maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
42,014 Commits 23 Branches 0 Tags 2.2 GiB
Commit Graph

37770 Commits

Author SHA1 Message Date
Evan Cheng 4a3c5eab34 - Fix a subtle bug in RemoveCopyByCommutingDef. ALR is the live range where the source is defined; BLR is the live range which is defined by the copy. 
If ALR and BLR overlaps and end of BLR extends beyond end of ALR, e.g.                                                                                                 
 A = or A, B                                                                                                                                                            
 ...                                                                                                                                                                    
 B = A                                                                                                                                                                  
 ...                                                                                                                                                                    
 C = A<kill>                                                                                                                                                            
 ...                                                                                                                                                                    
   = B                                                                                                                                                                  
                                                                                                                                                                        
then do not add kills of A to the newly created B interval.
- Also fix some kill info update bug.

llvm-svn: 48141
 2008-03-10 08:11:32 +00:00
Evan Cheng 831ae49599 Doh 
llvm-svn: 48140
 2008-03-10 07:59:01 +00:00
Tanya Lattner f7c52a1d52 Add nosubmit option to control what is sent back to the results server. 
Patch by Joachim Durchholz. Thanks!

llvm-svn: 48139
 2008-03-10 07:28:08 +00:00
Owen Anderson 75d04819a6 Move StrongPHIElimination after live interval analysis. This will make things happier down the road. 
llvm-svn: 48138
 2008-03-10 07:22:36 +00:00
Tanya Lattner 5f4b355f20 Remove llvm-upgrade and update tests. 
llvm-svn: 48137
 2008-03-10 07:21:50 +00:00
Evan Cheng b5d11980d9 Avoid creating BUILD_VECTOR of all zero elements of "non-normalized" type (e.g. v8i16 on x86) after legalizer. Instruction selection does not expect to see them. In all likelihood this can only be an issue in a bugpoint reduced test case. 
llvm-svn: 48136
 2008-03-10 07:19:13 +00:00
Chris Lattner 176542a970 remove an execution test. 
llvm-svn: 48135
 2008-03-10 06:53:14 +00:00
Chris Lattner 853b6ba425 switch from hard coded g++/as tools to the ones detected from 
the build system.  Patch by Joachim Durchholz  for PR2121

llvm-svn: 48134
 2008-03-10 06:52:10 +00:00
Chris Lattner 0c1d7b43a6 Fix a typo 
llvm-svn: 48133
 2008-03-10 06:49:40 +00:00
Chris Lattner ca43433493 Fix PR2120 by changing the replacement order to change compile_cxx 
before compile_c.  Patch by Joachim Durchholz!

llvm-svn: 48132
 2008-03-10 06:45:35 +00:00
Christopher Lamb 4ba3f0430b Allow insert_subreg into implicit, target-specific values. 
Change insert/extract subreg instructions to be able to be used in TableGen patterns.
Use the above features to reimplement an x86-64 pseudo instruction as a pattern.

llvm-svn: 48130
 2008-03-10 06:12:08 +00:00
Nick Lewycky 5cb21155e4 Update the .cvs files to match today's asm syntax change. 
llvm-svn: 48128
 2008-03-10 05:01:34 +00:00
Christopher Lamb 80b928e600 Revert accidentally committed local changes. 
llvm-svn: 48126
 2008-03-10 04:16:09 +00:00
Christopher Lamb 0faed2a9a6 Add support in TableGen for unknown operands that infer their type from the pattern their used in. This will be used to allow insert/extract subreg patterns in .td files! 
llvm-svn: 48125
 2008-03-10 04:13:41 +00:00
Nick Lewycky fb2c1a999a Turn unwind_to into "unwinds to". 
llvm-svn: 48123
 2008-03-10 02:20:00 +00:00
Dale Johannesen 4e622ec86d Increase ISD::ParamFlags to 64 bits. Increase the ByValSize 
field to 32 bits, thus enabling correct handling of ByVal
structs bigger than 0x1ffff.  Abstract interface a bit.
Fixes gcc.c-torture/execute/pr23135.c and 
gcc.c-torture/execute/pr28982b.c in gcc testsuite (were ICE'ing
on ppc32, quietly producing wrong code on x86-32.)

llvm-svn: 48122
 2008-03-10 02:17:22 +00:00
Anton Korobeynikov 148562f13e This works on x86_64 
llvm-svn: 48121
 2008-03-09 22:24:03 +00:00
Chris Lattner aed9406bb7 Darwin PPC64 indirect call target goes in X12, not R12. This fixes these 
two regression tests:
test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert.ll
test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert2.ll

llvm-svn: 48120
 2008-03-09 20:49:33 +00:00
Chris Lattner 78e817d7fc cell really does support cross-regclass moves, because R3 is in lots of different regclasses, and the code is not consistent when it comes to value tracking. 
llvm-svn: 48119
 2008-03-09 20:31:11 +00:00
Chris Lattner 3ba79ed9e8 make sure ar.pfs is in a register class, this fixes test/CodeGen/IA64/ret-0.ll 
llvm-svn: 48118
 2008-03-09 20:12:44 +00:00
Chris Lattner 4c4234b59c remove an extraneous (and ugly) default argument, thanks Duncan. 
llvm-svn: 48117
 2008-03-09 20:04:36 +00:00
Evan Cheng b46ec99b7e Fix RUN line. 
llvm-svn: 48116
 2008-03-09 19:55:22 +00:00
Ted Kremenek d48ed17d18 Fix some compilation errors on msvc: 
- "Redefinition of I" (iterator masks previous definition)
- include missing header file

Patch by Argiris Kirtzidis!

llvm-svn: 48115
 2008-03-09 18:32:50 +00:00
Nick Lewycky 0ac65c3b99 And again. 
llvm-svn: 48112
 2008-03-09 17:13:05 +00:00
Nick Lewycky 929703b29e Braces belong here. No functionality change. 
llvm-svn: 48111
 2008-03-09 17:11:18 +00:00
Anton Korobeynikov 971a4660a0 Temporary XFAIL the test, until I'll commit a fix 
llvm-svn: 48110
 2008-03-09 16:24:04 +00:00
Nick Lewycky 83750d9caf SCCP also needs to be taught to follow unwind_to 
llvm-svn: 48109
 2008-03-09 09:44:38 +00:00
Chris Lattner ce5f841bb5 fp_round's produced by getCopyFromParts should always be exact, because 
they are produced by calls (which are known exact) and by cross block copies
which are known to be produced by extends.

This improves:

define double @test2() {
	%tmp85 = call double asm sideeffect "fld0", "={st(0)}"()
	ret double %tmp85
}

from:

_test2:
	subl	$20, %esp
	# InlineAsm Start
	fld0
	# InlineAsm End
	fstpl	8(%esp)
	movsd	8(%esp), %xmm0
	movsd	%xmm0, (%esp)
	fldl	(%esp)
	addl	$20, %esp
	#FP_REG_KILL
	ret

to:

_test2:
	# InlineAsm Start
	fld0
	# InlineAsm End
	#FP_REG_KILL
	ret

by avoiding a f64 <-> f80 trip

llvm-svn: 48108
 2008-03-09 09:38:46 +00:00
Chris Lattner 86829f0ff7 teach X86InstrInfo::copyRegToReg how to copy into ST(0) from 
an RFP register class.

Teach ScheduleDAG how to handle CopyToReg with different src/dst 
reg classes.

This allows us to compile trivial inline asms that expect stuff
on the top of x87-fp stack.

llvm-svn: 48107
 2008-03-09 09:15:31 +00:00
Nick Lewycky 271506f29c Don't eliminate blocks that are only reachable by unwind_to. 
llvm-svn: 48106
 2008-03-09 08:50:23 +00:00
Chris Lattner 9e07537e8c Add ScheduleDAG support for copytoreg where the src/dst register are 
in different register classes, e.g. copy of ST(0) to RFP*.  This gets
some really trivial inline asm working that plops things on the top of
stack (PR879)

llvm-svn: 48105
 2008-03-09 08:49:15 +00:00
Chris Lattner b79bafcec8 add some code to support cross-register class copying from 
RST -> RFP{32/64/80}.  We only handle ST(0) for now.

llvm-svn: 48104
 2008-03-09 08:46:19 +00:00
Tanya Lattner aa6f5c9ddd Remove llvm-upgrade and update tests. 
llvm-svn: 48103
 2008-03-09 08:16:40 +00:00
Chris Lattner c4c9dde04c rearrange some code, no functionality change. 
llvm-svn: 48101
 2008-03-09 07:58:04 +00:00
Chris Lattner 381bbdb924 fix 80 col violation 
llvm-svn: 48100
 2008-03-09 07:51:01 +00:00
Nick Lewycky 42445be0df Firstly, having a BranchInst isn't exclusive with having an unwind_to. 
Secondly, we have to check whether the branch is actually pointing to the block
with the unwind in it. We could have gotten here because of the unwind_to alone.

llvm-svn: 48099
 2008-03-09 07:50:37 +00:00
Chris Lattner 459f518703 claim ST(x) registers are 80 bits, which is true. This doesn't affect 
codegen yet because these can't be spilled (they don't exist until after RA).

llvm-svn: 48098
 2008-03-09 07:49:01 +00:00
Chris Lattner 83b3473dd8 extend fp values with FP_EXTEND not FP_ROUND. 
llvm-svn: 48097
 2008-03-09 07:47:22 +00:00
Nick Lewycky f3d637fa14 A BB that unwind_to an "unwind" inst is that same as one that doesn't unwind_to 
at all.

llvm-svn: 48096
 2008-03-09 07:36:38 +00:00
Gordon Henriksen 95f4b77b0d This patch cleans up the OCaml bindings so that they format nicely with 
ocamldoc. It does not yet hook into the build system, though.

