[PATCH] D34559: [X86][DAG] Switch X86 Target to post-legalized store merge

[Nirav Dave via Phabricator via llvm-commits]

niravd marked an inline comment as done.
niravd added inline comments.


================
Comment at: test/CodeGen/X86/extract-store.ll:528
+; SSE-X32-NEXT:    popl %edi
+; SSE-X32-NEXT:    retl
 ;
----------------
RKSimon wrote:
> Any idea why this still doesn't merge? Same for @extract_f128_1 as well.
At merge time the component stores have an alignment of 1 and TLI.allowsMisalignedMemoryAccess believe this is legal but slow because  Subtarget.isUnalignedMem16Slow() is set so we don't bother merging. 







https://reviews.llvm.org/D34559