Commit Graph

16742 Commits

Author SHA1 Message Date
Craig Topper 2dca3b287b [X86] Make the FMA3 instruction names consistent between VEX and EVEX encoded versions.
This places the 132/213/231 form number in front of the SS/SD/PS/PD. Move the Y for 256-bit versions to be after the PS/PD. Change the AVX512 scalar forms to include a Z in the their name. This new format should be consistent with the general naming of instructions.

llvm-svn: 276559
2016-07-24 08:26:38 +00:00
Simon Pilgrim 9e99dd9e8b [X86][SSE] Added float widened broadcast tests
llvm-svn: 276535
2016-07-23 21:24:02 +00:00
Simon Pilgrim 8d18969716 [X86][SSE] Added more widened broadcast tests
Added more vXi16 and vXi8 tests

llvm-svn: 276534
2016-07-23 21:15:31 +00:00
Simon Pilgrim b9e47a8cd0 [X86][SSE] Added tests where we should be trying to widen a load+splat into a broadcast
llvm-svn: 276527
2016-07-23 16:19:17 +00:00
Simon Pilgrim 8aa6f34455 [X86][SSE] Regenerated uitofp <2 x i32> -> <2 x float> conversion tests
Demonstrate difference in codegen discussed on PR14760

llvm-svn: 276526
2016-07-23 15:55:42 +00:00
Craig Topper b6519db90d [AVX512] Implement commuting support for EVEX encoded FMA3 instructions.
llvm-svn: 276521
2016-07-23 07:16:56 +00:00
Tom Stellard b8253c88b6 Revert "[AMDGPU] Emit read-only data to .rodata for hsa"
This reverts commit r276298.

Data stored in .rodata can have a negative offset from .text, but we
don't support negative values in relocations yet.

This caused a regression in one of the amp conformance tests:
5_Data_Cont/5_2_a_v/5_2_3_m/Assignment/Test.02.01

llvm-svn: 276498
2016-07-22 23:46:40 +00:00
Tim Northover 98a56eb7f4 GlobalISel: allow multiple types on MachineInstrs.
llvm-svn: 276481
2016-07-22 22:13:36 +00:00
Tim Northover 33b07d6725 GlobalISel: implement legalization pass, with just one transformation.
This adds the actual MachineLegalizeHelper to do the work and a trivial pass
wrapper that legalizes all instructions in a MachineFunction. Currently the
only transformation supported is splitting up a vector G_ADD into one acting on
smaller vectors.

llvm-svn: 276461
2016-07-22 20:03:43 +00:00
Anna Thomas 0be4a0e6a4 Invariant start/end intrinsics overloaded for address space
Summary:
The llvm.invariant.start and llvm.invariant.end intrinsics currently
support specifying invariant memory objects only in the default address
space.

With this change, these intrinsics are overloaded for any adddress space
for memory objects
and we can use these llvm invariant intrinsics in non-default address
spaces.

Example: llvm.invariant.start.p1i8(i64 4, i8 addrspace(1)* %ptr)

This overloaded intrinsic is needed for representing final or invariant
memory in managed languages.

Reviewers: apilipenko, reames

Subscribers: llvm-commits
llvm-svn: 276447
2016-07-22 17:49:40 +00:00
Matt Arsenault e2fe67b951 AMDGPU: Remove redundant test
llvm-svn: 276439
2016-07-22 17:01:36 +00:00
Matt Arsenault 3c07c813c0 AMDGPU: Fix groupstaticsize for large LDS
The size can exceed s_movk_i32's limit, and we don't
want to use it this early since it inhibits optimizations.

This should probably be merged to the release branch.

llvm-svn: 276438
2016-07-22 17:01:33 +00:00
Matt Arsenault 8d718dcfda AMDGPU: Add HSA dispatch id intrinsic
llvm-svn: 276437
2016-07-22 17:01:30 +00:00
Matt Arsenault 7fb961f3e6 AMDGPU: Fix i1 fp_to_int
R600's i1 fp_to_uint selected but was incorrect according to
what instcombine constant folds to.

llvm-svn: 276435
2016-07-22 17:01:21 +00:00
Tim Northover bd5054602e GlobalISel: implement alloca instruction
llvm-svn: 276433
2016-07-22 16:59:52 +00:00
Simon Pilgrim 820f87a72d [SelectionDAG] Optimization of BITREVERSE legalization for power-of-2 integer scalar/vector types
An extension of D19978, this patch replaces the default BITREVERSE evaluation of individual bit masks+shifts with block mask+shifts when we have integer elements of power-of-2 bits in size.