Patch by Erick Tryzelaar!

llvm-svn: 48095
 2008-03-09 07:17:38 +00:00
Chris Lattner 4c869594bc rename FP_SETRESULT -> FP_SET_ST0 
llvm-svn: 48094
 2008-03-09 07:08:44 +00:00
Chris Lattner d587e580a6 rename FpGETRESULT32 -> FpGET_ST0_32 etc. Add support for 
isel'ing value preserving FP roundings from one fp stack reg to another
into a noop, instead of stack traffic.

llvm-svn: 48093
 2008-03-09 07:05:32 +00:00
Chris Lattner a6ce71fb84 reduce this testcase more 
llvm-svn: 48092
 2008-03-09 06:57:21 +00:00
Chris Lattner b6387c8a74 Finish implementing a readme entry: when inserting an i64 variable 
into a vector of zeros or undef, and when the top part is obviously
zero, we can just use movd + shuffle.  This allows us to compile
vec_set-B.ll into:

_test3:
	movl	$1234567, %eax
	andl	4(%esp), %eax
	movd	%eax, %xmm0
	ret

instead of:

_test3:
	subl	$28, %esp
	movl	$1234567, %eax
	andl	32(%esp), %eax
	movl	%eax, (%esp)
	movl	$0, 4(%esp)
	movq	(%esp), %xmm0
	addl	$28, %esp
	ret

llvm-svn: 48090
 2008-03-09 05:42:06 +00:00
Nick Lewycky 11fc6f8765 Update the block cloner which fixes bugpoint on code using unwind_to (phew!) 
and also update the cloning interface's major user, the loop optimizations.

llvm-svn: 48088
 2008-03-09 05:24:34 +00:00
Nick Lewycky 5ce9b521d7 Update the inliner and simplifycfg to handle unwind_to. 
llvm-svn: 48086
 2008-03-09 05:10:13 +00:00
Nick Lewycky cc24104703 Two things. Preserve the unwind_to when splitting a BB. 
Add the ability to remove just one instance of a BB from a phi node. This fixes
the compile error in the tree now.

llvm-svn: 48085
 2008-03-09 05:04:48 +00:00
Nick Lewycky 4d0ed842b1 Prune the unwind_to labels on BBs that don't need them. Another step in the 
removal of invoke, PR1269.

llvm-svn: 48084
 2008-03-09 04:55:16 +00:00
Chris Lattner 93930dc28c add a note 
llvm-svn: 48064
 2008-03-09 01:08:22 +00:00
Chris Lattner eef374c197 Implement a readme entry, compiling 
#include <xmmintrin.h>
__m128i doload64(short x) {return _mm_set_epi16(0,0,0,0,0,0,0,1);}

into:
	movl	$1, %eax
	movd	%eax, %xmm0
	ret

instead of a constant pool load.

llvm-svn: 48063
 2008-03-09 01:05:04 +00:00
Chris Lattner 3ed815a3a5 upgrade this test 
llvm-svn: 48062
 2008-03-09 00:32:10 +00:00
Chris Lattner 031e04b7a3 make this test harder 
llvm-svn: 48061
 2008-03-09 00:30:06 +00:00
Chris Lattner 322c826c9d Fix two problems in SelectionDAGLegalize::ExpandBUILD_VECTOR's handling 
of BUILD_VECTORS that only have two unique elements:

1. The previous code was nondeterminstic, because it walked a map in
   SDOperand order, which isn't determinstic.
2. The previous code didn't handle the case when one element was undef
   very well.  Now we ensure that the generated shuffle mask has the
   undef vector on the RHS (instead of potentially being on the LHS)
   and that any elements that refer to it are themselves undef.  This
   allows us to compile CodeGen/X86/vec_set-9.ll into:

_test3:
	movd	%rdi, %xmm0
	punpcklqdq	%xmm0, %xmm0
	ret

instead of:

_test3:
	movd	%rdi, %xmm1
	#IMPLICIT_DEF %xmm0
	punpcklqdq	%xmm1, %xmm0
	ret

... saving a register.

llvm-svn: 48060
 2008-03-09 00:29:42 +00:00
Chris Lattner a1f25b0020 Teach SD some vector identities, allowing us to compile vec_set-9 into: 
_test3:
	movd	%rdi, %xmm1
	#IMPLICIT_DEF %xmm0
	punpcklqdq	%xmm1, %xmm0
	ret

instead of:

_test3:
	#IMPLICIT_DEF %rax
	movd	%rax, %xmm0
	movd	%rdi, %xmm1
	punpcklqdq	%xmm1, %xmm0
	ret

This is still not ideal.  There is no reason to two xmm regs.

llvm-svn: 48058
 2008-03-08 23:43:36 +00:00
Chris Lattner ad58828354 1) Improve comments. 
2) Don't try to insert an i64 value into the low part of a 
   vector with movq on an x86-32 target.  This allows us to 
   compile:

__m128i doload64(short x) {return _mm_set_epi16(0,0,0,0,0,0,0,1);}

into:

_doload64:
	movaps	LCPI1_0, %xmm0
	ret

instead of:

_doload64:
	subl	$28, %esp
	movl	$0, 4(%esp)
	movl	$1, (%esp)
	movq	(%esp), %xmm0
	addl	$28, %esp
	ret

llvm-svn: 48057
 2008-03-08 22:59:52 +00:00
Chris Lattner 8a6ebd23a8 minor simplifications to this code, don't create a dead 
SCALAR_TO_VECTOR on paths that end up not using it.

llvm-svn: 48056
 2008-03-08 22:48:29 +00:00
Chris Lattner 35adf46967 This one looks easy, add a note. 
llvm-svn: 48055
 2008-03-08 22:32:39 +00:00
Chris Lattner a76e23a935 move these to the appropriate file 
llvm-svn: 48054
 2008-03-08 22:28:45 +00:00
Nick Lewycky 35e92c7bea Not all users of a BB are Instructions any more. 
llvm-svn: 48047
 2008-03-08 07:48:41 +00:00
Nick Lewycky a53414fd79 Load the symbols first so that the interpreter constructor can find them when 
it tries to initialize them.

llvm-svn: 48046
 2008-03-08 02:49:45 +00:00
Andrew Lenharth e2e69ff426 much simpler test case 
llvm-svn: 48045
 2008-03-08 02:05:22 +00:00
Dan Gohman 39b0ac4e59 Remove unused runPass methods. 
llvm-svn: 48044
 2008-03-08 01:43:56 +00:00
Dale Johannesen 21a8f144a6 More ppc32 byval handling (bug fixes). Things 
are looking pretty good now.

llvm-svn: 48043
 2008-03-08 01:41:42 +00:00
Evan Cheng 95cf661534 Implement x86 support for @llvm.prefetch. It corresponds to prefetcht{0|1|2} and prefetchnta instructions. 
llvm-svn: 48042
 2008-03-08 00:58:38 +00:00
Dan Gohman d2ab5fd713 Add support for calls with i128 return values on ppc64. 
llvm-svn: 48041
 2008-03-08 00:19:12 +00:00
Bill Wendling d6951455e4 Something that kills a super-register also 
kills the sub-register.

llvm-svn: 48038
 2008-03-07 23:45:15 +00:00
Evan Cheng 4e97e94809 Fix a typo. It's causing consumer-typeset to miscompile. Perhaps more. 
llvm-svn: 48035
 2008-03-07 22:39:49 +00:00
Dan Gohman 6c11f47c54 There is no killUse. 
llvm-svn: 48034
 2008-03-07 22:24:41 +00:00
Andrew Lenharth b04bde94c1 add dropped section test case for PR2123 
llvm-svn: 48033
 2008-03-07 21:19:43 +00:00
Devang Patel 2fbf7eef9c Add new sretpromotion pass. 
llvm-svn: 48032
 2008-03-07 21:07:34 +00:00
Bill Wendling 968f32ce3d PPC64 passes arguments of integral type in i64 registers, not i32. Reflect this 
by promoting smaller integral values (i32 at this point) to i64, then truncating
to get the wanted size.

llvm-svn: 48030
 2008-03-07 20:49:02 +00:00
Dan Gohman 8d2ead2e34 Add support for lowering 128-bit shifts on ppc64. 
llvm-svn: 48029
 2008-03-07 20:36:53 +00:00
Dale Johannesen bfa252d402 Next bits of PPC byval handling. Basically functional 
but there are bugs.

llvm-svn: 48028
 2008-03-07 20:27:40 +00:00
Chris Lattner 20b5a2b037 Add support for ppc64 shifts with 7-bit (oversized) shift amount (e.g. PPCshl). 
llvm-svn: 48027
 2008-03-07 20:18:24 +00:00
Chris Lattner 25ff7e217d Replace SDT_PPCShiftOp in favor of SDTIntBinOps. This allows it to work 
with 32 or 64-bit operands/results.

