[PATCH] D54095: [X86] X86DAGToDAGISel::matchBitExtract(): extract 'lshr' from `X`

[Roman Lebedev via Phabricator via llvm-commits]

lebedev.ri updated this revision to Diff 174038.
lebedev.ri marked an inline comment as done.
lebedev.ri added a comment.

In https://reviews.llvm.org/D54095#1298444, @andreadb wrote:

> ...
>  I think the problem is that you are constructing a 8-bit quantity by inserting on an implicit definition, instead of a zero.


Yes. Sorry about that.
I have now resorted to truncating to i8, zero-extending to i16, and inserting into the proper [undef] register.
On these tests it has the same effect as directly zero-extending to the proper register size, so not sure maybe i'm again over-engineering this?


Repository:
  rL LLVM

https://reviews.llvm.org/D54095

Files:
  lib/Target/X86/X86ISelDAGToDAG.cpp
  test/CodeGen/X86/extract-bits.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D54095.174038.patch
Type: text/x-patch
Size: 32200 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20181114/370dab76/attachment.bin>