llvm-svn: 74019

It is not a good idea to have data member's name match argument's name. In fact, it is a simple receipe to waste an hour or so.

llvm-svn: 74018

llvm-svn: 74017

llvm-svn: 74016

llvm-svn: 74015

 - Lives inside new library lib/MC (LLVMMC.a)

llvm-svn: 74013

Use atomic operations when accessing statistics, and make the lazy initialization of statistics actually threadsafe.

llvm-svn: 74005

llvm-svn: 74004

llvm-svn: 74003

Make timers threadsafe again.  This isn't quite as nice as I'd hoped (it uses locking rather than atomic arithmetic),
but should work on all the platforms we care about.

I might revisit this if a totally awesome way to do it occurs to me.

llvm-svn: 74002

llvm-svn: 74001

Chris recently broke llvmc with his Makefile changes (r75379). That patch made
the global change .o -> .a, which caused built-in llvmc plugins to stop working
since plugin initialization in llvmc is based on static variables not referenced
from the main executable. This patch implements auto-generated forced references
to the plugin libraries.

llvm-svn: 74000

llvm-svn: 73999

llvm-svn: 73998

llvm-svn: 73997

llvm-svn: 73994

we care about are capable of supporting it.

llvm-svn: 73993

llvm-svn: 73991

llvm-svn: 73990

instructions, which implies that there is an explicit memory operand.  There is
(however) no explicit memory operand; although this is a store, the only memory
operand is implicit, indicated by DS:EDI.  This causes the table-generation code
for the disassembler to report an error."

Patch by Sean Callanan!

llvm-svn: 73989

llvm-svn: 73988

llvm-svn: 73987

llvm-svn: 73986

Add IsThumb1Only to most 16-bit thumb instructions since we want to isel 32-bit instructions when they are available.

llvm-svn: 73985

llvm-svn: 73984

different for ppc; add another version of the test.

llvm-svn: 73983

llvm-svn: 73982

generating LLVM IR; it is correct in the code as written
to use 8-byte-aligned operations to copy Key in bar.  Formerly
the gcc inliner was run, now it isn't.  I don't think it's
possible to preserve this as a pure FE test.  Adding -O2 lets
the llvm optimizers get rid of the 8-byte-aligned stores, at least.

llvm-svn: 73981

llvm-svn: 73980

llvm-svn: 73979

llvm-svn: 73978

llvm-svn: 73975

Use 64-bit integer counters for tracking time, rather than doubles.  This will be more atomic op friendly.

llvm-svn: 73974

generated code was apparently doing stores directly into the return value
aggregate; now, it's doing a copy from a compiler-generated static object.
That object is initialized using [4 x i8] which breaks the test.  I believe
this change preserves the original point of the test.

Of course it would be better for the code to do stores directly into the
return aggregate, but that is not what happens at -O0; the llvm optimizers
seem to do that on x86 but not on ppc32, possibly because of the explicit
padding (which is unavoidable).  I think it must have been being done by
a gcc optimizer pass before.

llvm-svn: 73972

the near future.

llvm-svn: 73971

llvm-svn: 73970

cleans up the CMake-based build system a bit. Started by a patch from
Xerxes Rånby.

llvm-svn: 73969

llvm-svn: 73967

llvm-svn: 73966

also some contribution from Jim Grosbach, Bob Wilson, and Evan Cheng.

I've done my best to consolidate the patches with those that were done by
Viktor Kutuzov and Anton Korzh from Access Softek, Inc. Let me know if missed
anything. I've completely reorganized the thumb2 td file, made more extensive
uses of multiclass, etc.

Test cases will be contributed later after I re-organize what's in svn first.

llvm-svn: 73965