OSDN Git Service
Sanjoy Das [Sat, 11 Mar 2017 01:15:48 +0000 (01:15 +0000)]
Use a WeakVH for UnknownInstructions in AliasSetTracker
Summary:
This change solves the same problem as D30726, except that this only
throws out the bathwater.
AST was not correctly tracking and deleting UnknownInstructions via
handles. The existing code only tracks "pointers" in its
`ASTCallbackVH`, so an UnknownInstruction (that isn't also def'ing a
pointer used by another memory instruction) never gets a
`ASTCallbackVH`.
There are two other ways to solve this problem:
- Use the `PointerRec` scheme for both known and unknown instructions.
- Use a `CallbackVH` that erases the offending Instruction from the
UnknownInstruction list.
Both of the above changes seemed to be significantly (and unnecessarily
IMO) more complex than this.
Reviewers: chandlerc, dberlin, hfinkel, reames
Subscribers: mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D30849
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297539
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Sat, 11 Mar 2017 00:51:01 +0000 (00:51 +0000)]
VNCoercion: Make the function signatures all consistent
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297537
91177308-0d34-0410-b5e6-
96231b3b80d8
Stanislav Mekhanoshin [Sat, 11 Mar 2017 00:29:27 +0000 (00:29 +0000)]
[AMDGPU] Remove getBidirectionalReasonRank
This method inverts the Reason field of a scheduling candidate.
It does right comparison between RegCritical and RegExcess, but
everything else is broken. In fact it can prefer less strong reason
such as Weak over RegCritical because Weak > -RegCritical.
The CandReason enum is properly sorted, so just remove artificial
ranking.
Differential Revision: https://reviews.llvm.org/D30557
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297536
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Sat, 11 Mar 2017 00:28:33 +0000 (00:28 +0000)]
[IRTranslator] Simplify error handling for translating constants. NFC.
We don't need to check whether the fallback path is enabled to return
false. Just do that all the time on error cases, the caller knows (or
at least should know!) how to handle the failing case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297535
91177308-0d34-0410-b5e6-
96231b3b80d8
Stanislav Mekhanoshin [Sat, 11 Mar 2017 00:14:52 +0000 (00:14 +0000)]
Fix subreg value numbers in handleMoveUp
The problem can occur in presence of subregs. If we are swapping two
instructions defining different subregs of the same register we will
get a new liveout from a block. We need to preserve value number for
block's liveout for successor block's livein to match.
Differential Revision: https://reviews.llvm.org/D30558
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297534
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 22:53:19 +0000 (22:53 +0000)]
Strip trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297529
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 22:44:47 +0000 (22:44 +0000)]
Fix redundant condition (PR32138)
'!A || (A && B)' is equivalent to '!A || B'
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297527
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Fri, 10 Mar 2017 22:44:24 +0000 (22:44 +0000)]
[RDF] Remove the map of reaching defs from copy propagation
Use Liveness::getNearestAliasedRef to find the reaching def instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297526
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Fri, 10 Mar 2017 22:42:17 +0000 (22:42 +0000)]
[RDF] Implement Liveness::getNearestAliasedRef(Reg, Inst)
This function will find the closest ref node aliased to Reg that is
in an instruction preceding Inst. This could be used to identify the
hypothetical reaching def of Reg, if Reg was a member of Inst.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297524
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 22:35:07 +0000 (22:35 +0000)]
[X86][SSE] Fix load folding for (V)CVTDQ2PD
This only requires a 64-bit memory source, not the whole 128-bits. But the 128-bit case is still supported via X86InstrInfo::foldMemoryOperandImpl
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297523
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 22:10:34 +0000 (22:10 +0000)]
[X86] Fix Wunused-lambda-capture warning
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297521
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 21:55:24 +0000 (21:55 +0000)]
[X86][RTM] Regenerate RTM intrinsic tests for 32/64-bit targets.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297518
91177308-0d34-0410-b5e6-
96231b3b80d8
Jonathan Roelofs [Fri, 10 Mar 2017 21:44:16 +0000 (21:44 +0000)]
Respect CMAKE_INSTALL_MANDIR for sphinx generated manpages
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297516
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Fri, 10 Mar 2017 21:37:10 +0000 (21:37 +0000)]
LTO: Hash type identifier resolutions for WholeProgramDevirt.
Differential Revision: https://reviews.llvm.org/D30555
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297514
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Fri, 10 Mar 2017 21:35:17 +0000 (21:35 +0000)]
LTO: Hash type identifier resolutions for LowerTypeTests.
