Commit Graph

109494 Commits

Author SHA1 Message Date
Craig Topper 7c2abdd249 [X86] Remove unnecessary isel pattern that is a combination of two other patterns.
The pattern was this

 def : Pat<(i32 (zext (i8 (bitconvert (v8i1 VK8:$src))))),
           (MOVZX32rr8 (EXTRACT_SUBREG (i32 (COPY_TO_REGCLASS VK8:$src, GR32)), sub_8bit))>, Requires<[NoDQI]>;

but if you just let (i32 (zext X)) match byte itself you'll get MOVZX32rr8. And if you let (i8 (bitconvert (v8i1 VK8:$src))) match by itself you'll get (EXTRACT_SUBREG (i32 (COPY_TO_REGCLASS VK8:$src, GR32)), sub_8bit).

So we can just let isel do the two patterns naturally.

llvm-svn: 322049
2018-01-09 00:50:42 +00:00
Jessica Paquette 3291e7353e [MachineOutliner] AArch64: Handle instrs that use SP and will never need fixups
This commit does two things. Firstly, it adds a collection of flags which can
be passed along to the target to encode information about the MBB that an
instruction lives in to the outliner.

Second, it adds some of those flags to the AArch64 outliner in order to add
more stack instructions to the list of legal instructions that are handled
by the outliner. The two flags added check if

- There are calls in the MachineBasicBlock containing the instruction
- The link register is available in the entire block

If the link register is available and there are no calls, then a stack
instruction can always be outlined without fixups, regardless of what it is,
since in this case, the outliner will never modify the stack to create a
call or outlined frame.

The motivation for doing this was checking which instructions are most often
missed by the outliner. Instructions like, say

%sp<def> = ADDXri %sp, 32, 0; flags: FrameDestroy

are very common, but cannot be outlined in the case that the outliner might
modify the stack. This commit allows us to outline instructions like this.
  

llvm-svn: 322048
2018-01-09 00:26:18 +00:00
Stefan Pintilie 55bfdd040a [PowerPC] Manually schedule the prologue and epilogue
This patch makes the following changes to the schedule of instructions in the
prologue and epilogue.

The stack pointer update is moved down in the prologue so that the callee saves
do not have to wait for the update to happen.
Saving the lr is moved down in the prologue to hide the latency of the mflr.
The stack pointer is moved up in the epilogue so that restoring of the lr can
happen sooner.
The mtlr is moved up in the epilogue so that it is away form the blr at the end
of the epilogue. The latency of the mtlr can now be hidden by the loads of the
callee saved registers.

Differential Revision: https://reviews.llvm.org/D41737

llvm-svn: 322036
2018-01-08 22:23:10 +00:00
Justin Bogner 6f6846fc9d AlwaysInliner: Alow setting InsertLifetime in the new-style pass
llvm-svn: 322033
2018-01-08 22:07:42 +00:00
Zachary Turner 59468f5a1e Fix uninitialized read error reported by MSAN.
The problem was that our Obj -> Yaml dumper had not been taught
to handle certain types of records.  This meant that when I
generated the test input files, the records were still there but
none of its fields were filled out.  So when it did the
Yaml -> Obj conversion as part of the test, it generated records
with garbage in them.

The patch here fixes the Obj <-> Yaml converter, and additionally
updates the test file with fresh Yaml generated by the fixed
converter.

llvm-svn: 322029
2018-01-08 21:38:50 +00:00
Justin Bogner 92fe563b57 ArgPromotion: Allow setting MaxElements in the new-style pass
llvm-svn: 322025
2018-01-08 21:13:35 +00:00
Sanjay Patel 7dfe96ad16 [ValueTracking] remove overzealous assert
The test is derived from a failing fuzz test:
https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=5008

Credit to @rksimon for pointing out the problem.

llvm-svn: 322016
2018-01-08 18:31:13 +00:00
Petar Jovanovic e9500ba745 [LiveDebugValues] Change condition for block termination recognition
The last iterator of MBB should be recognized as MBB.end() not as
MBB.instr_end() which could return bundled instruction that is not iterable
with basic iterator.

Patch by Nikola Prica.

Differential Revision: https://reviews.llvm.org/D41626

llvm-svn: 322015
2018-01-08 18:21:15 +00:00
Sanjay Patel 52149f0305 [TargetLibraryInfo] fix finite mathlib function availability
This patch was part of:
https://reviews.llvm.org/D41338
...but we can expose the bug in IR via constant propagation
as shown in the test. Unless the triple includes 'linux', we
should not fold these because the functions don't exist on
other platforms (yet?).

llvm-svn: 322010
2018-01-08 17:38:09 +00:00
Adrian McCarthy ce63a925cc Revert "Emit Function IDs table for Control Flow Guard"
The new test fails on the Hexagon bot.  Reverting while I investigate.

