maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/lib/CodeGen/TargetLoweringBase.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

2181 lines
82 KiB
C++
Raw Normal View History

[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
//===- TargetLoweringBase.cpp - Implement the TargetLoweringBase class ----===//
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
//
Update the file headers across all of the LLVM projects in the monorepo to reflect the new license. We understand that people may be surprised that we're moving the header entirely to discuss the new license. We checked this carefully with the Foundation's lawyer and we believe this is the correct approach. Essentially, all code in the project is now made available by the LLVM project under our new license, so you will see that the license headers include that license only. Some of our contributors have contributed code under our old license, and accordingly, we have retained a copy of our old license notice in the top-level files in each project and repository. llvm-svn: 351636
2019-01-19 16:50:56 +08:00
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
//
//===----------------------------------------------------------------------===//
//
// This implements the TargetLoweringBase class.
//
//===----------------------------------------------------------------------===//
#include "llvm/ADT/BitVector.h"
#include "llvm/ADT/STLExtras.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/ADT/SmallVector.h"
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
#include "llvm/ADT/StringExtras.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/ADT/StringRef.h"
enable SDISel sincos optimization for GNU environments - add sincos to runtime library if target triple environment is GNU - added canCombineSinCosLibcall() which checks that sincos is in the RTL and if the environment is GNU then unsafe fpmath is enabled (required to preserve errno) - extended sincos-opt lit test Reviewed by: Hal Finkel llvm-svn: 175283
2013-02-16 02:45:18 +08:00
#include "llvm/ADT/Triple.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/ADT/Twine.h"
GlobalISel: Preserve load/store metadata in IRTranslator This was dropping the invariant metadata on dead argument loads, so they weren't deleted. Atomics still need to be fixed the same way. Also, apparently store was never preserving dereferencable which should also be fixed.
2020-01-13 03:10:42 +08:00
#include "llvm/Analysis/Loads.h"
GlobalISel: Move Localizer::shouldLocalize(..) to TargetLowering Add a new target hook for shouldLocalize so that targets can customize the logic. https://reviews.llvm.org/D75207
2020-03-03 01:15:40 +08:00
#include "llvm/Analysis/TargetTransformInfo.h"
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
#include "llvm/CodeGen/Analysis.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/CodeGen/ISDOpcodes.h"
#include "llvm/CodeGen/MachineBasicBlock.h"
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineFunction.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/CodeGen/MachineInstr.h"
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
#include "llvm/CodeGen/MachineInstrBuilder.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/CodeGen/MachineMemOperand.h"
#include "llvm/CodeGen/MachineOperand.h"
TargetLowering: Add finalizeLowering() function; NFC Adds a new method finalizeLowering to TargetLoweringBase. This is in preparation for an upcoming commit. This function is meant for target specific adjustments to MachineFrameInfo or register reservations. Move the freezeRegisters() and the hasCopyImplyingStackAdjustment() handling into the new function to prove the concept. As an added bonus GlobalISel no longer missed the hasCopyImplyingStackAdjustment() handling with this. Differential Revision: https://reviews.llvm.org/D32621 llvm-svn: 301679
2017-04-29 04:25:05 +08:00
#include "llvm/CodeGen/MachineRegisterInfo.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/CodeGen/RuntimeLibcalls.h"
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
#include "llvm/CodeGen/StackMaps.h"
Fix a bunch more layering of CodeGen headers that are in Target All these headers already depend on CodeGen headers so moving them into CodeGen fixes the layering (since CodeGen depends on Target, not the other way around). llvm-svn: 318490
2017-11-17 09:07:10 +08:00
#include "llvm/CodeGen/TargetLowering.h"
#include "llvm/CodeGen/TargetOpcodes.h"
#include "llvm/CodeGen/TargetRegisterInfo.h"
[IR][CodeGen] Remove dependency on EVT from IR/Function.cpp. Move EVT to CodeGen layer. Currently EVT is in the IR layer only because of Function.cpp needing a very small piece of the functionality of EVT::getEVTString(). The rest of EVT is used in codegen making CodeGen a better place for it. The previous code converted a Type* to EVT and then called getEVTString. This was only expected to handle the primitive types from Type*. Since there only a few primitive types, we can just print them as strings directly. Differential Revision: https://reviews.llvm.org/D45017 llvm-svn: 328806
2018-03-30 01:21:10 +08:00
#include "llvm/CodeGen/ValueTypes.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/IR/Attributes.h"
#include "llvm/IR/CallingConv.h"
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DerivedTypes.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/IR/Function.h"
#include "llvm/IR/GlobalValue.h"
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
#include "llvm/IR/GlobalVariable.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/Type.h"
[CodeGenPrepare] use branch weight metadata to decide if a select should be turned into a branch This is part of solving PR27344: https://llvm.org/bugs/show_bug.cgi?id=27344 CGP should undo the SimplifyCFG transform for the same reason that earlier patches have used this same mechanism: it's possible that passes between SimplifyCFG and CGP may be able to optimize the IR further with a select in place. For the TLI hook default, >99% taken or not taken is chosen as the default threshold for a highly predictable branch. Even the most limited HW branch predictors will be correct on this branch almost all the time, so even a massive mispredict penalty perf loss would be overcome by the win from all the times the branch was predicted correctly. As a follow-up, we could make the default target hook less conservative by using the SchedMachineModel's MispredictPenalty. Or we could just let targets override the default by implementing the hook with that and other target-specific options. Note that trying to statically determine mispredict rates for close-to-balanced profile weight data is generally impossible if the HW is sufficiently advanced. Ie, 50/50 taken/not-taken might still be 100% predictable. Finally, note that this patch as-is will not solve PR27344 because the current __builtin_unpredictable() branch weight default values are 4 and 64. A proposal to change that is in D19435. Differential Revision: http://reviews.llvm.org/D19488 llvm-svn: 267572
2016-04-27 01:11:17 +08:00
#include "llvm/Support/BranchProbability.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/Support/Casting.h"
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
#include "llvm/Support/CommandLine.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include "llvm/Support/Compiler.h"
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
#include "llvm/Support/ErrorHandling.h"
Fix layering of MachineValueType.h by moving it from CodeGen to Support This is used by llvm tblgen as well as by LLVM Targets, so the only common place is Support for now. (maybe we need another target for these sorts of things - but for now I'm at least making them correct & we can make them better if/when people have strong feelings) llvm-svn: 328395
2018-03-24 07:58:25 +08:00
#include "llvm/Support/MachineValueType.h"
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
#include "llvm/Support/MathExtras.h"
#include "llvm/Target/TargetMachine.h"
TargetLowering.h - remove unnecessary includes. NFC. Replace with forward declarations and move SizeOpts.h down to TargetLoweringBase.cpp
2020-05-22 21:22:56 +08:00
#include "llvm/Transforms/Utils/SizeOpts.h"
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include <algorithm>
#include <cassert>
#include <cstddef>
#include <cstdint>
Fix a bunch more layering of CodeGen headers that are in Target All these headers already depend on CodeGen headers so moving them into CodeGen fixes the layering (since CodeGen depends on Target, not the other way around). llvm-svn: 318490
2017-11-17 09:07:10 +08:00
#include <cstring>
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
#include <iterator>
#include <string>
#include <tuple>
#include <utility>
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
using namespace llvm;
add a cl::opt override for TargetLoweringBase's JumpIsExpensive This patch is not intended to change existing codegen behavior for any target. It just exposes the JumpIsExpensive setting on the command-line to allow for easier testing and emergency overrides. Also, change the existing regression test to use FileCheck, explicitly specify the jump-is-expensive option, and use more precise checks. Differential Revision: http://reviews.llvm.org/D10846 llvm-svn: 241179
2015-07-02 02:10:20 +08:00
static cl::opt<bool> JumpIsExpensiveOverride(
"jump-is-expensive", cl::init(false),
cl::desc("Do not create extra branches to split comparison logic."),
cl::Hidden);
Add option to specify minimum number of entries for jump tables Add an option to allow easier experimentation by target maintainers with the minimum number of entries to create jump tables. Also clarify the name of the other existing option governing the creation of jump tables. Differential revision: https://reviews.llvm.org/D25883 llvm-svn: 285104
2016-10-26 03:53:51 +08:00
static cl::opt<unsigned> MinimumJumpTableEntries
("min-jump-table-entries", cl::init(4), cl::Hidden,
cl::desc("Set minimum number of entries to use a jump table."));
Revert r372893 "[CodeGen] Replace -max-jump-table-size with -max-jump-table-targets" This caused severe compile-time regressions, see PR43455. > Modern processors predict the targets of an indirect branch regardless of > the size of any jump table used to glean its target address. Moreover, > branch predictors typically use resources limited by the number of actual > targets that occur at run time. > > This patch changes the semantics of the option `-max-jump-table-size` to limit > the number of different targets instead of the number of entries in a jump > table. Thus, it is now renamed to `-max-jump-table-targets`. > > Before, when `-max-jump-table-size` was specified, it could happen that > cluster jump tables could have targets used repeatedly, but each one was > counted and typically resulted in tables with the same number of entries. > With this patch, when specifying `-max-jump-table-targets`, tables may have > different lengths, since the number of unique targets is counted towards the > limit, but the number of unique targets in tables is the same, but for the > last one containing the balance of targets. > > Differential revision: https://reviews.llvm.org/D60295 llvm-svn: 373060
2019-09-27 17:54:26 +08:00
static cl::opt<unsigned> MaximumJumpTableSize
("max-jump-table-size", cl::init(UINT_MAX), cl::Hidden,
cl::desc("Set maximum size of jump tables."));
Add support to optionally limit the size of jump tables. Many high-performance processors have a dedicated branch predictor for indirect branches, commonly used with jump tables. As sophisticated as such branch predictors are, they tend to have well defined limits beyond which their effectiveness is hampered or even nullified. One such limit is the number of possible destinations for a given indirect branches that such branch predictors can handle. This patch considers a limit that a target may set to the number of destination addresses in a jump table. Patch by: Evandro Menezes <e.menezes@samsung.com>, Aditya Kumar <aditya.k7@samsung.com>, Sebastian Pop <s.pop@samsung.com>. Differential revision: https://reviews.llvm.org/D21940 llvm-svn: 282412
2016-09-26 23:32:33 +08:00
[InlineCost] Improve the cost heuristic for Switch Summary: The motivation example is like below which has 13 cases but only 2 distinct targets ``` lor.lhs.false2: ; preds = %if.then switch i32 %Status, label %if.then27 [ i32 -7012, label %if.end35 i32 -10008, label %if.end35 i32 -10016, label %if.end35 i32 15000, label %if.end35 i32 14013, label %if.end35 i32 10114, label %if.end35 i32 10107, label %if.end35 i32 10105, label %if.end35 i32 10013, label %if.end35 i32 10011, label %if.end35 i32 7008, label %if.end35 i32 7007, label %if.end35 i32 5002, label %if.end35 ] ``` which is compiled into a balanced binary tree like this on AArch64 (similar on X86) ``` .LBB853_9: // %lor.lhs.false2 mov w8, #10012 cmp w19, w8 b.gt .LBB853_14 // BB#10: // %lor.lhs.false2 mov w8, #5001 cmp w19, w8 b.gt .LBB853_18 // BB#11: // %lor.lhs.false2 mov w8, #-10016 cmp w19, w8 b.eq .LBB853_23 // BB#12: // %lor.lhs.false2 mov w8, #-10008 cmp w19, w8 b.eq .LBB853_23 // BB#13: // %lor.lhs.false2 mov w8, #-7012 cmp w19, w8 b.eq .LBB853_23 b .LBB853_3 .LBB853_14: // %lor.lhs.false2 mov w8, #14012 cmp w19, w8 b.gt .LBB853_21 // BB#15: // %lor.lhs.false2 mov w8, #-10105 add w8, w19, w8 cmp w8, #9 // =9 b.hi .LBB853_17 // BB#16: // %lor.lhs.false2 orr w9, wzr, #0x1 lsl w8, w9, w8 mov w9, #517 and w8, w8, w9 cbnz w8, .LBB853_23 .LBB853_17: // %lor.lhs.false2 mov w8, #10013 cmp w19, w8 b.eq .LBB853_23 b .LBB853_3 .LBB853_18: // %lor.lhs.false2 mov w8, #-7007 add w8, w19, w8 cmp w8, #2 // =2 b.lo .LBB853_23 // BB#19: // %lor.lhs.false2 mov w8, #5002 cmp w19, w8 b.eq .LBB853_23 // BB#20: // %lor.lhs.false2 mov w8, #10011 cmp w19, w8 b.eq .LBB853_23 b .LBB853_3 .LBB853_21: // %lor.lhs.false2 mov w8, #14013 cmp w19, w8 b.eq .LBB853_23 // BB#22: // %lor.lhs.false2 mov w8, #15000 cmp w19, w8 b.ne .LBB853_3 ``` However, the inline cost model estimates the cost to be linear with the number of distinct targets and the cost of the above switch is just 2 InstrCosts. The function containing this switch is then inlined about 900 times. This change use the general way of switch lowering for the inline heuristic. It etimate the number of case clusters with the suitability check for a jump table or bit test. Considering the binary search tree built for the clusters, this change modifies the model to be linear with the size of the balanced binary tree. The model is off by default for now : -inline-generic-switch-cost=false This change was originally proposed by Haicheng in D29870. Reviewers: hans, bmakam, chandlerc, eraman, haicheng, mcrosier Reviewed By: hans Subscribers: joerg, aemerson, llvm-commits, rengolin Differential Revision: https://reviews.llvm.org/D31085 llvm-svn: 301649
2017-04-29 00:04:03 +08:00
/// Minimum jump table density for normal functions.
static cl::opt<unsigned>
JumpTableDensity("jump-table-density", cl::init(10), cl::Hidden,
cl::desc("Minimum density for building a jump table in "
"a normal function"));
/// Minimum jump table density for -Os or -Oz functions.
static cl::opt<unsigned> OptsizeJumpTableDensity(
"optsize-jump-table-density", cl::init(40), cl::Hidden,
cl::desc("Minimum density for building a jump table in "
"an optsize function"));
[FPEnv] Add an option to disable strict float node mutating to an normal float node This patch add an option 'disable-strictnode-mutation' to prevent strict node mutating to an normal node. So we can make sure that the patch which sets strict-node as legal works correctly. Patch by Chen Liu(LiuChen3) Differential Revision: https://reviews.llvm.org/D70226
2019-11-22 09:17:19 +08:00
// FIXME: This option is only to test if the strict fp operation processed
// correctly by preventing mutating strict fp operation to normal fp operation
// during development. When the backend supports strict float operation, this
// option will be meaningless.
static cl::opt<bool> DisableStrictNodeMutation("disable-strictnode-mutation",
cl::desc("Don't mutate strict-float node to a legalize node"),
cl::init(false), cl::Hidden);
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
static bool darwinHasSinCos(const Triple &TT) {
assert(TT.isOSDarwin() && "should be called with darwin triple");
TargetLoweringBase: Fix darwinHasSinCos() Another followup to my refactoring in r321036: Turns out we can end up with an x86 darwin target that is not macos (simulator triples can look like i386-apple-ios) so we need the x86/32bit check in all cases. llvm-svn: 321104
2017-12-20 04:24:12 +08:00
// Don't bother with 32 bit x86.
if (TT.getArch() == Triple::x86)
return false;
// Macos < 10.9 has no sincos_stret.
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
if (TT.isMacOSX())
return !TT.isMacOSXVersionLT(10, 9) && TT.isArch64Bit();
TargetLoweringBase: Remove unnecessary watchos exception; NFC WatchOS isn't report as iOS (as opposed to tvos) so the exception I added in my last commit wasn't necessary after all. llvm-svn: 321041
2017-12-19 07:33:28 +08:00
// iOS < 7.0 has no sincos_stret.
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
if (TT.isiOS())
TargetLoweringBase: Remove unnecessary watchos exception; NFC WatchOS isn't report as iOS (as opposed to tvos) so the exception I added in my last commit wasn't necessary after all. llvm-svn: 321041
2017-12-19 07:33:28 +08:00
return !TT.isOSVersionLT(7, 0);
// Any other darwin such as WatchOS/TvOS is new enough.
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
return true;
}
[CodeGenPrepare] use branch weight metadata to decide if a select should be turned into a branch This is part of solving PR27344: https://llvm.org/bugs/show_bug.cgi?id=27344 CGP should undo the SimplifyCFG transform for the same reason that earlier patches have used this same mechanism: it's possible that passes between SimplifyCFG and CGP may be able to optimize the IR further with a select in place. For the TLI hook default, >99% taken or not taken is chosen as the default threshold for a highly predictable branch. Even the most limited HW branch predictors will be correct on this branch almost all the time, so even a massive mispredict penalty perf loss would be overcome by the win from all the times the branch was predicted correctly. As a follow-up, we could make the default target hook less conservative by using the SchedMachineModel's MispredictPenalty. Or we could just let targets override the default by implementing the hook with that and other target-specific options. Note that trying to statically determine mispredict rates for close-to-balanced profile weight data is generally impossible if the HW is sufficiently advanced. Ie, 50/50 taken/not-taken might still be 100% predictable. Finally, note that this patch as-is will not solve PR27344 because the current __builtin_unpredictable() branch weight default values are 4 and 64. A proposal to change that is in D19435. Differential Revision: http://reviews.llvm.org/D19488 llvm-svn: 267572
2016-04-27 01:11:17 +08:00
// Although this default value is arbitrary, it is not random. It is assumed
// that a condition that evaluates the same way by a higher percentage than this
// is best represented as control flow. Therefore, the default value N should be
// set such that the win from N% correct executions is greater than the loss
// from (100 - N)% mispredicted executions for the majority of intended targets.
static cl::opt<int> MinPercentageForPredictableBranch(
"min-predictable-branch", cl::init(99),
cl::desc("Minimum percentage (0-100) that a condition must be either true "
"or false to assume that the condition is predictable"),
cl::Hidden);
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
void TargetLoweringBase::InitLibcalls(const Triple &TT) {
Move Runtime libcall definitions to a .def file This will allow eliminating the duplication of the names, and allow adding extra information such as signatures in a future commit. Differential Revision: https://reviews.llvm.org/D35522 llvm-svn: 308531
2017-07-20 05:53:30 +08:00
#define HANDLE_LIBCALL(code, name) \
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
setLibcallName(RTLIB::code, name);
Put "built-in" function definitions in global Used list, for LTO. (fix bug 34169) When building with LTO, builtin functions that are defined but whose calls have not been inserted yet, get internalized. The Global Dead Code Elimination phase in the new LTO implementation then removes these function definitions. Later optimizations add calls to those functions, and the linker then dies complaining that there are no definitions. This CL fixes the new LTO implementation to check if a function is builtin, and if so, to not internalize (and later DCE) the function. As part of this fix I needed to move the RuntimeLibcalls.{def,h} files from the CodeGen subidrectory to the IR subdirectory. I have updated all the files that accessed those two files to access their new location. Fixes PR34169 Patch by Caroline Tice! Differential Revision: https://reviews.llvm.org/D49434 llvm-svn: 337847
2018-07-25 03:34:37 +08:00
#include "llvm/IR/RuntimeLibcalls.def"
Move Runtime libcall definitions to a .def file This will allow eliminating the duplication of the names, and allow adding extra information such as signatures in a future commit. Differential Revision: https://reviews.llvm.org/D35522 llvm-svn: 308531
2017-07-20 05:53:30 +08:00
#undef HANDLE_LIBCALL
TargetLowering: Fix InitLibcallCallingConvs() overriding things set in InitLibcalls() I missed the fact that the later called InitLibcallCallingConvs() overrides some things set in InitLibcalls() when I did the refactoring in r321036. Fix by merging InitLibcallCallingConvs() into InitLibcalls() and doing the initialization earlier. llvm-svn: 321045
2017-12-19 08:20:33 +08:00
// Initialize calling conventions to their default.
for (int LC = 0; LC < RTLIB::UNKNOWN_LIBCALL; ++LC)
setLibcallCallingConv((RTLIB::Libcall)LC, CallingConv::C);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
[PowerPC] Support fp128 libcalls On PowerPC, IEEE 754 quadruple-precision libcall names use "kf" instead of "tf". In libgcc, libgcc/config/rs6000/float128-sed converts TF names to KF names. This patch implements its 24 substitution rules. Reviewed By: hfinkel Differential Revision: https://reviews.llvm.org/D64282 llvm-svn: 366039
2019-07-15 13:02:32 +08:00
// For IEEE quad-precision libcall names, PPC uses "kf" instead of "tf".
if (TT.getArch() == Triple::ppc || TT.isPPC64()) {
setLibcallName(RTLIB::ADD_F128, "__addkf3");
setLibcallName(RTLIB::SUB_F128, "__subkf3");
setLibcallName(RTLIB::MUL_F128, "__mulkf3");
setLibcallName(RTLIB::DIV_F128, "__divkf3");
setLibcallName(RTLIB::FPEXT_F32_F128, "__extendsfkf2");
setLibcallName(RTLIB::FPEXT_F64_F128, "__extenddfkf2");
setLibcallName(RTLIB::FPROUND_F128_F32, "__trunckfsf2");
setLibcallName(RTLIB::FPROUND_F128_F64, "__trunckfdf2");
setLibcallName(RTLIB::FPTOSINT_F128_I32, "__fixkfsi");
setLibcallName(RTLIB::FPTOSINT_F128_I64, "__fixkfdi");
setLibcallName(RTLIB::FPTOUINT_F128_I32, "__fixunskfsi");
setLibcallName(RTLIB::FPTOUINT_F128_I64, "__fixunskfdi");
setLibcallName(RTLIB::SINTTOFP_I32_F128, "__floatsikf");
setLibcallName(RTLIB::SINTTOFP_I64_F128, "__floatdikf");
setLibcallName(RTLIB::UINTTOFP_I32_F128, "__floatunsikf");
setLibcallName(RTLIB::UINTTOFP_I64_F128, "__floatundikf");
setLibcallName(RTLIB::OEQ_F128, "__eqkf2");
setLibcallName(RTLIB::UNE_F128, "__nekf2");
setLibcallName(RTLIB::OGE_F128, "__gekf2");
setLibcallName(RTLIB::OLT_F128, "__ltkf2");
setLibcallName(RTLIB::OLE_F128, "__lekf2");
setLibcallName(RTLIB::OGT_F128, "__gtkf2");
setLibcallName(RTLIB::UO_F128, "__unordkf2");
}
Move Runtime libcall definitions to a .def file This will allow eliminating the duplication of the names, and allow adding extra information such as signatures in a future commit. Differential Revision: https://reviews.llvm.org/D35522 llvm-svn: 308531
2017-07-20 05:53:30 +08:00
// A few names are different on particular architectures or environments.
Pre-fill LibcallRoutineNames with nullptr. And rearrange InitLibcallNames slightly. llvm-svn: 266142
2016-04-13 06:32:47 +08:00
if (TT.isOSDarwin()) {
// For f16/f32 conversions, Darwin uses the standard naming scheme, instead
// of the gnueabi-style __gnu_*_ieee.
// FIXME: What about other targets?
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
setLibcallName(RTLIB::FPEXT_F16_F32, "__extendhfsf2");
setLibcallName(RTLIB::FPROUND_F32_F16, "__truncsfhf2");
AArch64/X86: Factor out common bzero logic; NFC llvm-svn: 321035
2017-12-19 07:14:28 +08:00
TargetLoweringBase: The ios simulator has no bzero function. Make sure I really get back to the beahvior before my rewrite in r321035 which turned out not to be completely NFC as I changed the behavior for the ios simulator environment. llvm-svn: 322223
2018-01-11 04:49:57 +08:00
// Some darwins have an optimized __bzero/bzero function.
switch (TT.getArch()) {
case Triple::x86:
case Triple::x86_64:
if (TT.isMacOSX() && !TT.isMacOSXVersionLT(10, 6))
setLibcallName(RTLIB::BZERO, "__bzero");
break;
case Triple::aarch64:
AArch64: support arm64_32, an ILP32 slice for watchOS. This is the main CodeGen patch to support the arm64_32 watchOS ABI in LLVM. FastISel is mostly disabled for now since it would generate incorrect code for ILP32. llvm-svn: 371722
2019-09-12 18:22:23 +08:00
case Triple::aarch64_32:
TargetLoweringBase: The ios simulator has no bzero function. Make sure I really get back to the beahvior before my rewrite in r321035 which turned out not to be completely NFC as I changed the behavior for the ios simulator environment. llvm-svn: 322223
2018-01-11 04:49:57 +08:00
setLibcallName(RTLIB::BZERO, "bzero");
break;
default:
break;
TargetLoweringBase: Followup to r321035 I missed some prefixes and the fact that on AArch64 we use "bzero" instead of "__bzero" as on X86 when doing my refactoring in r321035. Improve tests for bzero. llvm-svn: 321046
2017-12-19 08:43:00 +08:00
}
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
if (darwinHasSinCos(TT)) {
setLibcallName(RTLIB::SINCOS_STRET_F32, "__sincosf_stret");
setLibcallName(RTLIB::SINCOS_STRET_F64, "__sincos_stret");
if (TT.isWatchABI()) {
setLibcallCallingConv(RTLIB::SINCOS_STRET_F32,
CallingConv::ARM_AAPCS_VFP);
setLibcallCallingConv(RTLIB::SINCOS_STRET_F64,
CallingConv::ARM_AAPCS_VFP);
}
}
Pre-fill LibcallRoutineNames with nullptr. And rearrange InitLibcallNames slightly. llvm-svn: 266142
2016-04-13 06:32:47 +08:00
} else {
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
setLibcallName(RTLIB::FPEXT_F16_F32, "__gnu_h2f_ieee");
setLibcallName(RTLIB::FPROUND_F32_F16, "__gnu_f2h_ieee");
Pre-fill LibcallRoutineNames with nullptr. And rearrange InitLibcallNames slightly. llvm-svn: 266142
2016-04-13 06:32:47 +08:00
}
Add __atomic_* lowering to AtomicExpandPass. (Recommit of r266002, with r266011, r266016, and not accidentally including an extra unused/uninitialized element in LibcallRoutineNames) AtomicExpandPass can now lower atomic load, atomic store, atomicrmw, and cmpxchg instructions to __atomic_* library calls, when the target doesn't support atomics of a given size. This is the first step towards moving all atomic lowering from clang into llvm. When all is done, the behavior of __sync_* builtins, __atomic_* builtins, and C11 atomics will be unified. Previously LLVM would pass everything through to the ISelLowering code. There, unsupported atomic instructions would turn into __sync_* library calls. Because of that behavior, Clang currently avoids emitting llvm IR atomic instructions when this would happen, and emits __atomic_* library functions itself, in the frontend. This change makes LLVM able to emit __atomic_* libcalls, and thus will eventually allow clang to depend on LLVM to do the right thing. It is advantageous to do the new lowering to atomic libcalls in AtomicExpandPass, before ISel time, because it's important that all atomic operations for a given size either lower to __atomic_* libcalls (which may use locks), or native instructions which won't. No mixing and matching. At the moment, this code is enabled only for SPARC, as a demonstration. The next commit will expand support to all of the other targets. Differential Revision: http://reviews.llvm.org/D18200 llvm-svn: 266115
2016-04-13 04:18:48 +08:00
[TargetLowering] Android has sincos functions Since Android API version 9 the Android libm has had the sincos functions, so they should be recognised as libcalls and sincos optimisation should be applied. Differential Revision: https://reviews.llvm.org/D52025 llvm-svn: 342471
2018-09-18 21:18:21 +08:00
if (TT.isGNUEnvironment() || TT.isOSFuchsia() ||
(TT.isAndroid() && !TT.isAndroidVersionLT(9))) {
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
setLibcallName(RTLIB::SINCOS_F32, "sincosf");
setLibcallName(RTLIB::SINCOS_F64, "sincos");
setLibcallName(RTLIB::SINCOS_F80, "sincosl");
setLibcallName(RTLIB::SINCOS_F128, "sincosl");
setLibcallName(RTLIB::SINCOS_PPCF128, "sincosl");
enable SDISel sincos optimization for GNU environments - add sincos to runtime library if target triple environment is GNU - added canCombineSinCosLibcall() which checks that sincos is in the RTL and if the environment is GNU then unsafe fpmath is enabled (required to preserve errno) - extended sincos-opt lit test Reviewed by: Hal Finkel llvm-svn: 175283
2013-02-16 02:45:18 +08:00
}
[stackprotector] Add in the stackprotector libcall. We support this libcall on all platforms except for OpenBSD (See lib/Codegen/StackProtector.cpp). llvm-svn: 188193
2013-08-13 02:45:38 +08:00
[TargetLowering][PS4] Add sincos(f) lib functions when target is PS4 PS4 supports sincosf and sincos. Adding the library functions enables the sin(f)+cos(f) -> sincos(f) optimization. Differential Revision: https://reviews.llvm.org/D67009 llvm-svn: 370675
2019-09-03 00:53:32 +08:00
if (TT.isPS4CPU()) {
setLibcallName(RTLIB::SINCOS_F32, "sincosf");
setLibcallName(RTLIB::SINCOS_F64, "sincos");
}
Move Runtime libcall definitions to a .def file This will allow eliminating the duplication of the names, and allow adding extra information such as signatures in a future commit. Differential Revision: https://reviews.llvm.org/D35522 llvm-svn: 308531
2017-07-20 05:53:30 +08:00
if (TT.isOSOpenBSD()) {
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
setLibcallName(RTLIB::STACKPROTECTOR_CHECK_FAIL, nullptr);
[CodeGen] Use standard -not gnueabi- naming for f16 libcalls on Darwin. Other targets probably should as well. Since r237161, compiler-rt has both, but I don't see why anything other than gnueabi would use a gnueabi naming scheme. llvm-svn: 237324
2015-05-14 09:00:51 +08:00
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
/// getFPEXT - Return the FPEXT_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPEXT(EVT OpVT, EVT RetVT) {
CodeGen: emit IR-level f16 conversion intrinsics as fptrunc/fpext This makes the first stage DAG for @llvm.convert.to.fp16 an fptrunc, and correspondingly @llvm.convert.from.fp16 an fpext. The legalisation path is now uniform, regardless of the input IR: fptrunc -> FP_TO_FP16 (if f16 illegal) -> libcall fpext -> FP16_TO_FP (if f16 illegal) -> libcall Each target should be able to select the version that best matches its operations and not be required to duplicate patterns for both fptrunc and FP_TO_FP16 (for example). As a result we can remove some redundant AArch64 patterns. llvm-svn: 213507
2014-07-21 17:13:56 +08:00
if (OpVT == MVT::f16) {
if (RetVT == MVT::f32)
return FPEXT_F16_F32;
} else if (OpVT == MVT::f32) {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
if (RetVT == MVT::f64)
return FPEXT_F32_F64;
if (RetVT == MVT::f128)
return FPEXT_F32_F128;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
if (RetVT == MVT::ppcf128)
return FPEXT_F32_PPCF128;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
} else if (OpVT == MVT::f64) {
if (RetVT == MVT::f128)
return FPEXT_F64_F128;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
else if (RetVT == MVT::ppcf128)
return FPEXT_F64_PPCF128;
Add support for emitting libcalls for x86_fp80 -> fp128 and vice-versa compiler_rt doesn't provide them (yet), but libgcc does. PR34076. llvm-svn: 322772
2018-01-18 06:29:16 +08:00
} else if (OpVT == MVT::f80) {
if (RetVT == MVT::f128)
return FPEXT_F80_F128;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
return UNKNOWN_LIBCALL;
}
/// getFPROUND - Return the FPROUND_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPROUND(EVT OpVT, EVT RetVT) {
CodeGen: generate single libcall for fptrunc -> f16 operations. Previously we asserted on this code. Currently compiler-rt doesn't actually implement any of these new libcalls, but external help is pretty much the only viable option for LLVM. I've followed the much more generic "__truncST2" naming, as opposed to the odd name for f32 -> f16 truncation. This can obviously be changed later, or overridden by any targets that need to. llvm-svn: 213252
2014-07-17 19:12:12 +08:00
if (RetVT == MVT::f16) {
if (OpVT == MVT::f32)
return FPROUND_F32_F16;
if (OpVT == MVT::f64)
return FPROUND_F64_F16;
if (OpVT == MVT::f80)
return FPROUND_F80_F16;
if (OpVT == MVT::f128)
return FPROUND_F128_F16;
if (OpVT == MVT::ppcf128)
return FPROUND_PPCF128_F16;
} else if (RetVT == MVT::f32) {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
if (OpVT == MVT::f64)
return FPROUND_F64_F32;
if (OpVT == MVT::f80)
return FPROUND_F80_F32;
if (OpVT == MVT::f128)
return FPROUND_F128_F32;
if (OpVT == MVT::ppcf128)
return FPROUND_PPCF128_F32;
} else if (RetVT == MVT::f64) {
if (OpVT == MVT::f80)
return FPROUND_F80_F64;
if (OpVT == MVT::f128)
return FPROUND_F128_F64;
if (OpVT == MVT::ppcf128)
return FPROUND_PPCF128_F64;
Add support for emitting libcalls for x86_fp80 -> fp128 and vice-versa compiler_rt doesn't provide them (yet), but libgcc does. PR34076. llvm-svn: 322772
2018-01-18 06:29:16 +08:00
} else if (RetVT == MVT::f80) {
if (OpVT == MVT::f128)
return FPROUND_F128_F80;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
return UNKNOWN_LIBCALL;
}
/// getFPTOSINT - Return the FPTOSINT_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPTOSINT(EVT OpVT, EVT RetVT) {
if (OpVT == MVT::f32) {
if (RetVT == MVT::i32)
return FPTOSINT_F32_I32;
if (RetVT == MVT::i64)
return FPTOSINT_F32_I64;
if (RetVT == MVT::i128)
return FPTOSINT_F32_I128;
} else if (OpVT == MVT::f64) {
if (RetVT == MVT::i32)
return FPTOSINT_F64_I32;
if (RetVT == MVT::i64)
return FPTOSINT_F64_I64;
if (RetVT == MVT::i128)
return FPTOSINT_F64_I128;
} else if (OpVT == MVT::f80) {
if (RetVT == MVT::i32)
return FPTOSINT_F80_I32;
if (RetVT == MVT::i64)
return FPTOSINT_F80_I64;
if (RetVT == MVT::i128)
return FPTOSINT_F80_I128;
} else if (OpVT == MVT::f128) {
if (RetVT == MVT::i32)
return FPTOSINT_F128_I32;
if (RetVT == MVT::i64)
return FPTOSINT_F128_I64;
if (RetVT == MVT::i128)
return FPTOSINT_F128_I128;
} else if (OpVT == MVT::ppcf128) {
if (RetVT == MVT::i32)
return FPTOSINT_PPCF128_I32;
if (RetVT == MVT::i64)
return FPTOSINT_PPCF128_I64;
if (RetVT == MVT::i128)
return FPTOSINT_PPCF128_I128;
}
return UNKNOWN_LIBCALL;
}
/// getFPTOUINT - Return the FPTOUINT_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getFPTOUINT(EVT OpVT, EVT RetVT) {
if (OpVT == MVT::f32) {
if (RetVT == MVT::i32)
return FPTOUINT_F32_I32;
if (RetVT == MVT::i64)
return FPTOUINT_F32_I64;
if (RetVT == MVT::i128)
return FPTOUINT_F32_I128;
} else if (OpVT == MVT::f64) {
if (RetVT == MVT::i32)
return FPTOUINT_F64_I32;
if (RetVT == MVT::i64)
return FPTOUINT_F64_I64;
if (RetVT == MVT::i128)
return FPTOUINT_F64_I128;
} else if (OpVT == MVT::f80) {
if (RetVT == MVT::i32)
return FPTOUINT_F80_I32;
if (RetVT == MVT::i64)
return FPTOUINT_F80_I64;
if (RetVT == MVT::i128)
return FPTOUINT_F80_I128;
} else if (OpVT == MVT::f128) {
if (RetVT == MVT::i32)
return FPTOUINT_F128_I32;
if (RetVT == MVT::i64)
return FPTOUINT_F128_I64;
if (RetVT == MVT::i128)
return FPTOUINT_F128_I128;
} else if (OpVT == MVT::ppcf128) {
if (RetVT == MVT::i32)
return FPTOUINT_PPCF128_I32;
if (RetVT == MVT::i64)
return FPTOUINT_PPCF128_I64;
if (RetVT == MVT::i128)
return FPTOUINT_PPCF128_I128;
}
return UNKNOWN_LIBCALL;
}
/// getSINTTOFP - Return the SINTTOFP_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getSINTTOFP(EVT OpVT, EVT RetVT) {
if (OpVT == MVT::i32) {
if (RetVT == MVT::f32)
return SINTTOFP_I32_F32;
if (RetVT == MVT::f64)
return SINTTOFP_I32_F64;
if (RetVT == MVT::f80)
return SINTTOFP_I32_F80;
if (RetVT == MVT::f128)
return SINTTOFP_I32_F128;
if (RetVT == MVT::ppcf128)
return SINTTOFP_I32_PPCF128;
} else if (OpVT == MVT::i64) {
if (RetVT == MVT::f32)
return SINTTOFP_I64_F32;
if (RetVT == MVT::f64)
return SINTTOFP_I64_F64;
if (RetVT == MVT::f80)
return SINTTOFP_I64_F80;
if (RetVT == MVT::f128)
return SINTTOFP_I64_F128;
if (RetVT == MVT::ppcf128)
return SINTTOFP_I64_PPCF128;
} else if (OpVT == MVT::i128) {
if (RetVT == MVT::f32)
return SINTTOFP_I128_F32;
if (RetVT == MVT::f64)
return SINTTOFP_I128_F64;
if (RetVT == MVT::f80)
return SINTTOFP_I128_F80;
if (RetVT == MVT::f128)
return SINTTOFP_I128_F128;
if (RetVT == MVT::ppcf128)
return SINTTOFP_I128_PPCF128;
}
return UNKNOWN_LIBCALL;
}
/// getUINTTOFP - Return the UINTTOFP_*_* value for the given types, or
/// UNKNOWN_LIBCALL if there is none.
