676130f2e1
Fix PR3423: Link llvm on ARM EABI machines. Patch by Robert Schuster.
...
llvm-svn: 63489
2009-02-01 06:42:27 +00:00
dfbb6a1a9a
DebugLoc propagation.
...
llvm-svn: 63488
2009-01-31 22:04:51 +00:00
f23908151a
Reinstate this optimization to fold icmp of xor when possible. Don't try to
...
turn icmp eq a+x, b+x into icmp eq a, b if a+x or b+x has other uses. This
may have been increasing register pressure leading to the bzip2 slowdown.
llvm-svn: 63487
2009-01-31 21:30:05 +00:00
5f98ea28ca
DebugLoc propagation. Done with file.
...
llvm-svn: 63486
2009-01-31 21:04:24 +00:00
4d9fa9e71d
DebugLoc propagation. Done with file.
...
llvm-svn: 63485
2009-01-31 20:01:02 +00:00
9e2b9f3234
Fix PR3452 (an infinite loop bootstrapping) by disabling the recent
...
improvements to the EvaluateInDifferentType code. This code works
by just inserted a bunch of new code and then seeing if it is
useful. Instcombine is not allowed to do this: it can only insert
new code if it is useful, and only when it is converging to a more
canonical fixed point. Now that we iterate when DCE makes progress,
this causes an infinite loop when the code ends up not being used.
llvm-svn: 63483
2009-01-31 19:05:27 +00:00
41826036b1
Fix PR3401: when using large integers, the type
...
returned by getShiftAmountTy may be too small
to hold shift values (it is an i8 on x86-32).
Before and during type legalization, use a large
but legal type for shift amounts: getPointerTy;
afterwards use getShiftAmountTy, fixing up any
shift amounts with a big type during operation
legalization. Thanks to Dan for writing the
original patch (which I shamelessly pillaged).
llvm-svn: 63482
2009-01-31 15:50:11 +00:00
76a63ed099
now that all the pieces are in place, teach instcombine's
...
simplifydemandedbits to simplify instructions with *multiple
uses* in contexts where it can get away with it. This allows
it to simplify the code in multi-use-or.ll into a single 'add
double'.
This change is particularly interesting because it will cover
up for some common codegen bugs with large integers created due
to the recent SROA patch. When working on fixing those bugs,
this should be disabled.
llvm-svn: 63481
2009-01-31 08:40:03 +00:00
3e2cb66c56
simplify/clarify control flow and improve comments, no functionality change.
...
llvm-svn: 63480
2009-01-31 08:24:16 +00:00
83c6a141b8
make some fairly meaty internal changes to how SimplifyDemandedBits works.
...
Now, if it detects that "V" is the same as some other value,
SimplifyDemandedBits returns the new value instead of RAUW'ing it immediately.
This has two benefits:
1) simpler code in the recursive SimplifyDemandedBits routine.
2) it allows future fun stuff in instcombine where an operation has multiple
uses and can be simplified in one context, but not all.
#2 isn't implemented yet, this patch should have no functionality change.
llvm-svn: 63479
2009-01-31 08:15:18 +00:00
19efe30264
add accessors
...
llvm-svn: 63478
2009-01-31 07:34:19 +00:00
585cfb2ce7
minor cleanups
...
llvm-svn: 63477
2009-01-31 07:26:06 +00:00
94cfb281c3
make sure to set Changed=true when instcombine hacks on the code,
...
not doing so prevents it from properly iterating and prevents it
from deleting the entire body of dce-iterate.ll
llvm-svn: 63476
2009-01-31 07:04:22 +00:00
b6080cf943
Used "-enable-unsafe-fp-math" to allow this transformation - (a * b -c) = c - a *b.
...
llvm-svn: 63475
2009-01-31 06:50:54 +00:00
cf9ba82324
If unsafe FP optimization is not set, don't allow -(A-B) => B-A because
...
when A==B, -0.0 != +0.0.
llvm-svn: 63474
2009-01-31 06:07:45 +00:00
3b585af0ec
Don't use DebugLoc::getUnknownLoc(). Default to something hopefully sensible.
...
llvm-svn: 63473
2009-01-31 03:12:48 +00:00
ec99c46d44
Simplify and generalize the SROA "convert to scalar" transformation to
...
be able to handle *ANY* alloca that is poked by loads and stores of
bitcasts and GEPs with constant offsets. Before the code had a number
of annoying limitations and caused it to miss cases such as storing into
holes in structs and complex casts (as in bitfield-sroa) where we had
unions of bitfields etc. This also handles a number of important cases
that are exposed due to the ABI lowering stuff we do to pass stuff by
value.
One case that is pretty great is that we compile
2006-11-07-InvalidArrayPromote.ll into:
define i32 @func(<4 x float> %v0, <4 x float> %v1) nounwind {
%tmp10 = call <4 x i32> @llvm.x86.sse2.cvttps2dq(<4 x float> %v1)
%tmp105 = bitcast <4 x i32> %tmp10 to i128
%tmp1056 = zext i128 %tmp105 to i256
%tmp.upgrd.43 = lshr i256 %tmp1056, 96
%tmp.upgrd.44 = trunc i256 %tmp.upgrd.43 to i32
ret i32 %tmp.upgrd.44
}
which turns into:
_func:
subl $28, %esp
cvttps2dq %xmm1, %xmm0
movaps %xmm0, (%esp)
movl 12(%esp), %eax
addl $28, %esp
ret
Which is pretty good code all things considering :).
One effect of this is that SROA will start generating arbitrary bitwidth
integers that are a multiple of 8 bits. In the case above, we got a
256 bit integer, but the codegen guys assure me that it can handle the
simple and/or/shift/zext stuff that we're doing on these operations.
This addresses rdar://6532315
llvm-svn: 63469
2009-01-31 02:28:54 +00:00
db7c5f6a7b
Move CurDebugLoc into SelectionDAGLowering.
...
llvm-svn: 63468
2009-01-31 02:22:37 +00:00
dc588689bb
back out my previous change, it exposes a latent bug. investigating
...
llvm-svn: 63463
2009-01-31 01:31:59 +00:00
f88092ee06
use precise accessors
...
llvm-svn: 63459
2009-01-31 00:58:14 +00:00
dc0f124429
Propagate debug info in LegalizeFloatTypes.
...
Complete (modulo bugs).
llvm-svn: 63458
2009-01-31 00:43:08 +00:00
42aa385e20
Propagate debug info. This file complete
...
(modulo bugs)
llvm-svn: 63457
2009-01-31 00:20:43 +00:00
c910889511
Propagate debug info through MakeLibCall and a
...
couple of things that use it.
llvm-svn: 63456
2009-01-31 00:11:23 +00:00
31b50991cb
More DebugLoc propagation.
...
llvm-svn: 63454
2009-01-30 23:59:18 +00:00
27d9dd4b57
More DebugLoc propagation.
...
llvm-svn: 63452
2009-01-30 23:36:47 +00:00
306bfc2213
More DebugLoc propagation in LOAD etc. methods.
...
llvm-svn: 63451
2009-01-30 23:27:35 +00:00
9d7c69d832
Fix build on case-sensitive filesystems (i.e. everybody else)
...
llvm-svn: 63448
2009-01-30 23:20:42 +00:00
0bd29743e3
More DebugLoc propagation in floating-point methods.
...
llvm-svn: 63446
2009-01-30 23:15:49 +00:00
555a375bb6
Make LowerCallTo and LowerArguments take a DebugLoc
...
argument. Adjust all callers and overloaded versions.
llvm-svn: 63444
2009-01-30 23:10:59 +00:00
6fbf5495f8
Standardize comments about folding xforms.
...
llvm-svn: 63443
2009-01-30 23:10:18 +00:00
8fb81f1b3d
Get rid of the non-DebugLoc-ified getNOT() method.
...
llvm-svn: 63442
2009-01-30 23:03:19 +00:00
3dc5d2454e
Propagate debug loc info for some FP arithmetic methods.
...
llvm-svn: 63441
2009-01-30 22:57:07 +00:00
cb9be5d174
Propagate debug loc info for some FP arithmetic methods.
...
llvm-svn: 63440
2009-01-30 22:53:48 +00:00
4e0a61514b
Propagate debug loc info for BIT_CONVERT.
...
llvm-svn: 63439
2009-01-30 22:44:24 +00:00
7bfa43b022
Propagate debug loc info for more *_EXTEND methods.
...
llvm-svn: 63437
2009-01-30 22:33:24 +00:00
9b3dc8d848
Propagate debug loc info for ANY_EXTEND.
...
llvm-svn: 63436
2009-01-30 22:27:33 +00:00
c409318562
Propagate debug loc info for some of the *_EXTEND functions.
...
llvm-svn: 63434
2009-01-30 22:23:15 +00:00
cab9a2eef5
DebugLoc form of getNOT().