llvm-svn: 48026
 2008-03-07 20:13:51 +00:00
Evan Cheng 39a3221e27 Fixed a register scavenger bug. If a def is re-defining part of a super register, there must be an implicit def of the super-register on the MI. 
llvm-svn: 48024
 2008-03-07 20:12:54 +00:00
Andrew Lenharth fc42740952 only extract main if the user didn't specify anything to extract 
llvm-svn: 48023
 2008-03-07 20:10:54 +00:00
Devang Patel 6c8219e732 RetVal is not used when there are more then one return operands. 
llvm-svn: 48022
 2008-03-07 20:08:07 +00:00
Andrew Lenharth 3906c6aea3 make error message reflect default search function name 
llvm-svn: 48021
 2008-03-07 20:07:24 +00:00
Devang Patel 780b3ca64b Update inliner to handle functions that return multiple values. 
llvm-svn: 48020
 2008-03-07 20:06:16 +00:00
Chris Lattner c1969de3e9 fix 80 col violations 
llvm-svn: 48019
 2008-03-07 20:05:43 +00:00
Devang Patel 47d774b2c8 Place for sret promotion tests. 
llvm-svn: 48016
 2008-03-07 20:00:15 +00:00
Andrew Lenharth 3f13b6630a add a pass that can extract all kinds of global values, not just functions. Update llvm-extract to use it and optionally extract a global variable if you want it too 
llvm-svn: 48015
 2008-03-07 19:51:57 +00:00
Gordon Henriksen a49d435b27 Cleanup some comments in the OCaml bindings. 
Patch by Erick Tryzelaar.

llvm-svn: 48014
 2008-03-07 19:13:06 +00:00
Gordon Henriksen 67b619609c Fix a typo. 'make clean' in bindings/ocaml would leave an output. 
llvm-svn: 48012
 2008-03-07 18:43:51 +00:00
Anton Korobeynikov e8a3e74b6c Clarify some important bits 
llvm-svn: 48010
 2008-03-07 18:34:50 +00:00
Anton Korobeynikov 0074ea3ae7 Small cleanup: propagate thread-localness via generic routine. 
No functionality change.

llvm-svn: 48009
 2008-03-07 18:32:18 +00:00
Gordon Henriksen 53aaaaff16 Regenerate. 
llvm-svn: 48008
 2008-03-07 18:20:01 +00:00
Gordon Henriksen eb967e37f2 Prefer to use ocamlc.opt to ocamlc and soforth. 
These natively compiled versions are faster.

Patch by Erick Tryzelaar!

llvm-svn: 48007
 2008-03-07 18:19:47 +00:00
Chris Lattner d4defb00df mark frem as expand for all legal fp types on x86, regardless of whether 
we're using SSE or not.  This fixes PR2122.

llvm-svn: 48006
 2008-03-07 06:36:32 +00:00
Bill Wendling 92e52c636f Add testcase. 
llvm-svn: 48005
 2008-03-06 23:34:22 +00:00
Bill Wendling 55bfd8c3f7 When setting the "unused" info, take into account something like this: 
%r3<def> = OR %x3<kill>, %x3

We don't want to mark the %r3 as unused even though it's a sub-register of %x3.

llvm-svn: 48003
 2008-03-06 23:22:43 +00:00
Evan Cheng 34173f0a43 80 col violation. 
llvm-svn: 47998
 2008-03-06 17:42:34 +00:00
Gabor Greif 636ab19205 some more spelling changes 
llvm-svn: 47996
 2008-03-06 10:51:21 +00:00
Gabor Greif ad19df0687 fix typos 
llvm-svn: 47994
 2008-03-06 10:36:00 +00:00
Evan Cheng a3cb090446 Constant fold SIGN_EXTEND_INREG with ashr not lshr. 
llvm-svn: 47992
 2008-03-06 08:20:51 +00:00
Nick Lewycky b2e3c31587 Exercise the new CFG change. 
llvm-svn: 47990
 2008-03-06 06:55:58 +00:00
Nick Lewycky 19b0df202e Treat BBs that use BBs as proper predecessors and successors in the CFG. 
llvm-svn: 47989
 2008-03-06 06:54:53 +00:00
Nick Lewycky 3e2d7c9f85 Commit the testcase too. 
llvm-svn: 47988
 2008-03-06 06:50:03 +00:00
Nick Lewycky d0b62a1552 Don't try to simplify urem and srem using arithmetic rules that don't work 
under modulo (overflow). Fixes PR1933.

llvm-svn: 47987
 2008-03-06 06:48:30 +00:00
Scott Michel e9b690b7f1 Refine Cell's i64 constant generation code to cover more constants where the 
upper and lower 32-bits are the same (in addition to 0 and -1 previously.)

llvm-svn: 47985
 2008-03-06 04:02:54 +00:00
Andrew Lenharth dcabdb1b38 gcc likes things spelled correctly 
llvm-svn: 47981
 2008-03-05 23:41:37 +00:00
Devang Patel 26cabc447c Skip, for now, callsites where use of sret argument is not dominated by callsite. 
llvm-svn: 47980
 2008-03-05 23:39:23 +00:00
Dale Johannesen 0d235052d1 Next bit of PPC ByVal handling; call-site code seems 
correct now.

llvm-svn: 47978
 2008-03-05 23:31:27 +00:00
Anton Korobeynikov 2609888f29 Missed patch from my last commit 
llvm-svn: 47977
 2008-03-05 23:21:39 +00:00
Anton Korobeynikov e43711d392 Resolve aliases to aliasees, where possible 
llvm-svn: 47975
 2008-03-05 23:08:47 +00:00
Anton Korobeynikov 24f65c3ab7 Handle functions as targets during linking of aliases as well 
llvm-svn: 47974
 2008-03-05 23:08:16 +00:00
Scott Michel 42f56b4838 - Fix support for "special" i64 immediates that can be loaded 
using IL, ILA, et. al. v2i64 and i64 are now supported by
  the select bits (SELB) instruction.

- Add missing comparison operations (testcase forthcoming)

- More multiclass refactoring.

llvm-svn: 47973
 2008-03-05 23:02:02 +00:00
Scott Michel 48072bf179 - Expand tabs to spaces. 
- select_bits.ll now fully functional now that PR1993 is closed. It was
  previously broken by refactoring in SPUInstrInfo.td and using multiclasses.
- Same for eqv.ll

llvm-svn: 47972
 2008-03-05 23:00:19 +00:00
Steve Naroff 1c703ef97f Update vcproj file. 
llvm-svn: 47969
 2008-03-05 22:30:39 +00:00
Anton Korobeynikov dac5fa94ef Try hard to link aliases. Checks can be too strict by now. 
llvm-svn: 47968
 2008-03-05 22:22:46 +00:00
Evan Cheng 29b502e0e0 Fix a coalescer bug wrt how dead copy interval is shortened. 
llvm-svn: 47966
 2008-03-05 22:09:42 +00:00
Devang Patel 3b1c95f885 Handle 'ret' with multiple values. 
llvm-svn: 47965
 2008-03-05 21:50:24 +00:00
Andrew Lenharth 7fbece5d8a test for something more interesting than not crashing 
llvm-svn: 47962
 2008-03-05 20:24:26 +00:00
Andrew Lenharth 04bd14db4d sync ops on ptrs, was breaking libgomp 
llvm-svn: 47960
 2008-03-05 19:48:27 +00:00
Daniel Berlin c3d98074a9 Add Hybrid Cycle Detection to Andersen's analysis. 
Patch by Curtis Dunham.

llvm-svn: 47959
 2008-03-05 19:31:47 +00:00
Dale Johannesen 8ee39c61f2 Clarify that CALLSEQ_START..END may not be nested, 
and add some protection against creating such.

llvm-svn: 47957
 2008-03-05 19:14:03 +00:00
Scott Michel 944207423b This patch fixes a problem encountered by the CellSPU backend where variants 
were being pruned in patterns where a variable was used more than once, e.g.:

  (or (and R32C:$rA, R32C:$rC), (and R32C:$rB, (not R32C:$rC)))

In this example, $rC is used more than once and is actually significant to
instruction selection pattern matching when commuted variants are produced.
This patch scans the pattern's clauses and collects the variables, creating
a set of variables that are used more than once. TreePatternNode::isIsomorphicTo()
also understands that multiply-used variables are significant.

llvm-svn: 47950
 2008-03-05 17:49:05 +00:00
Chris Lattner 7c08a01698 evan implemented this. 
llvm-svn: 47948
 2008-03-05 17:11:51 +00:00
Anton Korobeynikov 6be42d2e23 Fix test not to emit junk into source directory 
llvm-svn: 47947
 2008-03-05 15:44:25 +00:00
Anton Korobeynikov f8e03a9b3b Testcase for PR2054 
llvm-svn: 47946
 2008-03-05 15:43:58 +00:00
Anton Korobeynikov 2e13597f9d Remember the source->dest mapping when copying aliases. This fixes PR2054 
llvm-svn: 47945
 2008-03-05 15:27:21 +00:00
Anton Korobeynikov 12cee7195f Clarify the state-of-the-art 
llvm-svn: 47944
 2008-03-05 15:11:00 +00:00
Bill Wendling 5ddd03d88b Use AC_PATH_PROG correctly: 
http://www.gnu.org/software/autoconf/manual/autoconf.html#Generic-Programs

llvm-svn: 47943
 2008-03-05 09:28:02 +00:00
Evan Cheng 3ea44e4ee9 isTwoAddress = 1 -> Constraints. 
llvm-svn: 47941
 2008-03-05 08:19:16 +00:00
Evan Cheng 6ec7dc6bea PSLLWri etc. are two-address instructions. 
llvm-svn: 47940
 2008-03-05 08:11:27 +00:00
Chris Lattner 2acd0c25f6 add a note 
llvm-svn: 47939
 2008-03-05 07:22:39 +00:00
Chris Lattner 78e9cab229 Generalize FP constant shrinking optimization to apply to any vt 
except ppc long double.  This allows us to shrink constant pool
entries for x86 long double constants, which in turn allows us to
use flds/fldl instead of fldt.

llvm-svn: 47938
 2008-03-05 06:48:13 +00:00
Chris Lattner 3dc3899007 Improve comment, pass in the original VT so that we can shrink a long double constant 
all the way to float, not stopping at double.