After calling BSWAP to reverse the order of the constituent bytes (which typically follows a similar approach), every neighbouring 4-bits, 2-bits and finally 1-bit pairs are masked off and swapped over with shifts.

In doing so we can significantly reduce the number of operations required.

Differential Revision: https://reviews.llvm.org/D21578

llvm-svn: 276432
2016-07-22 16:46:25 +00:00
Krzysztof Parzyszek d3d0a4bda3 [Hexagon] Use loop data prefetch on Hexagon
llvm-svn: 276422
2016-07-22 14:22:43 +00:00
Simon Pilgrim ea0d4f9962 [X86][AVX] Added support for lowering to VBROADCASTF128/VBROADCASTI128 (reapplied)
As reported on PR26235, we don't currently make use of the VBROADCASTF128/VBROADCASTI128 instructions (or the AVX512 equivalents) to load+splat a 128-bit vector to both lanes of a 256-bit vector.

This patch enables lowering from subvector insertion/concatenation patterns and auto-upgrades the llvm.x86.avx.vbroadcastf128.pd.256 / llvm.x86.avx.vbroadcastf128.ps.256 intrinsics to match.

We could possibly investigate using VBROADCASTF128/VBROADCASTI128 to load repeated constants as well (similar to how we already do for scalar broadcasts).

Reapplied with fix for PR28657 - removed intrinsic definitions (clang companion patch to be be submitted shortly).

Differential Revision: https://reviews.llvm.org/D22460

llvm-svn: 276416
2016-07-22 13:58:44 +00:00
Ahmed Bougacha 29333c9de6 [FastISel] Ignore @llvm.assume.
llvm-svn: 276410
2016-07-22 12:54:53 +00:00
Benjamin Kramer 5ba0e20315 Revert "[X86][AVX] Added support for lowering to VBROADCASTF128/VBROADCASTI128"
It caused PR28657.

This reverts commit r276281.

llvm-svn: 276405
2016-07-22 11:03:10 +00:00
Hrvoje Varga 2db00ce4b6 [mips][microMIPS] Implement SLT, SLTI, SLTIU, SLTU microMIPS32r6 instructions
Differential Revision: https://reviews.llvm.org/D19906

llvm-svn: 276397
2016-07-22 07:18:33 +00:00
Craig Topper 52e2e8381b [AVX512] Add ExeDomain to vector extend and truncate instructions.
llvm-svn: 276394
2016-07-22 05:46:44 +00:00
Craig Topper f4151bea72 [AVX512] Add initial support for the Execution Domain fixing pass to change some EVEX instructions.
llvm-svn: 276393
2016-07-22 05:00:52 +00:00
Craig Topper 0b90756b0a [AVX512] Add load folding for some AVX512VL logic and arithmetic instructions.
llvm-svn: 276391
2016-07-22 05:00:39 +00:00
Craig Topper ab13b33ded [AVX512] Update X86InstrInfo::foldMemoryOperandCustom to handle the EVEX encoded instructions too.
llvm-svn: 276390
2016-07-22 05:00:35 +00:00
Quentin Colombet ecd81a3d1b [MIRTesting] Abort when failing to parse a function.
When we failed to parse a function in the mir parser, we should abort
the whole compilation instead of continuing in a weird state. Indeed,
this was creating strange machine function passes failures that were
hard to understand, until we notice that the function actually did not
get parsed correctly!

llvm-svn: 276348
2016-07-21 22:25:57 +00:00
Michael Kuperstein c523333bbf [X86] Do not use AND8ri8 in AVX512 pattern
This variant is (as documented in the TD) for disassembler use only, and should
not be used in patterns - it is longer, and is broken on 64-bit.

llvm-svn: 276347
2016-07-21 22:24:08 +00:00
Akira Hatanaka b8d2873d93 [AArch64][Inline-Asm] Return the 32-bit floating point register class
when constraint "w" is used on a 32-bit operand.

This enables compiling the following code, which used to error out in
the backend:

void foo1(int a) {
  asm volatile ("sqxtn h0, %s0\n" : : "w"(a):);
}

Fixes PR28633.

llvm-svn: 276344
2016-07-21 21:39:05 +00:00
Anna Thomas c858faa244 Revert "Invariant start/end intrinsics overloaded for address space"
This reverts commit r276316.

llvm-svn: 276320
2016-07-21 19:06:28 +00:00
Anna Thomas 29b24dfe44 Invariant start/end intrinsics overloaded for address space
Summary:
The llvm.invariant.start and llvm.invariant.end intrinsics currently
support specifying invariant memory objects only in the default address space.

