10fc6e9650
Separate const from non-const stuff during mergeing
...
llvm-svn: 100642
2010-04-07 18:19:13 +00:00
0453de0133
Some initial version of global merger
...
llvm-svn: 100641
2010-04-07 18:19:07 +00:00
dd1c82141c
Fix memory leaks for external symbol name strings.
...
llvm-svn: 100601
2010-04-07 03:36:01 +00:00
6ac5cc973c
Clean up some signedness oddities in this code noticed by clang.
...
llvm-svn: 100599
2010-04-07 01:49:15 +00:00
5d7f0a0fdd
Move printing of target-indepedent DEBUG_VALUE comments
...
into AsmPrinter. Target-dependent form is still generated
by FastISel and still handled in X86 code.
llvm-svn: 100596
2010-04-07 01:15:14 +00:00
796583eec0
Fix a number of clang -Wsign-compare warnings that didn't have an obvious
...
solution. The only reason these don't fire with gcc-4.2 is that gcc turns off
part of -Wsign-compare in C++ on accident.
llvm-svn: 100581
2010-04-06 23:35:53 +00:00
b36c70913b
Revert 100573, it's causing some testsuite problems.
...
llvm-svn: 100578
2010-04-06 22:45:26 +00:00
85b35b6214
Move printing of DEBUG_VALUE comments to target-independent place.
...
There is probably a more elegant way to do this.
llvm-svn: 100573
2010-04-06 22:21:07 +00:00
5202269dc4
Expand SELECT and SELECT_CC for NEON vector types.
...
Radar 7770501.
llvm-svn: 100568
2010-04-06 22:02:24 +00:00
4dac890600
Fix PR6696 and PR6663
...
When a frame pointer is not otherwise required, and dynamic stack alignment
is necessary solely due to the spilling of a register with larger alignment
requirements than the default stack alignment, the frame pointer can be both
used as a general purpose register and a frame pointer. That goes poorly, for
obvious reasons. This patch brings back a bit of old logic for identifying
the use of such registers and conservatively reserves the frame pointer
during register allocation in such cases.
For now, implement for X86 only since it's 32-bit linux which is hitting this,
and we want a targeted fix for 2.7. As a follow-on, this will be expanded
to handle other targets, as theoretically the problem could arise elsewhere
as well.
llvm-svn: 100559
2010-04-06 20:26:37 +00:00
41051a0bfe
Don't try to collapse DomainValues onto an incompatible SSE domain.
...
This fixes the Bullet regression on i386/nocona.
llvm-svn: 100553
2010-04-06 19:48:56 +00:00
1a9b3f3484
Properly enable load clustering.
...
Operand 2 on a load instruction does not have to be a RegisterSDNode for this to
work.
llvm-svn: 100497
2010-04-05 23:48:02 +00:00
23d16d5b86
Fix ADD32rr_alt instruction encoding bug. Patch by Marius Wachtler.
...
llvm-svn: 100480
2010-04-05 22:21:09 +00:00
1290fa0f72
Remove FIXME.
...
llvm-svn: 100466
2010-04-05 21:14:32 +00:00
6e39c4a097
don't use emitlabel in the arm asm printer yet, the order
...
isn't well specified. ARM really needs to have its instprinter
finished at some point.
llvm-svn: 100439
2010-04-05 17:52:31 +00:00
6a0e89aefb
fix a couple problems I introduced handling symbols
...
with spaces in them. Sym->getName() != OS << *Sym
llvm-svn: 100434
2010-04-05 16:32:14 +00:00
0151d7b025
Disambiguate else.
...
llvm-svn: 100423
2010-04-05 10:17:15 +00:00
305f2efb63
unthread MMI from FastISel
...
llvm-svn: 100416
2010-04-05 06:05:26 +00:00
82ff9af068
remove the MMI pointer from MachineFrameInfo.
...
llvm-svn: 100415
2010-04-05 05:57:52 +00:00
50b1bf63a7
simplify code.
...
llvm-svn: 100412
2010-04-05 05:48:36 +00:00
dacfd2c6d4
Get rid of traling whitespaces. No functionality change.
