Commit Graph

11761 Commits

Author SHA1 Message Date
Dan Gohman
26719acdb7 Add a const.
llvm-svn: 101777
2010-04-19 15:39:27 +00:00
Dan Gohman
f4e99e3f84 These functions don't need to access this.
llvm-svn: 101776
2010-04-19 15:38:38 +00:00
Anton Korobeynikov
e325c693a5 Make processor FUs unique for given itinerary. This extends the limit of 32
FU per CPU arch to 32 per intinerary allowing precise modelling of quite
complex pipelines in the future.

llvm-svn: 101754
2010-04-18 20:31:01 +00:00
Bill Wendling
9218175b5f Add a "PadTo" field to the emitULEB128Bytes method. This will pad out to the
indicated number of bytes.

llvm-svn: 101684
2010-04-18 00:51:49 +00:00
Dale Johannesen
f044ed7cf9 Add comment (lost when reverting and reapplying 101503).
llvm-svn: 101664
2010-04-17 19:56:46 +00:00
Chris Lattner
65d1e40895 teach the x86 asm parser how to handle segment prefixes
in memory operands.  rdar://7874844

llvm-svn: 101661
2010-04-17 18:56:34 +00:00
Chris Lattner
a5798151a3 refactor .if handling code a bit.
llvm-svn: 101659
2010-04-17 18:14:27 +00:00
Dan Gohman
3dde4fe35f Add const qualifiers to TargetLoweringObjectFile usage.
llvm-svn: 101640
2010-04-17 16:44:48 +00:00
Dan Gohman
a0f855157e Use const qualifiers with TargetLowering. This eliminates several
const_casts, and it reinforces the design of the Target classes being
immutable.

SelectionDAGISel::IsLegalToFold is now a static member function, because
PIC16 uses it in an unconventional way. There is more room for API
cleanup here.

And PIC16's AsmPrinter no longer uses TargetLowering.

llvm-svn: 101635
2010-04-17 15:26:15 +00:00
Chris Lattner
bef627e798 reenable r101565, removing a problematic assertion.
CGSCC can delete nodes in regions of the callgraph that
have already been visited.  If new CG nodes are allocated
to the same pointer, we shouldn't abort, just handle it
correctly by assigning a new number.  This should restore
stability by removing invalidated pointers that *will* be
reused from the densemap in the iterator.

llvm-svn: 101628
2010-04-17 07:17:19 +00:00
Evan Cheng
6442d111dd More work to allow dag combiner to promote 16-bit ops to 32-bit.
llvm-svn: 101621
2010-04-17 06:13:15 +00:00
Bob Wilson
ad00f21093 Re-commit my previous SSAUpdater changes. The previous version naively tried
to determine where to place PHIs by iteratively comparing reaching definitions
at each block.  That was just plain wrong.  This version now computes the
dominator tree within the subset of the CFG where PHIs may need to be placed,
and then places the PHIs in the iterated dominance frontier of each definition.
The rest of the patch is mostly the same, with a few more performance
improvements added in.

llvm-svn: 101612
2010-04-17 03:08:24 +00:00
Dale Johannesen
9968eef300 Reapply 101503+101520. These are "obviously correct" [Chris]
and don't cause any problems on Darwin.

llvm-svn: 101584
2010-04-17 00:08:21 +00:00
Eric Christopher
e78496e5f1 Revert 101465, it broke internal OpenGL testing.
Probably the best way to know that all getOperand() calls have been handled
is to replace that API instead of updating.

llvm-svn: 101579
2010-04-16 23:37:20 +00:00
Devang Patel
1929d78e71 Add support to emit dwarf ranges.
llvm-svn: 101575
2010-04-16 23:33:45 +00:00
Dan Gohman
b03b56ff82 Add a getSelectionDAGInfo member to TargetMachine.
llvm-svn: 101567
2010-04-16 23:09:10 +00:00
Chris Lattner
d28dcb3ffe building on the new CallGraphSCC abstraction, teach CallGraphSCCPassManager
to keep the node entries in scc_iterator up to date instead of dangling as
the SCC mutates.

