| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Fix a bug in the x86-32 code generated for segmented stacks. | Sanjoy Das | 2011-12-03 | 1 | -8/+0 |
| | | | | | | | | | | | Currently LLVM pads the call to __morestack with a add and sub of 8 bytes to esp. This isn't correct since __morestack expects the call to be followed directly by a ret. This commit also adjusts the relevant test-case. llvm-svn: 145765 | ||||
| * | Creating multiple JITs on X86 in multiple threads causes multiple writes (of | Nick Lewycky | 2011-12-03 | 1 | -0/+2 |
| | | | | | | | | | the same value) to this variable. This code could be refactored, but it doesn't matter since the old JIT is going away. Add tsan annotations to ignore the race. llvm-svn: 145745 | ||||
| * | [arm-fast-isel] Unaligned stores of floats require special care. | Chad Rosier | 2011-12-03 | 1 | -3/+18 |
| | | | | | | | rdar://10510150 llvm-svn: 145742 | ||||
| * | Fixed deadstoreelimination bug where negative indices were incorrectly ↵ | Pete Cooper | 2011-12-03 | 1 | -1/+1 |
| | | | | | | | | | | | causing the optimisation to occur Turns out long long + unsigned long long is unsigned. Doh! Fixes http://llvm.org/bugs/show_bug.cgi?id=11455 llvm-svn: 145731 | ||||
| * | Add support for constant folding the pow intrinsic. | Chad Rosier | 2011-12-03 | 1 | -3/+6 |
| | | | | | | | rdar://10514247 llvm-svn: 145730 | ||||
| * | ARM NEON VEXT aliases for data type suffices. | Jim Grosbach | 2011-12-02 | 3 | -11/+56 |
| | | | | | llvm-svn: 145726 | ||||
| * | ARM VEXT tighten up operand classes a bit. | Jim Grosbach | 2011-12-02 | 1 | -2/+2 |
| | | | | | llvm-svn: 145722 | ||||
| * | ARM VST1 single lane assembly parsing. | Jim Grosbach | 2011-12-02 | 2 | -4/+179 |
| | | | | | llvm-svn: 145718 | ||||
| * | Move global variables in TargetMachine into new TargetOptions class. As an API | Nick Lewycky | 2011-12-02 | 54 | -463/+391 |
| | | | | | | | | | | | | | change, now you need a TargetOptions object to create a TargetMachine. Clang patch to follow. One small functionality change in PTX. PTX had commented out the machine verifier parts in their copy of printAndVerify. That now calls the version in LLVMTargetMachine. Users of PTX who need verification disabled should rely on not passing the command-line flag to enable it. llvm-svn: 145714 | ||||
| * | ARM VLD1 single lane assembly parsing. | Jim Grosbach | 2011-12-02 | 3 | -9/+124 |
| | | | | | llvm-svn: 145712 | ||||
| * | ARM encoder method needs the physical register number, not the enum. | Jim Grosbach | 2011-12-02 | 1 | -1/+1 |
| | | | | | llvm-svn: 145711 | ||||
| * | [arm-fast-isel] After promoting a function parameter be sure to update the | Chad Rosier | 2011-12-02 | 1 | -2/+4 |
| | | | | | | | | | argument value type. Otherwise, the sign/zero-extend has no effect on arguments passed via the stack (i.e., undefined high-order bits). rdar://10515467 llvm-svn: 145701 | ||||
| * | Clean up aliases for ARM VLD1 single-lane assembly parsing a bit. | Jim Grosbach | 2011-12-02 | 3 | -26/+133 |
| | | | | | | | Add the 16-bit lane variants while I'm at it. llvm-svn: 145693 | ||||
| * | Fix quadratic behavior in InlineFunction by fetching the personality ↵ | Benjamin Kramer | 2011-12-02 | 1 | -29/+33 |
