- Jun 18, 2013
-
-
Nadav Rotem authored
llvm-svn: 184228
-
Kevin Enderby authored
"When assembling to the ARM instruction set, the .N qualifier produces an assembler error and the .W qualifier has no effect." In the pre-matcher handler in the asm parser the ".w" (wide) qualifier when in ARM mode is now discarded. And an error message is now produced when the ".n" (narrow) qualifier is used in ARM mode. Test cases for these were added. rdar://14064574 llvm-svn: 184224
-
Quentin Colombet authored
value is zero. This allows optmizations to kick in more easily. Fix some test cases so that they remain meaningful (i.e., not completely dead coded) when optimizations apply. <rdar://problem/14096009> superfluous multiply by high part of zero-extended value. llvm-svn: 184222
-
Jack Carter authored
When producing objects that are abi compliant we are marking neither the object file nor the assembly file correctly and thus generate warnings. We need to set the EF_CPIC flag in the ELF header when generating direct object. Note that the warning is only generated when compiling without PIC. When compiling with clang the warning will be suppressed by supplying: -Wa,-mno-shared -Wa,-call_nonpic Also the following directive should also be added: .option pic0 when compiling without PIC, This eliminates the need for supplying: -mno-shared -call_nonpic on the assembler command line. Patch by Douglas Gilmore llvm-svn: 184220
-
Rafael Espindola authored
llvm-svn: 184219
-
Rafael Espindola authored
llvm-svn: 184217
-
David Blaikie authored
llvm-svn: 184213
-
Stefanus Du Toit authored
For decoding, keep the current behavior of always decoding these as their REP versions. In the future, this could be improved to recognize the cases where these behave as XACQUIRE and XRELEASE and decode them as such. llvm-svn: 184207
-
Rafael Espindola authored
llvm-svn: 184206
-
Nadav Rotem authored
llvm-svn: 184201
-
Nadav Rotem authored
llvm-svn: 184200
-
Alexey Samsonov authored
llvm-svn: 184191
-
Rafael Espindola authored
This allows the compiler to see the enum and warn about it. While in here, fix a switch to not use a default and fix style violations. llvm-svn: 184186
-
Amaury de la Vieuville authored
When using a positive offset, literal loads where encoded as if it was negative, because: - The sign bit was not assigned to an operand - The addrmode_imm12 operand was not encoding the sign bit correctly This patch also makes the assembler look at the .w/.n specifier for loads. llvm-svn: 184182
-
Amaury de la Vieuville authored
llvm-svn: 184181
-
Amaury de la Vieuville authored
This fixes two previous issues: - Negative offsets were not correctly disassembled - The decoded opcodes were not the right one llvm-svn: 184180
-
Amaury de la Vieuville authored
llvm-svn: 184179
-
Timur Iskhodzhanov authored
llvm-svn: 184178
-
Bill Wendling authored
Someone may want to do something crazy, like replace these objects if they change or something. No functionality change intended. llvm-svn: 184175
-
Nick Lewycky authored
llvm-svn: 184174
-
Bill Wendling authored
llvm-svn: 184173
-
Bill Wendling authored
llvm-svn: 184172
-
- Jun 17, 2013
-
-
Andrew Trick authored
llvm-svn: 184135
-
Andrew Trick authored
llvm-svn: 184133
-
Andrew Trick authored
A complex, expensive heuristic with little value in the current design. llvm-svn: 184132
-
Andrew Trick authored
llvm-svn: 184131
-
Andrew Trick authored
llvm-svn: 184130
-
Andrew Trick authored
This eliminates the MultiPressure scheduling "reason". It was sensitive to queue order. We don't like being sensitive to queue order. llvm-svn: 184129
-
Bill Wendling authored
llvm-svn: 184121
-
Rafael Espindola authored
llvm-svn: 184120
-
Rafael Espindola authored
This ports a missing feature from PathV1.h. I am not sure how to test this with the regular infrastructure, but an Apple bot should check this when r183985 is reapplied. llvm-svn: 184119
-
Vincent Lejeune authored
llvm-svn: 184117
-
Vincent Lejeune authored
Fixes rv7x0 bug in Heaven reported here: https://bugs.freedesktop.org/show_bug.cgi?id=64257 llvm-svn: 184116
-
Rafael Espindola authored
This reverts commit r183985. We were missing the checks for not deleting things like /dev/null. llvm-svn: 184111
-
Benjamin Kramer authored
The main advantages here are way better heuristics, taking into account not just loop depth but also __builtin_expect and other static heuristics and will eventually learn how to use profile info. Most of the work in this patch is pushing the MachineBlockFrequencyInfo analysis into the right places. This is good for a 5% speedup on zlib's deflate (x86_64), there were some very unfortunate spilling decisions in its hottest loop in longest_match(). Other benchmarks I tried were mostly neutral. This changes register allocation in subtle ways, update the tests for it. 2012-02-20-MachineCPBug.ll was deleted as it's very fragile and the instruction it looked for was gone already (but the FileCheck pattern picked up unrelated stuff). llvm-svn: 184105
-
Pekka Jaaskelainen authored
vectorizing loops with memory accesses to non-zero address spaces. It simply dropped the AS info. Fixes PR16306. llvm-svn: 184103
-
Rafael Espindola authored
This patch also adds a simpler version of sys::fs::remove and a tool_output_file constructor for when we already have an open file. llvm-svn: 184095
-
Nadav Rotem authored
llvm-svn: 184089
-
Nadav Rotem authored
llvm-svn: 184084
-
Rafael Espindola authored
llvm-ar is the only tool that needs to write archive files. Every other tool should be able to use the lib/Object interface. llvm-svn: 184083
-