| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Fixed scalarizing an extract subvector and prevent an infinite loop | Mon P Wang | 2008-11-06 | 2 | -3/+9 |
| | | | | | | | when simplify a vector. llvm-svn: 58820 | ||||
| * | The size limit is for individual arrays. So if any array has more than 8 bytes | Bill Wendling | 2008-11-06 | 1 | -5/+3 |
| | | | | | | | in it, then emit stack protectors. llvm-svn: 58819 | ||||
| * | Encode extend instructions; more clean up. | Evan Cheng | 2008-11-06 | 4 | -75/+142 |
| | | | | | llvm-svn: 58818 | ||||
| * | Fix a use of an invalid iterator when -debug-pass=Details is used. | Dan Gohman | 2008-11-06 | 1 | -1/+1 |
| | | | | | llvm-svn: 58816 | ||||
| * | Don't recalculate the stack position of the stack protector. | Bill Wendling | 2008-11-06 | 1 | -0/+2 |
| | | | | | llvm-svn: 58815 | ||||
| * | Emit label for llvm.dbg.func.start of the inlined function. | Devang Patel | 2008-11-06 | 2 | -3/+9 |
| | | | | | llvm-svn: 58814 | ||||
| * | Clarify documentation. A module pass MP can require a function paqss FP only ↵ | Devang Patel | 2008-11-06 | 1 | -1/+2 |
| | | | | | | | if FP does not require any module pass. llvm-svn: 58813 | ||||
| * | - Improve naming consistency: Branch -> BrFrm, BranchMisc -> BrMiscFrm. | Evan Cheng | 2008-11-06 | 4 | -76/+48 |
| | | | | | | | | - Consolidate instruction formats. - Other clean up. llvm-svn: 58808 | ||||
| * | Improve JIT debugging outputs format consistency. | Evan Cheng | 2008-11-06 | 2 | -16/+18 |
| | | | | | llvm-svn: 58807 | ||||
| * | plug leakage of mutex data. pthread_mutex_destroy() doesnt free our ↵ | Nuno Lopes | 2008-11-06 | 1 | -1/+1 |
| | | | | | | | malloc'ed memory. llvm-svn: 58805 | ||||
| * | Formating/comment changes - no functionality change. | Duncan Sands | 2008-11-06 | 1 | -13/+8 |
| | | | | | llvm-svn: 58801 | ||||
| * | Remove opcode from instruction TS flags; add MOVCC support; fix addrmode3 ↵ | Evan Cheng | 2008-11-06 | 6 | -234/+228 |
| | | | | | | | encoding bug. llvm-svn: 58800 | ||||
| * | - Rename stackprotector_{prologue,epilogue} to stackprotector_{create,check}. | Bill Wendling | 2008-11-06 | 5 | -19/+11 |
| | | | | | | | | - Get rid of "HasStackProtector" in MachineFrameInfo. - Modify intrinsics to tell which are doing what with memory. llvm-svn: 58799 | ||||
| * | Update VC++ projects. | Steve Naroff | 2008-11-06 | 3 | -36/+52 |
| | | | | | llvm-svn: 58798 | ||||
| * | Widening cleanup | Mon P Wang | 2008-11-06 | 4 | -25/+22 |
| | | | | | llvm-svn: 58796 | ||||
| * | Handle smul<x><y>, smulw<y>, smla<x><y>, smlaw<y>. | Evan Cheng | 2008-11-06 | 3 | -64/+101 |
| | | | | | llvm-svn: 58793 | ||||
| * | Adjust the stack protector heuristic to care about only arrays or calls to | Bill Wendling | 2008-11-06 | 1 | -1/+9 |
| | | | | | | | "alloca". llvm-svn: 58792 | ||||
| * | Implement the stack protector stack accesses via intrinsics: | Bill Wendling | 2008-11-06 | 5 | -9/+105 |
| | | | | | | | | | | | | | - stackprotector_prologue creates a stack object and stores the guard there. - stackprotector_epilogue reads the stack guard from the stack position created by stackprotector_prologue. - The PrologEpilogInserter was changed to make sure that the stack guard is first on the stack frame. llvm-svn: 58791 | ||||
| * | Fix so_imm encoding bug; add support for MOVi2pieces. | Evan Cheng | 2008-11-06 | 3 | -16/+56 |
| | | | | | llvm-svn: 58790 | ||||
| * | Fix encoding of multiple instructions with 3 src operands; also handle ↵ | Evan Cheng | 2008-11-06 | 4 | -77/+96 |
| | | | | | | | smmul, smmla, and smmls. llvm-svn: 58789 | ||||
| * | Need a \n. | Evan Cheng | 2008-11-06 | 1 | -1/+2 |
| | | | | | llvm-svn: 58788 | ||||
| * | InstructionNamer preserves everything. | Devang Patel | 2008-11-06 | 1 | -0/+4 |
| | | | | | llvm-svn: 58787 | ||||
| * | Emit label for llvm.dbg.func.start of the inlined function. | Devang Patel | 2008-11-06 | 2 | -3/+77 |
| | | | | | llvm-svn: 58786 | ||||
| * | Add post-build event for clangDriver (to copy clang.exe into "dstroot"). | Steve Naroff | 2008-11-06 | 1 | -3/+9 |
| | | | | | llvm-svn: 58783 | ||||
| * | Undo 58778 but makes the binary dump prettier. | Evan Cheng | 2008-11-05 | 1 | -3/+21 |
| | | | | | llvm-svn: 58782 | ||||
| * | Add comments to function. | Bill Wendling | 2008-11-05 | 1 | -0/+6 |
| | | | | | llvm-svn: 58781 | ||||
| * | Encode pic load / store instructions; fix some encoding bugs. | Evan Cheng | 2008-11-05 | 3 | -89/+153 |