With this change, these intrinsics are overloaded for any adddress space for memory objects
and we can use these llvm invariant intrinsics in non-default address spaces.

Example: llvm.invariant.start.p1i8(i64 4, i8 addrspace(1)* %ptr)

This overloaded intrinsic is needed for representing final or invariant memory in managed languages.

Reviewers: tstellarAMD, reames, apilipenko

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D22519

llvm-svn: 276316
2016-07-21 18:41:44 +00:00
Quentin Colombet 2b59eab79f [IRTranslator] Add G_SUB opcode.
This commit adds a generic SUB opcode to global-isel.

llvm-svn: 276308
2016-07-21 17:26:50 +00:00
Konstantin Zhuravlyov 3c0d8d22fe [AMDGPU] Emit read-only data to .rodata for hsa
Differential Revision: https://reviews.llvm.org/D22538

llvm-svn: 276298
2016-07-21 15:59:23 +00:00
Quentin Colombet 7bcc921dd8 [IRTranslator] Add G_AND opcode.
This commit adds a generic AND opcode to global-isel.

llvm-svn: 276297
2016-07-21 15:50:42 +00:00
Geoff Berry 4ff2e36d32 [AArch64] Load/store opt: Don't count transient instructions towards search limits.
Summary:
This change also changes findMatchingInsn and
findMatchingUpdateInsnForward to take DBG_VALUE opcodes into account
when tracking register defs and uses, which could potentially inhibit
these optimizations in the presence of debug information.

Reviewers: mcrosier

Subscribers: aemerson, rengolin, mcrosier, llvm-commits

Differential Revision: https://reviews.llvm.org/D22582

llvm-svn: 276293
2016-07-21 15:20:25 +00:00
Simon Pilgrim 88e0940d3b [X86][SSE] Allow folding of store/zext with PEXTRW of 0'th element
Under normal circumstances we prefer the higher performance MOVD to extract the 0'th element of a v8i16 vector instead of PEXTRW.

But as detailed on PR27265, this prevents the SSE41 implementation of PEXTRW from folding the store of the 0'th element. Additionally it prevents us from making use of the fact that the (SSE2) reg-reg version of PEXTRW implicitly zero-extends the i16 element to the i32/i64 destination register.

This patch only preferentially lowers to MOVD if we will not be zero-extending the extracted i16, nor prevent a store from being folded (on SSSE41).

Fix for PR27265.

Differential Revision: https://reviews.llvm.org/D22509

llvm-svn: 276289
2016-07-21 14:54:17 +00:00
Simon Pilgrim 4caefdf834 Fixed line endings
llvm-svn: 276287
2016-07-21 14:36:41 +00:00
Simon Pilgrim c8e20b1150 [X86][AVX] Added support for lowering to VBROADCASTF128/VBROADCASTI128
As reported on PR26235, we don't currently make use of the VBROADCASTF128/VBROADCASTI128 instructions (or the AVX512 equivalents) to load+splat a 128-bit vector to both lanes of a 256-bit vector.

This patch enables lowering from subvector insertion/concatenation patterns and auto-upgrades the llvm.x86.avx.vbroadcastf128.pd.256 / llvm.x86.avx.vbroadcastf128.ps.256 intrinsics to match.

We could possibly investigate using VBROADCASTF128/VBROADCASTI128 to load repeated constants as well (similar to how we already do for scalar broadcasts).

Differential Revision: https://reviews.llvm.org/D22460

llvm-svn: 276281
2016-07-21 14:10:54 +00:00
Marina Yatsina c1fa163392 ExecutionDepsFix - Fix bug in clearance calculation
The clearance calculation did not take into account registers defined as outputs or clobbers in inline assembly machine instructions because these register defs are implicit.

Differential Revision: http://reviews.llvm.org/D22580

llvm-svn: 276266
2016-07-21 12:37:07 +00:00
Matt Arsenault f0ba86a4d5 AMDGPU: Fix phis from blocks split due to register indexing
llvm-svn: 276257
2016-07-21 09:40:57 +00:00
Matthias Braun d9fdad72ae IPRA: Fix RegMask calculation for alias registers
This patch fixes a very subtle bug in regmask calculation. Thanks to zan
jyu Wong <zyfwong@gmail.com> for bringing this to notice.
For example if CL is only clobbered than CH should not be marked
clobbered but CX, RCX and ECX should be mark clobbered. Previously for
each modified register all of its aliases are marked clobbered by
markRegClobbred() in RegUsageInfoCollector.cpp but that is wrong because
when CL is clobbered then MRI::isPhysRegModified() will return true for
CL, CX, ECX, RCX which is correct behavior but then for CX, EXC, RCX we
mark CH also clobbered as CH is aliased to CX,ECX,RCX so
markRegClobbred() is not required because isPhysRegModified already take
cares of proper aliasing register. A very simple test case has been
added to verify this change.
Please find relevant bug report here :
http://llvm.org/PR28567