Differential Revision: https://reviews.llvm.org/D30553
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297513
91177308-0d34-0410-b5e6-
96231b3b80d8
Volkan Keles [Fri, 10 Mar 2017 21:25:09 +0000 (21:25 +0000)]
[GlobalISel] LegalizerHelper: Lower (G_FSUB X, Y) to (G_FADD X, (G_FNEG Y))
Summary: No test case as none of the in-tree targets with GlobalISel support has this condition.
Reviewers: qcolombet, aditya_nandakumar, dsanders, t.p.northover, ab
Reviewed By: qcolombet
Subscribers: dberris, rovka, kristof.beyls, llvm-commits, igorb
Differential Revision: https://reviews.llvm.org/D30786
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297512
91177308-0d34-0410-b5e6-
96231b3b80d8
Volkan Keles [Fri, 10 Mar 2017 21:23:13 +0000 (21:23 +0000)]
GlobalISel: Translate ConstantAggregateZero vectors
Reviewers: qcolombet, aditya_nandakumar, dsanders, ab, t.p.northover, javed.absar
Reviewed By: qcolombet
Subscribers: dberris, rovka, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D30259
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297509
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 10 Mar 2017 21:20:17 +0000 (21:20 +0000)]
Sink accessing TII to fix release Werror builds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297507
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Fri, 10 Mar 2017 20:50:51 +0000 (20:50 +0000)]
[ProfileSummaryInfo] Remove unneeded braces. NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297506
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Fri, 10 Mar 2017 20:44:39 +0000 (20:44 +0000)]
Fix all these headers to properly mark the doxygen comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297505
91177308-0d34-0410-b5e6-
96231b3b80d8
Evandro Menezes [Fri, 10 Mar 2017 20:20:04 +0000 (20:20 +0000)]
[AArch64, X86] Additional debug information for MacroFusion
In order to make it easier to parse information about the performance of
MacroFusion, this patch adds the function and the instruction names to the
debug output of this pass.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297504
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Fri, 10 Mar 2017 20:13:58 +0000 (20:13 +0000)]
WholeProgramDevirt: Implement export/import support for VCP.
Differential Revision: https://reviews.llvm.org/D30017
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297503
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Fri, 10 Mar 2017 20:09:11 +0000 (20:09 +0000)]
WholeProgramDevirt: Implement export/import support for unique ret val opt.
Differential Revision: https://reviews.llvm.org/D29917
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297502
91177308-0d34-0410-b5e6-
96231b3b80d8
Dehao Chen [Fri, 10 Mar 2017 19:45:16 +0000 (19:45 +0000)]
Refactor the PSI to extract getCallSiteCount and remove checks for profile type.
Summary: There is no need to check profile count as only CallInst will have metadata attached.
Reviewers: eraman
Reviewed By: eraman
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30799
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297500
91177308-0d34-0410-b5e6-
96231b3b80d8
Konstantin Zhuravlyov [Fri, 10 Mar 2017 19:39:07 +0000 (19:39 +0000)]
[AMDGPU] Split R600/SI getFrameIndexReference and emit stack object offsets for SI
Differential Revision: https://reviews.llvm.org/D29674
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297499
91177308-0d34-0410-b5e6-
96231b3b80d8
Yaxun Liu [Fri, 10 Mar 2017 19:35:43 +0000 (19:35 +0000)]
Rename PT_NOTE namespace name used in AMDGPUPTNote.h
Patch by Guansong Zhang.
Differential Revision: https://reviews.llvm.org/D30750
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297498
91177308-0d34-0410-b5e6-
96231b3b80d8
Volkan Keles [Fri, 10 Mar 2017 19:08:28 +0000 (19:08 +0000)]
[GlobalISel] Translate insertelement and extractelement
Reviewers: qcolombet, aditya_nandakumar, dsanders, ab, t.p.northover, javed.absar
Reviewed By: qcolombet
Subscribers: dberris, rovka, llvm-commits, kristof.beyls
Differential Revision: https://reviews.llvm.org/D30761
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297495
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Fri, 10 Mar 2017 19:05:04 +0000 (19:05 +0000)]
NewGVN: Rename InitialClass to TOP, which is what most people would expect it to be called
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297494
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Kuperstein [Fri, 10 Mar 2017 18:59:07 +0000 (18:59 +0000)]
[SLP] Revert everything that has to do with memory access sorting.
This reverts r293386, r294027, r294029 and r296411.
Turns out the SLP tree isn't actually a "tree" and we don't handle
accessing the same packet of loads in several different orders well,
causing miscompiles.