This is a really terrible problem which was causing -g to affect codegen 
because it would permute the memory image of the compiler process.

Thanks to Dale for expertly hunting it down.

llvm-svn: 101565
2010-04-16 23:04:30 +00:00
Chris Lattner
a9a7c28cd2 move ReplaceNode out of line, rename scc_iterator::fini -> isAtEnd().
No functionality change.

llvm-svn: 101562
2010-04-16 22:59:24 +00:00
Chris Lattner
6a038be777 introduce a new CallGraphSCC class, and pass it around
to CallGraphSCCPass's instead of passing around a
std::vector<CallGraphNode*>.  No functionality change,
but now we have a much tidier interface.

llvm-svn: 101558
2010-04-16 22:42:17 +00:00
Chris Lattner
7e6e8651b7 tidy up
llvm-svn: 101548
2010-04-16 21:57:10 +00:00
Dan Gohman
b3afc2a215 Create a new TargetSelectionDAGInfo class. This will eventually acquire
SelectionDAG-specific parts of TargetLowering.

llvm-svn: 101537
2010-04-16 21:12:11 +00:00
Dan Gohman
1119e6060b Eliminate an unnecessary SelectionDAG dependency in getOptimalMemOpType.
llvm-svn: 101531
2010-04-16 20:11:05 +00:00
Dan Gohman
56bc819d8e Fix this code to avoid implicit assumptions about the length of the array.
llvm-svn: 101530
2010-04-16 20:08:45 +00:00
Dan Gohman
2520b6e199 Revert 101520, which depended on 101503, which was reverted.
llvm-svn: 101528
2010-04-16 19:36:41 +00:00
Duncan Sands
7fa5f38be8 Revert commit 101503 (johannes), in the hope of fixing the dragonegg build,
see http://google1.osuosl.org:8011/builders/dragonegg-x86_64-linux/builds/693
Original commit text:
Use a ValueMap not a std::map for the reason indicated
in the comment.  This was causing nondeterministic changes
in inlining decisions.

llvm-svn: 101525
2010-04-16 19:28:59 +00:00
Dan Gohman
5d99e7df1c Trim a #include.
llvm-svn: 101520
2010-04-16 18:45:11 +00:00
Dale Johannesen
759091323b Use a ValueMap not a std::map for the reason indicated
in the comment.  This was causing nondeterministic changes
in inlining decisions.

llvm-svn: 101503
2010-04-16 18:16:08 +00:00
Gabor Greif
e7d6812008 reapply r101434
with a fix for self-hosting

rotate CallInst operands, i.e. move callee to the back
of the operand array

the motivation for this patch are laid out in my mail to llvm-commits:
more efficient access to operands and callee, faster callgraph-construction,
smaller compiler binary

llvm-svn: 101465
2010-04-16 15:33:14 +00:00
Dan Gohman
f754ff8ee9 Make callIsSmall accessible as a utility function.
llvm-svn: 101463
2010-04-16 15:14:50 +00:00
Bill Wendling
7ede9396f4 The JIT calls TidyLandingPads to tidy up the landing pads. However, because the
JIT doesn't use the MC back-end asm printer to emit labels that it uses, the
section for the MCSymbol is never set. And thus the MCSymbol for the EH label
isn't marked as "defined". Because of that, TidyLandingPads removes the needed
landing pads from the JIT output. This breaks EH for every JIT program.

This is a work-around for this limitation. We pass in the label locations
map. If the label has a non-zero value, then it was "emitted" by the JIT and
TidyLandingPads shouldn't remove that label.

A nicer solution would be to mark the MCSymbol as "used" by the JIT and not rely
upon the section being set to determine if it's defined or not.

llvm-svn: 101453
2010-04-16 08:46:10 +00:00
Evan Cheng
d143bfe0a4 Adding support for dag combiner to promote operations for profit. This requires target specific queries. For example, x86 should promote i16 to i32 when it does not impact load folding.
x86 support is off by default. It can be enabled with -promote-16bit.

