summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Test that a call to a trampoline is turned into a call toDuncan Sands2007-09-111-0/+24
| | | | | | the underlying nested function. llvm-svn: 41846
* Two ParamAttrsVectors which differ by a permutationDuncan Sands2007-09-111-0/+4
| | | | | | | | | | of their elements do not yield the same ParamAttrsList, though they should. On the other hand, everyone seems to pass such vectors with elements ordered by increasing index, so rather than sorting the elements simply assert that the elements are ordered in this way. llvm-svn: 41845
* Turn calls to trampolines into calls to the underlyingDuncan Sands2007-09-111-0/+138
| | | | | | nested function. llvm-svn: 41844
* Fold the adjust_trampoline intrinsic intoDuncan Sands2007-09-1114-112/+46
| | | | | | | init_trampoline. There is now only one trampoline intrinsic. llvm-svn: 41841
* My compiler warns about the semicolon.Duncan Sands2007-09-111-1/+1
| | | | llvm-svn: 41840
* The personality function on Darwin needs a global stub. We then refer toBill Wendling2007-09-117-22/+38
| | | | | | that global stub instead of doing the ".set" thingy we were doing before. llvm-svn: 41838
* Add a ValueInfoT template parameter to DenseMap so that it can properly make ↵Owen Anderson2007-09-111-8/+22
| | | | | | | | decisions based on whether the key AND the value require ctors/dtors. llvm-svn: 41837
* Don't bother to initialize values corresponding to empty or tombstone Owen Anderson2007-09-111-1/+3
| | | | | | keys. llvm-svn: 41834
* Fix a typo in memdep, which was causing PR1648.Owen Anderson2007-09-111-1/+1
| | | | llvm-svn: 41833
* Fix bugs with &=, intersect with complement. Add three argument version of ↵Daniel Berlin2007-09-111-48/+131
| | | | | | intersect with complement. llvm-svn: 41832
* Fix non-deterministic behavior in the DenseMap copy constructor.Owen Anderson2007-09-111-2/+8
| | | | llvm-svn: 41831
* Add remaining functions necessary for andersen'sDaniel Berlin2007-09-111-12/+194
| | | | llvm-svn: 41830
* Avoid negative logic.Devang Patel2007-09-111-4/+4
| | | | llvm-svn: 41829
* Refactor code into a separate method.Devang Patel2007-09-111-28/+47
| | | | llvm-svn: 41826
* Testcase for PR1634Chris Lattner2007-09-111-0/+88
| | | | llvm-svn: 41824
* Clear split info object.Devang Patel2007-09-111-0/+1
| | | | llvm-svn: 41823
* Split condition does not have to be ICmpInst in all cases.Devang Patel2007-09-111-5/+8
| | | | llvm-svn: 41822
* Check all terminators inside loop.Devang Patel2007-09-101-4/+3
| | | | llvm-svn: 41821
* remove obsolete testcaseChris Lattner2007-09-101-9/+0
| | | | llvm-svn: 41820
* remove some dead code, this is handled by constant folding.Chris Lattner2007-09-101-8/+1
| | | | llvm-svn: 41819
* Fix a buggy constant folding transformation when handling aliases.Chris Lattner2007-09-102-4/+20
| | | | llvm-svn: 41818
* Swap exit condition operands if it works.Devang Patel2007-09-101-5/+15
| | | | llvm-svn: 41817
* regenerateChris Lattner2007-09-102-266/+308
| | | | llvm-svn: 41816
* Fix PR1645 by resolving forward alias references.Chris Lattner2007-09-102-0/+30
| | | | llvm-svn: 41815
* Fixed isFunctionPointerType to correctly handle typedefs.Ted Kremenek2007-09-101-1/+1
| | | | llvm-svn: 41814
* Add missing llvm-dis.Dale Johannesen2007-09-101-1/+1
| | | | llvm-svn: 41813
* 80 col.Evan Cheng2007-09-101-4/+6
| | | | llvm-svn: 41812
* Fix the following bug submitted by Justin Handville.Steve Naroff2007-09-101-2/+19