This reverts https://reviews.llvm.org/rL322005

This reverts commit b7e0026b4385180c378edc658ec91a39566f2942.

llvm-svn: 322008
2018-01-08 17:12:01 +00:00
Aleksandar Beserminji a734d409c6 [mips] Remove duplicated R6 EVA instructions
This patch removes duplicated EVA instructions in R6.

Differential Revision: https://reviews.llvm.org/D41769

llvm-svn: 322007
2018-01-08 16:50:33 +00:00
Davide Italiano 9a60d2c157 [CVP] Replace incoming values from unreachable blocks with undef.
This is an attempt of fixing PR35807.
Due to the non-standard definition of dominance in LLVM, where uses in
unreachable blocks are dominated by anything, you can have, in an
unreachable block:

  %patatino = OP1 %patatino, CONSTANT

When `SimplifyInstruction` receives a PHI where an incoming value is of
the aforementioned form, in some cases, loops indefinitely.

What I propose here instead is keeping track of the incoming values
from unreachable blocks, and replacing them with undef. It fixes this
case, and it seems to be good regardless (even if we can't prove that
the value is constant, as it's coming from an unreachable block, we
can ignore it).

Differential Revision:  https://reviews.llvm.org/D41812

llvm-svn: 322006
2018-01-08 16:34:06 +00:00
Adrian McCarthy cf6e6c82c1 Emit Function IDs table for Control Flow Guard
Adds option /guard:cf to clang-cl and -cfguard to cc1 to emit function IDs
of functions that have their address taken into a section named .gfids$y for
compatibility with Microsoft's Control Flow Guard feature.

Differential Revision: https://reviews.llvm.org/D40531

llvm-svn: 322005
2018-01-08 16:33:42 +00:00
Nirav Dave 6e2d03d410 [DAG] Teach BaseIndexOffset to correctly handle with indexed operations
BaseIndexOffset address analysis incorrectly ignores offsets folded
into indexed memory operations causing potential errors in alias
analysis of pre-indexed operations.

Reviewers: efriedma, RKSimon, hfinkel, jyknight

Subscribers: hiraditya, javed.absar, llvm-commits

Differential Revision: https://reviews.llvm.org/D41701

llvm-svn: 322003
2018-01-08 16:21:35 +00:00
Sanjay Patel 31b4b76f99 [InstCombine] fold min/max tree with common operand (PR35717)
There is precedence for factorization transforms in instcombine for FP ops with fast-math. 
We also have similar logic in foldSPFofSPF().

It would take more work to add this to reassociate because that's specialized for binops, 
and min/max are not binops (or even single instructions). Also, I don't have evidence that 
larger min/max trees than this exist in real code, but if we find that's true, we might
want to reorganize where/how we do this optimization.

In the motivating example from https://bugs.llvm.org/show_bug.cgi?id=35717 , we have:

int test(int xc, int xm, int xy) {
  int xk;
  if (xc < xm)
    xk = xc < xy ? xc : xy;
  else
    xk = xm < xy ? xm : xy;
  return xk;
}

This patch solves that problem because we recognize more min/max patterns after rL321672

https://rise4fun.com/Alive/Qjne
https://rise4fun.com/Alive/3yg

Differential Revision: https://reviews.llvm.org/D41603

llvm-svn: 321998
2018-01-08 15:05:34 +00:00
Momchil Velikov ac7c5c1d92 [ARM] Fix PR35379 - incorrect unwind information when compiling with -Oz
The patch makes the unwind information not mention registers, which were pushed
solely for the purpose of saving stack adjustment instructions.

Differential revision: https://reviews.llvm.org/D41300
Fixes https://bugs.llvm.org/show_bug.cgi?id=35379

llvm-svn: 321996
2018-01-08 14:47:19 +00:00
Alexey Bataev 5b9a77d4ea [SLP] Fix PR35777: Incorrect handling of aggregate values.
Summary:
Fixes the bug with incorrect handling of InsertValue|InsertElement
instrucions in SLP vectorizer. Currently, we may use incorrect
ExtractElement instructions as the operands of the original
InsertValue|InsertElement instructions.

Reviewers: mkuper, hfinkel, RKSimon, spatel

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D41767

llvm-svn: 321994
2018-01-08 14:43:06 +00:00
Alexey Bataev 118a0a2c38 [SLP] Fix PR35628: Count external uses on extra reduction arguments.
Summary:
If the vectorized value is marked as extra reduction argument, its users
are not considered as external users. Patch fixes this.

