Commit Graph

38075 Commits

Author SHA1 Message Date
Owen Anderson 53336d8055 Fix for PR2190. Memdep's non-local caching was checking dirtied blocks in the
wrong order.

llvm-svn: 49499
2008-04-10 22:13:32 +00:00
Dan Gohman f4bc782495 Make several symbols static.
llvm-svn: 49496
2008-04-10 21:11:47 +00:00
Dan Gohman 99b7b3f03b Teach InstCombine's ComputeMaskedBits to handle pointer expressions
in addition to integer expressions. Rewrite GetOrEnforceKnownAlignment
as a ComputeMaskedBits problem, moving all of its special alignment
knowledge to ComputeMaskedBits as low-zero-bits knowledge.

Also, teach ComputeMaskedBits a few basic things about Mul and PHI
instructions.

This improves ComputeMaskedBits-based simplifications in a few cases,
but more noticeably it significantly improves instcombine's alignment
detection for loads, stores, and memory intrinsics.

llvm-svn: 49492
2008-04-10 18:43:06 +00:00
Evan Cheng 16ea87d6ee A copy instruction may use a register multiple times on some targets. Change them all.
llvm-svn: 49491
2008-04-10 18:38:47 +00:00
Evan Cheng 5e1971c894 Add comment.
llvm-svn: 49469
2008-04-10 08:03:14 +00:00
Chris Lattner aeb23a8a34 add a note, this is actually not too bad to implement.
llvm-svn: 49466
2008-04-10 05:54:50 +00:00
Chris Lattner c692188075 move the x86-32 part of PR2108 here.
llvm-svn: 49465
2008-04-10 05:37:47 +00:00
Chris Lattner ad75302497 Fix the x86-64 side of PR2108 by adding a v2f64 version of
MOVZQI2PQIrr.  This would be better handled as a dag combine 
(with the goal of eliminating the bitconvert) but I don't know
how to do that safely.  Thoughts welcome.

llvm-svn: 49463
2008-04-10 05:13:43 +00:00
Chris Lattner 57dd77462b produce an error on invalid input instead of asserting:
def : Pat<((v2f64 (vector_shuffle immAllZerosV_bc,
           ^

llvm-svn: 49462
2008-04-10 04:48:34 +00:00
Evan Cheng 9d339849ee Teach branch folding pass about implicit_def instructions. Unfortunately we can't just eliminate them since register scavenger expects every register use to be defined. However, we can delete them when there are no intra-block uses. Carefully removing some implicit def's which enable more blocks to be optimized away.
llvm-svn: 49461
2008-04-10 02:32:10 +00:00
Chris Lattner a29d2536aa Disable an xform we've had for a long time, pow(x,0.5) -> sqrt.
This is not safe for all inputs.

llvm-svn: 49458
2008-04-10 02:07:51 +00:00
Evan Cheng 2025a490d0 Move memcpy / memset optimization pass after GVN.
llvm-svn: 49457
2008-04-10 01:33:05 +00:00
Evan Cheng c8eeb752a3 - More aggressively coalescing away copies whose source is defined by an implicit_def.
- Added insert_subreg coalescing support.

llvm-svn: 49448
2008-04-09 20:57:25 +00:00
Dan Gohman 33b3300178 Make isVectorClearMaskLegal's operand list const.
llvm-svn: 49446
2008-04-09 20:09:42 +00:00
Dan Gohman 44b3005e86 Fix some minor errors in comments.
llvm-svn: 49445
2008-04-09 20:08:06 +00:00
Dan Gohman 3049516bd5 Add const qualifiers.
llvm-svn: 49443
2008-04-09 18:31:41 +00:00
Dan Gohman bb0022b810 Update comments to use 2.0 syntax type names.
llvm-svn: 49442
2008-04-09 18:24:25 +00:00
Dan Gohman 3d074a3125 Add XMM1 as a second return value register for f32 and f64 on x86-64. This
is needed for the x86-64-ABI handling of structs that contain floating-point
members that are returned by value.

llvm-svn: 49441
2008-04-09 17:54:37 +00:00
Dan Gohman cbf87313a2 Add DX as a second return value register for i16 on x86.
llvm-svn: 49440
2008-04-09 17:53:38 +00:00
Chris Lattner 802134fc02 Generalize getUnaryFloatFunction to handle any FP unary function, automatically
figuring out the suffix to use.  implement pow(2,x) -> exp2(x).

llvm-svn: 49437
2008-04-09 17:48:11 +00:00
Chris Lattner 091afc7714 remove capital letter from test name.
llvm-svn: 49436
2008-04-09 17:46:36 +00:00
Chris Lattner cca74e5ab9 use the new ConstantFP::get method to make this work with
long double and simplify the code.