llvm-svn: 47937
 2008-03-05 06:46:58 +00:00
Evan Cheng 3bd59641ac Ignore debugging related instructions if they get this far. 
llvm-svn: 47934
 2008-03-05 02:34:36 +00:00
Evan Cheng 801bfb2cf7 Rather than asserting. Dump out the MI that we are not able to encode and abort. 
llvm-svn: 47933
 2008-03-05 02:08:03 +00:00
Dan Gohman da7897c4e1 Codegen support for i128 UINT_TO_FP. This just fixes a 
bug in r47928 (Int64Ty is the correct type for the constant
pool entry here) and removes the asserts, now that the code
is capable of handling i128.

llvm-svn: 47932
 2008-03-05 02:07:31 +00:00
Evan Cheng 0a62cb44ce Add a target lowering hook to control whether it's worthwhile to compress fp constant. 
For x86, if sse2 is available, it's not a good idea since cvtss2sd is slower than a movsd load and it prevents load folding. On x87, it's important to shrink fp constant since fldt is very expensive.

llvm-svn: 47931
 2008-03-05 01:30:59 +00:00
Andrew Lenharth 357061a74d 64bit CAS on 32bit x86. 
llvm-svn: 47929
 2008-03-05 01:15:49 +00:00
Dan Gohman d9d874b0cd Codegen support for i128 SINT_TO_FP. 
llvm-svn: 47928
 2008-03-05 01:08:17 +00:00
Evan Cheng 6325446666 Refactor code. Remove duplicated functions that basically do the same thing as 
findRegisterUseOperandIdx, findRegisterDefOperandIndx. Fix some naming inconsistencies.

llvm-svn: 47927
 2008-03-05 00:59:57 +00:00
Devang Patel e516aa1127 Skip functions that return multiple values. 
llvm-svn: 47924
 2008-03-05 00:36:59 +00:00
Devang Patel b3ca38c1d5 Check struct return type first. 
llvm-svn: 47922
 2008-03-05 00:27:05 +00:00
Bill Wendling 3fcd946c26 Removed spurious EnablePPCRS check. 
llvm-svn: 47918
 2008-03-04 23:27:33 +00:00
Dale Johannesen 85d41a1c44 Move PPC lowering functions into PPCTargetLowering 
class (cosmetic).  First piece of byval implementation;
this doesn't work yet.  No functional change.

llvm-svn: 47917
 2008-03-04 23:17:14 +00:00
Bill Wendling 09de8e494e Use -enable-ppc-regscavenger flag for these checks 
llvm-svn: 47916
 2008-03-04 23:16:26 +00:00
Bill Wendling 1af20ad336 Use a command-line option to turn register scavenging on/off for PPC. 
llvm-svn: 47915
 2008-03-04 23:13:51 +00:00
Devang Patel ce556d908b Print types for all ret operands. 
llvm-svn: 47911
 2008-03-04 22:05:14 +00:00
Devang Patel 4566d885dd Use while loop. 
llvm-svn: 47909
 2008-03-04 21:59:49 +00:00
Devang Patel b2fae1d13f Revert SmallVector ctor variants. 
llvm-svn: 47908
 2008-03-04 21:54:56 +00:00
Devang Patel 941ab37ea8 Use cast instead of dyn_cast. 
Update test to use multiple return value directly, instead of relying on -sretpromotion.

llvm-svn: 47907
 2008-03-04 21:45:28 +00:00
Devang Patel ea4cb89d34 Filter nested structs 
llvm-svn: 47906
 2008-03-04 21:32:09 +00:00
Devang Patel 841322b32a Handle multiple return values. 
llvm-svn: 47904
 2008-03-04 21:15:15 +00:00
Andrew Lenharth 4fee9f35b5 x86-64 atomics 
llvm-svn: 47903
 2008-03-04 21:13:33 +00:00
Anton Korobeynikov 28179f764c Properly populate lists of defined/undefined symbols in presence of aliases 
llvm-svn: 47900
 2008-03-04 20:16:11 +00:00
Anton Korobeynikov d72ade3b3b Add alias with non-external aliasee to the list of exported symbols of the library. 
llvm-svn: 47899
 2008-03-04 20:15:35 +00:00
Devang Patel 297ea0de75 Use appropriate index to get the result value. 
llvm-svn: 47897
 2008-03-04 19:22:54 +00:00
Devang Patel 748c77f530 Skip sret attribute while preparing attribute list for 
new function and new call sites.

llvm-svn: 47896
 2008-03-04 19:12:58 +00:00
Devang Patel 5dc63480db Add FunctionType ctor variant that takes SmallVector params. 
llvm-svn: 47895
 2008-03-04 18:57:05 +00:00
Devang Patel ed6e4b0b7d Increment counter that keeps track of total number of sret promoted. 
llvm-svn: 47892
 2008-03-04 17:48:11 +00:00
Devang Patel 6bb36e686d Skip declarations. 
llvm-svn: 47890
 2008-03-04 17:47:06 +00:00
Devang Patel 5b69f3fa33 Process externally visible functions also. Later on code generator will do the right thing. 
llvm-svn: 47889
 2008-03-04 17:46:26 +00:00
Devang Patel ecce5af83d Collect statistics. 
llvm-svn: 47888
 2008-03-04 17:44:37 +00:00
Gordon Henriksen 099278e8f5 Fix a typo noticed by Erick Tryzelaar, 
llvm-svn: 47886
 2008-03-04 14:52:05 +00:00
Roman Levenstein c62c2bb4d0 Some improvements related to the computation of heights, depths of SUnits. 
The basic idea is that all these algorithms are computing the longest paths from the root node or to the exit node. Therefore the existing implementation that uses and iterative and potentially
exponential algorithm was changed to a well-known graph algorithm based on dynamic programming. It has a linear run-time.

llvm-svn: 47884
 2008-03-04 11:19:43 +00:00
Evan Cheng 38caf77419 Refactor ExpandConstantFP so it can optimize load from constpool of types larger than f64 into extload from smaller types. 
llvm-svn: 47883
 2008-03-04 08:05:30 +00:00
Evan Cheng 16e2cf6db1 Really fix the test. 
llvm-svn: 47882
 2008-03-04 08:01:56 +00:00
Evan Cheng 7473c74d52 Fix broken test. 
llvm-svn: 47881
 2008-03-04 07:59:13 +00:00
Evan Cheng 59d58ab8c4 80 column violations. 
llvm-svn: 47878
 2008-03-04 03:20:06 +00:00
Evan Cheng b9bd5a879d Fix 80 column violations. 
llvm-svn: 47877
 2008-03-04 03:19:19 +00:00
Bill Wendling 2ae707888b Did I say 'e = getNumOperands()'? I meant --e, of course. 
llvm-svn: 47875
 2008-03-04 00:48:15 +00:00
Evan Cheng 62240d65fd Add PR1501 test case. 
llvm-svn: 47874
 2008-03-04 00:47:45 +00:00
Evan Cheng 567d2e5b57 Rename isOperand() to isOperandOf() (and other similar methods). It always confuses me. 
llvm-svn: 47872
 2008-03-04 00:41:45 +00:00
Evan Cheng 33ff36321e Remove -always-fold-and-in-test. 
llvm-svn: 47871
 2008-03-04 00:40:35 +00:00
Bill Wendling 0e541ea730 Miscellaneous clean-ups based on Evan's feedback: 
- Cleaned up how the prologue-epilogue inserter loops over the instructions.
- Instead of restarting the processing of an instruction if we remove an
  implicit kill, just update the end iterator and make sure that the iterator
  isn't incremented.

llvm-svn: 47870
 2008-03-03 23:57:28 +00:00
Dan Gohman e1c4f99549 Misc. APInt-ification in the DAGCombiner. 
llvm-svn: 47869
 2008-03-03 23:51:38 +00:00
Dan Gohman 10f34077f1 More APInt-ification. 
llvm-svn: 47868
 2008-03-03 23:35:36 +00:00
Dan Gohman 0e238dc813 Yet more APInt-ification. 
llvm-svn: 47867
 2008-03-03 22:37:52 +00:00
Dan Gohman 2fa65b7997 More APInt-ification. 
llvm-svn: 47866
 2008-03-03 22:22:56 +00:00
Dan Gohman a986eea82f Add support for lowering i64 SRA_PARTS and friends on x86-64. 
llvm-svn: 47865
 2008-03-03 22:22:09 +00:00
Dan Gohman f2bbfa3ba0 More APInt-ification. 
llvm-svn: 47864
 2008-03-03 22:20:46 +00:00
Bill Wendling 632ea65072 This is the initial check-in for adding register scavenging to PPC. (Currently, 
PPC-64 doesn't work.) This also lowers the spilling of the CR registers so that
it uses a register other than the default R0 register (the scavenger scrounges
for one). A significant part of this patch fixes how kill information is
handled.

llvm-svn: 47863
 2008-03-03 22:19:16 +00:00
Bill Wendling 7921ad0d67 Go through the machine instruction's operands to make sure that we're not 
marking both a super- and sub-register as "killed". This removes implicit uses
that are marked as "killed".

llvm-svn: 47862
 2008-03-03 22:14:33 +00:00
Bill Wendling 528083bc28 Make the register scavenger update the bookkeeping values for sub/super 
registers.

llvm-svn: 47861
 2008-03-03 22:12:25 +00:00
Bill Wendling 4836d58f89 Multiple instructions can be inserted when eliminating frame indexes. We need 
the register scavenger to process all of those new instructions instead of just
the last one inserted.

