Commit Graph

13019 Commits

Author SHA1 Message Date
Evan Cheng 70f714fdbe Allow more tailcall optimization: calls with inputs that are all passed in registers.
llvm-svn: 94873
2010-01-30 01:22:00 +00:00
Evan Cheng 2d5a75b42a Don't forget to transfer target flag when inserting a tailcall instruction.
llvm-svn: 94872
2010-01-30 01:16:15 +00:00
Daniel Dunbar 76e5d70c57 MC/X86 AsmParser: Handle absolute memory operands correctly. We were doing
something totally broken and parsing them as immediates, but the .td file also
had the wrong match class so things sortof worked. Except, that is, that we
would parse
  movl $0, %eax
as
  movl 0, %eax
Feel free to guess how well that worked.

llvm-svn: 94869
2010-01-30 01:02:48 +00:00
Daniel Dunbar f05d00787a X86.td: Refactor to bring operands that use print_pcrel_imm together.
llvm-svn: 94861
2010-01-30 00:24:12 +00:00
Daniel Dunbar a97adee959 AsmMatcher/X86: Separate out sublass for memory operands that have no segment
register, and use to cleanup a FIXME in X86AsmParser.cpp.

llvm-svn: 94859
2010-01-30 00:24:00 +00:00
Johnny Chen 34a6afc68d Modified encoding bits specification for VFP instructions. In particular, the D
bit (Inst{22}) and the M bit (Inst{5}) should be left unspecified.  For binary
format instructions, Inst{6} and Inst{4} need to specified for proper decodings.

llvm-svn: 94855
2010-01-29 23:21:10 +00:00
Evan Cheng 25217ffaed PPC is not ready for sibcall optimization.
llvm-svn: 94853
2010-01-29 23:05:56 +00:00
Dale Johannesen ad00f03e86 Add assertion to humor the paranoid.
llvm-svn: 94843
2010-01-29 21:21:28 +00:00
Bob Wilson 7c42b9d51e Improve isSafeToLoadUnconditionally to recognize that GEPs with constant
indices are safe if the result is known to be within the bounds of the
underlying object.

llvm-svn: 94829
2010-01-29 19:19:08 +00:00
Evan Cheng 297a494f55 Catch more trivial tail call opportunities: no inputs and output types match.
llvm-svn: 94804
2010-01-29 06:45:59 +00:00
Sean Callanan 82436d1666 Added a custom TableGen backend to support the
enhanced disassembler, and the necessary makefile
rules to build the table for X86.

llvm-svn: 94764
2010-01-29 00:21:04 +00:00
Benjamin Kramer 29063eac23 Replace strcpy with memcpy when we have the length around anyway.
llvm-svn: 94746
2010-01-28 18:04:38 +00:00
Chris Lattner cc9a6f0580 convert the last 3 targets to use EmitFunctionBody() now that
it has before/end body hooks.

 lib/Target/Alpha/AsmPrinter/AlphaAsmPrinter.cpp |   49 ++-----------
 lib/Target/Mips/AsmPrinter/MipsAsmPrinter.cpp   |   87 ++++++------------------
 lib/Target/XCore/AsmPrinter/XCoreAsmPrinter.cpp |   56 +++------------
 test/CodeGen/XCore/ashr.ll                      |    2 
 4 files changed, 48 insertions(+), 146 deletions(-)

llvm-svn: 94741
2010-01-28 06:22:43 +00:00
Evan Cheng 346af88396 Fix a bug introduced by r94490 where it created a X86ISD::CMP whose output type is different from its inputs.
This fixes PR6146.

llvm-svn: 94731
2010-01-28 01:57:22 +00:00
Chris Lattner 125c118cd8 switch blackfin to the default runOnMachineFunction
llvm-svn: 94729
2010-01-28 01:54:33 +00:00
Chris Lattner b662b7c801 eliminate a now-useless class.
llvm-svn: 94728
2010-01-28 01:50:22 +00:00
Chris Lattner fd97a3369e Switch MSP430, SPU, Sparc, and SystemZ to use EmitFunctionBody().
Diffstat:
 6 files changed, 30 insertions(+), 284 deletions(-)

llvm-svn: 94727
2010-01-28 01:48:52 +00:00
Chris Lattner 73de5fbfc3 Give AsmPrinter the most common expected implementation of
runOnMachineFunction, and switch PPC to use EmitFunctionBody.
The two ppc asmprinters now don't heave to define 
runOnMachineFunction.

