maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
78,269 Commits 23 Branches 0 Tags 2.2 GiB
Commit Graph

78269 Commits

Author SHA1 Message Date
Chris Lattner b02cdaaa9c Check in the first big step of rewriting DAGISelEmitter to 
produce a table based matcher instead of gobs of C++ Code.

Though it's not done yet, the shrinkage seems promising,
the table for the X86 ISel is 75K and still has a lot of 
optimization to come (compare to the ~1.5M of .o generated
the old way, much of which will go away).

The code is currently disabled by default (the #if 0 in
DAGISelEmitter.cpp).  When enabled it generates a dead
SelectCode2 function in the DAGISel Header which will
eventually replace SelectCode.

There is still a lot of stuff left to do, which are
documented with a trail of FIXMEs.

llvm-svn: 96215
 2010-02-15 08:04:42 +00:00
Chris Lattner 8fd1fe1612 give SDValue an operator->, allowing V->isTargetOpcode() and 
many other natural things.

llvm-svn: 96214
 2010-02-15 07:11:34 +00:00
Chris Lattner f0520da7a1 don't make insanely large node numbers for no reason, 
packing somewhat densely is better than not.

llvm-svn: 96213
 2010-02-15 06:39:31 +00:00
Chris Lattner d4df64d15f no need to add the instruction count anymore. 
llvm-svn: 96212
 2010-02-15 06:38:41 +00:00
Zhongxing Xu ec345b7ecb Simplify code. If we can reach here, the base must be virtual. 
llvm-svn: 96211
 2010-02-15 04:28:35 +00:00
Mikhail Glushenkov 5352f07f2c Revert r96130 ("Forward parameter options as '-option=param'"). 
This behaviour must be configurable.

llvm-svn: 96210
 2010-02-15 03:17:06 +00:00
Chris Lattner 115158fa4f enhance raw_svector_ostream::write_impl to work with unbuffered streams, 
which may call write_impl on things that are not the usual buffer.

llvm-svn: 96209
 2010-02-15 02:18:26 +00:00
Chris Lattner f733d75892 make PadToColumn return the stream so you can use: 
OS.PadToColumn(42) << "foo";

llvm-svn: 96208
 2010-02-15 02:17:50 +00:00
Dale Johannesen c704ea1c51 Ignore DBG_VALUE in a couple more places. 
llvm-svn: 96207
 2010-02-15 01:45:47 +00:00
John McCall 7538eec67b When emitting an aggregate into a temporary, make sure we set the alignment 
on the alloca.

The fact that codegen makes this class of bug so wonderfully easy to make is
embarrassing.

llvm-svn: 96204
 2010-02-15 01:23:36 +00:00
Dan Gohman 4d8feb11dd When restoring a saved insert location, check to see if the saved 
insert location has become an "inserted" instruction since the time
it was saved. If so, advance to the first non-"inserted" instruction.

llvm-svn: 96203
 2010-02-15 00:21:43 +00:00
Ted Kremenek 4ba99bee2a clang::Attr objects should be treated as AST nodes as well by 
the LLVMConventionsChecker.

llvm-svn: 96202
 2010-02-14 22:58:16 +00:00
Chris Lattner 450d50467c constize 
llvm-svn: 96199
 2010-02-14 22:33:49 +00:00
Chris Lattner 53c39ba3c1 clean up a bunch of code, move some random predicates 
on TreePatternNode to be methods on TreePatternNode.

llvm-svn: 96197
 2010-02-14 22:22:58 +00:00
Chris Lattner 4aa57ba3ef mark "addr" as having type "iPTR", eliminating some type comparisons 
in hte generated dag isel fil.

llvm-svn: 96193
 2010-02-14 21:53:19 +00:00
Chris Lattner b3d9942c51 remove the DisablePatternForFastISel predicate, which is a check 
that predated -fast-isel which attempted to speed up the dag pattern
matchers at -O0.  Since fast-isel is around, this is basically 
obsolete and removing it shrinks the generated dag isels.

