| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | When isel is emitting instructions for an x86 target without CMOV, the CFG is | Jakob Stoklund Olesen | 2010-09-30 | 3 | -1/+24 | |
| | | | | | | | | | | | | | | | | | | edited during emission. If the basic block ends in a switch that gets lowered to a jump table, any phis at the default edge were getting updated wrong. The jump table data structure keeps a pointer to the header blocks that wasn't getting updated after the MBB is split. This bug was exposed on 32-bit Linux when disabling critical edge splitting in codegen prepare. The fix is to uipdate stale MBB pointers whenever a block is split during emission. llvm-svn: 115191 | |||||
| * | Let FE mark a variable as artificial variable. | Devang Patel | 2010-09-29 | 1 | -0/+2 | |
| | | | | | llvm-svn: 115102 | |||||
| * | Model Cortex-a9 load to SUB, RSB, ADD, ADC, SBC, RSC, CMN, MVN, or CMP | Evan Cheng | 2010-09-29 | 2 | -37/+34 | |
| | | | | | | | pipeline forwarding path. llvm-svn: 115098 | |||||
| * | Silence msvc warnings. | Benjamin Kramer | 2010-09-29 | 1 | -6/+6 | |
| | | | | | llvm-svn: 115097 | |||||
| * | Add support to let FE encode method access specifier. | Devang Patel | 2010-09-29 | 1 | -1/+12 | |
| | | | | | llvm-svn: 115089 | |||||
| * | UnreachableBlockElim could incorrectly return false when it had not modified ↵ | Owen Anderson | 2010-09-29 | 1 | -1/+4 | |
| | | | | | | | the CFG, but HAD modified some PHI nodes. Fixes PR8174. llvm-svn: 115083 | |||||
| * | Assign DW_ACCESS_public accessibility attribute to members by default. | Devang Patel | 2010-09-29 | 1 | -1/+10 | |
| | | | | | llvm-svn: 115067 | |||||
| * | Revert r114997. It was causing a failure on darwin10-selfhost. | Bill Wendling | 2010-09-28 | 1 | -5/+13 | |
| | | | | | llvm-svn: 115002 | |||||
| * | Removed a bunch of unnecessary target_link_libraries. | Oscar Fuentes | 2010-09-28 | 3 | -5/+0 | |
| | | | | | llvm-svn: 114999 | |||||
| * | Fix a FIXME. _foo.eh symbols are currently always exported so that the linker | Bill Wendling | 2010-09-28 | 1 | -13/+5 | |
| | | | | | | | knows about them. This is not necessary on 10.6 and later. llvm-svn: 114997 | |||||
| * | Give the if-converter access to MachineLoopInfo, and use it to generate ↵ | Owen Anderson | 2010-09-28 | 1 | -16/+50 | |
| | | | | | | | | | plausible branch prediction estimates. llvm-svn: 114981 | |||||
| * | Part one of switching to using a more sane heuristic for determining ↵ | Owen Anderson | 2010-09-28 | 1 | -4/+4 | |
| | | | | | | | | | | | | if-conversion profitability. Rather than having arbitrary cutoffs, actually try to cost model the conversion. For now, the constants are tuned to more or less match our existing behavior, but these will be changed to reflect realistic values as this work proceeds. llvm-svn: 114973 | |||||
| * | Provide an interface to let FEs anchor debug info for types. | Devang Patel | 2010-09-28 | 1 | -0/+4 | |
| | | | | | llvm-svn: 114969 | |||||
| * | Remove dead argument. | Devang Patel | 2010-09-27 | 2 | -5/+4 | |
| | | | | | llvm-svn: 114920 | |||||
| * | Don't try to make a vector of x86mmx; this won't work, | Dale Johannesen | 2010-09-27 | 1 | -2/+4 | |
| | | | | | | | and asserts. llvm-svn: 114843 | |||||
| * | the latest assembler that runs on powerpc 10.4 machines doesn't | Chris Lattner | 2010-09-27 | 2 | -2/+24 | |
| | | | | | | | | support aligned comm. Detect when compiling for 10.4 and don't emit an alignment for comm. THis will hopefully fix PR8198. llvm-svn: 114817 | |||||
| * | Fixed some tests to avoid LiveIntervals::getInstructionFromIndex(..) ↵ | Lang Hames | 2010-09-26 | 3 | -9/+8 | |
| | | | | | | | overhead where possible. Thanks to Jakob for the suggestions. llvm-svn: 114798 | |||||