...
llvm-svn: 63433
2009-01-30 22:11:22 +00:00
b6b6f46fe4
- Propagate debug loc info for SELECT.
...
- Added xform for (select X, 1, Y) and (select X, Y, 0), which was commented on,
but missing.
llvm-svn: 63428
2009-01-30 22:02:18 +00:00
d51e3ff540
Propagate debug loc info for Shifts.
...
llvm-svn: 63424
2009-01-30 21:37:17 +00:00
35972a9460
Propagate debug loc info for XOR and MatchRotate.
...
llvm-svn: 63420
2009-01-30 21:14:50 +00:00
f29b6e1318
Propagate debug loc info for OR. Also clean up some comments.
...
llvm-svn: 63419
2009-01-30 20:59:34 +00:00
ff8acd684f
Perform obvious constant arithmetic folding.
...
llvm-svn: 63417
2009-01-30 20:50:00 +00:00
8617191302
Propagate debug loc info for AND. Also clean up some comments.
...
llvm-svn: 63416
2009-01-30 20:43:18 +00:00
781db7a1ad
Propagate debug loc info in SimplifyBinOpWithSameOpcodeHands.
...
llvm-svn: 63411
2009-01-30 19:25:47 +00:00
1933b00c12
use precise getters
...
llvm-svn: 63403
2009-01-30 18:27:21 +00:00
97f1720621
use precise getters
...
llvm-svn: 63402
2009-01-30 18:21:13 +00:00
c094970cd2
Each input file is encoded as a separate compile unit in LLVM debugging
...
information output. However, many target specific tool chains prefer to encode
only one compile unit in an object file. In this situation, the LLVM code
generator will include debugging information entities in the compile unit
that is marked as main compile unit. The code generator accepts maximum one main
compile unit per module. If a module does not contain any main compile unit
then the code generator will emit multiple compile units in the output object
file.
[Part 1]
Update DebugInfo APIs to accept optional boolean value while creating DICompileUnit to mark the unit as "main" unit. By defaults all units are considered non-main. Update SourceLevelDebugging.html to document "main" compile unit.
Update DebugInfo APIs to not accept and encode separate source file/directory entries while creating various llvm.dbg.* entities. There was a recent, yet to be documented, change to include this additional information so no documentation changes are required here.
Update DwarfDebug to handle "main" compile unit. If "main" compile unit is seen then all DIEs are inserted into "main" compile unit. All other compile units are used to find source location for llvm.dbg.* values. If there is not any "main" compile unit then create unique compile unit DIEs for each llvm.dbg.compile_unit.
[Part 2]
Create separate llvm.dbg.compile_unit for each input file. Mark compile unit create for main_input_filename as "main" compile unit. Use appropriate compile unit, based on source location information collected from the tree node, while creating llvm.dbg.* values using DebugInfo APIs.
---
This is Part 1.
llvm-svn: 63400
2009-01-30 18:20:31 +00:00
4b23938fb9
As Duncan suggested, add braces for the one-line "else branch".
...
Thanks, Duncan.
llvm-svn: 63389
2009-01-30 09:44:49 +00:00
25b6f8153e
This is to fix the bug in IntrinsicLowering.cpp,
...
the LowerPartSet(). It didn't handle the situation correctly when
the low, high argument values are in reverse order (low > high)
with 'Val' type i32 (a corner case).
llvm-svn: 63388
2009-01-30 09:02:50 +00:00
c10d810303
Fixed the comment. No functionality change.
...
llvm-svn: 63387
2009-01-30 09:01:44 +00:00
1e36fbb5ed
This is case is to uncover the bug in IntrinsicLowering.cpp,
...
the LowerPartSet(). It didn't handle the situation correctly when
the low, high argument values are in reverse order (low > high)
with 'Val' type is i32 (a corner case).
llvm-svn: 63386
2009-01-30 08:59:51 +00:00
ecbe4a0860
Add opposite_of and inverse_opt to support -fno- style options. This
...
is necessary for eventual gcc commmand line compatibility.
llvm-svn: 63384
2009-01-30 08:19:46 +00:00
082174cb78
Use sublw for comparison with literals instead of subwf.
...
llvm-svn: 63382
2009-01-30 07:55:25 +00:00
cbb20a6ee1
When PerformBuildVectorCombine, avoid creating a X86ISD::VZEXT_LOAD of
...
an illegal type.
llvm-svn: 63380
2009-01-30 07:07:40 +00:00
964a29f671
Enable emitting of constant values in non-default address space as well. The APIs emitting constants now take an additional parameter signifying the address space in which to emit. The APIs like getData8BitsDirective() etc are made virtual enabling targets to be able to define appropirate directivers for various sizes and address spaces.
...
llvm-svn: 63377
2009-01-30 04:25:10 +00:00
9b3407e5bb
Propagate debug loc info in SimplifyNodeWithTwoResults.
...
llvm-svn: 63376
2009-01-30 03:08:40 +00:00
faed065e5c
Propagate debug loc info for MULHS.
...
llvm-svn: 63375
2009-01-30 03:00:18 +00:00
d033af09fd
Propagate debug loc info for SREM and UREM.
...
llvm-svn: 63374
2009-01-30 02:57:00 +00:00
aff3e03765
Propagate debug loc info for UDIV.
...
llvm-svn: 63373
2009-01-30 02:55:25 +00:00
5b663e7b53
Propagate debug loc info for SDIV.
...
llvm-svn: 63372
2009-01-30 02:52:17 +00:00
b48dcf67e5
Forgot to propagate debug loc info here.
...
llvm-svn: 63371
2009-01-30 02:49:26 +00:00
1ee0d41ef8
Fix a post-RA scheduling dependency bug.
...
If a MachineInstr doesn't have a memoperand but has an opcode that
is known to load or store, assume its memory reference may alias
*anything*, including stack slots which the compiler completely
controls.
To partially compensate for this, teach the ScheduleDAG building
code to do basic getUnderlyingValue analysis. This greatly
reduces the number of instructions that require restrictive
dependencies. This code will need to be revisited when we start
doing real alias analysis, but it should suffice for now.
llvm-svn: 63370
2009-01-30 02:49:14 +00:00
091f92f568
Propagate debug loc info for MUL.
...
llvm-svn: 63369
2009-01-30 02:45:56 +00:00
48ff08ef3e
Propagate debug loc info in SUB.
...
llvm-svn: 63368
2009-01-30 02:42:10 +00:00
6127757920
Propagate debug loc info in ADDC and ADDE.
...
llvm-svn: 63367
2009-01-30 02:38:00 +00:00
c442348dd7
Propagate debug loc info in DAG combine's "ADD".
...
llvm-svn: 63366
2009-01-30 02:31:17 +00:00
cdd96133bd
- Propagate debug loc info in combineSelectAndUse().
...
- Modify ReassociateOps so that the resulting SDValue is what the comment claims
it is.
llvm-svn: 63365
2009-01-30 02:23:43 +00:00
c7d6951c92
Add some comments.
...
llvm-svn: 63364
2009-01-30 02:12:57 +00:00
caffe92d86
Fix a bug in getMaxValue/getMinValue to pass the right signedness the
...
the constructed APSInt, patch suggested by Ben Laurie!
llvm-svn: 63362
2009-01-30 01:58:19 +00:00
acbb381cc4
Enable target tripple.
...
llvm-svn: 63361
2009-01-30 01:40:58 +00:00
a103ee4f0d
Linux and other target's encoding for DW_AT_declaration may not match.
...
llvm-svn: 63360
2009-01-30 01:37:30 +00:00
ed255b3d8e
Propagate debug info when building SelectionDAG.
...
llvm-svn: 63359
2009-01-30 01:34:22 +00:00
33a21670f4
Remove dead code, again.
...
llvm-svn: 63358
2009-01-30 01:27:49 +00:00
776a4c5baa
Remove dead code.
...
llvm-svn: 63357
2009-01-30 01:25:47 +00:00
4ba91058d2
Add DW_AT_declaration for class methods.
...
llvm-svn: 63356
2009-01-30 01:21:46 +00:00
9c9a3b6665
Propagate debug location info for the token factor.
...
llvm-svn: 63355
2009-01-30 01:13:16 +00:00
199740651c
Add dump() routines to help debug debug info :)
...
llvm-svn: 63353
2009-01-30 01:03:10 +00:00
f6d0aff0bd
Add DebugLoc propagation to some of the methods in DAG combiner.
...
llvm-svn: 63350
2009-01-30 00:45:56 +00:00
33dddf3235
Explain why this is here.
...
llvm-svn: 63342
2009-01-29 23:19:43 +00:00
ad89c410e6
XFAIL this test. It only worked before because of a bug in the spill point selection code. Not deleting because
...
it should be possible to enhance the selection code to handle this in the future.
llvm-svn: 63340
2009-01-29 22:27:56 +00:00
bf77d2eb9d
Correct the algorithms for choosing spill and restore points so that we don't try to insert loads/stores between call frame setup and the actual call.