Patch by Vivek Pandya <vivekvpandya@gmail.com>

Differential Revision: https://reviews.llvm.org/D22400

llvm-svn: 276235
2016-07-21 03:50:39 +00:00
Justin Lebar cd564c6b46 [NVPTX] Enable the load-store vectorizer on nvptx.
Reviewers: tra

Subscribers: jholewinski, arsenm, asbirlea

Differential Revision: https://reviews.llvm.org/D22592

llvm-svn: 276196
2016-07-20 22:11:36 +00:00
Artem Belevich 7e9c9a6582 [NVPTX] Renamed NVPTXLowerKernelArgs -> NVPTXLowerArgs. NFC.
After r276153 the pass applies to both kernels and regular functions.

Differential Revision: https://reviews.llvm.org/D22583

llvm-svn: 276189
2016-07-20 21:44:07 +00:00
Ahmed Bougacha a0cdd79070 [AArch64][FastISel] Select -O0 legal cmpxchg.
At -O0, cmpxchg survives AtomicExpand: it's mostly straightforward
to select it in fast-isel, and let the pseudo be expanded later.

extractvalues on the result are the tricky part: the generic logic
only works for legal types (and it would be painful to make it
support illegal types), so we can only support i32/i64 cmpxchg.

llvm-svn: 276183
2016-07-20 21:12:32 +00:00
Ahmed Bougacha b0674d1143 [AArch64][FastISel] Select atomic stores into STLR.
llvm-svn: 276182
2016-07-20 21:12:27 +00:00
Tim Northover 62ae568bbb GlobalISel: implement low-level type with just size & vector lanes.
This should be all the low-level instruction selection needs to determine how
to implement an operation, with the remaining context taken from the opcode
(e.g. G_ADD vs G_FADD) or other flags not based on type (e.g. fast-math).

llvm-svn: 276158
2016-07-20 19:09:30 +00:00
Artem Belevich 74158b5061 [NVPTX] deal with all aggregate return types.
Fixes a crash in llvm_unreachable when a function has array return type.

Differential Revision: https://reviews.llvm.org/D22524

llvm-svn: 276154
2016-07-20 18:39:52 +00:00
Artem Belevich b2e76a5e7a [NVPTX] Improve lowering of byval args of device functions.
Avoid unnecessary spills of byval arguments of device functions to
local space on SASS level and subsequent pointer conversion to generic
address space that follows. Instead, make a local copy in IR, provide
a way to access arguments directly, and let LLVM optimize the copy away
when possible.

Differential Review: https://reviews.llvm.org/D21421

llvm-svn: 276153
2016-07-20 18:39:47 +00:00
Matt Arsenault f14db7a933 AMDGPU: Add missing test coverage for control flow breaks
None of the current lit tests hit si_break handling.

llvm-svn: 276129
2016-07-20 15:20:35 +00:00
Yaxun Liu 4b1d9f7f18 AMDGPU: Fix bug causing crash due to invalid opencl version metadata.
Differential Revision: https://reviews.llvm.org/D22526

llvm-svn: 276119
2016-07-20 14:38:06 +00:00
Diana Picus f345d40ae2 [ARM] Skip inline asm memory operands in DAGToDAGISel
Retry r275776 (no changes, we suspect the issue was with another commit).

The current logic for handling inline asm operands in DAGToDAGISel interprets
the operands by looking for constants, which should represent the flags
describing the kind of operand we're dealing with (immediate, memory, register
def etc). The operands representing actual data are skipped only if they are
non-const, with the exception of immediate operands which are skipped explicitly
when a flag describing an immediate is found.

The oversight is that memory operands may be const too (e.g. for device drivers
reading a fixed address), so we should explicitly skip the operand following a
flag describing a memory operand. If we don't, we risk interpreting that
constant as a flag, which is definitely not intended.

Fixes PR26038

Differential Revision: https://reviews.llvm.org/D22103

llvm-svn: 276101
2016-07-20 09:48:24 +00:00
David Majnemer 5d26127752 Revert "Disable this-return argument forwarding on ARM/AArch64"
Inference of the 'returned' attribute was fixed in r276008, lets try
turning the backend support back on.

