[llvm] [AMDGPU] Validate WMMA scale/format combination (PR #203074)

[Stanislav Mekhanoshin via llvm-commits]

================
@@ -0,0 +1,20 @@
+; RUN: not llc -mtriple=amdgcn -mcpu=gfx1250 < %s 2>&1 | FileCheck -check-prefix=ERR %s
+
+; ERR: error: invalid matrix and scale format combination in wmma call
----------------
rampitec wrote:

It is user's input sanitization and the input is exact immediate values used in a call to intrinsic. Like if for matrix_a_fmt was used 1, then matrix_a_scale_fmt cannot be 1 or 2 but only 0, etc.

https://github.com/llvm/llvm-project/pull/203074