llvm-svn: 49435
2008-04-09 17:17:35 +00:00
Chris Lattner 0704e3acff ConstantFP::get should be static.
llvm-svn: 49434
2008-04-09 17:16:28 +00:00
Nate Begeman a6063118c6 Fix a bug where an incorrect bit mask would be generated if a target's last asm
string began at a power of 2 in the string index.  For example, if "ret" started
at position 16, the ret instruction would be assigned code 16, but the mask would be AsmChars[] + Code & 15, not Code & 31.

llvm-svn: 49433
2008-04-09 16:24:11 +00:00
Devang Patel a7dfbc0366 Be conservative if getresult operand is neither call nor invoke.
llvm-svn: 49430
2008-04-09 15:58:24 +00:00
Owen Anderson ef9a6fd5c2 Factor a bunch of functionality related to memcpy and memset transforms out of
GVN and into its own pass.

llvm-svn: 49419
2008-04-09 08:23:16 +00:00
Owen Anderson 8ee792d1b6 Remove accidentally duplicated code.
llvm-svn: 49418
2008-04-09 07:55:01 +00:00
Evan Cheng 0515472470 Unbreak teh build.
llvm-svn: 49417
2008-04-09 07:06:01 +00:00
Chris Lattner 98bd939af8 add a simplified accessor for creating an fp constant of a
particular value but variable type.

llvm-svn: 49416
2008-04-09 06:38:30 +00:00
Evan Cheng aa3b55f842 Missed a hasInterval check.
llvm-svn: 49415
2008-04-09 01:30:15 +00:00
Chris Lattner b5b3e3136b add a version of ConstantFP::get that doesn't take a redundant Type* value,
start migrating code over to use it.

llvm-svn: 49413
2008-04-09 00:45:01 +00:00
Chris Lattner b859fb49ed many cleanups to the pow optimizer. Allow it to handle powf,
add support for  pow(x, 2.0) -> x*x.

llvm-svn: 49411
2008-04-09 00:07:45 +00:00
Chris Lattner 8ca9450885 make ConstantFP::isExactlyValue work for long double as well.
llvm-svn: 49410
2008-04-09 00:03:58 +00:00
Devang Patel 0a14b5b649 Add CreateGetResult()
llvm-svn: 49398
2008-04-08 20:41:22 +00:00
Duncan Sands 470ab1a04d Check that bodies and calls but not declarations
are marked nounwind when compiling without
-fexceptions.

llvm-svn: 49393
2008-04-08 19:31:52 +00:00
Dale Johannesen 5169fa17b5 Rename -disable-required-unwind-tables to -unwind-tables-optional.
llvm-svn: 49391
2008-04-08 18:10:08 +00:00
Dale Johannesen 4c0c018bc5 Rename -disable-required-unwind-tables to unwind-tables-optional.
llvm-svn: 49389
2008-04-08 18:07:49 +00:00
Gabor Greif 00fcdeddd3 merge r48768 from branches/ggreif/parallelized-test
llvm-svn: 49382
2008-04-08 15:22:41 +00:00
Devang Patel 3d840ded0d Add multiple value return instruction constructor.
llvm-svn: 49374
2008-04-08 07:30:13 +00:00
Duncan Sands 57d3a36386 Convenience method for setting the nounwind
attribute for a function.

llvm-svn: 49373
2008-04-08 07:23:58 +00:00
Chris Lattner 166c458df4 silence warning when assertions are disabled.
llvm-svn: 49371
2008-04-08 05:49:09 +00:00
Devang Patel 8cd2a3ae2a Fix insert point handling for multiple return values.
llvm-svn: 49367
2008-04-08 02:24:08 +00:00
Dale Johannesen fe767621ca Handle the situation in 2008-01-25-EmptyFunction.ll
correctly when unwind info is being generated.

llvm-svn: 49366
2008-04-08 00:37:56 +00:00
Dale Johannesen 399c0e63af Missed one.
llvm-svn: 49365
2008-04-08 00:14:59 +00:00
Dale Johannesen da298f9107 Add -disable-required-unwind-tables to tests
that need it (usually, grepping for some string
found in unwind info)

llvm-svn: 49364
2008-04-08 00:14:17 +00:00
Dale Johannesen 344aec2952 Implement new llc flag -disable-required-unwind-tables.
Corresponds to -fno-unwind-tables (usually default in gcc).

llvm-svn: 49361
2008-04-08 00:10:24 +00:00
Ted Kremenek a518bb79b1 Make getDirnameSep a static method (not part of Path's interface).
llvm-svn: 49354
2008-04-07 22:01:32 +00:00
Ted Kremenek c042bf1db5 Added method Path::getDirname().
llvm-svn: 49352
2008-04-07 21:53:57 +00:00
Sam Bishop 3a13067557 Added support for Create() calls that take an argument besides the
deserializer.

llvm-svn: 49350
2008-04-07 21:36:46 +00:00
Dan Gohman 3bc3ddd638 Rename MemOperand to MachineMemOperand. This was suggested by
review feedback from Chris quite a while ago. No functionality
change.

llvm-svn: 49348
2008-04-07 19:35:22 +00:00