This reverts commit r275677.

llvm-svn: 276081
2016-07-20 04:13:01 +00:00
Matthias Braun 5b9722d6c7 Revert "RegScavenging: Add scavengeRegisterBackwards()"
Reverting this commit for now as it seems to be causing failures on
test-suite tests on the clang-ppc64le-linux-lnt bot.

This reverts commit r276044.

llvm-svn: 276068
2016-07-20 00:21:32 +00:00
Matt Arsenault a1fe17c9ad AMDGPU: Change fdiv lowering based on !fpmath metadata
If 2.5 ulp is acceptable, denormals are not required, and
isn't a reciprocal which will already be handled, replace
with a faster fdiv.

Simplify the lowering tests by using per function
subtarget features.

llvm-svn: 276051
2016-07-19 23:16:53 +00:00
Matthias Braun 84fd4bee6c RegScavenging: Add scavengeRegisterBackwards()
This is a variant of scavengeRegister() that works for
enterBasicBlockEnd()/backward(). The benefit of the backward mode is
that it is not affected by incomplete kill flags.

This patch also changes
PrologEpilogInserter::doScavengeFrameVirtualRegs() to use the register
scavenger in backwards mode.

Differential Revision: http://reviews.llvm.org/D21885

llvm-svn: 276044
2016-07-19 22:37:09 +00:00
Evandro Menezes 238fa76574 [AArch64] Properly validate the reciprocal estimation.
Add check for legal data types when expanding into a Newton series.

Differential Revision: https://reviews.llvm.org/D22267

llvm-svn: 276041
2016-07-19 22:31:11 +00:00
Ahmed Bougacha 5a59b24bdd [GlobalISel] Mark newly-created gvregs as having a bank.
Also verify that we never try to set the size of a vreg associated
to a register class.

Report an error when we encounter that in MIR. Fix a testcase that
hit that error and had a size for no reason.

llvm-svn: 276012
2016-07-19 19:48:36 +00:00
Simon Pilgrim 5366d0e0bc [X86][AVX512] Added AVX512 subvector broadcast tests
llvm-svn: 275994
2016-07-19 17:04:28 +00:00
Simon Pilgrim f2d02cb0f6 [X86][AVX] Fixed typo in test names
llvm-svn: 275992
2016-07-19 16:52:05 +00:00
Simon Pilgrim 0ea8d275cc [X86][SSE] Reimplement SSE fp2si conversion intrinsics instead of using generic IR
D20859 and D20860 attempted to replace the SSE (V)CVTTPS2DQ and VCVTTPD2DQ truncating conversions with generic IR instead.

It turns out that the behaviour of these intrinsics is different enough from generic IR that this will cause problems, INF/NAN/out of range values are guaranteed to result in a 0x80000000 value - which plays havoc with constant folding which converts them to either zero or UNDEF. This is also an issue with the scalar implementations (which were already generic IR and what I was trying to match).

This patch changes both scalar and packed versions back to using x86-specific builtins.

It also deals with the other scalar conversion cases that are runtime rounding mode dependent and can have similar issues with constant folding.

A companion clang patch is at D22105

Differential Revision: https://reviews.llvm.org/D22106

llvm-svn: 275981
2016-07-19 15:07:43 +00:00
Sam Parker 6ca4bbb00d [ARM] Refactor Thumb2 Mul and Mla instr descs
Recommitting after r274347 was reverted. This patch introduces some
classes to refactor the 3 and 4 register Thumb2 multiplication
instruction descriptions, plus improved tests for some of those
instructions.

Differential Revision: https://reviews.llvm.org/D21929

llvm-svn: 275979
2016-07-19 14:44:05 +00:00
Simon Pilgrim b87a21f1c3 [AARCH64] Fix linu triple typo
As promised in D22191

llvm-svn: 275976
2016-07-19 14:12:45 +00:00
Simon Pilgrim fc4d4b251d [AARCH64] Enable AARCH64 lit tests on windows dev machines
As discussed on PR27654, this patch fixes the triples of a lot of aarch64 tests and enables lit tests on windows

This will hopefully help stop cases where windows developers break the aarch64 target

Differential Revision: https://reviews.llvm.org/D22191

llvm-svn: 275973
2016-07-19 13:35:11 +00:00
Daniel Sanders 6a73883c48 [mips] Correct label prefixes for N32 and N64.
Summary:
N32 and N64 follow the standard ELF conventions (.L) whereas O32 uses its own
($).

