process memory operands with a parenthesized expression for a displacement,
like "(4+5)(%eax)".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73878 91177308-0d34-0410-b5e6-96231b3b80d8

Implement full support for parsing primary expressions.  We can now parse
all of health and voronoi (ignoring directives).  We only get 409 lines into
176.gcc though because we don't have binary operators yet:

Parsing 176.gcc.llc.s:409: unexpected token in operand list
movsbl _arityvec+1(,%edi,8), %eax
                ^

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73877 91177308-0d34-0410-b5e6-96231b3b80d8

implement parser support for '*' operands, as in "call *%eax".

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73876 91177308-0d34-0410-b5e6-96231b3b80d8

implement memory operand parsing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73875 91177308-0d34-0410-b5e6-96231b3b80d8

Testcase for (llvm-gcc) 73873.  Usually
std::pair<double, float*>
is 16 bytes on darwin-powerpc, but not always.
See testcase for full weirdness.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73874 91177308-0d34-0410-b5e6-96231b3b80d8

Factor out code for computing umin and smin for SCEV expressions into
helper functions. Based on a patch by Nick Lewycky.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73869 91177308-0d34-0410-b5e6-96231b3b80d8

start implementing some simple operand parsing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73867 91177308-0d34-0410-b5e6-96231b3b80d8

Teach ScalarEvolution how to analyze loops with multiple exit
blocks, and also exit blocks with multiple conditions (combined
with (bitwise) ands and ors). It's often infeasible to compute an
exact trip count in such cases, but a useful upper bound can often
be found.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73866 91177308-0d34-0410-b5e6-96231b3b80d8

Delete an unused variable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73865 91177308-0d34-0410-b5e6-96231b3b80d8

Fix this code to correctly handle loops with multiple exits. Until
now, this hasn't mattered, because ScalarEvolution hasn't been able
to compute trip counts for loops with multiple exits. But it will
soon.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73864 91177308-0d34-0410-b5e6-96231b3b80d8

Rename a variable for consistency with the ExitBlock vs ExitingBlock
terminology that LoopInfo uses.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73863 91177308-0d34-0410-b5e6-96231b3b80d8

Fix ScalarEvolution's backedge-taken count computations to check for
overflow when computing a integer division to round up.

Thanks to Nick Lewycky for noticing this!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73862 91177308-0d34-0410-b5e6-96231b3b80d8

rename SourceMgr::PrintError to PrintMessage.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73861 91177308-0d34-0410-b5e6-96231b3b80d8

set up the top-level parsing loop.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73860 91177308-0d34-0410-b5e6-96231b3b80d8

stub out parser for asm files.  Change invariant on lexer to always
print its error message when it returns an asmtok::Error token.
Compute a proper error code for llvm-mc in 'lex' mode.  Add new
-as-lex option to enable lexing mode (vs parsing mode).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73859 91177308-0d34-0410-b5e6-96231b3b80d8

add string literals.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73858 91177308-0d34-0410-b5e6-96231b3b80d8

hopefully fix the build on linux.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73857 91177308-0d34-0410-b5e6-96231b3b80d8

simplify some error recovery stuff.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73856 91177308-0d34-0410-b5e6-96231b3b80d8

implement enough of a lexer to get through Olden/health/Output/health.llc.s
without errors.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73855 91177308-0d34-0410-b5e6-96231b3b80d8

Fix PR4185.
Handle FpSET_ST0_80 being used when ST0 is still alive.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73850 91177308-0d34-0410-b5e6-96231b3b80d8

oh yeah, cmake needs to be told explicitly about new files :)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73849 91177308-0d34-0410-b5e6-96231b3b80d8

some baby steps.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73848 91177308-0d34-0410-b5e6-96231b3b80d8

prune #include, fix warning.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73847 91177308-0d34-0410-b5e6-96231b3b80d8

start wiring up support for asm parsing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73846 91177308-0d34-0410-b5e6-96231b3b80d8

move include searching logic from TGLexer to SourceMgr.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73845 91177308-0d34-0410-b5e6-96231b3b80d8