llvm-svn: 96188
 2010-02-14 21:11:53 +00:00
Chris Lattner dd2ec58276 add an insertion operator. 
llvm-svn: 96187
 2010-02-14 21:10:33 +00:00
Chris Lattner 78291e3be8 tidy up 
llvm-svn: 96186
 2010-02-14 21:10:15 +00:00
Ted Kremenek 4853131459 Add LLVM conventions check that scans for AST elements (types, stmts, decls) 
that allocate heap memory.

llvm-svn: 96184
 2010-02-14 19:09:13 +00:00
Ted Kremenek 988805c81e Change LLVMConventionsChecker to accept an entire translation unit instead 
of operating on each code decl.  This exposes two flaws in AnalysisConsumer
that should eventually be fixed:

(1) It is not possible to associate multiple "actions" with a single
    command line argument.  This will require the notion of an
	"analysis" group, and possibly tablegen support.  (although eventually
    we want to support dynamically loading analyses as well)

(2) AnalysisConsumer may not actually be scanning the declarations in namespaces.
    We'll experiment first in LLVMConventionsChecker before changing the
    behavior in AnalysisConsumer.

llvm-svn: 96183
 2010-02-14 19:09:05 +00:00
Ted Kremenek 39df94b00d Rework translation unit actions to actually take an entire translation unit 
as imput.

llvm-svn: 96182
 2010-02-14 19:08:51 +00:00
Ted Kremenek c968e5e9b9 For the StringRef check, also visit the children of DeclStmts. 
llvm-svn: 96181
 2010-02-14 19:08:43 +00:00
Ted Kremenek 4806a833a7 Place type-checking static methods at type of file (where they will congregate). 
No functionality change.

llvm-svn: 96180
 2010-02-14 19:08:36 +00:00
Dan Gohman e4e51a63da Fix whitespace. 
llvm-svn: 96179
 2010-02-14 18:51:39 +00:00
Dan Gohman e7f74bb16c Fix a comment. 
llvm-svn: 96178
 2010-02-14 18:51:20 +00:00
Dan Gohman bb7d52213c When complicated expressions are broken down into subexpressions 
with multiplication by constants distributed through, occasionally
those subexpressions can include both x and -x. For now, if this
condition is discovered within LSR, just prune such cases away,
as they won't be profitable. This fixes a "zero allocated in a
base register" assertion failure.

llvm-svn: 96177
 2010-02-14 18:50:49 +00:00
Chris Lattner d0413848cc temporarily revert 96173, it is causing test failures. 
llvm-svn: 96176
 2010-02-14 18:38:38 +00:00
Sanjiv Gupta d6cbe9f23f fixes to pagesel/banksel inserter. 
1. restore these across direct/indirect calls.
2. restore pagesel for any macros with gotos.

llvm-svn: 96175
 2010-02-14 18:27:42 +00:00
Anton Korobeynikov 0a5e0371b0 Forgot to commit the header 
llvm-svn: 96174
 2010-02-14 18:25:41 +00:00
Sanjiv Gupta 1ef8cdd29b renamed pic16 specifiic macros. 
llvm-svn: 96173
 2010-02-14 18:20:18 +00:00
Chris Lattner 3b38fd6488 follow-on to PR6280 
llvm-svn: 96172
 2010-02-14 18:20:09 +00:00
Anders Carlsson 033d48697f Don't compute final overriders or build vtables for bases that don't need a vtable. 
llvm-svn: 96171
 2010-02-14 17:05:59 +00:00
Anton Korobeynikov ce77ce3072 Drop winmcasminfo and use normal AT&T COFF for all windows targets. 
Otherwise AT&T asm printer is used with non-compatible MCAsmInfo and
there is no way to override this behaviour.

