- 09afa1e LiveInterval: Use range based for loops for subregister ranges. by Matthias Braun · 11 years ago
- 9676195 LiveInterval: Use more range based for loops for value numbers and segments. by Matthias Braun · 11 years ago
- 279f836 VirtRegMap: Improve block live-in info if subregister liveness is available. by Matthias Braun · 11 years ago
- d70caaf VirtRegMap: No implicit defs/uses for super registers with subreg liveness tracking. by Matthias Braun · 11 years ago
- cf0444b Move register class name strings to a single array in MCRegisterInfo to reduce static table size and number of relocation entries. by Craig Topper · 11 years ago
- 1c5fce0 Migrate another set of getSubtargetImpl away. by Eric Christopher · 11 years ago
- fc6de42 Have MachineFunction cache a pointer to the subtarget to make lookups by Eric Christopher · 11 years ago
- d913448 Remove the TargetMachine forwards for TargetSubtargetInfo based by Eric Christopher · 11 years ago
- 1b9dde0 [Modules] Remove potential ODR violations by sinking the DEBUG_TYPE by Chandler Carruth · 12 years ago
- b36376e Switch a number of loops in lib/CodeGen over to range-based for-loops, now that by Owen Anderson · 12 years ago
- 16c6bf4 Phase 2 of the great MachineRegisterInfo cleanup. This time, we're changing by Owen Anderson · 12 years ago
- 4584cd5 [C++11] Add 'override' keyword to virtual methods that override their base class. by Craig Topper · 12 years ago
- 442f784 [cleanup] Re-sort all the includes with utils/sort_includes.py. by Chandler Carruth · 12 years ago
- 18583d7 Keep the link register for uwtable. by Logan Chien · 12 years ago
- efbcf49 Yet another patch to reduce compile time for small programs: by Puyan Lotfi · 12 years ago
- b06a0ed [VirtRegMap] Fix for PR17825. Do not ignore noreturn definitions when setting by Quentin Colombet · 12 years ago
- fa403ab [PR16882] Ignore noreturn definitions when setting isPhysRegUsed. by Quentin Colombet · 12 years ago
- 084665f Remove VirtRegMap::getRegAllocPref(). by Jakob Stoklund Olesen · 13 years ago
- 1dd82dd Use MRI::getSimpleHint() instead of getRegAllocPref() in remaining cases. by Jakob Stoklund Olesen · 13 years ago
- 74052b0 Add VirtRegMap::hasKnownPreference(). by Jakob Stoklund Olesen · 13 years ago
- ed0881b Use the new script to sort the includes of every file under lib. by Chandler Carruth · 13 years ago
- 26c9d70 Make the LiveRegMatrix analysis available to targets. by Jakob Stoklund Olesen · 13 years ago
- c30a9af Switch most getReservedRegs() clients to the MRI equivalent. by Jakob Stoklund Olesen · 13 years ago
- b53825b Fix a significant recent(?) regression. StackSlotColoring no longer did anything by Evan Cheng · 13 years ago
- 19f49ac Release build: guard dump functions with by Manman Ren · 13 years ago
- 742534c Release build: guard dump functions with "ifndef NDEBUG" by Manman Ren · 13 years ago
- bb4bdd8 Handle overlapping regunit intervals in LiveIntervals::addKillFlags(). by Jakob Stoklund Olesen · 13 years ago
- a538d83 Add a getName function to MachineFunction. Use it in places that previously did getFunction()->getName(). Remove includes of Function.h that are no longer needed. by Craig Topper · 13 years ago
- be33629 Also compute MBB live-in lists in the new rewriter pass. by Jakob Stoklund Olesen · 13 years ago
- 1224312 Reintroduce VirtRegRewriter. by Jakob Stoklund Olesen · 13 years ago
- a0cf42f Transfer regmasks to MRI. by Jakob Stoklund Olesen · 14 years ago
- d42aba5 Rewriter should definitly rewrite instructions inside bundles. by Evan Cheng · 14 years ago
- a887908 Use the 'regalloc' debug tag for most register allocator tracing. by Jakob Stoklund Olesen · 14 years ago
- 4043d92 Assert when reserved registers have been assigned. by Jakob Stoklund Olesen · 14 years ago
- c601d8c More dead code elimination in VirtRegMap. by Jakob Stoklund Olesen · 14 years ago
