| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Teach ScalarEvolution how to reason about no-wrap flags on loops | Dan Gohman | 2009-09-17 | 1 | -0/+76 |
| | | | | | | | | | where the induction variable has a non-unit stride, such as {0,+,2}, and there are expressions such as {1,+,2} inside the loop formed with or or add nsw operators. llvm-svn: 82151 | ||||
| * | Remove test cases using -regalloc=simple. | Daniel Dunbar | 2009-09-17 | 2 | -22/+0 |
| | | | | | llvm-svn: 82130 | ||||
| * | Fix PR4910: Broken logic in coalescer means when a physical register ↵ | Evan Cheng | 2009-09-17 | 1 | -0/+64 |
| | | | | | | | liveness is being shortened, the sub-registers were not. The symptom is the register allocator could not find a free register for this particular test. llvm-svn: 82108 | ||||
| * | Fix the comment in this test. | Dan Gohman | 2009-09-16 | 1 | -1/+2 |
| | | | | | llvm-svn: 82051 | ||||
| * | Don't sink gep operators through phi nodes if the result would require | Dan Gohman | 2009-09-16 | 1 | -0/+101 |
| | | | | | | | | | more than one phi, since that leads to higher register pressure on entry to the phi. This is especially problematic when the phi is in a loop header, as it increases register pressure throughout the loop. llvm-svn: 81993 | ||||
| * | Convert more tests to FileCheck. | Bob Wilson | 2009-09-15 | 6 | -45/+153 |
| | | | | | llvm-svn: 81915 | ||||
| * | fix PR4984 by ensuring that fastisel adds properly sign extended GEP ↵ | Chris Lattner | 2009-09-15 | 1 | -2/+22 |
| | | | | | | | | | displacement values to machineinstrs. llvm-svn: 81886 | ||||
| * | rename test | Chris Lattner | 2009-09-15 | 1 | -0/+0 |
| | | | | | llvm-svn: 81884 | ||||
| * | convert to filecheck | Chris Lattner | 2009-09-15 | 1 | -4/+20 |
| | | | | | llvm-svn: 81882 | ||||
| * | Fix superreg use in ARMAsmPrinter. Approved by Anton Korobeynikov. | Sandeep Patel | 2009-09-15 | 1 | -0/+41 |
| | | | | | llvm-svn: 81878 | ||||
| * | several major improvements to the sparc backend: support for weak linkage | Chris Lattner | 2009-09-15 | 2 | -0/+15 |
| | | | | | | | and PIC codegen. Patch by Venkatraman Govindaraju! llvm-svn: 81877 | ||||
| * | When a constant's type is refined, update the constant in place | Dan Gohman | 2009-09-15 | 2 | -0/+44 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | instead of cloning and RAUWing it. - Make AbstractTypeUser a friend of Value so that it can offer its subclasses a way to update a Value's type in place. This is better than a universally visible setType method on Value, and it's sufficient for the immediate need. - Eliminate the constant "convert" functions. This eliminates a lot of logic duplication, and fixes a complicated bug where a constant can't actually be cloned during the type refinement process because some of the types that its folder needs are half-destroyed, being in the middle of refinement themselves. - Move the getValType functions from being static overloaded functions in Constants.cpp to be members of class template specializations in ConstantsContext.h. This means that the code ends up getting instantiated twice, however it also makes it possible to eliminate all "convert" functions, so it's not a big net code size increase. And if desired, the duplicate instantiations could be eliminated with some reorganization. llvm-svn: 81861 | ||||