llvm-svn: 94722
2010-01-28 01:28:58 +00:00
Chris Lattner fa1f67ea38 switch ARM to EmitFunctionBody().
llvm-svn: 94719
2010-01-28 01:10:34 +00:00
Chris Lattner 94a946cac4 Remove the argument from EmitJumpTableInfo, because it doesn't need it.
Move the X86 implementation of function body emission up to 
AsmPrinter::EmitFunctionBody, which works by calling the virtual
EmitInstruction method.

llvm-svn: 94716
2010-01-28 01:02:27 +00:00
Chris Lattner 2b796241a7 Drop the argument to AsmPrinter::EmitConstantPool and make it virtual.
Overload it in the ARM backend to do nothing, since is does insane
constant pool emission.

llvm-svn: 94708
2010-01-28 00:19:24 +00:00
Chris Lattner 80108921be don't emit constant pools twice.
llvm-svn: 94706
2010-01-28 00:15:18 +00:00
Chris Lattner 56db8c3008 switch ARM to use EmitFunctionHeader.
llvm-svn: 94703
2010-01-27 23:58:11 +00:00
Chris Lattner feba1e2495 eliminate the ARMFunctionInfo::Align member, using
MachineFunction::Alignment instead.

llvm-svn: 94701
2010-01-27 23:37:36 +00:00
Chris Lattner 1aef3717bd switch blackfin to use EmitFunctionHeader. BlackfinAsmPrinter.cpp
is now less than 200 LOC!

llvm-svn: 94699
2010-01-27 23:26:37 +00:00
Chris Lattner 5d9fb4b746 switch mips to use the shared EmitFunctionHeader() function
llvm-svn: 94698
2010-01-27 23:23:58 +00:00
Dale Johannesen 77b108e17b Treat MO_REG 0 location as undefined in DEBUG_VALUE,
per document.

llvm-svn: 94693
2010-01-27 22:11:16 +00:00
Jeffrey Yasskin 091217be6f Kill ModuleProvider and ghost linkage by inverting the relationship between
Modules and ModuleProviders. Because the "ModuleProvider" simply materializes
GlobalValues now, and doesn't provide modules, it's renamed to
"GVMaterializer". Code that used to need a ModuleProvider to materialize
Functions can now materialize the Functions directly. Functions no longer use a
magic linkage to record that they're materializable; they simply ask the
GVMaterializer.

Because the C ABI must never change, we can't remove LLVMModuleProviderRef or
the functions that refer to it. Instead, because Module now exposes the same
functionality ModuleProvider used to, we store a Module* in any
LLVMModuleProviderRef and translate in the wrapper methods.  The bindings to
other languages still use the ModuleProvider concept.  It would probably be
worth some time to update them to follow the C++ more closely, but I don't
intend to do it.

Fixes http://llvm.org/PR5737 and http://llvm.org/PR5735.

llvm-svn: 94686
2010-01-27 20:34:15 +00:00
Chris Lattner bc1e6f0873 add a new AsmPrinter::EmitFunctionEntryLabel virtual function,
which allows targets to override function entry label emission.
Use it to convert linux/ppc to use EmitFunctionHeader().

llvm-svn: 94667
2010-01-27 07:21:55 +00:00
Evan Cheng 85476f304c Perform trivial tail call optimization for callees with "C" ABI. These are done
even when -tailcallopt is not specified and it does not require changing ABI.
First case is the most trivial one. Perform tail call optimization when both
the caller and callee do not return values and when the callee does not take
any input arguments.

llvm-svn: 94664
2010-01-27 06:25:16 +00:00
Chris Lattner 57f064722a ppc/linux isn't ready for this and it was an accident that it was included.
This should fix a bunch of linux buildbot failures.

llvm-svn: 94643
2010-01-27 01:02:43 +00:00
Chris Lattner 2ab11001cc use existing basic block numbers instead of recomputing
a new set of them.

llvm-svn: 94631
2010-01-27 00:20:02 +00:00
Chris Lattner 9da5e1f650 Switch MSP430, CellSPU, SystemZ, Darwin/PPC, Alpha, and Sparc to
EmitFunctionHeader:

7 files changed, 16 insertions(+), 210 deletions(-)

llvm-svn: 94630
2010-01-27 00:17:20 +00:00
Evan Cheng 640b58b8d7 Clarify what -tailcallopt option actually do.
llvm-svn: 94628
2010-01-27 00:10:09 +00:00
Jim Grosbach a3575ca846 Adjust setjmp instruction sequence to not need 32-bit alignment padding
llvm-svn: 94627
2010-01-27 00:07:20 +00:00
Evan Cheng 67a69dd2ed Eliminate target hook IsEligibleForTailCallOptimization.
Target independent isel should always pass along the "tail call" property. Change
target hook LowerCall's parameter "isTailCall" into a refernce. If the target
decides it's impossible to honor the tail call request, it should set isTailCall
to false to make target independent isel happy.

llvm-svn: 94626
2010-01-27 00:07:07 +00:00
Evan Cheng d10089a5b4 Restore to pre-94570 state.
llvm-svn: 94625
2010-01-27 00:00:57 +00:00
Evan Cheng 75766332f9 Ignore 'forced' tailcall opt in fastisel mode.
llvm-svn: 94617
2010-01-26 23:28:40 +00:00
Chris Lattner c7b91156e4 now that enough stuff is constified, move function header printing
logic up from X86 into the common code.  The other targets will 
hopefully start using this soon.

llvm-svn: 94614
2010-01-26 23:18:44 +00:00
Chris Lattner f13e5f7323 constify a method argument.
llvm-svn: 94612
2010-01-26 23:15:09 +00:00
Evan Cheng 1b7b9e852c Delete blank lines that bug me.
llvm-svn: 94610
2010-01-26 23:07:57 +00:00
Chris Lattner 3ba6da49f5 call emitconstantpool and emitjumptable like other targets.
llvm-svn: 94601
2010-01-26 22:06:58 +00:00
Chris Lattner ff234e09b9 Eliminate SetDirective, and replace it with HasSetDirective.
Default HasSetDirective to true, since most targets have it.

The targets that claim to not have it probably do, or it is
spelled differently. These include Blackfin, Mips, Alpha, and
PIC16.  All of these except pic16 are normal ELF targets, so
they almost certainly have it.

llvm-svn: 94585
2010-01-26 20:40:54 +00:00
Evan Cheng 345ec3166e Delete dead code.
llvm-svn: 94583
2010-01-26 20:36:21 +00:00
Rafael Espindola dcb03f0f6b Emit .comm alignment in bytes but .align in powers of 2 for ARM ELF.
Original patch by Sandeep Patel and updated by me.

llvm-svn: 94582
2010-01-26 20:21:43 +00:00
Chris Lattner 3dd38a8112 eliminate MCAsmInfo::NeedsSet: we now just use .set on any platform
that has it.

llvm-svn: 94581
2010-01-26 20:20:43 +00:00
Chris Lattner 526f560424 don't set to the default value.
llvm-svn: 94580
2010-01-26 20:17:34 +00:00
Evan Cheng 8703c412f4 Code refactoring, no functionality change.
llvm-svn: 94570
2010-01-26 19:04:47 +00:00
Dan Gohman 1c7e3837fb Remove SIL, DIL, and BPL from the GR8_NOREX allocation order also.
llvm-svn: 94560
2010-01-26 18:30:24 +00:00
Dan Gohman ed71d8d70b SIL, DIL, BPL, and SPL require a REX prefix.
llvm-svn: 94558
2010-01-26 18:14:22 +00:00
Chris Lattner 547c761dd6 eliminate the TargetLowering::UsesGlobalOffsetTable bool, which is
subsumed by TargetLowering::getJumpTableEncoding().  Change uses of
it to be more specific.

llvm-svn: 94529
2010-01-26 06:53:37 +00:00
Chris Lattner 8a785d7a67 Move getJTISymbol from MachineJumpTableInfo to MachineFunction,
which is more convenient, and change getPICJumpTableRelocBaseExpr
to take a MachineFunction to match.

Next, move the X86 code that create a PICBase symbol to
X86TargetLowering::getPICBaseSymbol from 
X86MCInstLower::GetPICBaseSymbol, which was an asmprinter specific 
library.  This eliminates a 'gross hack', and allows us to
implement X86ISelLowering::getPICJumpTableRelocBaseExpr which now
calls it.