Reviewers: mkuper, hfinkel, RKSimon, spatel

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D41786

llvm-svn: 321993
2018-01-08 14:33:11 +00:00
Sam Parker 3800f0f11d [DAGCombine] Fix for PR35761
I had falsely assumed that constant operands would be operand(1) of
the bin ops that may need their constant operand to be masked.

Bugzilla: https://bugs.llvm.org/show_bug.cgi?id=35761

Differential Revision: https://reviews.llvm.org/D41667

llvm-svn: 321991
2018-01-08 13:21:24 +00:00
Jonas Paulsson 22f208f034 [SystemZ] Comment fix in SystemZElimCompare.cpp
NFC

Review: Ulrich Weigand
llvm-svn: 321990
2018-01-08 12:52:40 +00:00
Momchil Velikov d17dabca31 [ARM] Fix PR35481
This patch allows `r7` to be used, regardless of its use as a frame pointer, as
a temporary register when popping `lr`, and also falls back to using a high
temporary register if, for some reason, we weren't able to find a suitable low
one.

Differential revision: https://reviews.llvm.org/D40961
Fixes https://bugs.llvm.org/show_bug.cgi?id=35481

llvm-svn: 321989
2018-01-08 11:32:37 +00:00
Francis Visoiu Mistrih d52da12822 [X86] Remove side-effects from determineCalleeSaves
(Target)FrameLowering::determineCalleeSaves can be called multiple
times. I don't think it should have side-effects as creating stack
objects and setting global MachineFunctionInfo state as it is doing
today (in other back-ends as well).

This moves the creation of stack objects from determineCalleeSaves to
assignCalleeSavedSpillSlots.

Differential Revision: https://reviews.llvm.org/D41703

llvm-svn: 321987
2018-01-08 10:46:05 +00:00
Craig Topper f090e8a89a [X86] Replace CVT2MASK ISD opcode with PCMPGTM compared to zero.
CVT2MASK is just checking the sign bit which can be represented with a comparison with zero.

llvm-svn: 321985
2018-01-08 06:53:54 +00:00
Craig Topper a2018e799a [X86] Add patterns to allow 512-bit BWI compare instructions to be used for 128/256-bit compares when VLX is not available.
llvm-svn: 321984
2018-01-08 06:53:52 +00:00
Craig Topper 9f5859e3ee [X86] Simplify some code in lower1BitVectorShuffle by relying on getNode's ability to constant fold vector SIGN_EXTEND.
llvm-svn: 321979
2018-01-07 23:56:37 +00:00
Craig Topper 03d8e516cf [X86] Add VSHUFF32X4 and similar instructions to load folding tables.
llvm-svn: 321978
2018-01-07 23:30:20 +00:00
Davide Italiano e15bffe9ea Revert "[SCCP] Manually fold branches on undef."
I thought this was responsible for PR35723, but I was
wrong, the issue lies elsewhere. Revert while I debug.

llvm-svn: 321975
2018-01-07 22:09:44 +00:00
Davide Italiano 4c39758a38 [SLPVectorizer] Reintroduce std::stable_sort(properlyDominates()).
The approach was never discussed, I wasn't able to reproduce this
non-determinism, and the original author went AWOL.
After a discussion on the ML, Philip suggested to revert this.

llvm-svn: 321974
2018-01-07 22:06:24 +00:00
Craig Topper e9f44e1b80 [X86] Revert accidental change to CMakeLists.txt in r321952
I had removed the qualifiers around the autogenerated folding table so I could compare with the manual table, but didn't intend to commit the change.

llvm-svn: 321971
2018-01-07 21:03:43 +00:00
Simon Pilgrim 998180dad3 [DAG] Fix for Bug PR34620 - Allow SimplifyDemandedBits to look through bitcasts
Allow SimplifyDemandedBits to use TargetLoweringOpt::computeKnownBits to look through bitcasts. This can help simplifying in some cases where bitcasts of constants generated during or after legalization can't be folded away, and thus didn't get picked up by SimplifyDemandedBits. This fixes PR34620, where a redundant pand created during legalization from lowering and lshr <16xi8> wasn't being simplified due to the presence of a bitcasted build_vector as an operand.

