| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Re-materialization is for uses only. | Evan Cheng | 2008-04-01 | 1 | -4/+4 |
| | | | | | llvm-svn: 49053 | ||||
| * | Don't use __bzero for memset if the second argument isn't zero. | Dan Gohman | 2008-04-01 | 1 | -1/+2 |
| | | | | | llvm-svn: 49050 | ||||
| * | Speculatively micro-optimize memory-zeroing calls on Darwin 10. | Dan Gohman | 2008-04-01 | 3 | -5/+35 |
| | | | | | llvm-svn: 49048 | ||||
| * | Revert 49006 for the moment. | Dale Johannesen | 2008-04-01 | 8 | -49/+25 |
| | | | | | llvm-svn: 49046 | ||||
| * | Don't dereference MBB->end(). | Owen Anderson | 2008-04-01 | 1 | -1/+1 |
| | | | | | llvm-svn: 49043 | ||||
| * | Change the MemoryBuffer::getFile* methods to take just a pointer to the | Chris Lattner | 2008-04-01 | 8 | -31/+20 |
| | | | | | | | | | start of a filename, not a filename+length. All clients can produce a null terminated name, and the system api's require null terminated strings anyway. llvm-svn: 49041 | ||||
| * | Disabling remat of load from gv stub (temporarily) again to fix llvmgcc ↵ | Evan Cheng | 2008-04-01 | 1 | -0/+3 |
| | | | | | | | bootstrap miscompare. llvm-svn: 49037 | ||||
| * | add some #includes. | Chris Lattner | 2008-04-01 | 1 | -2/+8 |
| | | | | | llvm-svn: 49036 | ||||
| * | MappedFile is dead, remove it. | Chris Lattner | 2008-04-01 | 3 | -235/+0 |
| | | | | | llvm-svn: 49035 | ||||
| * | Implement Path::MapInFilePages/UnMapFilePages on unix, which | Chris Lattner | 2008-04-01 | 2 | -7/+11 |
| | | | | | | | provides fast MappedFile::getFile for large files. llvm-svn: 49034 | ||||
| * | Reimplement MemoryBuffer::getFile with three enhancements: | Chris Lattner | 2008-04-01 | 1 | -85/+64 |
| | | | | | | | | | | | | | | | | | | | | | | 1) stop using MappedFile. 2) if profitable use the sys::path::MapInFilePages api to read the file. 3) otherwise fallback to read. When sys::path::MapInFilePages is implemented, this provides several benefits: #1: this avoids fragmenting memory for small files. #2: this avoids extraneous stat calls when the file size is known. #3: this only keeps the file descriptor open while reading the file, not for the duration of the lifetime of the memory buffer. This fixes a serious clang FD 'leak' problem. I believe that this will work on a win32 machine, but I don't have one to test on. I'd appreciate it if someone could check. llvm-svn: 49031 | ||||
| * | Stub out some sys::Path::MapInFilePages/UnMapFilePages methods. | Chris Lattner | 2008-04-01 | 2 | -0/+20 |
| | | | | | llvm-svn: 49030 | ||||
| * | change the archive stuff to use MemoryBuffer instead of mappedfile. | Chris Lattner | 2008-04-01 | 3 | -47/+34 |
| | | | | | | | | MemoryBuffer is higher level and more closely matches the model needed. llvm-svn: 49029 | ||||
| * | rewrite SourceFile to be in terms of MemoryBuffer, not MappedFile. | Chris Lattner | 2008-04-01 | 1 | -13/+19 |
| | | | | | llvm-svn: 49027 | ||||
| * | Make MappedFile::map return a const correct pointer, don't leak address ↵ | Chris Lattner | 2008-04-01 | 2 | -2/+3 |
| | | | | | | | space on Unix platforms. llvm-svn: 49026 | ||||
| * | Remove the MappedFile::charBase member, rename base -> getBase() and | Chris Lattner | 2008-04-01 | 2 | -7/+8 |
| | | | | | | | make getBase() return a const-correct pointer. llvm-svn: 49025 | ||||