...
llvm-svn: 100404
2010-04-05 04:51:50 +00:00
dba13e7922
The disassembler impl. of MCDisassembler::getInstruction() was using the pattern
...
uint32_t insn;
MemoryObject.readBytes(Address, 4, (uint8_t*)&insn, NULL)
to read 4 bytes of memory contents into a 32-bit uint variable. This leaves the
interpretation of byte order up to the host machine and causes PPC test cases of
arm-tests, neon-tests, and thumb-tests to fail. Fixed to use a byte array for
reading the memory contents and shift the bytes into place for the 32-bit uint
variable in the ARM case and 16-bit halfword in the Thumb case.
llvm-svn: 100403
2010-04-05 04:46:17 +00:00
f0ef4e4019
implement EmitFunctionEntryLabel to emit the .cc_top directive,
...
allowing xcore to use the normal runOnMachineFunction
implementation.
llvm-svn: 100402
2010-04-05 04:44:02 +00:00
a49ac8ace0
prune some #includes.
...
llvm-svn: 100399
2010-04-05 04:04:10 +00:00
b93331f3be
Replace TSFlagsFields and TSFlagsShifts with a simpler TSFlags field.
...
When a target instruction wants to set target-specific flags, it should simply
set bits in the TSFlags bit vector defined in the Instruction TableGen class.
This works well because TableGen resolves member references late:
class I : Instruction {
AddrMode AM = AddrModeNone;
let TSFlags{3-0} = AM.Value;
}
let AM = AddrMode4 in
def ADD : I;
TSFlags gets the expected bits from AddrMode4 in this example.
llvm-svn: 100384
2010-04-05 03:10:20 +00:00
7cfa70e9b3
fastisel doesn't need DwarfWriter, remove some tendricles.
...
llvm-svn: 100381
2010-04-05 02:19:28 +00:00
0b8adb0652
Temporarily remove to disable building of ARM disassembler.
...
llvm-svn: 100380
2010-04-05 01:57:50 +00:00
492a82e426
Re-apply 100265 but instead disable building of ARM disassembly for now.
...
llvm-svn: 100379
2010-04-05 01:34:00 +00:00
876a5015af
Reverting 100265 to try to get buildbots green again. Lots of self-hosting buildbots started complaining since this commit. Also xfail ARM disassembly tests.
...
llvm-svn: 100378
2010-04-05 01:04:27 +00:00
626cb66fdb
just have all targets create the DwarfWriter.
...
llvm-svn: 100377
2010-04-05 00:42:55 +00:00
8b30492da3
simplify various getAnalysisUsage implementations.
...
llvm-svn: 100376
2010-04-05 00:38:44 +00:00
324c86600d
eliminate the magic AbsoluteDebugSectionOffsets MAI hook,
...
which is really a property of the section being referenced.
Add a predicate to MCSection to replace it.
Yay for reduction in magic.
llvm-svn: 100367
2010-04-04 23:22:29 +00:00
8964b838e4
revert my patch, need to reconsider this and figure out what is really going on.
...
llvm-svn: 100358
2010-04-04 21:49:31 +00:00
407f848835
fix pasto, this is the wrong setting for arm elf.
...
llvm-svn: 100357
2010-04-04 21:37:20 +00:00
d03ac95d5d
Clean up SSEDomainFix pass.
...
Restrict bit mask operations to the DomainValue class. Rename methods for
clarity.
llvm-svn: 100353
2010-04-04 21:27:26 +00:00
4af7c5a650
don't reset the default.
...
llvm-svn: 100352
2010-04-04 21:06:50 +00:00
7bde8c07a7
clean up the asmprinter header and privatize some stuff.
...
llvm-svn: 100342
2010-04-04 18:52:31 +00:00
21dc46e256
remove TargetMachine.h #include, also, TRI isn't used frequently
...
enough to warrant caching in AsmPrinter, so remove it.
llvm-svn: 100336
2010-04-04 18:06:11 +00:00
42caaa4f5b
Switch SSEDomainFix to SpecificBumpPtrAllocator.
...
llvm-svn: 100332
2010-04-04 18:00:21 +00:00
d20699bc87
Momentous day: remove the "O" member from AsmPrinter. Now all
...
"asm printering" happens through MCStreamer. This also
Streamerizes PIC16 debug info, which escaped my attention.