- 28df7ef Stop tracking spill slot uses in VirtRegMap. by Jakob Stoklund Olesen · 14 years ago
- 92255f2 Remove dead code and data from VirtRegMap. by Jakob Stoklund Olesen · 14 years ago
- 38b3f31 Stop tracking unused registers in VirtRegMap. by Jakob Stoklund Olesen · 14 years ago
- d5d39bb Also add <imp-use,kill> flags for redefined super-registers. by Jakob Stoklund Olesen · 14 years ago
- 53e2e48 VirtRegMap is counting spill slots, not register spills. by Jakob Stoklund Olesen · 14 years ago
- 6cc4e4d Also count identity copies. by Jakob Stoklund Olesen · 15 years ago
- 71d3b89 Also add <imp-def> operands for defined and dead super-registers when rewriting. by Jakob Stoklund Olesen · 15 years ago
- e6e6750 Don't completely eliminate identity copies that also modify super register liveness. by Jakob Stoklund Olesen · 15 years ago
- 710656d Dump the register map before rewriting. by Jakob Stoklund Olesen · 15 years ago
- 5bfec69 Add VirtRegMap::rewrite() and use it in the new register allocators. by Jakob Stoklund Olesen · 15 years ago
- 2fb5b31 Simplify a bunch of isVirtualRegister() and isPhysicalRegister() logic. by Jakob Stoklund Olesen · 15 years ago
- d82ac37 Remove MachineRegisterInfo::getLastVirtReg(), it was giving wrong results by Jakob Stoklund Olesen · 15 years ago
- 1331a15 Replace TargetRegisterInfo::printReg with a PrintReg class that also works without a TRI instance. by Jakob Stoklund Olesen · 15 years ago
- cf4d5ce Fix VirtRegMap to use TRI::index2VirtReg and TRI::virtReg2Index instead of by Jakob Stoklund Olesen · 15 years ago
- 39aed73 Remember to resize SpillSlotToUsesMap when allocating an emergency spill slot. by Jakob Stoklund Olesen · 15 years ago
- df7a4f2 Now with fewer extraneous semicolons! by Owen Anderson · 15 years ago
- a57b97e Fix batch of converting RegisterPass<> to INTIALIZE_PASS(). by Owen Anderson · 15 years ago
- ddbf7a8 Use the right floating point load/store instructions in PPCInstrInfo::foldMemoryOperandImpl(). by Jakob Stoklund Olesen · 16 years ago
- 4a61882 Fix "the the" and similar typos. by Dan Gohman · 16 years ago
- 91c42f8 Change errs() to dbgs(). by David Greene · 16 years ago
- 1fbe054 by David Greene · 16 years ago
- 05fb963 The Indexes Patch. by Lang Hames · 16 years ago
- 4729191 Distinquish stack slots from other stack objects. They (and fixed objects) get FixedStack PseudoSourceValues. by Evan Cheng · 16 years ago
- 471ba48 remove some uses of llvm/Support/Streams.h by Chris Lattner · 16 years ago
- 796e43e Move more to raw_ostream, provide support for writing MachineBasicBlock, by Daniel Dunbar · 16 years ago
- 1283c6a Part 1. by Evan Cheng · 16 years ago
- 085caf1 Move register allocation preference (or hint) from LiveInterval to MachineRegisterInfo. This allows more passes to set them. by Evan Cheng · 16 years ago
- 3f77805 Make sure to color with only allocatable registers for the specific register class. by Evan Cheng · 17 years ago
- 210fc62 In some rare cases, the register allocator can spill registers but end up not utilizing registers at all. The fundamental problem is linearscan's backtracking can end up freeing more than one allocated registers. However, reloads and restores might be folded into uses / defs and freed registers might not be used at all. by Evan Cheng · 17 years ago
- b8017e0 Oy! When reverting r68073, I added in experimental code. Sorry... by Bill Wendling · 17 years ago
- c4b08e5 Revert r68073. It's causing a failure in the Apple-style builds. by Bill Wendling · 17 years ago
- f24f26c Add newlines at end of file (this can annoy gcov) by Daniel Dunbar · 17 years ago
- d37ddf5 Convert VirtRegMap to a MachineFunctionPass. by Owen Anderson · 17 years ago
- aabe06d Reorganization: Move the Spiller out of VirtRegMap.cpp into its own files. No (intended) functionality change. by Owen Anderson · 17 years ago