Rename TGSourceMgr -> SourceMgr.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73844 91177308-0d34-0410-b5e6-96231b3b80d8

rename TGLoc -> SMLoc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73843 91177308-0d34-0410-b5e6-96231b3b80d8

move TGSourceMgr class out of TableGen into libsupport.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73842 91177308-0d34-0410-b5e6-96231b3b80d8

simplify closing paren and (rip) printing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73841 91177308-0d34-0410-b5e6-96231b3b80d8

indentation fix

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73840 91177308-0d34-0410-b5e6-96231b3b80d8

Expand this test to handle more cases (remainder and shifts) of zero.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73839 91177308-0d34-0410-b5e6-96231b3b80d8

fit in 80 cols

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73838 91177308-0d34-0410-b5e6-96231b3b80d8

remove #if 0 code

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73837 91177308-0d34-0410-b5e6-96231b3b80d8

cosmetic changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73836 91177308-0d34-0410-b5e6-96231b3b80d8

implement PR4424: 0/x is always 0 for integer division.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73835 91177308-0d34-0410-b5e6-96231b3b80d8

change TLS_ADDR lowering to lower to a real mem operand, instead of matching as
a global with that gets printed with the :mem modifier.  All operands to lea's
should be handled with the lea32mem operand kind, and this allows the TLS stuff
to do this.  There are several better ways to do this, but I went for the minimal
change since I can't really test this (beyond make check).

This also makes the use of EBX explicit in the operand list in the 32-bit,
instead of implicit in the instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73834 91177308-0d34-0410-b5e6-96231b3b80d8

no need for unwind info

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73832 91177308-0d34-0410-b5e6-96231b3b80d8

no need for unwind info here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73831 91177308-0d34-0410-b5e6-96231b3b80d8

remove dead code, the main codepath never has to deal
with MBB's.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73830 91177308-0d34-0410-b5e6-96231b3b80d8

eliminate the "call" operand modifier from the asm descriptions, modeling
it as a pcrel immediate instead.  This gets pc-rel weirdness out of the
main printoperand codepath.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73829 91177308-0d34-0410-b5e6-96231b3b80d8

Clarify that if the alignment is zero then an alignment
compatible with the type will be used (at least the ABI
alignment).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73827 91177308-0d34-0410-b5e6-96231b3b80d8

fix edit-o

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73824 91177308-0d34-0410-b5e6-96231b3b80d8

some comments and cleanup

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73818 91177308-0d34-0410-b5e6-96231b3b80d8

implement support for lowering subregs when preparing to print
LEA64_32r, eliminating a bunch of modifier logic stuff on addr modes.

Implement support for printing mbb labels as operands.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73817 91177308-0d34-0410-b5e6-96231b3b80d8

Fix PR4419: handle defs of partial uses.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73816 91177308-0d34-0410-b5e6-96231b3b80d8

Remove unused field.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73815 91177308-0d34-0410-b5e6-96231b3b80d8

Initialize MMI

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73813 91177308-0d34-0410-b5e6-96231b3b80d8

hook up printMemReference.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73811 91177308-0d34-0410-b5e6-96231b3b80d8

Start implementing translation of MachineInstr to MCInst.  Next
step is to make tblgen generate something more appropriate for MCInst,
and generate calls to operand translation routines where needed.
This includes a bunch of #if 0 code which will slowly be refactored into
something sensible.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73810 91177308-0d34-0410-b5e6-96231b3b80d8

make immediates be int64_t like machineoperand.  Add some apis

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73809 91177308-0d34-0410-b5e6-96231b3b80d8

Tweak this test to be a little less unusual.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73808 91177308-0d34-0410-b5e6-96231b3b80d8

Generalize isLoopGuardedByCond's checking to consider two
SCEVUnknowns with identical Instructions to be equal. This allows
it to analze cases such as the attached testcase, where the front-end
has cloned the loop controlling expression. Along with r73805, this
lets IndVarSimplify eliminate all the sign-extend casts in the
loop in the attached testcase.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73807 91177308-0d34-0410-b5e6-96231b3b80d8

Workaround for an... interesting bug in Darwin's pthread_rwlock_init.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73806 91177308-0d34-0410-b5e6-96231b3b80d8

Use ScalarEvolution's new GetMinSignBits and GetMinLeadingZeros
in the loop backedge-taken count computation of the maximum
possible trip count.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73805 91177308-0d34-0410-b5e6-96231b3b80d8