RTLIB::Libcall RTLIB::getUINTTOFP(EVT OpVT, EVT RetVT) {
if (OpVT == MVT::i32) {
if (RetVT == MVT::f32)
return UINTTOFP_I32_F32;
if (RetVT == MVT::f64)
return UINTTOFP_I32_F64;
if (RetVT == MVT::f80)
return UINTTOFP_I32_F80;
if (RetVT == MVT::f128)
return UINTTOFP_I32_F128;
if (RetVT == MVT::ppcf128)
return UINTTOFP_I32_PPCF128;
} else if (OpVT == MVT::i64) {
if (RetVT == MVT::f32)
return UINTTOFP_I64_F32;
if (RetVT == MVT::f64)
return UINTTOFP_I64_F64;
if (RetVT == MVT::f80)
return UINTTOFP_I64_F80;
if (RetVT == MVT::f128)
return UINTTOFP_I64_F128;
if (RetVT == MVT::ppcf128)
return UINTTOFP_I64_PPCF128;
} else if (OpVT == MVT::i128) {
if (RetVT == MVT::f32)
return UINTTOFP_I128_F32;
if (RetVT == MVT::f64)
return UINTTOFP_I128_F64;
if (RetVT == MVT::f80)
return UINTTOFP_I128_F80;
if (RetVT == MVT::f128)
return UINTTOFP_I128_F128;
if (RetVT == MVT::ppcf128)
return UINTTOFP_I128_PPCF128;
}
return UNKNOWN_LIBCALL;
}
Tweak some atomics functions in preparation for larger changes; NFC. - Rename getATOMIC to getSYNC, as llvm will soon be able to emit both '__sync' libcalls and '__atomic' libcalls, and this function is for the '__sync' ones. - getInsertFencesForAtomic() has been replaced with shouldInsertFencesForAtomic(Instruction), so that the decision can be made per-instruction. This functionality will be used soon. - emitLeadingFence/emitTrailingFence are no longer called if shouldInsertFencesForAtomic returns false, and thus don't need to check the condition themselves. llvm-svn: 263665
2016-03-17 06:12:04 +08:00
RTLIB::Libcall RTLIB::getSYNC(unsigned Opc, MVT VT) {
SDAG: Merge the meat of two ExpandAtomic implementations. The copies already diverged, don't let them become any worse. Reduce redundancy in code with a little macro metaprogramming. llvm-svn: 231401
2015-03-06 04:04:29 +08:00
#define OP_TO_LIBCALL(Name, Enum) \
case Name: \
switch (VT.SimpleTy) { \
default: \
return UNKNOWN_LIBCALL; \
case MVT::i8: \
return Enum##_1; \
case MVT::i16: \
return Enum##_2; \
case MVT::i32: \
return Enum##_4; \
case MVT::i64: \
return Enum##_8; \
case MVT::i128: \
return Enum##_16; \
}
switch (Opc) {
OP_TO_LIBCALL(ISD::ATOMIC_SWAP, SYNC_LOCK_TEST_AND_SET)
OP_TO_LIBCALL(ISD::ATOMIC_CMP_SWAP, SYNC_VAL_COMPARE_AND_SWAP)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_ADD, SYNC_FETCH_AND_ADD)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_SUB, SYNC_FETCH_AND_SUB)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_AND, SYNC_FETCH_AND_AND)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_OR, SYNC_FETCH_AND_OR)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_XOR, SYNC_FETCH_AND_XOR)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_NAND, SYNC_FETCH_AND_NAND)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_MAX, SYNC_FETCH_AND_MAX)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_UMAX, SYNC_FETCH_AND_UMAX)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_MIN, SYNC_FETCH_AND_MIN)
OP_TO_LIBCALL(ISD::ATOMIC_LOAD_UMIN, SYNC_FETCH_AND_UMIN)
}
#undef OP_TO_LIBCALL
return UNKNOWN_LIBCALL;
}
[Atomics] Rename and change prototype for atomic memcpy intrinsic Summary: Background: http://lists.llvm.org/pipermail/llvm-dev/2017-May/112779.html This change is to alter the prototype for the atomic memcpy intrinsic. The prototype itself is being changed to more closely resemble the semantics and parameters of the llvm.memcpy intrinsic -- to ease later combination of the llvm.memcpy and atomic memcpy intrinsics. Furthermore, the name of the atomic memcpy intrinsic is being changed to make it clear that it is not a generic atomic memcpy, but specifically a memcpy is unordered atomic. Reviewers: reames, sanjoy, efriedma Reviewed By: reames Subscribers: mzolotukhin, anna, llvm-commits, skatkov Differential Revision: https://reviews.llvm.org/D33240 llvm-svn: 305558
2017-06-16 22:43:59 +08:00
RTLIB::Libcall RTLIB::getMEMCPY_ELEMENT_UNORDERED_ATOMIC(uint64_t ElementSize) {
Introduce element-wise atomic memcpy intrinsic This change adds a new intrinsic which is intended to provide memcpy functionality with additional atomicity guarantees. Please refer to the review thread or language reference for further details. Differential Revision: https://reviews.llvm.org/D27133 llvm-svn: 290708
2016-12-29 22:31:07 +08:00
switch (ElementSize) {
case 1:
[Atomics] Rename and change prototype for atomic memcpy intrinsic Summary: Background: http://lists.llvm.org/pipermail/llvm-dev/2017-May/112779.html This change is to alter the prototype for the atomic memcpy intrinsic. The prototype itself is being changed to more closely resemble the semantics and parameters of the llvm.memcpy intrinsic -- to ease later combination of the llvm.memcpy and atomic memcpy intrinsics. Furthermore, the name of the atomic memcpy intrinsic is being changed to make it clear that it is not a generic atomic memcpy, but specifically a memcpy is unordered atomic. Reviewers: reames, sanjoy, efriedma Reviewed By: reames Subscribers: mzolotukhin, anna, llvm-commits, skatkov Differential Revision: https://reviews.llvm.org/D33240 llvm-svn: 305558
2017-06-16 22:43:59 +08:00
return MEMCPY_ELEMENT_UNORDERED_ATOMIC_1;
Introduce element-wise atomic memcpy intrinsic This change adds a new intrinsic which is intended to provide memcpy functionality with additional atomicity guarantees. Please refer to the review thread or language reference for further details. Differential Revision: https://reviews.llvm.org/D27133 llvm-svn: 290708
2016-12-29 22:31:07 +08:00
case 2:
[Atomics] Rename and change prototype for atomic memcpy intrinsic Summary: Background: http://lists.llvm.org/pipermail/llvm-dev/2017-May/112779.html This change is to alter the prototype for the atomic memcpy intrinsic. The prototype itself is being changed to more closely resemble the semantics and parameters of the llvm.memcpy intrinsic -- to ease later combination of the llvm.memcpy and atomic memcpy intrinsics. Furthermore, the name of the atomic memcpy intrinsic is being changed to make it clear that it is not a generic atomic memcpy, but specifically a memcpy is unordered atomic. Reviewers: reames, sanjoy, efriedma Reviewed By: reames Subscribers: mzolotukhin, anna, llvm-commits, skatkov Differential Revision: https://reviews.llvm.org/D33240 llvm-svn: 305558
2017-06-16 22:43:59 +08:00
return MEMCPY_ELEMENT_UNORDERED_ATOMIC_2;
Introduce element-wise atomic memcpy intrinsic This change adds a new intrinsic which is intended to provide memcpy functionality with additional atomicity guarantees. Please refer to the review thread or language reference for further details. Differential Revision: https://reviews.llvm.org/D27133 llvm-svn: 290708
2016-12-29 22:31:07 +08:00
case 4:
[Atomics] Rename and change prototype for atomic memcpy intrinsic Summary: Background: http://lists.llvm.org/pipermail/llvm-dev/2017-May/112779.html This change is to alter the prototype for the atomic memcpy intrinsic. The prototype itself is being changed to more closely resemble the semantics and parameters of the llvm.memcpy intrinsic -- to ease later combination of the llvm.memcpy and atomic memcpy intrinsics. Furthermore, the name of the atomic memcpy intrinsic is being changed to make it clear that it is not a generic atomic memcpy, but specifically a memcpy is unordered atomic. Reviewers: reames, sanjoy, efriedma Reviewed By: reames Subscribers: mzolotukhin, anna, llvm-commits, skatkov Differential Revision: https://reviews.llvm.org/D33240 llvm-svn: 305558
2017-06-16 22:43:59 +08:00
return MEMCPY_ELEMENT_UNORDERED_ATOMIC_4;
Introduce element-wise atomic memcpy intrinsic This change adds a new intrinsic which is intended to provide memcpy functionality with additional atomicity guarantees. Please refer to the review thread or language reference for further details. Differential Revision: https://reviews.llvm.org/D27133 llvm-svn: 290708
2016-12-29 22:31:07 +08:00
case 8:
[Atomics] Rename and change prototype for atomic memcpy intrinsic Summary: Background: http://lists.llvm.org/pipermail/llvm-dev/2017-May/112779.html This change is to alter the prototype for the atomic memcpy intrinsic. The prototype itself is being changed to more closely resemble the semantics and parameters of the llvm.memcpy intrinsic -- to ease later combination of the llvm.memcpy and atomic memcpy intrinsics. Furthermore, the name of the atomic memcpy intrinsic is being changed to make it clear that it is not a generic atomic memcpy, but specifically a memcpy is unordered atomic. Reviewers: reames, sanjoy, efriedma Reviewed By: reames Subscribers: mzolotukhin, anna, llvm-commits, skatkov Differential Revision: https://reviews.llvm.org/D33240 llvm-svn: 305558
2017-06-16 22:43:59 +08:00
return MEMCPY_ELEMENT_UNORDERED_ATOMIC_8;
Introduce element-wise atomic memcpy intrinsic This change adds a new intrinsic which is intended to provide memcpy functionality with additional atomicity guarantees. Please refer to the review thread or language reference for further details. Differential Revision: https://reviews.llvm.org/D27133 llvm-svn: 290708
2016-12-29 22:31:07 +08:00
case 16:
[Atomics] Rename and change prototype for atomic memcpy intrinsic Summary: Background: http://lists.llvm.org/pipermail/llvm-dev/2017-May/112779.html This change is to alter the prototype for the atomic memcpy intrinsic. The prototype itself is being changed to more closely resemble the semantics and parameters of the llvm.memcpy intrinsic -- to ease later combination of the llvm.memcpy and atomic memcpy intrinsics. Furthermore, the name of the atomic memcpy intrinsic is being changed to make it clear that it is not a generic atomic memcpy, but specifically a memcpy is unordered atomic. Reviewers: reames, sanjoy, efriedma Reviewed By: reames Subscribers: mzolotukhin, anna, llvm-commits, skatkov Differential Revision: https://reviews.llvm.org/D33240 llvm-svn: 305558
2017-06-16 22:43:59 +08:00
return MEMCPY_ELEMENT_UNORDERED_ATOMIC_16;
Introduce element-wise atomic memcpy intrinsic This change adds a new intrinsic which is intended to provide memcpy functionality with additional atomicity guarantees. Please refer to the review thread or language reference for further details. Differential Revision: https://reviews.llvm.org/D27133 llvm-svn: 290708
2016-12-29 22:31:07 +08:00
default:
return UNKNOWN_LIBCALL;
}
}
Add element atomic memmove intrinsic Summary: Continuing the work from https://reviews.llvm.org/D33240, this change introduces an element unordered-atomic memmove intrinsic. This intrinsic is essentially memmove with the implementation requirement that all loads/stores used for the copy are done with unordered-atomic loads/stores of a given element size. Reviewers: eli.friedman, reames, mkazantsev, skatkov Reviewed By: reames Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D34884 llvm-svn: 307796
2017-07-12 23:25:26 +08:00
RTLIB::Libcall RTLIB::getMEMMOVE_ELEMENT_UNORDERED_ATOMIC(uint64_t ElementSize) {
switch (ElementSize) {
case 1:
return MEMMOVE_ELEMENT_UNORDERED_ATOMIC_1;
case 2:
return MEMMOVE_ELEMENT_UNORDERED_ATOMIC_2;
case 4:
return MEMMOVE_ELEMENT_UNORDERED_ATOMIC_4;
case 8:
return MEMMOVE_ELEMENT_UNORDERED_ATOMIC_8;
case 16:
return MEMMOVE_ELEMENT_UNORDERED_ATOMIC_16;
default:
return UNKNOWN_LIBCALL;
}
}
Add element atomic memset intrinsic Summary: Continuing the work from https://reviews.llvm.org/D33240, this change introduces an element unordered-atomic memset intrinsic. This intrinsic is essentially memset with the implementation requirement that all stores used for the assignment are done with unordered-atomic stores of a given element size. Reviewers: eli.friedman, reames, mkazantsev, skatkov Reviewed By: reames Subscribers: jfb, dschuff, sbc100, jgravelle-google, aheejin, efriedma, llvm-commits Differential Revision: https://reviews.llvm.org/D34885 llvm-svn: 307854
2017-07-13 05:57:23 +08:00
RTLIB::Libcall RTLIB::getMEMSET_ELEMENT_UNORDERED_ATOMIC(uint64_t ElementSize) {
switch (ElementSize) {
case 1:
return MEMSET_ELEMENT_UNORDERED_ATOMIC_1;
case 2:
return MEMSET_ELEMENT_UNORDERED_ATOMIC_2;
case 4:
return MEMSET_ELEMENT_UNORDERED_ATOMIC_4;
case 8:
return MEMSET_ELEMENT_UNORDERED_ATOMIC_8;
case 16:
return MEMSET_ELEMENT_UNORDERED_ATOMIC_16;
default:
return UNKNOWN_LIBCALL;
}
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
/// InitCmpLibcallCCs - Set default comparison libcall CC.
static void InitCmpLibcallCCs(ISD::CondCode *CCs) {
memset(CCs, ISD::SETCC_INVALID, sizeof(ISD::CondCode)*RTLIB::UNKNOWN_LIBCALL);
CCs[RTLIB::OEQ_F32] = ISD::SETEQ;
CCs[RTLIB::OEQ_F64] = ISD::SETEQ;
CCs[RTLIB::OEQ_F128] = ISD::SETEQ;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
CCs[RTLIB::OEQ_PPCF128] = ISD::SETEQ;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
CCs[RTLIB::UNE_F32] = ISD::SETNE;
CCs[RTLIB::UNE_F64] = ISD::SETNE;
CCs[RTLIB::UNE_F128] = ISD::SETNE;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
CCs[RTLIB::UNE_PPCF128] = ISD::SETNE;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
CCs[RTLIB::OGE_F32] = ISD::SETGE;
CCs[RTLIB::OGE_F64] = ISD::SETGE;
CCs[RTLIB::OGE_F128] = ISD::SETGE;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
CCs[RTLIB::OGE_PPCF128] = ISD::SETGE;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
CCs[RTLIB::OLT_F32] = ISD::SETLT;
CCs[RTLIB::OLT_F64] = ISD::SETLT;
CCs[RTLIB::OLT_F128] = ISD::SETLT;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
CCs[RTLIB::OLT_PPCF128] = ISD::SETLT;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
CCs[RTLIB::OLE_F32] = ISD::SETLE;
CCs[RTLIB::OLE_F64] = ISD::SETLE;
CCs[RTLIB::OLE_F128] = ISD::SETLE;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
CCs[RTLIB::OLE_PPCF128] = ISD::SETLE;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
CCs[RTLIB::OGT_F32] = ISD::SETGT;
CCs[RTLIB::OGT_F64] = ISD::SETGT;
CCs[RTLIB::OGT_F128] = ISD::SETGT;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
CCs[RTLIB::OGT_PPCF128] = ISD::SETGT;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
CCs[RTLIB::UO_F32] = ISD::SETNE;
CCs[RTLIB::UO_F64] = ISD::SETNE;
CCs[RTLIB::UO_F128] = ISD::SETNE;
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
CCs[RTLIB::UO_PPCF128] = ISD::SETNE;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
We can get the TLOF from the TargetMachine - so constructor no longer requires TargetLoweringObjectFile to be passed. llvm-svn: 221926
2014-11-14 05:29:21 +08:00
/// NOTE: The TargetMachine owns TLOF.
DataLayout is mandatory, update the API to reflect it with references. Summary: Now that the DataLayout is a mandatory part of the module, let's start cleaning the codebase. This patch is a first attempt at doing that. This patch is not exactly NFC as for instance some places were passing a nullptr instead of the DataLayout, possibly just because there was a default value on the DataLayout argument to many functions in the API. Even though it is not purely NFC, there is no change in the validation. I turned as many pointer to DataLayout to references, this helped figuring out all the places where a nullptr could come up. I had initially a local version of this patch broken into over 30 independant, commits but some later commit were cleaning the API and touching part of the code modified in the previous commits, so it seemed cleaner without the intermediate state. Test Plan: Reviewers: echristo Subscribers: llvm-commits From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 231740
2015-03-10 10:37:25 +08:00
TargetLoweringBase::TargetLoweringBase(const TargetMachine &tm) : TM(tm) {
Use the target options specified on a function to reset the back-end. During LTO, the target options on functions within the same Module may change. This would necessitate resetting some of the back-end. Do this for X86, because it's a Friday afternoon. llvm-svn: 178917
2013-04-06 05:52:40 +08:00
initActions();
// Perform these initializations only once.