...
This fixes the last known failure for the pre-alloc-splitter.
llvm-svn: 63339
2009-01-29 22:13:06 +00:00
14d55f0a5c
Explicitly add PseudoSourceValue information when lowering
...
BUILD_VECTOR and conversions to stack operations.
llvm-svn: 63333
2009-01-29 21:02:43 +00:00
60d6844aa8
Make a few things const, fix some comments, and simplify
...
some assertions.
llvm-svn: 63328
2009-01-29 19:49:27 +00:00
dfb6735385
Fix windows build, patch by Marius Wachtler!
...
llvm-svn: 63325
2009-01-29 18:53:28 +00:00
a160d4af82
Local register allocator shouldn't assume only the entry and landing pad basic blocks have live-ins.
...
llvm-svn: 63323
2009-01-29 18:37:30 +00:00
8b437ccbbe
Fix two typos that Duncan spotted in a comment.
...
llvm-svn: 63312
2009-01-29 16:18:12 +00:00
ef04ed5477
In the case of an extractelement on an insertelement value,
...
the element indices may be equal if either one is not a
constant.
llvm-svn: 63311
2009-01-29 16:10:46 +00:00
a115859df0
Add a always_inline test case.
...
llvm-svn: 63304
2009-01-29 09:31:54 +00:00
a434d930ff
Revert r63273. This was already implemented by Dale. There's no need for my
...
change.
llvm-svn: 63301
2009-01-29 09:01:55 +00:00
45799abe61
Add a test case for Chris lvalue alignment fixes.
...
llvm-svn: 63300
2009-01-29 08:59:46 +00:00
c0924a0632
Fix an issue where restores could be inserted after a terminator instruction,
...
and an iterator invalidation issue.
FreeBench/pifft no longer miscompiles with these fixes!
llvm-svn: 63293
2009-01-29 08:22:06 +00:00
a590802ed6
Comments are good. :-)
...
llvm-svn: 63276
2009-01-29 05:41:02 +00:00
b05ce78c51
Add support for aggressive load-use-store folding. This takes care of the
...
vast majority of code size regressions introduced by pre-alloc-splitting.
llvm-svn: 63274
2009-01-29 05:28:55 +00:00
50338007b9
- Add DebugLoc to getTargetNode().
...
- Modify TableGen to add the DebugLoc when calling getTargetNode.
(The light-weight wrappers are only temporary. The non-DebugLoc version will be
removed once the whole debug info stuff is finished with.)
llvm-svn: 63273
2009-01-29 05:27:31 +00:00
c0474013cb
Fix PR3424, a static constructor ordering issue. Patch by Robert Schuster!
...
llvm-svn: 63269
2009-01-29 04:43:42 +00:00
76a2736c74
Exit with nice warnings when register allocator run out of registers.
...
llvm-svn: 63267
2009-01-29 02:20:59 +00:00
e58ab79f33
Make x86's BT instruction matching more thorough, and add some
...
dagcombines that help it match in several more cases. Add
several more cases to test/CodeGen/X86/bt.ll. This doesn't
yet include matching for BT with an immediate operand, it
just covers more register+register cases.
llvm-svn: 63266
2009-01-29 01:59:02 +00:00
84f0165273
Move the code that starts printing the Select_* functions
...
after the code that sorts the patterns. This doesn't
affect the output, but it makes the code a little easier
to follow.
llvm-svn: 63265
2009-01-29 01:37:18 +00:00
89ba594975
A slight compile time optimization. If the caller knows there isn't a free register getReg() should not call getFreeReg().
...
llvm-svn: 63263
2009-01-29 01:13:00 +00:00
839acbb089
Add DebugLoc-sensitive versions of many node creation
...
functions. Currently omitted: memcpy, memmove, memset.
llvm-svn: 63259
2009-01-29 00:47:48 +00:00
55ca1d38ea
Fix comment about removeRange.
...
llvm-svn: 63255
2009-01-29 00:06:09 +00:00
9150f735fa
Fixed lowering of v816 shuffles.
...
llvm-svn: 63252
2009-01-28 23:11:14 +00:00
42b63bc175
Make test platform agnostic.
...
llvm-svn: 63247
2009-01-28 22:20:56 +00:00
1b6a3bce82
Add DebugLoc to the getNode() methods.
...
llvm-svn: 63245
2009-01-28 22:17:52 +00:00
d21775ae0e
Give this test an explicit target, to make it host-independent.
...
llvm-svn: 63244
2009-01-28 22:14:58 +00:00
511d5ece2b
SDOperand has been renamed to SDValue. SDNode::Val is now
...
accessed via SDNode::getNode.
llvm-svn: 63240
2009-01-28 21:36:46 +00:00
f52ac41566
Add more comments describing SDNode operator codes.
...
llvm-svn: 63239
2009-01-28 21:34:36 +00:00
666bf20441
Add DebugLoc-aware constructors for SDNode derived
...
classes (those that reasonably have a DebugLoc
associated with them).
llvm-svn: 63236
2009-01-28 21:18:29 +00:00
7a1b4538a0
Add some comments on ISD::NodeType.
...
llvm-svn: 63234
2009-01-28 21:13:08 +00:00
d7ecb3b661
Do not forget to derived type while constructing an array type.
...
llvm-svn: 63233
2009-01-28 21:08:20 +00:00
df17987c19
Fix some issues with volatility, move "CanConvertToScalar" check
...
after the others.
llvm-svn: 63227
2009-01-28 20:16:43 +00:00
1498e62117
strengthen this test.
...
llvm-svn: 63222
2009-01-28 19:29:30 +00:00
a15ea78ea6
Fixed extract element when the result needs to be promoted and the input widened.
...
llvm-svn: 63217
2009-01-28 18:53:39 +00:00
e7d7b0c2fd
Delete unnecessary elses.
...
llvm-svn: 63214
2009-01-28 18:03:09 +00:00
ff2d99d94b
Enable PACKAGE_VERSION in cmake builds, this unbreaks the clang build with cmake.
...
Patch by Piotr Rak!
llvm-svn: 63213
2009-01-28 17:49:03 +00:00
4aa1846215
Make isOperationLegal do what its name suggests, and introduce a
...
new isOperationLegalOrCustom, which does what isOperationLegal
previously did.
Update a bunch of callers to use isOperationLegalOrCustom
instead of isOperationLegal. In some case it wasn't obvious
which behavior is desired; when in doubt I changed then to
isOperationLegalOrCustom as that preserves their previous
behavior.
This is for the second half of PR3376.
llvm-svn: 63212
2009-01-28 17:46:25 +00:00
ba21b7d57a
Formatting.
...
llvm-svn: 63199
2009-01-28 14:42:54 +00:00
5a913d61e3
Rename getAnalysisToUpdate to getAnalysisIfAvailable.
...
llvm-svn: 63198
2009-01-28 13:14:17 +00:00
08e53d041f
Fix PR3415 (infinite loop in EscapeAnalysis) by
...
deleting the escape analysis pass.
llvm-svn: 63197
2009-01-28 11:33:59 +00:00
f31f288863
The memory alignment requirement on some of the mov{h|l}p{d|s} patterns are 16-byte. That is overly strict. These instructions read / write f64 memory locations without alignment requirement.
...
llvm-svn: 63195
2009-01-28 08:35:02 +00:00
d880efc005
Added sse test patterns for r62979 and r63193.
...
llvm-svn: 63194
2009-01-28 08:13:56 +00:00
5a685a52c1
Add shuffle splat pattern for x86 sse shifts.
...
llvm-svn: 63193
2009-01-28 08:12:05 +00:00
02b47b5e7a
Typo.
...
llvm-svn: 63174
2009-01-28 03:47:58 +00:00
84172f7dca
Update the generated docs.
...
llvm-svn: 63173
2009-01-28 03:47:38 +00:00
2115d09a10
Add three new option properties.
...
Adds new option properties 'multi_val', 'one_or_more' and 'zero_or_one'.
llvm-svn: 63172
2009-01-28 03:47:20 +00:00
57cf3964f3
Clarify comment.
...
llvm-svn: 63171
2009-01-28 03:46:22 +00:00
b3bbde3e62
Use ValueType::bitsLT to simplify some code.
...
llvm-svn: 63170
2009-01-28 03:10:52 +00:00
172ad92b29
Use ZERO_EXTEND instead of ANY_EXTEND when promoting
...
shift amounts, to avoid implicitly assuming that
target architectures will ignore the high bits.
llvm-svn: 63169
2009-01-28 02:58:31 +00:00
3cf43ca571
Comment fixes.
...
llvm-svn: 63164
2009-01-28 01:19:52 +00:00
e4510972a6
Suppress a compile time warning.
...
llvm-svn: 63161
2009-01-28 00:53:34 +00:00
b3e0931401
Embalm my ideas of how things should work. Not that
...
anyone will pay attention.
llvm-svn: 63155
2009-01-27 23:32:10 +00:00
5adc3dbc96
Add type DIE into appropriate context DIE.
