Commits

Show all
Author Commit Message Labels Comments Date
Bill Wendling
Rename Bender to Xcode-3.2.
Branches
Apple/Xcode-3.2
Bill Wendling
Merge r75857: Skip special LLVM prefix '1' while emitting linknage name. This prefix is used by LLVM to inform the asm printer to not emit usual global symbol prefix before the symbol name.
Branches
Apple/Bender
Bill Wendling
--- Merging r75183 into '.': U lib/Target/X86/X86RegisterInfo.cpp If -fomit-frame-pointer is used, we still need to record when the %esp register is modified. Otherwise, the unwinder will get confused. The old code (before I started my hacking) did this. It dropped on the floor, because I wasn't aware of this requirement. On the plus side, if we use "alloca" in a function, we create frame pointers even with -fomit-frame-pointer is enabled! This is a Good Thing…
Tags
llvmCore-2118
Branches
Apple/Bender
Bill Wendling
--- Merging r75047 into '.': U lib/Target/X86/X86RegisterInfo.cpp U lib/Target/X86/X86RegisterInfo.h DWARF requires frame moves be specified at specific times. If you have a prologue like this: __Z3fooi: Leh_func_begin1: LBB1_0: ## entry pushl %ebp Llabel1: movl %esp, %ebp Llabel2: pushl %esi Llabel3: subl $20, %esp call "L1$pb" "L1$pb": popl %esi The "pushl %ebp" needs a table entry specifying the offset. The "movl %esp, %ebp" makes %eb…
Branches
Apple/Bender
Bill Wendling
In the current AMD backend (AMDILISelLowering.cpp), they currently set extract subvector to be custom lowered but LLVM 2115 doesn't currently allow custom lowering of that node, e.g., in select_uchar8_uchar8 0x100879198: v2i8 = extract_subvector 0x100878fa8, 0x100879290 This gets transformed to 0x10087f328: v2i8,ch = load 0x10030de58, 0x10087ef48, 0x100878408 which they don't expect.
Tags
llvmCore-2117
Branches
Apple/Bender
Bill Wendling
--- Merging r74053 into '.': A test/CodeGen/X86/inline-asm-tied.ll U lib/CodeGen/MachineInstr.cpp Fix support for inline asm input / output operand tying when operand spans across multiple registers (e.g. two i64 operands in 32-bit mode).
Tags
llvmCore-2115
Branches
Apple/Bender
Bill Wendling
Second half or r73891.
Tags
llvmCore-2114
Branches
Apple/Bender
Bill Wendling
Apply patch to fix "structure assignment drops address space qualifier on the llvm side".
Branches
Apple/Bender
Bill Wendling
--- Merging r72816 into '.': U tools/lto/LTOCodeGenerator.cpp
Branches
Apple/Bender
Bill Wendling
--- Merging r73565 into '.': A test/FrontendC++/2009-06-16-DebugInfoCrash.cpp Test for llvm-gcc patch 73564.
Branches
Apple/Bender
Bill Wendling
There doesn't seem to be a reason to move the save FP stuff.
Tags
llvmCore-2113
Branches
Apple/Bender
Bill Wendling
Merge in r73465: The DWARF to compact encoding converter assumes that the DW_CFA_def_cfa_offset comes after the DW_CFA_def_cfa_register, because the CFA is really ESP from the start of the function and only gets an offset when the "subl $xxx,%esp" instruction happens, not the other way around. And reapply r72898: The DWARF unwind info was incorrect. While compiling with `-fomit-frame-pointer', we would lack the …
Branches
Apple/Bender
Bill Wendling
--- Merging r73457 into '.': A test/DebugInfo/2009-06-15-abstract_origin.ll M lib/CodeGen/AsmPrinter/DwarfWriter.cpp Use MainCU if it is available.
Branches
Apple/Bender
convert-repo
update tags
Dan Gohman
Support vector casts in more places, fixing a variety of assertion failures. To support this, add some utility functions to Type to help support vector/scalar-independent code. Change ConstantInt::get and ConstantFP::get to support vector types, and add an overload to ConstantInt::get that uses a static IntegerType type, for convenience. Introduce a new getConstant method for ScalarEvolution, to simplify common use cases.
Anton Korobeynikov
Address more comments :)
Dan Gohman
Fix a typo in a diagnostic.
Anton Korobeynikov
Rename methods for the sake of consistency.
Anton Korobeynikov
Document ABIType enum. Patch by Sandeep Patel!
Devang Patel
Gracefully handle imbalanced inline function begin and end markers.
Evan Cheng
ifcvt should ignore cfg where true and false successors are the same.
Evan Cheng
Typo.
Dan Gohman
Rewrite the noredzone description, attempting to avoid confusing language.
Bill Wendling
The Ls and Qs were mixed up. Patch by Sean.
Tags
checker-0.211
Dale Johannesen
Fix the crash in this test. This is basically the same problem addressed in 31284, but the patch there only addressed the case where an invoke is the first thing in a block.
Evan Cheng
Do not form ldrd / strd if the two dests / srcs are the same. Code clean up.
Owen Anderson
Merge PartialInliner changes.
Bill Wendling
"The Intel instruction tables should include the 64-bit and 32-bit instructions that push immediate operands of 1, 2, and 4 bytes (extended to the native register size in each case). The assembly mnemonics are "pushl" and "pushq." One such instruction appears at the beginning of the "start" function , so this is essential for accurate disassembly when unwinding." Patch by Sean Callanan!
Evan Cheng
Silence a warning.
Bill Wendling
This test is failing. Revert for now.
  1. Prev
  2. Next