projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3de0a38) | patch
Do not assume that FP vector operands are never legalized by expanding
authorNemanja Ivanovic <nemanja.i.ibm@gmail.com>
Wed, 26 Oct 2016 19:51:35 +0000 (19:51 +0000)
committerNemanja Ivanovic <nemanja.i.ibm@gmail.com>
Wed, 26 Oct 2016 19:51:35 +0000 (19:51 +0000)
commit275853e777a2a2f7069d7b10b61ebc80bc699d5f
treea751f5b48cc0f87d3f2c49fb282a8d10801aaeaftree | snapshot
parent3de0a385c96fa94dd7ed3c1df8b10dd65cf738c8commit | diff
Do not assume that FP vector operands are never legalized by expanding

This patch ensures that if a floating point vector operand is legalized by
expanding, it is legalized through the stack rather than by calling
DAGTypeLegalizer::IntegerToVector which will cause a failure since the operand
is a non-integer type.

This fixes PR 30715.

llvm-svn: 285231
llvm/lib/CodeGen/SelectionDAG/LegalizeTypesGeneric.cpp diff | blob | history
llvm/test/CodeGen/PowerPC/pr30715.ll [new file with mode: 0644] blob
Domain: System / Toolchain;