...
llvm-svn: 63154
2009-01-27 23:22:55 +00:00
4e04ef3c55
Add a DebugLoc field and some simple accessors.
...
llvm-svn: 63152
2009-01-27 23:20:29 +00:00
675869aef0
Reorder args, constify.
...
llvm-svn: 63151
2009-01-27 23:19:41 +00:00
fd03bdd00c
Add testcase for r63142.
...
llvm-svn: 63149
2009-01-27 23:00:53 +00:00
70d4c08cfe
Treat [1 x i8] zeroinitializer as a C string, placing such stuff into
...
mergeable string section. I don't see any bad impact of such decision
(rather then placing it into mergeable const section, as it was before),
but at least Darwin linker won't complain anymore.
The problem in LLVM is that we don't have special type for string constants
(like gcc does). Even more, we have two separate types: ConstatArray for non-null
strings and ConstantAggregateZero for null stuff.... It's a bit weird :)
llvm-svn: 63142
2009-01-27 22:29:24 +00:00
b841dc2957
Use .empty() instead of comparing .size() with 0.
...
llvm-svn: 63139
2009-01-27 22:12:23 +00:00
aa064e6cd6
Update to latest spelling.
...
llvm-svn: 63138
2009-01-27 22:09:11 +00:00
f041127bda
Add DebugLoc field and simple accessors.
...
llvm-svn: 63136
2009-01-27 21:41:04 +00:00
573e529127
Refine DebugLoc per review comments.
...
llvm-svn: 63132
2009-01-27 21:15:07 +00:00
fb58faf29e
Add an assertion to the form of SelectionDAG::getConstant that takes
...
a uint64_t to verify that the value is in range for the given type,
to help catch accidental overflow. Fix a few places that relied on
getConstant implicitly truncating the value.
llvm-svn: 63128
2009-01-27 20:39:34 +00:00
0ca1e7c6b6
Reformat the allocation-order arrays to a more conventional style.
...
llvm-svn: 63121
2009-01-27 19:25:38 +00:00
0bd9546039
Delete redundant return statements.
...
llvm-svn: 63120
2009-01-27 19:23:22 +00:00
13f18e853f
Respect the DisableRedZone flag on PowerPC.
...
llvm-svn: 63119
2009-01-27 19:19:28 +00:00
f77f0ce21a
Simplify findNonImmUse; return the result using the return value
...
instead of via a by-reference argument. No functionality change.
llvm-svn: 63118
2009-01-27 19:04:30 +00:00
ec39eb8519
APInt's countLeadingOnes() was broken for negative i128 values,
...
causing assertion failures in getSExtValue().
Fix it by making highWordBits actually contain what its name says,
and add some more unit-tests for APInt.
This fixes PR3419.
llvm-svn: 63107
2009-01-27 18:06:03 +00:00
ed2ba0fb01
Make some comments doxygen-friendly.
...
llvm-svn: 63104
2009-01-27 17:28:23 +00:00
431053ed2c
Make the pre-split-limit option more useful by using a per-function counter.
...
llvm-svn: 63091
2009-01-27 05:01:15 +00:00
1bc8af207e
Implement multiple with overflow by 2 with an add instruction.
...
llvm-svn: 63090
2009-01-27 03:30:42 +00:00
ce95cddd0f
Forgot this test case.
...
llvm-svn: 63089
2009-01-27 02:59:39 +00:00
7740523a89
Eliminate unnecessary operands-list traversals.
...
llvm-svn: 63088
2009-01-27 02:37:43 +00:00
52e907a780
Add a FrontendC testcase for the x86-64 Red Zone feature,
...
to help verify that the feature may be disabled through
the -mno-red-zone option.
llvm-svn: 63079
2009-01-27 00:59:55 +00:00
75cee3a93c
Enable the red zone on x86-64 by default.
...
llvm-svn: 63078
2009-01-27 00:58:47 +00:00
45c899cd15
Assorted debug info fixes.
...
- DW_AT_bit_size is only suitable for bitfields.
- Encode source location info for derived types.
- Source location and type size info is not useful for subroutine_type (info is included in respective DISubprogram) and array_type.
llvm-svn: 63077
2009-01-27 00:45:04 +00:00
8738997c11
Add a regression test for x86-64 red zone usage.
...
llvm-svn: 63075
2009-01-27 00:40:27 +00:00
1cd2a2c9f8
Fix the Red Zone calculation for functions with frame pointers.
...
Don't use the Red Zone when dynamic stack realignment is needed.
This could be implemented, but most x86-64 ABIs don't require
dynamic stack realignment so it isn't urgent.
llvm-svn: 63074
2009-01-27 00:40:06 +00:00
b8c79bdc31
No need to keep size of DebugLocations vector separately.
...
llvm-svn: 63070
2009-01-26 23:47:30 +00:00
03490f0ce1
Testcase for 6522054.
...
llvm-svn: 63067
2009-01-26 23:22:19 +00:00
49483188c3
CellSPU:
...
- Update DWARF debugging support.
llvm-svn: 63059
2009-01-26 22:33:37 +00:00
3789a13c30
Make the Dwarf macro information section optional; CellSPU's assembler
...
doesn't support it. The default is set to 'true', so this should not
impact any other target backends.
llvm-svn: 63058
2009-01-26 22:32:51 +00:00
b6d36e1d14
Implement Red Zone utilization on x86-64. This is currently
...
disabled by default; I'll enable it when I hook it up with
the llvm-gcc flag which controls it.
llvm-svn: 63056
2009-01-26 22:22:31 +00:00
8cf220398b
More updates to VC proj...
...
llvm-svn: 63050
2009-01-26 22:03:42 +00:00
e84c3f261e
Reapply r63025 and r63026, with fixes for the failing testcases.
...
llvm-svn: 63049
2009-01-26 21:57:31 +00:00
d77e476921
Fix PR3393, which amounts to a bug in the expensive
...
checking logic. Rather than make the checking more
complicated, I've tweaked some logic to make things
conform to how the checking thought things ought to
be, since this results in a simpler "mental model".
llvm-svn: 63048
2009-01-26 21:54:18 +00:00
a26699211d
Add method raw_fd_ostream::seek() for random access within a file.
...
llvm-svn: 63044
2009-01-26 21:42:04 +00:00
ac272eaf13
At Nick Lewycky's request, rename this test with a more informative name.
...
llvm-svn: 63042
2009-01-26 21:36:31 +00:00
e7fa3c37c6
Also revert r63206
...
llvm-svn: 63041
2009-01-26 21:30:17 +00:00
a39c83d118
Temporarily revert r63025 until the testsuite failures can be fixed.
...
llvm-svn: 63040
2009-01-26 21:27:03 +00:00
0b46005b19
Fix the name of an argument.
...
llvm-svn: 63037
2009-01-26 20:27:11 +00:00
4b4622454c
During bittest switch lowering emit shift in the test block, which should (theoretically)
...
allow us to generate more efficient code. We don't do this now though :)
llvm-svn: 63027
2009-01-26 19:26:01 +00:00
5e235d3a7e
Get rid of a bunch of dead code now that interval reconstruction is enabled.
...
llvm-svn: 63026
2009-01-26 19:18:06 +00:00
feffdec244
Fix an issue where LiveIntervals was trying to be smart about removing kill
...
markers, and ended up foiling the interval reconstruction.
This allows us to turn on reconstruction in the pre alloc splitter, which
fixes a number of miscompilations.
llvm-svn: 63025
2009-01-26 19:12:06 +00:00
6c7e85142b
Enhance logic in X86DAGToDAGISel::PreprocessForRMW which move load inside callseq_start to allow it to be folded into a call. It was not considering the cases where a token factor is between the load and the callseq_start.
...
llvm-svn: 63022
2009-01-26 18:43:34 +00:00
72287a599f
Silence a bogus compiler warning.
...
llvm-svn: 63021
2009-01-26 18:33:51 +00:00
0113928f94
Update VS project files.
...
llvm-svn: 63017
2009-01-26 18:08:55 +00:00
179dd918b9
Fix a bug in BitVector.h. All assignment operations (except the usual
...
assignment operator) were returning a copy of the bit vector, instead of a
reference! This old semantics probably did not meet the expectations.
With this patch, chained assignments happen to the right object.
llvm-svn: 63012
2009-01-26 11:07:20 +00:00
4e6871da09
Actually source file has already been uniquified into an id during isel. Eliminate the StringMap.
...
llvm-svn: 63009
2009-01-26 07:53:42 +00:00
b5706c45fa
Add data structure to define and track debug location during codegen.
...
llvm-svn: 63008
2009-01-26 07:41:49 +00:00
fd0ab12327
Looks like comments were chopped off.
