Add test for rdar://6394879.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60079 91177308-0d34-0410-b5e6-96231b3b80d8

add a comment, make save/restore logic more obvious.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60076 91177308-0d34-0410-b5e6-96231b3b80d8

This adds in some code (currently disabled unless you pass
-enable-smarter-addr-folding to llc) that gives CGP a better
cost model for when to sink computations into addressing modes.
The basic observation is that sinking increases register
pressure when part of the addr computation has to be available
for other reasons, such as having a use that is a non-memory
operation.  In cases where it works, it can substantially reduce
register pressure.

This code is currently an overall win on 403.gcc and 255.vortex
(the two things I've been looking at), but there are several
things I want to do before enabling it by default:

1. This isn't doing any caching of results, so it is much slower
   than it could be.  It currently slows down release-asserts llc
   by 1.7% on 176.gcc: 27.12s -> 27.60s.
2. This doesn't think about inline asm memory operands yet.
3. The cost model botches the case when the needed value is live
   across the computation for other reasons.

I'll continue poking at this, and eventually turn it on as llcbeta.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60074 91177308-0d34-0410-b5e6-96231b3b80d8

Revert r60042. IndVarSimplify should check if APFloat is PPCDoubleDouble first before trying to convert it to an integer.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60072 91177308-0d34-0410-b5e6-96231b3b80d8

Teach CodeGenPrepare to look through Bitcast instructions when attempting to
optimize addressing modes.  This allows us to optimize things like isel-sink2.ll
into:

movl 4(%esp), %eax
cmpb $0, 4(%eax)
jne LBB1_2 ## F
LBB1_1: ## TB
movl $4, %eax
ret
LBB1_2: ## F
movzbl 7(%eax), %eax
ret

instead of:

_test:
movl 4(%esp), %eax
cmpb $0, 4(%eax)
leal 4(%eax), %eax
jne LBB1_2 ## F
LBB1_1: ## TB
movl $4, %eax
ret
LBB1_2: ## F
movzbl 3(%eax), %eax
ret

This shrinks (e.g.) 403.gcc from 1133510 to 1128345 lines of .s.

Note that the 2008-10-16-SpillerBug.ll testcase is dubious at best, I doubt
it is really testing what it thinks it is.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60068 91177308-0d34-0410-b5e6-96231b3b80d8

fix an over-reduced test.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60067 91177308-0d34-0410-b5e6-96231b3b80d8

this doesn't need EH

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60066 91177308-0d34-0410-b5e6-96231b3b80d8

ignore tests output

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60065 91177308-0d34-0410-b5e6-96231b3b80d8

change AnnotationManager to use 'const char*' instead of std::string. this fixes the leakage of those strings and avoids the creation of such strings in static cosntructors (should result in a little improvement of startup time)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60064 91177308-0d34-0410-b5e6-96231b3b80d8

CMake: Removed tools/llvmc2 directory hierarchy.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60054 91177308-0d34-0410-b5e6-96231b3b80d8

CMake: llvmc2 is now known as llvmc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60052 91177308-0d34-0410-b5e6-96231b3b80d8

Add a man page for llvmc. Really basic for now, will be updated later.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60049 91177308-0d34-0410-b5e6-96231b3b80d8

Since the old llvmc was removed, rename llvmc2 to llvmc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60048 91177308-0d34-0410-b5e6-96231b3b80d8

Make -fsyntax-only, -include and -emit-llvm work for C++ and Objective-C/C++.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60047 91177308-0d34-0410-b5e6-96231b3b80d8

docs: Add author info + fix incorrect code example.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60046 91177308-0d34-0410-b5e6-96231b3b80d8

Small documentation update.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60045 91177308-0d34-0410-b5e6-96231b3b80d8

Document the plugin priority feature.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60044 91177308-0d34-0410-b5e6-96231b3b80d8

A simplification for checking whether the signs of the operands and sum differ. Thanks, Duncan.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60043 91177308-0d34-0410-b5e6-96231b3b80d8

convertToSignExtendedInteger should return opInvalidOp instead of asserting if sematics of float does not allow arithmetics.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60042 91177308-0d34-0410-b5e6-96231b3b80d8

Suppress warnings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60041 91177308-0d34-0410-b5e6-96231b3b80d8

This method got renamed, thanks to Mattias Holm for pointing this out.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60039 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU:
(a) Remove conditionally removed code in SelectXAddr. Basically, hope for the
    best that the A-form and D-form address predicates catch everything before
    the code decides to emit a X-form address.
(b) Expand vector store test cases to include the usual suspects.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60034 91177308-0d34-0410-b5e6-96231b3b80d8

add info about how to run the tests with valgrind

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60030 91177308-0d34-0410-b5e6-96231b3b80d8

Now with the correct type for the 0.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60016 91177308-0d34-0410-b5e6-96231b3b80d8

Get rid of unused variable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60015 91177308-0d34-0410-b5e6-96231b3b80d8

Hacker's Delight says, "Signed integer overflow of addition occurs if and only
if the operands have the same sign and the sum has sign opposite to that of the
operands."