This fixes the majority of object differences between -fintegrated-as and
-fno-integrated-as.

Reviewers: sdardis

Subscribers: dsanders, sdardis, llvm-commits

Differential Revision: https://reviews.llvm.org/D22412

llvm-svn: 275967
2016-07-19 10:49:03 +00:00
Elena Demikhovsky 2c0780b8e5 AVX-512: Fixed BT instruction selection.
The following condition expression ( a >> n) & 1 is converted to "bt a, n" instruction. It works on all intel targets.
But on AVX-512 it was broken because the expression is modified to (truncate (a >>n) to i1).

I added the new sequence (truncate (a >>n) to i1) to the BT pattern.

Differential Revision: https://reviews.llvm.org/D22354

llvm-svn: 275950
2016-07-19 07:14:21 +00:00
Craig Topper d6ca1dc45e [AVX512] Give priority to EVEX encoded PSHUFB over the VEX versions.
llvm-svn: 275942
2016-07-19 02:00:38 +00:00
Matt Arsenault cb540bc03c AMDGPU: Expand register indexing pseudos in custom inserter
This is to help moveSILowerControlFlow to before regalloc.
There are a couple of tradeoffs with this. The complete CFG
is visible to more passes, the loop body avoids an extra copy of m0,
vcc isn't required, and immediate offsets can be shrunk into s_movk_i32.

The disadvantage is the register allocator doesn't understand that
the single lane's vector is dead within the loop body, so an extra
register is used to outlive the loop block when expanding the
VGPR -> m0 loop. This also now results in worse waitcnt insertion
before the loop instead of after for pending operations at the point
of the indexing, but that should be fixed by future improvements to
cross block waitcnt insertion.

v_movreld_b32's operands are now modeled more correctly since vdst
is not a true output. This is kind of a hack to treat vdst as a
use operand. Extra checking is required in the verifier since
I can't seem to get tablegen to emit an implicit operand for a
virtual register.

llvm-svn: 275934
2016-07-19 00:35:03 +00:00
Matt Arsenault 50b76399ed AMDGPU: Fix test name and broken CHECK-LABEL
llvm-svn: 275928
2016-07-18 23:09:51 +00:00
Artem Belevich 9f97dcb018 [NVPTX] Make sure we adjust alignment at all call sites
.. including calls from kernel functions that were
ignored by mistake before.

llvm-svn: 275920
2016-07-18 21:58:48 +00:00
Artem Belevich 052b1ed2fd [NVPTX] Force minimum alignment of 4 for byval arguments of device-side functions.
Taking address of a byval variable in PTX is legal, but currently runs
into miscompilation by ptxas on sm_50+ (NVIDIA issue 1789042).
Work around the issue by enforcing minimum alignment on byval arguments
of device functions.

The change is a no-op on SASS level for sm_3x where ptxas already aligns
local copy by at least 4.

Differential Revision: https://reviews.llvm.org/D22428

llvm-svn: 275893
2016-07-18 19:54:56 +00:00
Vitaly Buka c93e10fcbb Revert "[ARM] Skip inline asm memory operands in DAGToDAGISel"
Breaks asan, see https://reviews.llvm.org/D22103

This reverts commit r275776.

llvm-svn: 275890
2016-07-18 19:44:01 +00:00
Vitaly Buka fa474e3eb9 Revert "[ARM] Update test to use CHECK-LABEL. NFCI."
Breaks asan, see https://reviews.llvm.org/D22103

This reverts commit r275777.

llvm-svn: 275889
2016-07-18 19:43:58 +00:00
Simon Pilgrim 069c732f82 [X86][SSE] Regenerate extraction from promotion test
Added tests for SSE2 as well as SSE41

llvm-svn: 275878
2016-07-18 18:53:15 +00:00
Simon Pilgrim a68b8df3a7 [X86][SSE] Regenerate extraction+store memop tests
Added tests for SSE2 as well as SSE41+AVX

llvm-svn: 275876
2016-07-18 18:44:01 +00:00
Simon Pilgrim b21b47ba61 [X86][SSE] Regenerate truncate+extension memop tests
Added tests for SSE2 as well as SSE41

llvm-svn: 275875
2016-07-18 18:42:33 +00:00
Simon Pilgrim 600baaed89 Regenerate test
llvm-svn: 275872
2016-07-18 18:38:51 +00:00
Matt Arsenault c96e1deffa AMDGPU: Add intrinsic for s_flbit_i32/v_ffbh_i32
llvm-svn: 275871
2016-07-18 18:35:05 +00:00
Matt Arsenault 4c519d3518 AMDGPU/R600: Replace barrier intrinsics
llvm-svn: 275870
2016-07-18 18:34:59 +00:00
Matt Arsenault efb24540b1 AMDGPU: Remove dead check in AMDGPUPromoteAlloca
This is currently only called with GEP users. A direct
alloca would only happen with current typed pointers
for arrays which are a perverse case.