[PPC] Inline expansion of memcmp This patch does an inline expansion of memcmp. It changes the memcmp library call into an inline expansion when the size is known at compile time and is under a target specified threshold. This expansion is implemented in CodeGenPrepare and expands into straight line code. The target specifies a maximum load size and the expansion works by using this size to load the two sources, compare, and exit early if a difference is found. It also has a special case when the memcmp result is used in a compare to zero equality. Differential Revision: https://reviews.llvm.org/D28637 llvm-svn: 304313
2017-06-01 01:12:38 +08:00
MaxStoresPerMemset = MaxStoresPerMemcpy = MaxStoresPerMemmove =
MaxLoadsPerMemcmp = 8;
[AArch64] Gangup loads and stores for pairing. Keep loads and stores together (target defines how many loads and stores to gang up), such that it will help in pairing and vectorization. Differential Revision https://reviews.llvm.org/D46477 llvm-svn: 332482
2018-05-16 23:36:52 +08:00
MaxGluedStoresPerMemcpy = 0;
[PPC] Inline expansion of memcmp This patch does an inline expansion of memcmp. It changes the memcmp library call into an inline expansion when the size is known at compile time and is under a target specified threshold. This expansion is implemented in CodeGenPrepare and expands into straight line code. The target specifies a maximum load size and the expansion works by using this size to load the two sources, compare, and exit early if a difference is found. It also has a special case when the memcmp result is used in a compare to zero equality. Differential Revision: https://reviews.llvm.org/D28637 llvm-svn: 304313
2017-06-01 01:12:38 +08:00
MaxStoresPerMemsetOptSize = MaxStoresPerMemcpyOptSize =
MaxStoresPerMemmoveOptSize = MaxLoadsPerMemcmpOptSize = 4;
Disable compare sinking in CodeGenPrepare when multiple condition registers are available As noted in the comment above CodeGenPrepare::OptimizeInst, which aggressively sinks compares to reduce pressure on the condition register(s), for targets such as PowerPC with multiple condition registers, this may not be the right thing to do. This adds an HasMultipleConditionRegisters boolean to TLI, and CodeGenPrepare::OptimizeInst is skipped when HasMultipleConditionRegisters is true. This functionality will be used by the PowerPC backend in an upcoming commit. Especially when the PowerPC backend starts tracking individual condition register bits as separate allocatable entities (which will happen in this upcoming commit), this sinking from CodeGenPrepare::OptimizeInst is significantly suboptimial. llvm-svn: 198354
2014-01-03 05:13:43 +08:00
HasMultipleConditionRegisters = false;
Set default value of HasExtractBitsInsn to false llvm-svn: 206803
2014-04-22 06:22:44 +08:00
HasExtractBitsInsn = false;
add a cl::opt override for TargetLoweringBase's JumpIsExpensive This patch is not intended to change existing codegen behavior for any target. It just exposes the JumpIsExpensive setting on the command-line to allow for easier testing and emergency overrides. Also, change the existing regression test to use FileCheck, explicitly specify the jump-is-expensive option, and use more precise checks. Differential Revision: http://reviews.llvm.org/D10846 llvm-svn: 241179
2015-07-02 02:10:20 +08:00
JumpIsExpensive = JumpIsExpensiveOverride;
Use the target options specified on a function to reset the back-end. During LTO, the target options on functions within the same Module may change. This would necessitate resetting some of the back-end. Do this for X86, because it's a Friday afternoon. llvm-svn: 178917
2013-04-06 05:52:40 +08:00
PredictableSelectIsExpensive = false;
[CodeGenPrepare] Reapply r224351 with a fix for the assertion failure: The type promotion helper does not support vector type, so when make such it does not kick in in such cases. Original commit message: [CodeGenPrepare] Move sign/zero extensions near loads using type promotion. This patch extends the optimization in CodeGenPrepare that moves a sign/zero extension near a load when the target can combine them. The optimization may promote any operations between the extension and the load to make that possible. Although this optimization may be beneficial for all targets, in particular AArch64, this is enabled for X86 only as I have not benchmarked it for other targets yet. ** Context ** Most targets feature extended loads, i.e., loads that perform a zero or sign extension for free. In that context it is interesting to expose such pattern in CodeGenPrepare so that the instruction selection pass can form such loads. Sometimes, this pattern is blocked because of instructions between the load and the extension. When those instructions are promotable to the extended type, we can expose this pattern. ** Motivating Example ** Let us consider an example: define void @foo(i8* %addr1, i32* %addr2, i8 %a, i32 %b) { %ld = load i8* %addr1 %zextld = zext i8 %ld to i32 %ld2 = load i32* %addr2 %add = add nsw i32 %ld2, %zextld %sextadd = sext i32 %add to i64 %zexta = zext i8 %a to i32 %addza = add nsw i32 %zexta, %zextld %sextaddza = sext i32 %addza to i64 %addb = add nsw i32 %b, %zextld %sextaddb = sext i32 %addb to i64 call void @dummy(i64 %sextadd, i64 %sextaddza, i64 %sextaddb) ret void } As it is, this IR generates the following assembly on x86_64: [...] movzbl (%rdi), %eax # zero-extended load movl (%rsi), %es # plain load addl %eax, %esi # 32-bit add movslq %esi, %rdi # sign extend the result of add movzbl %dl, %edx # zero extend the first argument addl %eax, %edx # 32-bit add movslq %edx, %rsi # sign extend the result of add addl %eax, %ecx # 32-bit add movslq %ecx, %rdx # sign extend the result of add [...] The throughput of this sequence is 7.45 cycles on Ivy Bridge according to IACA. Now, by promoting the additions to form more extended loads we would generate: [...] movzbl (%rdi), %eax # zero-extended load movslq (%rsi), %rdi # sign-extended load addq %rax, %rdi # 64-bit add movzbl %dl, %esi # zero extend the first argument addq %rax, %rsi # 64-bit add movslq %ecx, %rdx # sign extend the second argument addq %rax, %rdx # 64-bit add [...] The throughput of this sequence is 6.15 cycles on Ivy Bridge according to IACA. This kind of sequences happen a lot on code using 32-bit indexes on 64-bit architectures. Note: The throughput numbers are similar on Sandy Bridge and Haswell. ** Proposed Solution ** To avoid the penalty of all these sign/zero extensions, we merge them in the loads at the beginning of the chain of computation by promoting all the chain of computation on the extended type. The promotion is done if and only if we do not introduce new extensions, i.e., if we do not degrade the code quality. To achieve this, we extend the existing “move ext to load” optimization with the promotion mechanism introduced to match larger patterns for addressing mode (r200947). The idea of this extension is to perform the following transformation: ext(promotableInst1(...(promotableInstN(load)))) => promotedInst1(...(promotedInstN(ext(load)))) The promotion mechanism in that optimization is enabled by a new TargetLowering switch, which is off by default. In other words, by default, the optimization performs the “move ext to load” optimization as it was before this patch. ** Performance ** Configuration: x86_64: Ivy Bridge fixed at 2900MHz running OS X 10.10. Tested Optimization Levels: O3/Os Tests: llvm-testsuite + externals. Results: - No regression beside noise. - Improvements: CINT2006/473.astar: ~2% Benchmarks/PAQ8p: ~2% Misc/perlin: ~3% The results are consistent for both O3 and Os. <rdar://problem/18310086> llvm-svn: 224402
2014-12-17 09:36:17 +08:00
EnableExtLdPromotion = false;
Use the target options specified on a function to reset the back-end. During LTO, the target options on functions within the same Module may change. This would necessitate resetting some of the back-end. Do this for X86, because it's a Friday afternoon. llvm-svn: 178917
2013-04-06 05:52:40 +08:00
StackPointerRegisterToSaveRestore = 0;
BooleanContents = UndefinedBooleanContent;
Make it possible for ints/floats to return different values from getBooleanContents() Summary: On MIPS32r6/MIPS64r6, floating point comparisons return 0 or -1 but integer comparisons return 0 or 1. Updated the various uses of getBooleanContents. Two simplifications had to be disabled when float and int boolean contents differ: - ScalarizeVecRes_VSELECT except when the kind of boolean contents is trivially discoverable (i.e. when the condition of the VSELECT is a SETCC node). - visitVSELECT (select C, 0, 1) -> (xor C, 1). Come to think of it, this one could test for the common case of 'C' being a SETCC too. Preserved existing behaviour for all other targets and updated the affected MIPS32r6/MIPS64r6 tests. This also fixes the pi benchmark where the 'low' variable was counting in the wrong direction because it thought it could simply add the result of the comparison. Reviewers: hfinkel Reviewed By: hfinkel Subscribers: hfinkel, jholewinski, mcrosier, llvm-commits Differential Revision: http://reviews.llvm.org/D4389 llvm-svn: 212697
2014-07-10 18:18:12 +08:00
BooleanFloatContents = UndefinedBooleanContent;
Use the target options specified on a function to reset the back-end. During LTO, the target options on functions within the same Module may change. This would necessitate resetting some of the back-end. Do this for X86, because it's a Friday afternoon. llvm-svn: 178917
2013-04-06 05:52:40 +08:00
BooleanVectorContents = UndefinedBooleanContent;
SchedPreferenceInfo = Sched::ILP;
In visitSTORE, always use FindBetterChain, rather than only when UseAA is enabled. Recommiting with compiler time improvements Recommitting after fixup of 32-bit aliasing sign offset bug in DAGCombiner. * Simplify Consecutive Merge Store Candidate Search Now that address aliasing is much less conservative, push through simplified store merging search and chain alias analysis which only checks for parallel stores through the chain subgraph. This is cleaner as the separation of non-interfering loads/stores from the store-merging logic. When merging stores search up the chain through a single load, and finds all possible stores by looking down from through a load and a TokenFactor to all stores visited. This improves the quality of the output SelectionDAG and the output Codegen (save perhaps for some ARM cases where we correctly constructs wider loads, but then promotes them to float operations which appear but requires more expensive constant generation). Some minor peephole optimizations to deal with improved SubDAG shapes (listed below) Additional Minor Changes: 1. Finishes removing unused AliasLoad code 2. Unifies the chain aggregation in the merged stores across code paths 3. Re-add the Store node to the worklist after calling SimplifyDemandedBits. 4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is arbitrary, but seems sufficient to not cause regressions in tests. 5. Remove Chain dependencies of Memory operations on CopyfromReg nodes as these are captured by data dependence 6. Forward loads-store values through tokenfactors containing {CopyToReg,CopyFromReg} Values. 7. Peephole to convert buildvector of extract_vector_elt to extract_subvector if possible (see CodeGen/AArch64/store-merge.ll) 8. Store merging for the ARM target is restricted to 32-bit as some in some contexts invalid 64-bit operations are being generated. This can be removed once appropriate checks are added. This finishes the change Matt Arsenault started in r246307 and jyknight's original patch. Many tests required some changes as memory operations are now reorderable, improving load-store forwarding. One test in particular is worth noting: CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store forwarding converts a load-store pair into a parallel store and a memory-realized bitcast of the same value. However, because we lose the sharing of the explicit and implicit store values we must create another local store. A similar transformation happens before SelectionDAG as well. Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle llvm-svn: 297695
2017-03-14 08:34:14 +08:00
GatherAllAliasesMaxDepth = 18;
[FPEnv] Add an option to disable strict float node mutating to an normal float node This patch add an option 'disable-strictnode-mutation' to prevent strict node mutating to an normal node. So we can make sure that the patch which sets strict-node as legal works correctly. Patch by Chen Liu(LiuChen3) Differential Revision: https://reviews.llvm.org/D70226
2019-11-22 09:17:19 +08:00
IsStrictFPEnabled = DisableStrictNodeMutation;
Add __atomic_* lowering to AtomicExpandPass. (Recommit of r266002, with r266011, r266016, and not accidentally including an extra unused/uninitialized element in LibcallRoutineNames) AtomicExpandPass can now lower atomic load, atomic store, atomicrmw, and cmpxchg instructions to __atomic_* library calls, when the target doesn't support atomics of a given size. This is the first step towards moving all atomic lowering from clang into llvm. When all is done, the behavior of __sync_* builtins, __atomic_* builtins, and C11 atomics will be unified. Previously LLVM would pass everything through to the ISelLowering code. There, unsupported atomic instructions would turn into __sync_* library calls. Because of that behavior, Clang currently avoids emitting llvm IR atomic instructions when this would happen, and emits __atomic_* library functions itself, in the frontend. This change makes LLVM able to emit __atomic_* libcalls, and thus will eventually allow clang to depend on LLVM to do the right thing. It is advantageous to do the new lowering to atomic libcalls in AtomicExpandPass, before ISel time, because it's important that all atomic operations for a given size either lower to __atomic_* libcalls (which may use locks), or native instructions which won't. No mixing and matching. At the moment, this code is enabled only for SPARC, as a demonstration. The next commit will expand support to all of the other targets. Differential Revision: http://reviews.llvm.org/D18200 llvm-svn: 266115
2016-04-13 04:18:48 +08:00
// TODO: the default will be switched to 0 in the next commit, along
// with the Target-specific changes necessary.
MaxAtomicSizeInBitsSupported = 1024;
Use the target options specified on a function to reset the back-end. During LTO, the target options on functions within the same Module may change. This would necessitate resetting some of the back-end. Do this for X86, because it's a Friday afternoon. llvm-svn: 178917
2013-04-06 05:52:40 +08:00
Support expanding partial-word cmpxchg to full-word cmpxchg in AtomicExpandPass. Many CPUs only have the ability to do a 4-byte cmpxchg (or ll/sc), not 1 or 2-byte. For those, you need to mask and shift the 1 or 2 byte values appropriately to use the 4-byte instruction. This change adds support for cmpxchg-based instruction sets (only SPARC, in LLVM). The support can be extended for LL/SC-based PPC and MIPS in the future, supplanting the ISel expansions those architectures currently use. Tests added for the IR transform and SPARCv9. Differential Revision: http://reviews.llvm.org/D21029 llvm-svn: 273025
2016-06-18 02:11:48 +08:00
MinCmpXchgSizeInBits = 0;
Relax unaligned access assertion when type is byte aligned Summary: This relaxes an assertion inside SelectionDAGBuilder which is overly restrictive on targets which have no concept of alignment (such as AVR). In these architectures, all types are aligned to 8-bits. After this, LLVM will only assert that accesses are aligned on targets which actually require alignment. This patch follows from a discussion on llvm-dev a few months ago http://llvm.1065342.n5.nabble.com/llvm-dev-Unaligned-atomic-load-store-td112815.html Reviewers: bogner, nemanjai, joerg, efriedma Reviewed By: efriedma Subscribers: efriedma, cactus, llvm-commits Differential Revision: https://reviews.llvm.org/D39946 llvm-svn: 320243
2017-12-09 14:45:36 +08:00
SupportsUnalignedAtomics = false;
Support expanding partial-word cmpxchg to full-word cmpxchg in AtomicExpandPass. Many CPUs only have the ability to do a 4-byte cmpxchg (or ll/sc), not 1 or 2-byte. For those, you need to mask and shift the 1 or 2 byte values appropriately to use the 4-byte instruction. This change adds support for cmpxchg-based instruction sets (only SPARC, in LLVM). The support can be extended for LL/SC-based PPC and MIPS in the future, supplanting the ISel expansions those architectures currently use. Tests added for the IR transform and SPARCv9. Differential Revision: http://reviews.llvm.org/D21029 llvm-svn: 273025
2016-06-18 02:11:48 +08:00
Pre-fill LibcallRoutineNames with nullptr. And rearrange InitLibcallNames slightly. llvm-svn: 266142
2016-04-13 06:32:47 +08:00
std::fill(std::begin(LibcallRoutineNames), std::end(LibcallRoutineNames), nullptr);
X86/AArch64/ARM: Factor out common sincos_stret logic; NFCI Note: - X86ISelLowering: setLibcallName(SINCOS) was superfluous as InitLibcalls() already does it. - ARMISelLowering: Setting libcallnames for sincos/sincosf seemed superfluous as in the darwin case it wouldn't be used while for all other cases InitLibcalls already does it. llvm-svn: 321036
2017-12-19 07:19:42 +08:00
InitLibcalls(TM.getTargetTriple());
Use the target options specified on a function to reset the back-end. During LTO, the target options on functions within the same Module may change. This would necessitate resetting some of the back-end. Do this for X86, because it's a Friday afternoon. llvm-svn: 178917
2013-04-06 05:52:40 +08:00
InitCmpLibcallCCs(CmpLibcallCCs);
}
void TargetLoweringBase::initActions() {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// All operations default to being supported.
memset(OpActions, 0, sizeof(OpActions));
memset(LoadExtActions, 0, sizeof(LoadExtActions));
memset(TruncStoreActions, 0, sizeof(TruncStoreActions));
memset(IndexedModeActions, 0, sizeof(IndexedModeActions));
memset(CondCodeActions, 0, sizeof(CondCodeActions));
Use std::fill to simplify some code. NFC llvm-svn: 265771
2016-04-08 15:10:46 +08:00
std::fill(std::begin(RegClassForVT), std::end(RegClassForVT), nullptr);
std::fill(std::begin(TargetDAGCombineArray),
std::end(TargetDAGCombineArray), 0);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
Allow FP types for atomicrmw xchg llvm-svn: 351427
2019-01-17 18:49:01 +08:00
for (MVT VT : MVT::fp_valuetypes()) {
[llvm][CodeGen] Avoid implicit cast of TypeSize to integer in `initActions`. Reviewers: sdesmalen, efriedma Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77317
2020-04-07 02:25:17 +08:00
MVT IntVT = MVT::getIntegerVT(VT.getSizeInBits().getFixedSize());
Allow FP types for atomicrmw xchg llvm-svn: 351427
2019-01-17 18:49:01 +08:00
if (IntVT.isValid()) {
setOperationAction(ISD::ATOMIC_SWAP, VT, Promote);
AddPromotedToType(ISD::ATOMIC_SWAP, VT, IntVT);
}
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Set default actions for various operations.
[CodeGen] Use MVT iterator_ranges in legality loops. NFC intended. A few loops do trickier things than just iterating on an MVT subset, so I'll leave them be for now. Follow-up of r225387. llvm-svn: 225392
2015-01-08 05:27:10 +08:00
for (MVT VT : MVT::all_valuetypes()) {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Default all indexed load / store to expand.
for (unsigned IM = (unsigned)ISD::PRE_INC;
IM != (unsigned)ISD::LAST_INDEXED_MODE; ++IM) {
[CodeGen] Use MVT iterator_ranges in legality loops. NFC intended. A few loops do trickier things than just iterating on an MVT subset, so I'll leave them be for now. Follow-up of r225387. llvm-svn: 225392
2015-01-08 05:27:10 +08:00
setIndexedLoadAction(IM, VT, Expand);
setIndexedStoreAction(IM, VT, Expand);
[Codegen][ARM] Add addressing modes from masked loads and stores MVE has a basic symmetry between it's normal loads/store operations and the masked variants. This means that masked loads and stores can use pre-inc and post-inc addressing modes, just like the standard loads and stores already do. To enable that, this patch adds all the relevant infrastructure for treating masked loads/stores addressing modes in the same way as normal loads/stores. This involves: - Adding an AddressingMode to MaskedLoadStoreSDNode, along with an extra Offset operand that is added after the PtrBase. - Extending the IndexedModeActions from 8bits to 16bits to store the legality of masked operations as well as normal ones. This array is fairly small, so doubling the size still won't make it very large. Offset masked loads can then be controlled with setIndexedMaskedLoadAction, similar to standard loads. - The same methods that combine to indexed loads, such as CombineToPostIndexedLoadStore, are adjusted to handle masked loads in the same way. - The ARM backend is then adjusted to make use of these indexed masked loads/stores. - The X86 backend is adjusted to hopefully be no functional changes. Differential Revision: https://reviews.llvm.org/D70176
2019-11-21 22:56:37 +08:00
setIndexedMaskedLoadAction(IM, VT, Expand);
setIndexedMaskedStoreAction(IM, VT, Expand);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
IR: add "cmpxchg weak" variant to support permitted failure. This commit adds a weak variant of the cmpxchg operation, as described in C++11. A cmpxchg instruction with this modifier is permitted to fail to store, even if the comparison indicated it should. As a result, cmpxchg instructions must return a flag indicating success in addition to their original iN value loaded. Thus, for uniformity *all* cmpxchg instructions now return "{ iN, i1 }". The second flag is 1 when the store succeeded. At the DAG level, a new ATOMIC_CMP_SWAP_WITH_SUCCESS node has been added as the natural representation for the new cmpxchg instructions. It is a strong cmpxchg. By default this gets Expanded to the existing ATOMIC_CMP_SWAP during Legalization, so existing backends should see no change in behaviour. If they wish to deal with the enhanced node instead, they can call setOperationAction on it. Beware: as a node with 2 results, it cannot be selected from TableGen. Currently, no use is made of the extra information provided in this patch. Test updates are almost entirely adapting the input IR to the new scheme. Summary for out of tree users: ------------------------------ + Legacy Bitcode files are upgraded during read. + Legacy assembly IR files will be invalid. + Front-ends must adapt to different type for "cmpxchg". + Backends should be unaffected by default. llvm-svn: 210903
2014-06-13 22:24:07 +08:00
// Most backends expect to see the node which just returns the value loaded.
[CodeGen] Use MVT iterator_ranges in legality loops. NFC intended. A few loops do trickier things than just iterating on an MVT subset, so I'll leave them be for now. Follow-up of r225387. llvm-svn: 225392
2015-01-08 05:27:10 +08:00
setOperationAction(ISD::ATOMIC_CMP_SWAP_WITH_SUCCESS, VT, Expand);
IR: add "cmpxchg weak" variant to support permitted failure. This commit adds a weak variant of the cmpxchg operation, as described in C++11. A cmpxchg instruction with this modifier is permitted to fail to store, even if the comparison indicated it should. As a result, cmpxchg instructions must return a flag indicating success in addition to their original iN value loaded. Thus, for uniformity *all* cmpxchg instructions now return "{ iN, i1 }". The second flag is 1 when the store succeeded. At the DAG level, a new ATOMIC_CMP_SWAP_WITH_SUCCESS node has been added as the natural representation for the new cmpxchg instructions. It is a strong cmpxchg. By default this gets Expanded to the existing ATOMIC_CMP_SWAP during Legalization, so existing backends should see no change in behaviour. If they wish to deal with the enhanced node instead, they can call setOperationAction on it. Beware: as a node with 2 results, it cannot be selected from TableGen. Currently, no use is made of the extra information provided in this patch. Test updates are almost entirely adapting the input IR to the new scheme. Summary for out of tree users: ------------------------------ + Legacy Bitcode files are upgraded during read. + Legacy assembly IR files will be invalid. + Front-ends must adapt to different type for "cmpxchg". + Backends should be unaffected by default. llvm-svn: 210903
2014-06-13 22:24:07 +08:00
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// These operations default to expand.
[CodeGen] Use MVT iterator_ranges in legality loops. NFC intended. A few loops do trickier things than just iterating on an MVT subset, so I'll leave them be for now. Follow-up of r225387. llvm-svn: 225392
2015-01-08 05:27:10 +08:00
setOperationAction(ISD::FGETSIGN, VT, Expand);
setOperationAction(ISD::CONCAT_VECTORS, VT, Expand);
setOperationAction(ISD::FMINNUM, VT, Expand);
setOperationAction(ISD::FMAXNUM, VT, Expand);
DAG: Change behavior of fminnum/fmaxnum nodes Introduce new versions that follow the IEEE semantics to help with legalization that may need quieted inputs. There are some regressions from inserting unnecessary canonicalizes when these are matched from fast math fcmp + select which should be fixed in a future commit. llvm-svn: 344914
2018-10-23 00:27:27 +08:00
setOperationAction(ISD::FMINNUM_IEEE, VT, Expand);
setOperationAction(ISD::FMAXNUM_IEEE, VT, Expand);
[NFC] Rename minnan and maxnan to minimum and maximum Summary: Changes all uses of minnan/maxnan to minimum/maximum globally. These names emphasize that the semantic difference between these operations is more than just NaN-propagation. Reviewers: arsenm, aheejin, dschuff, javed.absar Subscribers: jholewinski, sdardis, wdng, sbc100, jgravelle-google, jrtc27, atanasyan, llvm-commits Differential Revision: https://reviews.llvm.org/D53112 llvm-svn: 345218
2018-10-25 06:49:55 +08:00
setOperationAction(ISD::FMINIMUM, VT, Expand);
setOperationAction(ISD::FMAXIMUM, VT, Expand);
Add generic fmad DAG node. This allows sharing of FMA forming combines to work with instructions that have the same semantics as a separate multiply and add. This is expand by default, and only formed post legalization so it shouldn't have much impact on targets that do not want it. llvm-svn: 230070
2015-02-21 06:10:33 +08:00
setOperationAction(ISD::FMAD, VT, Expand);
Add SDNodes for umin, umax, smin and smax. This adds new SDNodes for signed/unsigned min/max. These nodes are built from select/icmp pairs matched at SDAGBuilder stage. This patch adds the nodes, as well as legalization support and sets them to be "expand" for all targets. NFC for now; this will be tested when I switch AArch64 to using these new nodes. llvm-svn: 237423
2015-05-15 17:03:15 +08:00
setOperationAction(ISD::SMIN, VT, Expand);
setOperationAction(ISD::SMAX, VT, Expand);
setOperationAction(ISD::UMIN, VT, Expand);
setOperationAction(ISD::UMAX, VT, Expand);
[SelectionDAG] Add a signed integer absolute ISD node Reduced version of D26357 - based on the discussion on llvm-dev about canonicalization of UMIN/UMAX/SMIN/SMAX as well as ABS I've reduced that patch to just the ABS ISD node (with x86/sse support) to improve basic combines and lowering. ARM/AArch64, Hexagon, PowerPC and NVPTX all have similar instructions allowing us to make this a generic opcode and move away from the hard coded tablegen patterns which makes it tricky to match more complex patterns. At the moment this patch doesn't attempt legalization as we only create an ABS node if its legal/custom. Differential Revision: https://reviews.llvm.org/D29639 llvm-svn: 297780
2017-03-15 05:26:58 +08:00
setOperationAction(ISD::ABS, VT, Expand);
[SelectionDAG] Initial support for FSHL/FSHR funnel shift opcodes (PR39467) This is an initial patch to add a minimum level of support for funnel shifts to the SelectionDAG and to begin wiring it up to the X86 SHLD/SHRD instructions. Some partial legalization code has been added to handle the case for 'SlowSHLD' where we want to expand instead and I've added a few DAG combines so we don't get regressions from the existing DAG builder expansion code. Differential Revision: https://reviews.llvm.org/D54698 llvm-svn: 348353
2018-12-05 19:12:12 +08:00
setOperationAction(ISD::FSHL, VT, Expand);
setOperationAction(ISD::FSHR, VT, Expand);
[Intrinsic] Signed Saturation Addition Intrinsic Add an intrinsic that takes 2 integers and perform saturation addition on them. This is a part of implementing fixed point arithmetic in clang where some of the more complex operations will be implemented as intrinsics. Differential Revision: https://reviews.llvm.org/D53053 llvm-svn: 344629
2018-10-17 01:35:41 +08:00
setOperationAction(ISD::SADDSAT, VT, Expand);
[Intrinsic] Unigned Saturation Addition Intrinsic Add an intrinsic that takes 2 integers and perform unsigned saturation addition on them. This is a part of implementing fixed point arithmetic in clang where some of the more complex operations will be implemented as intrinsics. Differential Revision: https://reviews.llvm.org/D53340 llvm-svn: 344971
2018-10-23 07:08:40 +08:00
setOperationAction(ISD::UADDSAT, VT, Expand);
[Intrinsic] Signed and Unsigned Saturation Subtraction Intirnsics Add an intrinsic that takes 2 integers and perform saturation subtraction on them. This is a part of implementing fixed point arithmetic in clang where some of the more complex operations will be implemented as intrinsics. Differential Revision: https://reviews.llvm.org/D53783 llvm-svn: 345512
2018-10-30 00:54:37 +08:00
setOperationAction(ISD::SSUBSAT, VT, Expand);
setOperationAction(ISD::USUBSAT, VT, Expand);
[Intrinsic] Signed Fixed Point Multiplication Intrinsic Add an intrinsic that takes 2 signed integers with the scale of them provided as the third argument and performs fixed point multiplication on them. This is a part of implementing fixed point arithmetic in clang where some of the more complex operations will be implemented as intrinsics. Differential Revision: https://reviews.llvm.org/D54719 llvm-svn: 348912
2018-12-12 14:29:14 +08:00
setOperationAction(ISD::SMULFIX, VT, Expand);
[Intrinsic] Signed Fixed Point Saturation Multiplication Intrinsic Add an intrinsic that takes 2 signed integers with the scale of them provided as the third argument and performs fixed point multiplication on them. The result is saturated and clamped between the largest and smallest representable values of the first 2 operands. This is a part of implementing fixed point arithmetic in clang where some of the more complex operations will be implemented as intrinsics. Differential Revision: https://reviews.llvm.org/D55720 llvm-svn: 361289
2019-05-22 03:17:19 +08:00
setOperationAction(ISD::SMULFIXSAT, VT, Expand);
[Intrinsic] Unsigned Fixed Point Multiplication Intrinsic Add an intrinsic that takes 2 unsigned integers with the scale of them provided as the third argument and performs fixed point multiplication on them. This is a part of implementing fixed point arithmetic in clang where some of the more complex operations will be implemented as intrinsics. Differential Revision: https://reviews.llvm.org/D55625 llvm-svn: 353059
2019-02-05 01:18:11 +08:00
setOperationAction(ISD::UMULFIX, VT, Expand);
[Intrinsic] Add the llvm.umul.fix.sat intrinsic Summary: Add an intrinsic that takes 2 unsigned integers with the scale of them provided as the third argument and performs fixed point multiplication on them. The result is saturated and clamped between the largest and smallest representable values of the first 2 operands. This is a part of implementing fixed point arithmetic in clang where some of the more complex operations will be implemented as intrinsics. Patch by: leonardchan, bjope Reviewers: RKSimon, craig.topper, bevinh, leonardchan, lebedev.ri, spatel Reviewed By: leonardchan Subscribers: ychen, wuzish, nemanjai, MaskRay, jsji, jdoerfert, Ka-Ka, hiraditya, rjmccall, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D57836 llvm-svn: 371308
2019-09-07 20:16:14 +08:00
setOperationAction(ISD::UMULFIXSAT, VT, Expand);
[Intrinsic] Add fixed point division intrinsics. Summary: This patch adds intrinsics and ISelDAG nodes for signed and unsigned fixed-point division: llvm.sdiv.fix.* llvm.udiv.fix.* These intrinsics perform scaled division on two integers or vectors of integers. They are required for the implementation of the Embedded-C fixed-point arithmetic in Clang. Patch by: ebevhan Reviewers: bjope, leonardchan, efriedma, craig.topper Reviewed By: craig.topper Subscribers: Ka-Ka, ilya, hiraditya, jdoerfert, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D70007
2020-01-08 22:05:03 +08:00
setOperationAction(ISD::SDIVFIX, VT, Expand);
[Intrinsic] Add fixed point saturating division intrinsics. Summary: This patch adds intrinsics and ISelDAG nodes for signed and unsigned fixed-point division: ``` llvm.sdiv.fix.sat.* llvm.udiv.fix.sat.* ``` These intrinsics perform scaled, saturating division on two integers or vectors of integers. They are required for the implementation of the Embedded-C fixed-point arithmetic in Clang. Reviewers: bjope, leonardchan, craig.topper Subscribers: hiraditya, jdoerfert, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D71550
2019-12-16 22:25:52 +08:00
setOperationAction(ISD::SDIVFIXSAT, VT, Expand);
[Intrinsic] Add fixed point division intrinsics. Summary: This patch adds intrinsics and ISelDAG nodes for signed and unsigned fixed-point division: llvm.sdiv.fix.* llvm.udiv.fix.* These intrinsics perform scaled division on two integers or vectors of integers. They are required for the implementation of the Embedded-C fixed-point arithmetic in Clang. Patch by: ebevhan Reviewers: bjope, leonardchan, efriedma, craig.topper Reviewed By: craig.topper Subscribers: Ka-Ka, ilya, hiraditya, jdoerfert, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D70007
2020-01-08 22:05:03 +08:00
setOperationAction(ISD::UDIVFIX, VT, Expand);
[Intrinsic] Add fixed point saturating division intrinsics. Summary: This patch adds intrinsics and ISelDAG nodes for signed and unsigned fixed-point division: ``` llvm.sdiv.fix.sat.* llvm.udiv.fix.sat.* ``` These intrinsics perform scaled, saturating division on two integers or vectors of integers. They are required for the implementation of the Embedded-C fixed-point arithmetic in Clang. Reviewers: bjope, leonardchan, craig.topper Subscribers: hiraditya, jdoerfert, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D71550
2019-12-16 22:25:52 +08:00
setOperationAction(ISD::UDIVFIXSAT, VT, Expand);
Set ISD::FROUND to Expand by default for all types For most libm ISD nodes, TargetLoweringBase::initActions sets the default scalar-type action to Expand, and leaves the vector-type action default as Legal. This is not appropriate for the new ISD::FROUND node (which no backend but PowerPC handles explicitly). Fixes PR16842. llvm-svn: 188048
2013-08-09 12:13:44 +08:00
CodeGen: Default overflow operations to expand so we don't have to assume targets are lying Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu> Reviewed-by: ab Differential Revision: http://reviews.llvm.org/D9265 llvm-svn: 236119
2015-04-30 00:30:46 +08:00
// Overflow operations default to expand
setOperationAction(ISD::SADDO, VT, Expand);
setOperationAction(ISD::SSUBO, VT, Expand);
setOperationAction(ISD::UADDO, VT, Expand);
setOperationAction(ISD::USUBO, VT, Expand);
setOperationAction(ISD::SMULO, VT, Expand);
setOperationAction(ISD::UMULO, VT, Expand);
Revert r248483, r242546, r242545, and r242409 - absdiff intrinsics After much discussion, ending here: http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151123/315620.html it has been decided that, instead of having the vectorizer directly generate special absdiff and horizontal-add intrinsics, we'll recognize the relevant reduction patterns during CodeGen. Accordingly, these intrinsics are not needed (the operations they represent can be pattern matched, as is already done in some backends). Thus, we're backing these out in favor of the current development work. r248483 - Codegen: Fix llvm.*absdiff semantic. r242546 - [ARM] Use [SU]ABSDIFF nodes instead of intrinsics for VABD/VABA r242545 - [AArch64] Use [SU]ABSDIFF nodes instead of intrinsics for ABD/ABA r242409 - [Codegen] Add intrinsics 'absdiff' and corresponding SDNodes for absolute difference operation llvm-svn: 255387
2015-12-12 07:11:52 +08:00
Do not legalize large add with addc/adde, introduce addcarry and do it with uaddo/addcarry Summary: As per discution on how to get better codegen an large int legalization, it became clear that using a glue for the carry was preventing several desirable optimizations. Passing the carry down as a value allow for more flexibility. Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer Subscribers: igorb, llvm-commits Differential Revision: https://reviews.llvm.org/D29872 llvm-svn: 301775
2017-05-01 03:24:09 +08:00
// ADDCARRY operations default to expand
setOperationAction(ISD::ADDCARRY, VT, Expand);
setOperationAction(ISD::SUBCARRY, VT, Expand);
Do not legalize large setcc with setcce, introduce setcccarry and do it with usubo/setcccarry. Summary: This is a continuation of the work started in D29872 . Passing the carry down as a value rather than as a glue allows for further optimizations. Introducing setcccarry makes the use of addc/subc unecessary and we can start the removal process. This patch only introduce the optimization strictly required to get the same level of optimization as was available before nothing more. Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D33374 llvm-svn: 304404
2017-06-01 19:14:17 +08:00
setOperationAction(ISD::SETCCCARRY, VT, Expand);
Do not legalize large add with addc/adde, introduce addcarry and do it with uaddo/addcarry Summary: As per discution on how to get better codegen an large int legalization, it became clear that using a glue for the carry was preventing several desirable optimizations. Passing the carry down as a value allow for more flexibility. Reviewers: jyknight, nemanjai, mkuper, spatel, RKSimon, zvi, bkramer Subscribers: igorb, llvm-commits Differential Revision: https://reviews.llvm.org/D29872 llvm-svn: 301775
2017-05-01 03:24:09 +08:00
Set ADDE/ADDC/SUBE/SUBC to expand by default Summary: They've been deprecated in favor of UADDO/ADDCARRY or USUBO/SUBCARRY for a while. Target that uses these opcodes are changed in order to ensure their behavior doesn't change. Reviewers: efriedma, craig.topper, dblaikie, bkramer Subscribers: jholewinski, arsenm, jyknight, sdardis, nemanjai, nhaehnle, kbarton, fedor.sergeev, asb, rbar, johnrusso, simoncook, jordy.potman.lists, apazos, sabuasal, niosHD, jrtc27, zzheng, edward-jones, mgrang, atanasyan, llvm-commits Differential Revision: https://reviews.llvm.org/D47422 llvm-svn: 333748
2018-06-01 21:21:33 +08:00
// ADDC/ADDE/SUBC/SUBE default to expand.
setOperationAction(ISD::ADDC, VT, Expand);
setOperationAction(ISD::ADDE, VT, Expand);
setOperationAction(ISD::SUBC, VT, Expand);
setOperationAction(ISD::SUBE, VT, Expand);
[CodeGen] Default CTTZ_ZERO_UNDEF/CTLZ_ZERO_UNDEF to Expand in TargetLoweringBase. This is what the majority of the targets want and removes a bunch of code. Set it to Legal explicitly in the few cases where that's the desired behavior. llvm-svn: 267853
2016-04-28 11:34:31 +08:00
// These default to Expand so they will be expanded to CTLZ/CTTZ by default.
setOperationAction(ISD::CTLZ_ZERO_UNDEF, VT, Expand);
setOperationAction(ISD::CTTZ_ZERO_UNDEF, VT, Expand);
[SDAG] Introduce a new BITREVERSE node along with a corresponding LLVM intrinsic Several backends have instructions to reverse the order of bits in an integer. Conceptually matching such patterns is similar to @llvm.bswap, and it was mentioned in http://reviews.llvm.org/D14234 that it would be best if these patterns were matched in InstCombine instead of reimplemented in every different target. This patch introduces an intrinsic @llvm.bitreverse.i* that operates similarly to @llvm.bswap. For plumbing purposes there is also a new ISD node ISD::BITREVERSE, with simple expansion and promotion support. The intention is that InstCombine's BSWAP detection logic will be extended to support BITREVERSE too, and @llvm.bitreverse intrinsics emitted (if the backend supports lowering it efficiently). llvm-svn: 252878
2015-11-12 20:29:09 +08:00
setOperationAction(ISD::BITREVERSE, VT, Expand);
Remove trailing space sed -Ei 's/[[:space:]]+$//' include/**/*.{def,h,td} lib/**/*.{cpp,h} llvm-svn: 338293
2018-07-31 03:41:25 +08:00
Set ISD::FROUND to Expand by default for all types For most libm ISD nodes, TargetLoweringBase::initActions sets the default scalar-type action to Expand, and leaves the vector-type action default as Legal. This is not appropriate for the new ISD::FROUND node (which no backend but PowerPC handles explicitly). Fixes PR16842. llvm-svn: 188048
2013-08-09 12:13:44 +08:00
// These library functions default to expand.
