Testcase for a new instcombine xform, patch contributed by Nick Lewycky!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28449 91177308-0d34-0410-b5e6-96231b3b80d8

Don't make zero-sized static arrays

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28448 91177308-0d34-0410-b5e6-96231b3b80d8

Patches to make the LLVM sources more -pedantic clean.  Patch provided
by Anton Korobeynikov!  This is a step towards closing PR786.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28447 91177308-0d34-0410-b5e6-96231b3b80d8

One of these xforms is only safe with unsafe math enabled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28446 91177308-0d34-0410-b5e6-96231b3b80d8

Fix CodeGen/Generic/vector.ll:test_div with altivec.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28445 91177308-0d34-0410-b5e6-96231b3b80d8

New testcase

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28444 91177308-0d34-0410-b5e6-96231b3b80d8

Handle SETO* like we handle SET*, restoring behavior after Evan's setcc
change.  This fixes PowerPC/fnegsel.ll.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28443 91177308-0d34-0410-b5e6-96231b3b80d8

Print struct return functions and calls as actually returning the hidden
argument struct pointer, enabling ABI compatibility for the CBE with
platforms with strange struct-return ABIs.  This fixes 252.eon and
CoyoteBench/fftbench on Darwin/X86 among other things.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28442 91177308-0d34-0410-b5e6-96231b3b80d8

Fix file header comment

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28441 91177308-0d34-0410-b5e6-96231b3b80d8

Better way to check for vararg.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28440 91177308-0d34-0410-b5e6-96231b3b80d8

Remove PreprocessCCCArguments and PreprocessFastCCArguments now that
FORMAL_ARGUMENTS nodes include a token operand.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28439 91177308-0d34-0410-b5e6-96231b3b80d8

Implement an annoying part of the Darwin/X86 abi: the callee of a struct
return argument pops the hidden struct pointer if present, not the caller.

For example, in this testcase:

struct X { int D, E, F, G; };
struct X bar() {
  struct X a;
  a.D = 0;
  a.E = 1;
  a.F = 2;
  a.G = 3;
  return a;
}
void foo(struct X *P) {
  *P = bar();
}

We used to emit:

_foo:
        subl $28, %esp
        movl 32(%esp), %eax
        movl %eax, (%esp)
        call _bar
        addl $28, %esp
        ret
_bar:
        movl 4(%esp), %eax
        movl $0, (%eax)
        movl $1, 4(%eax)
        movl $2, 8(%eax)
        movl $3, 12(%eax)
        ret

This is correct on Linux/X86 but not Darwin/X86.  With this patch, we now
emit:

_foo:
        subl $28, %esp
        movl 32(%esp), %eax
        movl %eax, (%esp)
        call _bar
***     addl $24, %esp
        ret
_bar:
        movl 4(%esp), %eax
        movl $0, (%eax)
        movl $1, 4(%eax)
        movl $2, 8(%eax)
        movl $3, 12(%eax)
***     ret $4

For the record, GCC emits (which is functionally equivalent to our new code):

_bar:
        movl    4(%esp), %eax
        movl    $3, 12(%eax)
        movl    $2, 8(%eax)
        movl    $1, 4(%eax)
        movl    $0, (%eax)
        ret     $4
_foo:
        pushl   %esi
        subl    $40, %esp
        movl    48(%esp), %esi
        leal    16(%esp), %eax
        movl    %eax, (%esp)
        call    _bar
        subl    $4, %esp
        movl    16(%esp), %eax
        movl    %eax, (%esi)
        movl    20(%esp), %eax
        movl    %eax, 4(%esi)
        movl    24(%esp), %eax
        movl    %eax, 8(%esi)
        movl    28(%esp), %eax
        movl    %eax, 12(%esi)
        addl    $40, %esp
        popl    %esi
        ret

This fixes SingleSource/Benchmarks/CoyoteBench/fftbench with LLC and the
JIT, and fixes the X86-backend portion of PR729.  The CBE still needs to
be updated.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28438 91177308-0d34-0410-b5e6-96231b3b80d8