Revert until we can fix this properly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297493
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 18:36:46 +0000 (18:36 +0000)]
[SelectionDAG] Add support for BUILD_VECTOR to ComputeNumSignBits
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297492
91177308-0d34-0410-b5e6-
96231b3b80d8
Volkan Keles [Fri, 10 Mar 2017 18:34:57 +0000 (18:34 +0000)]
[GlobalISel] Make LegalizerInfo accessible in LegalizerHelper
Summary:
We don’t actually use LegalizerInfo in Legalizer pass, it’s just passed
as an argument.
In order to check if an instruction is legal or not, we need to get LegalizerInfo
by calling `MI.getParent()->getParent()->getSubtarget().getLegalizerInfo()`.
Instead, make LegalizerInfo accessible in LegalizerHelper.
Reviewers: qcolombet, aditya_nandakumar, dsanders, ab, t.p.northover, kristof.beyls
Reviewed By: qcolombet
Subscribers: dberris, llvm-commits, rovka
Differential Revision: https://reviews.llvm.org/D30838
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297491
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Fri, 10 Mar 2017 18:34:37 +0000 (18:34 +0000)]
CODE_OWNERS: Take code ownership of instruction scheduling.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297490
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Fri, 10 Mar 2017 18:34:33 +0000 (18:34 +0000)]
CODE_OWNERS: Sort names
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297489
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Fri, 10 Mar 2017 18:33:41 +0000 (18:33 +0000)]
[Support] Don't return an error if realPath fails.
In openFileForRead, we would not previously return an error
if real_path resolution failed. After a recent patch, we
started propagating this error up. This caused a failure
in clang when trying to call openFileForRead("nul"). This
patch restores the previous behavior of not propagating this
error up.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297488
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 18:01:53 +0000 (18:01 +0000)]
[X86][SSE] Added tests showing missed truncations for sitofp conversion
SelectionDAG::ComputeNumSignBits is poor at build_vector handling, meaning that we can't see that all the vXi64 sources are in fact sign extended i32 or smaller.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297486
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Fri, 10 Mar 2017 17:47:13 +0000 (17:47 +0000)]
Fix test failure when Home directory cannot be found.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297484
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Fri, 10 Mar 2017 17:39:21 +0000 (17:39 +0000)]
Add llvm::sys::fs::real_path.
LLVM already has real_path like functionality, but it is
cumbersome to use and involves clean up after (e.g. you have
to call openFileForRead, then close the resulting FD).
Furthermore, on Windows it doesn't work for directories since
opening a directory and opening a file require slightly
different flags.
So I add a simple function `real_path` which works for all
paths on all platforms and has a simple to use interface.
In doing so, I add the ability to opt in to resolving tilde
expressions (e.g. ~/foo), which are normally handled by
the shell.
Differential Revision: https://reviews.llvm.org/D30668
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297483
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Fri, 10 Mar 2017 17:26:44 +0000 (17:26 +0000)]
[SelectionDAG] Make SelectionDAG aware of the known bits in USUBO and SSUBO and SUBC.
Summary:
Depends on D30379
This improves the state of things for the sub class of operation.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30436
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297482
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 17:23:55 +0000 (17:23 +0000)]
[X86][MMX] Add tests showing missed opportunities to use MMX sitofp conversions
If we are transferring MMX registers to XMM for conversion we could use the MMX equivalents (CVTPI2PD + CVTPI2PS) without affecting rounding/exceptions etc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297481
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Fri, 10 Mar 2017 17:06:52 +0000 (17:06 +0000)]
[SelectionDAG] Make SelectionDAG aware of the known bits in UADDO and SADDO.
Summary: As per title. This is extracted from D29872 and I threw SADDO in.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30379
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297479
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 16:59:43 +0000 (16:59 +0000)]
[X86][MMX] Add tests showing missed opportunities to use MMX fptosi conversions
If we are transferring XMM conversion results to MMX registers we could use the MMX equivalents (CVTPD2PI/CVTTPD2PI + CVTPS2PI/CVTTPS2PI) with affecting rounding/expections etc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297476
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 16:18:50 +0000 (16:18 +0000)]
[X86][MMX] Updated bad stack spill shift value test to actually show the problem
Cleaning up the ir had stopped showing the issue.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297475
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 16:07:39 +0000 (16:07 +0000)]
[X86][MMX] Regenerate mmx bitcast tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297474
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 15:53:41 +0000 (15:53 +0000)]
[X86][MMX] Add test showing bad stack spill of shift value
i32 is spilled to stack but 64-bit mmx is reloaded - leaving garbage in the other half of the register
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297471
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 15:41:05 +0000 (15:41 +0000)]
[X86][MMX] Regenerate mmx load folding tests
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297470
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 14:16:55 +0000 (14:16 +0000)]
Fix signed/unsigned comparison warnings
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297460
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 14:01:50 +0000 (14:01 +0000)]
Fix Wdocumentation warning
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297459
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Fri, 10 Mar 2017 13:44:32 +0000 (13:44 +0000)]
[APInt] Add APInt::insertBits() method to insert an APInt into a larger APInt
We currently have to insert bits via a temporary variable of the same size as the target with various shift/mask stages, resulting in further temporary variables, all of which require the allocation of memory for large APInts (MaskSizeInBits > 64).