Also fix crashes on 0 x and 1 x arrays.

llvm-svn: 275869
2016-07-18 18:34:53 +00:00
Tim Northover 918f05063c CodeGenPrep: use correct function to determine Global's alignment.
Elsewhere (particularly computeKnownBits) we assume that a global will be
aligned to the value returned by Value::getPointerAlignment. This is used to
boost the alignment on memcpy/memset, so any target-specific request can only
increase that value.

llvm-svn: 275866
2016-07-18 18:28:52 +00:00
Simon Pilgrim c941f6b329 [X86][AVX] Add target shuffle decode support for VBROADCAST
Currently we only decode broadcasts from a vector of the same size.

llvm-svn: 275823
2016-07-18 17:32:59 +00:00
Krzysztof Parzyszek 5948ea78b9 [Hexagon] Handle returning small structures by value
This is compliant with the official ABI, but allows experimentation with
calling conventions.

llvm-svn: 275822
2016-07-18 17:30:41 +00:00
Chih-Hung Hsieh 4d9f2c154d [X86] Accept SELECT op code for x86-64 fp128 type
DAGTypeLegalizer::CanSkipSoftenFloatOperand should allow
SELECT op code for x86_64 fp128 type for MME targets,
so SoftenFloatOperand does not abort on SELECT op code.

Differential Revision: http://reviews.llvm.org/D21758

llvm-svn: 275818
2016-07-18 17:20:09 +00:00
Simon Pilgrim 4ac7420618 [X86][AVX2] Added tests that demonstrate duplicate broadcasts
We don't yet decode broadcasts as a target shuffle

llvm-svn: 275808
2016-07-18 16:17:34 +00:00
Krzysztof Parzyszek 786333ffcc [Hexagon] Enable .cur formation in MISched for Hexagon V60
Schedule a load and its use in the same packet in MISched. Previously,
isResourceAvailable was returning false for dependences in the same
packet, which prevented MISched from packetizing a load and its use in
the same packet for v60.

Patch by Ikhlas Ajbar.

llvm-svn: 275804
2016-07-18 16:05:27 +00:00
Nemanja Ivanovic d3c284f645 [PowerPC] Remove redundant direct moves when extracting integers and converting to FP
This patch corresponds to review:
https://reviews.llvm.org/D21354

We use direct moves for extracting integer elements from vectors. We also use
direct moves when converting integers to FP. When these operations are chained,
we get a direct move out of a VSR followed by a direct move back into a VSR.
These are redundant - all we need to do is line up the element and convert.

llvm-svn: 275796
2016-07-18 15:30:00 +00:00
Krzysztof Parzyszek 393b37937b [Hexagon] Use timing class info as tie-breaker in machine scheduler
Patch by Sirish Pande.

llvm-svn: 275794
2016-07-18 15:17:10 +00:00
Krzysztof Parzyszek 3467e9d0a9 [Hexagon] HexagonMachineScheduler should account for resources
The machine scheduler needs to account for available resources
more accurately in order to avoid scheduling an instruction that
forces a new packet to be created.

This occurs in two ways: First, an instruction without an available
resource may have a large priority due to other metrics and be
scheduled when there are other instructions with available resources.
Second, an instruction with a non-zero latency may become available
prematurely. In both these cases, we attempt change the priority
in order to allow a better instruction to be scheduled.

Patch by Brendon Cahoon.

llvm-svn: 275793
2016-07-18 14:52:13 +00:00
Krzysztof Parzyszek 748d3efec6 [Hexagon] Fix zero latency instructions with multiple predecessors
An instruction may have multiple predecessors that are candidates
for using .cur. However, only one of them can use .cur in the
packet. When this case occurs, we need to make sure that only
one of the dependences gets a 0 latency value.

Patch by Brendon Cahoon.

llvm-svn: 275790
2016-07-18 14:23:10 +00:00
Simon Dardis d32a2d30cb [inlineasm] Propagate operand constraints to the backend
When SelectionDAGISel transforms a node representing an inline asm
block, memory constraint information is not preserved. This can cause
constraints to be broken when a memory offset is of the form:

offset + frame index

when the frame is resolved.