[CodeGen] Use MVT iterator_ranges in legality loops. NFC intended. A few loops do trickier things than just iterating on an MVT subset, so I'll leave them be for now. Follow-up of r225387. llvm-svn: 225392
2015-01-08 05:27:10 +08:00
setOperationAction(ISD::FROUND, VT, Expand);
[FPEnv] Intrinsic llvm.roundeven This intrinsic implements IEEE-754 operation roundToIntegralTiesToEven, and performs rounding to the nearest integer value, rounding halfway cases to even. The intrinsic represents the missed case of IEEE-754 rounding operations and now llvm provides full support of the rounding operations defined by the standard. Differential Revision: https://reviews.llvm.org/D75670
2020-05-26 20:24:05 +08:00
setOperationAction(ISD::FROUNDEVEN, VT, Expand);
[SelectionDAG] Set ISD::FPOWI to Expand by default Summary: Currently FPOWI defaults to Legal and LegalizeDAG.cpp turns Legal into Expand for this opcode because Legal is a "lie". This patch changes the default for this opcode to Expand and removes the hack from LegalizeDAG.cpp. It also removes all the code in the targets that set this opcode to Expand themselves since they can just rely on the default. Reviewers: spatel, RKSimon, efriedma Reviewed By: RKSimon Subscribers: jfb, dschuff, sbc100, jgravelle-google, nemanjai, javed.absar, andrew.w.kaylor, llvm-commits Differential Revision: https://reviews.llvm.org/D33530 llvm-svn: 304215
2017-05-30 23:27:55 +08:00
setOperationAction(ISD::FPOWI, VT, Expand);
Add a llvm.copysign intrinsic This adds a llvm.copysign intrinsic; We already have Libfunc recognition for copysign (which is turned into the FCOPYSIGN SDAG node). In order to autovectorize calls to copysign in the loop vectorizer, we need a corresponding intrinsic as well. In addition to the expected changes to the language reference, the loop vectorizer, BasicTTI, and the SDAG builder (the intrinsic is transformed into an FCOPYSIGN node, just like the function call), this also adds FCOPYSIGN to a few lists in LegalizeVector{Ops,Types} so that vector copysigns can be expanded. In TargetLoweringBase::initActions, I've made the default action for FCOPYSIGN be Expand for vector types. This seems correct for all in-tree targets, and I think is the right thing to do because, previously, there was no way to generate vector-values FCOPYSIGN nodes (and most targets don't specify an action for vector-typed FCOPYSIGN). llvm-svn: 188728
2013-08-20 07:35:46 +08:00
// These operations default to expand for vector types.
[CodeGen] Use MVT iterator_ranges in legality loops. NFC intended. A few loops do trickier things than just iterating on an MVT subset, so I'll leave them be for now. Follow-up of r225387. llvm-svn: 225392
2015-01-08 05:27:10 +08:00
if (VT.isVector()) {
setOperationAction(ISD::FCOPYSIGN, VT, Expand);
[DAGCombine] Initialize the default operation action for SIGN_EXTEND_INREG for vector type as 'expand' instead of 'legal' For now, we didn't set the default operation action for SIGN_EXTEND_INREG for vector type, which is 0 by default, that is legal. However, most target didn't have native instructions to support this opcode. It should be set as expand by default, as what we did for ANY_EXTEND_VECTOR_INREG. Differential Revision: https://reviews.llvm.org/D70000
2020-01-03 11:26:41 +08:00
setOperationAction(ISD::SIGN_EXTEND_INREG, VT, Expand);
[CodeGen] Use MVT iterator_ranges in legality loops. NFC intended. A few loops do trickier things than just iterating on an MVT subset, so I'll leave them be for now. Follow-up of r225387. llvm-svn: 225392
2015-01-08 05:27:10 +08:00
setOperationAction(ISD::ANY_EXTEND_VECTOR_INREG, VT, Expand);
setOperationAction(ISD::SIGN_EXTEND_VECTOR_INREG, VT, Expand);
setOperationAction(ISD::ZERO_EXTEND_VECTOR_INREG, VT, Expand);
[AArch64][SVE] Add SPLAT_VECTOR ISD Node Adds a new ISD node to replicate a scalar value across all elements of a vector. This is needed for scalable vectors, since BUILD_VECTOR cannot be used. Fixes up default type legalization for scalable vectors after the new MVT type ranges were introduced. At present I only use this node for scalable vectors. A DAGCombine has been added to transform a BUILD_VECTOR into a SPLAT_VECTOR if all elements are the same, but only if the default operation action of Expand has been overridden by the target. I've only added result promotion legalization for scalable vector i8/i16/i32/i64 types in AArch64 for now. Reviewers: t.p.northover, javed.absar, greened, cameron.mcinally, jmolloy Reviewed By: jmolloy Differential Revision: https://reviews.llvm.org/D47775 llvm-svn: 375222
2019-10-18 19:48:35 +08:00
setOperationAction(ISD::SPLAT_VECTOR, VT, Expand);
[SDAG] Make the new zext-vector-inreg node default to expand so targets don't need to set it manually. This is based on feedback from Tom who pointed out that if every target needs to handle this we need to reach out to those maintainers. In fact, it doesn't make sense to duplicate everything when anything other than expand seems unlikely at this stage. llvm-svn: 212661
2014-07-10 06:53:04 +08:00
}
Introduce new @llvm.get.dynamic.area.offset.i{32, 64} intrinsics. The @llvm.get.dynamic.area.offset.* intrinsic family is used to get the offset from native stack pointer to the address of the most recent dynamic alloca on the caller's stack. These intrinsics are intendend for use in combination with @llvm.stacksave and @llvm.restore to get a pointer to the most recent dynamic alloca. This is useful, for example, for AddressSanitizer's stack unpoisoning routines. Patch by Max Ostapenko. Differential Revision: http://reviews.llvm.org/D14983 llvm-svn: 254404
2015-12-01 19:40:55 +08:00
Allow target to handle STRICT floating-point nodes The ISD::STRICT_ nodes used to implement the constrained floating-point intrinsics are currently never passed to the target back-end, which makes it impossible to handle them correctly (e.g. mark instructions are depending on a floating-point status and control register, or mark instructions as possibly trapping). This patch allows the target to use setOperationAction to switch the action on ISD::STRICT_ nodes to Legal. If this is done, the SelectionDAG common code will stop converting the STRICT nodes to regular floating-point nodes, but instead pass the STRICT nodes to the target using normal SelectionDAG matching rules. To avoid having the back-end duplicate all the floating-point instruction patterns to handle both strict and non-strict variants, we make the MI codegen explicitly aware of the floating-point exceptions by introducing two new concepts: - A new MCID flag "mayRaiseFPException" that the target should set on any instruction that possibly can raise FP exception according to the architecture definition. - A new MI flag FPExcept that CodeGen/SelectionDAG will set on any MI instruction resulting from expansion of any constrained FP intrinsic. Any MI instruction that is *both* marked as mayRaiseFPException *and* FPExcept then needs to be considered as raising exceptions by MI-level codegen (e.g. scheduling). Setting those two new flags is straightforward. The mayRaiseFPException flag is simply set via TableGen by marking all relevant instruction patterns in the .td files. The FPExcept flag is set in SDNodeFlags when creating the STRICT_ nodes in the SelectionDAG, and gets inherited in the MachineSDNode nodes created from it during instruction selection. The flag is then transfered to an MIFlag when creating the MI from the MachineSDNode. This is handled just like fast-math flags like no-nans are handled today. This patch includes both common code changes required to implement the new features, and the SystemZ implementation. Reviewed By: andrew.w.kaylor Differential Revision: https://reviews.llvm.org/D55506 llvm-svn: 362663
2019-06-06 06:33:10 +08:00
// Constrained floating-point operations default to expand.
[FPEnv] Divide macro INSTRUCTION into INSTRUCTION and DAG_INSTRUCTION, and macro FUNCTION likewise. NFCI. Some functions like fmuladd don't really have a node, we should divide the declaration form those have node to avoid introducing fake nodes. Differential Revision: https://reviews.llvm.org/D72871
2020-01-17 10:32:30 +08:00
#define DAG_INSTRUCTION(NAME, NARG, ROUND_MODE, INTRINSIC, DAGN) \
[FEnv] File with properties of constrained intrinsics Summary In several places we need to enumerate all constrained intrinsics or IR nodes that should be represented by them. It is easy to miss some of the cases. To make working with these intrinsics more convenient and robust, this change introduces file containing definitions of all constrained intrinsics and some of their properties. This file can be included to generate constrained intrinsics processing code. Reviewers: kpn, andrew.w.kaylor, cameron.mcinally, uweigand Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D69887
2019-11-05 21:42:16 +08:00
setOperationAction(ISD::STRICT_##DAGN, VT, Expand);
#include "llvm/IR/ConstrainedOps.def"
Allow target to handle STRICT floating-point nodes The ISD::STRICT_ nodes used to implement the constrained floating-point intrinsics are currently never passed to the target back-end, which makes it impossible to handle them correctly (e.g. mark instructions are depending on a floating-point status and control register, or mark instructions as possibly trapping). This patch allows the target to use setOperationAction to switch the action on ISD::STRICT_ nodes to Legal. If this is done, the SelectionDAG common code will stop converting the STRICT nodes to regular floating-point nodes, but instead pass the STRICT nodes to the target using normal SelectionDAG matching rules. To avoid having the back-end duplicate all the floating-point instruction patterns to handle both strict and non-strict variants, we make the MI codegen explicitly aware of the floating-point exceptions by introducing two new concepts: - A new MCID flag "mayRaiseFPException" that the target should set on any instruction that possibly can raise FP exception according to the architecture definition. - A new MI flag FPExcept that CodeGen/SelectionDAG will set on any MI instruction resulting from expansion of any constrained FP intrinsic. Any MI instruction that is *both* marked as mayRaiseFPException *and* FPExcept then needs to be considered as raising exceptions by MI-level codegen (e.g. scheduling). Setting those two new flags is straightforward. The mayRaiseFPException flag is simply set via TableGen by marking all relevant instruction patterns in the .td files. The FPExcept flag is set in SDNodeFlags when creating the STRICT_ nodes in the SelectionDAG, and gets inherited in the MachineSDNode nodes created from it during instruction selection. The flag is then transfered to an MIFlag when creating the MI from the MachineSDNode. This is handled just like fast-math flags like no-nans are handled today. This patch includes both common code changes required to implement the new features, and the SystemZ implementation. Reviewed By: andrew.w.kaylor Differential Revision: https://reviews.llvm.org/D55506 llvm-svn: 362663
2019-06-06 06:33:10 +08:00
[stack-protection] Add support for MSVC buffer security check Summary: This patch is adding support for the MSVC buffer security check implementation The buffer security check is turned on with the '/GS' compiler switch. * https://msdn.microsoft.com/en-us/library/8dbf701c.aspx * To be added to clang here: http://reviews.llvm.org/D20347 Some overview of buffer security check feature and implementation: * https://msdn.microsoft.com/en-us/library/aa290051(VS.71).aspx * http://www.ksyash.com/2011/01/buffer-overflow-protection-3/ * http://blog.osom.info/2012/02/understanding-vs-c-compilers-buffer.html For the following example: ``` int example(int offset, int index) { char buffer[10]; memset(buffer, 0xCC, index); return buffer[index]; } ``` The MSVC compiler is adding these instructions to perform stack integrity check: ``` push ebp mov ebp,esp sub esp,50h [1] mov eax,dword ptr [__security_cookie (01068024h)] [2] xor eax,ebp [3] mov dword ptr [ebp-4],eax push ebx push esi push edi mov eax,dword ptr [index] push eax push 0CCh lea ecx,[buffer] push ecx call _memset (010610B9h) add esp,0Ch mov eax,dword ptr [index] movsx eax,byte ptr buffer[eax] pop edi pop esi pop ebx [4] mov ecx,dword ptr [ebp-4] [5] xor ecx,ebp [6] call @__security_check_cookie@4 (01061276h) mov esp,ebp pop ebp ret ``` The instrumentation above is: * [1] is loading the global security canary, * [3] is storing the local computed ([2]) canary to the guard slot, * [4] is loading the guard slot and ([5]) re-compute the global canary, * [6] is validating the resulting canary with the '__security_check_cookie' and performs error handling. Overview of the current stack-protection implementation: * lib/CodeGen/StackProtector.cpp * There is a default stack-protection implementation applied on intermediate representation. * The target can overload 'getIRStackGuard' method if it has a standard location for the stack protector cookie. * An intrinsic 'Intrinsic::stackprotector' is added to the prologue. It will be expanded by the instruction selection pass (DAG or Fast). * Basic Blocks are added to every instrumented function to receive the code for handling stack guard validation and errors handling. * Guard manipulation and comparison are added directly to the intermediate representation. * lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp * lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp * There is an implementation that adds instrumentation during instruction selection (for better handling of sibbling calls). * see long comment above 'class StackProtectorDescriptor' declaration. * The target needs to override 'getSDagStackGuard' to activate SDAG stack protection generation. (note: getIRStackGuard MUST be nullptr). * 'getSDagStackGuard' returns the appropriate stack guard (security cookie) * The code is generated by 'SelectionDAGBuilder.cpp' and 'SelectionDAGISel.cpp'. * include/llvm/Target/TargetLowering.h * Contains function to retrieve the default Guard 'Value'; should be overriden by each target to select which implementation is used and provide Guard 'Value'. * lib/Target/X86/X86ISelLowering.cpp * Contains the x86 specialisation; Guard 'Value' used by the SelectionDAG algorithm. Function-based Instrumentation: * The MSVC doesn't inline the stack guard comparison in every function. Instead, a call to '__security_check_cookie' is added to the epilogue before every return instructions. * To support function-based instrumentation, this patch is * adding a function to get the function-based check (llvm 'Value', see include/llvm/Target/TargetLowering.h), * If provided, the stack protection instrumentation won't be inlined and a call to that function will be added to the prologue. * modifying (SelectionDAGISel.cpp) do avoid producing basic blocks used for inline instrumentation, * generating the function-based instrumentation during the ISEL pass (SelectionDAGBuilder.cpp), * if FastISEL (not SelectionDAG), using the fallback which rely on the same function-based implemented over intermediate representation (StackProtector.cpp). Modifications * adding support for MSVC (lib/Target/X86/X86ISelLowering.cpp) * adding support function-based instrumentation (lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp, .h) Results * IR generated instrumentation: ``` clang-cl /GS test.cc /Od /c -mllvm -print-isel-input ``` ``` *** Final LLVM Code input to ISel *** ; Function Attrs: nounwind sspstrong define i32 @"\01?example@@YAHHH@Z"(i32 %offset, i32 %index) #0 { entry: %StackGuardSlot = alloca i8* <<<-- Allocated guard slot %0 = call i8* @llvm.stackguard() <<<-- Loading Stack Guard value call void @llvm.stackprotector(i8* %0, i8** %StackGuardSlot) <<<-- Prologue intrinsic call (store to Guard slot) %index.addr = alloca i32, align 4 %offset.addr = alloca i32, align 4 %buffer = alloca [10 x i8], align 1 store i32 %index, i32* %index.addr, align 4 store i32 %offset, i32* %offset.addr, align 4 %arraydecay = getelementptr inbounds [10 x i8], [10 x i8]* %buffer, i32 0, i32 0 %1 = load i32, i32* %index.addr, align 4 call void @llvm.memset.p0i8.i32(i8* %arraydecay, i8 -52, i32 %1, i32 1, i1 false) %2 = load i32, i32* %index.addr, align 4 %arrayidx = getelementptr inbounds [10 x i8], [10 x i8]* %buffer, i32 0, i32 %2 %3 = load i8, i8* %arrayidx, align 1 %conv = sext i8 %3 to i32 %4 = load volatile i8*, i8** %StackGuardSlot <<<-- Loading Guard slot call void @__security_check_cookie(i8* %4) <<<-- Epilogue function-based check ret i32 %conv } ``` * SelectionDAG generated instrumentation: ``` clang-cl /GS test.cc /O1 /c /FA ``` ``` "?example@@YAHHH@Z": # @"\01?example@@YAHHH@Z" # BB#0: # %entry pushl %esi subl $16, %esp movl ___security_cookie, %eax <<<-- Loading Stack Guard value movl 28(%esp), %esi movl %eax, 12(%esp) <<<-- Store to Guard slot leal 2(%esp), %eax pushl %esi pushl $204 pushl %eax calll _memset addl $12, %esp movsbl 2(%esp,%esi), %esi movl 12(%esp), %ecx <<<-- Loading Guard slot calll @__security_check_cookie@4 <<<-- Epilogue function-based check movl %esi, %eax addl $16, %esp popl %esi retl ``` Reviewers: kcc, pcc, eugenis, rnk Subscribers: majnemer, llvm-commits, hans, thakis, rnk Differential Revision: http://reviews.llvm.org/D20346 llvm-svn: 272053
2016-06-08 04:15:35 +08:00
// For most targets @llvm.get.dynamic.area.offset just returns 0.
Introduce new @llvm.get.dynamic.area.offset.i{32, 64} intrinsics. The @llvm.get.dynamic.area.offset.* intrinsic family is used to get the offset from native stack pointer to the address of the most recent dynamic alloca on the caller's stack. These intrinsics are intendend for use in combination with @llvm.stacksave and @llvm.restore to get a pointer to the most recent dynamic alloca. This is useful, for example, for AddressSanitizer's stack unpoisoning routines. Patch by Max Ostapenko. Differential Revision: http://reviews.llvm.org/D14983 llvm-svn: 254404
2015-12-01 19:40:55 +08:00
setOperationAction(ISD::GET_DYNAMIC_AREA_OFFSET, VT, Expand);
[SDAG][AArch64] Legalize VECREDUCE Fixes https://bugs.llvm.org/show_bug.cgi?id=36796. Implement basic legalizations (PromoteIntRes, PromoteIntOp, ExpandIntRes, ScalarizeVecOp, WidenVecOp) for VECREDUCE opcodes. There are more legalizations missing (esp float legalizations), but there's no way to test them right now, so I'm not adding them. This also includes a few more changes to make this work somewhat reasonably: * Add support for expanding VECREDUCE in SDAG. Usually experimental.vector.reduce is expanded prior to codegen, but if the target does have native vector reduce, it may of course still be necessary to expand due to legalization issues. This uses a shuffle reduction if possible, followed by a naive scalar reduction. * Allow the result type of integer VECREDUCE to be larger than the vector element type. For example we need to be able to reduce a v8i8 into an (nominally) i32 result type on AArch64. * Use the vector operand type rather than the scalar result type to determine the action, so we can control exactly which vector types are supported. Also change the legalize vector op code to handle operations that only have vector operands, but no vector results, as is the case for VECREDUCE. * Default VECREDUCE to Expand. On AArch64 (only target using VECREDUCE), explicitly specify for which vector types the reductions are supported. This does not handle anything related to VECREDUCE_STRICT_*. Differential Revision: https://reviews.llvm.org/D58015 llvm-svn: 355860
2019-03-12 04:22:13 +08:00
// Vector reduction default to expand.
setOperationAction(ISD::VECREDUCE_FADD, VT, Expand);
setOperationAction(ISD::VECREDUCE_FMUL, VT, Expand);
setOperationAction(ISD::VECREDUCE_ADD, VT, Expand);
setOperationAction(ISD::VECREDUCE_MUL, VT, Expand);
setOperationAction(ISD::VECREDUCE_AND, VT, Expand);
setOperationAction(ISD::VECREDUCE_OR, VT, Expand);
setOperationAction(ISD::VECREDUCE_XOR, VT, Expand);
setOperationAction(ISD::VECREDUCE_SMAX, VT, Expand);
setOperationAction(ISD::VECREDUCE_SMIN, VT, Expand);
setOperationAction(ISD::VECREDUCE_UMAX, VT, Expand);
setOperationAction(ISD::VECREDUCE_UMIN, VT, Expand);
setOperationAction(ISD::VECREDUCE_FMAX, VT, Expand);
setOperationAction(ISD::VECREDUCE_FMIN, VT, Expand);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
// Most targets ignore the @llvm.prefetch intrinsic.
setOperationAction(ISD::PREFETCH, MVT::Other, Expand);
[CodeGen] Support (and default to) expanding READCYCLECOUNTER to 0. For targets that didn't support this, this will let us respect the langref instead of failing to select. Note that we don't need to change the 32-bit x86/PPC lowerings (to account for the result type/# difference) because they're both custom and bypass type legalization. llvm-svn: 246258
2015-08-28 09:49:59 +08:00
// Most targets also ignore the @llvm.readcyclecounter intrinsic.
setOperationAction(ISD::READCYCLECOUNTER, MVT::i64, Expand);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// ConstantFP nodes default to expand. Targets can either change this to
// Legal, in which case all fp constants are legal, or use isFPImmLegal()
// to optimize expansions for certain constants.
setOperationAction(ISD::ConstantFP, MVT::f16, Expand);
setOperationAction(ISD::ConstantFP, MVT::f32, Expand);
setOperationAction(ISD::ConstantFP, MVT::f64, Expand);
setOperationAction(ISD::ConstantFP, MVT::f80, Expand);
setOperationAction(ISD::ConstantFP, MVT::f128, Expand);
// These library functions default to expand.
Deduplicate a bunch of setOpActions into an MVT range-for. NFC. llvm-svn: 233330
2015-03-27 07:21:03 +08:00
for (MVT VT : {MVT::f32, MVT::f64, MVT::f128}) {
[DAGCombiner] try to convert pow(x, 1/3) to cbrt(x) This is a follow-up suggested in D51630 and originally proposed as an IR transform in D49040. Copying the motivational statement by @evandro from that patch: "This transformation helps some benchmarks in SPEC CPU2000 and CPU2006, such as 188.ammp, 447.dealII, 453.povray, and especially 300.twolf, as well as some proprietary benchmarks. Otherwise, no regressions on x86-64 or A64." I'm proposing to add only the minimum support for a DAG node here. Since we don't have an LLVM IR intrinsic for cbrt, and there are no other DAG ways to create a FCBRT node yet, I don't think we need to worry about DAG builder, legalization, a strict variant, etc. We should be able to expand as needed when adding more functionality/transforms. For reference, these are transform suggestions currently listed in SimplifyLibCalls.cpp: // * cbrt(expN(X)) -> expN(x/3) // * cbrt(sqrt(x)) -> pow(x,1/6) // * cbrt(cbrt(x)) -> pow(x,1/9) Also, given that we bail out on long double for now, there should not be any logical differences between platforms (unless there's some platform out there that has pow() but not cbrt()). Differential Revision: https://reviews.llvm.org/D51753 llvm-svn: 342348
2018-09-17 00:50:26 +08:00
setOperationAction(ISD::FCBRT, VT, Expand);
Deduplicate a bunch of setOpActions into an MVT range-for. NFC. llvm-svn: 233330
2015-03-27 07:21:03 +08:00
setOperationAction(ISD::FLOG , VT, Expand);
setOperationAction(ISD::FLOG2, VT, Expand);
setOperationAction(ISD::FLOG10, VT, Expand);
setOperationAction(ISD::FEXP , VT, Expand);
setOperationAction(ISD::FEXP2, VT, Expand);
setOperationAction(ISD::FFLOOR, VT, Expand);
setOperationAction(ISD::FNEARBYINT, VT, Expand);
setOperationAction(ISD::FCEIL, VT, Expand);
setOperationAction(ISD::FRINT, VT, Expand);
setOperationAction(ISD::FTRUNC, VT, Expand);
setOperationAction(ISD::FROUND, VT, Expand);
[FPEnv] Intrinsic llvm.roundeven This intrinsic implements IEEE-754 operation roundToIntegralTiesToEven, and performs rounding to the nearest integer value, rounding halfway cases to even. The intrinsic represents the missed case of IEEE-754 rounding operations and now llvm provides full support of the rounding operations defined by the standard. Differential Revision: https://reviews.llvm.org/D75670
2020-05-26 20:24:05 +08:00
setOperationAction(ISD::FROUNDEVEN, VT, Expand);
[CodeGen] Add lround/llround builtins This patch add the ISD::LROUND and ISD::LLROUND along with new intrinsics. The changes are straightforward as for other floating-point rounding functions, with just some adjustments required to handle the return value being an interger. The idea is to optimize lround/llround generation for AArch64 in a subsequent patch. Current semantic is just route it to libm symbol. llvm-svn: 360889
2019-05-16 21:15:27 +08:00
setOperationAction(ISD::LROUND, VT, Expand);
setOperationAction(ISD::LLROUND, VT, Expand);
[CodeGen] Add lrint/llrint builtins This patch add the ISD::LRINT and ISD::LLRINT along with new intrinsics. The changes are straightforward as for other floating-point rounding functions, with just some adjustments required to handle the return value being an interger. The idea is to optimize lrint/llrint generation for AArch64 in a subsequent patch. Current semantic is just route it to libm symbol. Reviewed By: craig.topper Differential Revision: https://reviews.llvm.org/D62017 llvm-svn: 361875
2019-05-29 04:47:44 +08:00
setOperationAction(ISD::LRINT, VT, Expand);
setOperationAction(ISD::LLRINT, VT, Expand);
Deduplicate a bunch of setOpActions into an MVT range-for. NFC. llvm-svn: 233330
2015-03-27 07:21:03 +08:00
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Default ISD::TRAP to expand (which turns it into abort).
setOperationAction(ISD::TRAP, MVT::Other, Expand);
// On most systems, DEBUGTRAP and TRAP have no difference. The "Expand"
// here is to inform DAG Legalizer to replace DEBUGTRAP with TRAP.
setOperationAction(ISD::DEBUGTRAP, MVT::Other, Expand);
}
Re-instate the EVT parameter to getScalarShiftAmountTy() for OOT user A documentation for this function would be nice by the way. From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241807
2015-07-09 23:12:23 +08:00
MVT TargetLoweringBase::getScalarShiftAmountTy(const DataLayout &DL,
EVT) const {
[CodeGen] Use "DL.getPointerSizeInBits" instead of "8 * DL.getPointerSize". NFC llvm-svn: 360315
2019-05-09 16:07:36 +08:00
return MVT::getIntegerVT(DL.getPointerSizeInBits(0));
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
[SelectionDAG] Add LegalTypes flag to getShiftAmountTy. Use it to unify and simplify DAGCombiner and simplifySetCC code and fix a bug. DAGCombiner and SimplifySetCC both use getPointerTy for shift amounts pre-legalization. DAGCombiner uses a single helper function to hide this. SimplifySetCC does it in multiple places. This patch adds a defaulted parameter to getShiftAmountTy that can make it return getPointerTy for scalar types. Use this parameter to simplify the SimplifySetCC and DAGCombiner. Additionally, there were two places in SimplifySetCC that were creating shifts using the target's preferred shift amount pre-legalization. If the target uses a narrow type and the type is illegal, this can cause SimplfiySetCC to create a shift with an amount that can't represent all possible shift values for the type. To fix this we should use pointer type there too. Alternatively we could make getScalarShiftAmountTy for each target return a safe value for large types as proposed in D43445. And maybe we should still do that, but fixing the SimplifySetCC code keeps other targets from tripping over this in the future. Fixes PR36250. Differential Revision: https://reviews.llvm.org/D43449 llvm-svn: 325602
2018-02-21 01:41:05 +08:00
EVT TargetLoweringBase::getShiftAmountTy(EVT LHSTy, const DataLayout &DL,
bool LegalTypes) const {
Fix PR10475 - ISD::SHL/SRL/SRA must have either both scalar or both vector operands but TLI.getShiftAmountTy() so far only return scalar type. As a result, backend logic assuming that breaks. - Rename the original TLI.getShiftAmountTy() to TLI.getScalarShiftAmountTy() and re-define TLI.getShiftAmountTy() to return target-specificed scalar type or the same vector type as the 1st operand. - Fix most TICG logic assuming TLI.getShiftAmountTy() a simple scalar type. llvm-svn: 176364
2013-03-02 02:40:30 +08:00
assert(LHSTy.isInteger() && "Shift amount is not an integer type!");
if (LHSTy.isVector())
return LHSTy;
[SelectionDAG] Add LegalTypes flag to getShiftAmountTy. Use it to unify and simplify DAGCombiner and simplifySetCC code and fix a bug. DAGCombiner and SimplifySetCC both use getPointerTy for shift amounts pre-legalization. DAGCombiner uses a single helper function to hide this. SimplifySetCC does it in multiple places. This patch adds a defaulted parameter to getShiftAmountTy that can make it return getPointerTy for scalar types. Use this parameter to simplify the SimplifySetCC and DAGCombiner. Additionally, there were two places in SimplifySetCC that were creating shifts using the target's preferred shift amount pre-legalization. If the target uses a narrow type and the type is illegal, this can cause SimplfiySetCC to create a shift with an amount that can't represent all possible shift values for the type. To fix this we should use pointer type there too. Alternatively we could make getScalarShiftAmountTy for each target return a safe value for large types as proposed in D43445. And maybe we should still do that, but fixing the SimplifySetCC code keeps other targets from tripping over this in the future. Fixes PR36250. Differential Revision: https://reviews.llvm.org/D43449 llvm-svn: 325602
2018-02-21 01:41:05 +08:00
return LegalTypes ? getScalarShiftAmountTy(DL, LHSTy)
: getPointerTy(DL);
Fix PR10475 - ISD::SHL/SRL/SRA must have either both scalar or both vector operands but TLI.getShiftAmountTy() so far only return scalar type. As a result, backend logic assuming that breaks. - Rename the original TLI.getShiftAmountTy() to TLI.getScalarShiftAmountTy() and re-define TLI.getShiftAmountTy() to return target-specificed scalar type or the same vector type as the 1st operand. - Fix most TICG logic assuming TLI.getShiftAmountTy() a simple scalar type. llvm-svn: 176364
2013-03-02 02:40:30 +08:00
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
bool TargetLoweringBase::canOpTrap(unsigned Op, EVT VT) const {
assert(isTypeLegal(VT));
switch (Op) {
default:
return false;
case ISD::SDIV:
case ISD::UDIV:
case ISD::SREM:
case ISD::UREM:
return true;
}
}
add a cl::opt override for TargetLoweringBase's JumpIsExpensive This patch is not intended to change existing codegen behavior for any target. It just exposes the JumpIsExpensive setting on the command-line to allow for easier testing and emergency overrides. Also, change the existing regression test to use FileCheck, explicitly specify the jump-is-expensive option, and use more precise checks. Differential Revision: http://reviews.llvm.org/D10846 llvm-svn: 241179
2015-07-02 02:10:20 +08:00
void TargetLoweringBase::setJumpIsExpensive(bool isExpensive) {
// If the command-line option was specified, ignore this request.
