Stop uppercasing build attribute data.
authorCharlie Turner <charlie.turner@arm.com>
Thu, 27 Nov 2014 12:13:56 +0000 (12:13 +0000)
committerCharlie Turner <charlie.turner@arm.com>
Thu, 27 Nov 2014 12:13:56 +0000 (12:13 +0000)
commit8d43369163182f5176ed6d9f652a227dc0d7e4ce
tree24cdac65c25bc27012b3e31e64af2db6a2548bd4
parentf8516e16629e0cbbfcb201f6ae48a1cb66a36757
Stop uppercasing build attribute data.

The string data for string-valued build attributes were being unconditionally
uppercased. There is no mention in the ARM ABI addenda about case conventions,
so it's technically implementation defined as to whether the data are
capitialised in some way or not. However, there are good reasons not to
captialise the data.

  * It's less work.
  * Some vendors may legitimately have case-sensitive checks for these
    attributes which would fail on LLVM generated object files.
  * There could be locale issues with uppercasing.

The original reasons for uppercasing appear to have stemmed from an
old codesourcery toolchain behaviour, see

http://comments.gmane.org/gmane.comp.compilers.llvm.cvs/87133

This patch makes the object file emitted no longer captialise string
data, it encodes as seen in the assembly source.

Change-Id: Ibe20dd6e60d2773d57ff72a78470839033aa5538
llvm-svn: 222882
llvm/lib/Target/ARM/MCTargetDesc/ARMELFStreamer.cpp
llvm/test/CodeGen/ARM/build-attributes-encoding.s
llvm/test/MC/ARM/directive-arch-iwmmxt.s
llvm/test/MC/ARM/directive-arch-iwmmxt2.s
llvm/test/MC/ARM/directive-cpu.s
llvm/test/MC/ARM/directive-eabi_attribute-overwrite.s
llvm/test/MC/ARM/directive-eabi_attribute.s
llvm/test/tools/llvm-readobj/ARM/attributes.s