This in turn allows us to eliminate the 
X86AsmPrinter::printPICJumpTableSetLabel method, which was the
only overload of printPICJumpTableSetLabel.

llvm-svn: 94526
2010-01-26 06:28:43 +00:00
Chris Lattner cc0f734cd0 simplify asmprinter: only emit .set directives when entries have
EK_LabelDifference32 kind and the target has .set support.  Simplify
X86AsmPrinter::printPICJumpTableSetLabel to make use of recent helpers.

llvm-svn: 94518
2010-01-26 05:15:20 +00:00
Chris Lattner 4bfbe93437 implement X86 @GOTOFF jump table entries with the new EK_Custom32
jump table entry kind, instead of overloading 
AsmPrinter::printPICJumpTableEntry.

This has a pretty horrible and inefficient FIXME around how @GOTOFF
is currently smashed into the mcsymbol name, but otherwise this is
much cleaner.

llvm-svn: 94516
2010-01-26 05:02:42 +00:00
Chris Lattner d051af7551 add a new MachineBasicBlock::getSymbol method, replacing
the AsmPrinter::GetMBBSymbol.

llvm-svn: 94515
2010-01-26 04:55:51 +00:00
Chris Lattner 170442fa12 don't bother setting the AsmPrinter::MF ivar, now that
AsmPrinter::SetupMachineFunction sets it.  Note that systemz
and msp430 didn't.  Yay for reduced inconsistency! :)

llvm-svn: 94510
2010-01-26 04:38:11 +00:00
Chris Lattner 6715952c25 make MachineFunction keep track of its ID and make
MachineFunctionAnalysis dole them out, instead of having
AsmPrinter do both.  Have the AsmPrinter::SetupMachineFunction
method set the 'AsmPrinter::MF' variable.

llvm-svn: 94509
2010-01-26 04:35:26 +00:00
Daniel Dunbar a8d7385d99 Unbreak MSVC/CMake build.
llvm-svn: 94502
2010-01-26 03:56:22 +00:00
Evan Cheng 555f61bf58 Implement cond ? -1 : 0 with sbb.
llvm-svn: 94490
2010-01-26 02:00:44 +00:00
Dale Johannesen 0563fe3445 Accept immediate as value of a dbg_value.
llvm-svn: 94489
2010-01-26 01:54:26 +00:00
Sean Callanan eb182e9c88 Added the implementation of the Intel-specific
TargetAsmLexer.

llvm-svn: 94482
2010-01-26 01:00:10 +00:00
Dale Johannesen d5575f29f1 Generate DEBUG_VALUE comments on x86. The (limited)
dbg.declare's we currently generate go through both
register allocators without perturbing the results.

llvm-svn: 94480
2010-01-26 00:09:58 +00:00
Sean Callanan ad857fc070 Added the TargetAsmLexer implementation for AT&T syntax.
llvm-svn: 94479
2010-01-26 00:08:25 +00:00
Dale Johannesen e5a4134d11 use findDebugLoc in more places.
llvm-svn: 94477
2010-01-26 00:03:12 +00:00
Jim Grosbach db67493f76 Minor jump table cleanup.
llvm-svn: 94475
2010-01-25 23:50:13 +00:00
Chris Lattner 9c1efcd4f6 in 32-bit pic mode for targets with a GOT, x86 emits jump table
entries with @GOTOFF whih is EK_GPRel32BlockAddress.

llvm-svn: 94474
2010-01-25 23:38:14 +00:00
Chris Lattner 8186eecf0f fix quoting problem jim noticed!
llvm-svn: 94472
2010-01-25 23:28:03 +00:00
Chris Lattner b6db2c6b31 Rearrange handling of jump tables. Highlights:
1. MachineJumpTableInfo is now created lazily for a function the first time
   it actually makes a jump table instead of for every function.
2. The encoding of jump table entries is now described by the
   MachineJumpTableInfo::JTEntryKind enum.  This enum is determined by the
   TLI::getJumpTableEncoding() hook, instead of by lots of code scattered
   throughout the compiler that "knows" that jump table entries are always
   32-bits in pic mode (for example).
3. The size and alignment of jump table entries is now calculated based on
   their kind, instead of at machinefunction creation time.

