[PATCH] D62423: [globalisel][legalizer] Attempt to write down the minimal legalization rules

[Petar Avramovic via Phabricator via llvm-commits]

Petar.Avramovic added a comment.

Producer/Consumer type set and minimum rules for scalars are assuming that set of available integer scalars is the same as set of all available floating point scalars. This does not have to be the case.
Could we define Producer/Consumer type set in terms of available sizes in register banks?
That is Producer/Consumer type set for G_ANYEXT, G_SEXT, G_ZEXT, G_TRUNC corresponds to available sizes in GPR register bank (integer instructions). 
That is Producer/Consumer type set for G_FPEXT, G_FPTRUNC corresponds to available sizes in FPR register bank (floating point instructions).

This is the problem with GMIR in general since type (integer/float) is not available in low level type, but in most cases it is possible to figure out the type from opcode (float opcodes start with G_F , integer start with G_).
There are also ambiguous opcodes that are used for both integer and floating point instructions assuming they are available for same sizes (G_PHI, G_LOAD, G_STORE, G_IMPLICIT_DEF).



================
Comment at: llvm/docs/GlobalISel.rst:422
+
+Minimum Rules For Scalars
+"""""""""""""""""""""""""
----------------
Here we are assuming that set of available integer scalars is the same as set of all available floating point scalars. This does not have to be the case. 
e.g. "-mtriple=mipsel-linux-gnu" "-mtriple=i386-linux-gnu -mattr=+sse2"(http://lists.llvm.org/pipermail/llvm-dev/2017-July/114930.html) have i32, float/double available but i64 is not available.
G_FPTRUNC is legal for {s32, s64} but G_TRUNC is not (for -mtriple=mipsel-linux-gnu I would say that it is unsupported since there are no register classes that hold less then 32 bits, and would expect artifact combiner to deal with G_TRUNC).



Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D62423/new/

https://reviews.llvm.org/D62423