| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Clean up the public interface of RegScavenger. | Jakob Stoklund Olesen | 2009-08-16 | 2 | -44/+20 |
| | | | | | | | Remove unused methods and make others private. llvm-svn: 79194 | ||||
| * | Be more clever about regclasses in ScheduleDAGSDNodes::EmitCopyFromReg. | Jakob Stoklund Olesen | 2009-08-16 | 1 | -5/+5 |
| | | | | | | | | | If two uses of a CopyFromReg want different regclasses, first try a common sub-class, then fall back on the copy emitted in AddRegisterOperand. There is no need for an assert here. The cross-class joiner usually cleans up nicely. llvm-svn: 79193 | ||||
| * | Fix use after free in Thumb2SizeReduction (PR4707). A MachineInstr was used ↵ | Benjamin Kramer | 2009-08-16 | 1 | -1/+4 |
| | | | | | | | after erasure. llvm-svn: 79189 | ||||
| * | Styalistic and format changes. No functionality change. | Bill Wendling | 2009-08-16 | 1 | -73/+85 |
| | | | | | llvm-svn: 79187 | ||||
| * | Fix the MSVC build again. | Benjamin Kramer | 2009-08-16 | 1 | -1/+1 |
| | | | | | llvm-svn: 79186 | ||||
| * | CMake: Hopefully this will fix the build on VS. I can't replicate the | Oscar Fuentes | 2009-08-16 | 2 | -1/+1 |
| | | | | | | | | | | | | | failure with VS 9.0, nmake and cmake 2.6.4. The buildbot output does not show the patch level of cmake, it just says 2.6. Sadly, parallel builds are broken due to recent changes on LLVM Target libraries and its auxiliaries (TargetInfo, AsmPrinter, AsmParser). I have a patch for stablishing the correct dependencies, but cmake is buggy and generates makefiles that can't handle them. llvm-svn: 79180 | ||||
| * | CMake: Revert r79144. It reverted a change necessary for correct | Oscar Fuentes | 2009-08-16 | 2 | -0/+9 |
| | | | | | | | parallel builds. llvm-svn: 79177 | ||||
| * | add support for external symbols + X86::MOVPC32r. | Chris Lattner | 2009-08-16 | 2 | -59/+123 |
| | | | | | llvm-svn: 79175 | ||||
| * | Fix for PR3016: detect the tricky case, where there are | Eli Friedman | 2009-08-16 | 2 | -60/+37 |
| | | | | | | | | | | | | | | | | | unfoldable references to a PHI node in the block being folded, and disable the transformation in that case. The correct transformation of such PHI nodes depends on whether BB dominates Succ, and dominance is expensive to compute here. (Alternatively, it's possible to check whether any uses are live, but that's also essentially a dominance calculation. Another alternative is to use reg2mem, but it probably isn't a good idea to use that in simplifycfg.) Also, remove some incorrect code from CanPropagatePredecessorsForPHIs which is made unnecessary with this patch: it didn't consider the case where a PHI node in BB has multiple uses. llvm-svn: 79174 | ||||
| * | implement support for lowering references to global addresses. For example, ↵ | Chris Lattner | 2009-08-16 | 2 | -5/+53 |
| | | | | | | | | | | | | | | we now can asmprint: NEW: movl "L___stack_chk_guard$non_lazy_ptr", %eax OLD: movl L___stack_chk_guard$non_lazy_ptr, %eax where 'new' is coming out of the MCInst version of the printer. llvm-svn: 79170 | ||||
| * | more formatting improvements, no functionality change. | Chris Lattner | 2009-08-16 | 1 | -57/+65 |
| | | | | | llvm-svn: 79167 | ||||
