Commits · 2bcec1297efb3cb763a4f6ae8217118e72e61e30 · Lorenzo Albano / LLVM bpEVL · GitLab

Jan 24, 2011
- fix PR9017, a bug where we'd assert when promoting in unreachable · b4017769
  Chris Lattner authored Jan 24, 2011
  b4017769
- fix PR9015, a crash linking recursive metadata. · 23289c38
  Chris Lattner authored Jan 24, 2011
  23289c38
- this isn't a memset, we do convert dest[i] to one though :) · 96856032
  Chris Lattner authored Jan 24, 2011
  96856032
- with recent work, we now optimize this into: · b830ee52
  Chris Lattner authored Jan 24, 2011
  b830ee52
- enhance SRoA to promote allocas that are used by PHI nodes. This often · d83e7b0f
  Chris Lattner authored Jan 24, 2011
  d83e7b0f
Jan 23, 2011
- Enhance SRoA to promote allocas that are used by selects in some · a960725d
  Chris Lattner authored Jan 23, 2011
  a960725d
- teach Value::isDereferenceablePointer that byval arguments are always · 9879965f
  Chris Lattner authored Jan 23, 2011
  9879965f
- Add a memset loop that LoopIdiomRecognize doesn't recognize. · 773bc67e
  Anders Carlsson authored Jan 23, 2011
  773bc67e
- Simplify some code with no functionality change. Make the test a lot more · d4192f71
  Nick Lewycky authored Jan 23, 2011
  d4192f71
- Initialize MCNoExecStack. · c5efca47
  Rafael Espindola authored Jan 23, 2011
  c5efca47
- Add support for the --noexecstack option. · b3eca9bb
  Rafael Espindola authored Jan 23, 2011
  b3eca9bb
- Null initialize a few variables flagged by · 3c4408ce
  Ted Kremenek authored Jan 23, 2011
  3c4408ce
- Add support for lowercase variants. · 8bac423d
  Rafael Espindola authored Jan 23, 2011
  8bac423d
- Enhance SRoA to be more aggressive about scalarization of aggregate allocas · 9491dee2
  Chris Lattner authored Jan 23, 2011
  9491dee2
- Convert two std::vectors to SmallVectors for a 3.4% speedup running -scalarrepl · 07d6fe34
  Cameron Zwarich authored Jan 23, 2011
  07d6fe34
- have AllocaInfo store the alloca being inspected, simplifying callers. · 8acbb795
  Chris Lattner authored Jan 23, 2011
  8acbb795
- Rearrange some code a bit. Change MarkUnsafe to · 3e56c290
  Chris Lattner authored Jan 23, 2011
  3e56c290
- remove an old hack that avoided creating MMX datatypes. The · a587ab7b
  Chris Lattner authored Jan 23, 2011
  a587ab7b
- Use value ranges to fold ext(trunc) in SCEV when possible. · bc98f5b7
  Nick Lewycky authored Jan 23, 2011
  bc98f5b7
- Delay the creation of eh_frame so that the user can change the defaults. · 4b7b7fba
  Rafael Espindola authored Jan 23, 2011
  4b7b7fba
- Remove more duplicated code. · 0e7e34e4
  Rafael Espindola authored Jan 23, 2011
  0e7e34e4
- Remove duplicated code. · aea4958e
  Rafael Espindola authored Jan 23, 2011
  aea4958e
Jan 22, 2011
- Have SCEV turn sext(x) into zext(x) when x is s>= 0. This applies many times in · b32c8943
  Nick Lewycky authored Jan 22, 2011
  b32c8943
- Add a FIXME explaining the move to a single indirect call bonus per function · c70e037b
  Eric Christopher authored Jan 22, 2011
  c70e037b
- Only apply the devirtualization bonus once instead of per-call site in the · 08e8b3b6
  Eric Christopher authored Jan 22, 2011
  08e8b3b6
- Pass sret arguments through the stack instead of through registers in Sparc... · cc91b7a3
  Venkatraman Govindaraju authored Jan 22, 2011
  cc91b7a3
- Added ICC, FCC as uses of movcc instruction to generate correct code when -mattr=v9 is used. · 7a0c3500
  Venkatraman Govindaraju authored Jan 22, 2011
  7a0c3500
Jan 21, 2011
- Actually check memcpy lengths, instead of just commenting about · 19e30d5a
  Dan Gohman authored Jan 21, 2011
  19e30d5a
- Sparc backend: · ef8cf45e
  Venkatraman Govindaraju authored Jan 21, 2011
  ef8cf45e
- Just because we have determined that an (fcmp | fcmp) is true for A < B, · a834200d
  Owen Anderson authored Jan 21, 2011
  a834200d
- Last round of fixes for movw + movt global address codegen. · 2f2435d0
  Evan Cheng authored Jan 21, 2011
  2f2435d0
- Clang was not parsing target triples involving EABI and was generating wrong... · 83758d5c
  Renato Golin authored Jan 21, 2011
  83758d5c
- Handles libffi on the CMake build. · 64955953
  Oscar Fuentes authored Jan 21, 2011
  64955953
- Fix the encoding of QADD/SUB, QDADD/SUB. While qadd16, qadd8 use "rd, rn, rm", · 4bd61238
  Bruno Cardoso Lopes authored Jan 21, 2011
  4bd61238
- Implement support for byval arguments in Sparc backend. · 0594789f
  Venkatraman Govindaraju authored Jan 21, 2011
  0594789f
- SCCP doesn't actually preserve the CFG. It will delete and insert terminator · ae0275e0
  Nick Lewycky authored Jan 21, 2011
  ae0275e0
- Enable support for precise scheduling of the instruction selection · bd428ec5
  Andrew Trick authored Jan 21, 2011
  bd428ec5
- Convert -enable-sched-cycles and -enable-sched-hazard to -disable · 47ff14b0
  Andrew Trick authored Jan 21, 2011
  47ff14b0
- fix PR9013, an infinite loop in instcombine. · b5e15d19
  Chris Lattner authored Jan 21, 2011
  b5e15d19
- update obsolete comment. · f4ca47bd
  Chris Lattner authored Jan 21, 2011
  f4ca47bd