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60014 91177308-0d34-0410-b5e6-96231b3b80d8

Teach MatchScaledValue to handle Scales by 1 with MatchAddr (which
can recursively match things) and scales by 0 by ignoring them.
This triggers once in 403.gcc, saving 1 (!!!!) instruction in the
whole huge app.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60013 91177308-0d34-0410-b5e6-96231b3b80d8

significantly refactor all the addressing mode matching logic
into a new AddressingModeMatcher class.  This makes it easier
to reason about and reduces passing around of stuff, but has
no functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60012 91177308-0d34-0410-b5e6-96231b3b80d8

refactor all the constantexpr/instruction handling code out into a
new FindMaximalLegalAddressingModeForOperation helper method.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60011 91177308-0d34-0410-b5e6-96231b3b80d8

another minor tweak

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60010 91177308-0d34-0410-b5e6-96231b3b80d8

minor cleanups no functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60009 91177308-0d34-0410-b5e6-96231b3b80d8

CMake: Documented two cmake variables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60008 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU: Relax constraints on when to generate a X-form address, evidently
they were too tight according to bug 3126.

Fix bug 3126.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60006 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU: test should use shlqby, not shlqbyi

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60001 91177308-0d34-0410-b5e6-96231b3b80d8

XFAIL this test. A recent CellSPU check-in broke it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@60000 91177308-0d34-0410-b5e6-96231b3b80d8

Initial support for anti-dependence breaking. Currently this code does not
introduce any new spilling; it just uses unused registers.

Refactor the SUnit topological sort code out of the RRList scheduler and
make use of it to help with the post-pass scheduler.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59999 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU: Fix mnemonic typo in pattern; "shlqbyi" -> "shlqby".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59998 91177308-0d34-0410-b5e6-96231b3b80d8

Testcase for constant CFStrings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59992 91177308-0d34-0410-b5e6-96231b3b80d8

rearrange and tidy some code, no functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59990 91177308-0d34-0410-b5e6-96231b3b80d8

minor cleanups to debug code, no functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59989 91177308-0d34-0410-b5e6-96231b3b80d8

reenable test

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59986 91177308-0d34-0410-b5e6-96231b3b80d8

reenable the right part of the code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59985 91177308-0d34-0410-b5e6-96231b3b80d8

Temporarily XFAIL this test. r59976 and r59972 broke it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59981 91177308-0d34-0410-b5e6-96231b3b80d8

Move the scheduler constructor functions to SchedulerRegistry.h, to
simplify header dependencies for front-ends that just want to choose
a scheduler and don't need all the scheduling machinery declarations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59978 91177308-0d34-0410-b5e6-96231b3b80d8

Fix an unused-parameter warning.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59977 91177308-0d34-0410-b5e6-96231b3b80d8

revert an accidental commit, this fixes the regression on test/CodeGen/X86/isel-sink.ll

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59976 91177308-0d34-0410-b5e6-96231b3b80d8

Fix 3113: If we have a dead cyclic PHI, replace the whole thing
with an undef.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59972 91177308-0d34-0410-b5e6-96231b3b80d8

- Make lowering of "add with overflow" customizable by back-ends.
- Mark "add with overflow" as having a custom lowering for X86. Give it a null
  lowering representation for now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59971 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU:
(a) Slight rethink on i64 zero/sign/any extend code - use a shuffle to
    directly zero-extend i32 to i64, but use rotates and shifts for
    sign extension. Also ensure unified register consistency.
(b) Add new test harness for i64 operations: i64ops.ll

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59970 91177308-0d34-0410-b5e6-96231b3b80d8

Check in the rest of this change. The isAntiDep flag needs to be passed
to removePred because an SUnit can both data-depend and anti-depend
on the same SUnit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59969 91177308-0d34-0410-b5e6-96231b3b80d8

Pass the isAntiDep argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59968 91177308-0d34-0410-b5e6-96231b3b80d8

Run post-RA scheduling after branch folding, as it tends to
obscure tail-merging opportunities.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59967 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a few HTML tidiness issues.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59966 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU:
(a) Improve the extract element code: there's no need to do gymnastics with
    rotates into the preferred slot if a shuffle will do the same thing.
(b) Rename a couple of SPUISD pseudo-instructions for readability and better
    semantic correspondence.
(c) Fix i64 sign/any/zero extension lowering.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59965 91177308-0d34-0410-b5e6-96231b3b80d8

Delete some spurious characters.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59964 91177308-0d34-0410-b5e6-96231b3b80d8