This is another of the compile time issues identified in PR32037 (see also D30265).
This patch adds the APInt::insertBits() helper method which avoids the temporary memory allocation and masks/inserts the raw bits directly into the target.
Differential Revision: https://reviews.llvm.org/D30780
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297458
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Dardis [Fri, 10 Mar 2017 13:27:14 +0000 (13:27 +0000)]
[mips][msa] Accept more values for constant splats
This patches teaches the MIPS backend to accept more values for constant
splats. Previously, only 10 bit signed immediates or values that could be
loaded using an ldi.[bhwd] instruction would be acceptted. This patch relaxes
that constraint so that any constant value that be splatted is accepted.
As a result, the constant pool is used less for vector operations, and the
suite of bit manipulation instructions b(clr|set|neg)i can now be used with
the full range of their immediate operand.
Reviewers: slthakur
Differential Revision: https://reviews.llvm.org/D30640
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297457
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Fri, 10 Mar 2017 13:21:12 +0000 (13:21 +0000)]
imm_comp_XFORM (defined in ARMInstrThumb.td) duplicates imm_not_XFORM (defined in ARMInstrInfo.td)
Reviewers: grosbach, rengolin, jmolloy
Reviewed By: jmolloy
Subscribers: aemerson, llvm-commits
Differential Revision: https://reviews.llvm.org/D30782
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297456
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanne Wouda [Fri, 10 Mar 2017 13:08:20 +0000 (13:08 +0000)]
[Assembler] Add location info to unary expressions.
Summary:
This is a continuation of D28861. Add an SMLoc to MCUnaryExpr such that
a better diagnostic can be given in case of an error in later stages of
assembling.
Reviewers: rengolin, grosbach, javed.absar, olista01
Reviewed By: olista01
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30581
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297454
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Fri, 10 Mar 2017 12:41:33 +0000 (12:41 +0000)]
Refactor the multiply-accumulate combines to act on
ARMISD::ADD[CE] nodes, instead of the generic ISD::ADD[CE].
Summary:
This allows for some simplification because the combines
are no longer limited to just one go at the node before
it gets legalized into an ARM target-specific one.
Reviewers: jmolloy, rogfer01
Subscribers: aemerson, llvm-commits, rengolin
Differential Revision: https://reviews.llvm.org/D30401
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297453
91177308-0d34-0410-b5e6-
96231b3b80d8
George Rimar [Fri, 10 Mar 2017 10:31:56 +0000 (10:31 +0000)]
WholeProgramDevirt: Fixed compilation error under MSVS2015.
It was introduced in:
r296945
WholeProgramDevirt: Implement exporting for single-impl devirtualization.
---------------------
r296939
WholeProgramDevirt: Add any unsuccessful llvm.type.checked.load devirtualizations to the list of llvm.type.test users.
---------------------
Microsoft Visual Studio Community 2015
Version 14.0.23107.0 D14REL
Does not compile that code without additional brackets, showing multiple error like below:
WholeProgramDevirt.cpp(1216): error C2958: the left bracket '[' found at 'c:\access_softek\llvm\lib\transforms\ipo\wholeprogramdevirt.cpp(1216)' was not matched correctly
WholeProgramDevirt.cpp(1216): error C2143: syntax error: missing ']' before '}'
WholeProgramDevirt.cpp(1216): error C2143: syntax error: missing ';' before '}'
WholeProgramDevirt.cpp(1216): error C2059: syntax error: ']'
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297451
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Fri, 10 Mar 2017 08:22:25 +0000 (08:22 +0000)]
[llvm-readobj] Support SHT_MIPS_DWARF section type flag
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297448
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Fri, 10 Mar 2017 08:22:20 +0000 (08:22 +0000)]
[MC] Set SHT_MIPS_DWARF section type for all .debug_* sections on MIPS
All MIPS .debug_* sections should be marked with ELF type SHT_MIPS_DWARF
accordingly the specification [1]. Also the same section type is assigned
to these sections by GNU tools.
