summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* Fix PR1090:Reid Spencer2007-01-081-42/+98
| | | | | | | | | Implemented some llvm_fcmp_{pred} functions at the start of the function bodies and use them for fcmp instructions and constant expressions. These help implement the ordered and unordered comparisons necessary for correct exectuion of these comparisons. llvm-svn: 33007
* Fix a bug in an assert that would never trigger.Reid Spencer2007-01-081-1/+1
| | | | llvm-svn: 33005
* Convert uses of getPrimitiveSize that should be getPrimitiveSizeInBits.Reid Spencer2007-01-082-15/+14
| | | | llvm-svn: 33003
* Types should be const.Reid Spencer2007-01-071-2/+2
| | | | llvm-svn: 33001
* this pass is unusedChris Lattner2007-01-071-436/+0
| | | | llvm-svn: 32998
* llvm 2.0 doesn't support llvm.isunordered.*Chris Lattner2007-01-071-17/+0
| | | | llvm-svn: 32994
* remove support for llvm.isunorderedChris Lattner2007-01-071-22/+0
| | | | llvm-svn: 32992
* remove llvm.isunorderedChris Lattner2007-01-071-5/+0
| | | | llvm-svn: 32991
* Change the interface to Module::getOrInsertFunction to be easier to use,to ↵Chris Lattner2007-01-074-256/+119
| | | | | | | | resolve PR1088, and to help PR411. This simplifies many clients also llvm-svn: 32989
* Change the interface to Module::getOrInsertFunction to be easier to use,Chris Lattner2007-01-071-12/+24
| | | | | | to resolve PR1088, and to help PR411. llvm-svn: 32988
* prepare for adjustment to getOrInsertFunction methodChris Lattner2007-01-071-6/+8
| | | | llvm-svn: 32985
* relax typeChris Lattner2007-01-071-1/+1
| | | | llvm-svn: 32983
* relax some typesChris Lattner2007-01-074-21/+20
| | | | llvm-svn: 32982
* relax typesChris Lattner2007-01-071-7/+7
| | | | llvm-svn: 32981
* relax some typesChris Lattner2007-01-072-4/+3
| | | | llvm-svn: 32980
* remove support for old-style varargs upgradingChris Lattner2007-01-071-106/+6
| | | | llvm-svn: 32979
* For PR1086:Reid Spencer2007-01-071-30/+54
| | | | | | | Parameter attributes do have to be specially handled in the CBE. Implement their handling. llvm-svn: 32976
* Fix PR1015 and Transforms/IndVarsSimplify/2007-01-06-TripCount.ll, aChris Lattner2007-01-071-1/+15
| | | | | | miscompilation of Qt. llvm-svn: 32974
* cast of int to bool no longer does a compare, rendering this fixmeChris Lattner2007-01-071-1/+0
| | | | | | obsolete llvm-svn: 32972
* add -debug output for -indvars.Chris Lattner2007-01-071-6/+19
| | | | llvm-svn: 32971
* As PR1085 was fixed, back out workaroundAnton Korobeynikov2007-01-071-6/+1
| | | | llvm-svn: 32969
* wow, the link was already broken :)Chris Lattner2007-01-061-1/+1
| | | | llvm-svn: 32963
* add a noteChris Lattner2007-01-061-1/+4
| | | | llvm-svn: 32962
* Disable the macho writer until it is 100% functional. Enabling it whenChris Lattner2007-01-061-0/+3
| | | | | | broken invites bug reports. llvm-svn: 32961
* Fix regressions in InstCombine/call-cast-target.ll and ↵Chris Lattner2007-01-061-1/+5
| | | | | | InstCombine/2003-11-13-ConstExprCastCall.ll llvm-svn: 32959
* gcc often inserts it's own names for sections (e.g.Anton Korobeynikov2007-01-061-2/+7
| | | | | | | gnu.linkonce.t.FunctionName). Convert them to "normal" LLVM names, otherwise linker won't be able to merge them. llvm-svn: 32958
* For PR411:Reid Spencer2007-01-0623-264/+202
| | | | | | | | | Take an incremental step towards type plane elimination. This change separates types from values in the symbol tables by finally making use of the TypeSymbolTable class. This yields more natural interfaces for dealing with types and unclutters the SymbolTable class. llvm-svn: 32956
* this final call to canLosslesslyBitCastTo is dead, because ValueRequiresCastChris Lattner2007-01-061-4/+0
| | | | | | is only called on integers. llvm-svn: 32949