Fix some "</p>"s.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59963 91177308-0d34-0410-b5e6-96231b3b80d8

Some documentation for LegalizeTypes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59962 91177308-0d34-0410-b5e6-96231b3b80d8

Minor fix debug for register allocation debug output.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59961 91177308-0d34-0410-b5e6-96231b3b80d8

If the type legalizer actually legalized anything
(this doesn't happen that often, since most code
does not use illegal types) then follow it by a
DAG combiner run that is allowed to generate
illegal operations but not illegal types.  I didn't
modify the target combiner code to distinguish like
this between illegal operations and illegal types,
so it will not produce illegal operations as well
as not producing illegal types.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59960 91177308-0d34-0410-b5e6-96231b3b80d8

Fix comments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59958 91177308-0d34-0410-b5e6-96231b3b80d8

NULL, unique pointers from malloc(0), and freed pointers are legal values for
noalias attribute parameters/return values.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59955 91177308-0d34-0410-b5e6-96231b3b80d8

Move target independent td files from lib/Target/ to include/llvm/Target so they can be distributed along with the header files.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59953 91177308-0d34-0410-b5e6-96231b3b80d8

Eliminate some unused variable compile time warnings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59952 91177308-0d34-0410-b5e6-96231b3b80d8

Test add-with-overflow with fast ISel.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59945 91177308-0d34-0410-b5e6-96231b3b80d8

Seriously strengthen the guarantee offered by noalias on a function's return
value. It must now be as if the pointer were allocated and has not escaped to
the caller. Thanks to Dan Gohman for pointing out the error in the original
and helping devise this definition.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59940 91177308-0d34-0410-b5e6-96231b3b80d8

Extend the 'noalias' attribute to function return values. This is intended to
indicate functions that allocate, such as operator new, or list::insert. The
actual definition is slightly less strict (for now).

No changes to the bitcode reader/writer, asm printer or verifier were needed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59934 91177308-0d34-0410-b5e6-96231b3b80d8

Added missing description for -disable-mmx option.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59929 91177308-0d34-0410-b5e6-96231b3b80d8

Add support for llvm.uadd.with.overflow.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59926 91177308-0d34-0410-b5e6-96231b3b80d8

Roll back rev 59890, since Chris says this can never happen.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59914 91177308-0d34-0410-b5e6-96231b3b80d8

Rename SetCCResultContents to BooleanContents.  In
practice these booleans are mostly produced by SetCC,
however the concept is more general.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59911 91177308-0d34-0410-b5e6-96231b3b80d8

Make JIT::runFunction handle functions with non-C calling conventions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59904 91177308-0d34-0410-b5e6-96231b3b80d8

Added -disable-mmx using a patch from Preston Gurd.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59901 91177308-0d34-0410-b5e6-96231b3b80d8

Added check to avoid generating extract subvector beyond the end of the vector when normalizing vector shuffles.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59900 91177308-0d34-0410-b5e6-96231b3b80d8

Fixed named anchor for llvm.stackprotector intrinsic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59893 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU: Fix bug 3056. Varadic extract_element was not implemented (nor was it
ever conceived to occur).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59891 91177308-0d34-0410-b5e6-96231b3b80d8

Check for NULL before traversing the isa<> type hierarchy checking with a
NULL-based reference.

Note: Encountered this a few times on Tiger + gcc 4.0.1. Might just be a
platform-specific compiler issue, but it's good defensive programming in any
case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59890 91177308-0d34-0410-b5e6-96231b3b80d8

Describe how the JIT maps fields to MachineOperands, patch by
JP Bonn!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59876 91177308-0d34-0410-b5e6-96231b3b80d8

Make a convenient helper for printing offsets.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59872 91177308-0d34-0410-b5e6-96231b3b80d8

Cleanup of the [SU]ADDO type legalization code. Patch by Duncan!

"It simplifies the type legalization part a bit, and produces better code by
teaching SelectionDAG about the extra bits in an i8 SADDO/UADDO node.  In
essence, I spontaneously decided that on x86 this i8 boolean result would be
either 0 or 1, and on other platforms 0/1 or 0/-1, depending on whether the
platform likes it's boolean zero extended or sign extended."