| | | | | | | | | | | | function of the callee once and not for every invoke in the caller. The callee is usually smaller than the caller, too. This reduces the compile time of ARMDisassembler.cpp by 32% (Release build). It still takes ages to compile though. llvm-svn: 145690 | ||||
| * | Check for error after InstantiateMultclassDef. | Jim Grosbach | 2011-12-02 | 1 | -0/+2 |
| | | | | | llvm-svn: 145689 | ||||
| * | Add XOP feature flag. | Jan Sjödin | 2011-12-02 | 3 | -2/+12 |
| | | | | | llvm-svn: 145682 | ||||
| * | Reduce duplicate code in isHorizontalBinOp and add some asserts to protect ↵ | Craig Topper | 2011-12-02 | 1 | -32/+24 |
| | | | | | | | assumptions llvm-svn: 145681 | ||||
| * | Add instruction selection support for horizontal add/sub of 256-bit floating ↵ | Craig Topper | 2011-12-02 | 1 | -2/+4 |
| | | | | | | | point vectors. Also add the test case for 256-bit integer vectors. llvm-svn: 145680 | ||||
| * | remove unneeded FIXME comment | Hal Finkel | 2011-12-02 | 1 | -1/+0 |
| | | | | | llvm-svn: 145679 | ||||
| * | make sure ScheduleDAGInstrs::EmitSchedule does not crash when the first ↵ | Hal Finkel | 2011-12-02 | 1 | -5/+5 |
| | | | | | | | instruction in Sequence is a Noop llvm-svn: 145677 | ||||
| * | update PPC 940 hazard rec. to function in postRA mode | Hal Finkel | 2011-12-02 | 6 | -142/+93 |
| | | | | | llvm-svn: 145676 | ||||
| * | Fix a few more places where TargetData/TargetLibraryInfo is not being passed. | Chad Rosier | 2011-12-02 | 11 | -37/+76 |
| | | | | | | | Add FIXMEs to places that are non-trivial to fix. llvm-svn: 145661 | ||||
| * | ARM start parsing VLD1 single lane instructions. | Jim Grosbach | 2011-12-02 | 2 | -14/+115 |
| | | | | | | | | The alias pseudos need cleaned up for size suffix handling, but this gets the basics working. Will be cleaning up and adding more. llvm-svn: 145655 | ||||
| * | Abuse of mass replace isn't warranted even when the build is failing. Thanks | Chad Rosier | 2011-12-01 | 1 | -25/+29 |
| | | | | | | | for the suggestion, Eric. llvm-svn: 145643 | ||||
| * | Fix build by not assuming TLI is guaranteed. Will have to track down cases where | Chad Rosier | 2011-12-01 | 1 | -24/+23 |
| | | | | | | | TLI isn't being passed to ensure we don't miss opportunities to fold calls. llvm-svn: 145641 | ||||
| * | Prevent library calls from being folded if -fno-builtin has been specified. | Chad Rosier | 2011-12-01 | 1 | -23/+24 |
| | | | | | | | rdar://10500969 llvm-svn: 145639 | ||||
| * | CodeGen: fix CMake build | Dylan Noblesmith | 2011-12-01 | 1 | -0/+1 |
| | | | | | | | Missing file from r145629. llvm-svn: 145634 | ||||
| * | ExecutionEngine: honor optimization level | Dylan Noblesmith | 2011-12-01 | 3 | -8/+9 |
| | | | | | | | | | | | It was getting ignored after r144788. Also fix an accidental implicit cast from the OptLevel enum to an optional bool argument. MSVC warned on this, but gcc didn't. llvm-svn: 145633 | ||||
| * | Last bit of TargetLibraryInfo propagation. Also fixed a case for TargetData | Chad Rosier | 2011-12-01 | 7 | -39/+78 |
| | | | | | | | | where it appeared beneficial to pass. More of rdar://10500969 llvm-svn: 145630 | ||||
| * | Add a deterministic finite automaton based packetizer for VLIW architectures | Anshuman Dasgupta | 2011-12-01 | 1 | -0/+98 |
| | | | | | llvm-svn: 145629 | ||||
| * | Fix unreachable return & simplify some branches. | David Blaikie | 2011-12-01 | 1 | -9/+5 |