| * | Use llvm-link -S instead of using llvm-dis. | Dan Gohman | 2009-09-15 | 21 | -24/+24 |
| | | | | | llvm-svn: 81860 | ||||
| * | Restore a comment that was lost in the merge. | Dan Gohman | 2009-09-15 | 1 | -0/+4 |
| | | | | | llvm-svn: 81857 | ||||
| * | convert to filecheck | Chris Lattner | 2009-09-15 | 1 | -9/+10 |
| | | | | | llvm-svn: 81848 | ||||
| * | Forbid arrays of function-type and structures with function-typed fields. | Nick Lewycky | 2009-09-15 | 1 | -0/+6 |
| | | | | | | | | | | | While I'm there, change code that does: SomeTy == Type::getFooType(Context) into: SomeTy->getTypeID() == FooTyID to decrease the amount of useless type creation which may involve locking, etc. llvm-svn: 81846 | ||||
| * | fix PR4963: folding insertvalue would sometimes turn a packed struct into | Chris Lattner | 2009-09-15 | 1 | -0/+6 |
| | | | | | | | an unpacked one. llvm-svn: 81845 | ||||
| * | this is failing on linux hosts, force a triple. | Chris Lattner | 2009-09-15 | 1 | -1/+1 |
| | | | | | llvm-svn: 81833 | ||||
| * | merge one more in. | Chris Lattner | 2009-09-15 | 2 | -12/+10 |
| | | | | | llvm-svn: 81824 | ||||
| * | merge some more cmov tests into cmov.ll | Chris Lattner | 2009-09-15 | 3 | -87/+95 |
| | | | | | llvm-svn: 81823 | ||||
| * | merge two cmov tests into one. | Chris Lattner | 2009-09-15 | 3 | -49/+48 |
| | | | | | llvm-svn: 81822 | ||||
| * | Don't pull a load through a callseq_start if the load's chain | Dan Gohman | 2009-09-15 | 1 | -0/+23 |
| | | | | | | | | | | has multiple uses, as one of the other uses may be on a path to a different node above the callseq_start, because that leads to a cyclic graph. This problem is exposed when -combiner-global-alias-analysis is used. This fixes PR4880. llvm-svn: 81821 | ||||
| * | Added the first bits of the ARM target assembler to llvm-mc. For now it only | Kevin Enderby | 2009-09-15 | 3 | -5/+12 |
| | | | | | | | | | parses the .word directive as 4 bytes and ARMAsmParser::ParseInstruction will give an error is called. Broke out the test of the .word directive into two different test cases, one for x86 and one for arm. llvm-svn: 81817 | ||||
| * | On x86-64, the 32-bit cmov doesn't actually clear the high 32-bit of | Dan Gohman | 2009-09-15 | 1 | -0/+19 |
| | | | | | | | its result if the condition is false. llvm-svn: 81814 | ||||
| * | When extending a memset range past the front, set the alignment of the | Dan Gohman | 2009-09-14 | 1 | -0/+18 |
| | | | | | | | memset region to the alignment of the new start address. llvm-svn: 81810 | ||||
| * | Add a VALGRIND_EXTRA_ARGS makefile variable, with the obvious semantics. | Daniel Dunbar | 2009-09-14 | 1 | -1/+1 |
| | | | | | llvm-svn: 81764 | ||||
| * | Teach 'make check-lit' to run unittests. | Daniel Dunbar | 2009-09-14 | 2 | -1/+81 |
| | | | | | llvm-svn: 81753 | ||||
| * | unbreak this test by working around an asmparser bug. | Chris Lattner | 2009-09-13 | 1 | -1/+1 |
| | | | | | llvm-svn: 81724 | ||||
| * | Fix a small issue with recent changes to this code. | Mikhail Glushenkov | 2009-09-13 | 2 | -12/+7 |
| | | | | | | | The 'false.c' file wasn't being used. llvm-svn: 81709 | ||||
| * | merge the linux cpool/jtbl pic tests into pic.ll and convert to filecheck. | Chris Lattner | 2009-09-13 | 3 | -92/+100 |