* simplify some more code now that there are not multiple different integerChris Lattner2007-01-061-8/+4
| | | | | | types of the same size llvm-svn: 32948
* eliminate some uses of canLosslesslyBitCastTo, this actually makes the code ↵Chris Lattner2007-01-061-15/+14
| | | | | | | | stronger, by nuking relational pointer comparisons with casts. llvm-svn: 32947
* no need to worry about int vs uint any more.Chris Lattner2007-01-061-2/+1
| | | | llvm-svn: 32946
* new noteChris Lattner2007-01-061-0/+5
| | | | llvm-svn: 32945
* setSetCCIsExpensive is gone.Evan Cheng2007-01-051-2/+0
| | | | llvm-svn: 32941
* Expand fcopysign to the bitwise sequence if select is marked as expensive.Evan Cheng2007-01-051-4/+6
| | | | llvm-svn: 32940
* 1) Remove old AnalysisResolver.Devang Patel2007-01-052-16/+9
| | | | | | 2) Rename AnalysisResolver_New as AnalysisResolver llvm-svn: 32938
* Regenerate.Reid Spencer2007-01-053-319/+281
| | | | llvm-svn: 32934
* For PR1077:Reid Spencer2007-01-051-26/+7
| | | | | | | | Disallow merging of dupliate global variables. It is now illegal to declare or define two global variables of the same name and same type. llvm-gcc3 is dead in 2.0 and llvm-gcc4 doesn't have that problem nor need the hack. llvm-svn: 32933
* - FCOPYSIGN custom lowering bug. Clear the sign bit of operand 0 first beforeEvan Cheng2007-01-053-21/+39
| | | | | | | | or'ing in the sign bit of operand 1. - Tweaking: rather than left shift the sign bit, fp_extend operand 1 first before taking its sign bit if its type is smaller than that of operand 0. llvm-svn: 32932
* Bug in ExpandFCOPYSIGNToBitwiseOps(). Clear the old sign bit of operand 0Evan Cheng2007-01-051-4/+14
| | | | | | before or'ing in the sign bit of operand 1. llvm-svn: 32930
* CopyToReg source operand can be a register as well. e.g. Copy from ↵Evan Cheng2007-01-051-2/+6
| | | | | | GlobalBaseReg. llvm-svn: 32929
* Remove PassManagerT.hDevang Patel2007-01-051-0/+0
| | | | llvm-svn: 32928
* Remove old pass manager.Devang Patel2007-01-053-1099/+1
| | | | llvm-svn: 32927
* Regenerate.Reid Spencer2007-01-053-1170/+1159
| | | | llvm-svn: 32905
* Change the syntax for parameter attributes:Reid Spencer2007-01-053-72/+79
| | | | | | | | | | | | | | | | | | | 1. The @ sign is no longer necessary. 2. We now support "function attributes" as parameter attribute 0. 3. Instead of locating the return type attributes after the type of a function result, they are now located after the function header's closing paranthesis and before any alignment or section options. 4. The way has been prepared for a new "noreturn" function attribute but there is no support for recognizing it in the lexer nor doing anything with it if it does get set. 5. The FunctionType::getParamAttrsText method now has support for returning multiple attributes. This required a change in its interface. I'm unhappy that this change leads to 6 new shift/reduce conflicts, but in each case bison's decision to choose the shift is correct so there shouldn't be any damage from these conflicts. llvm-svn: 32904
* TypoEvan Cheng2007-01-051-1/+1
| | | | llvm-svn: 32902
* With SSE2, expand FCOPYSIGN to a series of SSE bitwise operations.Evan Cheng2007-01-053-8/+93
| | | | llvm-svn: 32900
* Implement InstCombine/vec_shuffle.ll:%test7, simplifying shuffles withChris Lattner2007-01-051-1/+23
| | | | | | undef operands. llvm-svn: 32899
* fold things like a^b != c^a -> b != c. This implements ↵Chris Lattner2007-01-051-12/+33
| | | | | | InstCombine/xor.ll:test27 llvm-svn: 32893
* Compile X + ~X to -1. This implements Instcombine/add.ll:test34Chris Lattner2007-01-051-1/+7
| | | | llvm-svn: 32890
* GEP subscript is interpreted as a signed value.Evan Cheng2007-01-051-1/+1
| | | | llvm-svn: 32888
OpenPOWER on IntegriCloud