[PATCH] D142636: AMDGPU/MC: Refactor decoders. Rework decoders for float immediates

Tue Jan 31 05:38:00 PST 2023

Petar.Avramovic added inline comments.

================
Comment at: llvm/lib/Target/AMDGPU/Disassembler/AMDGPUDisassembler.cpp:1489

-  // ToDo: case 248: 1/(2*PI) - is allowed only on VI
-  switch (Width) {
----------------
Joe_Nash wrote:
> This comment should probably be maintained?
I should not have deleted that comment.
This patch does not affect since it does not considers values of Imm (case 248).
Apart from that it is most probably fine, case 248 is available on vi and gfx9+, before it was listed as reserved. 
There is a feature for it FeatureInv2PiInlineImm that is used in a few places 

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142636/new/

https://reviews.llvm.org/D142636