This removes a leak from LLVMTargetMachine of the 'legacy'
output stream.
llvm-svn: 100327
2010-04-04 08:18:47 +00:00
3cb9086c26
mc'ize the remaining uses of O.
...
llvm-svn: 100322
2010-04-04 07:23:00 +00:00
4f63f7ee8e
finish eliminating uses of O.
...
llvm-svn: 100321
2010-04-04 07:17:25 +00:00
794b2f1b37
mcize more of ppc stub printing.
...
llvm-svn: 100320
2010-04-04 07:12:28 +00:00
5e5961864f
mcize a bunch more stuff, using EmitRawText for things we
...
don't have mcstreamer support for yet.
llvm-svn: 100319
2010-04-04 07:05:53 +00:00
3d86cd6710
convert the non-MCInstPrinter'ized EmitInstruction
...
implementations to use EmitRawText instead of writing
directly to "O".
llvm-svn: 100318
2010-04-04 06:12:20 +00:00
83a093183c
streamerize the rest of PIC16 asm printer.
...
llvm-svn: 100317
2010-04-04 05:53:03 +00:00
d479317d65
streamerize printing of dbg_value, the x86 backend is now fully
...
streamerized for everything.
llvm-svn: 100316
2010-04-04 05:40:34 +00:00
bf43d4b6e9
split DEBUG_VALUE printing stuff out to its own method.
...
llvm-svn: 100315
2010-04-04 05:38:19 +00:00
9b13639f45
mc'ize elf stub printing, convert cygwin stuff to EmitRawText,
...
which will abort in .o file writing mode.
llvm-svn: 100314
2010-04-04 05:35:04 +00:00
3bb09768cb
fix PrintAsmOperand and PrintAsmMemoryOperand to pass down
...
raw_ostream to print to.
llvm-svn: 100313
2010-04-04 05:29:35 +00:00
787253819a
use predicates in DBG_VALUE printing code to simplify it.
...
llvm-svn: 100312
2010-04-04 05:21:31 +00:00
562e02e4e1
remove more implicit uses of "O".
...
llvm-svn: 100311
2010-04-04 05:19:20 +00:00
7012916275
fix an ugly wart in the MCInstPrinter api where the
...
raw_ostream to print an instruction to had to be specified
at MCInstPrinter construction time instead of being able
to pick at each call to printInstruction.
llvm-svn: 100307
2010-04-04 05:04:31 +00:00
76c564b1bb
change a ton of code to not implicitly use the "O" raw_ostream
...
member of AsmPrinter. Instead, pass it in explicitly.
llvm-svn: 100306
2010-04-04 04:47:45 +00:00
c576ee9040
Reapply address space patch after fixing an issue in MemCopyOptimizer.
...
Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)
llvm-svn: 100304
2010-04-04 03:10:48 +00:00
f33c7fcc28
asmstreamerize the .size directive for function bodies, force clients
...
of printOffset to pass in a stream to print to.
llvm-svn: 100296
2010-04-03 22:28:33 +00:00
fc8c3b7547
Get rid of the middleman (ARMAlgorithm), which causes more trouble than the
...
abstraction it brings. And also get rid of the atexit() handler, it does not
belong in the lib directory. :-)
llvm-svn: 100265
2010-04-03 04:10:56 +00:00
884e66a545
Fix comment.
...
llvm-svn: 100259
2010-04-03 01:17:30 +00:00
58e83eb50d
Register ARMAlgorithm::DoCleanup() to be called on exit to free the memory
...
occuplied by the cached ARMAlgorithm objects.
llvm-svn: 100258
2010-04-03 01:09:47 +00:00
000e502eb1
Rewrite aesimc handling. It only takes a single input and has a single
...
dest.
llvm-svn: 100252
2010-04-02 23:48:33 +00:00
a0d74064fe
Fix another build warning.
...
llvm-svn: 100251
2010-04-02 23:43:38 +00:00
7b999ea7b7
Second try of initial ARM/Thumb disassembler check-in. It consists of a tablgen
...
backend (ARMDecoderEmitter) which emits the decoder functions for ARM and Thumb,
and the disassembler core which invokes the decoder function and builds up the
MCInst based on the decoded Opcode.