| * | Avoid using VNInfo::getCopy as much as possible. I want to get rid of it. | Jakob Stoklund Olesen | 2010-09-25 | 1 | -10/+6 | |
| | | | | | llvm-svn: 114794 | |||||
| * | Removed VNInfo::isDefAccurate(). Def "accuracy" can be checked by testing ↵ | Lang Hames | 2010-09-25 | 10 | -76/+75 | |
| | | | | | | | whether LiveIntervals::getInstructionFromIndex(def) returns NULL. llvm-svn: 114791 | |||||
| * | Remove SlotIndex::PHI_BIT. It is no longer used by anything. | Jakob Stoklund Olesen | 2010-09-25 | 3 | -11/+4 | |
| | | | | | llvm-svn: 114779 | |||||
| * | Remove the only use of SlotIndex::isPHI. This bit is not being set consistently | Jakob Stoklund Olesen | 2010-09-25 | 1 | -1/+1 | |
| | | | | | | | and it will be removed shortly. llvm-svn: 114778 | |||||
| * | Terminator gaps were unused. Might as well delete them. | Jakob Stoklund Olesen | 2010-09-24 | 1 | -16/+1 | |
| | | | | | llvm-svn: 114776 | |||||
| * | Fix for test/CodeGen/PowerPC/2008-10-17-AsmMatchingOperands.ll crash. | John Thompson | 2010-09-24 | 1 | -0/+20 | |
| | | | | | llvm-svn: 114767 | |||||
| * | Get rid of pop_macro warnings on MSVC. | Michael J. Spencer | 2010-09-24 | 2 | -6/+10 | |
| | | | | | llvm-svn: 114750 | |||||
| * | Attach a DebugLoc to a GC point in order to get precise information in the ↵ | Nicolas Geoffray | 2010-09-24 | 1 | -6/+8 | |
| | | | | | | | JIT of a GC point. llvm-svn: 114736 | |||||
| * | Revert 114634 for now since buildbot claim it broke Clang self-hosting. I ↵ | Evan Cheng | 2010-09-23 | 1 | -4/+6 | |
| | | | | | | | doubt it but it's possible it's exposing another bug somewhere. llvm-svn: 114681 | |||||
| * | Fix VS 2010 build. | Oscar Fuentes | 2010-09-23 | 1 | -4/+4 | |
| | | | | | | | Patch by Nathan Jeffords! llvm-svn: 114661 | |||||
| * | Follow up to r114630. Do not optimize away unconditional branch following a ↵ | Evan Cheng | 2010-09-23 | 1 | -6/+4 | |
| | | | | | | | conditional one. llvm-svn: 114634 | |||||
| * | Don't sink insert_subreg, subreg_to_reg, reg_sequence. They are meant to be | Evan Cheng | 2010-09-23 | 1 | -1/+10 | |
| | | | | | | | close to their sources to facilitate coalescing. llvm-svn: 114631 | |||||
| * | SDISel should not optimize a unconditional branch following a conditional branch | Evan Cheng | 2010-09-23 | 1 | -4/+5 | |
| | | | | | | | | | | | | | when the unconditional branch destination is the fallthrough block. The canonicalization makes it easier to allow optimizations on DAGs to invert conditional branches. The branch folding pass (and AnalyzeBranch) will clean up the unnecessary unconditional branches later. This is one of the patches leading up to disabling codegen prepare critical edge splitting. llvm-svn: 114630 | |||||
| * | Moved the PBQP allocator class out of the header and back in to the cpp file ↵ | Lang Hames | 2010-09-23 | 1 | -60/+190 | |
| | | | | | | | | | | | | | | | to hide the gory details. Allocator instances can now be created by calling createPBQPRegisterAllocator. Tidied up use of CoalescerPair as per Jakob's suggestions. Made the new PBQPBuilder based construction process the default. The internal construction process remains in-place and available via -pbqp-builder=false for now. It will be removed shortly if the new process doesn't cause any regressions. llvm-svn: 114626 | |||||
| * | A select between a constant and zero, when fed by a bit test, can be efficiently | Owen Anderson | 2010-09-22 | 1 | -0/+29 | |
| | | | | | | | | lowered using a series of shifts. Fixes <rdar://problem/8285015>. llvm-svn: 114599 | |||||
| * | Use DW_OP_fbreg when offset is based on frame register. | Devang Patel | 2010-09-22 | 1 | -0/+10 | |
| | | | | | llvm-svn: 114585 | |||||
| * | Build the complement interval dupli after the split intervals instead of | Jakob Stoklund Olesen | 2010-09-21 | 2 | -35/+145 | |
