1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-25 14:02:52 +02:00
Commit Graph

1045 Commits

Author SHA1 Message Date
Dale Johannesen
fa28929927 Reapply 63765. Patches for clang and llvm-gcc to follow.
llvm-svn: 63812
2009-02-05 01:49:45 +00:00
Dale Johannesen
d27bc65e74 Remove non-DebugLoc forms of CopyToReg and CopyFromReg.
Adjust callers.

llvm-svn: 63789
2009-02-04 23:02:30 +00:00
Dale Johannesen
f6e1822ccd Reverting 63765. This broke the build of both clang
and llvm-gcc.

llvm-svn: 63786
2009-02-04 22:47:25 +00:00
Nate Begeman
66f10b55ed New feature: add support for target intrinsics being defined in the
target directories themselves.  This also means that VMCore no longer
needs to know about every target's list of intrinsics.  Future work
will include converting the PowerPC target to this interface as an
example implementation.

llvm-svn: 63765
2009-02-04 19:47:21 +00:00
Duncan Sands
cac6cf74f9 Fix PR3453 and probably a bunch of other potential
crashes or wrong code with codegen of large integers:
eliminate the legacy getIntegerVTBitMask and
getIntegerVTSignBit methods, which returned their
value as a uint64_t, so couldn't handle huge types.

llvm-svn: 63494
2009-02-01 18:06:53 +00:00
Bill Wendling
da5e35aa5f Explain why this is here.
llvm-svn: 63342
2009-01-29 23:19:43 +00:00
Bill Wendling
c65110ba7e - 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
Dan Gohman
21a780e19c 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
Mikhail Glushenkov
cd3b13fc7b Typo.
llvm-svn: 63174
2009-01-28 03:47:58 +00:00
Mikhail Glushenkov
920118470d 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
Evan Cheng
c971801ae1 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
Chris Lattner
fb1a7b3501 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
Mikhail Glushenkov
bf9e875d8e Change the hook API back to prevent memory leaks.
llvm-svn: 62686
2009-01-21 13:04:33 +00:00
Mikhail Glushenkov
9153777db7 Allow hooks with arguments.
llvm-svn: 62685
2009-01-21 13:04:00 +00:00
Dan Gohman
89103ff4d5 Avoid triggering an assertion failure when an instruction pattern
is a leaf node. Patch by Brandner!

llvm-svn: 62361
2009-01-16 21:30:55 +00:00
Dan Gohman
1810c6366d Add support for instructions with multiple ComplexPatterns, by
adding more information to the temporary variables names so that
they don't conflict.

llvm-svn: 62296
2009-01-16 02:05:52 +00:00
Duncan Sands
bcdbfb63dc Rename getABITypeSize to getTypePaddedSize, as
suggested by Chris.

llvm-svn: 62099
2009-01-12 20:38:59 +00:00
Chris Lattner
ed4ee47a0d make tblgen autogenerate the nocapture intrinsics for
llvm.memcpy/memset/memmove.  This allows removal of some 
hackish code from basicaa.

llvm-svn: 62071
2009-01-12 02:41:37 +00:00
Chris Lattner
c26f718218 add scaffolding to emit argument attributes. No functionality
change.

llvm-svn: 62067
2009-01-12 01:27:55 +00:00
Chris Lattner
a844423a1b make tblgen emit the entire Intrinsic::getAttributes method,
not a random piece of it.  No functionality change.

llvm-svn: 62066
2009-01-12 01:18:58 +00:00
Chris Lattner
c31c59a3be add nocapture attribute to llvm.mem* intrinsics and have tblgen
parse them.  tblgen doesn't yet do anything with this info though.

llvm-svn: 62065
2009-01-12 01:12:03 +00:00
Bob Wilson
98fdf706f7 Improve support for type-generic vector intrinsics by teaching TableGen how
to handle LLVMMatchType intrinsic parameters, and by adding new subclasses
of LLVMMatchType to match vector types with integral elements that are 
either twice as wide or half as wide as the elements of the matched type.

llvm-svn: 61834
2009-01-07 00:09:01 +00:00
Dan Gohman
7934babf16 Delete an unused variable and simplify the code.
llvm-svn: 61732
2009-01-05 19:31:28 +00:00
Bob Wilson
50a8e4457d Handle iAny and fAny types in TreePatternNode::UpdateNodeType.
llvm-svn: 61713
2009-01-05 17:52:54 +00:00
Bob Wilson
265ddb0b65 Fix spelling in some comments.
llvm-svn: 61702
2009-01-05 17:23:09 +00:00
Dan Gohman
84db870cc3 Use dyn_cast intead of isa + cast in the generated DAGISel code. This
reduces the amount of code slightly when assertions are enabled.

llvm-svn: 61249
2008-12-19 18:13:39 +00:00
Mon P Wang
84ad2a383d Added support for vector widening.
llvm-svn: 61209
2008-12-18 20:03:17 +00:00
Mikhail Glushenkov
964b643cf8 Fix typo in error message.
llvm-svn: 61191
2008-12-18 04:06:58 +00:00
Mikhail Glushenkov
90baa08c2b Some enhancements for the 'case' expression.
Add (error) and (empty).