llvm-svn: 47860
 2008-03-03 22:11:16 +00:00
Chris Lattner be750abfd4 new testcase 
llvm-svn: 47858
 2008-03-03 21:59:00 +00:00
Devang Patel 9d91785987 s/isReturnStruct()/hasStructRetAttr()/g 
llvm-svn: 47857
 2008-03-03 21:46:28 +00:00
Devang Patel 83558f9f46 regenerate. 
llvm-svn: 47849
 2008-03-03 18:58:47 +00:00
Devang Patel c6b56c79aa If a function uses multive values in ret instruction then it 
is returning a struct value.

llvm-svn: 47848
 2008-03-03 18:58:16 +00:00
Devang Patel b0faf89133 Preserve paramater attributes. 
llvm-svn: 47847
 2008-03-03 18:36:03 +00:00
Tanya Lattner 46be50c668 Renable lto2 build. 
llvm-svn: 47845
 2008-03-03 17:32:40 +00:00
Chris Lattner e209be4985 Stub out a Path::GetMainExecutable call to find the path to the 
main executable of a program.  This needs to be implemented on windows.

llvm-svn: 47835
 2008-03-03 02:55:43 +00:00
Chris Lattner d51372aa96 another random note 
llvm-svn: 47831
 2008-03-02 19:29:42 +00:00
Chris Lattner 6b0a189225 add a note 
llvm-svn: 47830
 2008-03-02 19:27:34 +00:00
Chris Lattner a70df9e2ee Evan implemented these. 
llvm-svn: 47828
 2008-03-02 18:05:14 +00:00
Chris Lattner bd0bb3f07f Evan implemented this. 
llvm-svn: 47827
 2008-03-02 17:56:29 +00:00
Chris Lattner a8b5ed8d3a add a testcase for misc vector stuff 
llvm-svn: 47826
 2008-03-02 08:57:59 +00:00
Chris Lattner c93b48caeb add support for lvsl, fixing PR1481 
llvm-svn: 47825
 2008-03-02 08:54:27 +00:00
Chris Lattner 8fb4c98284 add support for the sse.cmp.* intrinsics, which fixes sse.isamax with the CBE. 
llvm-svn: 47824
 2008-03-02 08:47:13 +00:00
Chris Lattner 9044719798 refactor intrinsic handling code out into its own method. 
llvm-svn: 47823
 2008-03-02 08:29:41 +00:00
Chris Lattner 2541f1dac6 fix printing of undef vectors, this fixes "simple" and "build" in UnitTests/Vector. 
Now they all pass.

llvm-svn: 47820
 2008-03-02 08:14:45 +00:00
Chris Lattner 4c058b39e4 insertelement got the wrong operands. 
llvm-svn: 47819
 2008-03-02 08:10:16 +00:00
Chris Lattner 051709436e Several changes: 
* Simplify handling of byval, making it easier to understand and more 
    consistent.  This fixes PR2065.
  * Clean up and simplify handling of GEPs.  I can actually understand it now!
  * Implement support for GEP'ing into vectors, this fixes
    SingleSource/UnitTests/Vector/build2 among others.

llvm-svn: 47818
 2008-03-02 08:07:24 +00:00
Nick Kledzik e8338e52ed suppress extra -install_name on Tiger. The 10.4 tools don't support it. Fix whitespace. 
llvm-svn: 47817
 2008-03-02 06:33:06 +00:00
Chris Lattner 9140fba6ad Fix a bug I introduced in constant array and constant vector handling. 
llvm-svn: 47816
 2008-03-02 05:46:57 +00:00
Chris Lattner 7295bc15a0 implement shufflevector. 
llvm-svn: 47815
 2008-03-02 05:41:07 +00:00
Chris Lattner bf6acfb71d make operand accessors const-correct 
llvm-svn: 47814
 2008-03-02 05:32:05 +00:00
Chris Lattner f724e34ddc Add a new ShuffleVectorInst::getMaskValue method. 
llvm-svn: 47813
 2008-03-02 05:28:33 +00:00
Chris Lattner 689d830b6d implement extractelement. 
llvm-svn: 47812
 2008-03-02 03:57:08 +00:00
Chris Lattner d9cc003b0d implement insertelement. 
llvm-svn: 47811
 2008-03-02 03:52:39 +00:00
Chris Lattner 92d929c21d respect isSigned for vector types, fixing sdiv of vectors etc. 
llvm-svn: 47810
 2008-03-02 03:41:23 +00:00
Chris Lattner fe1083be61 print the attribute in the right place, this fixes function returning vectors. 
llvm-svn: 47809
 2008-03-02 03:39:43 +00:00
Nick Lewycky 8bbec39262 Oops, can't unwind to the entry block (entry block may have no preds). 
llvm-svn: 47808
 2008-03-02 03:34:55 +00:00
Chris Lattner d861066d15 vector types are simple types. This fixes div/rem of vectors. 
llvm-svn: 47807
 2008-03-02 03:33:31 +00:00
Chris Lattner b104be599a Print vector types appropriately. This gets basic vector code working 
(PR1126)

llvm-svn: 47806
 2008-03-02 03:29:50 +00:00
Chris Lattner c8d974a323 rename PT -> VT for VectorTypes. 
llvm-svn: 47805
 2008-03-02 03:18:46 +00:00
Chris Lattner 895aef4720 Print i32/i64 integer constants as 1u instead of ((unsigned int)1). 
Use dyn_cast better.

llvm-svn: 47804
 2008-03-02 03:16:38 +00:00
Chris Lattner 647c6642e7 Move pr717 to here. 
llvm-svn: 47803
 2008-03-02 02:51:40 +00:00
Nick Lewycky 3cc9be0b59 Add an unwind_to field to basic blocks, making them Users instead of Values. 
This is the first checkin for PR1269, the new EH infrastructure.

llvm-svn: 47802
 2008-03-02 02:48:09 +00:00
Andrew Lenharth 20bcdba9ca good catch anton 
llvm-svn: 47800
 2008-03-01 23:18:21 +00:00
Andrew Lenharth f5c90ec12c make CAS work 
llvm-svn: 47799
 2008-03-01 22:27:48 +00:00
Andrew Lenharth d032c33300 all but CAS working on x86 
llvm-svn: 47798
 2008-03-01 21:52:34 +00:00
Nick Lewycky 36c00f834a Also handle "getresult". 
llvm-svn: 47797
 2008-03-01 21:47:06 +00:00
Nick Lewycky 7f6e95a973 Print the name, not a pointer. 
llvm-svn: 47796
 2008-03-01 17:20:55 +00:00
Andrew Lenharth 0070dd1de3 Add lock prefix support to x86. Also add the instructions necessary for the atomic ops. They are still marked pseudo, since I cannot figure out what format to use, but they are the correct opcode. 
llvm-svn: 47795
 2008-03-01 13:37:02 +00:00
Tanya Lattner 5640bd186a Remove llvm-upgrade and update test cases. 
llvm-svn: 47793
 2008-03-01 09:15:35 +00:00
Chris Lattner ffe0da0eb2 Fix PR2113 by verifying allocations. 
llvm-svn: 47792
 2008-03-01 09:01:57 +00:00
Chris Lattner 903ee68639 Fix this test. 
llvm-svn: 47791
 2008-03-01 09:00:21 +00:00
Chris Lattner a683edb2d8 allow specified inline threshold to be negative, as the value is 
itself sometimes negative.

llvm-svn: 47786
 2008-03-01 08:09:51 +00:00
Tanya Lattner b7e02fc0bf Remove llvm-upgrade and update tests. 
llvm-svn: 47784
 2008-03-01 07:38:40 +00:00
Tanya Lattner bb20900184 Disable lto because its broken on tiger. 
llvm-svn: 47783
 2008-03-01 07:36:30 +00:00
Dale Johannesen 208cc8f1b9 Add MVT::is128BitVector and is64BitVector. Shrink 
unaligned load/store code using them.  Per review
of unaligned load/store vector patch.

llvm-svn: 47782
 2008-03-01 03:40:57 +00:00
Evan Cheng 73bdf043a1 Refactor / clean up code; remove td list scheduler special tie breaker (no real benefit). 
llvm-svn: 47779
 2008-03-01 00:39:47 +00:00
Devang Patel e8ddaba169 Fix cut-n-pasto. 
llvm-svn: 47777
 2008-02-29 23:41:13 +00:00
Devang Patel 4eedfd2421 Add pass to promote sret. 
This pass transforms 

  %struct._Point = type { i32, i32, i32, i32, i32, i32 }
  define internal void @foo(%struct._Point* sret  %agg.result)

into

  %struct._Point = type { i32, i32, i32, i32, i32, i32 }
  define internal %struct._Point @foo()

This pass updates foo() clients appropriately to use
getresult instruction to extract return values.