...
llvm-svn: 63007
2009-01-26 07:40:13 +00:00
3975bbec58
LLVM_SUPPORT_DEBUGINFO_H -> LLVM_ANALYSIS_DEBUGINFO_H since DebugInfo.h is under Analysis.
...
llvm-svn: 63005
2009-01-26 07:31:20 +00:00
3537a62704
Fixed optimization of combining two shuffles where the first shuffle inputs
...
has a different number of elements than the output.
llvm-svn: 62998
2009-01-26 04:39:00 +00:00
8e4ac9b71a
Take the next steps in making SDUse more consistent with LLVM Use, and
...
tidy up SDUse and related code.
- Replace the operator= member functions with a set method, like
LLVM Use has, and variants setInitial and setNode, which take
care up updating use lists, like LLVM Use's does. This simplifies
code that calls these functions.
- getSDValue() is renamed to get(), as in LLVM Use, though most
places can either use the implicit conversion to SDValue or the
convenience functions instead.
- Fix some more node vs. value terminology issues.
Also, eliminate the one remaining use of SDOperandPtr, and
SDOperandPtr itself.
llvm-svn: 62995
2009-01-26 04:35:06 +00:00
95b2a206ee
Untabify code.
...
llvm-svn: 62991
2009-01-26 03:37:41 +00:00
9e3e4a9219
CellSPU:
...
- Rename fcmp.ll test to fcmp32.ll, start adding new double tests to fcmp64.ll
- Fix select_bits.ll test
- Capitulate to the DAGCombiner and move i64 constant loads to instruction
selection (SPUISelDAGtoDAG.cpp).
<rant>DAGCombiner will insert all kinds of 64-bit optimizations after
operation legalization occurs and now we have to do most of the work that
instruction selection should be doing twice (once to determine if v2i64
build_vector can be handled by SelectCode(), which then runs all of the
predicates a second time to select the necessary instructions.) But,
CellSPU is a good citizen.</rant>
llvm-svn: 62990
2009-01-26 03:31:40 +00:00
624801e87e
Fix a typo
...
llvm-svn: 62989
2009-01-26 03:15:54 +00:00
a2550a8e96
De-identifying per sabre review
...
llvm-svn: 62988
2009-01-26 03:15:31 +00:00
5c8fc9d549
Build libLTO on any platform so long as PIC is enabled.
...
llvm-svn: 62987
2009-01-26 03:04:57 +00:00
7ea4e861ca
Add some documentation for address-space-based access to the segment registers.
...
llvm-svn: 62986
2009-01-26 02:54:45 +00:00
9449991c4f
Handle single-entry phi nodes gracefully in condprop.
...
llvm-svn: 62985
2009-01-26 02:18:20 +00:00
7b6647c178
Fix PR3408 by making a non-obvious assumption very obvious, and
...
handling the flaw inherent in that assumption. :)
llvm-svn: 62984
2009-01-26 02:11:30 +00:00
57cb472b56
More cleanups and simplifications, no functionality change.
...
llvm-svn: 62983
2009-01-26 01:57:01 +00:00
d67aaa6560
tidy asserts
...
llvm-svn: 62982
2009-01-26 01:38:24 +00:00
5eca265519
Map address space 256 to gs; similar mappings could be supported for the
...
other x86 segments. address space 0 is stack/default, 1-255 are reserved for
client use.
llvm-svn: 62980
2009-01-26 01:24:32 +00:00
8a51d8c8f7
Support pattern matching various x86 sse shifts.
...
llvm-svn: 62979
2009-01-26 00:52:55 +00:00
80b283c1cd
silence a warning when assertions are disabled.
...
llvm-svn: 62976
2009-01-25 23:08:00 +00:00
692ed0f67d
should have removed the + when manually applying a patch!
...
llvm-svn: 62973
2009-01-25 20:29:34 +00:00
97be2f5840
revert this patch for now, because Codegen does still want to generate SSE code,
...
for example in the case of va-args. XFAIL associated tests.
llvm-svn: 62972
2009-01-25 20:21:24 +00:00
3cc1940003
testcase for llvm-gcc part of PR3402.
...
llvm-svn: 62969
2009-01-25 18:00:06 +00:00
a23c73bbdc
If user explicitly asks not to use SSE, don't force it. This fixes LLVM part of PR3402.
...
llvm-svn: 62967
2009-01-25 17:58:56 +00:00
f1d38be265
Eliminate the loop that searches through each of the operands
...
of each use in the SelectionDAG ReplaceAllUses* functions. Thanks
to Chris for spotting this opportunity.
Also, factor out code from all 5 of the ReplaceAllUses* functions
into AddNonLeafNodeToCSEMaps, which is now renamed
AddModifiedNodeToCSEMaps to more accurately reflect its purpose.
llvm-svn: 62964
2009-01-25 16:29:12 +00:00
3a113ec468
Whitespace tidiments.
...
llvm-svn: 62963
2009-01-25 16:21:38 +00:00
e7b0dde2ee
Move the N->use_empty() assert from DeleteNode to
...
DeleteNodeNotInCSEMaps, since DeleteNode just calls
DeleteNodeNotInCSEMaps.
llvm-svn: 62962
2009-01-25 16:20:37 +00:00
e3f2278de1
Reality-check the FAQ entry for "Can I use LLVM to convert C++ to C?"
...
llvm-svn: 62961
2009-01-25 16:04:50 +00:00
21add8f983
Start generating arbitrary precision integer SCEVs. This removes the temporary
...
code that rounded up and capped the size.
llvm-svn: 62958
2009-01-25 08:16:27 +00:00
cb7a10ab63
Actually run the test in this directory.
...
llvm-svn: 62957
2009-01-25 08:05:07 +00:00
5647c5d1a4
The function that does nothing but call malloc is noalias return.
...
llvm-svn: 62956
2009-01-25 07:59:57 +00:00
1c7c019229
Private linkage support for PPC / Darwin.
...
llvm-svn: 62955
2009-01-25 06:32:01 +00:00
abda665f5f
Teach 2addr pass to be do more commuting. If both uses of a two-address instruction are killed, but the first operand has a use before and after the def, commute if the second operand does not suffer from the same issue.
...
%reg1028<def> = EXTRACT_SUBREG %reg1027<kill>, 1
%reg1029<def> = MOV8rr %reg1028
%reg1029<def> = SHR8ri %reg1029, 7, %EFLAGS<imp-def,dead>
insert => %reg1030<def> = MOV8rr %reg1028
%reg1030<def> = ADD8rr %reg1028<kill>, %reg1029<kill>, %EFLAGS<imp-def,dead>
In this case, it might not be possible to coalesce the second MOV8rr
instruction if the first one is coalesced. So it would be profitable to
commute it:
%reg1028<def> = EXTRACT_SUBREG %reg1027<kill>, 1
%reg1029<def> = MOV8rr %reg1028
%reg1029<def> = SHR8ri %reg1029, 7, %EFLAGS<imp-def,dead>
insert => %reg1030<def> = MOV8rr %reg1029
%reg1030<def> = ADD8rr %reg1029<kill>, %reg1028<kill>, %EFLAGS<imp-def,dead>
llvm-svn: 62954
2009-01-25 03:53:59 +00:00
b09b0242ca
Fix an indent and a typo.
...
llvm-svn: 62940
2009-01-24 22:12:48 +00:00
2b3389a626
Revert previous change; even this mild and clearly
...
more accurate change loses more than it gains on
benchmarks.
llvm-svn: 62938
2009-01-24 21:49:34 +00:00
3cedd4dc64
add note about possible GEP improvement with fields of size 0.
...
llvm-svn: 62925
2009-01-24 19:30:25 +00:00
73fd5a5651
Add a PR comment to this test.
...
llvm-svn: 62921
2009-01-24 17:32:54 +00:00
f4395ea97a
testcase for PR3381.
...
Also it was an empty struct, not a void after all.
llvm-svn: 62920
2009-01-24 17:16:04 +00:00
142ccc0e15
Add more documentation mentioning the limitations due to PR2660.
...
llvm-svn: 62919
2009-01-24 15:58:40 +00:00
73ff92272f
void* is represented as pointer to empty struct {}.
...
Thus we need to check whether the struct is empty before trying to index into
it. This fixes PR3381.
llvm-svn: 62918
2009-01-24 11:30:49 +00:00
ee34b8f179
Some cleanups. No functional changes.
...
llvm-svn: 62917
2009-01-24 10:07:43 +00:00
899ecdbbba
Improve the inlining cost function a bit.
...
Little practical effect.
llvm-svn: 62908
2009-01-24 01:27:33 +00:00
72cd68fe64
Make InstCombineStoreToCast handle aggregates more aggressively,
...
handling the case in Transforms/InstCombine/cast-store-gep.ll, which
is a heavily reduced testcase from Clang on x86-64.
llvm-svn: 62904
2009-01-24 01:00:13 +00:00
82221d62a9
Perform optional clang building.
