This reverts r192454
Apparently FileCheck isn't as smart as I though and does not enforce a
topological order between variable defs+uses.
llvm-svn: 192472
other in memory and the target has paired load and performs post-isel loads
combining.
E.g., this optimization will transform something like this:
a = load i64* addr
b = trunc i64 a to i32
c = lshr i64 a, 32
d = trunc i64 c to i32
into:
b = load i32* addr1
d = load i32* addr2
Where addr1 = addr2 +/- sizeof(i32), if the target supports paired load and
performs post-isel loads combining.
One should overload TargetLowering::hasPairedLoad to provide this information.
The default is false.
<rdar://problem/14477220>
llvm-svn: 192471
There are aliases for --start-group/--end-group options represented
by -( and -) respectively in the command line.
This change adds and improves the test for the alias options to be
tested.
Looks like users use this option widely than explicitly using
--start-group/--end-group.
llvm-svn: 192470
Just pass a Target* into ObjectFileELF::GetImageInfoAddress so that
it can do the extra dereference necessary on MIPS, instead of passing
a flag back to the caller.
Review: http://llvm-reviews.chandlerc.com/D1899
llvm-svn: 192469
Currently data-race-test unittests fail with the following false positive:
WARNING: ThreadSanitizer: data race (pid=20365)
Write of size 8 at 0x7da000008050 by thread T54:
#0 close tsan_interceptors.cc:1483 (racecheck_unittest-linux-amd64-O0+0x0000000eb34a)
#1 NegativeTests_epoll::Worker2() unittest/posix_tests.cc:1148 (racecheck_unittest-linux-amd64-O0+0x0000000cc6b1)
#2 MyThread::ThreadBody(MyThread*) unittest/./thread_wrappers_pthread.h:367 (racecheck_unittest-linux-amd64-O0+0x000000097500)
Previous read of size 8 at 0x7da000008050 by thread T49:
#0 epoll_ctl tsan_interceptors.cc:1646 (racecheck_unittest-linux-amd64-O0+0x0000000e9fee)
#1 NegativeTests_epoll::Worker1() unittest/posix_tests.cc:1140 (racecheck_unittest-linux-amd64-O0+0x0000000cc5b5)
#2 MyThread::ThreadBody(MyThread*) unittest/./thread_wrappers_pthread.h:367 (racecheck_unittest-linux-amd64-O0+0x000000097500)
llvm-svn: 192448
INSERT is the first type of MSA instruction that requires a change to the way MSA registers are parsed.
This happens because MSA registers may be suffixed by an index in the form of an immediate or a
general purpose register. The changes to parseMSARegs reflect that requirement.
llvm-svn: 192447
For NVPTX, this fixes a crash where the emitImplicitDef implementation was expecting physical registers,
while NVPTX uses virtual registers (with a couple of exceptions). Now, the implicit def comment will be
emitted as a true PTX register name. Other targets can use this to customize the output of implicit def
comments.
Fixes PR17519
llvm-svn: 192444
When a ConstantExpr which uses a thread local is part of a PHI node
instruction, the insruction that replaces the ConstantExpr must
be inserted in the predecessor block, in front of the terminator instruction.
If the predecessor block has multiple successors, the edge is first split.
llvm-svn: 192432
allowRemainingUndefines() is already checked in Resolver::resolve(), so we
don't need to check it again after returning from it. It's actually not only
superfluous but buggy because a failure of resolve() does not always mean that
there is a remaining undefines.
llvm-svn: 192423
Quentin Colombet