[PATCH] D82983: AMDGPU: Move handling of AGPR copies to a separate function

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: rampitec, kzhuravl.
Herald added subscribers: kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely.
Herald added a project: LLVM.

This is in preparation for fixing multiple problems with the way AGPR
copies are handled, but this change is NFC itself. First, it's relying
on recursively calling copyPhysReg, which is losing information
necessary to get correct super register handling.

      

Second, it's constructing a new RegScavenger and doing a O(N^2) walk
on every single sub-spill for every AGPR tuple copy. Third, it's using
the forward form of the scavenger, and not using the preferred
backwards scan.


https://reviews.llvm.org/D82983

Files:
  llvm/lib/Target/AMDGPU/SIInstrInfo.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D82983.274856.patch
Type: text/x-patch
Size: 6312 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200701/7b9670b4/attachment-0001.bin>