| * | Change DiffFilesWithTolerance to be written in terms of MemoryBuffer, | Chris Lattner | 2008-04-01 | 1 | -73/+56 |
| | | | | | | | | | not an mmapped file. This more closely matches its requirements and provides an implicitly null terminated buffer, something this routine had to emulate itself before. llvm-svn: 49024 | ||||
| * | Remove MappedFile support for mapping files for write and exec | Chris Lattner | 2008-04-01 | 6 | -112/+18 |
| | | | | | | | | and shared. This complicates the design, is not used, and probably doesn't even work. llvm-svn: 49022 | ||||
| * | add missing #includes | Chris Lattner | 2008-04-01 | 3 | -0/+3 |
| | | | | | llvm-svn: 49020 | ||||
| * | More soft fp fixes. | Evan Cheng | 2008-04-01 | 1 | -1/+2 |
| | | | | | llvm-svn: 49016 | ||||
| * | Pasto. | Evan Cheng | 2008-04-01 | 1 | -1/+1 |
| | | | | | llvm-svn: 49014 | ||||
| * | Add comment. | Evan Cheng | 2008-04-01 | 1 | -0/+1 |
| | | | | | llvm-svn: 49013 | ||||
| * | Unbreak ARM / Thumb soft FP support. | Evan Cheng | 2008-04-01 | 2 | -9/+21 |
| | | | | | llvm-svn: 49012 | ||||
| * | Accept 'y' constraint (MMX) in inline asm. | Dale Johannesen | 2008-04-01 | 1 | -0/+1 |
| | | | | | llvm-svn: 49011 | ||||
| * | cleanup the MappedFile API and comments. This removes and updates | Chris Lattner | 2008-04-01 | 2 | -118/+119 |
| | | | | | | | | | tons of out of date comments (really nothing throws here!) and fixes some other fairly glaring issues: "size" used to return the size of the file *and* change it, depending on how you called it. llvm-svn: 49009 | ||||
| * | remove DEFINING_FILE_FOR for MappedFile.h | Chris Lattner | 2008-04-01 | 1 | -1/+0 |
| | | | | | llvm-svn: 49008 | ||||
| * | Emit exception handling info for functions which are | Dale Johannesen | 2008-03-31 | 8 | -25/+49 |
| | | | | | | | | | | | | not marked nounwind, or for all functions when -enable-eh is set, provided the target supports Dwarf EH. llvm-gcc generates nounwind in the right places; other FEs will need to do so also. Given such a FE, -enable-eh should no longer be needed. llvm-svn: 49006 | ||||
| * | It's not safe to fold a load from GV stub or constantpool into a two-address ↵ | Evan Cheng | 2008-03-31 | 2 | -2/+11 |
| | | | | | | | use. llvm-svn: 49002 | ||||
| * | Actually disable crash reporting on Mac OS X, returning bugpoint to speedy | Nate Begeman | 2008-03-31 | 1 | -23/+6 |
| | | | | | | | crash miscompilations. llvm-svn: 49000 | ||||
| * | Set blockBegin to point to the beginning of the block, | Dan Gohman | 2008-03-31 | 1 | -2/+2 |
| | | | | | | | not the end. llvm-svn: 48999 | ||||
| * | Move reMaterialize() from TargetRegisterInfo to TargetInstrInfo. | Evan Cheng | 2008-03-31 | 22 | -141/+87 |
| | | | | | llvm-svn: 48995 | ||||
| * | Fix a DAGCombiner optimization to respect volatile qualification. | Dan Gohman | 2008-03-31 | 1 | -1/+2 |
| | | | | | llvm-svn: 48994 | ||||
| * | Expose Function::viewCFG and Function::viewCFGOnly to bindings. | Erick Tryzelaar | 2008-03-31 | 1 | -0/+9 |
| | | | | | llvm-svn: 48982 | ||||
| * | Re-apply 48911. | Evan Cheng | 2008-03-31 | 1 | -8/+12 |
| | | | | | llvm-svn: 48977 | ||||
| * | The support for remat of instructions with a register operand is hackish, to ↵ | Evan Cheng | 2008-03-31 | 1 | -18/+7 |