...
llvm-svn: 62895
2009-01-24 00:00:41 +00:00
6d8970695b
Refactor code. No functionality change.
...
llvm-svn: 62893
2009-01-23 23:27:33 +00:00
ccfacfff9a
Introduce two DWARF attribute extentions DW_AT_APPLE_optimized, DW_AT_APPLE_flags.
...
DW_AT_APPLE_optimized flag is set when a compile_unit is optimized. The debugger takes advantage of this information some way.
DW_AT_APPLE_flags encodes command line options when certain env. variable is set. This is used by build engineers to track various gcc command lines used by by a project, irrespective of whether the project used makefile, Xcode or something else.
llvm-gcc patch is next.
llvm-svn: 62888
2009-01-23 22:33:47 +00:00
832959536a
hopefully address PR3379 by making the P modifier work in x86 inline asm.
...
llvm-svn: 62887
2009-01-23 22:33:40 +00:00
486d309b34
Fix test case. Use valid file name and directory in global variable's debug info entry.
...
llvm-svn: 62883
2009-01-23 21:54:18 +00:00
59c431347f
use CallSite::isCalle instead of slow getOperandNo
...
llvm-svn: 62877
2009-01-23 21:17:04 +00:00
eb61fcf2a1
Simplify the logic of getting hold of a PHI predecessor block.
...
There is now a direct way from value-use-iterator to incoming block in PHINode's API.
This way we avoid the iterator->index->iterator trip, and especially the costly
getOperandNo() invocation. Additionally there is now an assertion that the iterator
really refers to one of the PHI's Uses.
llvm-svn: 62869
2009-01-23 19:40:15 +00:00
617001d842
Add support for deleting a module provider from a JIT in such a way that it does not cause the owned module to be fully materialized.
...
llvm-svn: 62864
2009-01-23 19:27:28 +00:00
dd7aa37895
Empty DIType represents void. In this case no need to construct any type DIE.
...
llvm-svn: 62861
2009-01-23 19:13:31 +00:00
1275e28ded
Fold x-0 to x in unsafe-fp-math mode. This comes up in the
...
testcase from PR3376, and in fact is sufficient to completely
avoid the problem in that testcase.
There's an underlying problem though; TLI.isOperationLegal
considers Custom to be Legal, which might be ok in some
cases, but that's what DAGCombiner is using in many places
to test if something is legal when LegalOperations is true.
When DAGCombiner is running after legalize, this isn't
sufficient. I'll address this in a separate commit.
llvm-svn: 62860
2009-01-23 19:10:37 +00:00
d4aae43e79
Patch to build llvm with srcdir != objdir from
...
llvm-top. By Michael Schuerig, with some small
tweaks to make the "don't configure if already
configured" logic work.
llvm-svn: 62842
2009-01-23 08:42:38 +00:00
3f4591c89f
fix two more cases where we could let the NLPDI cache get unsorted.
...
With this, sqlite3 now passes.
llvm-svn: 62839
2009-01-23 07:12:16 +00:00
e3ea48c71e
Unconditionally reset 'cache' to zero, even if we don't need to resort it.
...
This avoids using a dangling pointer.
Reset NumSortedEntries after restoring Cache to avoid extraneous sorts.
This fixes the reduced sqlite3 testcase, but apparently not the whole app.
llvm-svn: 62838
2009-01-23 06:48:41 +00:00
9db227999a
Only check if coalescing is worthwhile when the result is targeting a more restrictive register class.
...
llvm-svn: 62837
2009-01-23 05:48:59 +00:00
f347c3615b
Update test to reflect command line option name change.
...
llvm-svn: 62836
2009-01-23 05:45:31 +00:00
a2ca64a39a
Stage two of fixing pre-alloc-splitting's code size issues: filter out restores that are just
...
going to be re-spilled again.
This also helps performance. Pre-alloc-splitting now seems to be an overall win on SPEC.
llvm-svn: 62834
2009-01-23 03:28:53 +00:00
ab7bf14fb4
Cross register class coalescing. Not yet enabled.
...
llvm-svn: 62832
2009-01-23 02:15:19 +00:00
7d537c23c1
Code did not follow associated comment. not a good idea.
...
llvm-svn: 62828
2009-01-23 01:25:58 +00:00
4c8f55abda
Set appropriate tag for the composite type.
...
llvm-svn: 62827
2009-01-23 01:19:09 +00:00
706d40e662
a minor tweak to my previous patch, handle the invalidation case
...
when there are multiple iterations of the loop. This fixes PR3375.
llvm-svn: 62822
2009-01-23 00:27:03 +00:00
48abf2380b
Add an initial pass at dead spill/restore removal for pre alloc splitting.
...
llvm-svn: 62821
2009-01-23 00:23:32 +00:00
1f9053a83c
do not sign extend characters input to isprint. This improves
...
compatibility with VC++. Patch by Max Burke!
llvm-svn: 62813
2009-01-22 23:38:45 +00:00
0681693277
Handful of fixes to the VC project files.
...
llvm-svn: 62807
2009-01-22 23:05:37 +00:00
c2dc7ee5d0
Fix a minor bug in DAGCombiner's folding of SELECT. Folding "select C, 0, 1"
...
to "C ^ 1" is only valid when C is known to be either 0 or 1. Most of the
similar foldings in this function only handle "i1" types, but this one appears
intentionally written to handle larger integer types. If C has an integer
type larger than "i1", this needs to check if the high bits of a boolean
are known to be zero. I also changed the comment to describe this folding as
"C ^ 1" instead of "~C", since that is what the code does and since the latter
would only be valid for "i1" types. The good news is that most LLVM targets
use TargetLowering::ZeroOrOneBooleanContent so this change will not disable
the optimization; the bad news is that I've been unable to come up with a
testcase to demonstrate the problem.
I have also removed a "FIXME" comment for folding "select C, X, 0" to "C & X",
since the code looks correct to me. It could be made more aggressive by not
limiting the type to "i1", but that would then require checking for
TargetLowering::ZeroOrNegativeOneBooleanContent. Similar changes could be
done for the other SELECT foldings, but it was decided to be not worth the
trouble and complexity (see e.g., r44663).
llvm-svn: 62790
2009-01-22 22:05:48 +00:00
1f3411de47
Don't create ISD::FNEG nodes after legalize if they aren't legal.
...
Simplify x+0 to x in unsafe-fp-math mode. This avoids a bunch of
redundant work in many cases, because in unsafe-fp-math mode,
ISD::FADD with a constant is considered free to negate, so the
DAGCombiner often negates x+0 to -0-x thinking it's free, when
in reality the end result is -x, which is more expensive than x.
Also, combine x*0 to 0.
This fixes PR3374.
llvm-svn: 62789
2009-01-22 21:58:43 +00:00
f4013373cd
introduce a useful abstraction to find out if a Use is in the call position of an instruction
...
llvm-svn: 62788
2009-01-22 21:35:57 +00:00
483e12e09e
Regenerated configure after backing out 62553 and r62616.
...
llvm-svn: 62778
2009-01-22 20:11:17 +00:00
9a8602ce50
Revert r62553 and r62616 due to issues with portability.
...
llvm-svn: 62777
2009-01-22 20:09:20 +00:00
aa1212271a
rename methods in System/Host to be more consistent.
...
llvm-svn: 62776
2009-01-22 19:53:00 +00:00
dec7fe2e71
Do not use buggy llvm-gcc to generate testcases.
...
llvm-svn: 62770
2009-01-22 18:28:11 +00:00
c58900504b
Add SelectionDAG::getNOT method to construct bitwise NOT operations,
...
corresponding to the "not" and "vnot" PatFrags. Use the new method
in some places where it seems appropriate.
llvm-svn: 62768
2009-01-22 17:39:32 +00:00
e3a26635fb
Remove no-longer relevant comment. Pointed out
...
by Gabor.
llvm-svn: 62765
2009-01-22 15:37:29 +00:00
ac6f7eeb05
This passes on linux.
...
llvm-svn: 62764
2009-01-22 15:07:15 +00:00
4186ddfc5a
Few targets like the tiny little PIC16 have only 16-bit pointers.
...
llvm-svn: 62763
2009-01-22 10:14:21 +00:00
4a0bf66eb8
Eliminate a couple of fields from TargetRegisterClass: SubRegClasses and SuperRegClasses. These are not necessary. Also eliminate getSubRegisterRegClass and getSuperRegisterRegClass. These are slow and their results can change if register file names change. Just use TargetLowering::getRegClassFor() to get the right TargetRegisterClass instead.
...
llvm-svn: 62762
2009-01-22 09:10:11 +00:00
e09d631d8e
fix a typo
...
llvm-svn: 62761
2009-01-22 07:21:55 +00:00
705ac7082b
add a note
...
llvm-svn: 62760
2009-01-22 07:16:03 +00:00
bed6be62e4
fix a testcase.