| | | | | | llvm-svn: 145627 | ||||
| * | Dummy commit to check commit access. | Sanjoy Das | 2011-12-01 | 1 | -1/+1 |
| | | | | | llvm-svn: 145619 | ||||
| * | Improved fix for abs(val) != 0 to check other similar case. Also fixed ↵ | Pete Cooper | 2011-12-01 | 1 | -6/+10 |
| | | | | | | | style issues and confusing comment llvm-svn: 145618 | ||||
| * | [asan] two minor fixes: use UnreachableInst after the neverreturn function ↵ | Kostya Serebryany | 2011-12-01 | 1 | -2/+3 |
| | | | | | | | call; use report_fatal_error when blacklist file can not be found llvm-svn: 145611 | ||||
| * | Add missing functions. | Chad Rosier | 2011-12-01 | 1 | -1/+4 |
| | | | | | llvm-svn: 145608 | ||||
| * | Autodetect bulldozers. | Benjamin Kramer | 2011-12-01 | 1 | -0/+2 |
| | | | | | llvm-svn: 145607 | ||||
| * | Add a few more functions to TargetLibraryInfo. More of rdar://10500969. | Chad Rosier | 2011-12-01 | 1 | -1/+16 |
| | | | | | llvm-svn: 145596 | ||||
| * | For 64-bit the rest of the general regs are ok for the q constraint. Make | Eric Christopher | 2011-12-01 | 1 | -1/+17 |
| | | | | | | | | | sure we can emit both the high and low versions of those registers. Fixes rdar://10392864 llvm-svn: 145579 | ||||
| * | Add some missing anchors. | David Blaikie | 2011-12-01 | 3 | -0/+7 |
| | | | | | llvm-svn: 145578 | ||||
| * | Pass AVX vectors which are arguments to varargs functions on the stack. ↵ | Eli Friedman | 2011-12-01 | 1 | -4/+9 |
| | | | | | | | <rdar://problem/10463281>. llvm-svn: 145573 | ||||
| * | Added instcombine pattern to spot comparing -val or val against 0. | Pete Cooper | 2011-12-01 | 1 | -0/+14 |
| | | | | | | | | | (val != 0) == (-val != 0) so "abs(val) != 0" becomes "val != 0" Fixes <rdar://problem/10482509> llvm-svn: 145563 | ||||
| * | Propagate TargetLibraryInfo throughout ConstantFolding.cpp and | Chad Rosier | 2011-12-01 | 13 | -259/+394 |
| | | | | | | | | InstructionSimplify.cpp. Other fixups as needed. Part of rdar://10500969 llvm-svn: 145559 | ||||
| * | Make use of "getScalarType()". No functionality change. | Nick Lewycky | 2011-12-01 | 1 | -8/+3 |
| | | | | | llvm-svn: 145556 | ||||
| * | Small fix for assembler generation on Darwin PPC64. Patch by Michael ↵ | Eli Friedman | 2011-12-01 | 1 | -1/+1 |
| | | | | | | | Kostylev. PR11437. llvm-svn: 145553 | ||||
| * | make asan work at -O0, llvm part. Patch by glider@google.com | Kostya Serebryany | 2011-11-30 | 1 | -1/+1 |
| | | | | | llvm-svn: 145530 | ||||
| * | Support for encoding all FMA4 instructions and tablegen patterns for all | Jan Sjödin | 2011-11-30 | 1 | -0/+349 |
| | | | | | | | remaining FMA4 instructions and intrinsics with tests. llvm-svn: 145525 | ||||
| * | Make GlobalMerge honor the preferred alignment on globals without an ↵ | Eli Friedman | 2011-11-30 | 1 | -1/+1 |
| | | | | | | | | | explicitly specified alignment. <rdar://problem/10497732>. llvm-svn: 145523 | ||||
| * | Remove unused variable | Matt Beaumont-Gay | 2011-11-30 | 1 | -2/+0 |
| | | | | | llvm-svn: 145517 | ||||
| * | ARM parsing for VLD1 all lanes, with writeback. | Jim Grosbach | 2011-11-30 | 4 | -43/+82 |
| | | | | | llvm-svn: 145510 | ||||
| * | Add a few functions to TargetLibraryInfo. | Chad Rosier | 2011-11-30 | 1 | -2/+50 |
| | | | | | llvm-svn: 145508 | ||||