Add debugging code to test for various locking faux-pas's, when running in single threaded mode.  This should help improve testing coverage for
threading support, without having extensive actually concurrent clients yet.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73803 91177308-0d34-0410-b5e6-96231b3b80d8

Forgot this file.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73802 91177308-0d34-0410-b5e6-96231b3b80d8

Revert r73790, and replace it with a significantly less ugly solution.  Rather than trying to make the global reader-writer lock work,
create separate recursive mutexes for each value map.  The recursive-ness fixes the double-acquiring issue, which having one per ValueMap
lets us continue to maintain some concurrency.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73801 91177308-0d34-0410-b5e6-96231b3b80d8

stub out some hacky code for wiring up the new asmprinter interfaces
on X86.  Not useful yet.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73799 91177308-0d34-0410-b5e6-96231b3b80d8

allow clients of the asmprinter to opt-out of the boilerplate with a #define.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73798 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a typo in a comment that Frits von Bommel noticed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73796 91177308-0d34-0410-b5e6-96231b3b80d8

Make GetMinTrailingZeros a member function of ScalarEvolution,
so that it can access the TargetData member (when available) and
use ValueTracking.h information to compute information for
SCEVUnknown Values.

Also add GetMinLeadingZeros and GetMinSignBits functions,
with minimal implementations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73794 91177308-0d34-0410-b5e6-96231b3b80d8

Re-apply r73718, now that the fix in r73787 is in, and add a
hand-crafted testcase which demonstrates the bug that was exposed
in 254.gap.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73793 91177308-0d34-0410-b5e6-96231b3b80d8

DwarfWriter is used to emit EH info also.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73792 91177308-0d34-0410-b5e6-96231b3b80d8

Enable arm pre-allocation load / store multiple optimization pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73791 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a serious bug that would cause deadlock during abstract type refinement.  The constant creation
gets involved, and we end up trying to recursively acquire a writer lock.  The fix for this is slightly horrible,
and involves passing a boolean "locked" parameter around in Constants.cpp, but it's better than having locked and
unlocked versions of most of the code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73790 91177308-0d34-0410-b5e6-96231b3b80d8

Fix LSR's OptimizeSMax to ignore max operators with more than 2 operands,
which it isn't prepared to handle.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73787 91177308-0d34-0410-b5e6-96231b3b80d8

mv CodeGen/DebugLoc.h Support/DebugLoc.h

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73786 91177308-0d34-0410-b5e6-96231b3b80d8

Move up dwarf writer initialization in common AsmPrinter class.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73784 91177308-0d34-0410-b5e6-96231b3b80d8

Revert 73718. It's breaking 254.gap.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73783 91177308-0d34-0410-b5e6-96231b3b80d8

Unbreak cyclic deps

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73781 91177308-0d34-0410-b5e6-96231b3b80d8

Fix bad paste-o in the MSVC atomics.

Patch by Ð\98ван Ð¡Ð¾Ñ\80окин.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73779 91177308-0d34-0410-b5e6-96231b3b80d8

Forgot to remove some explicit locking when it became implicit in the ValueMap.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73778 91177308-0d34-0410-b5e6-96231b3b80d8

Fences are expensive.  Avoid doing them when we're not in multithreaded mode.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73777 91177308-0d34-0410-b5e6-96231b3b80d8

Make ScalarEvolution::print print getSCEVAtScope values for
expressions, when interesting.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73776 91177308-0d34-0410-b5e6-96231b3b80d8

Move the memory fences out of the path for single-threaded mode.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73775 91177308-0d34-0410-b5e6-96231b3b80d8

Don't (unconditionally) use getSCEVAtScope to simplify the step
expression in IVUsers, because in the case of a use of a non-linear
addrec outside of a loop, this causes the addrec to be evaluated as
a linear addrec.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73774 91177308-0d34-0410-b5e6-96231b3b80d8

Fix incorrect comment pointed out by Duncan.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73773 91177308-0d34-0410-b5e6-96231b3b80d8

Clean up nightly tester to use Perl's File::Find.
Patch by Benjamin Kramer.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73772 91177308-0d34-0410-b5e6-96231b3b80d8

Fix a major typo.

