- ee50a1a Don't special case XS, XD prefixes. by Evan Cheng · 20 years ago
- 576c141 Bug fix: XS, XD prefixes were being emitted twice. by Evan Cheng · 20 years ago
- 33d5082 Make sure that weak functions are aligned properly by Chris Lattner · 20 years ago
- ddf7532 Duh by Evan Cheng · 20 years ago
- 968c178 Remove -disable-x86-sse by Evan Cheng · 20 years ago
- 60330ff add an assert by Chris Lattner · 20 years ago
- c4f8eee Keep to < 80 cols by Evan Cheng · 20 years ago
- ac940ab Missed a break so memcpy cases fell through to memset. Doh. by Evan Cheng · 20 years ago
- 80e89d7 Fixed a build breakage. by Evan Cheng · 20 years ago
- a03a5dc Rename maxStoresPerMemSet to maxStoresPerMemset, etc. by Evan Cheng · 20 years ago
- 952b7d6 Add a entry. by Evan Cheng · 20 years ago
- 87ed716 Set maxStoresPerMemSet to 16. Ditto for maxStoresPerMemCpy and by Evan Cheng · 20 years ago
- 1db92f9 Expand memset dst, c, size to a series of stores if size falls below the by Evan Cheng · 20 years ago
- 82eaf62 Enable SSE (for the right subtargets) by Evan Cheng · 20 years ago
- 976e110 Match changes to unix build system. by Jeff Cohen · 20 years ago
- 06a248c9 now that libcalls don't suck, we can remove this hack by Chris Lattner · 20 years ago
- 1828ee1 Adjust to new style "generated files in CVS" mechanism for lex output by Chris Lattner · 20 years ago
- 32eecb0 adjust to new style of handling lexer changes by Chris Lattner · 20 years ago
- e8242b1 Adjust to new form of handling lexer dependencies, this way shouldn't have by Chris Lattner · 20 years ago
- 342567c Implement an alternative way of handling generated lex files in CVS. This by Chris Lattner · 20 years ago
- 77efe27 Wrap a couple more long lines by Chris Lattner · 20 years ago
- d25ad61 wrap long lines by Chris Lattner · 20 years ago
- 3dd4c40 Use statistics to keep track of what flavors of loops we are unswitching by Chris Lattner · 20 years ago
- 4b653a0 Fix a latent bug in the call sequence handling stuff. Some targets (e.g. x86) by Chris Lattner · 20 years ago
- 2082ad9 Implement Instcombine/and.ll:test34 by Chris Lattner · 20 years ago
- 1f7b4e6 new testcase by Chris Lattner · 20 years ago
- c70cc35 Fix typo that caused build failures for things trying to use m_Or. by Chris Lattner · 20 years ago
- f345fe4 If any of the sign extended bits are demanded, the input sign bit is demanded by Chris Lattner · 20 years ago
- 3fcb664 testcase reduced from Ptrdist/bc that is miscompiled by recent instcombine changes. by Chris Lattner · 20 years ago
- f9a0532 Another hack due to allowing multiple symbols with the same name. by Chris Lattner · 20 years ago
- ed1e1e2 Another work around for the 'symbols with different types can have the same by Chris Lattner · 20 years ago
- 40ec503 improved zap discovery by Andrew Lenharth · 20 years ago
- bd7c22c test for improved zap discovery by Andrew Lenharth · 20 years ago
- f198eb9 Sync up the tag numbers with gcc4. by Jim Laskey · 20 years ago
- d96185a Rename to better reflect usage (current and planned.) by Jim Laskey · 20 years ago
- 6831a81 Completely rewrite libcall insertion by the legalizer, providing the by Chris Lattner · 20 years ago
- 94dd292 Switch targets over to using SelectionDAG::getCALLSEQ_START to create by Chris Lattner · 20 years ago
- 45b3976 Switch to using getCALLSEQ_START instead of using our own creation calls by Chris Lattner · 20 years ago
- 6a5b6d7 Add a method by Chris Lattner · 20 years ago
- 3b60592 this passes now, due to Nate's recent efforts by Chris Lattner · 20 years ago
- d1157dd Reduce this testcase a bit more, with the help of llvm-extract and some hand tweaks by Chris Lattner · 20 years ago
- c15637b Be careful not to request or look at bits shifted in from outside the size by Chris Lattner · 20 years ago
- 50b5c2e Added a test case for a libcall insertion bug. by Evan Cheng · 20 years ago
- 789fd42 Add missing patterns for andi. and andis., fixing test/Regression/CodeGen/ by Nate Begeman · 20 years ago
- 8d6bbdb remove some more dead special case code by Chris Lattner · 20 years ago
- f8c36f5 Eliminate special case hacks that are superceded by general purpose hacks by Chris Lattner · 20 years ago
- 6261d76 tweaks by Chris Lattner · 20 years ago
- bf5d8a8 Three changes: by Chris Lattner · 20 years ago
- e7df1a2 new testcase by Chris Lattner · 20 years ago
- ecd9653 move a failing testcase from bit-tracking.ll to narrow.ll, and move the by Chris Lattner · 20 years ago
- ee628cf Revert my last patch. It too breaks stuff by Chris Lattner · 20 years ago