llvm-svn: 61117
2008-12-17 02:47:01 +00:00
Mikhail Glushenkov
e50ddcc42a Put Edge* classes into anonymous namespace.
Prevents conflicts between plugins.

llvm-svn: 60871
2008-12-11 10:34:18 +00:00
Mikhail Glushenkov
c75a4df77c Make 'extern' an option property.
Makes (forward) work better.

llvm-svn: 60667
2008-12-07 16:47:12 +00:00
Mikhail Glushenkov
e0c9f63882 Better error message.
llvm-svn: 60664
2008-12-07 16:45:12 +00:00
Mikhail Glushenkov
85098f0f90 Re-apply Cedric's changes.
Use B instead of Beg (for consistency), but NodeA and NodeB instead of A
and B.

llvm-svn: 60663
2008-12-07 16:44:47 +00:00
Mikhail Glushenkov
4522fd9b8a Try to guess when the auto-generated cl::Sink option should be marked 'extern'.
This would be much easier to do if the CommandLine library didn't use
global state. Global state is evil.

llvm-svn: 60659
2008-12-07 16:42:47 +00:00
Mikhail Glushenkov
1348ac2185 Add a (progn)-like construct for (actions). Implemented as a DAG list.
llvm-svn: 60658
2008-12-07 16:42:22 +00:00
Mikhail Glushenkov
f01331152a Use (actions) instead of option properties, support external options.
Also includes a major refactoring. See documentation for more
information.

llvm-svn: 60656
2008-12-07 16:41:11 +00:00
Cedric Venet
c2986f3f3a The use of the construct:
for(Type1 B = ...;;) { Type2 B ; ... }
is bad: code is hard to read and VS VS don't like it (it ignore the second declaration of B).
This patch fix the problem in tablegen. Please don't write code like this. 

llvm-svn: 60590
2008-12-05 13:37:30 +00:00
Dan Gohman
5dad0993a9 Rename isSimpleLoad to canFoldAsLoad, to better reflect its meaning.
llvm-svn: 60487
2008-12-03 18:15:48 +00:00
Dan Gohman
ac6561793c Mark x86's V_SET0 and V_SETALLONES with isSimpleLoad, and teach X86's
foldMemoryOperand how to "fold" them, by converting them into constant-pool
loads. When they aren't folded, they use xorps/cmpeqd, but for example when
register pressure is high, they may now be folded as memory operands, which
reduces register pressure.

Also, mark V_SET0 isAsCheapAsAMove so that two-address-elimination will
remat it instead of copying zeros around (V_SETALLONES was already marked).

llvm-svn: 60461
2008-12-03 05:21:24 +00:00
Dan Gohman
6333d48459 Add a sanity-check to tablegen to catch the case where isSimpleLoad
is set but mayLoad is not set. Fix all the problems this turned up.

Change code to not use isSimpleLoad instead of mayLoad unless it
really wants isSimpleLoad.

llvm-svn: 60459
2008-12-03 02:30:17 +00:00
Mikhail Glushenkov
cd57fbc9c4 Support multiple compilation graph definitions. Not terribly useful, but makes the code more generic.
llvm-svn: 60199
2008-11-28 00:13:47 +00:00
Mikhail Glushenkov
187b171022 Add 'hidden' and 'really_hidden' option properties.
llvm-svn: 60198
2008-11-28 00:13:25 +00:00
Evan Cheng
f18016728c On x86 favors folding short immediate into some arithmetic operations (e.g. add, and, xor, etc.) because materializing an immediate in a register is expensive in turns of code size.
e.g.
movl 4(%esp), %eax
addl $4, %eax

is 2 bytes shorter than

movl $4, %eax
addl 4(%esp), %eax

llvm-svn: 60139
2008-11-27 00:49:46 +00:00
Mikhail Glushenkov
7534e81f38 Small fix: the error message was incorrect in some cases.
llvm-svn: 60099
2008-11-26 10:55:45 +00:00
Mikhail Glushenkov
0c4ad9c7cc Support dependencies between plugins by priority-sorting.
llvm-svn: 59449
2008-11-17 17:30:25 +00:00
Mikhail Glushenkov
c44e8c9f2b Filter ToolPropertiesList to exclude all Tools not mentioned in the compilation graph.
llvm-svn: 59448
2008-11-17 17:29:42 +00:00
Mikhail Glushenkov
773924299a Add a layer of indirection to make plugins more flexible.
Use strings instead of TableGen defs in the compilation graph
definition. Makes it easier for the plugins to modify an existing graph.

llvm-svn: 59447
2008-11-17 17:29:18 +00:00
Bill Wendling
9eaf125600 Fix to record comparator to make it work for return values > 1.
llvm-svn: 59242
2008-11-13 12:03:00 +00:00
Bill Wendling
ab7198eb4b Put comma in correct place for call to StructType::get
llvm-svn: 59241
2008-11-13 10:18:35 +00:00
Bill Wendling
9dc2bd7973 Modify the intrinsics pattern to separate out the "return" types from the
"parameter" types. An intrinsic can now return a multiple return values like
this:

  def add_with_overflow : Intrinsic<[llvm_i32_ty, llvm_i1_ty],
                                    [LLVMMatchType<0>, LLVMMatchType<0>]>;

llvm-svn: 59237
2008-11-13 09:08:33 +00:00