| | | | | | llvm-svn: 58780 | ||||
| * | Add command line option -entry-funcion to override entry function (default ↵ | Evan Cheng | 2008-11-05 | 1 | -5/+12 |
| | | | | | | | is main). llvm-svn: 58779 | ||||
| * | Remove debug output that's not really useful. | Evan Cheng | 2008-11-05 | 1 | -13/+0 |
| | | | | | llvm-svn: 58778 | ||||
| * | Make ISel ignore dead nodes. The DAGCombiner normally eliminates | Dan Gohman | 2008-11-05 | 2 | -0/+27 |
| | | | | | | | | dead nodes, but in this case its missing one. Fixing the DAGCombiner is desirable, but it's somewhat involved. llvm-svn: 58777 | ||||
| * | opt was not exporting the Mangler symbols | Andrew Lenharth | 2008-11-05 | 1 | -1/+2 |
| | | | | | llvm-svn: 58775 | ||||
| * | Add PR number. | Devang Patel | 2008-11-05 | 1 | -0/+1 |
| | | | | | llvm-svn: 58765 | ||||
| * | Restructure ARM code emitter to use instruction formats instead of ↵ | Evan Cheng | 2008-11-05 | 5 | -285/+302 |
| | | | | | | | addressing modes to determine how to encode instructions. llvm-svn: 58764 | ||||
| * | Use an assert to check that SelectCode isn't called on | Dan Gohman | 2008-11-05 | 1 | -5/+4 |
| | | | | | | | nodes that are already selected. llvm-svn: 58763 | ||||
| * | The HadDelete field is no longer used. | Dan Gohman | 2008-11-05 | 1 | -1/+0 |
| | | | | | llvm-svn: 58761 | ||||
| * | Reintroduce a comment that was removed with the AddToISelQueue | Dan Gohman | 2008-11-05 | 1 | -0/+1 |
| | | | | | | | changes. llvm-svn: 58760 | ||||
| * | Update some comments to reflect the new code. | Dan Gohman | 2008-11-05 | 1 | -3/+8 |
| | | | | | llvm-svn: 58759 | ||||
| * | Test commit, add Makefile for XCore target, more to follow. | Richard Osborne | 2008-11-05 | 1 | -0/+21 |
| | | | | | llvm-svn: 58755 | ||||
| * | Fix thinko in ppcf128 expansion of truncating store. | Duncan Sands | 2008-11-05 | 1 | -1/+1 |
| | | | | | llvm-svn: 58753 | ||||
| * | Add more vector move low and zero-extend patterns. | Evan Cheng | 2008-11-05 | 2 | -0/+24 |
| | | | | | llvm-svn: 58752 | ||||
| * | Type of shuffle mask has changed. | Evan Cheng | 2008-11-05 | 1 | -1/+2 |
| | | | | | llvm-svn: 58751 | ||||
| * | Indentation. | Evan Cheng | 2008-11-05 | 1 | -1/+1 |
| | | | | | llvm-svn: 58750 | ||||
| * | Eliminate the ISel priority queue, which used the topological order for a | Dan Gohman | 2008-11-05 | 11 | -308/+74 |
| | | | | | | | | | | | | | | | | | | | | | | | | | priority function. Instead, just iterate over the AllNodes list, which is already in topological order. This eliminates a fair amount of bookkeeping, and speeds up the isel phase by about 15% on many testcases. The impact on most targets is that AddToISelQueue calls can be simply removed. In the x86 target, there are two additional notable changes. The rule-bending AND+SHIFT optimization in MatchAddress that creates new pre-isel nodes during isel is now a little more verbose, but more robust. Instead of either creating an invalid DAG or creating an invalid topological sort, as it has historically done, it can now just insert the new nodes into the node list at a position where they will be consistent with the topological ordering. Also, the address-matching code has logic that checked to see if a node was "already selected". However, when a node is selected, it has all its uses taken away via ReplaceAllUsesWith or equivalent, so it won't recieve any further visits from MatchAddress. This code is now removed. llvm-svn: 58748 | ||||
| * | Use getTargetConstant instead of getConstant for nodes that should not be ↵ | Dan Gohman | 2008-11-05 | 1 | -8/+8 |
| | | | | | | | | | visited by isel and potentially forced into registers. llvm-svn: 58747 | ||||
| * | Rename isGVLazyPtr to isGVNonLazyPtr relocation. This represents Mac OS X | Evan Cheng | 2008-11-05 | 6 | -60/+62 |
| | | | | | | | indirect gv reference. Please don't call it lazy. llvm-svn: 58746 | ||||
| * | New test case. | Devang Patel | 2008-11-05 | 1 | -0/+6 |
| | | | | | llvm-svn: 58745 | ||||
| * | Do now allow InlineAlways pass to remove dead functions. | Devang Patel | 2008-11-05 | 3 | -0/+17 |
| | | | | | llvm-svn: 58744 | ||||
| * | Silence unused variable warnings. | Devang Patel | 2008-11-05 | 1 | -1/+3 |
| | | | | | llvm-svn: 58743 | ||||
| * | Check Attribute::NoInline. | Devang Patel | 2008-11-05 | 1 | -4/+2 |
| | | | | | llvm-svn: 58742 | ||||
| * | Remove dead variable. | Bill Wendling | 2008-11-05 | 1 | -1/+0 |
| | | | | | llvm-svn: 58741 | ||||

