llvm-svn: 137485

integer register to a floating point register. It is not valid to interpret
the value of a floating pointer register as part of a double precision
floating point value after a single precision floating point computational
or move instruction stores its result to the register.

- In the test case, the following code is generated before this patch is
  applied:
mtc1  $zero, $f2    ; unformatted copy to $f2
mov.s $f0, $f2      ; $f0 is in single format
sdc1  $f12, 0($sp)
mov.s $f1, $f2      ; $f1 is in single format
c.eq.d  $f12, $f0   ; $f0 cannot be interpreted as double

- The following code is generated after this patch is applied:
mtc1  $zero, $f0    ; unformatted copy to $f0
mtc1  $zero, $f1    ; unformatted copy to $f1
c.eq.d  $f12, $f0   ; $f0 can be interpreted as double

Bhanu Chetlapalli and Chris Dearman at MIPS technologies reported this bug and
provided the test case.

llvm-svn: 137484

llvm-svn: 137483

llvm-svn: 137482

llvm-svn: 137481

llvm-svn: 137480

llvm-svn: 137479

llvm-svn: 137478

llvm-svn: 137476

llvm-svn: 137471

llvm-svn: 137470

llvm-svn: 137468

llvm-svn: 137467

llvm-svn: 137466

llvm-svn: 137464

llvm-svn: 137463

when building with assertions disabled.

llvm-svn: 137460

llvm-svn: 137444

Thanks to Jeff Yasskin to reword.

llvm-svn: 137443

llvm-svn: 137442

.gitignore: Ignore /autom4te.cache. We can execute "PATH=/path/to/autotools/bin autoconf/AutoRegen.sh".

llvm-svn: 137441

llvm-svn: 137433

llvm-svn: 137432

when checking isNull(), we'd pick off the sentinel bit for the outer 
PointerUnion, but would not recursively convert the inner pointerunion to bool,
so if *its* sentinel bit is set, isNull() would incorrectly return false.

No testcase, because someone hit this when they were trying to refactor code
to use PointerUnion3, but they since found a better solution.

llvm-svn: 137428

llvm-svn: 137425

include/llvm/Config/config.h.cmake: Reorder to follow config.h.in, and import stuff from llvm-config.h.cmake and clang/config.h.

llvm-svn: 137424

STACK_DIRECTION
YYTEXT_POINTER
HAVE_NAMESPACES
HAVE_STD_ITERATOR
HAVE_FWD_ITERATOR
HAVE_BI_ITERATOR
HAVE_GLOBAL_HASH_MAP
HAVE_GLOBAL_HASH_SET
HAVE_GNU_EXT_HASH_MAP
HAVE_GNU_EXT_HASH_SET
HAVE_STD_EXT_HASH_MAP
HAVE_STD_EXT_HASH_SET

llvm-svn: 137423

MachOWriter: Don't crash on fixups with arithmetic, emit a relocation instead. This matches what as does.

llvm-svn: 137414

llvm-svn: 137413

llvm-svn: 137412

llvm-svn: 137408

Fix by Ivan Baev. Sorry I don't have a unit test, but the fix is obvious so I don't want to delay it.

llvm-svn: 137404

llvm-svn: 137403

is returned through a bitcast.

llvm-svn: 137402

llvm-svn: 137400

the retains and releases all use the same SSA pointer value.

Also, don't let CFG hazards disrupt nested retain+release pair
optimizations.

llvm-svn: 137399

rather than plain postorder, so that CFG constructs like single-exit loops
are reliably visited in a sensible order.

llvm-svn: 137398

This is possible now that we now longer provide an interface to iterate
the interference overlaps.

llvm-svn: 137397

llvm-svn: 137393

llvm-svn: 137389