-enable-unsafe-fp-math implies -enable-finite-only-fp-math

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28437 91177308-0d34-0410-b5e6-96231b3b80d8

Make class comment visible in Doxygen.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28436 91177308-0d34-0410-b5e6-96231b3b80d8

Fix missing include

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28435 91177308-0d34-0410-b5e6-96231b3b80d8

Added a test case for FP equality check.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28434 91177308-0d34-0410-b5e6-96231b3b80d8

Incorrect SETCC CondCode used for FP comparisons.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28433 91177308-0d34-0410-b5e6-96231b3b80d8

Added option -enable-finite-only-fp-math. When on, the codegen can assume that
FP arithmetic arguments and results are never NaNs or +=Infs. This includes
ignoring parity flag (PF) when checking for FP equality.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28432 91177308-0d34-0410-b5e6-96231b3b80d8

implement minimal versions of
ARMAsmPrinter::runOnMachineFunction
LowerFORMAL_ARGUMENTS
ARMInstrInfo::isMoveInstr

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28431 91177308-0d34-0410-b5e6-96231b3b80d8

Describe how to add a custom test.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28430 91177308-0d34-0410-b5e6-96231b3b80d8

Wrap long lines fix typo

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28429 91177308-0d34-0410-b5e6-96231b3b80d8

A isel deficiency.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28427 91177308-0d34-0410-b5e6-96231b3b80d8

For PR784:
Support Win32 platforms for llvm-gcc path. Patch by Anton Korobeynikov

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28426 91177308-0d34-0410-b5e6-96231b3b80d8

Back out indirect branch load folding hack. It broke some tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28425 91177308-0d34-0410-b5e6-96231b3b80d8

Add a note

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28424 91177308-0d34-0410-b5e6-96231b3b80d8

Make TargetData strings less redundant.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28423 91177308-0d34-0410-b5e6-96231b3b80d8

Silence a bogus gcc warning

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28422 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a parsing bug that caused 7 llvm-test regressions on PPC last night.
I'm suprised it didn't cause more!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28421 91177308-0d34-0410-b5e6-96231b3b80d8

- Use of load's chain result should be redirected to load's chain operand.
  If it reads the chain result of the call, then the use, callseq_start,
  and call would form a cycle!
- Don't forget handle node replacement!
- There could also be a TokenFactor between the load and the callseq_start.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28420 91177308-0d34-0410-b5e6-96231b3b80d8

A new entry

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28419 91177308-0d34-0410-b5e6-96231b3b80d8

Missing break statements.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28418 91177308-0d34-0410-b5e6-96231b3b80d8

Remove unused patterns.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28417 91177308-0d34-0410-b5e6-96231b3b80d8

Handle indirect call which folds a load manually. This never matches by
the TableGen generated code since the load's chain result is read by
the callseq_start node.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28416 91177308-0d34-0410-b5e6-96231b3b80d8

Sparc is big-endian.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28415 91177308-0d34-0410-b5e6-96231b3b80d8

Make all of the TargetMachine subclasses use the new string TargetData methods.

This is part of the on-going work on PR 761.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28414 91177308-0d34-0410-b5e6-96231b3b80d8

new testcase for csretcc

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28413 91177308-0d34-0410-b5e6-96231b3b80d8

Print csretcc calls like this:

        call csretcc void %structret( { sbyte }* %P )

instead of this:

        callcsretcc  void %structret( { sbyte }* %P )

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28412 91177308-0d34-0410-b5e6-96231b3b80d8

Fix misencoding of calling conventions

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28411 91177308-0d34-0410-b5e6-96231b3b80d8

pretty print csretcc for calls

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28410 91177308-0d34-0410-b5e6-96231b3b80d8

CSRet allows varargs

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28409 91177308-0d34-0410-b5e6-96231b3b80d8

Asmprint csret nicely

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28408 91177308-0d34-0410-b5e6-96231b3b80d8

Regenerate

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28407 91177308-0d34-0410-b5e6-96231b3b80d8

Add support for parsing csret

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28406 91177308-0d34-0410-b5e6-96231b3b80d8

csret functions can be varargs (as can target cc's).  Verify restrictions on
csret functions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28405 91177308-0d34-0410-b5e6-96231b3b80d8

