[llvm-bugs] [Bug 28230] New: [mc] flat_load -- with tfe, fetch-dest size shall be inreased by 1.

[via llvm-bugs]

https://llvm.org/bugs/show_bug.cgi?id=28230

            Bug ID: 28230
           Summary: [mc] flat_load -- with tfe, fetch-dest size shall be
                    inreased by 1.
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: AMDGPU
          Assignee: artem.tamazov at amd.com
          Reporter: artem.tamazov at amd.com
                CC: llvm-bugs at lists.llvm.org, nikolay.haustov at amd.com
    Classification: Unclassified

Created attachment 16611
  --> https://llvm.org/bugs/attachment.cgi?id=16611&action=edit
00008.tests_tg_gfx7_asm_lit1_bug.06.08.min.flat_load-tfe.zip

GCN ISA spec says:
TFE -- Texel Fail Enable for PRT (partially resident textures). When set to 1,
fetch can return a NACK that causes a VGPR write into DST+1 (first GPR after
all fetch-dest GPRs).

Examples:
> flat_load_dword v2, v[0:1] // correct
> flat_load_dword v[2:3], v[0:1] tfe // fails

Tested with r273139 (git # 4a94acfea8bfd6eb6b5809dd40f03e1fa42e673a)

Some GFX7 lit tests attached.

Similar problem with flat_atomic insns.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20160621/27117bc5/attachment-0001.html>