| * | code formatting improvements, no functionality change. | Chris Lattner | 2009-08-16 | 1 | -74/+85 |
| | | | | | llvm-svn: 79165 | ||||
| * | when emitting errors about CHECK-NEXT directives, show the line that the | Chris Lattner | 2009-08-16 | 1 | -2/+6 |
| | | | | | | | CHECK-NEXT is on. llvm-svn: 79164 | ||||
| * | Add an llvm-c function that lets you insert an instruction with a name. | Erick Tryzelaar | 2009-08-16 | 5 | -0/+22 |
| | | | | | llvm-svn: 79163 | ||||
| * | Expose most of the Constant creation functions to ocaml. | Erick Tryzelaar | 2009-08-16 | 3 | -4/+148 |
| | | | | | llvm-svn: 79162 | ||||
| * | Expose most of the IRBuilder functionality to ocaml. | Erick Tryzelaar | 2009-08-16 | 3 | -3/+315 |
| | | | | | llvm-svn: 79161 | ||||
| * | Expose most of the Constant creation functions to llvm-c. | Erick Tryzelaar | 2009-08-16 | 2 | -2/+108 |
| | | | | | llvm-svn: 79160 | ||||
| * | Expose most of the rest of IRBuilder's functions to llvm-c. | Erick Tryzelaar | 2009-08-16 | 2 | -0/+140 |
| | | | | | llvm-svn: 79159 | ||||
| * | Add more casts to the IRBuilder. | Erick Tryzelaar | 2009-08-16 | 2 | -1/+52 |
| | | | | | llvm-svn: 79158 | ||||
| * | Needs to check whether unaligned load / store of i64 is legal here. | Evan Cheng | 2009-08-15 | 1 | -4/+3 |
| | | | | | llvm-svn: 79150 | ||||
| * | Remove <iostream>. | Bill Wendling | 2009-08-15 | 1 | -20/+17 |
| | | | | | llvm-svn: 79146 | ||||
| * | Revert 78996. It was breaking MSVC builds. | Benjamin Kramer | 2009-08-15 | 2 | -9/+0 |
| | | | | | llvm-svn: 79144 | ||||
| * | Mingw also doesn't have st_blksize. | Dan Gohman | 2009-08-15 | 1 | -1/+1 |
| | | | | | llvm-svn: 79142 | ||||
| * | Fix test on Linux. | Eli Friedman | 2009-08-15 | 1 | -1/+1 |
| | | | | | llvm-svn: 79140 | ||||
| * | An overhaul of the exception handling code. This is arguably more correct than | Bill Wendling | 2009-08-15 | 1 | -55/+81 |
| | | | | | | | | | | | | | | what was there before. In "no FP mode", we weren't generating labels and unwind table entries after each "push" instruction. While more than likely "okay", it's not technically correct. The major thing was that the ordering of when to define a new CFA register and at what offset wasn't correct. This would cause the exception handling to fail in ways most miserable to users. I also cleaned up some code a bit. There's one function which has a "return" at the beginning, so it's never used. Should I just remove it? :-) llvm-svn: 79139 | ||||
| * | Reapply r79127. It was fixed by d0k. | Bill Wendling | 2009-08-15 | 7 | -33/+76 |
| | | | | | llvm-svn: 79136 | ||||
| * | Revert r79127. It was causing compilation errors. | Bill Wendling | 2009-08-15 | 7 | -76/+33 |
| | | | | | llvm-svn: 79135 | ||||
| * | Don't try to get the context from an erased Instruction. | Benjamin Kramer | 2009-08-15 | 1 | -1/+1 |
| | | | | | llvm-svn: 79134 | ||||
| * | Unbreak build. Evan, please make sure my changes are correct. | Benjamin Kramer | 2009-08-15 | 3 | -6/+5 |
| | | | | | llvm-svn: 79133 | ||||
| * | SSI construction should just go ahead and ignore instructions in unreachable | Nick Lewycky | 2009-08-15 | 2 | -2/+25 |
| | | | | | | | blocks. llvm-svn: 79132 | ||||
