Commit Graph

76928 Commits

Author SHA1 Message Date
Craig Topper
9bb36ed8d8 [X86] Add assembly parser support for mnemonic aliases for AVX-512 vpcmp instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229287 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-15 07:13:48 +00:00
Chandler Carruth
5ee516549f [x86] Fix PR22377, a regression with the new vector shuffle legality
test.

This was just a matter of the DAG combine for vector shuffles being too
aggressive. This is a bit of a grey area, but I think generally if we
can re-use intermediate shuffles, we should. Certainly, given the test
cases I have available, this seems like the right call.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229285 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-15 07:01:10 +00:00
Craig Topper
09ea4e976b [X86] Add assembler predicates for the rest of the AVX512 feature flags. This makes the assembly matching consistent across all AVX512 instructions. Without this we were allowing some AVX512 instructions to be parsed always, but not the foundation instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229280 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-15 04:54:55 +00:00
Craig Topper
e2f7231e45 [X86] Add the remaining 11 possible exact ModRM formats. This makes their encodings linear which can then be used to simplify some other code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229279 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-15 04:16:44 +00:00
Simon Pilgrim
6d5ee8a8b5 [X86][XOP] Enable commutation for XOP instructions
Patch to allow XOP instructions (integer comparison and integer multiply-add) to be commuted. The comparison instructions sometimes require the compare mode to be flipped but the remaining instructions can use default commutation modes.

This patch also sets the SSE domains of all the XOP instructions.

Differential Revision: http://reviews.llvm.org/D7646

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229267 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 22:40:46 +00:00
Craig Topper
32f60795f5 [X86] Improve parsing support AVX/SSE floating point compare instruction mnemonic aliases. They'll now print with the alias the parser received instead of converting to the explicit immediate form.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229266 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 21:54:03 +00:00
Ramkumar Ramachandra
0608cec657 InstCombine: propagate deref via new addDereferenceableAttr
The "dereferenceable" attribute cannot be added via .addAttribute(),
since it also expects a size in bytes. AttrBuilder#addAttribute or
AttributeSet#addAttribute is wrapped by classes Function, InvokeInst,
and CallInst. Add corresponding wrappers to
AttrBuilder#addDereferenceableAttr.

Having done this, propagate the dereferenceable attribute via
gc.relocate, adding a test to exercise it. Note that -datalayout is
required during execution over and above -instcombine, because
InstCombine only optionally requires DataLayoutPass.

Differential Revision: http://reviews.llvm.org/D7510

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229265 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 19:37:54 +00:00
Duncan P. N. Exon Smith
c7fe1ab250 Target: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229261 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 15:36:52 +00:00
Duncan P. N. Exon Smith
8cfbff509e NVPTX: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229260 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 15:35:43 +00:00
Andrea Di Biagio
47cd120a18 [optnone] Skip pass Constant Hoisting on optnone functions.
Added test CodeGen/X86/constant-hoisting-optnone.ll to verify that
pass Constant Hoisting is not run on optnone functions.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229258 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 15:11:48 +00:00
Simon Pilgrim
ee03ed8187 Line ending fix. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229256 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 13:27:53 +00:00
Chandler Carruth
96db150cec Remove a variable only used in an assert and sink its initializer into
the assert. Fixes -Wunused-variable on non-asserts builds.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229250 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 09:14:44 +00:00
Matt Arsenault
8a44761afe R600/SI: Implement correct f64 fdiv
This version passes the OpenCL conformance test.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229239 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 04:30:08 +00:00
Matt Arsenault
fd31a769ce R600/SI: Use complex operand folding for div_scale
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229238 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 04:24:28 +00:00
Matt Arsenault
9295d69bea R600/SI: Fix implicit vcc operand to v_div_fmas_*
This should allow finally fixing the f64 fdiv implementation.

Test is disabled for VI since there seems to be a problem with one
of the buffer load instructions on it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229236 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 04:22:00 +00:00
Matt Arsenault
f1b6c1f287 R600/SI: Fix schedule model for v_div_scale_{f32|f64}
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229235 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 04:03:18 +00:00
Matt Arsenault
958cf585ce R600/SI: Really fix size of VReg_1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229234 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 03:54:32 +00:00
Matt Arsenault
b79a67d967 R600/SI: Rename encoding field to match docs for VOP3b
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229233 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 03:54:29 +00:00
Zachary Turner
73a1e454d7 llvm-pdbdump: Only dump whitelisted global symbols.
Dumping the global scope contains a lot of very uninteresting
things and is generally polluted with a lot of random junk.
Furthermore, it dumps values unsorted, making it hard to read.
This patch dumps known interesting types only, and as a side
effect sorts the list by symbol type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229232 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 03:54:28 +00:00
Zachary Turner
f94896ffa0 llvm-pdbdump: Re-order header files according to LLVM style guide.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229231 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 03:53:56 +00:00
Matt Arsenault
abcfb2b907 R600/SI: Fix not encoding src2 for v_div_scale_{f32|f64}
This apparently got lost in the VI changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229230 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 03:40:35 +00:00
Matt Arsenault
124991cc67 R600/SI: Fix VOP3b encoding on VI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229228 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 03:02:23 +00:00
Matt Arsenault
e7ff4c1321 R600/SI: Fix phys reg copies in SIFoldOperands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229227 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:55:57 +00:00
Matt Arsenault
9b6d620f89 R600/SI: Fix copies from SGPR to VCC
This shows up without optimizations when vcc is required
to be used.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229226 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:55:56 +00:00
Matt Arsenault
582d603f17 R600/SI: Add hack to copy from a VGPR to VCC
This hopefully should be fixed when VReg_1 is removed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229225 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:55:54 +00:00
Duncan P. N. Exon Smith
4eba62d6c2 PowerPC: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229224 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:54:07 +00:00
Matt Arsenault
2d925cbe10 R600/SI: Fix size of VReg_1
This is really a 32-bit register, if we try to check the size of it,
we want 32-bits.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229223 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:51:44 +00:00
Duncan P. N. Exon Smith
20d6ad9de2 R600: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229222 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:45:45 +00:00
Duncan P. N. Exon Smith
39127e77ae Mips: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229221 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:37:48 +00:00
Duncan P. N. Exon Smith
f38384bcf1 ARM: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229220 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:24:44 +00:00
Duncan P. N. Exon Smith
6a1927d2c1 AArch64: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229218 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:09:06 +00:00
Duncan P. N. Exon Smith
894c8c514a X86: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229214 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 01:59:52 +00:00
Duncan P. N. Exon Smith
9de77c7eca CodeGen: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