| | | | | | | | | | | | | | | | | | | | int main(int argc, char* argv[]) { return 0; } After speaking briefly with Chris, we decided this should be a front-end fix. The fix...have Sema::GetTypeForDeclarator() do the default function/array conversion, as I outlined in the 9/9 email on this topic. Since this conversion is done before Sema::ParseParamDeclarator(), I thought I could remove the conversion from Sema::ParseParamDeclarator(). Unfortunately, this didn't work. The conversion apparently needs to be done in both places (which doesn't make sense to me). Will investigate. llvm-svn: 41811
* New entry.Evan Cheng2007-09-101-0/+3
| | | | llvm-svn: 41810
* Observation of rematerialization.Evan Cheng2007-09-101-0/+5
| | | | llvm-svn: 41809
* Add some notes about better flag handling.Chris Lattner2007-09-103-1/+101
| | | | llvm-svn: 41808
* Emit:Chris Lattner2007-09-101-0/+10
| | | | | | | | | | | | | | | | | cmpl %eax, %ecx setae %al movzbl %al, %eax instead of: cmpl %eax, %ecx setb %al xorb $1, %al movzbl %al, %eax when using logical not of a C comparison. llvm-svn: 41807
* this is not infinite recursion.Chris Lattner2007-09-101-0/+7
| | | | llvm-svn: 41806
* 1. Don't call Value::getName(), which is slow.Chris Lattner2007-09-101-38/+52
| | | | | | | | | | | | | | | | | | | 2. Lower calls to fabs and friends to FABS nodes etc unless the function has internal linkage. Before we wouldn't lower if it had a definition, which is incorrect. This allows us to compile: define double @fabs(double %f) { %tmp2 = tail call double @fabs( double %f ) ret double %tmp2 } into: _fabs: fabs f1, f1 blr llvm-svn: 41805
* Prevent tailcallelim from breaking "recursive" calls to builtins.Chris Lattner2007-09-102-0/+19
| | | | llvm-svn: 41804
* Some small tweaks to the recent Objc support...Steve Naroff2007-09-105-5/+13
| | | | llvm-svn: 41803
* It's not safe to rematerialize MOV32r0 etc. by simply cloning the originalEvan Cheng2007-09-101-3/+22
| | | | | | | instruction. These are implemented with xor which will modify the conditional code. They should be rematerialized as move instructions. llvm-svn: 41802
* Early patch to collect objective-c methods inserts them inFariborz Jahanian2007-09-107-27/+134
| | | | | | class object. llvm-svn: 41801
* Filter exit conditions which are not yet handled.Devang Patel2007-09-102-1/+53
| | | | llvm-svn: 41800
* New test.Devang Patel2007-09-101-0/+32
| | | | llvm-svn: 41799
* Require SCEV before LCSSA.Devang Patel2007-09-101-1/+1
| | | | llvm-svn: 41798
* Fixed LiveVariables to no longer track the liveness of function pointersTed Kremenek2007-09-104-37/+87
| | | | | | | | | that refer to direct function calls. Modified interface of LiveVariables to only track liveness of VarDecls. This cleans up a bunch of edge cases, and removed the bug just mentioned. llvm-svn: 41797
* Modified dumping of Decl information to including the Decl type.Ted Kremenek2007-09-101-2/+19
| | | | llvm-svn: 41796
* Added "isFunctionPointerType()" so that we can readily see if a DeclRefExprTed Kremenek2007-09-101-1/+8
| | | | | | refers to a function. llvm-svn: 41795
* Renaming of the LiveVariablesAuditor interface. Changed "Auditor" andTed Kremenek2007-09-103-37/+39
| | | | | | "Audit" to "Observer" and "Observe" llvm-svn: 41794
* Remove an un-needed dependence query. This improves compile time marginally ↵Owen Anderson2007-09-091-9/+10
| | | | | | on 401.bzip2. llvm-svn: 41792
* Add SparseBitVector implementationDaniel Berlin2007-09-092-0/+561
| | | | llvm-svn: 41790
* Turn GVN on by default.Owen Anderson2007-09-082-4/+2
| | | | llvm-svn: 41787
* Implement misaligned FP loads and stores.Dale Johannesen2007-09-082-11/+101
| | | | llvm-svn: 41786
* Fix description of the call instruction. There are two types, with one beingNick Lewycky2007-09-081-9/+15
| | | | | | optional. llvm-svn: 41785
OpenPOWER on IntegriCloud