Future work includes using the EntryKind in more places in the compiler,
eliminating other logic that "knows" the layout of jump tables in various
situations.

llvm-svn: 94470
2010-01-25 23:26:13 +00:00
Chris Lattner a14ac3fd80 prep work to support a future where getJumpTableInfo will return
a null pointer for functions with no jump tables.  No functionality
change.

llvm-svn: 94469
2010-01-25 23:22:00 +00:00
Chris Lattner 3072add73c add a method to get the alignment of an integer type even
when we don't have one laying around.  Useful if you don't
have an llvmcontext handy.

llvm-svn: 94468
2010-01-25 23:18:11 +00:00
Johnny Chen 5542c1edf9 Make it SP, LR, PC for GPR Register Class instead of LR, SP, PC.
llvm-svn: 94465
2010-01-25 22:54:29 +00:00
Chris Lattner a4074257ab eliminate redundant argument to EmitJumpTableInfo
llvm-svn: 94464
2010-01-25 22:41:33 +00:00
Johnny Chen ab2b1a7aeb Implemented ARMInstPrinter::printThumbS4ImmOperand().
llvm-svn: 94457
2010-01-25 22:13:10 +00:00
Sean Callanan 665493102e Implemented the dialect decision logic for the X86
TargetAsmLexer.  Dialect-specific lexing code will
be placed in the functions LexTokenATT() and
LexTokenIntel().

llvm-svn: 94456
2010-01-25 21:59:20 +00:00
Johnny Chen 495ac50b4e Fixed the order of GPR RegisterClass regs to be: ..., R10, R11, R12, ...
llvm-svn: 94455
2010-01-25 21:56:35 +00:00
Chris Lattner 00d188f379 handle the _set_ symbol with an MCSymbol.
llvm-svn: 94449
2010-01-25 21:17:10 +00:00
Chris Lattner 19bd039896 rename MAI::PICJumpTableDirective to MAI::GPRel32Directive to
make it clear what it is, instead of how it is used.

llvm-svn: 94448
2010-01-25 21:10:10 +00:00
Chris Lattner 360aeb76a2 pull the non-pic jump table case out of printPICJumpTableEntry
and MCize the non-pic case.  Now printPICJumpTableEntry really
is just about printing PIC entries.

llvm-svn: 94446
2010-01-25 21:01:58 +00:00
Chris Lattner ccabcd7f85 remove JumpTableDirective, it is always null.
llvm-svn: 94445
2010-01-25 20:52:54 +00:00
Chris Lattner 68d64aa28b mcize jump table symbol manipulation.
llvm-svn: 94441
2010-01-25 19:51:38 +00:00
Chris Lattner 6330d5330e sink an arm specific method out of asmprinter into the ARMAsmPrinter and
rename it to avoid shadowing.

llvm-svn: 94440
2010-01-25 19:39:52 +00:00
Chris Lattner a9ee93ef42 remove dead code: the x86 target never sets usesGlobalOffsetTable,
even on x86-32/elf which uses a GOT.

llvm-svn: 94439
2010-01-25 19:23:04 +00:00
Chris Lattner 601ef33c77 mcstreamerize .file and .file. This also fixes an issue where the
normal form of .file would fail if the filename had a weird character
in it.

llvm-svn: 94437
2010-01-25 18:58:59 +00:00
Chris Lattner bc696445e1 emit ELF .type directives through MCStreamer instead of doing it textually.
llvm-svn: 94436
2010-01-25 18:33:40 +00:00
Chris Lattner 48536010fb Remove support for i386 tiger tools for aligned common symbols.
Previously we would just silently miscompile code that used aligned
common's, now at least you'll get a build error.  tiger-ppc already
triggered the build error because it didn't have a version of this
logic.

llvm-svn: 94412
2010-01-25 07:26:17 +00:00
Chris Lattner 1588f18453 don't set value to its default.
llvm-svn: 94411
2010-01-25 07:23:14 +00:00
Jim Grosbach fa9ae6739a ARM does accept the .comm directive alignment.
llvm-svn: 94408
2010-01-25 04:59:07 +00:00
Rafael Espindola a1141dd6ab Fix PR6134.
We are not emitting alignments on Darwin for "bar". Not sure what is the
correct way to do it.

llvm-svn: 94400
2010-01-25 02:27:39 +00:00
Chris Lattner ecb457c4e2 linux/ppc does use alignment in bytes, not pow-2. This fixes PR6129.
It looks like linux/arm and linux/mips have the same setting, which 
are probably wrong.  Someone who cares about ARM and MIPS should 
investigate with the testcase in PR6129.