[1] ftp.software.ibm.com/software/os390/czos/dwarf/mips_extensions.pdf
Differential Revision: https://reviews.llvm.org/D29789
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297447
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Atanasyan [Fri, 10 Mar 2017 08:22:13 +0000 (08:22 +0000)]
[MC] Accept a numeric value as an ELF section header's type
GAS supports specification of section header's type using a numeric
value [1]. This patch brings the same functionality to LLVM. That allows
to setup some target-specific section types belong to the SHT_LOPROC -
SHT_HIPROC range. If we attempt to print unknown section type, MCSectionELF
class shows an error message. It's better than print sole '@' sign
without any section type name.
In case of MIPS, example of such section's type is SHT_MIPS_DWARF.
Without the patch we will have to implement some workarounds
in probably not-MIPS-specific part of code base to convert SHT_MIPS_DWARF
to the @progbits while printing assembly and to assign SHT_MIPS_DWARF for
@progbits sections named .debug_* if we encounter such section in
an input assembly.
[1] https://sourceware.org/binutils/docs/as/Section.html
Differential Revision: https://reviews.llvm.org/D29719
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297446
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Fri, 10 Mar 2017 07:40:27 +0000 (07:40 +0000)]
For Thumb1, lower ADDC/ADDE/SUBC/SUBE via the glueless ARMISD nodes,
same as already done for ARM and Thumb2.
Reviewers: jmolloy, rogfer01, efriedma
Subscribers: aemerson, llvm-commits, rengolin
Differential Revision: https://reviews.llvm.org/D30400
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297443
91177308-0d34-0410-b5e6-
96231b3b80d8
Yaron Keren [Fri, 10 Mar 2017 07:09:20 +0000 (07:09 +0000)]
Implement getPassName() for IR printing passes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297442
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 10 Mar 2017 05:25:49 +0000 (05:25 +0000)]
AMDGPU: Fix insertion point when reducing load intrinsics
The insertion point may be later than the next instruction,
so it is necessary to set it when replacing the call.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297439
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Fri, 10 Mar 2017 04:54:10 +0000 (04:54 +0000)]
Move memory coercion functions from GVN.cpp to VNCoercion.cpp so they can be shared between GVN and NewGVN.
Summary:
These are the functions used to determine when values of loads can be
extracted from stores, etc, and to perform the necessary insertions to
do this. There are no changes to the functions themselves except
reformatting, and one case where memdep was informed of a removed load
(which was pushed into the caller).
Reviewers: davide
Subscribers: mgorny, llvm-commits, Prazek
Differential Revision: https://reviews.llvm.org/D30478
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297438
91177308-0d34-0410-b5e6-
96231b3b80d8
Dehao Chen [Fri, 10 Mar 2017 01:44:37 +0000 (01:44 +0000)]
Do not use branch metadata to check if a basic block is hot.
Summary: We should not use that to check basic block hotness as optimization may mess it up.
Reviewers: eraman
Reviewed By: eraman
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30800
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297437
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Fri, 10 Mar 2017 00:55:29 +0000 (00:55 +0000)]
[x86] add tests for vec div/rem with 0 element in divisor; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297433
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 10 Mar 2017 00:47:35 +0000 (00:47 +0000)]
PatternMatch; Add m_ZExtOrSExt matcher
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297432
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Fri, 10 Mar 2017 00:32:33 +0000 (00:32 +0000)]
NewGVN: Rewrite DCE during elimination so we do it as well as old GVN did.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297428
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Fri, 10 Mar 2017 00:32:26 +0000 (00:32 +0000)]
NewGVN: Rename a few things for clarity
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297427
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Fri, 10 Mar 2017 00:25:44 +0000 (00:25 +0000)]
[GlobalISel] Use ImmutableCallSite instead of templates. NFC.
ImmutableCallSite abstracts away CallInst and InvokeInst. Use it!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297426
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Fri, 10 Mar 2017 00:25:35 +0000 (00:25 +0000)]
[GlobalISel] Fallback when failing to translate invoke.
We unintentionally stopped falling back in r293670.
While there, change an unusual construct.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297425
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Berlin [Fri, 10 Mar 2017 00:25:26 +0000 (00:25 +0000)]
Add support for DenseMap/DenseSet count and find using const pointers
Summary:
Similar to SmallPtrSet, this makes find and count work with both const
referneces and const pointers.