Committed on the behalf of @sameconrad (Sam Conrad)

Differential Revision: https://reviews.llvm.org/D41643

llvm-svn: 321969
2018-01-07 19:09:40 +00:00
Craig Topper c1ec57c3e2 [X86] Remove unneeded code from combineGatherScatter that used to delte SIGN_EXTEND_INREG nodes created during legalization of v2i1/v4i1 masks on KNL.
v2i1/v4i1 are now legal on KNL so no sign_extend_inreg is generated.

llvm-svn: 321968
2018-01-07 18:34:08 +00:00
Craig Topper d58c165545 [X86] Make v2i1 and v4i1 legal types without VLX
Summary:
There are few oddities that occur due to v1i1, v8i1, v16i1 being legal without v2i1 and v4i1 being legal when we don't have VLX. Particularly during legalization of v2i32/v4i32/v2i64/v4i64 masked gather/scatter/load/store. We end up promoting the mask argument to these during type legalization and then have to widen the promoted type to v8iX/v16iX and truncate it to get the element size back down to v8i1/v16i1 to use a 512-bit operation. Since need to fill the upper bits of the mask we have to fill with 0s at the promoted type.

It would be better if we could just have the v2i1/v4i1 types as legal so they don't undergo any promotion. Then we can just widen with 0s directly in a k register. There are no real v4i1/v2i1 instructions anyway. Everything is done on a larger register anyway.

This also fixes an issue that we couldn't implement a masked vextractf32x4 from zmm to xmm properly.

We now have to support widening more compares to 512-bit to get a mask result out so new tablegen patterns got added.

I had to hack the legalizer for widening the operand of a setcc a bit so it didn't try create a setcc returning v4i32, extract from it, then try to promote it using a sign extend to v2i1. Now we create the setcc with v4i1 if the original setcc's result type is v2i1. Then extract that and don't sign extend it at all.

There's definitely room for improvement with some follow up patches.

Reviewers: RKSimon, zvi, guyblank

Reviewed By: RKSimon

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D41560

llvm-svn: 321967
2018-01-07 18:20:37 +00:00
Hal Finkel 0f1314c5ee [LV][VPlan] NFC patch to move LoopVectorizationPlanner class out of LoopVectorize.cpp
Another small step forward to move VPlan stuff outside of LoopVectorize.cpp.

VPlanBuilder.h is renamed to LoopVectorizationPlanner.h
LoopVectorizationPlanner class is moved from LoopVectorize.cpp to
LoopVectorizationPlanner.h LoopVectorizationCostModel::VectorizationFactor
class is moved to LoopVectorizationPlanner.h (used by the planner class) ---
this needs further streamlining work in later patches and thus all I did was
take it out of the CostModel class and moved to the header file.  The callback
function had to stay inside LoopVectorize.cpp since it calls an
InnerLoopVectorizer member function declared in it.  Next Steps: Make
InnerLoopVectorizer, LoopVectorizationCostModel, and other classes more modular
and more aligned with VPlan direction, in small increments.

Previous step was: r320900 (https://reviews.llvm.org/D41045)

Patch by Hideki Saito, thanks!

Differential Revision: https://reviews.llvm.org/D41420

llvm-svn: 321962
2018-01-07 16:02:58 +00:00
Florian Hahn 55be37e7d4 [CodeExtractor] Use subset of function attributes for extracted function.
In addition to target-dependent attributes, we can also preserve a
white-listed subset of target independent function attributes. The white-list
excludes problematic attributes, most prominently:

* attributes related to memory accesses, as alloca instructions
  could be moved in/out of the extracted block

* control-flow dependent attributes, like no_return or thunk, as the
  relerelevant instructions might or might not get extracted.

Thanks @efriedma and @aemerson for providing a set of attributes that cannot be
propagated.


Reviewers: efriedma, davidxl, davide, silvas

Reviewed By: efriedma

Differential Revision: https://reviews.llvm.org/D41334

llvm-svn: 321961
2018-01-07 11:22:25 +00:00
Craig Topper d461aefe5f [PowerPC] Add an ISD::TRUNCATE to the legalization for ppc_is_decremented_ctr_nonzero
Summary:
I believe legalization is really expecting that ReplaceNodeResults will return something with the same type as the thing that's being legalized. Ultimately, it uses the output to replace the uses in the DAG so the type should match to make that work.

There are two relevant cases here. When crbits are enabled, then i1 is a legal type and getSetCCResultType should return i1. In this case, the truncate will be between i1 and i1 and should be removed (SelectionDAG::getNode does this). Otherwise, getSetCCResultType will be i32 and the legalizer will promote the truncate to be i32 -> i32 which will be similarly removed.

With this fixed we can remove some code from PromoteIntRes_SETCC that seemed to only exist to deal with the intrinsic being replaced with a larger type without changing the other operand. With the truncate being used for connectivity this doesn't happen anymore.