...
llvm-svn: 62758
2009-01-22 07:08:58 +00:00
f09619d533
Fix PR3358, a really nasty bug where recursive phi translated
...
analyses could be run without the caches properly sorted. This
can fix all sorts of weirdness. Many thanks to Bill for coming
up with the 'issorted' verification idea.
llvm-svn: 62757
2009-01-22 07:04:01 +00:00
c4b1d8295d
Bump to 2.6svn.
...
Regenerate configure (last regen was with the wrong version).
llvm-svn: 62751
2009-01-22 05:17:59 +00:00
9fc809e980
Make tblgen more portable, allowing it to build with ICC.
...
Patch by Robert Zeh!
llvm-svn: 62750
2009-01-22 05:10:16 +00:00
f7a16e9380
Don't create .dir files in installation directories. Switch to using
...
order-only dependancies for installation directories.
llvm-svn: 62746
2009-01-22 03:24:22 +00:00
8d176ed95b
The operator<() and operator>() were reversing their tests. Have the test the correct way.
...
llvm-svn: 62745
2009-01-22 03:05:10 +00:00
08bebedbab
Get rid of warning about implicit 64-to-32 bit conversions.
...
llvm-svn: 62741
2009-01-22 02:10:33 +00:00
a5127abeec
Add a few more notes about LLVM IR features that codegen doesn't
...
yet support.
llvm-svn: 62739
2009-01-22 01:39:38 +00:00
244d369d53
CMake: Tests PERL availability with PERL_EXECUTABLE variable.
...
On MSVC 64bits, does not put underscore before the symbol name on the
/INCLUDE linker parameter.
llvm-svn: 62738
2009-01-22 01:20:31 +00:00
1019bdef0c
Add explicit this-> (to make the VS compiler happy).
...
llvm-svn: 62731
2009-01-21 23:49:23 +00:00
31dd016f81
Recognize inline asm for bswap on x86-64 GLIBC. This allows it
...
to be supported in the JIT.
llvm-svn: 62730
2009-01-21 23:40:54 +00:00
3041abd608
Fix ilist two-phase name lookup
...
llvm-svn: 62724
2009-01-21 22:38:44 +00:00
d776b6e51d
Fix a missing word.
...
llvm-svn: 62720
2009-01-21 21:54:44 +00:00
01d7dd8607
Versions of VIM included with Intrepid and Leopard at least appear
...
to handle symlinks just fine, so reword the instructions in the
README accordingly.
llvm-svn: 62719
2009-01-21 21:52:42 +00:00
3155533003
Enable syntax highlighting of LLVM and tablegen files by default,
...
so that users don't have to copy text from the README to get this.
llvm-svn: 62718
2009-01-21 21:47:51 +00:00
455be5a0a0
Only set cindent for C and C++ source files.
...
llvm-svn: 62717
2009-01-21 21:30:25 +00:00
6cf1f8fd5b
Now with RUN line.
...
llvm-svn: 62716
2009-01-21 21:28:03 +00:00
ba11cd338b
Run this through -simplifycfg and -mem2reg to test only what we need to test.
...
llvm-svn: 62714
2009-01-21 21:02:27 +00:00
922bf52e8e
Add a comment to SelectionDAG::ReplaceAllUsesWith to describe a subtle
...
iteraction with SelectionDAG CSE.
llvm-svn: 62713
2009-01-21 20:50:09 +00:00
1f86498f93
Do not use host floating point types when emitting
...
ASCII IR; loading and storing these can change the
bits of NaNs on some hosts. Remove or add warnings
at a few other places using host floating point;
this is a bad thing to do in general.
llvm-svn: 62712
2009-01-21 20:32:55 +00:00
ec5eb161fd
Also favors NOT64r.
...
llvm-svn: 62710
2009-01-21 19:45:31 +00:00
a326520190
fix warning in release-asserts mode and spelling of assert.
...
llvm-svn: 62699
2009-01-21 18:38:18 +00:00
41b43da217
add getPointerToGlobal to the C bindings, patch by Lennart Augustsson!
...
PR3364
llvm-svn: 62697
2009-01-21 18:11:10 +00:00
77527f5812
Remove uses of uint32_t in favor of 'unsigned' for better
...
compatibility with cygwin. Patch by Jay Foad!
llvm-svn: 62695
2009-01-21 18:09:24 +00:00
7e6b932f18
Simplify ReduceLoadWidth's logic: it doesn't need several different
...
special cases after producing the new reduced-width load, because the
new load already has the needed adjustments built into it. This fixes
several bugs due to the special cases, including PR3317.
llvm-svn: 62692
2009-01-21 15:17:51 +00:00
b43c8996f2
Fix a recent regression. ClrOpcode is not set for i8; for i8, if
...
we want to clear %ah to zero before a division, just use a
zero-extending mov to %al. This fixes PR3366.
llvm-svn: 62691
2009-01-21 14:50:16 +00:00
6a4f729dc9
Mimic gcc behaviour with regard to response files.
...
llvm-svn: 62688
2009-01-21 13:14:02 +00:00
1a0c083748
Fix 'llvm-config --libs' output.
...
Change the naming scheme for llvmc plugins so that they do not appear in
'llvm-config --libs' output.
llvm-svn: 62687
2009-01-21 13:05:00 +00:00
fbc89cecad
Change the hook API back to prevent memory leaks.
...
llvm-svn: 62686
2009-01-21 13:04:33 +00:00
bf9716e15d
Allow hooks with arguments.
...
llvm-svn: 62685
2009-01-21 13:04:00 +00:00
d56cf3025f
This was causing invalid memory accesses when
...
generating debug info in the compiler.
llvm-svn: 62684
2009-01-21 11:51:17 +00:00
1de451d0d0
Let's try to have our cake and eat it to: move
...
this test into FrontendC to ensure that llvm-gcc
is available; assemble using "llvm-gcc -xassembler"
rather than "as".
llvm-svn: 62683
2009-01-21 11:37:31 +00:00
696f4a8598
Don't rely on grep -w working.
...
llvm-svn: 62682
2009-01-21 09:41:42 +00:00
335ea6cf2a
Fixed build warnings. Restoring changes done in 62600, they were lost in 62655.
...
llvm-svn: 62681
2009-01-21 09:02:46 +00:00
be7e41481b
Cleanup whitespace and comments, and tweak some
...
prototypes, in operand type legalization. No
functionality change.
llvm-svn: 62680
2009-01-21 09:00:29 +00:00
fe48dfa0f2
I accidentally removed this check in an earlier commit, which cause breakage in the pre alloc splitter.
...
llvm-svn: 62678
2009-01-21 08:18:03 +00:00
f737337707
Implement LowerOperationWrapper for legalizer.
...
Also a few signed comparison fixes.
llvm-svn: 62665
2009-01-21 05:44:05 +00:00
ed7d79fce4
CellSPU:
...
- Ensure that (operation) legalization emits proper FDIV libcall when needed.
- Fix various bugs encountered during llvm-spu-gcc build, along with various
cleanups.
- Start supporting double precision comparisons for remaining libgcc2 build.
Discovered interesting DAGCombiner feature, which is currently solved via
custom lowering (64-bit constants are not legal on CellSPU, but DAGCombiner
insists on inserting one anyway.)
- Update README.
llvm-svn: 62664
2009-01-21 04:58:48 +00:00
a70798cc9a
Allow targets to legalize operations (with illegal operands) that produces multiple values. For example, a load with an illegal operand (a load produces two values, a value and chain).
...
llvm-svn: 62663
2009-01-21 04:48:39 +00:00
201501995f
Favors generating "not" over "xor -1". For example.
...
unsigned test(unsigned a) {
return ~a;
}
llvm used to generate:
movl $4294967295, %eax
xorl 4(%esp), %eax
Now it generates:
movl 4(%esp), %eax
notl %eax
It's 3 bytes shorter.
llvm-svn: 62661
2009-01-21 02:09:05 +00:00
287b4bc44e
Disable on x86_64 until I figure out what's wrong.
...
llvm-svn: 62660
2009-01-21 02:08:30 +00:00
b5721632ee
Make special cases (0 inf nan) work for frem.
...
Besides APFloat, this involved removing code
from two places that thought they knew the
result of frem(0., x) but were wrong.
llvm-svn: 62645
2009-01-21 00:35:19 +00:00
be7a29de0b
Be more aggressive about renumbering vregs after splitting them.
...
llvm-svn: 62639
2009-01-21 00:13:28 +00:00
19f2dd794e
Encode member accessibility.
...
llvm-svn: 62638
2009-01-21 00:08:04 +00:00
6bbacbe372
Appropriately mark fowrad decls.
...
llvm-svn: 62625
2009-01-20 22:27:02 +00:00
6fbec1c230
Need compile unit to find location.