Reviewed by Chris Latter and Bob Wilson.
llvm-svn: 100233
2010-04-02 22:27:38 +00:00
7ad0ad0b9a
Added support for reporting operand token ranges
...
to the ARM AsmParser.
llvm-svn: 100232
2010-04-02 22:27:05 +00:00
2ef63183a5
Separate out the AES-NI instructions from the SSE4.2 instructions. Add
...
a new subtarget option for AES and check for the support. Add "westmere"
line of processors and add AES-NI support to the core i7.
Add a couple of TODOs for information I couldn't verify.
llvm-svn: 100231
2010-04-02 21:54:27 +00:00
010b373cf3
Fixes to the X86 disassembler. The disassembler will now
...
return an error status in all failure cases, printing
messages to debugs() only when debugging is enabled.
llvm-svn: 100229
2010-04-02 21:23:51 +00:00
593916d732
rename NewDebugLoc -> DebugLoc, prune #includes in DebugLoc.h.
...
This keeps around temporary typedef for clang/llvm-gcc so the
build won't break when I commit this :)
llvm-svn: 100218
2010-04-02 20:21:22 +00:00
6f306d7d30
use DebugLoc default ctor instead of DebugLoc::getUnknownLoc()
...
llvm-svn: 100214
2010-04-02 20:16:16 +00:00
915c5f9862
Switch the code generator (except the JIT) onto the new DebugLoc
...
representation. This eliminates the 'DILocation' MDNodes for
file/line/col tuples from -O0 -g codegen.
This remove the old DebugLoc class, making it a typedef for DebugLoc,
I'll rename NewDebugLoc next.
I didn't update the JIT to use the new apis, so it will continue to
work, but be as slow as before. Someone should eventually do this
or, better yet, rip out the JIT debug info stuff and build the JIT
on top of MC.
llvm-svn: 100209
2010-04-02 19:42:39 +00:00
61399375a2
Correctly lower memset / memcpy of undef. It should be a nop. PR6767.
...
llvm-svn: 100208
2010-04-02 19:36:14 +00:00
999c1b927b
Revert r100191 since it breaks objc in clang
...
llvm-svn: 100199
2010-04-02 18:43:02 +00:00
a972ab8564
Reapply address space patch after fixing an issue in MemCopyOptimizer.
...
Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)
llvm-svn: 100191
2010-04-02 18:04:15 +00:00
06a1639b98
Remove FIXME - if there's a better way to do this it isn't here.
...
llvm-svn: 100176
2010-04-02 04:32:37 +00:00
4bd755419f
Revert the recent alignment changes. They're broken for -Os because,
...
in particular, they end up aligning strings at 16-byte boundaries, and
there's no way for GlobalOpt to check OptForSize.
llvm-svn: 100172
2010-04-02 03:04:37 +00:00
4244d12769
Teach AnalyzeBranch, RemoveBranch and the branch
...
folder to be tolerant of debug info following the
branch(es) at the end of a block.
llvm-svn: 100168
2010-04-02 01:38:09 +00:00
8d6d0d4c58
Disambiguate conditional expression for newer GCCs.
...
llvm-svn: 100167
2010-04-02 01:31:24 +00:00
0e3218f6af
Change variables which are exactly 16 bytes to be 16-byte-aligned too.
...
This fixes test/Transforms/GlobalOpt/gv-align.ll.
llvm-svn: 100161
2010-04-02 00:46:07 +00:00
5342ddaadf
Revert r100143.
...
llvm-svn: 100146
2010-04-01 22:54:42 +00:00
f997c31598
In 64-bit mode, use i64 to lower memcpy / memset instead of f64.
...
llvm-svn: 100137
2010-04-01 20:27:45 +00:00
d9929f03cf
Add comments about DstAlign and SrcAlign.
...
llvm-svn: 100132
2010-04-01 20:10:42 +00:00
4c014c892a
- Avoid using floating point stores to implement memset unless the value is zero.
...
- Do not try to infer GV alignment unless its type is sized. It's not possible to infer alignment if it has opaque type.
llvm-svn: 100118
2010-04-01 18:19:11 +00:00
43cd9e3845
Fix sdisel memcpy, memset, memmove lowering:
...