Nicolas, this is probably the cause of the failures you were seeing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73770 91177308-0d34-0410-b5e6-96231b3b80d8

make jump threading handle lexically identical compare instructions
as if they were multiple uses of the same instruction.  This interacts
well with the existing loadpre that j-t does to open up many new jump
threads earlier.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73768 91177308-0d34-0410-b5e6-96231b3b80d8

fix the sparc codegen to not depend on the sparc asmprinter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73767 91177308-0d34-0410-b5e6-96231b3b80d8

Add support for AuroraUX.  Patch by evocallaghan.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73766 91177308-0d34-0410-b5e6-96231b3b80d8

Since -Wno-long-long is ignored without -pedantic,
place it with -pedantic.  Remove -Wunused since it
is implied by -Wall.  Group -Wno-unused-parameter
with -Wall -W since it is the combination of these
two that turns on -Wunused-parameter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73765 91177308-0d34-0410-b5e6-96231b3b80d8

Add comments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73761 91177308-0d34-0410-b5e6-96231b3b80d8

Fix for PR2484: add an SSE1 pattern for a shuffle we normally prefer to
handle with an SSE2 instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73760 91177308-0d34-0410-b5e6-96231b3b80d8

Should be using Bcc (average) latency to determine if-conversion threshold, not BL.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73759 91177308-0d34-0410-b5e6-96231b3b80d8

Minor cleanup; fixes review comments for a previous patch.  Sorry for
taking so long to get to this!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73757 91177308-0d34-0410-b5e6-96231b3b80d8

Teach jump threading to look at comparisons between phi nodes and non-constants.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73755 91177308-0d34-0410-b5e6-96231b3b80d8

Misc tweaks to Intel asm printing to make it more compatible with MASM.
Patch by Benedict Gaster.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73753 91177308-0d34-0410-b5e6-96231b3b80d8

Improve tail call elim to move loads above readonly calls
when it allows forming a tail call.  Patch by Frits van
Bommel.  This implements PR4323.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73752 91177308-0d34-0410-b5e6-96231b3b80d8

part of PR4405: disable a contentious optimization for
strcmp -> memcmp when the lengths of the strings are unknown.

Patch by Nick Lewycky!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73751 91177308-0d34-0410-b5e6-96231b3b80d8

More VNInfo tweaking, plus a little progress on intra-block splitting.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73750 91177308-0d34-0410-b5e6-96231b3b80d8

Transfer dead markers when a ldrd is changed into a ldm or a pair of ldr.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73749 91177308-0d34-0410-b5e6-96231b3b80d8

Include DataTypes.h for 'uint64_t'.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73748 91177308-0d34-0410-b5e6-96231b3b80d8

Latency information for ARM v6. It's rough and not yet hooked up. Right now  we are only using branch latency to determine if-conversion limits.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73747 91177308-0d34-0410-b5e6-96231b3b80d8

Mark a few Thumb instructions commutable; just happened to spot this
while experimenting.  I'm reasonably sure this is correct, but please
tell me if these instructions have some strange property which makes this
change unsafe.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73746 91177308-0d34-0410-b5e6-96231b3b80d8

Fix weird class-size-being-different problems.  At some level this is being caused by config.h not being
included everywhere.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73745 91177308-0d34-0410-b5e6-96231b3b80d8

add a file I missed, this goes with r73743

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73744 91177308-0d34-0410-b5e6-96231b3b80d8

Add some scaffolding for a new experimental asmprinter
implementation.  The idea is that we want asmprinting to
work by converting MachineInstrs into a new MCInst class,
then the per-instruction asmprinter works on MCInst.  MCInst
and the new asmprinters will not depend on most of the
llvm code generators.  This allows building diassemblers
that don't link in the whole llvm code generator.  This is
step #1 of many.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73743 91177308-0d34-0410-b5e6-96231b3b80d8

merge the common darwin settings from the X86/PPC/ARM targets
into DarwinTargetAsmInfo.cpp.  The remaining differences should
be evaluated.  It seems strange that x86/arm has .zerofill but ppc
doesn't, etc.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@73742 91177308-0d34-0410-b5e6-96231b3b80d8