Add new calling convention, as documented in LangRef.html

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28404 91177308-0d34-0410-b5e6-96231b3b80d8

New calling convention I will be adding shortly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28403 91177308-0d34-0410-b5e6-96231b3b80d8

Add a note

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28402 91177308-0d34-0410-b5e6-96231b3b80d8

Add a note

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28401 91177308-0d34-0410-b5e6-96231b3b80d8

Split the SSE readme items out into their own README.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28400 91177308-0d34-0410-b5e6-96231b3b80d8

Split FP-stack notes out of the main readme.  Next up: splitting out SSE.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28399 91177308-0d34-0410-b5e6-96231b3b80d8

Move a target-independent note out of the X86 readme.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28398 91177308-0d34-0410-b5e6-96231b3b80d8

Particularly ugly code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28397 91177308-0d34-0410-b5e6-96231b3b80d8

new testcase

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28396 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a doxygen problem and break lines at 80 columns

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28395 91177308-0d34-0410-b5e6-96231b3b80d8

Fix some doxygen usage in these headers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28394 91177308-0d34-0410-b5e6-96231b3b80d8

These can be transformed into lea as well. Not that we use this feature
currently...

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28393 91177308-0d34-0410-b5e6-96231b3b80d8

- Use exact-width integer types, e.g. int32_t, to avoid confusion.
- Fix a couple of minor bugs in i16immSExt8 and i16immZExt8.
- Added loadiPTR fragment used for indirect jumps and calls.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28392 91177308-0d34-0410-b5e6-96231b3b80d8

Use class tags instead of struct tags.  The coding standards specify this
for public classes for improved win32 compatibility.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28391 91177308-0d34-0410-b5e6-96231b3b80d8

Explicitly specify MOV32mi can only be used store 32-bit GV, etc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28390 91177308-0d34-0410-b5e6-96231b3b80d8

Now that iPTR is a fully resolved type. We end up losing the type check for
patterns that look like this:

def : Pat<(i32 (X86Wrapper tconstpool  :$dst)), (MOV32ri tconstpool  :$dst)>;

InsertOneTypeCheck should copy the type from the resolved pattern to the
unresolved one as long as there types are different.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28389 91177308-0d34-0410-b5e6-96231b3b80d8

implement movri
add a stub LowerFORMAL_ARGUMENTS

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28388 91177308-0d34-0410-b5e6-96231b3b80d8

Added a Flags field to TargetOperandInfo. Currently the only flag is
M_LOOK_UP_PTR_REG_CLASS which allows the register class of the operand to be
resolved via a callback at runtime.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28387 91177308-0d34-0410-b5e6-96231b3b80d8

lib/Target/Target.td

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28386 91177308-0d34-0410-b5e6-96231b3b80d8

Improve error output. Use the third parameter of the "catch" command to
capture the error output from the exec option. This generally will capture
the stderr messages generated by the tools. This information is then
printed if the test fails. This helps to recognize more quickly what the
error was. Otherwise, this information is lost.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28385 91177308-0d34-0410-b5e6-96231b3b80d8

add a note

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28384 91177308-0d34-0410-b5e6-96231b3b80d8

add a note

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28383 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a bogus gcc warning

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28382 91177308-0d34-0410-b5e6-96231b3b80d8

ImmMask should be 3 for a two-bit field; Compact X86II

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28381 91177308-0d34-0410-b5e6-96231b3b80d8

Fix some think-o's in my last commit. Thanks to Chris for pointing them out.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28380 91177308-0d34-0410-b5e6-96231b3b80d8

Change Module to use TargetData-compatible strings internally.