Work in progress.

llvm-svn: 101448
2010-04-16 06:14:10 +00:00
Eric Christopher
033524af28 Add a check for posix_spawn.
Regenerate configure and other autoconf files.

llvm-svn: 101444
2010-04-16 05:14:21 +00:00
Gabor Greif
cd116e8c6a back out r101423 and r101397, they break llvm-gcc self-host on darwin10
llvm-svn: 101434
2010-04-16 01:16:20 +00:00
Gabor Greif
2e18d34d80 reapply r101364, which has been backed out in r101368
with a fix

rotate CallInst operands, i.e. move callee to the back
of the operand array

the motivation for this patch are laid out in my mail to llvm-commits:
more efficient access to operands and callee, faster callgraph-construction,
smaller compiler binary

llvm-svn: 101397
2010-04-15 20:51:13 +00:00
Dan Gohman
034c0f7354 Make getPredecessorWithUniqueSuccessorForBB return the unique successor
in addition to the predecessor.

llvm-svn: 101374
2010-04-15 16:19:08 +00:00
Gabor Greif
6022150477 back out r101364, as it trips the linux nightlybot on some clang C++ tests
llvm-svn: 101368
2010-04-15 12:46:56 +00:00
Gabor Greif
428ca23bbd rotate CallInst operands, i.e. move callee to the back
of the operand array

the motivation for this patch are laid out in my mail to llvm-commits:
more efficient access to operands and callee, faster callgraph-construction,
smaller compiler binary

llvm-svn: 101364
2010-04-15 10:49:53 +00:00
Dan Gohman
64a84ceb3f Add more const qualifiers for LLVM IR pointers in CodeGen.
llvm-svn: 101342
2010-04-15 04:33:49 +00:00
Dan Gohman
0e0b8cf9fd Add const qualifiers to CodeGen's use of LLVM IR constructs.
llvm-svn: 101334
2010-04-15 01:51:59 +00:00
Dan Gohman
d9d73a414c Fix TrackingVH to handle const Value subclasses.
llvm-svn: 101300
2010-04-14 22:24:25 +00:00
Dan Gohman
a5a1f2f8d0 Remove a const here. This makes this function consistent with all the
other getOperand wrappers, and it makes it easier to use with DebugInfo
code, which isn't currently prepared to see const MDNode *.

llvm-svn: 101299
2010-04-14 22:23:05 +00:00
Dan Gohman
5c72d507d7 Constify GetConstantStringInfo.
llvm-svn: 101298
2010-04-14 22:20:45 +00:00
Dan Gohman
2c2f455dfa Move a bunch of methods from CallSite to CallSiteBase, so that they can
be used in ImmutableCallSite too.

llvm-svn: 101292
2010-04-14 21:47:32 +00:00
Dan Gohman
16ea461eb9 Use FunTy instead of hard-coding Function.
llvm-svn: 101283
2010-04-14 20:49:44 +00:00
Dan Gohman
94f168ddbe Oops, make these public.
llvm-svn: 101280
2010-04-14 20:31:28 +00:00
Dan Gohman
eeaee6bbe3 Move getType() and getCaller() into CallSiteBase so that
ImmutableCallSite can use them too.

llvm-svn: 101279
2010-04-14 20:28:44 +00:00
Dan Gohman
1b67547226 Delete unneeeded arguments.
llvm-svn: 101276
2010-04-14 20:17:22 +00:00
Dan Gohman
d5d199da5f Delete unused arguments.
llvm-svn: 101275
2010-04-14 20:05:00 +00:00
Dan Gohman
c64d1d02a3 Factor out EH landing pad code into a separate function, and constify
a bunch of stuff to support it.

llvm-svn: 101273
2010-04-14 19:53:31 +00:00
Dan Gohman
9ed196804f Fix a missing #include.
llvm-svn: 101270
2010-04-14 18:57:18 +00:00