This pass is not yet ready for prime time.

llvm-svn: 47776
 2008-02-29 23:34:08 +00:00
Anton Korobeynikov 0e8b146152 Use enumeration for preffered EH dwarf encoding reason 
llvm-svn: 47770
 2008-02-29 22:09:08 +00:00
Evan Cheng 26edb59d97 Don't fill eh frames even though these are text sections. 
llvm-svn: 47765
 2008-02-29 19:36:59 +00:00
Nick Kledzik 6555623ceb update to document new lto API 
llvm-svn: 47764
 2008-02-29 19:34:52 +00:00
Nick Kledzik 2204568281 install new lto dylib. pass down LLVM_SUBMIT_VERSION 
llvm-svn: 47763
 2008-02-29 19:32:13 +00:00
Nick Kledzik d5a2f4aee5 stop building lto on all platforms. Start building lto2 on Darwin 
llvm-svn: 47762
 2008-02-29 19:31:29 +00:00
Nick Kledzik 618ba73288 fix default RC_ProjectSourceSubversion to be zero 
llvm-svn: 47761
 2008-02-29 19:30:23 +00:00
Bill Wendling 16f1e0be3d DCE'ed this testcase. 
llvm-svn: 47760
 2008-02-29 19:28:11 +00:00
Bill Wendling 811153a551 If we reload a virtual register that's already been assigned, we want to mark 
that instruction as its "last use". This fixes PR1925.

llvm-svn: 47758
 2008-02-29 18:52:01 +00:00
Evan Cheng 2e26dc8051 Fix PR2112: don't run loop aligner if target doesn't have a TargetLowering object. 
llvm-svn: 47755
 2008-02-29 17:52:15 +00:00
Chris Lattner c966cebe93 fix a bug Anders ran into where scalarrepl would crash when promoting 
a union containing a vector and an array whose elements were smaller than
the vector elements.  this means we need to compile the load of the 
array elements into an extract element plus a truncate.

llvm-svn: 47752
 2008-02-29 07:12:06 +00:00
Chris Lattner 77205def10 Refactor some code out of ConvertUsesToScalar into their own methods, no 
functionality change.

llvm-svn: 47751
 2008-02-29 07:03:13 +00:00
Chris Lattner c612571555 Folding or(fcmp,fcmp) only works if the operands of the fcmps are the same fp type. 
llvm-svn: 47750
 2008-02-29 06:09:11 +00:00
Evan Cheng ca7c61e79a No need for coalescer to update kills. Only copies are coalesced and those instructions will be deleted. Doh. 
llvm-svn: 47749
 2008-02-29 02:50:03 +00:00
Evan Cheng 88f839944d Remove redundant #include. 
llvm-svn: 47748
 2008-02-29 02:49:15 +00:00
Nick Kledzik a28acfee92 also pass -compatibility_version when linking shared object for Darwin 
llvm-svn: 47747
 2008-02-29 02:01:34 +00:00
Dan Gohman bd2fa566e4 More APInt-ification. 
llvm-svn: 47746
 2008-02-29 01:47:35 +00:00
Dan Gohman 837a6dccd7 Use the new convertFromAPInt instead of convertFromZeroExtendedInteger, 
which allows more of the surrounding arithmetic to be done with APInt
instead of uint64_t.

llvm-svn: 47745
 2008-02-29 01:44:25 +00:00
Dan Gohman 06c45d516d Use the new convertFromAPInt instead of convertFromZeroExtendedInteger. 
llvm-svn: 47744
 2008-02-29 01:42:52 +00:00
Dan Gohman ec6be4a782 Use the new APInt-enabled form of getConstant instead of converting 
an APInt into a uint64_t to call getConstant.

llvm-svn: 47742
 2008-02-29 01:41:59 +00:00
Dan Gohman 105c1d4106 Add support to APInt for shift and rotate operations with APInt 
instead of uint32_t for the shift/rotate count operand type.

llvm-svn: 47741
 2008-02-29 01:40:47 +00:00
Anders Carlsson 17df4cd397 Use the correct instruction encodings for the 64-bit MMX movd. 
llvm-svn: 47740
 2008-02-29 01:35:12 +00:00
Dan Gohman ca24fd9096 Simplify code using convertFromZeroExtendedInteger with an APInt 
by using the new convertFromAPInt directly.

llvm-svn: 47739
 2008-02-29 01:27:13 +00:00
Dan Gohman 35723eb4f6 Add a method to APFloat to convert directly from APInt. 
llvm-svn: 47738
 2008-02-29 01:26:11 +00:00
Evan Cheng 95a7be473c Added option -align-loops=<true/false> to disable loop aligner pass. 
llvm-svn: 47736
 2008-02-28 23:29:57 +00:00
Lauro Ramos Venancio bb979f4b1c Update testcase. 
llvm-svn: 47735
 2008-02-28 23:13:15 +00:00
Dale Johannesen cbde4c2206 Interface of getByValTypeAlignment differed between 
generic & x86 versions; change generic to follow x86
and improve comments.  Add PPC version (not right
for non-Darwin.)

llvm-svn: 47734
 2008-02-28 22:31:51 +00:00
Andrew Lenharth 09b64a4ed1 Make llvm-ar behave like ar, if you create an empty archive, ar creates an empty archive. llvm-ar would not generate an output file in this case 
llvm-svn: 47733
 2008-02-28 22:24:48 +00:00
Lauro Ramos Venancio 714eeb2d55 Initialize TextAlignFillValue. Fix revision 47703. 
llvm-svn: 47732
 2008-02-28 22:14:09 +00:00
Lauro Ramos Venancio 9781aee03c Avoid collisions in value names. 
llvm-svn: 47725
 2008-02-28 20:26:04 +00:00
Gabor Greif a6e293a389 Run 'svn info' in the C locale. Also appropriately escape spaces and ? for gawk in darwin. 
llvm-svn: 47723
 2008-02-28 18:46:56 +00:00
Dale Johannesen c4c3de2b52 Fix an assertion message. 
llvm-svn: 47722
 2008-02-28 18:36:51 +00:00
Chris Lattner 643dce1146 add a note 
llvm-svn: 47720
 2008-02-28 17:21:27 +00:00
Duncan Sands e28886a913 Document that the shuffle mask may contain undef 
values.

llvm-svn: 47719
 2008-02-28 17:12:11 +00:00
Gabor Greif 376d2cedb7 Simplify and clean up "update" target. Also 
add an SVN-UPDATE-OPTIONS variable that controls
the update behaviour. This allows to go to a common
older revision of all svn directories:

  gmake update SVN-UPDATE-OPTIONS=-r47717

would rewind both llvm and clang to a common
revision (if clang is checked out into the llvm
tree).

llvm-svn: 47717
 2008-02-28 14:58:14 +00:00
Gabor Greif d71b410cb8 Now that targets are serialized, introduce 
two new convenience targets:

- update: svn update toplevel and try hard
          to locate updatable subdirectories
					using cunning tricks
- happiness: update then build and test

so what one wants to do now is:
  nice gmake --jobs happiness

Have fun!

llvm-svn: 47716
 2008-02-28 13:06:50 +00:00
Gabor Greif 76e7083f49 Treat all targets serially at the toplevel. This allows 
to specify
  nice gmake --jobs all check
and go to lunch, while a multiprocessor machine
will build everything using spare resources
and check the result thereafter.
Since concurrency of make is not restricted
in subdirectories, this should be a nearly
optimal way to do it.

Also teach the user about a configure switch.

llvm-svn: 47715
 2008-02-28 11:48:14 +00:00
Gabor Greif f77e6977a0 Fix http://llvm.org/bugs/show_bug.cgi?id=2104 by ordering lexicographically what gets printed. Be const-correct in PrintResults and uninline it too 
llvm-svn: 47712
 2008-02-28 08:38:45 +00:00
Evan Cheng 507713de08 Set to default: x86 no longer fold and into test if it has more than one use. 
llvm-svn: 47711
 2008-02-28 07:46:38 +00:00
Evan Cheng a465bfb87c Keep track how many commutes are performed by the scheduler. 
llvm-svn: 47710
 2008-02-28 07:40:24 +00:00
Chris Lattner e2f55f9fc5 Sparc backend doesn't support debug info yet, mark the nodes as expand. This fixes a crash on 
test/DebugInfo/funccall.ll

llvm-svn: 47709
 2008-02-28 05:54:25 +00:00
Chris Lattner 9824ffef0c implement expand for ISD::DECLARE by just deleting it. 
llvm-svn: 47708
 2008-02-28 05:53:40 +00:00
Chris Lattner f1444c35e8 Fix CodeGen/Generic/storetrunc-fp.ll on sparc, PR2105 
llvm-svn: 47707
 2008-02-28 05:48:04 +00:00
Chris Lattner beb6a5ba4f fix CodeGen/Generic/2008-01-25-dag-combine-mul.ll on sparc, PR2105 
llvm-svn: 47706
 2008-02-28 05:44:20 +00:00
Chris Lattner 765be8874d target-indep codegen memcpy lowering issue. 
llvm-svn: 47705
 2008-02-28 05:34:27 +00:00
Chris Lattner 83e80cd368 Add a random not very important note 
llvm-svn: 47704
 2008-02-28 04:52:59 +00:00
Evan Cheng c799065cc3 Add a quick and dirty "loop aligner pass". x86 uses it to align its loops to 16-byte boundaries. 
llvm-svn: 47703
 2008-02-28 00:43:03 +00:00
Eli Friedman 93e8b679a3 A few more small things I've run into. 
llvm-svn: 47702
 2008-02-28 00:21:43 +00:00
Anton Korobeynikov d73396bef9 EHPreferredDataFormat hook for PPC targets. Looks like Darwin 
uses the same encoding everywhere. Linux FIXME'ed.

llvm-svn: 47701
 2008-02-27 23:49:15 +00:00
Devang Patel a1a6882519 Remove dead code. 
llvm-svn: 47700
 2008-02-27 23:33:51 +00:00
Anton Korobeynikov ae24cca0e4 Preparation step for some cleanup/generalization in EH information emission: 
provide TAI hook for selection of EH data emission format. Currently unused.

llvm-svn: 47699
 2008-02-27 23:33:50 +00:00
Dale Johannesen 820de906b1 New test for misaligned Altivec laod/store. 
llvm-svn: 47698
 2008-02-27 23:14:50 +00:00
Anton Korobeynikov b3666760ce Assert on invalid PHINode::addIncoming() arguments. Patch by Erick Tryzelaar! 
llvm-svn: 47697
 2008-02-27 22:37:28 +00:00
Dale Johannesen bf76a08e7c Handle load/store of misaligned vectors that are the 
same size as an int type by doing a bitconvert of
load/store of the int type (same algorithm as floating point).
This makes them work for ppc Altivec.  There was some
code that purported to handle loads of (some) vectors
by splitting them into two smaller vectors, but getExtLoad
rejects subvector loads, so this could never have worked;
the patch removes it.