By propagating the constraints all the way to the backend, targets can
enforce memory operands of inline assembly to conform to their constraints.

For MIPSR6, some instructions had their offsets reduced to 9 bits from
16 bits such as ll/sc. This becomes problematic when using inline assembly
to perform atomic operations, as an offset can generated that is too big to
encode in the instruction.

Reviewers: dsanders, vkalintris

Differential Review: https://reviews.llvm.org/D21615

llvm-svn: 275786
2016-07-18 13:17:31 +00:00
Nicolai Haehnle bef1ceb815 AMDGPU: Disable AMDGPUPromoteAlloca pass for shader calling conventions.
Summary:
The work item intrinsics are not available for the shader
calling conventions. And even if we did hook them up most
shader stages haves some extra restrictions on the amount
of available LDS.

Reviewers: tstellarAMD, arsenm

Subscribers: nhaehnle, arsenm, llvm-commits, kzhuravl

Differential Revision: https://reviews.llvm.org/D20728

llvm-svn: 275779
2016-07-18 09:02:47 +00:00
Diana Picus 6731f13458 [ARM] Update test to use CHECK-LABEL. NFCI.
llvm-svn: 275777
2016-07-18 07:48:42 +00:00
Diana Picus 73ed44d328 [ARM] Skip inline asm memory operands in DAGToDAGISel
The current logic for handling inline asm operands in DAGToDAGISel interprets
the operands by looking for constants, which should represent the flags
describing the kind of operand we're dealing with (immediate, memory, register
def etc). The operands representing actual data are skipped only if they are
non-const, with the exception of immediate operands which are skipped explicitly
when a flag describing an immediate is found.

The oversight is that memory operands may be const too (e.g. for device drivers
reading a fixed address), so we should explicitly skip the operand following a
flag describing a memory operand. If we don't, we risk interpreting that
constant as a flag, which is definitely not intended.

Fixes PR26038

Differential Revision: https://reviews.llvm.org/D22103

llvm-svn: 275776
2016-07-18 07:35:14 +00:00
Craig Topper a3c55f5915 [AVX512] Add EVEX versions of scalar ADD/SUB/MUL/DIV to load folding tables.
llvm-svn: 275775
2016-07-18 06:49:32 +00:00
Craig Topper 83613bb436 [X86] Fix test checks to include leading 'v' on avx mnemonic names.
llvm-svn: 275774
2016-07-18 06:49:29 +00:00
Diana Picus 774d157a5d [ARM] Honour ABI for rem under -O0 for EABI, GNUEABI, Android and Musl
At higher optimization levels, we generate the libcall for DIVREM_Ix, which is
fine: aeabi_{u|i}divmod. At -O0 we generate the one for REM_Ix, which is the
default {u}mod{q|h|s|d}i3.

This commit makes sure that we don't generate REM_Ix calls for ABIs that
don't support them (i.e. where we need to use DIVREM_Ix instead). This is
achieved by bailing out of FastISel, which can't handle non-double multi-reg
returns, and letting the legalization infrastructure expand the REM_Ix calls.

It also updates the divmod-eabi.ll test to run under -O0 as well, and adds some
Windows checks to it to make sure we don't break things for it.

Fixes PR27068

Differential Revision: https://reviews.llvm.org/D21926

llvm-svn: 275773
2016-07-18 06:48:25 +00:00
Craig Topper 1af6cc00dc [X86] Add VPADD instructions to X86InstrInfo::isAssociativeAndCommutative.
llvm-svn: 275769
2016-07-18 06:14:54 +00:00
Craig Topper ba9b93d7f2 [X86] Add floating point packed logical ops to X86InstrInfo::isAssociativeAndCommutative.
llvm-svn: 275768
2016-07-18 06:14:50 +00:00
Craig Topper 3a99de4067 [X86] Add AVX512 instructions to X86InstrInfo::isAssociativeAndCommutative.
llvm-svn: 275767
2016-07-18 06:14:47 +00:00
Craig Topper f7a06c29bc [X86] Add AVX512 load opcodes and a couple AVX load opcodes to X86InstrInfo::areLoadsFromSameBasePtr.
llvm-svn: 275765
2016-07-18 06:14:43 +00:00
Craig Topper 650a15e2b3 [X86] Add more opcodes to isFrameLoadOpcode/isFrameStoreOpcode. Mainly AVX-512 related.
llvm-svn: 275764
2016-07-18 06:14:39 +00:00