                  -1.000000 0.000000 0.000000 1.000000 expected
FAILED
  Packed: 00 7f 00 00
  Unpacked (0,0): 0.000000 1.000000 0.000086 1.000000 obtained
                  0.000000 1.000000 0.000000 1.000000 expected
FAILED
  Packed: 00 81 00 00
  Unpacked (0,0): 0.000000 -1.000000 0.000086 1.000000 obtained
                  0.000000 -1.000000 0.000000 1.000000 expected

define void @fetch_r8g8bx_snorm_float(<4 x float>*, i8*, i32, i32) {
entry:
  %4 = alloca <4 x float>                         ; <<4 x float>*> [#uses=2]
  %5 = bitcast <4 x float>* %4 to float*          ; <float*> [#uses=1]
  %6 = bitcast i8* %1 to i8*                      ; <i8*> [#uses=1]
  call void @util_format_r8g8bx_snorm_fetch_rgba_float(float* %5, i8* %6, i32
%2, i32 %3)
  %7 = load <4 x float>* %4                       ; <<4 x float>> [#uses=1]
  store <4 x float> %7, <4 x float>* %0
  ret void
}

Testing PIPE_FORMAT_R8G8Bx_SNORM (unorm8) ...
Testing PIPE_FORMAT_R8G8B8X8_UNORM (float) ...
--

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.