if (!JumpIsExpensiveOverride.getNumOccurrences())
JumpIsExpensive = isExpensive;
}
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
TargetLoweringBase::LegalizeKind
TargetLoweringBase::getTypeConversion(LLVMContext &Context, EVT VT) const {
// If this is a simple type, use the ComputeRegisterProp mechanism.
if (VT.isSimple()) {
MVT SVT = VT.getSimpleVT();
assert((unsigned)SVT.SimpleTy < array_lengthof(TransformToType));
MVT NVT = TransformToType[SVT.SimpleTy];
LegalizeTypeAction LA = ValueTypeActions.getTypeAction(SVT);
assert((LA == TypeLegal || LA == TypeSoftenFloat ||
[LegalizeTypes][X86] Add a new strategy for type legalizing f16 type that softens it to i16, but promotes to f32 around arithmetic ops. This is based on this llvm-dev thread http://lists.llvm.org/pipermail/llvm-dev/2019-December/137521.html The current strategy for f16 is to promote type to float every except where the specific width is required like loads, stores, and bitcasts. This results in rounding occurring in odd places instead of immediately after arithmetic operations. This interacts in weird ways with the __fp16 type in clang which is a storage only type where arithmetic is always promoted to float. InstCombine can remove some fpext/fptruncs around such arithmetic and turn it into arithmetic on half. This wouldn't be so bad if SelectionDAG was able to put those fpext/fpround back in when it promotes. It is also not obvious how to handle to make the existing strategy work with STRICT fp. We need to use STRICT versions of the conversions which require chain operands. But if the conversions are created for a bitcast, there is no place to get an appropriate chain from. This patch implements a different strategy where conversions are emitted directly around arithmetic operations. And otherwise its passed around as an i16 including in arguments and return values. This can result in more conversions between arithmetic operations, but is closer to matching the IR the frontend generates for __fp16. And it will allow us to use the chain from constrained arithmetic nodes to link the STRICT_FP_TO_FP16/STRICT_FP16_TO_FP that will need to be added. I've set it up so that each target can opt into the new behavior. Converting all the targets myself was more than I was able to handle. Differential Revision: https://reviews.llvm.org/D73749
2020-02-01 14:42:07 +08:00
LA == TypeSoftPromoteHalf ||
MVT: Add v3i16/v3f16 vectors AMDGPU has some buffer intrinsics which theoretically could use this. Some of the generated tables include the 3 and 4 element vector versions of these rounded to 64-bits, which is ambiguous. Add these to help the table disambiguate these. Assertion change is for the path odd sized vectors now take for R600. v3i16 is widened to v4i16, which then needs to be promoted to v4i32. llvm-svn: 369038
2019-08-16 02:58:25 +08:00
(NVT.isVector() ||
ValueTypeActions.getTypeAction(NVT) != TypePromoteInteger)) &&
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
"Promote may not follow Expand or Promote");
if (LA == TypeSplitVector)
Revert "[CodeGen][SVE] Calculate correct type legalization for scalable vectors." Seems to break some buildbots, reverting the patch for now. This reverts commit 164f4b9d26fdf3cd640a09b63b5ec44d033cbe8a.
2020-06-05 23:03:13 +08:00
return LegalizeKind(LA,
EVT::getVectorVT(Context, SVT.getVectorElementType(),
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
SVT.getVectorElementCount() / 2));
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
if (LA == TypeScalarizeVector)
return LegalizeKind(LA, SVT.getVectorElementType());
return LegalizeKind(LA, NVT);
}
// Handle Extended Scalar Types.
if (!VT.isVector()) {
assert(VT.isInteger() && "Float types must be simple");
unsigned BitSize = VT.getSizeInBits();
// First promote to a power-of-two size, then expand if necessary.
if (BitSize < 8 || !isPowerOf2_32(BitSize)) {
EVT NVT = VT.getRoundIntegerType(Context);
assert(NVT != VT && "Unable to round integer VT");
LegalizeKind NextStep = getTypeConversion(Context, NVT);
// Avoid multi-step promotion.
if (NextStep.first == TypePromoteInteger)
return NextStep;
// Return rounded integer type.
return LegalizeKind(TypePromoteInteger, NVT);
}
return LegalizeKind(TypeExpandInteger,
EVT::getIntegerVT(Context, VT.getSizeInBits() / 2));
}
// Handle vector types.
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
ElementCount NumElts = VT.getVectorElementCount();
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
EVT EltVT = VT.getVectorElementType();
// Vectors with only one element are always scalarized.
if (NumElts == 1)
return LegalizeKind(TypeScalarizeVector, EltVT);
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
if (VT.getVectorElementCount() == ElementCount(1, true))
report_fatal_error("Cannot legalize this vector");
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
// Try to widen vector elements until the element type is a power of two and
// promote it to a legal type later on, for example:
// <3 x i8> -> <4 x i8> -> <4 x i32>
if (EltVT.isInteger()) {
// Vectors with a number of elements that is not a power of two are always
// widened, for example <3 x i8> -> <4 x i8>.
if (!VT.isPow2VectorType()) {
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
NumElts = NumElts.NextPowerOf2();
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
EVT NVT = EVT::getVectorVT(Context, EltVT, NumElts);
return LegalizeKind(TypeWidenVector, NVT);
}
// Examine the element type.
LegalizeKind LK = getTypeConversion(Context, EltVT);
// If type is to be expanded, split the vector.
// <4 x i140> -> <2 x i140>
if (LK.first == TypeExpandInteger)
return LegalizeKind(TypeSplitVector,
EVT::getVectorVT(Context, EltVT, NumElts / 2));
// Promote the integer element types until a legal vector type is found
// or until the element integer type is too big. If a legal type was not
// found, fallback to the usual mechanism of widening/splitting the
// vector.
EVT OldEltVT = EltVT;
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
while (true) {
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
// Increase the bitwidth of the element to the next pow-of-two
// (which is greater than 8 bits).
EltVT = EVT::getIntegerVT(Context, 1 + EltVT.getSizeInBits())
.getRoundIntegerType(Context);
// Stop trying when getting a non-simple element type.
// Note that vector elements may be greater than legal vector element
// types. Example: X86 XMM registers hold 64bit element on 32bit
// systems.
if (!EltVT.isSimple())
break;
// Build a new vector type and check if it is legal.
MVT NVT = MVT::getVectorVT(EltVT.getSimpleVT(), NumElts);
// Found a legal promoted vector type.
if (NVT != MVT() && ValueTypeActions.getTypeAction(NVT) == TypeLegal)
return LegalizeKind(TypePromoteInteger,
EVT::getVectorVT(Context, EltVT, NumElts));
}
// Reset the type to the unexpanded type if we did not find a legal vector
// type with a promoted vector element type.
EltVT = OldEltVT;
}
// Try to widen the vector until a legal type is found.
// If there is no wider legal type, split the vector.
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
while (true) {
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
// Round up to the next power of 2.
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
NumElts = NumElts.NextPowerOf2();
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
// If there is no simple vector type with this many elements then there
// cannot be a larger legal vector type. Note that this assumes that
// there are no skipped intermediate vector types in the simple types.
if (!EltVT.isSimple())
break;
MVT LargerVector = MVT::getVectorVT(EltVT.getSimpleVT(), NumElts);
if (LargerVector == MVT())
break;
// If this type is legal then widen the vector.
if (ValueTypeActions.getTypeAction(LargerVector) == TypeLegal)
return LegalizeKind(TypeWidenVector, LargerVector);
}
// Widen odd vectors to next power of two.
if (!VT.isPow2VectorType()) {
EVT NVT = VT.getPow2VectorType(Context);
return LegalizeKind(TypeWidenVector, NVT);
}
// Vectors with illegal element types are expanded.
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
EVT NVT = EVT::getVectorVT(Context, EltVT, VT.getVectorElementCount() / 2);
Move TargetLoweringBase::getTypeConversion to the .cpp file from the .h file. It's used in only one place (other than recursively) and there's no need to include it everywhere. Saves almost 900k from total llvm object file size. llvm-svn: 230561
2015-02-26 06:41:30 +08:00
return LegalizeKind(TypeSplitVector, NVT);
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
static unsigned getVectorTypeBreakdownMVT(MVT VT, MVT &IntermediateVT,
unsigned &NumIntermediates,
MVT &RegisterVT,
TargetLoweringBase *TLI) {
// Figure out the right, legal destination reg to copy into.
[llvm][Codegen] Make `getVectorTypeBreakdownMVT` work with scalable types. Reviewers: efriedma, andwar, sdesmalen Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77434
2020-04-10 07:10:52 +08:00
ElementCount EC = VT.getVectorElementCount();
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
MVT EltTy = VT.getVectorElementType();
unsigned NumVectorRegs = 1;
[CodeGen][SVE] CopyToReg: Split scalable EVTs that are not powers of 2 Scalable vectors cannot use 'BUILD_VECTOR', so it is necessary to properly split and widen scalable vectors when passing them to CopyToReg/CopyFromReg. This functionality is added to TargetLoweringBase::getVectorTypeBreakdown(). This patch only adds support for 'splitting' scalable vectors that are a multiple of some legal type, e.g. <vscale x 6 x i64> -> 3 x <vscale x 2 x i64> Reviewers: efriedma, c-rhodes Reviewed By: efriedma Tags: #llvm Differential Revision: https://reviews.llvm.org/D80139
2020-06-06 01:29:43 +08:00
// Scalable vectors cannot be scalarized, so splitting or widening is
// required.
if (VT.isScalableVector() && !isPowerOf2_32(EC.Min))
llvm_unreachable(
"Splitting or widening of non-power-of-2 MVTs is not implemented.");
[llvm][Codegen] Make `getVectorTypeBreakdownMVT` work with scalable types. Reviewers: efriedma, andwar, sdesmalen Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77434
2020-04-10 07:10:52 +08:00
// FIXME: We don't support non-power-of-2-sized vectors for now.
// Ideally we could break down into LHS/RHS like LegalizeDAG does.
if (!isPowerOf2_32(EC.Min)) {
// Split EC to unit size (scalable property is preserved).
NumVectorRegs = EC.Min;
EC = EC / NumVectorRegs;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
[llvm][Codegen] Make `getVectorTypeBreakdownMVT` work with scalable types. Reviewers: efriedma, andwar, sdesmalen Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77434
2020-04-10 07:10:52 +08:00
// Divide the input until we get to a supported size. This will
// always end up with an EC that represent a scalar or a scalable
// scalar.
while (EC.Min > 1 && !TLI->isTypeLegal(MVT::getVectorVT(EltTy, EC))) {
EC.Min >>= 1;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
NumVectorRegs <<= 1;
}
NumIntermediates = NumVectorRegs;
[llvm][Codegen] Make `getVectorTypeBreakdownMVT` work with scalable types. Reviewers: efriedma, andwar, sdesmalen Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77434
2020-04-10 07:10:52 +08:00
MVT NewVT = MVT::getVectorVT(EltTy, EC);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
if (!TLI->isTypeLegal(NewVT))
NewVT = EltTy;
IntermediateVT = NewVT;
[llvm][Codegen] Make `getVectorTypeBreakdownMVT` work with scalable types. Reviewers: efriedma, andwar, sdesmalen Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77434
2020-04-10 07:10:52 +08:00
unsigned LaneSizeInBits = NewVT.getScalarSizeInBits().getFixedSize();
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Convert sizes such as i33 to i64.
[llvm][Codegen] Make `getVectorTypeBreakdownMVT` work with scalable types. Reviewers: efriedma, andwar, sdesmalen Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77434
2020-04-10 07:10:52 +08:00
if (!isPowerOf2_32(LaneSizeInBits))
LaneSizeInBits = NextPowerOf2(LaneSizeInBits);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
MVT DestVT = TLI->getRegisterType(NewVT);
RegisterVT = DestVT;
if (EVT(DestVT).bitsLT(NewVT)) // Value is expanded, e.g. i64 -> i16.
[llvm][Codegen] Make `getVectorTypeBreakdownMVT` work with scalable types. Reviewers: efriedma, andwar, sdesmalen Reviewed By: efriedma Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77434
2020-04-10 07:10:52 +08:00
return NumVectorRegs *
(LaneSizeInBits / DestVT.getScalarSizeInBits().getFixedSize());
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Otherwise, promotion or legal types use the same number of registers as
// the vector decimated to the appropriate level.
return NumVectorRegs;
}
/// isLegalRC - Return true if the value types that can be represented by the
/// specified register class are all legal.
Move value type list from TargetRegisterClass to TargetRegisterInfo Differential Revision: https://reviews.llvm.org/D31937 llvm-svn: 301234
2017-04-25 03:51:12 +08:00
bool TargetLoweringBase::isLegalRC(const TargetRegisterInfo &TRI,
const TargetRegisterClass &RC) const {
for (auto I = TRI.legalclasstypes_begin(RC); *I != MVT::Other; ++I)
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
if (isTypeLegal(*I))
return true;
return false;
}
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
/// Replace/modify any TargetFrameIndex operands with a targte-dependent
/// sequence of memory operands that is recognized by PrologEpilogInserter.
CodeGen: Use MachineInstr& in TargetLowering, NFC This is a mechanical change to make TargetLowering API take MachineInstr& (instead of MachineInstr*), since the argument is expected to be a valid MachineInstr. In one case, changed a parameter from MachineInstr* to MachineBasicBlock::iterator, since it was used as an insertion point. As a side effect, this removes a bunch of MachineInstr* to MachineBasicBlock::iterator implicit conversions, a necessary step toward fixing PR26753. llvm-svn: 274287
2016-07-01 06:52:52 +08:00
MachineBasicBlock *
TargetLoweringBase::emitPatchPoint(MachineInstr &InitialMI,
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
MachineBasicBlock *MBB) const {
CodeGen: Use MachineInstr& in TargetLowering, NFC This is a mechanical change to make TargetLowering API take MachineInstr& (instead of MachineInstr*), since the argument is expected to be a valid MachineInstr. In one case, changed a parameter from MachineInstr* to MachineBasicBlock::iterator, since it was used as an insertion point. As a side effect, this removes a bunch of MachineInstr* to MachineBasicBlock::iterator implicit conversions, a necessary step toward fixing PR26753. llvm-svn: 274287
2016-07-01 06:52:52 +08:00
MachineInstr *MI = &InitialMI;
CodeGen: Minor cleanups to use MachineInstr::getMF. NFC Since r315388 we have a shorter way to say this, so we'll replace MI->getParent()->getParent() with MI->getMF() in a few places. llvm-svn: 315390
2017-10-11 07:50:49 +08:00
MachineFunction &MF = *MI->getMF();
MachineFunction: Return reference for getFrameInfo(); NFC getFrameInfo() never returns nullptr so we should use a reference instead of a pointer. llvm-svn: 277017
2016-07-29 02:40:00 +08:00
MachineFrameInfo &MFI = MF.getFrameInfo();
[Statepoints] Use Indirect operands for spill slots Teach the statepoint lowering code to emit Indirect stackmap entries for spill inserted by StatepointLowering (i.e. SelectionDAG), but Direct stackmap entries for in-IR allocas which represent manual stack slots. This is what the docs call for (http://llvm.org/docs/StackMaps.html#stack-map-format), but we've been emitting both as Direct. This was pointed out recently on the mailing list as a bug. It also blocks http://reviews.llvm.org/D15632 which extends the lowering to handle vector-of-pointers since only Indirect references can encode a variable sized slot. To implement this, I introduced a new flag on the StackObject class used to maintian information about stack slots. I original considered (and prototyped in http://reviews.llvm.org/D15632), the idea of using the existing isSpillSlot flag, but end up deciding that was a bit too risky and that the cost of adding a new flag was low. Having the new flag will also allow us - in the future - to emit better comments in verbose assembly which indicate where a particular stack spill around a call comes from. (deopt, gc, regalloc). Differential Revision: http://reviews.llvm.org/D15759 llvm-svn: 256352
2015-12-24 07:44:28 +08:00
// We're handling multiple types of operands here:
// PATCHPOINT MetaArgs - live-in, read only, direct
// STATEPOINT Deopt Spill - live-through, read only, indirect
// STATEPOINT Deopt Alloca - live-through, read only, direct
// (We're currently conservative and mark the deopt slots read/write in
Remove trailing space sed -Ei 's/[[:space:]]+$//' include/**/*.{def,h,td} lib/**/*.{cpp,h} llvm-svn: 338293
2018-07-31 03:41:25 +08:00
// practice.)
[Statepoints] Use Indirect operands for spill slots Teach the statepoint lowering code to emit Indirect stackmap entries for spill inserted by StatepointLowering (i.e. SelectionDAG), but Direct stackmap entries for in-IR allocas which represent manual stack slots. This is what the docs call for (http://llvm.org/docs/StackMaps.html#stack-map-format), but we've been emitting both as Direct. This was pointed out recently on the mailing list as a bug. It also blocks http://reviews.llvm.org/D15632 which extends the lowering to handle vector-of-pointers since only Indirect references can encode a variable sized slot. To implement this, I introduced a new flag on the StackObject class used to maintian information about stack slots. I original considered (and prototyped in http://reviews.llvm.org/D15632), the idea of using the existing isSpillSlot flag, but end up deciding that was a bit too risky and that the cost of adding a new flag was low. Having the new flag will also allow us - in the future - to emit better comments in verbose assembly which indicate where a particular stack spill around a call comes from. (deopt, gc, regalloc). Differential Revision: http://reviews.llvm.org/D15759 llvm-svn: 256352
2015-12-24 07:44:28 +08:00
// STATEPOINT GC Spill - live-through, read/write, indirect
// STATEPOINT GC Alloca - live-through, read/write, direct
// The live-in vs live-through is handled already (the live through ones are
// all stack slots), but we need to handle the different type of stackmap
// operands and memory effects here.
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
[TargetLowering][NFC] More efficient emitPatchpoint(). Current implementation of emitPatchpoint() is very inefficient: for every FrameIndex operand if creates new MachineInstr with that operand expanded and all other copied as is. Since PATCHPOINT/STATEPOINT instructions may have *a lot* of FrameIndex operands, we end up creating and erasing many machine instructions. But we can do it in single pass, with only one new machine instruction generated. Reviewed By: reames Differential Revision: https://reviews.llvm.org/D81181
2020-06-05 00:11:13 +08:00
if (!llvm::any_of(MI->operands(),
[](MachineOperand &Operand) { return Operand.isFI(); }))
return MBB;
MachineInstrBuilder MIB = BuildMI(MF, MI->getDebugLoc(), MI->getDesc());
// Inherit previous memory operands.
MIB.cloneMemRefs(*MI);
for (auto &MO : MI->operands()) {
if (!MO.isFI()) {
Fix build breakage caused by 66a1b83bf93ec46f6d7a06c47d5981ae154f9ea0
2020-06-05 20:52:14 +08:00
MIB.add(MO);
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
continue;
[TargetLowering][NFC] More efficient emitPatchpoint(). Current implementation of emitPatchpoint() is very inefficient: for every FrameIndex operand if creates new MachineInstr with that operand expanded and all other copied as is. Since PATCHPOINT/STATEPOINT instructions may have *a lot* of FrameIndex operands, we end up creating and erasing many machine instructions. But we can do it in single pass, with only one new machine instruction generated. Reviewed By: reames Differential Revision: https://reviews.llvm.org/D81181
2020-06-05 00:11:13 +08:00
}
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
// foldMemoryOperand builds a new MI after replacing a single FI operand
// with the canonical set of five x86 addressing-mode operands.
int FI = MO.getIndex();
[Statepoints] Use Indirect operands for spill slots Teach the statepoint lowering code to emit Indirect stackmap entries for spill inserted by StatepointLowering (i.e. SelectionDAG), but Direct stackmap entries for in-IR allocas which represent manual stack slots. This is what the docs call for (http://llvm.org/docs/StackMaps.html#stack-map-format), but we've been emitting both as Direct. This was pointed out recently on the mailing list as a bug. It also blocks http://reviews.llvm.org/D15632 which extends the lowering to handle vector-of-pointers since only Indirect references can encode a variable sized slot. To implement this, I introduced a new flag on the StackObject class used to maintian information about stack slots. I original considered (and prototyped in http://reviews.llvm.org/D15632), the idea of using the existing isSpillSlot flag, but end up deciding that was a bit too risky and that the cost of adding a new flag was low. Having the new flag will also allow us - in the future - to emit better comments in verbose assembly which indicate where a particular stack spill around a call comes from. (deopt, gc, regalloc). Differential Revision: http://reviews.llvm.org/D15759 llvm-svn: 256352
2015-12-24 07:44:28 +08:00
// Add frame index operands recognized by stackmaps.cpp
if (MFI.isStatepointSpillSlotObjectIndex(FI)) {
// indirect-mem-ref tag, size, #FI, offset.
// Used for spills inserted by StatepointLowering. This codepath is not
// used for patchpoints/stackmaps at all, for these spilling is done via
// foldMemoryOperand callback only.
assert(MI->getOpcode() == TargetOpcode::STATEPOINT && "sanity");
MIB.addImm(StackMaps::IndirectMemRefOp);
MIB.addImm(MFI.getObjectSize(FI));
Fix build breakage caused by 66a1b83bf93ec46f6d7a06c47d5981ae154f9ea0
2020-06-05 20:52:14 +08:00
MIB.add(MO);
[Statepoints] Use Indirect operands for spill slots Teach the statepoint lowering code to emit Indirect stackmap entries for spill inserted by StatepointLowering (i.e. SelectionDAG), but Direct stackmap entries for in-IR allocas which represent manual stack slots. This is what the docs call for (http://llvm.org/docs/StackMaps.html#stack-map-format), but we've been emitting both as Direct. This was pointed out recently on the mailing list as a bug. It also blocks http://reviews.llvm.org/D15632 which extends the lowering to handle vector-of-pointers since only Indirect references can encode a variable sized slot. To implement this, I introduced a new flag on the StackObject class used to maintian information about stack slots. I original considered (and prototyped in http://reviews.llvm.org/D15632), the idea of using the existing isSpillSlot flag, but end up deciding that was a bit too risky and that the cost of adding a new flag was low. Having the new flag will also allow us - in the future - to emit better comments in verbose assembly which indicate where a particular stack spill around a call comes from. (deopt, gc, regalloc). Differential Revision: http://reviews.llvm.org/D15759 llvm-svn: 256352
2015-12-24 07:44:28 +08:00
MIB.addImm(0);
} else {
// direct-mem-ref tag, #FI, offset.
// Used by patchpoint, and direct alloca arguments to statepoints
MIB.addImm(StackMaps::DirectMemRefOp);
Fix build breakage caused by 66a1b83bf93ec46f6d7a06c47d5981ae154f9ea0
2020-06-05 20:52:14 +08:00
MIB.add(MO);
[Statepoints] Use Indirect operands for spill slots Teach the statepoint lowering code to emit Indirect stackmap entries for spill inserted by StatepointLowering (i.e. SelectionDAG), but Direct stackmap entries for in-IR allocas which represent manual stack slots. This is what the docs call for (http://llvm.org/docs/StackMaps.html#stack-map-format), but we've been emitting both as Direct. This was pointed out recently on the mailing list as a bug. It also blocks http://reviews.llvm.org/D15632 which extends the lowering to handle vector-of-pointers since only Indirect references can encode a variable sized slot. To implement this, I introduced a new flag on the StackObject class used to maintian information about stack slots. I original considered (and prototyped in http://reviews.llvm.org/D15632), the idea of using the existing isSpillSlot flag, but end up deciding that was a bit too risky and that the cost of adding a new flag was low. Having the new flag will also allow us - in the future - to emit better comments in verbose assembly which indicate where a particular stack spill around a call comes from. (deopt, gc, regalloc). Differential Revision: http://reviews.llvm.org/D15759 llvm-svn: 256352
2015-12-24 07:44:28 +08:00
MIB.addImm(0);
}
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
assert(MIB->mayLoad() && "Folded a stackmap use to a non-load!");
// Add a new memory operand for this FI.
assert(MFI.getObjectOffset(FI) != -1);
[Statepoints 2/4] Statepoint infrastructure for garbage collection: MI & x86-64 Backend This is the second patch in a small series. This patch contains the MachineInstruction and x86-64 backend pieces required to lower Statepoints. It does not include the code to actually generate the STATEPOINT machine instruction and as a result, the entire patch is currently dead code. I will be submitting the SelectionDAG parts within the next 24-48 hours. Since those pieces are by far the most complicated, I wanted to minimize the size of that patch. That patch will include the tests which exercise the functionality in this patch. The entire series can be seen as one combined whole in http://reviews.llvm.org/D5683. The STATEPOINT psuedo node is generated after all gc values are explicitly spilled to stack slots. The purpose of this node is to wrap an actual call instruction while recording the spill locations of the meta arguments used for garbage collection and other purposes. The STATEPOINT is modeled as modifing all of those locations to prevent backend optimizations from forwarding the value from before the STATEPOINT to after the STATEPOINT. (Doing so would break relocation semantics for collectors which wish to relocate roots.) The implementation of STATEPOINT is closely modeled on PATCHPOINT. Eventually, much of the code in this patch will be removed. The long term plan is to merge the functionality provided by statepoints and patchpoints. Merging their implementations in the backend is likely to be a good starting point. Reviewed by: atrick, ributzka llvm-svn: 223085
2014-12-02 06:52:56 +08:00
[CodeGen] Add MMOs to statepoint nodes during SelectionDAG The existing statepoint lowering code does something odd; it adds machine memory operands post instruction selection. This was copied from the stackmap/patchpoint implementation, but appears to be non-idiomatic. This change is largely NFC. It moves the MMO creation logic into SelectionDAG building. It ends up not quite being NFC because the size of the stack slot is reflected in the MMO. The old code blindly used pointer size for the MMO size, which appears to have always been incorrect for larger values. It just happened nothing actually relied on the MMOs, so it worked out okay. For context, I'm planning on removing the MOVolatile flag from these in a future commit, and then removing the MOStore flag from deopt spill slots in a separate one. Doing so is motivated by a small test case where we should be able to better schedule spill slots, but don't do so due to a memory use/def implied by the statepoint. Differential Revision: https://reviews.llvm.org/D59106 llvm-svn: 355953
2019-03-13 03:12:33 +08:00
// Note: STATEPOINT MMOs are added during SelectionDAG. STACKMAP, and
// PATCHPOINT should be updated to do the same. (TODO)
if (MI->getOpcode() != TargetOpcode::STATEPOINT) {
auto Flags = MachineMemOperand::MOLoad;
MachineMemOperand *MMO = MF.getMachineMemOperand(
MachinePointerInfo::getFixedStack(MF, FI), Flags,
[Alignment][NFC] Transitionning more getMachineMemOperand call sites Summary: This is patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Reviewers: courbet Subscribers: hiraditya, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77127
2020-03-31 17:43:50 +08:00
MF.getDataLayout().getPointerSize(), MFI.getObjectAlign(FI));
[CodeGen] Add MMOs to statepoint nodes during SelectionDAG The existing statepoint lowering code does something odd; it adds machine memory operands post instruction selection. This was copied from the stackmap/patchpoint implementation, but appears to be non-idiomatic. This change is largely NFC. It moves the MMO creation logic into SelectionDAG building. It ends up not quite being NFC because the size of the stack slot is reflected in the MMO. The old code blindly used pointer size for the MMO size, which appears to have always been incorrect for larger values. It just happened nothing actually relied on the MMOs, so it worked out okay. For context, I'm planning on removing the MOVolatile flag from these in a future commit, and then removing the MOStore flag from deopt spill slots in a separate one. Doing so is motivated by a small test case where we should be able to better schedule spill slots, but don't do so due to a memory use/def implied by the statepoint. Differential Revision: https://reviews.llvm.org/D59106 llvm-svn: 355953
2019-03-13 03:12:33 +08:00
MIB->addMemOperand(MF, MMO);
[Statepoints 2/4] Statepoint infrastructure for garbage collection: MI & x86-64 Backend This is the second patch in a small series. This patch contains the MachineInstruction and x86-64 backend pieces required to lower Statepoints. It does not include the code to actually generate the STATEPOINT machine instruction and as a result, the entire patch is currently dead code. I will be submitting the SelectionDAG parts within the next 24-48 hours. Since those pieces are by far the most complicated, I wanted to minimize the size of that patch. That patch will include the tests which exercise the functionality in this patch. The entire series can be seen as one combined whole in http://reviews.llvm.org/D5683. The STATEPOINT psuedo node is generated after all gc values are explicitly spilled to stack slots. The purpose of this node is to wrap an actual call instruction while recording the spill locations of the meta arguments used for garbage collection and other purposes. The STATEPOINT is modeled as modifing all of those locations to prevent backend optimizations from forwarding the value from before the STATEPOINT to after the STATEPOINT. (Doing so would break relocation semantics for collectors which wish to relocate roots.) The implementation of STATEPOINT is closely modeled on PATCHPOINT. Eventually, much of the code in this patch will be removed. The long term plan is to merge the functionality provided by statepoints and patchpoints. Merging their implementations in the backend is likely to be a good starting point. Reviewed by: atrick, ributzka llvm-svn: 223085
2014-12-02 06:52:56 +08:00
}
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
}
[TargetLowering][NFC] More efficient emitPatchpoint(). Current implementation of emitPatchpoint() is very inefficient: for every FrameIndex operand if creates new MachineInstr with that operand expanded and all other copied as is. Since PATCHPOINT/STATEPOINT instructions may have *a lot* of FrameIndex operands, we end up creating and erasing many machine instructions. But we can do it in single pass, with only one new machine instruction generated. Reviewed By: reames Differential Revision: https://reviews.llvm.org/D81181
2020-06-05 00:11:13 +08:00
MBB->insert(MachineBasicBlock::iterator(MI), MIB);
MI->eraseFromParent();
Refactor a lot of patchpoint/stackmap related code to simplify and make it target independent. Most of the x86 specific stackmap/patchpoint handling was necessitated by the use of the native address-mode format for frame index operands. PEI has now been modified to treat stackmap/patchpoint similarly to DEBUG_INFO, allowing us to use a simple, platform independent register/offset pair for frame indexes on stackmap/patchpoints. Notes: - Folding is now platform independent and automatically supported. - Emiting patchpoints with direct memory references now just involves calling the TargetLoweringBase::emitPatchPoint utility method from the target's XXXTargetLowering::EmitInstrWithCustomInserter method. (See X86TargetLowering for an example). - No more ugly platform-specific operand parsers. This patch shouldn't change the generated output for X86. llvm-svn: 195944
2013-11-29 11:07:54 +08:00
return MBB;
}
[XRay][compiler-rt+llvm] Update XRay register stashing semantics Summary: This change expands the amount of registers stashed by the entry and `__xray_CustomEvent` trampolines. We've found that since the `__xray_CustomEvent` trampoline calls can show up in situations where the scratch registers are being used, and since we don't typically want to affect the code-gen around the disabled `__xray_customevent(...)` intrinsic calls, that we need to save and restore the state of even the scratch registers in the handling of these custom events. Reviewers: pcc, pelikan, dblaikie, eizan, kpw, echristo, chandlerc Reviewed By: echristo Subscribers: chandlerc, echristo, hiraditya, davide, dblaikie, llvm-commits Differential Revision: https://reviews.llvm.org/D40894 llvm-svn: 323940
2018-02-01 10:21:54 +08:00
MachineBasicBlock *
TargetLoweringBase::emitXRayCustomEvent(MachineInstr &MI,
MachineBasicBlock *MBB) const {
assert(MI.getOpcode() == TargetOpcode::PATCHABLE_EVENT_CALL &&
"Called emitXRayCustomEvent on the wrong MI!");
auto &MF = *MI.getMF();
auto MIB = BuildMI(MF, MI.getDebugLoc(), MI.getDesc());
for (unsigned OpIdx = 0; OpIdx != MI.getNumOperands(); ++OpIdx)
MIB.add(MI.getOperand(OpIdx));
MBB->insert(MachineBasicBlock::iterator(MI), MIB);
MI.eraseFromParent();
[XRay] Typed event logging intrinsic Summary: Add an LLVM intrinsic for type discriminated event logging with XRay. Similar to the existing intrinsic for custom events, but also accepts a type tag argument to allow plugins to be aware of different types and semantically interpret logged events they know about without choking on those they don't. Relies on a symbol defined in compiler-rt patch D43668. I may wait to submit before I can see demo everything working together including a still to come clang patch. Reviewers: dberris, pelikan, eizan, rSerge, timshen Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D45633 llvm-svn: 330219
2018-04-18 05:30:29 +08:00
return MBB;
}
MachineBasicBlock *
TargetLoweringBase::emitXRayTypedEvent(MachineInstr &MI,
MachineBasicBlock *MBB) const {
assert(MI.getOpcode() == TargetOpcode::PATCHABLE_TYPED_EVENT_CALL &&
"Called emitXRayTypedEvent on the wrong MI!");
auto &MF = *MI.getMF();
auto MIB = BuildMI(MF, MI.getDebugLoc(), MI.getDesc());
for (unsigned OpIdx = 0; OpIdx != MI.getNumOperands(); ++OpIdx)
MIB.add(MI.getOperand(OpIdx));
MBB->insert(MachineBasicBlock::iterator(MI), MIB);
MI.eraseFromParent();
[XRay][compiler-rt+llvm] Update XRay register stashing semantics Summary: This change expands the amount of registers stashed by the entry and `__xray_CustomEvent` trampolines. We've found that since the `__xray_CustomEvent` trampoline calls can show up in situations where the scratch registers are being used, and since we don't typically want to affect the code-gen around the disabled `__xray_customevent(...)` intrinsic calls, that we need to save and restore the state of even the scratch registers in the handling of these custom events. Reviewers: pcc, pelikan, dblaikie, eizan, kpw, echristo, chandlerc Reviewed By: echristo Subscribers: chandlerc, echristo, hiraditya, davide, dblaikie, llvm-commits Differential Revision: https://reviews.llvm.org/D40894 llvm-svn: 323940
2018-02-01 10:21:54 +08:00
return MBB;
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
/// findRepresentativeClass - Return the largest legal super-reg register class
/// of the register class for the specified type and its associated "cost".
