[llvm] [AMDGPU] Allow negative offsets in scratch instructions (PR #166979)

[Shilei Tian via llvm-commits]

================
@@ -1588,10 +1582,10 @@ def FeatureGFX9 : GCNSubtargetFeatureGeneration<"GFX9",
    FeatureScalarFlatScratchInsts, FeatureScalarAtomics, FeatureR128A16,
    FeatureA16, FeatureSMemTimeInst, FeatureFastDenormalF32, FeatureSupportsXNACK,
    FeatureUnalignedBufferAccess, FeatureUnalignedScratchAccess,
-   FeatureUnalignedDSAccess, FeatureNegativeScratchOffsetBug, FeatureGWS,
-   FeatureDefaultComponentZero,FeatureVmemWriteVgprInOrder, FeatureMemToLDSLoad,
-   FeatureCubeInsts, FeatureLerpInst, FeatureSadInsts, FeatureQsadInsts,
-   FeatureCvtNormInsts, FeatureCvtPkNormVOP2Insts,
+   FeatureUnalignedDSAccess, FeatureGWS, FeatureDefaultComponentZero,
+   FeatureVmemWriteVgprInOrder, FeatureMemToLDSLoad,
+   FeatureCubeInsts, FeatureLerpInst, FeatureSadInsts, FeatureQsadInsts, 
+   FeatureCvtNormInsts, FeatureCvtPkNormVOP2Insts, 
----------------
shiltian wrote:

don't add trailing whitespaces

https://github.com/llvm/llvm-project/pull/166979