Reviewers: dblaikie
Subscribers: llvm-commits, mzolotukhin
Differential Revision: https://reviews.llvm.org/D30713
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297424
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Thu, 9 Mar 2017 23:48:58 +0000 (23:48 +0000)]
[Unittests] Fix a build failure with clang 3.8. NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297423
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 9 Mar 2017 23:36:26 +0000 (23:36 +0000)]
GlobalISel: support trivial inlineasm calls.
They're used for nefarious purposes by ObjC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297422
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Friedman [Thu, 9 Mar 2017 23:33:36 +0000 (23:33 +0000)]
Refactor alias check from MISched into common helper. NFC.
Differential Revision: https://reviews.llvm.org/D30598
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297421
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 9 Mar 2017 23:08:21 +0000 (23:08 +0000)]
[WebAssembly] Fix the opcode numbers for floating-point le and gt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297420
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Thu, 9 Mar 2017 22:47:00 +0000 (22:47 +0000)]
[DAGCombiner] Do various combine on uaddo.
Summary: This essentially does the same transform as for ADC.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30417
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297416
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Thu, 9 Mar 2017 22:02:14 +0000 (22:02 +0000)]
[Hexagon] Fixes to the bitsplit generation
- Fix the insertion point, which occasionally could have been incorrect.
- Avoid creating multiple bitsplits with the same operands, if an old one
could be reused.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297414
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 9 Mar 2017 22:00:39 +0000 (22:00 +0000)]
GlobalISel: inform FrameLowering when we emit a function call.
Amongst other things (I expect) this is necessary to ensure decent backtraces
when an "unreachable" is involved.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297413
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 9 Mar 2017 21:56:03 +0000 (21:56 +0000)]
[InstSimplify] allow folds for bool vector div/rem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297411
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Thu, 9 Mar 2017 21:12:06 +0000 (21:12 +0000)]
GlobalISel: put debug info for static allocas in the MachineFunction.
The good reason to do this is that static allocas are pretty simple to handle
(especially at -O0) and avoiding tracking DBG_VALUEs throughout the pipeline
should give some kind of performance benefit.
The bad reason is that the debug pipeline is an unholy mess of implicit
contracts, where determining whether "DBG_VALUE %reg, imm" actually implies a
load or not involves the services of at least 3 soothsayers and the sacrifice
of at least one chicken. And it still gets it wrong if the variable is at SP
directly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297410
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 9 Mar 2017 20:42:30 +0000 (20:42 +0000)]
[ConstantFold] vector div/rem with any zero element in divisor is undef
Follow-up for:
https://reviews.llvm.org/D30665
https://reviews.llvm.org/rL297390
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297409
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 9 Mar 2017 20:34:27 +0000 (20:34 +0000)]
AMDGPU: Support for SimplifyDemandedVectorElts for load intrinsics
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297408
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 9 Mar 2017 20:31:20 +0000 (20:31 +0000)]
[InstSimplify] add tests for vector constant folding div/rem-by-0; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297407
91177308-0d34-0410-b5e6-
96231b3b80d8
Yaxun Liu [Thu, 9 Mar 2017 20:04:50 +0000 (20:04 +0000)]
AMDGPU: Add GCCBuiltin for ds_permute ds_bpermute
Differential Revision: https://reviews.llvm.org/D30580
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297406
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Thu, 9 Mar 2017 19:28:00 +0000 (19:28 +0000)]
[DAGCombiner] Do various combine on usubo.
Summary: This essentially does the same transform as for SUBC.
Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30437
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297404
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Thu, 9 Mar 2017 19:24:07 +0000 (19:24 +0000)]
CMake: Don't install llvm-tblgen twice
Summary:
The add_tablegen macros defines its own install target, and it was also calling
add_llvm_utility which adds another install target.
Configuring with -DLLVM_TOOLS_INSTALL_DIR set to something other than
'bin' along with -DLLVM_INSTALL_UTILS=ON was causing llvm-tablgen
to be installed to two separate directories.
Reviewers: beanz, hans
Reviewed By: beanz
Subscribers: llvm-commits, mgorny
Differential Revision: https://reviews.llvm.org/D30656
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297403
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Thu, 9 Mar 2017 19:14:23 +0000 (19:14 +0000)]
[Hexagon] Refactor the DAG preprocessing code, NFC
Extract individual transformations into their own functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297401
91177308-0d34-0410-b5e6-
96231b3b80d8
Rong Xu [Thu, 9 Mar 2017 19:08:55 +0000 (19:08 +0000)]
Minor format change. nfc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297400
91177308-0d34-0410-b5e6-
96231b3b80d8
Rong Xu [Thu, 9 Mar 2017 19:03:57 +0000 (19:03 +0000)]
[PGO] Refactor profile dumping function for ease of adding other profile kind
Refactor the dumping function so that we can add other value profile kind easily.