Add a comment above findRepresentativeClass explaining why it's where it is so that future generations can understand. llvm-svn: 231111
2015-03-04 03:47:14 +08:00
// This function is in TargetLowering because it uses RegClassForVT which would
// need to be moved to TargetRegisterInfo and would necessitate moving
// isTypeLegal over as well - a massive change that would just require
// TargetLowering having a TargetRegisterInfo class member that it would use.
Remove an argument-less call to getSubtargetImpl from TargetLoweringBase. This required plumbing a TargetRegisterInfo through computeRegisterProperties and into findRepresentativeClass which uses it for register class iteration. This required passing a subtarget into a few target specific initializations of TargetLowering. llvm-svn: 230583
2015-02-26 08:00:24 +08:00
std::pair<const TargetRegisterClass *, uint8_t>
TargetLoweringBase::findRepresentativeClass(const TargetRegisterInfo *TRI,
MVT VT) const {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
const TargetRegisterClass *RC = RegClassForVT[VT.SimpleTy];
if (!RC)
return std::make_pair(RC, 0);
// Compute the set of all super-register classes.
BitVector SuperRegRC(TRI->getNumRegClasses());
for (SuperRegClassIterator RCI(RC, TRI); RCI.isValid(); ++RCI)
SuperRegRC.setBitsInMask(RCI.getMask());
// Find the first legal register class with the largest spill size.
const TargetRegisterClass *BestRC = RC;
BitVector: add iterators for set bits Differential revision: https://reviews.llvm.org/D32060 llvm-svn: 303227
2017-05-17 09:07:53 +08:00
for (unsigned i : SuperRegRC.set_bits()) {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
const TargetRegisterClass *SuperRC = TRI->getRegClass(i);
// We want the largest possible spill size.
Move size and alignment information of regclass to TargetRegisterInfo 1. RegisterClass::getSize() is split into two functions: - TargetRegisterInfo::getRegSizeInBits(const TargetRegisterClass &RC) const; - TargetRegisterInfo::getSpillSize(const TargetRegisterClass &RC) const; 2. RegisterClass::getAlignment() is replaced by: - TargetRegisterInfo::getSpillAlignment(const TargetRegisterClass &RC) const; This will allow making those values depend on subtarget features in the future. Differential Revision: https://reviews.llvm.org/D31783 llvm-svn: 301221
2017-04-25 02:55:33 +08:00
if (TRI->getSpillSize(*SuperRC) <= TRI->getSpillSize(*BestRC))
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
continue;
Move value type list from TargetRegisterClass to TargetRegisterInfo Differential Revision: https://reviews.llvm.org/D31937 llvm-svn: 301234
2017-04-25 03:51:12 +08:00
if (!isLegalRC(*TRI, *SuperRC))
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
continue;
BestRC = SuperRC;
}
return std::make_pair(BestRC, 1);
}
/// computeRegisterProperties - Once all of the register classes are added,
/// this allows us to compute derived properties we expose.
Remove an argument-less call to getSubtargetImpl from TargetLoweringBase. This required plumbing a TargetRegisterInfo through computeRegisterProperties and into findRepresentativeClass which uses it for register class iteration. This required passing a subtarget into a few target specific initializations of TargetLowering. llvm-svn: 230583
2015-02-26 08:00:24 +08:00
void TargetLoweringBase::computeRegisterProperties(
const TargetRegisterInfo *TRI) {
Replace a couple asserts with static_asserts. llvm-svn: 222114
2014-11-17 08:26:50 +08:00
static_assert(MVT::LAST_VALUETYPE <= MVT::MAX_ALLOWED_VALUETYPE,
"Too many value types for ValueTypeActions to hold!");
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Everything defaults to needing one register.
for (unsigned i = 0; i != MVT::LAST_VALUETYPE; ++i) {
NumRegistersForVT[i] = 1;
RegisterTypeForVT[i] = TransformToType[i] = (MVT::SimpleValueType)i;
}
// ...except isVoid, which doesn't need any registers.
NumRegistersForVT[MVT::isVoid] = 0;
// Find the largest integer register class.
unsigned LargestIntReg = MVT::LAST_INTEGER_VALUETYPE;
[C++11] More 'nullptr' conversion. In some cases just using a boolean check instead of comparing to nullptr. llvm-svn: 206142
2014-04-14 08:51:57 +08:00
for (; RegClassForVT[LargestIntReg] == nullptr; --LargestIntReg)
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
assert(LargestIntReg != MVT::i1 && "No integer registers defined!");
// Every integer value type larger than this largest register takes twice as
// many registers to represent as the previous ValueType.
for (unsigned ExpandedReg = LargestIntReg + 1;
ExpandedReg <= MVT::LAST_INTEGER_VALUETYPE; ++ExpandedReg) {
NumRegistersForVT[ExpandedReg] = 2*NumRegistersForVT[ExpandedReg-1];
RegisterTypeForVT[ExpandedReg] = (MVT::SimpleValueType)LargestIntReg;
TransformToType[ExpandedReg] = (MVT::SimpleValueType)(ExpandedReg - 1);
ValueTypeActions.setTypeAction((MVT::SimpleValueType)ExpandedReg,
TypeExpandInteger);
}
// Inspect all of the ValueType's smaller than the largest integer
// register to see which ones need promotion.
unsigned LegalIntReg = LargestIntReg;
for (unsigned IntReg = LargestIntReg - 1;
IntReg >= (unsigned)MVT::i1; --IntReg) {
MVT IVT = (MVT::SimpleValueType)IntReg;
if (isTypeLegal(IVT)) {
LegalIntReg = IntReg;
} else {
RegisterTypeForVT[IntReg] = TransformToType[IntReg] =
Fix ignorded type qualifier warning [NFC] llvm-svn: 346332
2018-11-08 00:17:30 +08:00
(MVT::SimpleValueType)LegalIntReg;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
ValueTypeActions.setTypeAction(IVT, TypePromoteInteger);
}
}
// ppcf128 type is really two f64's.
if (!isTypeLegal(MVT::ppcf128)) {
[Power PC] softening long double type This patch implements softening of long double type (ppcf128) on ppc32 architecture and enables operations for this type for soft float. Patch by Strahinja Petrovic. Differential Revision: http://reviews.llvm.org/D15811 llvm-svn: 259791
2016-02-04 22:43:50 +08:00
if (isTypeLegal(MVT::f64)) {
NumRegistersForVT[MVT::ppcf128] = 2*NumRegistersForVT[MVT::f64];
RegisterTypeForVT[MVT::ppcf128] = MVT::f64;
TransformToType[MVT::ppcf128] = MVT::f64;
ValueTypeActions.setTypeAction(MVT::ppcf128, TypeExpandFloat);
} else {
NumRegistersForVT[MVT::ppcf128] = NumRegistersForVT[MVT::i128];
RegisterTypeForVT[MVT::ppcf128] = RegisterTypeForVT[MVT::i128];
TransformToType[MVT::ppcf128] = MVT::i128;
ValueTypeActions.setTypeAction(MVT::ppcf128, TypeSoftenFloat);
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
Set properties for f128 type. llvm-svn: 176378
2013-03-02 05:11:44 +08:00
// Decide how to handle f128. If the target does not have native f128 support,
// expand it to i128 and we will be generating soft float library calls.
if (!isTypeLegal(MVT::f128)) {
NumRegistersForVT[MVT::f128] = NumRegistersForVT[MVT::i128];
RegisterTypeForVT[MVT::f128] = RegisterTypeForVT[MVT::i128];
TransformToType[MVT::f128] = MVT::i128;
ValueTypeActions.setTypeAction(MVT::f128, TypeSoftenFloat);
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Decide how to handle f64. If the target does not have native f64 support,
// expand it to i64 and we will be generating soft float library calls.
if (!isTypeLegal(MVT::f64)) {
NumRegistersForVT[MVT::f64] = NumRegistersForVT[MVT::i64];
RegisterTypeForVT[MVT::f64] = RegisterTypeForVT[MVT::i64];
TransformToType[MVT::f64] = MVT::i64;
ValueTypeActions.setTypeAction(MVT::f64, TypeSoftenFloat);
}
[CodeGen] "PromoteInteger" f32 to f64 doesn't make sense. The original f32->f64 promotion logic was refactored into roughly the currently shape in r37781. However, starting with r132263, the legalizer has been split into different kinds, and the previous "Promote" (which did the right thing) was search-and-replace'd into "PromoteInteger". The divide gradually deepened, with type legalization ("PromoteInteger") being separated from ops legalization ("Promote", which still works for floating point ops). Fast-forward to today: there's no in-tree target with legal f64 but illegal f32 (rather: no tests were harmed in the making of this patch). With such a target, i.e., if you trick the legalizer into going through the PromoteInteger path for FP, you get the expected brokenness. For instance, there's no PromoteIntRes_FADD (the name itself sounds wrong), so we'll just hit some assert in the PromoteInteger path. Don't pretend we can promote f32 to f64. Instead, always soften. llvm-svn: 233464
2015-03-28 09:22:37 +08:00
// Decide how to handle f32. If the target does not have native f32 support,
// expand it to i32 and we will be generating soft float library calls.
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
if (!isTypeLegal(MVT::f32)) {
[CodeGen] "PromoteInteger" f32 to f64 doesn't make sense. The original f32->f64 promotion logic was refactored into roughly the currently shape in r37781. However, starting with r132263, the legalizer has been split into different kinds, and the previous "Promote" (which did the right thing) was search-and-replace'd into "PromoteInteger". The divide gradually deepened, with type legalization ("PromoteInteger") being separated from ops legalization ("Promote", which still works for floating point ops). Fast-forward to today: there's no in-tree target with legal f64 but illegal f32 (rather: no tests were harmed in the making of this patch). With such a target, i.e., if you trick the legalizer into going through the PromoteInteger path for FP, you get the expected brokenness. For instance, there's no PromoteIntRes_FADD (the name itself sounds wrong), so we'll just hit some assert in the PromoteInteger path. Don't pretend we can promote f32 to f64. Instead, always soften. llvm-svn: 233464
2015-03-28 09:22:37 +08:00
NumRegistersForVT[MVT::f32] = NumRegistersForVT[MVT::i32];
RegisterTypeForVT[MVT::f32] = RegisterTypeForVT[MVT::i32];
TransformToType[MVT::f32] = MVT::i32;
ValueTypeActions.setTypeAction(MVT::f32, TypeSoftenFloat);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
[CodeGen] Always promote f16 if not legal We don't currently have any runtime library functions for operations on f16 values (other than conversions to and from f32 and f64), so we should always promote it to f32, even if that is not a legal type. In that case, the f32 values would be softened to f32 library calls. SoftenFloatRes_FP_EXTEND now needs to check the promoted operand's type, as it may ne a no-op or require a different library call. getCopyFromParts and getCopyToParts now need to cope with a floating-point value stored in a larger integer part, as is the case for any target that needs to store an f16 value in a 32-bit integer register. Differential Revision: http://reviews.llvm.org/D12856 llvm-svn: 252459
2015-11-09 19:03:18 +08:00
// Decide how to handle f16. If the target does not have native f16 support,
// promote it to f32, because there are no f16 library calls (except for
// conversions).
CodeGen: soften f16 type by default instead of marking legal. Actual support for softening f16 operations is still limited, and can be added when it's needed. But Soften is much closer to being a useful thing to try than keeping it Legal when no registers can actually hold such values. Longer term, we probably want something between Soften and Promote semantics for most targets, it'll be more efficient to promote the 4 basic operations to f32 than libcall them. llvm-svn: 213372
2014-07-18 20:41:46 +08:00
if (!isTypeLegal(MVT::f16)) {
[LegalizeTypes][X86] Add a new strategy for type legalizing f16 type that softens it to i16, but promotes to f32 around arithmetic ops. This is based on this llvm-dev thread http://lists.llvm.org/pipermail/llvm-dev/2019-December/137521.html The current strategy for f16 is to promote type to float every except where the specific width is required like loads, stores, and bitcasts. This results in rounding occurring in odd places instead of immediately after arithmetic operations. This interacts in weird ways with the __fp16 type in clang which is a storage only type where arithmetic is always promoted to float. InstCombine can remove some fpext/fptruncs around such arithmetic and turn it into arithmetic on half. This wouldn't be so bad if SelectionDAG was able to put those fpext/fpround back in when it promotes. It is also not obvious how to handle to make the existing strategy work with STRICT fp. We need to use STRICT versions of the conversions which require chain operands. But if the conversions are created for a bitcast, there is no place to get an appropriate chain from. This patch implements a different strategy where conversions are emitted directly around arithmetic operations. And otherwise its passed around as an i16 including in arguments and return values. This can result in more conversions between arithmetic operations, but is closer to matching the IR the frontend generates for __fp16. And it will allow us to use the chain from constrained arithmetic nodes to link the STRICT_FP_TO_FP16/STRICT_FP16_TO_FP that will need to be added. I've set it up so that each target can opt into the new behavior. Converting all the targets myself was more than I was able to handle. Differential Revision: https://reviews.llvm.org/D73749
2020-02-01 14:42:07 +08:00
// Allow targets to control how we legalize half.
if (softPromoteHalfType()) {
NumRegistersForVT[MVT::f16] = NumRegistersForVT[MVT::i16];
RegisterTypeForVT[MVT::f16] = RegisterTypeForVT[MVT::i16];
TransformToType[MVT::f16] = MVT::f32;
ValueTypeActions.setTypeAction(MVT::f16, TypeSoftPromoteHalf);
} else {
NumRegistersForVT[MVT::f16] = NumRegistersForVT[MVT::f32];
RegisterTypeForVT[MVT::f16] = RegisterTypeForVT[MVT::f32];
TransformToType[MVT::f16] = MVT::f32;
ValueTypeActions.setTypeAction(MVT::f16, TypePromoteFloat);
}
CodeGen: soften f16 type by default instead of marking legal. Actual support for softening f16 operations is still limited, and can be added when it's needed. But Soften is much closer to being a useful thing to try than keeping it Legal when no registers can actually hold such values. Longer term, we probably want something between Soften and Promote semantics for most targets, it'll be more efficient to promote the 4 basic operations to f32 than libcall them. llvm-svn: 213372
2014-07-18 20:41:46 +08:00
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Loop over all of the vector value types to see which need transformations.
for (unsigned i = MVT::FIRST_VECTOR_VALUETYPE;
i <= (unsigned)MVT::LAST_VECTOR_VALUETYPE; ++i) {
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
MVT VT = (MVT::SimpleValueType) i;
if (isTypeLegal(VT))
continue;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
MVT EltVT = VT.getVectorElementType();
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
ElementCount EC = VT.getVectorElementCount();
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
bool IsLegalWiderType = false;
[AArch64][SVE] Add SPLAT_VECTOR ISD Node Adds a new ISD node to replicate a scalar value across all elements of a vector. This is needed for scalable vectors, since BUILD_VECTOR cannot be used. Fixes up default type legalization for scalable vectors after the new MVT type ranges were introduced. At present I only use this node for scalable vectors. A DAGCombine has been added to transform a BUILD_VECTOR into a SPLAT_VECTOR if all elements are the same, but only if the default operation action of Expand has been overridden by the target. I've only added result promotion legalization for scalable vector i8/i16/i32/i64 types in AArch64 for now. Reviewers: t.p.northover, javed.absar, greened, cameron.mcinally, jmolloy Reviewed By: jmolloy Differential Revision: https://reviews.llvm.org/D47775 llvm-svn: 375222
2019-10-18 19:48:35 +08:00
bool IsScalable = VT.isScalableVector();
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
LegalizeTypeAction PreferredAction = getPreferredVectorAction(VT);
switch (PreferredAction) {
[AArch64][SVE] Add SPLAT_VECTOR ISD Node Adds a new ISD node to replicate a scalar value across all elements of a vector. This is needed for scalable vectors, since BUILD_VECTOR cannot be used. Fixes up default type legalization for scalable vectors after the new MVT type ranges were introduced. At present I only use this node for scalable vectors. A DAGCombine has been added to transform a BUILD_VECTOR into a SPLAT_VECTOR if all elements are the same, but only if the default operation action of Expand has been overridden by the target. I've only added result promotion legalization for scalable vector i8/i16/i32/i64 types in AArch64 for now. Reviewers: t.p.northover, javed.absar, greened, cameron.mcinally, jmolloy Reviewed By: jmolloy Differential Revision: https://reviews.llvm.org/D47775 llvm-svn: 375222
2019-10-18 19:48:35 +08:00
case TypePromoteInteger: {
MVT::SimpleValueType EndVT = IsScalable ?
MVT::LAST_INTEGER_SCALABLE_VECTOR_VALUETYPE :
MVT::LAST_INTEGER_FIXEDLEN_VECTOR_VALUETYPE;
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
// Try to promote the elements of integer vectors. If no legal
// promotion was found, fall through to the widen-vector method.
[SVE][MVT] Fixed-length vector MVT ranges * Reordered MVT simple types to group scalable vector types together. * New range functions in MachineValueType.h to only iterate over the fixed-length int/fp vector types. * Stopped backends which don't support scalable vector types from iterating over scalable types. Reviewers: sdesmalen, greened Reviewed By: greened Differential Revision: https://reviews.llvm.org/D66339 llvm-svn: 372099
2019-09-17 18:19:23 +08:00
for (unsigned nVT = i + 1;
[AArch64][SVE] Add SPLAT_VECTOR ISD Node Adds a new ISD node to replicate a scalar value across all elements of a vector. This is needed for scalable vectors, since BUILD_VECTOR cannot be used. Fixes up default type legalization for scalable vectors after the new MVT type ranges were introduced. At present I only use this node for scalable vectors. A DAGCombine has been added to transform a BUILD_VECTOR into a SPLAT_VECTOR if all elements are the same, but only if the default operation action of Expand has been overridden by the target. I've only added result promotion legalization for scalable vector i8/i16/i32/i64 types in AArch64 for now. Reviewers: t.p.northover, javed.absar, greened, cameron.mcinally, jmolloy Reviewed By: jmolloy Differential Revision: https://reviews.llvm.org/D47775 llvm-svn: 375222
2019-10-18 19:48:35 +08:00
(MVT::SimpleValueType)nVT <= EndVT; ++nVT) {
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
MVT SVT = (MVT::SimpleValueType) nVT;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Promote vectors of integers to vectors with the same number
// of elements, with a wider element type.
getVectorElementType().getSizeInBits() -> getScalarSizeInBits() ; NFCI llvm-svn: 281495
2016-09-15 00:37:15 +08:00
if (SVT.getScalarSizeInBits() > EltVT.getSizeInBits() &&
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
SVT.getVectorElementCount() == EC && isTypeLegal(SVT)) {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
TransformToType[i] = SVT;
RegisterTypeForVT[i] = SVT;
NumRegistersForVT[i] = 1;
ValueTypeActions.setTypeAction(VT, TypePromoteInteger);
IsLegalWiderType = true;
break;
}
}
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
if (IsLegalWiderType)
break;
Added LLVM_FALLTHROUGH to address warning: this statement may fall through. NFC. llvm-svn: 304635
2017-06-03 13:11:14 +08:00
LLVM_FALLTHROUGH;
[AArch64][SVE] Add SPLAT_VECTOR ISD Node Adds a new ISD node to replicate a scalar value across all elements of a vector. This is needed for scalable vectors, since BUILD_VECTOR cannot be used. Fixes up default type legalization for scalable vectors after the new MVT type ranges were introduced. At present I only use this node for scalable vectors. A DAGCombine has been added to transform a BUILD_VECTOR into a SPLAT_VECTOR if all elements are the same, but only if the default operation action of Expand has been overridden by the target. I've only added result promotion legalization for scalable vector i8/i16/i32/i64 types in AArch64 for now. Reviewers: t.p.northover, javed.absar, greened, cameron.mcinally, jmolloy Reviewed By: jmolloy Differential Revision: https://reviews.llvm.org/D47775 llvm-svn: 375222
2019-10-18 19:48:35 +08:00
}
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
case TypeWidenVector:
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
if (isPowerOf2_32(EC.Min)) {
[TargetLowering] Teach computeRegisterProperties to only widen v3i16/v3f16 vectors to the next power of 2 type if that's legal. These were recently made simple types. This restores their behavior back to something like their EVT legalization. We might be able to fix the code in type legalization where the assert was failing, but I didn't investigate too much as I had already looked at the computeRegisterProperties code during the review for v3i16/v3f16. Most of the test changes restore the X86 codegen back to what it looked like before the recent change. The test case in vec_setcc.ll and is a reduced version of the reproducer from the fuzzer. Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=16490 llvm-svn: 369205
2019-08-18 14:28:06 +08:00
// Try to widen the vector.
for (unsigned nVT = i + 1; nVT <= MVT::LAST_VECTOR_VALUETYPE; ++nVT) {
MVT SVT = (MVT::SimpleValueType) nVT;
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
if (SVT.getVectorElementType() == EltVT &&
SVT.isScalableVector() == IsScalable &&
SVT.getVectorElementCount().Min > EC.Min && isTypeLegal(SVT)) {
[TargetLowering] Teach computeRegisterProperties to only widen v3i16/v3f16 vectors to the next power of 2 type if that's legal. These were recently made simple types. This restores their behavior back to something like their EVT legalization. We might be able to fix the code in type legalization where the assert was failing, but I didn't investigate too much as I had already looked at the computeRegisterProperties code during the review for v3i16/v3f16. Most of the test changes restore the X86 codegen back to what it looked like before the recent change. The test case in vec_setcc.ll and is a reduced version of the reproducer from the fuzzer. Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=16490 llvm-svn: 369205
2019-08-18 14:28:06 +08:00
TransformToType[i] = SVT;
RegisterTypeForVT[i] = SVT;
NumRegistersForVT[i] = 1;
ValueTypeActions.setTypeAction(VT, TypeWidenVector);
IsLegalWiderType = true;
break;
}
}
if (IsLegalWiderType)
break;
} else {
// Only widen to the next power of 2 to keep consistency with EVT.
MVT NVT = VT.getPow2VectorType();
if (isTypeLegal(NVT)) {
TransformToType[i] = NVT;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
ValueTypeActions.setTypeAction(VT, TypeWidenVector);
[TargetLowering] Teach computeRegisterProperties to only widen v3i16/v3f16 vectors to the next power of 2 type if that's legal. These were recently made simple types. This restores their behavior back to something like their EVT legalization. We might be able to fix the code in type legalization where the assert was failing, but I didn't investigate too much as I had already looked at the computeRegisterProperties code during the review for v3i16/v3f16. Most of the test changes restore the X86 codegen back to what it looked like before the recent change. The test case in vec_setcc.ll and is a reduced version of the reproducer from the fuzzer. Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=16490 llvm-svn: 369205
2019-08-18 14:28:06 +08:00
RegisterTypeForVT[i] = NVT;
NumRegistersForVT[i] = 1;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
break;
}
}
Added LLVM_FALLTHROUGH to address warning: this statement may fall through. NFC. llvm-svn: 304635
2017-06-03 13:11:14 +08:00
LLVM_FALLTHROUGH;
[CodeGen] Fix some Clang-tidy modernize-use-bool-literals and Include What You Use warnings; other minor fixes (NFC). llvm-svn: 313941
2017-09-22 07:20:16 +08:00
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
case TypeSplitVector:
case TypeScalarizeVector: {
MVT IntermediateVT;
MVT RegisterVT;
unsigned NumIntermediates;
[TargetLowering] Increase the storage size of NumRegistersForVT to allow the type break down for v256i1 and other types to be stored correctly v256i1 on X86 without avx512 breaks down to 256 i8 values when passed between basic blocks. But the NumRegistersForVT was sized at a byte for each VT. This results in 256 being stored as 0. This patch enlarges the type to 16 bits and adds an assert to ensure that no information is lost when the entry is stored. Differential Revision: https://reviews.llvm.org/D70138
2019-11-14 04:09:34 +08:00
unsigned NumRegisters = getVectorTypeBreakdownMVT(VT, IntermediateVT,
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
NumIntermediates, RegisterVT, this);
[TargetLowering] Increase the storage size of NumRegistersForVT to allow the type break down for v256i1 and other types to be stored correctly v256i1 on X86 without avx512 breaks down to 256 i8 values when passed between basic blocks. But the NumRegistersForVT was sized at a byte for each VT. This results in 256 being stored as 0. This patch enlarges the type to 16 bits and adds an assert to ensure that no information is lost when the entry is stored. Differential Revision: https://reviews.llvm.org/D70138
2019-11-14 04:09:34 +08:00
NumRegistersForVT[i] = NumRegisters;
assert(NumRegistersForVT[i] == NumRegisters &&
"NumRegistersForVT size cannot represent NumRegisters!");
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
RegisterTypeForVT[i] = RegisterVT;
MVT NVT = VT.getPow2VectorType();
if (NVT == VT) {
// Type is already a power of 2. The default action is to split.
TransformToType[i] = MVT::Other;
if (PreferredAction == TypeScalarizeVector)
ValueTypeActions.setTypeAction(VT, TypeScalarizeVector);
PR20557: Fix the bug that bogus cpu parameter crashes llc on AArch64 backend. Initial patch by Oleg Ranevskyy. llvm-svn: 220945
2014-10-31 10:35:34 +08:00
else if (PreferredAction == TypeSplitVector)
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
ValueTypeActions.setTypeAction(VT, TypeSplitVector);
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
else if (EC.Min > 1)
ValueTypeActions.setTypeAction(VT, TypeSplitVector);
PR20557: Fix the bug that bogus cpu parameter crashes llc on AArch64 backend. Initial patch by Oleg Ranevskyy. llvm-svn: 220945
2014-10-31 10:35:34 +08:00
else
Reland D80640: [CodeGen][SVE] Calculate correct type legalization for scalable vectors. This reverts commit 9bcef270d7a319c6c0fdffc6c80984a8f0a30ecb.