llvm-svn: 96165
 2010-02-14 15:19:54 +00:00
Benjamin Kramer 88ab94e7a3 Pass StringRefs to InsertText/ReplaceText in RewriteObjC and remove a ton of unnecessary length arguments. 
llvm-svn: 96164
 2010-02-14 14:14:16 +00:00
Daniel Dunbar 02968e53cf CIndex: Switch CXSourceRange to proper half-open intervals. 
- Doug, please review.

llvm-svn: 96162
 2010-02-14 10:02:57 +00:00
Daniel Dunbar edeb9b87e6 CIndex: Avoid an unnecessary getLocForEndOfToken call, the region of interest 
doesn't need to be a full token.
 - Doug, please review.

llvm-svn: 96161
 2010-02-14 10:02:42 +00:00
Daniel Dunbar 6092d507f0 c-index-test: Unify and always print half-open extents. 
llvm-svn: 96160
 2010-02-14 08:32:51 +00:00
Daniel Dunbar eb27e7d999 c-index-test: Simplify file scanning code. 
llvm-svn: 96159
 2010-02-14 08:32:32 +00:00
Daniel Dunbar 98c07e0e23 c-index-test: Unify syntax for printing extents. Yes, there were 4. 
llvm-svn: 96158
 2010-02-14 08:32:24 +00:00
Daniel Dunbar 2049933196 CIndex: Simplify (remove provably dead code). 
llvm-svn: 96157
 2010-02-14 08:32:11 +00:00
Daniel Dunbar 2f4ba179f5 CIndex: Inline CompareRegionOfInterest(CXSourceRange) into sole caller. 
llvm-svn: 96156
 2010-02-14 08:32:05 +00:00
Daniel Dunbar 80daf53600 Rename translateSourceRange(CXSourceRange) translateCXSourceRange, instead of 
having overloaded functions with inverse semantics.

llvm-svn: 96155
 2010-02-14 08:31:57 +00:00
Zhongxing Xu be36ecbb60 Fix pr6293. If ptr is NULL, no operation is preformed. 
llvm-svn: 96154
 2010-02-14 06:49:48 +00:00
Johnny Chen c95a814ec3 Try to factorize the specification of saturating add/subtract operations a bit, 
as suggested by Bob Wilson.

llvm-svn: 96153
 2010-02-14 06:32:20 +00:00
Dan Gohman 2d0f96d49a Actually, this code doesn't have to be quite so conservative in 
the no-TLI case. But it should still default to declining the
transformation.

llvm-svn: 96152
 2010-02-14 03:21:49 +00:00
Dan Gohman 6b7517342e In rememberInstruction, if the value being remembered is the 
current insertion point, advance the current insertion point.
This avoids a use-before-def situation in a testcase extracted
from clang which is difficult to reduce to a reasonable-sized
regression test.

llvm-svn: 96151
 2010-02-14 03:12:47 +00:00
Dan Gohman f446713fd0 Simplify this code; no need for a custom subclass if it doesn't need 
to override anything from the parent class.

llvm-svn: 96150
 2010-02-14 02:48:58 +00:00
Dan Gohman 6fafffb635 Remove a 'protected' keyword, now that SCEVExpander is no longer 
intended to be subclassed.

llvm-svn: 96149
 2010-02-14 02:47:26 +00:00
Dan Gohman cb76a806f0 Don't attempt aggressive post-inc uses if TargetLowering is not available, 
because profitability can't be sufficiently approximated.

llvm-svn: 96148
 2010-02-14 02:45:21 +00:00
Ted Kremenek 184b3383bf Add new static analyzer for checking LLVM coding conventions: -analyzer-check-llvm-conventions 
Currently these checks are intended to be largely syntactical, but may get more
sophisticated over time.

As an initial foray into this brave new world, emit a static analyzer warning
when binding a temporary 'std::string' to an 'llvm::StringRef' where the
lifetime of the 'std::string' does not outlive the 'llvm::StringRef'.

llvm-svn: 96147
 2010-02-14 02:45:18 +00:00