Differential Revision: https://reviews.llvm.org/D30752
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297399
91177308-0d34-0410-b5e6-
96231b3b80d8
Jan Sjodin [Thu, 9 Mar 2017 18:25:07 +0000 (18:25 +0000)]
Add front() method to SetVector.
Differential Revision: https://reviews.llvm.org/D27262
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297398
91177308-0d34-0410-b5e6-
96231b3b80d8
Artem Belevich [Thu, 9 Mar 2017 17:59:04 +0000 (17:59 +0000)]
[FileCheck] Added --enable-var-scope option to enable scope for regex variables.
If `--enable-var-scope` is in effect, variables with names that
start with `$` are considered to be global. All other variables are
local. All local variables get undefined at the beginning of each
CHECK-LABEL block. Global variables are not affected by CHECK-LABEL.
This makes it easier to ensure that individual tests are not affected
by variables set in preceding tests.
Differential Revision: https://reviews.llvm.org/D30749
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297396
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Thu, 9 Mar 2017 17:05:11 +0000 (17:05 +0000)]
[Hexagon] Add -mhvx option to the Hexagon backend
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297393
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Thu, 9 Mar 2017 16:29:30 +0000 (16:29 +0000)]
[Hexagon] Propagate zext of i1 into arithmetic code in selection DAG
(op ... (zext i1 c) ...) -> (select c (op ... 1 ...),
(op ... 0 ...))
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297391
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 9 Mar 2017 16:20:52 +0000 (16:20 +0000)]
[InstSimplify] vector div/rem with any zero element in divisor is undef
This was suggested as a DAG simplification in the review for rL297026 :
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-
20170306/435253.html
...but let's start with IR since we have actual docs for IR (LangRef).
Differential Revision:
https://reviews.llvm.org/D30665
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297390
91177308-0d34-0410-b5e6-
96231b3b80d8
Sam Parker [Thu, 9 Mar 2017 15:14:32 +0000 (15:14 +0000)]
[ARM] Remove t2xtpk feature from tests
I previously removed the T2XtPk feature from the ARM backend, but it
looks like I missed some of the tests that were using the feature.
Differential Revision: https://reviews.llvm.org/D30778
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297386
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 9 Mar 2017 15:02:25 +0000 (15:02 +0000)]
[DAG] recognize div/rem by 0 as undef before trying constant folding
As discussed in the review thread for rL297026, this is actually 2 changes that
would independently fix all of the test cases in the patch:
1. Return undef in FoldConstantArithmetic for div/rem by 0.
2. Move basic undef simplifications for div/rem (simplifyDivRem()) before
foldBinopIntoSelect() as a matter of efficiency.
I will handle the case of vectors with any zero element as a follow-up. That change
is the DAG sibling for D30665 + adding a check of vector elements to FoldConstantVectorArithmetic().
I'm deleting the test for PR30693 because it does not test for the actual bug any more
(dangers of using bugpoint).
Differential Revision:
https://reviews.llvm.org/D30741
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297384
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Thu, 9 Mar 2017 14:06:39 +0000 (14:06 +0000)]
[X86][SSE] Speed up constant pool shuffle mask decoding with direct copy (PR32037).
If the constants are already the correct size, we can copy them directly into the shuffle mask.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297381
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Dardis [Thu, 9 Mar 2017 14:03:26 +0000 (14:03 +0000)]
[mips] Revert fixes for PR32020.
The fix introduces segfaults and clobbers the value to be stored when
the atomic sequence loops.
Revert "[Target/MIPS] Kill dead code, no functional change intended."
This reverts commit r296153.
Revert "Recommit "[mips] Fix atomic compare and swap at O0.""
This reverts commit r296134.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297380
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Thu, 9 Mar 2017 13:57:04 +0000 (13:57 +0000)]
Fixed typos in comments. NFCI.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297379
91177308-0d34-0410-b5e6-
96231b3b80d8
Nuno Lopes [Thu, 9 Mar 2017 13:43:31 +0000 (13:43 +0000)]
fix build on Cygwin
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297378
91177308-0d34-0410-b5e6-
96231b3b80d8
Joey Gouly [Thu, 9 Mar 2017 13:38:06 +0000 (13:38 +0000)]
[SelectionDAG] Make SelectCode return void
SelectCode has been returning nullptr since
182dac0 ("SDAG: Make
SelectCodeCommon return void", 2016-05-10). Make SelectCode also
return void instead, as all callers have been updated.
