projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 0a9ff87) | patch
ARM: set default float ABI based on triple.
authorTim Northover <tnorthover@apple.com>
Wed, 18 Dec 2013 09:27:33 +0000 (09:27 +0000)
committerTim Northover <tnorthover@apple.com>
Wed, 18 Dec 2013 09:27:33 +0000 (09:27 +0000)
commite66345aed48362550467ae72af0b1cfb704b8f9f
tree9a4933467ce672a0ef661b9f57da34c50b109b17tree | snapshot
parent0a9ff8776b5fcffdb5292f261da04085ea8e116fcommit | diff
ARM: set default float ABI based on triple.

Clang sets the float-abi target option manually, but no longer
annotates each function with its ABI. This can lead to confusing
mistmatch between "clang -emit-llvm | llc" and normal clang
invocations.

Besides which, gnueabihf actually *is* hard-float. Defaulting to soft
was just perverse.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@197554 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/ARM/ARMSubtarget.h diff | blob | history
lib/Target/ARM/ARMTargetMachine.cpp diff | blob | history
test/CodeGen/ARM/default-float-abi.ll [new file with mode: 0644] blob
C/C++ LLVM-based compilers that forbids OOTA behaviors