- 82b98a0 Make these tests fail if opt crashes. by Chris Lattner · 20 years ago
- d308ddc Fix for my previously reverted patch by Chris Lattner · 20 years ago
- 539c337 Update comments to be actually accurate by Chris Lattner · 20 years ago
- 92aa95f This is implemented by the simplify-libcalls pass, not instcombine by Chris Lattner · 20 years ago
- 255d891 Port the recent innovations in ComputeMaskedBits to SimplifyDemandedBits. by Chris Lattner · 20 years ago
- 5077c7b revert my previous change, it exposed other problems. by Chris Lattner · 20 years ago
- b20f979 by Duraid Madina · 20 years ago
- 631a140 by Duraid Madina · 20 years ago
- 8587eb3 Make this check stricter. Disallow loop exit blocks from being shared by by Chris Lattner · 20 years ago
- 2486af1 Prevent certain nodes that have already been selected from being folded into by Evan Cheng · 20 years ago
- 441365c remove dead expr by Chris Lattner · 20 years ago
- ce72b17 Reorg for integration with gcc4. Old style debug info will not be passed though by Jim Laskey · 20 years ago
- c235809 implement unswitching of loops with switch stmts and selects in them by Chris Lattner · 20 years ago
- e825593 Update PHI nodes in successors of exit blocks. by Chris Lattner · 20 years ago
- b2bc315 Reform the unswitching code in terms of edge splitting, not block splitting. by Chris Lattner · 20 years ago
- 23addc0 Nicer code. :-) by Evan Cheng · 20 years ago
- f597dc7 Added X86 isel debugging stuff. by Evan Cheng · 20 years ago
- ebccc04 Remove a level of indirection. by Chris Lattner · 20 years ago
- 81be2e9 Fix a case where UnswitchTrivialCondition broke critical edges with by Chris Lattner · 20 years ago
- ad7a3e6 Use the auto-generated call matcher. Remove a broken impl of the frameaddr/returnaddr by Chris Lattner · 20 years ago
- f613fcb Update to new-style flags usage, simplifying the .td file by Chris Lattner · 20 years ago
- c57e5c1 Remove a completed entry; add a new entry about fisttp op by Evan Cheng · 20 years ago
- 708e1a5 add some notes, move some code around. Implement unswitching of loops by Chris Lattner · 20 years ago
- dd3ee6d Move code around to be more logical, no functionality change. by Chris Lattner · 20 years ago
- 4d1ca94 When unswitching a trivial loop, do admit we are doing it! :) by Chris Lattner · 20 years ago
- 4c41d49 Implement unconditional unswitching of 'trivial' loops, those loops that contain by Chris Lattner · 20 years ago
- f4f5f4e Simplify control flow a bit, note that unswitch preserves canonical loop form by Chris Lattner · 20 years ago
- 7d82d60 Match tblgen change. by Evan Cheng · 20 years ago
- 67212a0 Call InsertISelMapEntry rather than map insertion operator to prevent overly by Evan Cheng · 20 years ago
- fae9f1c Added SelectionDAG::InsertISelMapEntry(). This is used to workaround the gcc by Evan Cheng · 20 years ago
- e487abb Make the threshold a parameter by Chris Lattner · 20 years ago
- f262766 Done by Chris Lattner · 20 years ago
- 3ab48ed Enable LSR by default for SPARC: it is a clear win. by Chris Lattner · 20 years ago
- 2f4b898 Simplify the loop-unswitch pass, by not even trying to unswitch loops with by Chris Lattner · 20 years ago
- 9a4cacb Fix 80-column violations by Chris Lattner · 20 years ago
- 68d5ff2 Enhance MVIZ in three ways: by Chris Lattner · 20 years ago
- 12f7de8 new testcase by Chris Lattner · 20 years ago
- 7e9b26f Match getTargetNode() changes (now return SDNode* instead of SDOperand). by Evan Cheng · 20 years ago
- d7805a7 Match getTargetNode() changes (now returns SDNode* instead of SDOperand). by Evan Cheng · 20 years ago
- 6ae46c4 More changes to reduce frame size. by Evan Cheng · 20 years ago
- 2d99c81 this apparently passes on linux by Chris Lattner · 20 years ago
- 9413678 add an option to turn on LSR. by Chris Lattner · 20 years ago
- 239862c simplify this code now that each constant pool entry is not separately allocated by Chris Lattner · 20 years ago
- 3029f92 Adjust to MachineConstantPool interface change: instead of keeping a by Chris Lattner · 20 years ago
- 04a0f60 instead of keeping track of Constant/alignment pairs, actually compute the by Chris Lattner · 20 years ago
- fa77d43 rename fields of constant pool entries by Chris Lattner · 20 years ago
- 69a9563 Use a MachineConstantPoolEntry struct instead of a pair to hold by Chris Lattner · 20 years ago
- 74b7d27 Simplify code, alignment must be specified now. by Chris Lattner · 20 years ago
- e7a1e21 Assert invariants by Chris Lattner · 20 years ago