...
llvm-svn: 62624
2009-01-20 22:26:11 +00:00
e75fdb0510
Calls to fmod, it turns out, are constant-folded by
...
invoking the host fmod, not by lowering to frem and
constant-folding that. Fix this so it tests what I
want to test.
llvm-svn: 62622
2009-01-20 21:58:13 +00:00
f8a8c13c1e
Don't bother running the assembler, we don't know that it will be configured
...
for whatever llc defaults to. This fixes PR3363
llvm-svn: 62619
2009-01-20 21:41:53 +00:00
f1e873a221
Fix PR3243: a LiveVariables bug. When HandlePhysRegKill is checking whether the last reference is also the last def (i.e. dead def), it should also check if last reference is the current machine instruction being processed. This can happen when it is processing a physical register use and setting the current machine instruction as sub-register's last ref.
...
llvm-svn: 62617
2009-01-20 21:25:12 +00:00
363501d09b
Fix typo. Patch by Alexei Svitkine.
...
llvm-svn: 62616
2009-01-20 21:20:23 +00:00
2395916c87
Use "SINT_TO_FP" instead of "UINT_TO_FP" when getting the exponent. This was
...
causing the limited precision stuff to produce the wrong result for values in
the range [0, 1).
llvm-svn: 62615
2009-01-20 21:17:57 +00:00
bd7743d772
Fix struct member's debug info.
...
llvm-svn: 62610
2009-01-20 21:02:02 +00:00
4022b7c3f4
Add test case for PR3154.
...
llvm-svn: 62604
2009-01-20 19:29:54 +00:00
bf7c6a90f5
indentation...
...
llvm-svn: 62603
2009-01-20 19:23:29 +00:00
2333409d06
Need only one set of debug info versions enum.
...
llvm-svn: 62602
2009-01-20 19:22:03 +00:00
c544cb0eca
Change TargetInstrInfo::isMoveInstr to return source and destination sub-register indices as well.
...
llvm-svn: 62600
2009-01-20 19:12:24 +00:00
72854c3811
zap white spaces.
...
llvm-svn: 62598
2009-01-20 19:08:39 +00:00
6befc4d34e
Fix global variable's address in a DIE.
...
llvm-svn: 62596
2009-01-20 18:55:39 +00:00
f0dae1aae3
Enable debug info for enums.
...
llvm-svn: 62594
2009-01-20 18:35:14 +00:00
fe750179ff
Add an IEEE remainder function, which is not
...
fully implemented yet and not used. This is
mainly to clarify that APFloat::mod implements
C fmod, not remainder.
llvm-svn: 62593
2009-01-20 18:35:05 +00:00
17844c7a88
improve compatibility with various versions of graphviz, patch by
...
Patrick Boettcher!
llvm-svn: 62592
2009-01-20 18:25:03 +00:00
ca4881aeca
eliminate use of uint32_t to improve compatibility with cygwin
...
llvm-svn: 62591
2009-01-20 18:23:14 +00:00
4612ae1f8e
Eliminate use of uint32_t to improve compatibility with cygwin
...
llvm-svn: 62590
2009-01-20 18:22:57 +00:00
120e962243
Enable debug info for composite types.
...
llvm-svn: 62589
2009-01-20 18:13:03 +00:00
489c5484d3
Check that the "don't barf on k8" fix is not
...
accidentally reverted again.
llvm-svn: 62587
2009-01-20 18:08:39 +00:00
63f9e249ba
Spacing
...
llvm-svn: 62584
2009-01-20 17:30:40 +00:00
a908b60fb2
Temporarily XFAIL until this can be looked at. r62557 is what caused it to start failing.
...
llvm-svn: 62578
2009-01-20 10:28:39 +00:00
cd3fbfb460
If a vector is empty, you're not allowed to access any
...
elements, even if it is only to take the address. Test:
break-anti-dependencies.ll with ENABLE_EXPENSIVE_CHECKS.
llvm-svn: 62576
2009-01-20 09:05:19 +00:00
6cdcf1eb73
Refactor code. No functionality change.
...
llvm-svn: 62573
2009-01-20 06:44:16 +00:00
1d9c8e5522
Testcase for limited precision stuff.
...
llvm-svn: 62572
2009-01-20 06:23:59 +00:00
786a683441
Shift types need to match.
...
llvm-svn: 62571
2009-01-20 06:10:42 +00:00
c59945b4bd
another fix for PR3354
...
llvm-svn: 62561
2009-01-20 01:15:41 +00:00
83d2e066c1
Add a README entry noticed while investigating PR3216.
...
llvm-svn: 62558
2009-01-20 01:07:33 +00:00
161b7b66ac
Fix a dagcombine to not generate loads of non-round integer types,
...
as its comment says, even in the case where it will be generating
extending loads. This fixes PR3216.
llvm-svn: 62557
2009-01-20 01:06:45 +00:00
af6a3748c7
Do not use DenseMap because the iterator is invalidated while constructing types. After all there was a reason why std::map was used initially!
...
llvm-svn: 62555
2009-01-20 00:58:55 +00:00
c94134dc96
Regenerate.
...
BUILT WITH WRONG VERSION OF AUTOCONF! Somebody please regenerate with an
approved version. Thanks!
llvm-svn: 62554
2009-01-20 00:52:24 +00:00
766ac10452
Make the Interpreter use libffi if it's available. Patch from Alexei Svitkine!
...
This requires a rebuild of 'configure' itself. I will be committing that next, but
built with the wrong version of autoconf. Somebody who has the right one, please update
it.
As a side-note, because of the way autoconf works, all built tools will link against
libffi, not just lli. If you know how to fix this, please let me know ...
llvm-svn: 62553
2009-01-20 00:51:40 +00:00
8f79775a66
Make linear scan's trivial coalescer slightly more aggressive.
...
llvm-svn: 62547
2009-01-20 00:16:18 +00:00
caf1d22243
Doxygen-ify comments.
...
llvm-svn: 62546
2009-01-19 23:43:56 +00:00
44afc82ebe
Verify debug info.
...
llvm-svn: 62545
2009-01-19 23:21:49 +00:00
ea9f1d3c47
Fix a problem exposed by PR3354: simplifycfg was making a potentially
...
trapping instruction be executed unconditionally.
llvm-svn: 62541
2009-01-19 23:03:13 +00:00
534c8a2d72
Remove SDNode's virtual destructor. This makes it impossible for
...
SDNode subclasses to keep state that requires non-trivial
destructors, however it was already effectively impossible,
since the destructor isn't actually ever called. There currently
aren't any SDNode subclasses affected by this, and in general
it's desireable to keep SDNode objects light-weight.
This eliminates the last virtual member function in the SDNode
class, so it eliminates the need for a vtable pointer, making
SDNode smaller.
llvm-svn: 62539
2009-01-19 22:39:36 +00:00
d067ecd1c7
Move & restructure test per review.
...
llvm-svn: 62538
2009-01-19 22:33:12 +00:00
7eeb1cc605
convert this to an unfoldable potentially trapping constant expr.
...
llvm-svn: 62536
2009-01-19 22:12:33 +00:00
73d7fe5a34
improve compatibility with cygwin, patch by Jay Foad!
...
llvm-svn: 62535
2009-01-19 22:00:18 +00:00
dd8ef1bf30
div/rem by zero and div/rem overflow are both undefined according to
...
langref. Constant fold them to undef instead of trying to preserve
the trap. This fixes PR3354.
llvm-svn: 62534
2009-01-19 21:55:26 +00:00
cd0b1bf0a0
Fix SelectionDAG::ReplaceAllUsesWith to behave correctly when
...
uses are added to the From node while it is processing From's
use list, because of automatic local CSE. The fix is to avoid
visiting any new uses.
Fix a few places in the DAGCombiner that assumed that after
a RAUW call, the From node has no users and may be deleted.
This fixes PR3018.
llvm-svn: 62533
2009-01-19 21:44:21 +00:00
6f34e317e9
Fix PR3353, infinitely jump threading an infinite loop make from switches.
...
llvm-svn: 62529
2009-01-19 21:20:34 +00:00
740e98704d
compile-time fmod was done incorrectly. PR 3316.
...
llvm-svn: 62528
2009-01-19 21:17:05 +00:00
db3c05533c
DebugInfo is a lightweight APIs and consumers are expected to use light objects directly. There is no need to support isa<>, dyn_cast<> etc...
...
llvm-svn: 62527
2009-01-19 21:13:39 +00:00
8c8aa2ac29
Verify Intrinsic::dbg_declare.
...
llvm-svn: 62526
2009-01-19 21:00:48 +00:00
eb1e2006de
Add targets to support the installation of clang in isolation.
...
llvm-svn: 62522
2009-01-19 19:48:23 +00:00
44cc554311
DIVREM isel deficiency: If sign bit is known zero, zero out DX/EDX/RDX instead of sign extending the low part (in AX/EAX/RAX) into it.
...
llvm-svn: 62519
2009-01-19 19:06:11 +00:00
0346c04f39
Fix 80 col violations.
...
llvm-svn: 62518
2009-01-19 18:57:29 +00:00
Evan Cheng