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59864 91177308-0d34-0410-b5e6-96231b3b80d8

- Move conversion of [SU]ADDO from DAG combiner into legalizer.
- Add "promote integer type" stuff to the legalizer for these nodes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59847 91177308-0d34-0410-b5e6-96231b3b80d8

Fix build failure.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59844 91177308-0d34-0410-b5e6-96231b3b80d8

Silence unused variable warnings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59841 91177308-0d34-0410-b5e6-96231b3b80d8

Document TEST=dbgopt

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59839 91177308-0d34-0410-b5e6-96231b3b80d8

Correctly set the isCtrl flag for chain dependencies.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59837 91177308-0d34-0410-b5e6-96231b3b80d8

Update comments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59836 91177308-0d34-0410-b5e6-96231b3b80d8

Update comments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59834 91177308-0d34-0410-b5e6-96231b3b80d8

reapply Sanjiv's patch to genericize memcpy/memset/memmove to take an
arbitrary integer width for the count.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59823 91177308-0d34-0410-b5e6-96231b3b80d8

Revert r59802. It was breaking the build of llvm-gcc:

g++ -m32 -c -g -DIN_GCC -W -Wall -Wwrite-strings -Wmissing-format-attribute -fno-common -mdynamic-no-pic -DHAVE_CONFIG_H -Wno-unused -DTARGET_NAME=\"i386-apple-darwin9.5.0\" -I. -I. -I../../llvm-gcc.src/gcc -I../../llvm-gcc.src/gcc/. -I../../llvm-gcc.src/gcc/../include -I./../intl -I../../llvm-gcc.src/gcc/../libcpp/include  -I../../llvm-gcc.src/gcc/../libdecnumber -I../libdecnumber -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.obj/include -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.src/include -DENABLE_LLVM -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.obj/../llvm.src/include  -D_DEBUG  -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS   -I. -I. -I../../llvm-gcc.src/gcc -I../../llvm-gcc.src/gcc/. -I../../llvm-gcc.src/gcc/../include -I./../intl -I../../llvm-gcc.src/gcc/../libcpp/include  -I../../llvm-gcc.src/gcc/../libdecnumber -I../libdecnumber -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.obj/include -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.src/include ../../llvm-gcc.src/gcc/llvm-types.cpp -o llvm-types.o
../../llvm-gcc.src/gcc/llvm-convert.cpp: In member function 'void TreeToLLVM::EmitMemCpy(llvm::Value*, llvm::Value*, llvm::Value*, unsigned int)':
../../llvm-gcc.src/gcc/llvm-convert.cpp:1496: error: 'memcpy_i32' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp:1496: error: 'memcpy_i64' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp: In member function 'void TreeToLLVM::EmitMemMove(llvm::Value*, llvm::Value*, llvm::Value*, unsigned int)':
../../llvm-gcc.src/gcc/llvm-convert.cpp:1512: error: 'memmove_i32' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp:1512: error: 'memmove_i64' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp: In member function 'void TreeToLLVM::EmitMemSet(llvm::Value*, llvm::Value*, llvm::Value*, unsigned int)':
../../llvm-gcc.src/gcc/llvm-convert.cpp:1528: error: 'memset_i32' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp:1528: error: 'memset_i64' is not a member of 'llvm::Intrinsic'
make[3]: *** [llvm-convert.o] Error 1
make[3]: *** Waiting for unfinished jobs....
rm fsf-funding.pod gcov.pod gfdl.pod cpp.pod gpl.pod gcc.pod
make[2]: *** [all-stage1-gcc] Error 2
make[1]: *** [stage1-bubble] Error 2
make: *** [all] Error 2

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59809 91177308-0d34-0410-b5e6-96231b3b80d8

Make mem[cpy,move,set] intrinsics overloaded.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59802 91177308-0d34-0410-b5e6-96231b3b80d8

Default to converting UADDO to the generic form that SADDO is converted to.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59801 91177308-0d34-0410-b5e6-96231b3b80d8

Optimize (x/y)*y into x-(x%y) in general. Div and rem are about the same, and
a subtract is cheaper than a multiply. This generalizes an existing transform.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59800 91177308-0d34-0410-b5e6-96231b3b80d8

LLVM does have a fortran FE!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59795 91177308-0d34-0410-b5e6-96231b3b80d8

Clean up normalization of shuffles

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59792 91177308-0d34-0410-b5e6-96231b3b80d8

CellSPU:
(a) Fix bgs 3052, 3057
(b) Incorporate Duncan's suggestions re: i1 promotion
(c) Indentation updates.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59790 91177308-0d34-0410-b5e6-96231b3b80d8

Combine the two add with overflow intrinsics lowerings. They differ only in DAG node type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59788 91177308-0d34-0410-b5e6-96231b3b80d8

Set the isAntiDep flag in the MachineInstr scheduler.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59787 91177308-0d34-0410-b5e6-96231b3b80d8

Generate code for llvm.uadd.with.overflow intrinsic. No conversion support yet.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59786 91177308-0d34-0410-b5e6-96231b3b80d8

Add a flag to SDep for tracking which edges are anti-dependence edges.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59785 91177308-0d34-0410-b5e6-96231b3b80d8

Update comment to reflect a semblance of reality.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59784 91177308-0d34-0410-b5e6-96231b3b80d8

Remove chains. Unnecessary.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@59783 91177308-0d34-0410-b5e6-96231b3b80d8