1. Makes it possible to lower with floating point loads and stores.
2. Avoid unaligned loads / stores unless it's fast.
3. Fix some memcpy lowering logic bug related to when to optimize a
load from constant string into a constant.
4. Adjust x86 memcpy lowering threshold to make it more sane.
5. Fix x86 target hook so it uses vector and floating point memory
ops more effectively.
rdar://7774704
llvm-svn: 100090
2010-04-01 06:04:33 +00:00
738b0f9ec7
Nehalem unaligned memory access is fast.
...
llvm-svn: 100089
2010-04-01 05:58:17 +00:00
9002ac5d93
Add aeskeygenassist intrinsic and rename all of the aes intrinsics to
...
aes instead of sse4.2. Add a brief todo for a subtarget flag and rework
the aeskeygenassist instruction to more closely match the docs.
llvm-svn: 100078
2010-04-01 03:05:45 +00:00
7c90d22f4c
vml[as] are slow on 1136jf-s also.
...
llvm-svn: 100066
2010-04-01 00:13:43 +00:00
503a0ef6f4
reduce indentation, minor cleanups.
...
llvm-svn: 100042
2010-03-31 20:32:51 +00:00
58ca0a649c
Use spaces, not tabs
...
llvm-svn: 100037
2010-03-31 20:05:12 +00:00
d749aefbd5
Comment the changes for r98218 and friends inside the source code.
...
llvm-svn: 100033
2010-03-31 18:48:58 +00:00
866f5764a7
Comment the changes for r98218 and friends inside the source code.
...
llvm-svn: 100031
2010-03-31 18:47:10 +00:00
4cd5866f8e
Fix PR6750. Don't try to merge a DomainValue with itself.
...
llvm-svn: 100016
2010-03-31 17:13:16 +00:00
9986ba954c
Replace V_SET0 with variants for each SSE execution domain.
...
llvm-svn: 99975
2010-03-31 00:40:13 +00:00
710c6892be
Fix typo. Thank you, valgrind.
...
llvm-svn: 99974
2010-03-31 00:40:08 +00:00
6f6ebb663c
Enable -sse-domain-fix by default. Now with tests!
...
llvm-svn: 99954
2010-03-30 22:47:00 +00:00
3493398f13
V_SETALLONES is an integer instruction.
...
Since it is just a pxor in disguise, we should probably expand it to a full
polymorphic triple.
llvm-svn: 99953
2010-03-30 22:46:55 +00:00
dbff4e8103
Renumber SSE execution domains for better code size.
...
SSEDomainFix will collapse to the domain with the lower number when it has a
choice. The SSEPackedSingle domain often has smaller instructions, so prefer
that.
llvm-svn: 99952
2010-03-30 22:46:53 +00:00
6f7fd28824
Revert Mon Ping's change 99928, since it broke all the llvm-gcc buildbots.
...
llvm-svn: 99948
2010-03-30 22:27:04 +00:00
cf35648ebe
Revert "Enable -sse-domain-fix by default. What could possibly go wrong?"
...
Not running 'make check-all' before committing is a bad idea.
llvm-svn: 99933
2010-03-30 21:36:32 +00:00
a654df84e6
Enable -sse-domain-fix by default. What could possibly go wrong?
...
llvm-svn: 99931
2010-03-30 21:09:31 +00:00
7460571381
Added support for address spaces and added a isVolatile field to memcpy, memmove, and memset,
...
e.g., llvm.memcpy.i32(i8*, i8*, i32, i32) -> llvm.memcpy.p0i8.p0i8.i32(i8*, i8*, i32, i32, i1)
A update of langref will occur in a subsequent checkin.
llvm-svn: 99928
2010-03-30 20:55:56 +00:00
3b9af40938
Add cross-block inference to SSEDomainFix.
...
llvm-svn: 99916
2010-03-30 20:04:01 +00:00
6ad8167714
Remove the pmulld intrinsic and autoupdate it as a vector multiply.
...
Rewrite the pmulld patterns, and make sure that they fold in loads of
arguments into the instruction.
llvm-svn: 99910
2010-03-30 18:49:01 +00:00
Anton Korobeynikov