| | | | | | | | say the least. Since the register operand guaranteed to be PIC base and that it is already live at all uses, we are making sure it will not be spilled after its uses are rematerialized for both performance and correctness reasons. llvm-svn: 48976 | ||||
| * | Fix a major bug in the DFS calculation. Thanks for Christopher Lamb for ↵ | Owen Anderson | 2008-03-31 | 1 | -1/+1 |
| | | | | | | | pointing this out. llvm-svn: 48973 | ||||
| * | Don't eliminate bitcast instructions that change the type of a pointer | Nate Begeman | 2008-03-31 | 2 | -19/+26 |
| | | | | | llvm-svn: 48971 | ||||
| * | Fix comment typo. | Duncan Sands | 2008-03-30 | 1 | -1/+1 |
| | | | | | llvm-svn: 48967 | ||||
| * | Moved from PR1570. | Nick Lewycky | 2008-03-30 | 1 | -0/+26 |
| | | | | | llvm-svn: 48965 | ||||
| * | Fix "Control reaches the end of non-void function" warnings, | Chris Lattner | 2008-03-30 | 9 | -0/+15 |
| | | | | | | | patch by David Chisnall. llvm-svn: 48963 | ||||
| * | minor code cleanups, allow constant folding sinf/cosf. | Chris Lattner | 2008-03-30 | 1 | -20/+26 |
| | | | | | llvm-svn: 48961 | ||||
| * | Cosmetic changes. | Evan Cheng | 2008-03-29 | 1 | -24/+3 |
| | | | | | llvm-svn: 48947 | ||||
| * | change iterator invalidation avoidance to just move the iterator backward | Chris Lattner | 2008-03-29 | 1 | -18/+25 |
| | | | | | | | | | | | | | | | | | | | | | | when something changes, instead of moving forward. This allows us to simplify memset lowering, inserting the memset at the end of the range of stuff we're touching instead of at the start. This, in turn, allows us to make use of the addressing instructions already used in the function instead of inserting our own. For example, we now codegen: %tmp41 = getelementptr [8 x i8]* %ref_idx, i32 0, i32 0 ; <i8*> [#uses=2] call void @llvm.memset.i64( i8* %tmp41, i8 -1, i64 8, i32 1 ) instead of: %tmp20 = getelementptr [8 x i8]* %ref_idx, i32 0, i32 7 ; <i8*> [#uses=1] %ptroffset = getelementptr i8* %tmp20, i64 -7 ; <i8*> [#uses=1] call void @llvm.memset.i64( i8* %ptroffset, i8 -1, i64 8, i32 1 ) llvm-svn: 48940 | ||||
| * | make the common case of a single store (which clearly shouldn't be turned | Chris Lattner | 2008-03-29 | 1 | -3/+12 |
| | | | | | | | into a memset!) faster by avoiding an allocation of an std::list node. llvm-svn: 48939 | ||||
| * | give form-memset a significantly more sane heuristic, enable it by default. | Chris Lattner | 2008-03-29 | 1 | -7/+49 |
| | | | | | llvm-svn: 48937 | ||||
| * | Remove some unneeded code for LiveInterval joining, and fix a bug in the Phi ↵ | Owen Anderson | 2008-03-29 | 1 | -58/+25 |
| | | | | | | | | | elimination algorithm where we were accidentally reasoning about the source rather than the destination. llvm-svn: 48936 | ||||
| * | Fix a tokenfactor node to use the load chain rather than the | Dan Gohman | 2008-03-28 | 1 | -1/+1 |
| | | | | | | | load value. This fixes PR2177. llvm-svn: 48932 | ||||
| * | Backing out 48911 for now. It's breaking stuff. | Evan Cheng | 2008-03-28 | 1 | -12/+8 |
| | | | | | llvm-svn: 48922 | ||||
| * | ifdef out a dead function. Should this be removed? | Chris Lattner | 2008-03-28 | 1 | -0/+2 |
| | | | | | llvm-svn: 48916 | ||||
| * | Rename getAnyLoad to getLoad is suggested by Evan. | Duncan Sands | 2008-03-28 | 2 | -17/+17 |
| | | | | | llvm-svn: 48914 | ||||