2020-06-05 23:46:25 +08:00
ValueTypeActions.setTypeAction(VT, EC.Scalable
? TypeScalarizeScalableVector
: TypeScalarizeVector);
[codegen,aarch64] Add a target hook to the code generator to control vector type legalization strategies in a more fine grained manner, and change the legalization of several v1iN types and v1f32 to be widening rather than scalarization on AArch64. This fixes an assertion failure caused by scalarizing nodes like "v1i32 trunc v1i64". As v1i64 is legal it will fail to scalarize v1i32. This also provides a foundation for other targets to have more granular control over how vector types are legalized. Patch by Hao Liu, reviewed by Tim Northover. I'm committing it to allow some work to start taking place on top of this patch as it adds some really important hooks to the backend that I'd like to immediately start using. =] http://reviews.llvm.org/D4322 llvm-svn: 212242
2014-07-03 08:23:43 +08:00
} else {
TransformToType[i] = NVT;
ValueTypeActions.setTypeAction(VT, TypeWidenVector);
}
break;
}
default:
llvm_unreachable("Unknown vector legalization action!");
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
}
// Determine the 'representative' register class for each value type.
// An representative register class is the largest (meaning one which is
// not a sub-register class / subreg register class) legal register class for
// a group of value types. For example, on i386, i8, i16, and i32
// representative would be GR32; while on x86_64 it's GR64.
for (unsigned i = 0; i != MVT::LAST_VALUETYPE; ++i) {
const TargetRegisterClass* RRC;
uint8_t Cost;
Remove an argument-less call to getSubtargetImpl from TargetLoweringBase. This required plumbing a TargetRegisterInfo through computeRegisterProperties and into findRepresentativeClass which uses it for register class iteration. This required passing a subtarget into a few target specific initializations of TargetLowering. llvm-svn: 230583
2015-02-26 08:00:24 +08:00
std::tie(RRC, Cost) = findRepresentativeClass(TRI, (MVT::SimpleValueType)i);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
RepRegClassForVT[i] = RRC;
RepRegClassCostForVT[i] = Cost;
}
}
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT TargetLoweringBase::getSetCCResultType(const DataLayout &DL, LLVMContext &,
EVT VT) const {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
assert(!VT.isVector() && "No default SetCC type for vectors!");
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
return getPointerTy(DL).SimpleTy;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
MVT::SimpleValueType TargetLoweringBase::getCmpLibcallReturnType() const {
return MVT::i32; // return the default value
}
/// getVectorTypeBreakdown - Vector types are broken down into some number of
/// legal first class types. For example, MVT::v8f32 maps to 2 MVT::v4f32
/// with Altivec or SSE1, or 8 promoted MVT::f64 values with the X86 FP stack.
/// Similarly, MVT::v2i64 turns into 4 MVT::i32 values with both PPC and X86.
///
/// This method returns the number of registers needed, and the VT for each
/// register. It also returns the VT and quantity of the intermediate values
/// before they are promoted/expanded.
unsigned TargetLoweringBase::getVectorTypeBreakdown(LLVMContext &Context, EVT VT,
EVT &IntermediateVT,
unsigned &NumIntermediates,
MVT &RegisterVT) const {
[SVE][CodeGen] Fix legalisation for scalable types Summary: This patch handles illegal scalable types when lowering IR operations, addressing several places where the value of isScalableVector() is ignored. For types such as <vscale x 8 x i32>, this means splitting the operations. In this example, we would split it into two operations of type <vscale x 4 x i32> for the low and high halves. In cases such as <vscale x 2 x i32>, the elements in the vector will be promoted. In this case they will be promoted to i64 (with a vector of type <vscale x 2 x i64>) Reviewers: sdesmalen, efriedma, huntergr Reviewed By: efriedma Subscribers: david-arm, tschuett, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D78812
2020-05-07 17:01:31 +08:00
ElementCount EltCnt = VT.getVectorElementCount();
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// If there is a wider vector type with the same element type as this one,
// or a promoted vector type that has the same number of elements which
// are wider, then we should convert to that legal vector type.
// This handles things like <2 x float> -> <4 x float> and
// <4 x i1> -> <4 x i32>.
LegalizeTypeAction TA = getTypeAction(Context, VT);
[SVE][CodeGen] Fix legalisation for scalable types Summary: This patch handles illegal scalable types when lowering IR operations, addressing several places where the value of isScalableVector() is ignored. For types such as <vscale x 8 x i32>, this means splitting the operations. In this example, we would split it into two operations of type <vscale x 4 x i32> for the low and high halves. In cases such as <vscale x 2 x i32>, the elements in the vector will be promoted. In this case they will be promoted to i64 (with a vector of type <vscale x 2 x i64>) Reviewers: sdesmalen, efriedma, huntergr Reviewed By: efriedma Subscribers: david-arm, tschuett, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D78812
2020-05-07 17:01:31 +08:00
if (EltCnt.Min != 1 && (TA == TypeWidenVector || TA == TypePromoteInteger)) {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
EVT RegisterEVT = getTypeToTransformTo(Context, VT);
if (isTypeLegal(RegisterEVT)) {
IntermediateVT = RegisterEVT;
RegisterVT = RegisterEVT.getSimpleVT();
NumIntermediates = 1;
return 1;
}
}
// Figure out the right, legal destination reg to copy into.
EVT EltTy = VT.getVectorElementType();
unsigned NumVectorRegs = 1;
[CodeGen][SVE] CopyToReg: Split scalable EVTs that are not powers of 2 Scalable vectors cannot use 'BUILD_VECTOR', so it is necessary to properly split and widen scalable vectors when passing them to CopyToReg/CopyFromReg. This functionality is added to TargetLoweringBase::getVectorTypeBreakdown(). This patch only adds support for 'splitting' scalable vectors that are a multiple of some legal type, e.g. <vscale x 6 x i64> -> 3 x <vscale x 2 x i64> Reviewers: efriedma, c-rhodes Reviewed By: efriedma Tags: #llvm Differential Revision: https://reviews.llvm.org/D80139
2020-06-06 01:29:43 +08:00
// Scalable vectors cannot be scalarized, so handle the legalisation of the
// types like done elsewhere in SelectionDAG.
if (VT.isScalableVector() && !isPowerOf2_32(EltCnt.Min)) {
LegalizeKind LK;
EVT PartVT = VT;
do {
// Iterate until we've found a legal (part) type to hold VT.
LK = getTypeConversion(Context, PartVT);
PartVT = LK.second;
} while (LK.first != TypeLegal);
NumIntermediates =
VT.getVectorElementCount().Min / PartVT.getVectorElementCount().Min;
// FIXME: This code needs to be extended to handle more complex vector
// breakdowns, like nxv7i64 -> nxv8i64 -> 4 x nxv2i64. Currently the only
// supported cases are vectors that are broken down into equal parts
// such as nxv6i64 -> 3 x nxv2i64.
assert(NumIntermediates * PartVT.getVectorElementCount().Min ==
VT.getVectorElementCount().Min &&
"Expected an integer multiple of PartVT");
IntermediateVT = PartVT;
RegisterVT = getRegisterType(Context, IntermediateVT);
return NumIntermediates;
}
// FIXME: We don't support non-power-of-2-sized vectors for now. Ideally
// we could break down into LHS/RHS like LegalizeDAG does.
[SVE][CodeGen] Fix legalisation for scalable types Summary: This patch handles illegal scalable types when lowering IR operations, addressing several places where the value of isScalableVector() is ignored. For types such as <vscale x 8 x i32>, this means splitting the operations. In this example, we would split it into two operations of type <vscale x 4 x i32> for the low and high halves. In cases such as <vscale x 2 x i32>, the elements in the vector will be promoted. In this case they will be promoted to i64 (with a vector of type <vscale x 2 x i64>) Reviewers: sdesmalen, efriedma, huntergr Reviewed By: efriedma Subscribers: david-arm, tschuett, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D78812
2020-05-07 17:01:31 +08:00
if (!isPowerOf2_32(EltCnt.Min)) {
NumVectorRegs = EltCnt.Min;
EltCnt.Min = 1;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
// Divide the input until we get to a supported size. This will always
// end with a scalar if the target doesn't support vectors.
[SVE][CodeGen] Fix legalisation for scalable types Summary: This patch handles illegal scalable types when lowering IR operations, addressing several places where the value of isScalableVector() is ignored. For types such as <vscale x 8 x i32>, this means splitting the operations. In this example, we would split it into two operations of type <vscale x 4 x i32> for the low and high halves. In cases such as <vscale x 2 x i32>, the elements in the vector will be promoted. In this case they will be promoted to i64 (with a vector of type <vscale x 2 x i64>) Reviewers: sdesmalen, efriedma, huntergr Reviewed By: efriedma Subscribers: david-arm, tschuett, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D78812
2020-05-07 17:01:31 +08:00
while (EltCnt.Min > 1 &&
!isTypeLegal(EVT::getVectorVT(Context, EltTy, EltCnt))) {
EltCnt.Min >>= 1;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
NumVectorRegs <<= 1;
}
NumIntermediates = NumVectorRegs;
[SVE][CodeGen] Fix legalisation for scalable types Summary: This patch handles illegal scalable types when lowering IR operations, addressing several places where the value of isScalableVector() is ignored. For types such as <vscale x 8 x i32>, this means splitting the operations. In this example, we would split it into two operations of type <vscale x 4 x i32> for the low and high halves. In cases such as <vscale x 2 x i32>, the elements in the vector will be promoted. In this case they will be promoted to i64 (with a vector of type <vscale x 2 x i64>) Reviewers: sdesmalen, efriedma, huntergr Reviewed By: efriedma Subscribers: david-arm, tschuett, hiraditya, rkruppe, psnobl, cfe-commits, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D78812
2020-05-07 17:01:31 +08:00
EVT NewVT = EVT::getVectorVT(Context, EltTy, EltCnt);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
if (!isTypeLegal(NewVT))
NewVT = EltTy;
IntermediateVT = NewVT;
MVT DestVT = getRegisterType(Context, NewVT);
RegisterVT = DestVT;
[CodeGen] Fix warnings in getVectorTypeBreakdown Added NextPowerOf2() routine to TypeSize and rewritten the code in getVectorTypeBreakdown to avoid warnings being generated. Differential Revision: https://reviews.llvm.org/D81578
2020-06-11 14:47:23 +08:00
if (EVT(DestVT).bitsLT(NewVT)) { // Value is expanded, e.g. i64 -> i16.
TypeSize NewVTSize = NewVT.getSizeInBits();
// Convert sizes such as i33 to i64.
if (!isPowerOf2_32(NewVTSize.getKnownMinSize()))
NewVTSize = NewVTSize.NextPowerOf2();
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
return NumVectorRegs*(NewVTSize/DestVT.getSizeInBits());
[CodeGen] Fix warnings in getVectorTypeBreakdown Added NextPowerOf2() routine to TypeSize and rewritten the code in getVectorTypeBreakdown to avoid warnings being generated. Differential Revision: https://reviews.llvm.org/D81578
2020-06-11 14:47:23 +08:00
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Otherwise, promotion or legal types use the same number of registers as
// the vector decimated to the appropriate level.
return NumVectorRegs;
}
[PGO][PGSO] Fix -DBUILD_SHARED_LIBS=on builds after D69580/llvmorg-10-init-8797-g0d987e411ac Move TargetLoweringBase::isSuitableForJumpTable from llvm/CodeGen/TargetLowering.h to .cpp, to avoid the undefined reference from all LLVM${Target}ISelLowering.cpp. Another fix is to add a dependency on TransformUtils to all lib/Target/$Target/LLVMBuild.txt, but that is too disruptive.
2019-11-01 04:58:52 +08:00
bool TargetLoweringBase::isSuitableForJumpTable(const SwitchInst *SI,
uint64_t NumCases,
uint64_t Range,
ProfileSummaryInfo *PSI,
BlockFrequencyInfo *BFI) const {
// FIXME: This function check the maximum table size and density, but the
// minimum size is not checked. It would be nice if the minimum size is
// also combined within this function. Currently, the minimum size check is
// performed in findJumpTable() in SelectionDAGBuiler and
// getEstimatedNumberOfCaseClusters() in BasicTTIImpl.
const bool OptForSize =
SI->getParent()->getParent()->hasOptSize() ||
llvm::shouldOptimizeForSize(SI->getParent(), PSI, BFI);
const unsigned MinDensity = getMinimumJumpTableDensity(OptForSize);
const unsigned MaxJumpTableSize = getMaximumJumpTableSize();
// Check whether the number of cases is small enough and
// the range is dense enough for a jump table.
return (OptForSize || Range <= MaxJumpTableSize) &&
(NumCases * 100 >= Range * MinDensity);
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
/// Get the EVTs and ArgFlags collections that represent the legalized return
/// type of the given function. This does not require a DAG or a return value,
/// and is suitable for use before any DAGs for the function are constructed.
/// TODO: Move this out of TargetLowering.cpp.
DAG: Add calling convention argument to calling convention funcs This seems like a pretty glaring omission, and AMDGPU wants to treat kernels differently from other calling conventions. llvm-svn: 338194
2018-07-28 21:25:19 +08:00
void llvm::GetReturnInfo(CallingConv::ID CC, Type *ReturnType,
AttributeList attr,
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
SmallVectorImpl<ISD::OutputArg> &Outs,
Redirect DataLayout from TargetMachine to Module in ComputeValueVTs() Summary: Avoid using the TargetMachine owned DataLayout and use the Module owned one instead. This requires passing the DataLayout up the stack to ComputeValueVTs(). This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11019 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241773
2015-07-09 09:57:34 +08:00
const TargetLowering &TLI, const DataLayout &DL) {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
SmallVector<EVT, 4> ValueVTs;
Redirect DataLayout from TargetMachine to Module in ComputeValueVTs() Summary: Avoid using the TargetMachine owned DataLayout and use the Module owned one instead. This requires passing the DataLayout up the stack to ComputeValueVTs(). This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11019 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241773
2015-07-09 09:57:34 +08:00
ComputeValueVTs(TLI, DL, ReturnType, ValueVTs);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
unsigned NumValues = ValueVTs.size();
if (NumValues == 0) return;
for (unsigned j = 0, f = NumValues; j != f; ++j) {
EVT VT = ValueVTs[j];
ISD::NodeType ExtendKind = ISD::ANY_EXTEND;
Rename AttributeSet to AttributeList Summary: This class is a list of AttributeSetNodes corresponding the function prototype of a call or function declaration. This class used to be called ParamAttrListPtr, then AttrListPtr, then AttributeSet. It is typically accessed by parameter and return value index, so "AttributeList" seems like a more intuitive name. Rename AttributeSetImpl to AttributeListImpl to follow suit. It's useful to rename this class so that we can rename AttributeSetNode to AttributeSet later. AttributeSet is the set of attributes that apply to a single function, argument, or return value. Reviewers: sanjoy, javed.absar, chandlerc, pete Reviewed By: pete Subscribers: pete, jholewinski, arsenm, dschuff, mehdi_amini, jfb, nhaehnle, sbc100, void, llvm-commits Differential Revision: https://reviews.llvm.org/D31102 llvm-svn: 298393
2017-03-22 00:57:19 +08:00
if (attr.hasAttribute(AttributeList::ReturnIndex, Attribute::SExt))
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
ExtendKind = ISD::SIGN_EXTEND;
Rename AttributeSet to AttributeList Summary: This class is a list of AttributeSetNodes corresponding the function prototype of a call or function declaration. This class used to be called ParamAttrListPtr, then AttrListPtr, then AttributeSet. It is typically accessed by parameter and return value index, so "AttributeList" seems like a more intuitive name. Rename AttributeSetImpl to AttributeListImpl to follow suit. It's useful to rename this class so that we can rename AttributeSetNode to AttributeSet later. AttributeSet is the set of attributes that apply to a single function, argument, or return value. Reviewers: sanjoy, javed.absar, chandlerc, pete Reviewed By: pete Subscribers: pete, jholewinski, arsenm, dschuff, mehdi_amini, jfb, nhaehnle, sbc100, void, llvm-commits Differential Revision: https://reviews.llvm.org/D31102 llvm-svn: 298393
2017-03-22 00:57:19 +08:00
else if (attr.hasAttribute(AttributeList::ReturnIndex, Attribute::ZExt))
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
ExtendKind = ISD::ZERO_EXTEND;
// FIXME: C calling convention requires the return type to be promoted to
// at least 32-bit. But this is not necessary for non-C calling
// conventions. The frontend should mark functions whose return values
// require promoting with signext or zeroext attributes.
if (ExtendKind != ISD::ANY_EXTEND && VT.isInteger()) {
MVT MinVT = TLI.getRegisterType(ReturnType->getContext(), MVT::i32);
if (VT.bitsLT(MinVT))
VT = MinVT;
}
Reland "[SelectionDAG] Enable target specific vector scalarization of calls and returns" By target hookifying getRegisterType, getNumRegisters, getVectorBreakdown, backends can request that LLVM to scalarize vector types for calls and returns. The MIPS vector ABI requires that vector arguments and returns are passed in integer registers. With SelectionDAG's new hooks, the MIPS backend can now handle LLVM-IR with vector types in calls and returns. E.g. 'call @foo(<4 x i32> %4)'. Previously these cases would be scalarized for the MIPS O32/N32/N64 ABI for calls and returns if vector types were not legal. If vector types were legal, a single 128bit vector argument would be assigned to a single 32 bit / 64 bit integer register. By teaching the MIPS backend to inspect the original types, it can now implement the MIPS vector ABI which requires a particular method of scalarizing vectors. Previously, the MIPS backend relied on clang to scalarize types such as "call @foo(<4 x float> %a) into "call @foo(i32 inreg %1, i32 inreg %2, i32 inreg %3, i32 inreg %4)". This patch enables the MIPS backend to take either form for vector types. The previous version of this patch had a "conditional move or jump depends on uninitialized value". Reviewers: zoran.jovanovic, jaydeep, vkalintiris, slthakur Differential Revision: https://reviews.llvm.org/D27845 llvm-svn: 305083
2017-06-09 22:37:08 +08:00
unsigned NumParts =
DAG: Add calling convention argument to calling convention funcs This seems like a pretty glaring omission, and AMDGPU wants to treat kernels differently from other calling conventions. llvm-svn: 338194
2018-07-28 21:25:19 +08:00
TLI.getNumRegistersForCallingConv(ReturnType->getContext(), CC, VT);
Reland "[SelectionDAG] Enable target specific vector scalarization of calls and returns" By target hookifying getRegisterType, getNumRegisters, getVectorBreakdown, backends can request that LLVM to scalarize vector types for calls and returns. The MIPS vector ABI requires that vector arguments and returns are passed in integer registers. With SelectionDAG's new hooks, the MIPS backend can now handle LLVM-IR with vector types in calls and returns. E.g. 'call @foo(<4 x i32> %4)'. Previously these cases would be scalarized for the MIPS O32/N32/N64 ABI for calls and returns if vector types were not legal. If vector types were legal, a single 128bit vector argument would be assigned to a single 32 bit / 64 bit integer register. By teaching the MIPS backend to inspect the original types, it can now implement the MIPS vector ABI which requires a particular method of scalarizing vectors. Previously, the MIPS backend relied on clang to scalarize types such as "call @foo(<4 x float> %a) into "call @foo(i32 inreg %1, i32 inreg %2, i32 inreg %3, i32 inreg %4)". This patch enables the MIPS backend to take either form for vector types. The previous version of this patch had a "conditional move or jump depends on uninitialized value". Reviewers: zoran.jovanovic, jaydeep, vkalintiris, slthakur Differential Revision: https://reviews.llvm.org/D27845 llvm-svn: 305083
2017-06-09 22:37:08 +08:00
MVT PartVT =
DAG: Add calling convention argument to calling convention funcs This seems like a pretty glaring omission, and AMDGPU wants to treat kernels differently from other calling conventions. llvm-svn: 338194
2018-07-28 21:25:19 +08:00
TLI.getRegisterTypeForCallingConv(ReturnType->getContext(), CC, VT);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// 'inreg' on function refers to return value
ISD::ArgFlagsTy Flags = ISD::ArgFlagsTy();
Rename AttributeSet to AttributeList Summary: This class is a list of AttributeSetNodes corresponding the function prototype of a call or function declaration. This class used to be called ParamAttrListPtr, then AttrListPtr, then AttributeSet. It is typically accessed by parameter and return value index, so "AttributeList" seems like a more intuitive name. Rename AttributeSetImpl to AttributeListImpl to follow suit. It's useful to rename this class so that we can rename AttributeSetNode to AttributeSet later. AttributeSet is the set of attributes that apply to a single function, argument, or return value. Reviewers: sanjoy, javed.absar, chandlerc, pete Reviewed By: pete Subscribers: pete, jholewinski, arsenm, dschuff, mehdi_amini, jfb, nhaehnle, sbc100, void, llvm-commits Differential Revision: https://reviews.llvm.org/D31102 llvm-svn: 298393
2017-03-22 00:57:19 +08:00
if (attr.hasAttribute(AttributeList::ReturnIndex, Attribute::InReg))
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
Flags.setInReg();
// Propagate extension type if any
Rename AttributeSet to AttributeList Summary: This class is a list of AttributeSetNodes corresponding the function prototype of a call or function declaration. This class used to be called ParamAttrListPtr, then AttrListPtr, then AttributeSet. It is typically accessed by parameter and return value index, so "AttributeList" seems like a more intuitive name. Rename AttributeSetImpl to AttributeListImpl to follow suit. It's useful to rename this class so that we can rename AttributeSetNode to AttributeSet later. AttributeSet is the set of attributes that apply to a single function, argument, or return value. Reviewers: sanjoy, javed.absar, chandlerc, pete Reviewed By: pete Subscribers: pete, jholewinski, arsenm, dschuff, mehdi_amini, jfb, nhaehnle, sbc100, void, llvm-commits Differential Revision: https://reviews.llvm.org/D31102 llvm-svn: 298393
2017-03-22 00:57:19 +08:00
if (attr.hasAttribute(AttributeList::ReturnIndex, Attribute::SExt))
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
Flags.setSExt();
Rename AttributeSet to AttributeList Summary: This class is a list of AttributeSetNodes corresponding the function prototype of a call or function declaration. This class used to be called ParamAttrListPtr, then AttrListPtr, then AttributeSet. It is typically accessed by parameter and return value index, so "AttributeList" seems like a more intuitive name. Rename AttributeSetImpl to AttributeListImpl to follow suit. It's useful to rename this class so that we can rename AttributeSetNode to AttributeSet later. AttributeSet is the set of attributes that apply to a single function, argument, or return value. Reviewers: sanjoy, javed.absar, chandlerc, pete Reviewed By: pete Subscribers: pete, jholewinski, arsenm, dschuff, mehdi_amini, jfb, nhaehnle, sbc100, void, llvm-commits Differential Revision: https://reviews.llvm.org/D31102 llvm-svn: 298393
2017-03-22 00:57:19 +08:00
else if (attr.hasAttribute(AttributeList::ReturnIndex, Attribute::ZExt))
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
Flags.setZExt();
for (unsigned i = 0; i < NumParts; ++i)
Fix parameter name comments using clang-tidy. NFC. This patch applies clang-tidy's bugprone-argument-comment tool to LLVM, clang and lld source trees. Here is how I created this patch: $ git clone https://github.com/llvm/llvm-project.git $ cd llvm-project $ mkdir build $ cd build $ cmake -GNinja -DCMAKE_BUILD_TYPE=Debug \ -DLLVM_ENABLE_PROJECTS='clang;lld;clang-tools-extra' \ -DCMAKE_EXPORT_COMPILE_COMMANDS=On -DLLVM_ENABLE_LLD=On \ -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ ../llvm $ ninja $ parallel clang-tidy -checks='-*,bugprone-argument-comment' \ -config='{CheckOptions: [{key: StrictMode, value: 1}]}' -fix \ ::: ../llvm/lib/**/*.{cpp,h} ../clang/lib/**/*.{cpp,h} ../lld/**/*.{cpp,h} llvm-svn: 366177
2019-07-16 12:46:31 +08:00
Outs.push_back(ISD::OutputArg(Flags, PartVT, VT, /*isfixed=*/true, 0, 0));
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
}
/// getByValTypeAlignment - Return the desired alignment for ByVal aggregate
/// function arguments in the caller parameter area. This is the actual
/// alignment, not its logarithm.
Make getByValTypeAlignment() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: yaron.keren, rafael, llvm-commits, jholewinski Differential Revision: http://reviews.llvm.org/D11038 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241777
2015-07-09 10:09:28 +08:00
unsigned TargetLoweringBase::getByValTypeAlignment(Type *Ty,
const DataLayout &DL) const {
[Alignment][NFC] Use more Align versions of various functions Summary: This is patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Reviewers: courbet Subscribers: MatzeB, qcolombet, arsenm, sdardis, jvesely, nhaehnle, hiraditya, jrtc27, atanasyan, kerbowa, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D77291
2020-04-02 16:53:29 +08:00
return DL.getABITypeAlign(Ty).value();
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
[TargetLowering] Make allowsMemoryAccess methode virtual. Rename old function to explicitly show that it cares only about alignment. The new allowsMemoryAccess call the function related to alignment by default and can be overridden by target to inform whether the memory access is legal or not. Differential Revision: https://reviews.llvm.org/D67121 llvm-svn: 372935
2019-09-26 08:16:01 +08:00
bool TargetLoweringBase::allowsMemoryAccessForAlignment(
LLVMContext &Context, const DataLayout &DL, EVT VT, unsigned AddrSpace,
[Alignment][NFC] TargetLowering::allowsMemoryAccessForAlignment First patch of a series to adapt TargetLowering::allowsXXX functions This patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Differential Revision: https://reviews.llvm.org/D81372
2020-06-30 23:31:24 +08:00
Align Alignment, MachineMemOperand::Flags Flags, bool *Fast) const {
move DAGCombiner's allowableAlignment() helper function into the TLI Making allowableAlignment() more accessible was suggested as a predecessor patch for D10662, so I've pulled it into TargetLowering. This let's us remove 4 instances of duplicate logic in LegalizeDAG. There's a subtle functional change in the implementation: the existing allowableAlignment() code was using getPrefTypeAlignment() when checking alignment with the DataLayout and assumed that was fast. In this implementation, we use getABITypeAlignment() and assume that is fast. See the TODO comment or the discussion in the Phab review for future improvements in this implementation (don't use the data layout at all). There are no regression test changes from this difference, and I'm not sure how to expose it via a test. I think we actually do want to provide the 'Fast' param when checking this from DAGCombiner::MergeConsecutiveStores(). Ie, we shouldn't merge stores if the new stores are not going to be fast. But that change will require fixing allowsMisalignedMemoryAccess() overrides as noted in D10662. Differential Revision: http://reviews.llvm.org/D10905 llvm-svn: 243549
2015-07-30 02:24:18 +08:00
// Check if the specified alignment is sufficient based on the data layout.
// TODO: While using the data layout works in practice, a better solution
// would be to implement this check directly (make this a virtual function).
// For example, the ABI alignment may change based on software platform while
// this function should only be affected by hardware implementation.
Type *Ty = VT.getTypeForEVT(Context);
[Alignment][NFC] TargetLowering::allowsMemoryAccessForAlignment First patch of a series to adapt TargetLowering::allowsXXX functions This patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Differential Revision: https://reviews.llvm.org/D81372
2020-06-30 23:31:24 +08:00
if (Alignment >= DL.getABITypeAlign(Ty)) {
move DAGCombiner's allowableAlignment() helper function into the TLI Making allowableAlignment() more accessible was suggested as a predecessor patch for D10662, so I've pulled it into TargetLowering. This let's us remove 4 instances of duplicate logic in LegalizeDAG. There's a subtle functional change in the implementation: the existing allowableAlignment() code was using getPrefTypeAlignment() when checking alignment with the DataLayout and assumed that was fast. In this implementation, we use getABITypeAlignment() and assume that is fast. See the TODO comment or the discussion in the Phab review for future improvements in this implementation (don't use the data layout at all). There are no regression test changes from this difference, and I'm not sure how to expose it via a test. I think we actually do want to provide the 'Fast' param when checking this from DAGCombiner::MergeConsecutiveStores(). Ie, we shouldn't merge stores if the new stores are not going to be fast. But that change will require fixing allowsMisalignedMemoryAccess() overrides as noted in D10662. Differential Revision: http://reviews.llvm.org/D10905 llvm-svn: 243549
2015-07-30 02:24:18 +08:00
// Assume that an access that meets the ABI-specified alignment is fast.
if (Fast != nullptr)
*Fast = true;
return true;
}
Remove trailing space sed -Ei 's/[[:space:]]+$//' include/**/*.{def,h,td} lib/**/*.{cpp,h} llvm-svn: 338293
2018-07-31 03:41:25 +08:00
move DAGCombiner's allowableAlignment() helper function into the TLI Making allowableAlignment() more accessible was suggested as a predecessor patch for D10662, so I've pulled it into TargetLowering. This let's us remove 4 instances of duplicate logic in LegalizeDAG. There's a subtle functional change in the implementation: the existing allowableAlignment() code was using getPrefTypeAlignment() when checking alignment with the DataLayout and assumed that was fast. In this implementation, we use getABITypeAlignment() and assume that is fast. See the TODO comment or the discussion in the Phab review for future improvements in this implementation (don't use the data layout at all). There are no regression test changes from this difference, and I'm not sure how to expose it via a test. I think we actually do want to provide the 'Fast' param when checking this from DAGCombiner::MergeConsecutiveStores(). Ie, we shouldn't merge stores if the new stores are not going to be fast. But that change will require fixing allowsMisalignedMemoryAccess() overrides as noted in D10662. Differential Revision: http://reviews.llvm.org/D10905 llvm-svn: 243549
2015-07-30 02:24:18 +08:00
// This is a misaligned access.