| | | | | | | | | | | | | | | creating it before and subtracting split ranges. This way, the SSA update code in LiveIntervalMap can properly create and use new phi values in dupli. Now it is possible to create split regions where a value escapes along two different CFG edges, creating phi values outside the split region. This is a work in progress and probably quite broken. llvm-svn: 114492 | |||||
| * | Fixed pr20314-2.c failure, added E, F, p constraint letters. | John Thompson | 2010-09-21 | 1 | -6/+17 | |
| | | | | | llvm-svn: 114490 | |||||
| * | Rework passing parent pointers into complexpatterns, I forgot | Chris Lattner | 2010-09-21 | 1 | -27/+35 | |
| | | | | | | | | | that complex patterns are matched after the entire pattern has a structural match, therefore the NodeStack isn't in a useful state when the actual call to the matcher happens. llvm-svn: 114489 | |||||
| * | If only user of a vreg is an copy instruction to export copy of vreg out of ↵ | Devang Patel | 2010-09-21 | 1 | -0/+23 | |
| | | | | | | | | | | current basic block then insert DBG_VALUE so that debug value of the variable is also transfered to new vreg. Testcase is in r114476. This fixes radar 8412415. llvm-svn: 114478 | |||||
| * | correct this logic. | Chris Lattner | 2010-09-21 | 1 | -2/+2 | |
| | | | | | llvm-svn: 114474 | |||||
| * | Reimplement r114460 in target-independent DAGCombine rather than ↵ | Owen Anderson | 2010-09-21 | 1 | -0/+14 | |
| | | | | | | | | | | target-dependent, by using the predicate to discover the number of sign bits. Enhance X86's target lowering to provide a useful response to this query. llvm-svn: 114473 | |||||
| * | just like they can opt into getting the root of the pattern being | Chris Lattner | 2010-09-21 | 1 | -1/+5 | |
| | | | | | | | | matched, allow ComplexPatterns to opt into getting the parent node of the operand being matched. llvm-svn: 114472 | |||||
| * | Refix MSVC9 and upper_bound. It actually needs a fully symmetric comparator. | Jakob Stoklund Olesen | 2010-09-21 | 1 | -7/+5 | |
| | | | | | llvm-svn: 114469 | |||||
| * | finish pushing MachinePointerInfo through selectiondags. At this point, | Chris Lattner | 2010-09-21 | 2 | -16/+7 | |
| | | | | | | | | I think I've audited all uses, so it should be dependable for address spaces, and the pointer+offset info should also be accurate when there. llvm-svn: 114464 | |||||
| * | convert a couple more places to use the new getStore() | Chris Lattner | 2010-09-21 | 1 | -0/+4 | |
| | | | | | llvm-svn: 114463 | |||||
| * | update a bunch of code to use the MachinePointerInfo version of getStore. | Chris Lattner | 2010-09-21 | 5 | -34/+25 | |
| | | | | | llvm-svn: 114461 | |||||
| * | Don't pollute the global namespace. | Jakob Stoklund Olesen | 2010-09-21 | 1 | -0/+2 | |
| | | | | | llvm-svn: 114459 | |||||
| * | MSVC9 does not support upper_bound with an asymmetric comparator. | Jakob Stoklund Olesen | 2010-09-21 | 1 | -6/+10 | |
| | | | | | llvm-svn: 114455 | |||||
| * | Define the TargetLowering::getTgtMemIntrinsic hook for ARM so that NEON load | Bob Wilson | 2010-09-21 | 1 | -1/+2 | |
| | | | | | | | and store intrinsics are represented with MemIntrinsicSDNodes. llvm-svn: 114454 | |||||
| * | eliminate an old SelectionDAG::getTruncStore method, propagating | Chris Lattner | 2010-09-21 | 5 | -83/+63 | |
| | | | | | | | MachinePointerInfo around more. llvm-svn: 114452 | |||||
| * | eliminate last SelectionDAG::getLoad old entrypoint, on to stores. | Chris Lattner | 2010-09-21 | 3 | -22/+7 | |
| | | | | | llvm-svn: 114450 | |||||
| * | fix the code that infers SV info to be correct when dealing | Chris Lattner | 2010-09-21 | 1 | -28/+53 | |
| | | | | | | | with an indexed load/store that has an offset in the index. llvm-svn: 114449 | |||||