This is part of the on-going work on PR 761.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28379 91177308-0d34-0410-b5e6-96231b3b80d8

getCalleeSaveRegs and getCalleeSaveRegClasses are no long TableGen'd.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28378 91177308-0d34-0410-b5e6-96231b3b80d8

Remove CalleeSavedRegisters from class Target.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28377 91177308-0d34-0410-b5e6-96231b3b80d8

Don't generate getCalleeSaveReg and getCalleeSaveRegClasses anymore.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28376 91177308-0d34-0410-b5e6-96231b3b80d8

This test is buggy: printf is a varargs function.  This fixes the test with
the PPC JIT

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28375 91177308-0d34-0410-b5e6-96231b3b80d8

Make some changes suggested by Chris:
1. Remove the LLVM_DO_NOT_BUILD feature (not needed any more)
2. Ensure that lib/VMCore gets built first. This needs to be done because
   VMCore now uses tblgen to generate the Intrinsics header which are
   needed in other libraries. In parallel builds, this can cause problems.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28374 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a stupid bug when parsing TargetData strings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28373 91177308-0d34-0410-b5e6-96231b3b80d8

Use DIRS here instead of PARALLEL_DIRS, as VMCore has to be built before the
other dirs (for Intrinsics.gen).  :(

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28372 91177308-0d34-0410-b5e6-96231b3b80d8

Use generic iPTR instead i32 to represent pointer type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28371 91177308-0d34-0410-b5e6-96231b3b80d8

Another entry

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28370 91177308-0d34-0410-b5e6-96231b3b80d8

Remove the llvm-db and bugpoint restrictions from the win32 platform so
they can be compiled with ming32. The use of fork(2) has been removed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28369 91177308-0d34-0410-b5e6-96231b3b80d8

Remove PointerType from class Target

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28368 91177308-0d34-0410-b5e6-96231b3b80d8

Declare that lowerinvoke doesn't interact with other lowering passes.
Patch written by Domagoj Babic!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28367 91177308-0d34-0410-b5e6-96231b3b80d8

Typo

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28366 91177308-0d34-0410-b5e6-96231b3b80d8

Fix the result of the call to use a correct vbitconvert.  There is no need to
use getPackedTypeBreakdown at all here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28365 91177308-0d34-0410-b5e6-96231b3b80d8

Correct a previous patch which broke CodeGen/PowerPC/vec_call.ll

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28364 91177308-0d34-0410-b5e6-96231b3b80d8

Remove PointerType from target definition. Use abstract type MVT::iPTR to
represent pointer type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28363 91177308-0d34-0410-b5e6-96231b3b80d8

PR736 has already been fixed. Remove XFAIL marker.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28362 91177308-0d34-0410-b5e6-96231b3b80d8

New test case for vector type argument pass by value.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28361 91177308-0d34-0410-b5e6-96231b3b80d8

Fix call_adj.ll

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28360 91177308-0d34-0410-b5e6-96231b3b80d8

Added sanity check for obviously bogus immediates

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28359 91177308-0d34-0410-b5e6-96231b3b80d8

Added regression that breaks gcc4 build

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28358 91177308-0d34-0410-b5e6-96231b3b80d8

Should pass by reference.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28357 91177308-0d34-0410-b5e6-96231b3b80d8

Another entry

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28356 91177308-0d34-0410-b5e6-96231b3b80d8

Add a note about a note

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28355 91177308-0d34-0410-b5e6-96231b3b80d8

Make PPC call lowering more aggressive, making the isel matching code simple
enough to be autogenerated.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28354 91177308-0d34-0410-b5e6-96231b3b80d8

Another typo. Pointed out by Nate Begeman.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28353 91177308-0d34-0410-b5e6-96231b3b80d8

Fixed a LowerCallTo and LowerArguments bug. They were introducing illegal
VBIT_VECTOR nodes. There were some confusion about the semantics of
getPackedTypeBreakdown(). e.g. for <4 x f32> it returns 1 and v4f32, not 4,
and f32.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28352 91177308-0d34-0410-b5e6-96231b3b80d8

Fix an obvious bug in getPackedTypeBreakdown. Return 1 if type is legal.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28351 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a mis-leading comment.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28350 91177308-0d34-0410-b5e6-96231b3b80d8

Add a CloneModule call that exposes the mapping of values from the old module
to the new module.  Patch provided by Nick Lewycky!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@28349 91177308-0d34-0410-b5e6-96231b3b80d8