Patch by Sven van Haastregt.
Review: https://reviews.llvm.org/D30497
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297377
91177308-0d34-0410-b5e6-
96231b3b80d8
Sjoerd Meijer [Thu, 9 Mar 2017 13:28:37 +0000 (13:28 +0000)]
[ARM] remove FIXMEs and add vcmp MC test
Minor cleanup in ARMInstrVFP.td: removed some FIXMEs and added a MC test for
vcmp that was actually missing.
Differential Revision: https://reviews.llvm.org/D30745
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297376
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 9 Mar 2017 11:35:40 +0000 (11:35 +0000)]
[PM/Inliner] Make the new PM's inliner process call edges across an
entire SCC before iterating on newly-introduced call edges resulting
from any inlined function bodies.
This more closely matches the behavior of the old PM's inliner. While it
wasn't really clear to me initially, this behavior is actually essential
to the inliner behaving reasonably in its current design.
Because the inliner is fundamentally a bottom-up inliner and all of its
cost modeling is designed around that it often runs into trouble within
an SCC where we don't have any meaningful bottom-up ordering to use. In
addition to potentially cyclic, infinite inlining that we block with the
inline history mechanism, it can also take seemingly simple call graph
patterns within an SCC and turn them into *insanely* large functions by
accidentally working top-down across the SCC without any of the
threshold limitations that traditional top-down inliners use.
Consider this diabolical monster.cpp file that Richard Smith came up
with to help demonstrate this issue:
```
template <int N> extern const char *str;
void g(const char *);
template <bool K, int N> void f(bool *B, bool *E) {
if (K)
g(str<N>);
if (B == E)
return;
if (*B)
f<true, N + 1>(B + 1, E);
else
f<false, N + 1>(B + 1, E);
}
template <> void f<false, MAX>(bool *B, bool *E) { return f<false, 0>(B, E); }
template <> void f<true, MAX>(bool *B, bool *E) { return f<true, 0>(B, E); }
extern bool *arr, *end;
void test() { f<false, 0>(arr, end); }
```
When compiled with '-DMAX=N' for various values of N, this will create an SCC
with a reasonably large number of functions. Previously, the inliner would try
to exhaust the inlining candidates in a single function before moving on. This,
unfortunately, turns it into a top-down inliner within the SCC. Because our
thresholds were never built for that, we will incrementally decide that it is
always worth inlining and proceed to flatten the entire SCC into that one
function.
What's worse, we'll then proceed to the next function, and do the exact same
thing except we'll skip the first function, and so on. And at each step, we'll
also make some of the constant factors larger, which is awesome.
The fix in this patch is the obvious one which makes the new PM's inliner use
the same technique used by the old PM: consider all the call edges across the
entire SCC before beginning to process call edges introduced by inlining. The
result of this is essentially to distribute the inlining across the SCC so that
every function incrementally grows toward the inline thresholds rather than
allowing the inliner to grow one of the functions vastly beyond the threshold.
The code for this is a bit awkward, but it works out OK.
We could consider in the future doing something more powerful here such as
prioritized order (via lowest cost and/or profile info) and/or a code-growth
budget per SCC. However, both of those would require really substantial work
both to design the system in a way that wouldn't break really useful
abstraction decomposition properties of the current inliner and to be tuned
across a reasonably diverse set of code and workloads. It also seems really
risky in many ways. I have only found a single real-world file that triggers
the bad behavior here and it is generated code that has a pretty pathological
pattern. I'm not worried about the inliner not doing an *awesome* job here as
long as it does *ok*. On the other hand, the cases that will be tricky to get
right in a prioritized scheme with a budget will be more common and idiomatic
for at least some frontends (C++ and Rust at least). So while these approaches
are still really interesting, I'm not in a huge rush to go after them. Staying
even closer to the existing PM's behavior, especially when this easy to do,
seems like the right short to medium term approach.
I don't really have a test case that makes sense yet... I'll try to find a
variant of the IR produced by the monster template metaprogram that is both
small enough to be sane and large enough to clearly show when we get this wrong
in the future. But I'm not confident this exists. And the behavior change here
*should* be unobservable without snooping on debug logging. So there isn't
really much to test.
The test case updates come from two incidental changes:
1) We now visit functions in an SCC in the opposite order. I don't think there
really is a "right" order here, so I just update the test cases.
2) We no longer compute some analyses when an SCC has no call instructions that
we consider for inlining.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@297374
91177308-0d34-0410-b5e6-
96231b3b80d8