- fb8ded9 Yet another case where the spiller marked two uses of the same register on the same instruction as kill. This fixes PR3706. by Evan Cheng · 17 years ago
- de22116 If a MI uses the same register more than once, only mark one of them as 'kill'. by Evan Cheng · 17 years ago
- 398dee1 Last commit accidentially deleted this code. by Evan Cheng · 17 years ago
- ca2d654 The last commit was overly conservative. It's ok to reuse value that's already marked livein. by Evan Cheng · 17 years ago
- ee5fd03 If an available register falls through to a succ block, unset the last kill. Sorry, it's impossible to reduce a sensible test case. It basically requires the moon and stars to align in order to cause a failure. by Evan Cheng · 17 years ago
- f505cd5 A couple of places where reused use operands should be marked kill. This is exposed by recent availability fallthrough changes. by Evan Cheng · 17 years ago
- 65c0fd4 Revert this. It was breaking stuff. by Bill Wendling · 17 years ago
- 1c21ac3 Turn off the old way of handling debug information in the code generator. Use by Bill Wendling · 17 years ago
- e366c29 Adjust the sizes for a few SmallVectors to reflect their usage. by Dan Gohman · 17 years ago
- cf5cd6e It's (currently) not safe to keep certain physical registers live across basic blocks, e.g. x86 fp stack registers. by Evan Cheng · 17 years ago
- 8b39b54 If availability info is kept when fallthrough into a bb, add the available registers to live-in set. by Evan Cheng · 17 years ago
- eb5ec4a Remove a bogus assertion. It's possible a live-in available value is used by a previous instruction. by Evan Cheng · 17 years ago
- a1968b0 Implement PR3495: local spiller optimization. The local spiller can now keep availability information over BB boundaries. It visits BB's in depth first order. After visiting a BB if it find a successor which has a single predecessor it visits the successor next without clearing the availability information. This allows the successor to omit reloads or change them into copies. by Evan Cheng · 17 years ago
- 67cd395 Explicitly pass in debug location information to BuildMI. by Bill Wendling · 17 years ago
- c544cb0 Change TargetInstrInfo::isMoveInstr to return source and destination sub-register indices as well. by Evan Cheng · 17 years ago
- 1718fd4 Fix PR3124: overly strict assert. by Evan Cheng · 17 years ago
- 94169f1 Fix PR2898. Spiller delete a store for reuse before it knows for sure the reuse happened. by Evan Cheng · 17 years ago
- 08acb24 Fix a very subtle spiller bug: UpdateKills should not forget to track defs of aliases. by Evan Cheng · 17 years ago
- 9c4b7d5 Fix command-line option printing to print two spaces where needed, by Dan Gohman · 17 years ago
- dfe979b Don't dereference the end() iterator. Thanks to by Dan Gohman · 17 years ago
- 0d1e9a8 Switch the MachineOperand accessors back to the short names like by Dan Gohman · 17 years ago
- b401449 Propagate subreg index when promoting a load to a copy. by Evan Cheng · 17 years ago
- 534fe1c Fix PR2664 - spiller GetRegForReload wasn't respecting sub-register indices on machine operands. by Evan Cheng · 17 years ago
- 93d40ae Fix an overly strict assertion. Source register of a copy may not be killed, it may be killed by an implicit super-register use. by Evan Cheng · 17 years ago
- 7f3bfbd Make SimpleSpiller respect subregister indices. by Owen Anderson · 17 years ago
- aa33b93 Fix PR2596: out of bound reference. by Evan Cheng · 17 years ago
- 597bd16 Fix a LocalSpiller leak. This fixes tramp3d-v4. by Dan Gohman · 17 years ago
- 3b46030 Pool-allocation for MachineInstrs, MachineBasicBlocks, and by Dan Gohman · 17 years ago
- 50d5947 Cosmetic. by Evan Cheng · 17 years ago
- 51c75c0 Fix read after free found by valgrind. by Evan Cheng · 17 years ago
- fb79059 Teach the spiller to commute instructions in order to fold a reload. This hits 410 times on 444.namd and 122 times on 252.eon. by Evan Cheng · 17 years ago