Also, add `Function::getFnStackAlignment()`, and canonicalize:

getAttributes().getStackAlignment(AttributeSet::FunctionIndex)
  => getFnStackAlignment()

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229208 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 01:44:41 +00:00
Ahmed Bougacha
6a50342499 [X86] Factor out the CMOV pseudo definitions. NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229206 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 01:36:53 +00:00
Matthias Braun
821ec14add Revert "On ELF, put PIC jump tables in a non executable section."
This reverts commit r228939.

The commit broke something in the output of exception handling tables on
darwin x86-64.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229203 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 01:16:54 +00:00
Duncan P. N. Exon Smith
7520a90c75 Transforms: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229202 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 01:11:29 +00:00
Reid Kleckner
f7f5309424 Unify the two EH personality classification routines I wrote
We only need one.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229193 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 00:21:02 +00:00
Duncan P. N. Exon Smith
efd49ea0d2 Analysis: Canonicalize access to function attributes, NFC
Canonicalize access to function attributes to use the simpler API.

getAttributes().getAttribute(AttributeSet::FunctionIndex, Kind)
  => getFnAttribute(Kind)

getAttributes().hasAttribute(AttributeSet::FunctionIndex, Kind)
  => hasFnAttribute(Kind)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229192 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 00:12:15 +00:00
Eric Christopher
c21d153388 Use the template method to grab the target specific subtarget.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229191 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 00:09:46 +00:00
Philip Reames
d777c2c0c0 [InstCombine] When canonicalizing gep indices, prefer zext when possible
If we know that the sign bit of a value being sign extended is zero, we can use a zero extension instead.  This is motivated by the fact that zero extensions are generally cheaper on x86 (and most other architectures?).  We already apply a similar transform in DAGCombine, this just extends that to the IR level.

This comes up when we eagerly canonicalize gep indices to the width of a machine register (i64 on x86_64). To do so, we insert sign extensions (sext) to promote smaller types. 

Differential Revision: http://reviews.llvm.org/D7255



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229189 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 00:05:36 +00:00
Chris Bieneman
0213b090b8 Fixing broken bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229176 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 23:10:31 +00:00
Philip Reames
8c6097d490 Minor tweak to MDA
Two minor tweaks I noticed when reading through the code:
- No need to recompute begin() on every iteration.  We're not modifying the instructions in this loop.
- We can ignore PHINodes and Dbg intrinsics.  The current code does this anyways, but it will spend slightly more time doing so and will count towards the limit of instructions in the block.  It seems really silly to give up due the presence of PHIs...

Differential Revision: http://reviews.llvm.org/D7624



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229175 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 23:08:37 +00:00
Chris Bieneman
845c00e2e6 NFC. Moving the RegisteredOptionCategories global into the CommandLineParser class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229172 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:54:32 +00:00
Chris Bieneman
a87100c6d2 NFC. clang-format wants to change this from two lines to one.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229171 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:54:29 +00:00
Chris Bieneman
736c14962e NFC. More code cleanup making LookupOption a member of the CommandLineParser.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229170 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:54:27 +00:00
Eric Christopher
40ccb77781 The base pointer save offset can be computed at initialization time,
do so and fix up the calls.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229169 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:48:53 +00:00
Eric Christopher
f7f32530e0 Move the target machine variable so that it's initialized early
enough we can use it to initialize frame lowering.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229168 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:48:51 +00:00
Eric Christopher
7b93acde32 Stash the TargetMachine on the subtarget so we can access it later.
Clean up a subtarget function that has it passed in while we're at it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229164 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:23:04 +00:00
Eric Christopher
4015892a84 PPC LinkageSize can be computed at initialization time, do so.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229163 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:22:57 +00:00
Reid Kleckner
86b5c3a844 Triple: Make setEnvironment not override the object format
Discovered by Halide users who had C++ code like this:
  Triple.setArch(Triple::x86);
  Triple.setOS(Triple::Windows);
  Triple.setObjectFormat(Triple::ELF);
  Triple.setEnvironment(Triple::MSVC);

This would produce the stringified triple of x86-windows-msvc, instead
of the x86-windows-msvc-elf string needed to run MCJIT.

With this change, they retain the -elf suffix.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229160 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 22:05:50 +00:00