llvm-svn: 94381
2010-01-24 20:54:45 +00:00
Chris Lattner d96bcb5f22 don't reset the default.
llvm-svn: 94380
2010-01-24 20:53:12 +00:00
Chris Lattner 823aed16f9 make -fno-rtti the default unless a directory builds with REQUIRES_RTTI.
llvm-svn: 94378
2010-01-24 20:43:08 +00:00
Chris Lattner e3a68d1063 reassociate should do this.
llvm-svn: 94374
2010-01-24 20:17:09 +00:00
Chris Lattner 7e3f8b60d6 add a note.
llvm-svn: 94373
2010-01-24 20:01:41 +00:00
Chris Lattner e0359b4fe7 move PR5945 here.
llvm-svn: 94350
2010-01-24 02:27:03 +00:00
Chris Lattner 807a3bcbbb fix a parsing problem on instructions like:
movw	$8, (_cost_table_-L97$pb)+66(%eax)

After the parens, we could still have a binop.

llvm-svn: 94345
2010-01-24 01:07:33 +00:00
Chris Lattner 1b35bbe813 change the canonical form of "cond ? -1 : 0" to be
"sext cond" instead of a select.  This simplifies some instcombine
code, matches the policy for zext (cond ? 1 : 0 -> zext), and allows
us to generate better code for a testcase on ppc.

llvm-svn: 94339
2010-01-24 00:09:49 +00:00
Mon P Wang 586d997e98 Improved widening loads by adding support for wider loads if
the alignment allows.  Fixed a bug where we didn't use a
vector load/store for PR5626.

llvm-svn: 94338
2010-01-24 00:05:03 +00:00
Chris Lattner 249da5cb73 implement a simple instcombine xform that has been in the
readme forever.

llvm-svn: 94318
2010-01-23 18:49:30 +00:00
Chris Lattner 97331ae668 add a note
llvm-svn: 94317
2010-01-23 18:42:37 +00:00
Chris Lattner 082da53f9a add some notes, making posix-memalign be nocapture would be an easy improvement.
llvm-svn: 94312
2010-01-23 17:59:23 +00:00
Bill Wendling 8cbc25d945 Remove the '-disable-scheduling' flag and replace it with the 'source' option of
the '-pre-RA-sched' flag. It actually makes more sense to do it this way. Also,
keep track of the SDNode ordering by default. Eventually, we would like to make
this ordering a way to break a "tie" in the scheduler. However, doing that now
breaks the "CodeGen/X86/abi-isel.ll" test for 32-bit Linux.

llvm-svn: 94308
2010-01-23 10:26:57 +00:00
Chris Lattner b1301f7bce mcize lcomm, simplify .comm, extend both to support 64-bit sizes.
llvm-svn: 94299
2010-01-23 07:47:02 +00:00
Chris Lattner 76bdea3290 resolve a fixme: the "nonexecutable stack directive" is actually
a .section.  Switch to it with SwitchSection.

However, I think that this directive should be safe on any ELF target.
If so, we should hoist it up out of the X86 and SystemZ targets.

llvm-svn: 94298
2010-01-23 07:21:06 +00:00
Chris Lattner c55ea3fbcc use helpers.
llvm-svn: 94296
2010-01-23 07:00:21 +00:00
Chris Lattner 685508cf49 move the various directive enums out of the MCStreamer class
into a new MCDirectives.h file.

llvm-svn: 94294
2010-01-23 06:39:22 +00:00
Eric Christopher c1451d764f Don't lower splat vector load to relative to the esp if the
stack may be misaligned.

Update test accordingly.

Patch by Evan Cheng!

llvm-svn: 94291
2010-01-23 06:02:43 +00:00
Chris Lattner b065f1bc31 remove unneeded directive set.
llvm-svn: 94286
2010-01-23 05:28:25 +00:00
Chris Lattner 5241cde324 use some helpers instead of duplicating logic.
llvm-svn: 94285
2010-01-23 05:26:25 +00:00
Ted Kremenek 4a1aa119a3 Update CMake build.
llvm-svn: 94278
2010-01-23 04:01:37 +00:00
Sean Callanan 5051cb8010 Added the skeleton for the implementation of the X86
target-specific AsmLexer, and added the appropriate
registration calls.

llvm-svn: 94275
2010-01-23 02:43:15 +00:00
Sean Callanan 86c1181c3d Modified the register matcher function in AsmMatcher to
be static.  Also made it possible for clients to get it
and no other functions from ...GenAsmMatcher.inc by
defining REGISTERS_ONLY before including GenAsmMatcher.inc.
This sets the stage for target-specific lexers that can
identify registers and return AsmToken::Register as
appropriate.