llvm-svn: 47696
 2008-02-27 22:36:00 +00:00
Nick Kledzik 91a6dcff32 fixes from review of first commit 
llvm-svn: 47695
 2008-02-27 22:25:36 +00:00
Nick Kledzik 5f1db0a8de use PROJ_SRC_DIR so this builds with Apple style builds 
llvm-svn: 47694
 2008-02-27 22:23:46 +00:00
Dan Gohman 26854f242e Don't hard-code the mask size to be 32, which is incorrect on ppc64 
and was causing aborts with the new APInt changes. This may also be
fixing an obscure ppc64 bug.

llvm-svn: 47692
 2008-02-27 21:12:32 +00:00
Evan Cheng 3d17e4c427 This is done. 
llvm-svn: 47688
 2008-02-27 20:26:32 +00:00
Evan Cheng fdc732ab9a Fix a bug in dead spill slot elimination. 
llvm-svn: 47687
 2008-02-27 19:57:11 +00:00
Dan Gohman e5e32ec8f7 Remove the `else', at Evan's insistence. 
llvm-svn: 47686
 2008-02-27 19:44:57 +00:00
Dan Gohman 61377a3dc8 Add -analyze support to postdomtree. 
llvm-svn: 47680
 2008-02-27 18:38:29 +00:00
Chris Lattner 3df31ba41b actually run llc, thanks Dan :) 
llvm-svn: 47677
 2008-02-27 17:46:54 +00:00
Duncan Sands ef40c5b204 Add a FIXME about the VECTOR_SHUFFLE evil hack. 
llvm-svn: 47676
 2008-02-27 17:39:13 +00:00
Lauro Ramos Venancio 14241b29b2 Emit an error when a library is not found. It is the GNU ld behavior and it is expected by the configure scripts. 
llvm-svn: 47674
 2008-02-27 17:20:32 +00:00
Duncan Sands e158a82f26 LegalizeTypes support for EXTRACT_VECTOR_ELT. The 
approach taken is different to that in LegalizeDAG
when it is a question of expanding or promoting the
result type: for example, if extracting an i64 from
a <2 x i64>, when i64 needs expanding, it bitcasts
the vector to <4 x i32>, extracts the appropriate
two i32's, and uses those for the Lo and Hi parts.
Likewise, when extracting an i16 from a <4 x i16>,
and i16 needs promoting, it bitcasts the vector to
<2 x i32>, extracts the appropriate i32, twiddles
the bits if necessary, and uses that as the promoted
value.  This puts more pressure on bitcast legalization,
and I've added the appropriate cases.  They needed to
be added anyway since users can generate such bitcasts
too if they want to.  Also, when considering various
cases (Legal, Promote, Expand, Scalarize, Split) it is
a pain that expand can correspond to Expand, Scalarize
or Split, so I've changed the LegalizeTypes enum so it
lists those different cases - now Expand only means
splitting a scalar in two.
The code produced is the same as by LegalizeDAG for
all relevant testcases, except for
2007-10-31-extractelement-i64.ll, where the code seems
to have improved (see below; can an expert please tell
me if it is better or not).
Before < vs after >.

<       subl    $92, %esp
<       movaps  %xmm0, 64(%esp)
<       movaps  %xmm0, (%esp)
<       movl    4(%esp), %eax
<       movl    %eax, 28(%esp)
<       movl    (%esp), %eax
<       movl    %eax, 24(%esp)
<       movq    24(%esp), %mm0
<       movq    %mm0, 56(%esp)
---
>       subl    $44, %esp
>       movaps  %xmm0, 16(%esp)
>       pshufd  $1, %xmm0, %xmm1
>       movd    %xmm1, 4(%esp)
>       movd    %xmm0, (%esp)
>       movq    (%esp), %mm0
>       movq    %mm0, 8(%esp)

<       subl    $92, %esp
<       movaps  %xmm0, 64(%esp)
<       movaps  %xmm0, (%esp)
<       movl    12(%esp), %eax
<       movl    %eax, 28(%esp)
<       movl    8(%esp), %eax
<       movl    %eax, 24(%esp)
<       movq    24(%esp), %mm0
<       movq    %mm0, 56(%esp)
---
>       subl    $44, %esp
>       movaps  %xmm0, 16(%esp)
>       pshufd  $3, %xmm0, %xmm1
>       movd    %xmm1, 4(%esp)
>       movhlps %xmm0, %xmm0
>       movd    %xmm0, (%esp)
>       movq    (%esp), %mm0
>       movq    %mm0, 8(%esp)

<       subl    $92, %esp
<       movaps  %xmm0, 64(%esp)
---
>       subl    $44, %esp

<       movl    16(%esp), %eax
<       movl    %eax, 48(%esp)
<       movl    20(%esp), %eax
<       movl    %eax, 52(%esp)
<       movaps  %xmm0, (%esp)
<       movl    4(%esp), %eax
<       movl    %eax, 60(%esp)
<       movl    (%esp), %eax
<       movl    %eax, 56(%esp)
---
>       pshufd  $1, %xmm0, %xmm1
>       movd    %xmm1, 4(%esp)
>       movd    %xmm0, (%esp)
>       movd    %xmm1, 12(%esp)
>       movd    %xmm0, 8(%esp)

<       subl    $92, %esp
<       movaps  %xmm0, 64(%esp)
---
>       subl    $44, %esp

<       movl    24(%esp), %eax
<       movl    %eax, 48(%esp)
<       movl    28(%esp), %eax
<       movl    %eax, 52(%esp)
<       movaps  %xmm0, (%esp)
<       movl    12(%esp), %eax
<       movl    %eax, 60(%esp)
<       movl    8(%esp), %eax
<       movl    %eax, 56(%esp)
---
>       pshufd  $3, %xmm0, %xmm1
>       movd    %xmm1, 4(%esp)
>       movhlps %xmm0, %xmm0
>       movd    %xmm0, (%esp)
>       movd    %xmm1, 12(%esp)
>       movd    %xmm0, 8(%esp)

llvm-svn: 47672
 2008-02-27 13:34:40 +00:00
Gabor Greif c848f9ea4e Remove inappropriate whitespace, change CVS to SVN where it makes sense. 
llvm-svn: 47671
 2008-02-27 13:34:15 +00:00
Duncan Sands 2111bd2e37 LegalizeTypes support for legalizing the mask 
operand of a VECTOR_SHUFFLE.  The mask is a
vector of constant integers.  The code in
LegalizeDAG doesn't bother to legalize the
mask, since it's basically just storage for
a bunch of constants, however LegalizeTypes
is more picky.  The problem is that there may
not exist any legal vector-of-integers type
with a legal element type, so it is impossible
to create a legal mask!  Unless of course you
cheat by creating a BUILD_VECTOR where the
operands have a different type to the element
type of the vector being built...  This is
pretty ugly but works - all relevant tests in
the testsuite pass, and produce the same
assembler with and without LegalizeTypes.

llvm-svn: 47670
 2008-02-27 13:03:44 +00:00
Duncan Sands 5d5bc484d0 LegalizeTypes support for INSERT_VECTOR_ELT. 
llvm-svn: 47669
 2008-02-27 10:18:23 +00:00
Evan Cheng 8ae8e2d50b Don't track max alignment during stack object allocations since they can be deleted later. Let PEI compute it. 
llvm-svn: 47668
 2008-02-27 10:04:56 +00:00
Duncan Sands 96658d0189 Support for legalizing MEMBARRIER. 
llvm-svn: 47667
 2008-02-27 08:53:44 +00:00
Bill Wendling 97925ec704 Final de-tabification. 
llvm-svn: 47663
 2008-02-27 06:33:05 +00:00
Chris Lattner 6fca938971 Add path separator support, patch by Sam Bishop. 
llvm-svn: 47662
 2008-02-27 06:17:10 +00:00
Chris Lattner 83263b8cfb Make X86TargetLowering::LowerSINT_TO_FP return without creating a dead 
stack slot and store if the  SINT_TO_FP is actually legal.  This allows
us to compile:

double a(double b) {return (unsigned)b;}

to:

_a:
	cvttsd2siq	%xmm0, %rax
	movl	%eax, %eax
	cvtsi2sdq	%rax, %xmm0
	ret

instead of:

_a:
	subq	$8, %rsp
	cvttsd2siq	%xmm0, %rax
	movl	%eax, %eax
	cvtsi2sdq	%rax, %xmm0
	addq	$8, %rsp
	ret

crazy.

llvm-svn: 47660
 2008-02-27 05:57:41 +00:00
Chris Lattner 5fe95a04f5 this code is correct but strange looking ;-) 
llvm-svn: 47659
 2008-02-27 05:48:44 +00:00
Chris Lattner 3c7d3d5700 Compile x86-64-and-mask.ll into: 
_test:
	movl	%edi, %eax
	ret

instead of:

_test:
        movl    $4294967295, %ecx
        movq    %rdi, %rax
        andq    %rcx, %rax
        ret

It would be great to write this as a Pat pattern that used subregs 
instead of a 'pseudo' instruction, but I don't know how to do that
in td files.

