[llvm] [AMDGPU] IGLP: Fixes for VMEM load detection and unsigned int handling (PR #135090)

[Robert Imschweiler via llvm-commits]

================
@@ -2079,6 +2083,9 @@ class MFMASmallGemmSingleWaveOpt final : public IGLPStrategy {
 static unsigned DSWCount = 0;
 static unsigned DSWWithPermCount = 0;
 static unsigned DSWWithSharedVMEMCount = 0;
+static void resetDSWCounters() {
+  DSWCount = DSWWithPermCount = DSWWithSharedVMEMCount = 0;
+}
----------------
ro-i wrote:

Just to be sure: That would imply that values from previous stages (`AMDGPU::SchedulingPhase`) should not be cached and that I should remove this behavior? So that, for example, `DSWWithSharedVMEMCount` is recalculated on every call to `MFMASmallGemmSingleWaveOpt::applyIGLPStrategy()`?

https://github.com/llvm/llvm-project/pull/135090