| | | | | | | | | Change the picbase symbol on non-darwin systems from ".Lllvm$4.$piclabel" to ".L4$pb". The actual name doesn't matter and the darwin name is shorter. llvm-svn: 81688 | ||||
| * | Add LLVMGCCBINDIR to path, since LLVMC expects to find llvm-gcc in the path. | Daniel Dunbar | 2009-09-13 | 1 | -1/+2 |
| | | | | | llvm-svn: 81669 | ||||
| * | Switch Ocaml to use llvm_supports_binding. | Daniel Dunbar | 2009-09-13 | 3 | -5/+5 |
| | | | | | llvm-svn: 81665 | ||||
| * | tests: Add llvm_supports_binding predicate. | Daniel Dunbar | 2009-09-13 | 3 | -0/+17 |
| | | | | | llvm-svn: 81664 | ||||
| * | tests: Use %abs_tmp instead of ./%t to make these tests portable to 'lit'. | Daniel Dunbar | 2009-09-13 | 9 | -9/+9 |
| | | | | | llvm-svn: 81663 | ||||
| * | tests: Add a %abs_tmp substitution which is guaranteed to be a full path. | Daniel Dunbar | 2009-09-13 | 1 | -0/+2 |
| | | | | | llvm-svn: 81662 | ||||
| * | Sink llvm-gcc dependent tests into distinct subdirs. | Daniel Dunbar | 2009-09-13 | 17 | -21/+25 |
| | | | | | llvm-svn: 81661 | ||||
| * | Rewrite tests to not use Tcl substitution. | Daniel Dunbar | 2009-09-13 | 11 | -11/+11 |
| | | | | | llvm-svn: 81660 | ||||
| * | Simplify LLVMC tests. | Daniel Dunbar | 2009-09-13 | 3 | -10/+12 |
| | | | | | llvm-svn: 81659 | ||||
| * | Define proper subreg sets for arm - this should fix bunch of subtle problems | Anton Korobeynikov | 2009-09-13 | 1 | -0/+61 |
| | | | | | | | with subreg - superreg mapping and also fix PR4965. llvm-svn: 81657 | ||||
| * | Add -mattr=+sse2 to the -march=x86 version of this test. Without | Dan Gohman | 2009-09-12 | 1 | -1/+1 |
| | | | | | | | | sse, this code falls back to SelectionDAG isel which uses an x87 instruction, which is fine, but not what this test is testing for. llvm-svn: 81656 | ||||
| * | Remove an unnecessary -f. | Dan Gohman | 2009-09-11 | 1 | -1/+1 |
| | | | | | llvm-svn: 81546 | ||||
| * | Convert more tests to avoid llvm-as. | Dan Gohman | 2009-09-11 | 86 | -181/+112 |
| | | | | | llvm-svn: 81545 | ||||
| * | Eliminate more redundant llvm-as calls. | Dan Gohman | 2009-09-11 | 20 | -26/+26 |
| | | | | | llvm-svn: 81540 | ||||
| * | Fix this test to test what it was originally intended to test. | Dan Gohman | 2009-09-11 | 1 | -1/+1 |
| | | | | | llvm-svn: 81539 | ||||
| * | Change tests from "opt %s" to "opt < %s" so that opt doesn't see the | Dan Gohman | 2009-09-11 | 1273 | -1366/+1366 |
| | | | | | | | | | input filename so that opt doesn't print the input filename in the output so that grep lines in the tests don't unintentionally match strings in the input filename. llvm-svn: 81537 | ||||
| * | turn on -experimental-asm-printer for x86 / AT&T by default. | Chris Lattner | 2009-09-11 | 2 | -494/+494 |
| | | | | | llvm-svn: 81532 | ||||
| * | another random update | Chris Lattner | 2009-09-11 | 1 | -1/+1 |
| | | | | | llvm-svn: 81531 | ||||
| * | fix a bunch of spurious failures for people whose home directory | Chris Lattner | 2009-09-11 | 15 | -20/+16 |
| | | | | | | | is sabre. llvm-svn: 81528 | ||||
| * | this test is using invalid "intrinsics". | Chris Lattner | 2009-09-11 | 1 | -20/+0 |
| | | | | | llvm-svn: 81527 | ||||
| * | fix test to not get a moduleid that matches 'br' | Chris Lattner | 2009-09-11 | 1 | -1/+1 |
| | | | | | llvm-svn: 81526 | ||||