llvm-svn: 94266
2010-01-23 00:40:33 +00:00
Evan Cheng 16cf934381 Be more conservative with clustering f32 / f64 loads.
llvm-svn: 94254
2010-01-22 23:49:11 +00:00
Daniel Dunbar 862023fcb2 Mark EH_RETURN64 as CodeGenOnly.
llvm-svn: 94205
2010-01-22 20:16:37 +00:00
Chris Lattner efcadd659d add missing break, PR6114, patch by M Wahab.
llvm-svn: 94197
2010-01-22 19:11:53 +00:00
Chris Lattner e5f7214fac use both the Mangler and the CBE's mangler for identifiers,
this should fix a problem duncan saw handling symbols with '.' in them.

llvm-svn: 94192
2010-01-22 18:33:00 +00:00
Benjamin Kramer 94b993ed8a Simplify some uses of str(n)cmp with StringRef.
llvm-svn: 94189
2010-01-22 18:21:23 +00:00
Chris Lattner 7ba0661f27 Stop building RTTI information for *most* llvm libraries. Notable
missing ones are libsupport, libsystem and libvmcore.  libvmcore is
currently blocked on bugpoint, which uses EH.  Once it stops using
EH, we can switch it off.

This #if 0's out 3 unit tests, because gtest requires RTTI information.
Suggestions welcome on how to fix this.

llvm-svn: 94164
2010-01-22 06:49:46 +00:00
Evan Cheng 4f026f3750 Add two target hooks to determine whether two loads are near and should be scheduled together.
llvm-svn: 94147
2010-01-22 03:34:51 +00:00
Chris Lattner 81e2b9a237 add new file
llvm-svn: 94132
2010-01-22 01:58:44 +00:00
Chris Lattner 00646cfbd4 create a new MCParser library and move some stuff into it.
llvm-svn: 94129
2010-01-22 01:44:57 +00:00
Chris Lattner 88c46d713b stub out a new TargetAsmLexer interface.
llvm-svn: 94125
2010-01-22 00:58:59 +00:00
Jim Grosbach 267430f74d Fix PR5694. The CMN instructions set the flags differently from CMP, so they
cannot be directly interchanged for comparisons against negated values.
Disable the CMN instructions for the time being.

llvm-svn: 94119
2010-01-22 00:08:13 +00:00
Chris Lattner 43f2eecffa remove dead .erase.
llvm-svn: 94098
2010-01-21 20:43:39 +00:00
Chris Lattner 3785c11cd4 fix a problem with a missing _, testcase pending.
llvm-svn: 94095
2010-01-21 19:58:19 +00:00
Dan Gohman 51ad99d2c5 Re-implement the main strength-reduction portion of LoopStrengthReduction.
This new version is much more aggressive about doing "full" reduction in
cases where it reduces register pressure, and also more aggressive about
rewriting induction variables to count down (or up) to zero when doing so
reduces register pressure.

It currently uses fairly simplistic algorithms for finding reuse
opportunities, but it introduces a new framework allows it to combine
multiple strategies at once to form hybrid solutions, instead of doing
all full-reduction or all base+index.

llvm-svn: 94061
2010-01-21 02:09:26 +00:00
Evan Cheng 5d30f7c91c Fix a minor issue in x86 load / store folding table. movups does an unaligned load so it doesn't require 16-byte alignment.
llvm-svn: 94058
2010-01-21 00:55:14 +00:00
Chris Lattner 32fb9564bc eliminate FnStubInfo, using MachineModuleInfoMachO instead.
this makes function stub emission determinstic.

llvm-svn: 94033
2010-01-20 21:36:48 +00:00
Dale Johannesen c5db599813 make findDebugLoc a class method
llvm-svn: 94032
2010-01-20 21:36:02 +00:00
Chris Lattner 2f7c279257 split function stub printing out to its own function,
no functionality change.

llvm-svn: 94030
2010-01-20 21:19:44 +00:00
Chris Lattner 1ff4794cbb eliminate the GVStubs and HiddenGVStubs maps, and use
MachineModuleInfoMachO instead.  This eliminates two sources
of nondeterministic output in the ppc backend, but function
stubs are still bad.

llvm-svn: 94029
2010-01-20 21:16:14 +00:00
David Greene 0985160c54 When XDEBUG is enabled, check for SelectionDAG cycles at some key
points.  This will help us find future problems like the one
described in PR6019.