| * | Fix validation errors. | Bill Wendling | 2009-08-15 | 1 | -3/+3 |
| | | | | | llvm-svn: 79131 | ||||
| * | Fix validation errors. | Bill Wendling | 2009-08-15 | 1 | -11/+11 |
| | | | | | llvm-svn: 79130 | ||||
| * | Change allowsUnalignedMemoryAccesses to take type argument since some targets | Evan Cheng | 2009-08-15 | 7 | -33/+76 |
| | | | | | | | | | | | support unaligned mem access only for certain types. (Should it be size instead?) ARM v7 supports unaligned access for i16 and i32, some v6 variants support it as well. llvm-svn: 79127 | ||||
| * | wrap to 80 cols | Chris Lattner | 2009-08-15 | 1 | -2/+4 |
| | | | | | llvm-svn: 79124 | ||||
| * | implement support for CHECK-NEXT: in filecheck. | Chris Lattner | 2009-08-15 | 3 | -37/+157 |
| | | | | | llvm-svn: 79123 | ||||
| * | Refine EarlyClobber assert in register scavenger. | Jakob Stoklund Olesen | 2009-08-15 | 2 | -1/+43 |
| | | | | | | | | | | | | | | | It is legal for an inline asm operand to use an earlyclobber register if the use operand is tied to the earlyclobber operand. The issue is discussed here: http://gcc.gnu.org/ml/gcc/1999-04n/msg00431.html We should perhaps let only the machine code verifier worry about these finer details. EarlyClobber operands are not really interesting to the scavenger. This fixes PR4528 for the third time. llvm-svn: 79122 | ||||
| * | simplify some code. | Chris Lattner | 2009-08-15 | 1 | -8/+13 |
| | | | | | llvm-svn: 79121 | ||||
| * | rewrite FindStringInBuffer to use an explicit loop instead of | Chris Lattner | 2009-08-15 | 1 | -19/+27 |
| | | | | | | | | trying to wrap strstr which is just too inconvenient. Make it use a StringRef to avoid ".c_str()" calls. llvm-svn: 79120 | ||||
| * | Instead of using an std::pair, use a custom struct. | Chris Lattner | 2009-08-15 | 1 | -12/+21 |
| | | | | | llvm-svn: 79119 | ||||
| * | specify a target triple so global variable manglings are consistent etc. | Chris Lattner | 2009-08-15 | 1 | -1/+2 |
| | | | | | llvm-svn: 79118 | ||||
| * | convert to filecheck. | Chris Lattner | 2009-08-15 | 1 | -3/+18 |
| | | | | | llvm-svn: 79117 | ||||
| * | rename this test to sse2.ll | Chris Lattner | 2009-08-15 | 1 | -0/+0 |
| | | | | | llvm-svn: 79116 | ||||
| * | merge a bunch more sse3 tests into sse3.ll | Chris Lattner | 2009-08-15 | 11 | -174/+252 |
| | | | | | llvm-svn: 79115 | ||||
| * | convert test to filecheck format. | Chris Lattner | 2009-08-15 | 1 | -7/+17 |
| | | | | | llvm-svn: 79114 | ||||
| * | rename test | Chris Lattner | 2009-08-15 | 1 | -0/+0 |
| | | | | | llvm-svn: 79113 | ||||
| * | this is a test for sse3, simplify it. | Chris Lattner | 2009-08-15 | 1 | -1/+1 |
| | | | | | llvm-svn: 79112 | ||||
| * | the .eh_frame sections we generate need to be writable (which | Chris Lattner | 2009-08-15 | 1 | -1/+2 |
| | | | | | | | | is why they are datarel). This should fix PR4724, and is fallout from r78890. llvm-svn: 79111 | ||||
| * | document filecheck. | Chris Lattner | 2009-08-15 | 1 | -3/+87 |
| | | | | | llvm-svn: 79110 | ||||
| * | fix pasto noticed by Jakub | Chris Lattner | 2009-08-15 | 1 | -1/+1 |
| | | | | | llvm-svn: 79109 | ||||
| * | Use the new API for creating an OpaqueType. | Nicolas Geoffray | 2009-08-15 | 1 | -2/+4 |
| | | | | | llvm-svn: 79107 | ||||