Reviewers: hfinkel

Reviewed By: hfinkel

Subscribers: nemanjai, llvm-commits, kbarton

Differential Revision: https://reviews.llvm.org/D41654

llvm-svn: 321959
2018-01-07 07:51:36 +00:00
Craig Topper a21f551109 [X86] Add the 16 and 8-bit CRC32 instructions to the load folding tables.
llvm-svn: 321958
2018-01-07 06:48:20 +00:00
Craig Topper d0859a03b5 [X86] Correct the load folding flags for xmm fp->mmx conversion instructions.
The instructions that load 64-bits or an xmm register should be TB_NO_REVERSE to avoid the load being widened during unfold. The instructions that load 128-bits need to ensure 128-bit alignment.

llvm-svn: 321956
2018-01-07 06:24:30 +00:00
Craig Topper aa73941176 [X86] Add TB_NO_REVERSE to some scalar intrinsic instructions in the load folding table.
llvm-svn: 321955
2018-01-07 06:24:29 +00:00
Craig Topper 85657d59a9 [X86] Don't put any EVEX_B instructions in the tablegen generated load folding tables.
EVEX_B means different things for memory and register forms. The instructions should not be considered equivalent.

llvm-svn: 321954
2018-01-07 06:24:28 +00:00
Craig Topper 89293a2a94 [X86] Add 128 and 256-bit VPOPCNTD/Q instructions to load folding tables.
llvm-svn: 321953
2018-01-07 06:24:27 +00:00
Craig Topper a124ab10ef [X86] Add some 8 and 16-bit instructions to the load folding tables.
llvm-svn: 321952
2018-01-07 06:24:25 +00:00
Craig Topper 11aede13db [X86] Add EVEX vcvtph2ps to the load folding tables.
llvm-svn: 321951
2018-01-07 06:24:24 +00:00
Craig Topper 40cc8338f7 [X86] Remove cvtps2ph xmm->xmm from store folding tables. Add the evex versions of cvtps2ph to the store folding tables.
The memory form of the xmm->xmm version only writes 64-bits. If we use it in the folding tables and its get used for a stack spill, only half the slot will be written. Then a reload may read all 128-bits which will pull in garbage. But without the spill the upper bits of the register would have been zero. By not folding we would preserve the zeros.

llvm-svn: 321950
2018-01-07 06:24:23 +00:00
Craig Topper 8fa800b834 [X86] Add CMP8ri8 to load folding tables.
llvm-svn: 321949
2018-01-07 06:24:21 +00:00
Craig Topper cf93feb981 [X86] Remove assembler predicates from all AVX512 related feature flags.
We don't do fine grained feature control like this on features prior to AVX512.

We do still have checks in place in the assembly parser itself that prevents %zmm references or %xmm16-31 from being parsed without at least -mattr=avx512f. Same for rounding control and mask operands. That will prevent the table matcher from matching for any instructions that need those features and that's probably good enough.

llvm-svn: 321947
2018-01-06 21:45:30 +00:00
Craig Topper 61d8a60e23 [X86] Remove memory forms of EVEX encoded vcvttss2si/vcvttsd2si from asm matcher table.
This is also needed to fix PR35837.

llvm-svn: 321946
2018-01-06 21:27:25 +00:00
Craig Topper 0f4ccb7806 [X86] Add load folding pattern to EVEX vcvttss2si/vcvtsd2si.
llvm-svn: 321945
2018-01-06 21:02:26 +00:00
Craig Topper 90353a9f42 [X86] Remove an unnecessary VCVTTSD2SIrrb/VCVTSS2SIrrb instruction with no isel pattern that only existed for the assembler. Use VCVTTSD2SIrrb_Int instead.
For consistency use the _Int version of VCVTTSD2SIrr_Int and VCVTTSD2SIrm_Int for the assembler as well.

llvm-svn: 321944
2018-01-06 21:02:22 +00:00
Florian Hahn a82eef2363 [InlineFunction] Preserve calling convention when forwarding VarArgs.
Reviewers: efriedma, rnk, davide

Reviewed By: rnk, davide

Differential Revision: https://reviews.llvm.org/D41556

llvm-svn: 321943
2018-01-06 20:56:27 +00:00
Florian Hahn de10e6e064 [InlineFunction] Preserve attributes when forwarding VarArgs.
Reviewers: rnk, efriedma

Reviewed By: rnk

Differential Revision: https://reviews.llvm.org/D41555

llvm-svn: 321942
2018-01-06 20:46:00 +00:00