llvm-svn: 94019
2010-01-20 20:13:31 +00:00
Chris Lattner 6a7939243f reduce redundant are's.
llvm-svn: 94008
2010-01-20 17:53:51 +00:00
Chris Lattner db42c8d23d inline and radically simplify printDataDirective. It will eventually
go completely away.

llvm-svn: 93994
2010-01-20 07:33:29 +00:00
Chris Lattner 2b4364fe03 revert 93934, removing the MCAsmInfo endianness bit. I can't
stomache MCAsmInfo having this, and I found a better solution to
this layering issue.

llvm-svn: 93985
2010-01-20 06:34:14 +00:00
Dale Johannesen 91970b4ea2 Move findDebugLoc somewhere more central. Fix
more cases where debug declarations affect
debug line info.

llvm-svn: 93953
2010-01-20 00:19:24 +00:00
Bob Wilson 6a4491b8c7 Wrap some comments to 80 columns.
llvm-svn: 93940
2010-01-19 22:56:26 +00:00
Dale Johannesen aa6655557b Fix a case where debug_value was perturbing the
line number info.

llvm-svn: 93937
2010-01-19 22:50:05 +00:00
Chris Lattner 003b5c8472 give MCAsmInfo a 'has little endian' bit. This is unfortunate, but
I really want clients of the streamer to be able to say "emit this
64-bit integer" and have it get broken down right by the streamer.

I may change this in the future, we'll see how it works out.

llvm-svn: 93934
2010-01-19 22:42:28 +00:00
Chris Lattner ecaf0dcbc6 eliminate AsmPrinter::EmitZeros: just use MCStreamer directly.
llvm-svn: 93918
2010-01-19 21:51:22 +00:00
Sean Callanan 936b0d3144 Promoted the getTok() method to MCAsmParser so that
the two token accessor functions are declared consistently.
Modified the clients of MCAsmParser to reflect this change.

llvm-svn: 93916
2010-01-19 21:44:56 +00:00
Jakob Stoklund Olesen bdc17f6840 Remove predicates when changing an add into an unpredicable mov.
Since the mov is executed unconditionally, make sure that the add didn't have
any predicate.

llvm-svn: 93909
2010-01-19 21:08:28 +00:00
Sean Callanan a83fd7d52c Propagated the parser-side Lex function's declaration to
MCAsmParser, and changed the target-specific AsmParsers
to use it.

llvm-svn: 93900
2010-01-19 20:27:46 +00:00
Bruno Cardoso Lopes e1afaf5c22 Fix a bug introduced on r92564 where the name "Node" was already
in use by Mips.

llvm-svn: 93897
2010-01-19 19:57:07 +00:00
Chris Lattner c35681b298 Generalize mcasmstreamer data emission APIs to take an address space
identifier.  There is no way to work around it.

llvm-svn: 93896
2010-01-19 19:46:13 +00:00
Chris Lattner 9c31b0c695 remove MAI::ZeroDirectiveSuffix, which is only used by MASM,
which we don't support anymore.

llvm-svn: 93886
2010-01-19 18:37:01 +00:00
Jim Grosbach 04770f2aa1 For aligned load/store instructions, it's only required to know whether a
function can support dynamic stack realignment. That's a much easier question
to answer at instruction selection stage than whether the function actually
will have dynamic alignment prologue. This allows the removal of the
stack alignment heuristic pass, and improves code quality for cases where
the heuristic would result in dynamic alignment code being generated when
it was not strictly necessary.

llvm-svn: 93885
2010-01-19 18:31:11 +00:00
Bruno Cardoso Lopes 4f34432d09 On pic function calls some arguments were marked dead and
the instruction to load those args removed. This fix PR6071

llvm-svn: 93880
2010-01-19 17:00:43 +00:00
Bruno Cardoso Lopes c9818a77cb load f64 +0.0 in a cleaner way. This fix part of PR5445
llvm-svn: 93876
2010-01-19 12:53:04 +00:00
Bruno Cardoso Lopes 2f5c8e3495 Fix return registers for mips eabi
llvm-svn: 93875
2010-01-19 12:37:35 +00:00
Chris Lattner cd2915e467 stop using the .lcomm pseudoop on darwin, instead, directly use the
.zerofill directive.  Streamerize its generation.

llvm-svn: 93868
2010-01-19 06:25:51 +00:00