llvm-svn: 47658
 2008-02-27 05:47:54 +00:00
Evan Cheng 6d56368caf Spiller now remove unused spill slots. 
llvm-svn: 47657
 2008-02-27 03:04:06 +00:00
Dan Gohman 66272a545b Teach Legalize how to expand an EXTRACT_ELEMENT. 
llvm-svn: 47656
 2008-02-27 01:52:30 +00:00
Dan Gohman f19609abe8 Convert the last remaining users of the non-APInt form of 
ComputeMaskedBits to use the APInt form, and remove the
non-APInt form.

llvm-svn: 47654
 2008-02-27 01:23:58 +00:00
Devang Patel 5bc621c680 Add comment. 
llvm-svn: 47653
 2008-02-27 01:20:54 +00:00
Chris Lattner 3f86109fd1 add a note 
llvm-svn: 47652
 2008-02-27 01:17:20 +00:00
Dan Gohman ae2b6fbb8e Convert SimplifyDemandedMask and ShrinkDemandedConstant to use APInt. 
Change several cases in SimplifyDemandedMask that don't ever do any
simplifying to reuse the logic in ComputeMaskedBits instead of
duplicating it.

llvm-svn: 47648
 2008-02-27 00:25:32 +00:00
Bill Wendling 30075852ef Upgrade testcase. 
llvm-svn: 47644
 2008-02-26 23:22:40 +00:00
Devang Patel 0362de154e regenerate. 
llvm-svn: 47642
 2008-02-26 23:19:08 +00:00
Devang Patel 4db58fce82 Add assert. 
llvm-svn: 47641
 2008-02-26 23:17:50 +00:00
Owen Anderson e41c19c987 Add PR number to testcase. 
llvm-svn: 47640
 2008-02-26 23:16:11 +00:00
Owen Anderson d29ed0b122 Fix an issue where GVN had the sizes of the two memcpy's reverse, resulting 
in an invalid transformation.

llvm-svn: 47639
 2008-02-26 23:06:17 +00:00
Devang Patel f287e7dcc6 Add assert to check return type. 
llvm-svn: 47637
 2008-02-26 22:55:21 +00:00
Arnold Schwaighofer 3bfca3e942 Refactor according to Evan's and Anton's suggestions. 
llvm-svn: 47635
 2008-02-26 22:21:54 +00:00
Devang Patel 8490c6935b regenerate 
llvm-svn: 47634
 2008-02-26 22:17:48 +00:00
Devang Patel a58b1c9b6b Remove unncessary ReturnInst constructors. 
llvm-svn: 47633
 2008-02-26 22:12:58 +00:00
Chris Lattner d6bd311506 Use a smallvector for inactiveCounts and initialize it lazily 
instead of init'ing it maximally to zeros on entry.  getFreePhysReg
is pretty hot and only a few elements are typically used.  This speeds
up linscan by 5% on 176.gcc.

llvm-svn: 47631
 2008-02-26 22:08:41 +00:00
Bill Wendling d7a258d325 Rename PrintableName to Name. 
llvm-svn: 47629
 2008-02-26 21:47:57 +00:00
Anton Korobeynikov 9eeaae50ca Update per review. Patch by Mikhail Glushenkov! 
llvm-svn: 47628
 2008-02-26 21:44:24 +00:00
Bill Wendling c24ea4fb41 Change "Name" to "AsmName" in the target register info. Gee, a refactoring tool 
would have been a Godsend here!

llvm-svn: 47625
 2008-02-26 21:11:01 +00:00
Evan Cheng fa6b366892 Enable -coalescer-commute-instrs by default. 
llvm-svn: 47623
 2008-02-26 20:40:22 +00:00
Nick Kledzik 07b4a62234 first commit of new LTO system. It is not hooked up in the llvm/tools/Makefile, so no one will build it be default yet 
llvm-svn: 47621
 2008-02-26 20:26:43 +00:00
Devang Patel a736c001b6 Use SmallVector while constructing ReturnInst. 
llvm-svn: 47619
 2008-02-26 19:38:17 +00:00
Devang Patel 060e79c729 Avoid const_casts 
llvm-svn: 47616
 2008-02-26 19:15:26 +00:00
Gabor Greif aa2617206f fix http://llvm.org/bugs/show_bug.cgi?id=2097 
llvm-svn: 47615
 2008-02-26 19:13:21 +00:00
Devang Patel 85e38bfb2e Remove unnecessary getOperand/setOperand overriders. 
Simplify getReturnValue()

llvm-svn: 47614
 2008-02-26 19:08:13 +00:00
Dan Gohman 9db0aa86d9 Avoid aborting on invalid shift counts. 
llvm-svn: 47612
 2008-02-26 18:50:50 +00:00
Devang Patel c38eb52787 Unify to ReturnInst::init() member functions. 
llvm-svn: 47611
 2008-02-26 18:49:29 +00:00
Eli Friedman 666bbe34f4 Fix for pr2093: direct operands aren't necessarily addresses, so don't 
try to simplify them.

llvm-svn: 47610
 2008-02-26 18:37:49 +00:00
Chris Lattner a39cff3aaa fix this test so that the fn name doesn't match the regex 
llvm-svn: 47608
 2008-02-26 18:13:51 +00:00
Devang Patel ae682fb940 Optimize most common case by using single RetVal in ReturnInst. 
llvm-svn: 47607
 2008-02-26 17:56:20 +00:00
Arnold Schwaighofer 1f17bf6171 Correct function comments. 
llvm-svn: 47606
 2008-02-26 17:50:59 +00:00
Chris Lattner 07c83cc86e Fix PR2096, a regression introduced with my patch last night. This 
also fixes cfrac, flops, and 175.vpr

llvm-svn: 47605
 2008-02-26 17:09:59 +00:00
Gabor Greif 3d9755f6ca Really feed llvm-as with the testcase, do not let it read from stdin. This fixes the hangs seen on solaris10. 
llvm-svn: 47604
 2008-02-26 13:37:13 +00:00
Gabor Greif 0a43601153 unbreak check-one 
by supplying a dummy
"verbose" procedure

llvm-svn: 47603
 2008-02-26 13:27:49 +00:00
Gabor Greif d5691031f5 While hunting for two hanging tests, 
(on solaris10, which are:
  CodeGen/PowerPC/frounds.ll
  Transforms/InstCombine/2008-02-23-MulSub.ll)

I needed a tool to figure out which one is the guilty.

To this end I have added a verbosity
option to the test/Makefile.

It can be invoked thus:

gmake check TESTSUITE=CodeGen/PowerPC VERBOSE="-v -v"

(The number of "-v"s specifies the verbosity level.
 Instead of "-v" other aliases can be specified,
 please consult the dejagnu docs for info.)

At level >= 2 following line is logged for each
test, before running it:

ABOUT TO RUN: <test>.ll

llvm-svn: 47602
 2008-02-26 12:08:55 +00:00
Duncan Sands 7cdbbfd067 Fix a nasty bug in LegalizeTypes (spotted in 
CodeGen/PowerPC/illegal-element-type.ll): suppose
a node X is processed, and processing maps it to
a node Y.  Then X continues to exist in the DAG,
but with no users.  While processing some other
node, a new node may be created that happens to
be equal to X, and thus X will be reused rather
than a truly new node.  This can cause X to
"magically reappear", and since it is in the
Processed state in will not be reprocessed, so
at the end of type legalization the illegal node
X can still be present.  The solution is to replace
X with Y whenever X gets resurrected like this.

llvm-svn: 47601
 2008-02-26 11:21:42 +00:00
Bill Wendling 80d6b87934 De-tabify 
llvm-svn: 47600
 2008-02-26 10:57:23 +00:00
Bill Wendling d188e03715 De-tabify. 
llvm-svn: 47599
 2008-02-26 10:53:30 +00:00
Bill Wendling 7bb51dfbb1 De-tabify. 
llvm-svn: 47598
 2008-02-26 10:51:52 +00:00
Bill Wendling 293285f883 Detabify 
llvm-svn: 47597
 2008-02-26 10:49:39 +00:00
Bill Wendling 8ff4320071 Detabify. 
llvm-svn: 47596
 2008-02-26 10:46:10 +00:00
Bill Wendling e65d2afdb1 De-tabify. 
llvm-svn: 47595
 2008-02-26 10:45:29 +00:00
Arnold Schwaighofer 69a10f4112 Add support for intermodule tail calls on x86/32bit with 
GOT-style position independent code. Before only tail calls to
protected/hidden functions within the same module were optimized.
Now all function calls are tail call optimized.

llvm-svn: 47594
 2008-02-26 10:21:54 +00:00
Arnold Schwaighofer b01b99ec78 Change the lowering of arguments for tail call optimized 
calls. Before arguments that could overwrite each other were
explicitly lowered to a stack slot, not giving the register allocator
a chance to optimize. Now a sequence of copyto/copyfrom virtual
registers ensures that arguments are loaded in (virtual) registers
before they are lowered to the stack slot (and might overwrite each
other). Also parameter stack slots are marked mutable for
(potentially) tail calling functions.

llvm-svn: 47593
 2008-02-26 09:19:59 +00:00
Evan Cheng 2ff0b0e681 This is possible: 
vr1 = extract_subreg vr2, 3
...
vr3 = extract_subreg vr1, 2
The end result is vr3 is equal to vr2 with subidx 2.

llvm-svn: 47592
 2008-02-26 08:03:41 +00:00
Chris Lattner e7c14013f5 Fix isNegatibleForFree to not return true for ConstantFP nodes 
after legalize.  Just because a constant is legal (e.g. 0.0 in SSE) 
doesn't mean that its negated value is legal (-0.0).  We could make
this stronger by checking to see if the negated constant is actually
legal post negation, but it doesn't seem like a big deal.

llvm-svn: 47591
 2008-02-26 07:04:54 +00:00