[Alignment][NFC] TargetLowering::allowsMemoryAccessForAlignment First patch of a series to adapt TargetLowering::allowsXXX functions This patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Differential Revision: https://reviews.llvm.org/D81372
2020-06-30 23:31:24 +08:00
return allowsMisalignedMemoryAccesses(VT, AddrSpace, Alignment.value(), Flags,
Fast);
move DAGCombiner's allowableAlignment() helper function into the TLI Making allowableAlignment() more accessible was suggested as a predecessor patch for D10662, so I've pulled it into TargetLowering. This let's us remove 4 instances of duplicate logic in LegalizeDAG. There's a subtle functional change in the implementation: the existing allowableAlignment() code was using getPrefTypeAlignment() when checking alignment with the DataLayout and assumed that was fast. In this implementation, we use getABITypeAlignment() and assume that is fast. See the TODO comment or the discussion in the Phab review for future improvements in this implementation (don't use the data layout at all). There are no regression test changes from this difference, and I'm not sure how to expose it via a test. I think we actually do want to provide the 'Fast' param when checking this from DAGCombiner::MergeConsecutiveStores(). Ie, we shouldn't merge stores if the new stores are not going to be fast. But that change will require fixing allowsMisalignedMemoryAccess() overrides as noted in D10662. Differential Revision: http://reviews.llvm.org/D10905 llvm-svn: 243549
2015-07-30 02:24:18 +08:00
}
[TargetLowering] Make allowsMemoryAccess methode virtual. Rename old function to explicitly show that it cares only about alignment. The new allowsMemoryAccess call the function related to alignment by default and can be overridden by target to inform whether the memory access is legal or not. Differential Revision: https://reviews.llvm.org/D67121 llvm-svn: 372935
2019-09-26 08:16:01 +08:00
bool TargetLoweringBase::allowsMemoryAccessForAlignment(
LLVMContext &Context, const DataLayout &DL, EVT VT,
const MachineMemOperand &MMO, bool *Fast) const {
return allowsMemoryAccessForAlignment(Context, DL, VT, MMO.getAddrSpace(),
[Alignment][NFC] TargetLowering::allowsMemoryAccessForAlignment First patch of a series to adapt TargetLowering::allowsXXX functions This patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Differential Revision: https://reviews.llvm.org/D81372
2020-06-30 23:31:24 +08:00
MMO.getAlign(), MMO.getFlags(), Fast);
[TargetLowering] Make allowsMemoryAccess methode virtual. Rename old function to explicitly show that it cares only about alignment. The new allowsMemoryAccess call the function related to alignment by default and can be overridden by target to inform whether the memory access is legal or not. Differential Revision: https://reviews.llvm.org/D67121 llvm-svn: 372935
2019-09-26 08:16:01 +08:00
}
[Alignment][NFC] TargetLowering::allowsMemoryAccess Second patch of a series to adapt TargetLowering::allowsXXX functions This patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Differential Revision: https://reviews.llvm.org/D82785
2020-06-30 16:16:59 +08:00
bool TargetLoweringBase::allowsMemoryAccess(LLVMContext &Context,
const DataLayout &DL, EVT VT,
unsigned AddrSpace, Align Alignment,
MachineMemOperand::Flags Flags,
bool *Fast) const {
[Alignment][NFC] TargetLowering::allowsMemoryAccessForAlignment First patch of a series to adapt TargetLowering::allowsXXX functions This patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Differential Revision: https://reviews.llvm.org/D81372
2020-06-30 23:31:24 +08:00
return allowsMemoryAccessForAlignment(Context, DL, VT, AddrSpace, Alignment,
Flags, Fast);
[TargetLowering] Make allowsMemoryAccess methode virtual. Rename old function to explicitly show that it cares only about alignment. The new allowsMemoryAccess call the function related to alignment by default and can be overridden by target to inform whether the memory access is legal or not. Differential Revision: https://reviews.llvm.org/D67121 llvm-svn: 372935
2019-09-26 08:16:01 +08:00
}
[TargetLowering] Add allowsMemoryAccess(MachineMemOperand) helper wrapper. NFCI. As suggested by @arsenm on D63075 - this adds a TargetLowering::allowsMemoryAccess wrapper that takes a Load/Store node's MachineMemOperand to handle the AddressSpace/Alignment arguments and will also implicitly handle the MachineMemOperand::Flags change in D63075. llvm-svn: 363048
2019-06-11 19:00:23 +08:00
bool TargetLoweringBase::allowsMemoryAccess(LLVMContext &Context,
const DataLayout &DL, EVT VT,
const MachineMemOperand &MMO,
bool *Fast) const {
[Alignment][NFC] TargetLowering::allowsMemoryAccess Second patch of a series to adapt TargetLowering::allowsXXX functions This patch is part of a series to introduce an Alignment type. See this thread for context: http://lists.llvm.org/pipermail/llvm-dev/2019-July/133851.html See this patch for the introduction of the type: https://reviews.llvm.org/D64790 Differential Revision: https://reviews.llvm.org/D82785
2020-06-30 16:16:59 +08:00
return allowsMemoryAccess(Context, DL, VT, MMO.getAddrSpace(), MMO.getAlign(),
MMO.getFlags(), Fast);
[TargetLowering] Add allowsMemoryAccess(MachineMemOperand) helper wrapper. NFCI. As suggested by @arsenm on D63075 - this adds a TargetLowering::allowsMemoryAccess wrapper that takes a Load/Store node's MachineMemOperand to handle the AddressSpace/Alignment arguments and will also implicitly handle the MachineMemOperand::Flags change in D63075. llvm-svn: 363048
2019-06-11 19:00:23 +08:00
}
[CodeGenPrepare] use branch weight metadata to decide if a select should be turned into a branch This is part of solving PR27344: https://llvm.org/bugs/show_bug.cgi?id=27344 CGP should undo the SimplifyCFG transform for the same reason that earlier patches have used this same mechanism: it's possible that passes between SimplifyCFG and CGP may be able to optimize the IR further with a select in place. For the TLI hook default, >99% taken or not taken is chosen as the default threshold for a highly predictable branch. Even the most limited HW branch predictors will be correct on this branch almost all the time, so even a massive mispredict penalty perf loss would be overcome by the win from all the times the branch was predicted correctly. As a follow-up, we could make the default target hook less conservative by using the SchedMachineModel's MispredictPenalty. Or we could just let targets override the default by implementing the hook with that and other target-specific options. Note that trying to statically determine mispredict rates for close-to-balanced profile weight data is generally impossible if the HW is sufficiently advanced. Ie, 50/50 taken/not-taken might still be 100% predictable. Finally, note that this patch as-is will not solve PR27344 because the current __builtin_unpredictable() branch weight default values are 4 and 64. A proposal to change that is in D19435. Differential Revision: http://reviews.llvm.org/D19488 llvm-svn: 267572
2016-04-27 01:11:17 +08:00
BranchProbability TargetLoweringBase::getPredictableBranchThreshold() const {
return BranchProbability(MinPercentageForPredictableBranch, 100);
}
move DAGCombiner's allowableAlignment() helper function into the TLI Making allowableAlignment() more accessible was suggested as a predecessor patch for D10662, so I've pulled it into TargetLowering. This let's us remove 4 instances of duplicate logic in LegalizeDAG. There's a subtle functional change in the implementation: the existing allowableAlignment() code was using getPrefTypeAlignment() when checking alignment with the DataLayout and assumed that was fast. In this implementation, we use getABITypeAlignment() and assume that is fast. See the TODO comment or the discussion in the Phab review for future improvements in this implementation (don't use the data layout at all). There are no regression test changes from this difference, and I'm not sure how to expose it via a test. I think we actually do want to provide the 'Fast' param when checking this from DAGCombiner::MergeConsecutiveStores(). Ie, we shouldn't merge stores if the new stores are not going to be fast. But that change will require fixing allowsMisalignedMemoryAccess() overrides as noted in D10662. Differential Revision: http://reviews.llvm.org/D10905 llvm-svn: 243549
2015-07-30 02:24:18 +08:00
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
//===----------------------------------------------------------------------===//
// TargetTransformInfo Helpers
//===----------------------------------------------------------------------===//
int TargetLoweringBase::InstructionOpcodeToISD(unsigned Opcode) const {
enum InstructionOpcodes {
#define HANDLE_INST(NUM, OPCODE, CLASS) OPCODE = NUM,
#define LAST_OTHER_INST(NUM) InstructionOpcodesCount = NUM
#include "llvm/IR/Instruction.def"
};
switch (static_cast<InstructionOpcodes>(Opcode)) {
case Ret: return 0;
case Br: return 0;
case Switch: return 0;
case IndirectBr: return 0;
case Invoke: return 0;
Implementation of asm-goto support in LLVM This patch accompanies the RFC posted here: http://lists.llvm.org/pipermail/llvm-dev/2018-October/127239.html This patch adds a new CallBr IR instruction to support asm-goto inline assembly like gcc as used by the linux kernel. This instruction is both a call instruction and a terminator instruction with multiple successors. Only inline assembly usage is supported today. This also adds a new INLINEASM_BR opcode to SelectionDAG and MachineIR to represent an INLINEASM block that is also considered a terminator instruction. There will likely be more bug fixes and optimizations to follow this, but we felt it had reached a point where we would like to switch to an incremental development model. Patch by Craig Topper, Alexander Ivchenko, Mikhail Dvoretckii Differential Revision: https://reviews.llvm.org/D53765 llvm-svn: 353563
2019-02-09 04:48:56 +08:00
case CallBr: return 0;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
case Resume: return 0;
case Unreachable: return 0;
New EH representation for MSVC compatibility This introduces new instructions neccessary to implement MSVC-compatible exception handling support. Most of the middle-end and none of the back-end haven't been audited or updated to take them into account. Differential Revision: http://reviews.llvm.org/D11097 llvm-svn: 243766
2015-08-01 01:58:14 +08:00
case CleanupRet: return 0;
case CatchRet: return 0;
[IR] Reformulate LLVM's EH funclet IR While we have successfully implemented a funclet-oriented EH scheme on top of LLVM IR, our scheme has some notable deficiencies: - catchendpad and cleanupendpad are necessary in the current design but they are difficult to explain to others, even to seasoned LLVM experts. - catchendpad and cleanupendpad are optimization barriers. They cannot be split and force all potentially throwing call-sites to be invokes. This has a noticable effect on the quality of our code generation. - catchpad, while similar in some aspects to invoke, is fairly awkward. It is unsplittable, starts a funclet, and has control flow to other funclets. - The nesting relationship between funclets is currently a property of control flow edges. Because of this, we are forced to carefully analyze the flow graph to see if there might potentially exist illegal nesting among funclets. While we have logic to clone funclets when they are illegally nested, it would be nicer if we had a representation which forbade them upfront. Let's clean this up a bit by doing the following: - Instead, make catchpad more like cleanuppad and landingpad: no control flow, just a bunch of simple operands; catchpad would be splittable. - Introduce catchswitch, a control flow instruction designed to model the constraints of funclet oriented EH. - Make funclet scoping explicit by having funclet instructions consume the token produced by the funclet which contains them. - Remove catchendpad and cleanupendpad. Their presence can be inferred implicitly using coloring information. N.B. The state numbering code for the CLR has been updated but the veracity of it's output cannot be spoken for. An expert should take a look to make sure the results are reasonable. Reviewers: rnk, JosephTremoulet, andrew.w.kaylor Differential Revision: http://reviews.llvm.org/D15139 llvm-svn: 255422
2015-12-12 13:38:55 +08:00
case CatchPad: return 0;
case CatchSwitch: return 0;
case CleanupPad: return 0;
[IR] Add a dedicated FNeg IR Instruction The IEEE-754 Standard makes it clear that fneg(x) and fsub(-0.0, x) are two different operations. The former is a bitwise operation, while the latter is an arithmetic operation. This patch creates a dedicated FNeg IR Instruction to model that behavior. Differential Revision: https://reviews.llvm.org/D53877 llvm-svn: 346774
2018-11-14 02:15:47 +08:00
case FNeg: return ISD::FNEG;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
case Add: return ISD::ADD;
case FAdd: return ISD::FADD;
case Sub: return ISD::SUB;
case FSub: return ISD::FSUB;
case Mul: return ISD::MUL;
case FMul: return ISD::FMUL;
case UDiv: return ISD::UDIV;
X86TTI: Adjust sdiv cost now that we can lower it on plain SSE2. Includes a fix for a horrible typo that caused all SDIV costs to be slightly off :) llvm-svn: 207371
2014-04-28 02:47:54 +08:00
case SDiv: return ISD::SDIV;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
case FDiv: return ISD::FDIV;
case URem: return ISD::UREM;
case SRem: return ISD::SREM;
case FRem: return ISD::FREM;
case Shl: return ISD::SHL;
case LShr: return ISD::SRL;
case AShr: return ISD::SRA;
case And: return ISD::AND;
case Or: return ISD::OR;
case Xor: return ISD::XOR;
case Alloca: return 0;
case Load: return ISD::LOAD;
case Store: return ISD::STORE;
case GetElementPtr: return 0;
case Fence: return 0;
case AtomicCmpXchg: return 0;
case AtomicRMW: return 0;
case Trunc: return ISD::TRUNCATE;
case ZExt: return ISD::ZERO_EXTEND;
case SExt: return ISD::SIGN_EXTEND;
case FPToUI: return ISD::FP_TO_UINT;
case FPToSI: return ISD::FP_TO_SINT;
case UIToFP: return ISD::UINT_TO_FP;
case SIToFP: return ISD::SINT_TO_FP;
case FPTrunc: return ISD::FP_ROUND;
case FPExt: return ISD::FP_EXTEND;
case PtrToInt: return ISD::BITCAST;
case IntToPtr: return ISD::BITCAST;
case BitCast: return ISD::BITCAST;
Add addrspacecast instruction. Patch by Michele Scandale! llvm-svn: 194760
2013-11-15 09:34:59 +08:00
case AddrSpaceCast: return ISD::ADDRSPACECAST;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
case ICmp: return ISD::SETCC;
case FCmp: return ISD::SETCC;
case PHI: return 0;
case Call: return 0;
case Select: return ISD::SELECT;
case UserOp1: return 0;
case UserOp2: return 0;
case VAArg: return 0;
case ExtractElement: return ISD::EXTRACT_VECTOR_ELT;
case InsertElement: return ISD::INSERT_VECTOR_ELT;
case ShuffleVector: return ISD::VECTOR_SHUFFLE;
case ExtractValue: return ISD::MERGE_VALUES;
case InsertValue: return ISD::MERGE_VALUES;
case LandingPad: return 0;
[SelDag] Add FREEZE Summary: - Add FREEZE node to SelDag - Lower FreezeInst (in IR) to FREEZE node - Add Legalization for FREEZE node Reviewers: qcolombet, bogner, efriedma, lebedev.ri, nlopes, craig.topper, arsenm Reviewed By: lebedev.ri Subscribers: wdng, xbolva00, Petar.Avramovic, liuz, lkail, dylanmckay, hiraditya, Jim, arsenm, craig.topper, RKSimon, spatel, lebedev.ri, regehr, trentxintong, nlopes, mkuper, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D29014
2020-03-23 12:03:13 +08:00
case Freeze: return ISD::FREEZE;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
llvm_unreachable("Unknown instruction type encountered!");
}
[TTI] Make the cost APIs in TargetTransformInfo consistently use 'int' rather than 'unsigned' for their costs. For something like costs in particular there is a natural "negative" value, that of savings or saved cost. As a consequence, there is a lot of code that subtracts or creates negative values based on cost, all of which is prone to awkwardness or bugs when dealing with an unsigned type. Similarly, we *never* want these values to wrap, as that would cause Very Bad code generation (likely percieved as an infinite loop as we try to emit over 2^32 instructions or some such insanity). All around 'int' seems a much better fit for these basic metrics. I've added asserts to ensure that at least the TTI interface never returns negative numbers here. If we ever have a use case for negative numbers, we can remove this, but this way a bug where someone used '-1' to produce a 'very large' cost will be caught by the assert. This passes all tests, and is also UBSan clean. No functional change intended. Differential Revision: http://reviews.llvm.org/D11741 llvm-svn: 244080
2015-08-06 02:08:10 +08:00
std::pair<int, MVT>
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
TargetLoweringBase::getTypeLegalizationCost(const DataLayout &DL,
Type *Ty) const {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
LLVMContext &C = Ty->getContext();
Make TargetLowering::getPointerTy() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, ted, yaron.keren, rafael, llvm-commits Differential Revision: http://reviews.llvm.org/D11028 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241775
2015-07-09 10:09:04 +08:00
EVT MTy = getValueType(DL, Ty);
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
[TTI] Make the cost APIs in TargetTransformInfo consistently use 'int' rather than 'unsigned' for their costs. For something like costs in particular there is a natural "negative" value, that of savings or saved cost. As a consequence, there is a lot of code that subtracts or creates negative values based on cost, all of which is prone to awkwardness or bugs when dealing with an unsigned type. Similarly, we *never* want these values to wrap, as that would cause Very Bad code generation (likely percieved as an infinite loop as we try to emit over 2^32 instructions or some such insanity). All around 'int' seems a much better fit for these basic metrics. I've added asserts to ensure that at least the TTI interface never returns negative numbers here. If we ever have a use case for negative numbers, we can remove this, but this way a bug where someone used '-1' to produce a 'very large' cost will be caught by the assert. This passes all tests, and is also UBSan clean. No functional change intended. Differential Revision: http://reviews.llvm.org/D11741 llvm-svn: 244080
2015-08-06 02:08:10 +08:00
int Cost = 1;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// We keep legalizing the type until we find a legal kind. We assume that
// the only operation that costs anything is the split. After splitting
// we need to handle two types.
while (true) {
LegalizeKind LK = getTypeConversion(C, MTy);
if (LK.first == TypeLegal)
return std::make_pair(Cost, MTy.getSimpleVT());
if (LK.first == TypeSplitVector || LK.first == TypeExpandInteger)
Cost *= 2;
[X86] Part 1 to fix x86-64 fp128 calling convention. Almost all these changes are conditioned and only apply to the new x86-64 f128 type configuration, which will be enabled in a follow up patch. They are required together to make new f128 work. If there is any error, we should fix or revert them as a whole. These changes should have no impact to current configurations. * Relax type legalization checks to accept new f128 type configuration, whose TypeAction is TypeSoftenFloat, not TypeLegal, but also has TLI.isTypeLegal true. * Relax GetSoftenedFloat to return in some cases f128 type SDValue, which is TLI.isTypeLegal but not "softened" to i128 node. * Allow customized FABS, FNEG, FCOPYSIGN on new f128 type configuration, to generate optimized bitwise operators for libm functions. * Enhance related Lower* functions to handle f128 type. * Enhance DAGTypeLegalizer::run, SoftenFloatResult, and related functions to keep new f128 type in register, and convert f128 operators to library calls. * Fix Combiner, Emitter, Legalizer routines that did not handle f128 type. * Add ExpandConstant to handle i128 constants, ExpandNode to handle ISD::Constant node. * Add one more parameter to getCommonSubClass and firstCommonClass, to guarantee that returned common sub class will contain the specified simple value type. This extra parameter is used by EmitCopyFromReg in InstrEmitter.cpp. * Fix infinite loop in getTypeLegalizationCost when f128 is the value type. * Fix printOperand to handle null operand. * Enhance ISD::BITCAST node to handle f128 constant. * Expand new f128 type for BR_CC, SELECT_CC, SELECT, SETCC nodes. * Enhance X86AsmPrinter to emit f128 values in comments. Differential Revision: http://reviews.llvm.org/D15134 llvm-svn: 254653
2015-12-04 06:02:40 +08:00
// Do not loop with f128 type.
if (MTy == LK.second)
return std::make_pair(Cost, MTy.getSimpleVT());
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// Keep legalizing the type.
MTy = LK.second;
}
}
[safestack] Use non-thread-local unsafe stack pointer for Contiki OS Patch by Michael LeMay Differential revision: http://reviews.llvm.org/D19852 llvm-svn: 284254
2016-10-15 01:56:00 +08:00
Value *TargetLoweringBase::getDefaultSafeStackPointerLocation(IRBuilder<> &IRB,
bool UseTLS) const {
// compiler-rt provides a variable with a magic name. Targets that do not
// link with compiler-rt may also provide such a variable.
Module *M = IRB.GetInsertBlock()->getParent()->getParent();
const char *UnsafeStackPtrVar = "__safestack_unsafe_stack_ptr";
auto UnsafeStackPtr =
dyn_cast_or_null<GlobalVariable>(M->getNamedValue(UnsafeStackPtrVar));
Type *StackPtrTy = Type::getInt8PtrTy(M->getContext());
if (!UnsafeStackPtr) {
auto TLSModel = UseTLS ?
GlobalValue::InitialExecTLSModel :
GlobalValue::NotThreadLocal;
// The global variable is not defined yet, define it ourselves.
// We use the initial-exec TLS model because we do not support the
// variable living anywhere other than in the main executable.
UnsafeStackPtr = new GlobalVariable(
*M, StackPtrTy, false, GlobalValue::ExternalLinkage, nullptr,
UnsafeStackPtrVar, nullptr, TLSModel);
} else {
// The variable exists, check its type and attributes.
if (UnsafeStackPtr->getValueType() != StackPtrTy)
report_fatal_error(Twine(UnsafeStackPtrVar) + " must have void* type");
if (UseTLS != UnsafeStackPtr->isThreadLocal())
report_fatal_error(Twine(UnsafeStackPtrVar) + " must " +
(UseTLS ? "" : "not ") + "be thread-local");
}
return UnsafeStackPtr;
}
[safestack] Fast access to the unsafe stack pointer on AArch64/Android. Android libc provides a fixed TLS slot for the unsafe stack pointer, and this change implements direct access to that slot on AArch64 via __builtin_thread_pointer() + offset. This change also moves more code into TargetLowering and its target-specific subclasses to get rid of target-specific codegen in SafeStackPass. This change does not touch the ARM backend because ARM lowers builting_thread_pointer as aeabi_read_tp, which is not available on Android. The previous iteration of this change was reverted in r250461. This version leaves the generic, compiler-rt based implementation in SafeStack.cpp instead of moving it to TargetLoweringBase in order to allow testing without a TargetMachine. llvm-svn: 251324
2015-10-27 02:28:25 +08:00
Value *TargetLoweringBase::getSafeStackPointerLocation(IRBuilder<> &IRB) const {
if (!TM.getTargetTriple().isAndroid())
[safestack] Use non-thread-local unsafe stack pointer for Contiki OS Patch by Michael LeMay Differential revision: http://reviews.llvm.org/D19852 llvm-svn: 284254
2016-10-15 01:56:00 +08:00
return getDefaultSafeStackPointerLocation(IRB, true);
[safestack] Fast access to the unsafe stack pointer on AArch64/Android. Android libc provides a fixed TLS slot for the unsafe stack pointer, and this change implements direct access to that slot on AArch64 via __builtin_thread_pointer() + offset. This change also moves more code into TargetLowering and its target-specific subclasses to get rid of target-specific codegen in SafeStackPass. This change does not touch the ARM backend because ARM lowers builting_thread_pointer as aeabi_read_tp, which is not available on Android. The previous iteration of this change was reverted in r250461. This version leaves the generic, compiler-rt based implementation in SafeStack.cpp instead of moving it to TargetLoweringBase in order to allow testing without a TargetMachine. llvm-svn: 251324
2015-10-27 02:28:25 +08:00
// Android provides a libc function to retrieve the address of the current
// thread's unsafe stack pointer.
Module *M = IRB.GetInsertBlock()->getParent()->getParent();
Type *StackPtrTy = Type::getInt8PtrTy(M->getContext());
[opaque pointer types] Add a FunctionCallee wrapper type, and use it. Recommit r352791 after tweaking DerivedTypes.h slightly, so that gcc doesn't choke on it, hopefully. Original Message: The FunctionCallee type is effectively a {FunctionType*,Value*} pair, and is a useful convenience to enable code to continue passing the result of getOrInsertFunction() through to EmitCall, even once pointer types lose their pointee-type. Then: - update the CallInst/InvokeInst instruction creation functions to take a Callee, - modify getOrInsertFunction to return FunctionCallee, and - update all callers appropriately. One area of particular note is the change to the sanitizer code. Previously, they had been casting the result of `getOrInsertFunction` to a `Function*` via `checkSanitizerInterfaceFunction`, and storing that. That would report an error if someone had already inserted a function declaraction with a mismatching signature. However, in general, LLVM allows for such mismatches, as `getOrInsertFunction` will automatically insert a bitcast if needed. As part of this cleanup, cause the sanitizer code to do the same. (It will call its functions using the expected signature, however they may have been declared.) Finally, in a small number of locations, callers of `getOrInsertFunction` actually were expecting/requiring that a brand new function was being created. In such cases, I've switched them to Function::Create instead. Differential Revision: https://reviews.llvm.org/D57315 llvm-svn: 352827
2019-02-01 10:28:03 +08:00
FunctionCallee Fn = M->getOrInsertFunction("__safestack_pointer_address",
StackPtrTy->getPointerTo(0));
[safestack] Fast access to the unsafe stack pointer on AArch64/Android. Android libc provides a fixed TLS slot for the unsafe stack pointer, and this change implements direct access to that slot on AArch64 via __builtin_thread_pointer() + offset. This change also moves more code into TargetLowering and its target-specific subclasses to get rid of target-specific codegen in SafeStackPass. This change does not touch the ARM backend because ARM lowers builting_thread_pointer as aeabi_read_tp, which is not available on Android. The previous iteration of this change was reverted in r250461. This version leaves the generic, compiler-rt based implementation in SafeStack.cpp instead of moving it to TargetLoweringBase in order to allow testing without a TargetMachine. llvm-svn: 251324
2015-10-27 02:28:25 +08:00
return IRB.CreateCall(Fn);
}
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
//===----------------------------------------------------------------------===//
// Loop Strength Reduction hooks
//===----------------------------------------------------------------------===//
/// isLegalAddressingMode - Return true if the addressing mode represented
/// by AM is legal for this target, for a load/store of the specified type.
Make isLegalAddressingMode() taking DataLayout as an argument Summary: This change is part of a series of commits dedicated to have a single DataLayout during compilation by using always the one owned by the module. Reviewers: echristo Subscribers: jholewinski, llvm-commits, rafael, yaron.keren Differential Revision: http://reviews.llvm.org/D11040 From: Mehdi Amini <mehdi.amini@apple.com> llvm-svn: 241778
2015-07-09 10:09:40 +08:00
bool TargetLoweringBase::isLegalAddressingMode(const DataLayout &DL,
const AddrMode &AM, Type *Ty,
[SystemZ, LoopStrengthReduce] This patch makes LSR generate better code for SystemZ in the cases of memory intrinsics, Load->Store pairs or comparison of immediate with memory. In order to achieve this, the following common code changes were made: * New TTI hook: LSRWithInstrQueries(), which defaults to false. Controls if LSR should do instruction-based addressing evaluations by calling isLegalAddressingMode() with the Instruction pointers. * In LoopStrengthReduce: handle address operands of memset, memmove and memcpy as address uses, and call isFoldableMemAccessOffset() for any LSRUse::Address, not just loads or stores. SystemZ changes: * isLSRCostLess() implemented with Insns first, and without ImmCost. * New function supportedAddressingMode() that is a helper for TTI methods looking at Instructions passed via pointers. Review: Ulrich Weigand, Quentin Colombet https://reviews.llvm.org/D35262 https://reviews.llvm.org/D35049 llvm-svn: 308729
2017-07-21 19:59:37 +08:00
unsigned AS, Instruction *I) const {
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
// The default implementation of this implements a conservative RISCy, r+r and
// r+i addr mode.
// Allows a sign-extended 16-bit immediate field.
if (AM.BaseOffs <= -(1LL << 16) || AM.BaseOffs >= (1LL << 16)-1)
return false;
// No global is ever allowed as a base.
if (AM.BaseGV)
return false;
// Only support r+r,
switch (AM.Scale) {
case 0: // "r+i" or just "i", depending on HasBaseReg.
break;
case 1:
if (AM.HasBaseReg && AM.BaseOffs) // "r+r+i" is not allowed.
return false;
// Otherwise we have r+r or r+i.
break;
case 2:
if (AM.HasBaseReg || AM.BaseOffs) // 2*r+r or 2*r+i is not allowed.
return false;
// Allow 2*r as r+r.
break;
TargetLowering: n * r where n > 2 should be an illegal addressing mode llvm-svn: 201433
2014-02-15 05:10:34 +08:00
default: // Don't allow n * r
return false;
Split TargetLowering into a CodeGen and a SelectionDAG part. This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still a complete mess but as long as the edges consist of virtual call it doesn't cause breakage. BasicTTI did static calls and thus broke some build configurations. llvm-svn: 172246
2013-01-12 04:05:37 +08:00
}
return true;
}
[SSP] Remove llvm.stackprotectorcheck. This is a cleanup patch for SSP support in LLVM. There is no functional change. llvm.stackprotectorcheck is not needed, because SelectionDAG isn't actually lowering it in SelectBasicBlock; rather, it adds check code in FinishBasicBlock, ignoring the position where the intrinsic is inserted (See FindSplitPointForStackProtector()). llvm-svn: 265851
2016-04-09 05:26:31 +08:00
//===----------------------------------------------------------------------===//
// Stack Protector
//===----------------------------------------------------------------------===//
// For OpenBSD return its special guard variable. Otherwise return nullptr,
// so that SelectionDAG handle SSP.
Value *TargetLoweringBase::getIRStackGuard(IRBuilder<> &IRB) const {
if (getTargetMachine().getTargetTriple().isOSOpenBSD()) {
Module &M = *IRB.GetInsertBlock()->getParent()->getParent();
PointerType *PtrTy = Type::getInt8PtrTy(M.getContext());
[SSP] Do not set __guard_local to hidden for OpenBSD SSP __guard_local is defined as long on OpenBSD. If the source file contains a definition of __guard_local, it mismatches with the int8 pointer type used in LLVM. In that case, Module::getOrInsertGlobal() returns a cast operation instead of a GlobalVariable. Trying to set the visibility on the cast operation leads to random segfaults (seen when compiling the OpenBSD kernel, which also runs with stack protection). In the kernel, the hidden attribute does not matter. For userspace code, __guard_local is defined as hidden in the startup code. If a program re-defines __guard_local, the definition from the startup code will either win or the linker complains about multiple definitions (depending on whether the re-defined __guard_local is placed in the common segment or not). It also matches what gcc on OpenBSD does. Thanks Stefan Kempf <sisnkemp@gmail.com> for the patch! Differential Revision: http://reviews.llvm.org/D23674 llvm-svn: 279449
2016-08-23 02:26:27 +08:00
return M.getOrInsertGlobal("__guard_local", PtrTy);
[SSP] Remove llvm.stackprotectorcheck. This is a cleanup patch for SSP support in LLVM. There is no functional change. llvm.stackprotectorcheck is not needed, because SelectionDAG isn't actually lowering it in SelectBasicBlock; rather, it adds check code in FinishBasicBlock, ignoring the position where the intrinsic is inserted (See FindSplitPointForStackProtector()). llvm-svn: 265851
2016-04-09 05:26:31 +08:00
}
return nullptr;
}
// Currently only support "standard" __stack_chk_guard.
// TODO: add LOAD_STACK_GUARD support.
void TargetLoweringBase::insertSSPDeclarations(Module &M) const {
[AArch64] Don't crash trying to resolve __stack_chk_guard. In certain cases, the compiler might try to merge __stack_chk_guard with another global variable. (Or someone could theoretically define __stack_chk_guard as an alias.) In that case, make sure we don't crash. Differential Revision: https://reviews.llvm.org/D45746 llvm-svn: 330495
2018-04-21 08:07:46 +08:00
if (!M.getNamedValue("__stack_chk_guard"))
new GlobalVariable(M, Type::getInt8PtrTy(M.getContext()), false,
GlobalVariable::ExternalLinkage,
nullptr, "__stack_chk_guard");
[SSP] Remove llvm.stackprotectorcheck. This is a cleanup patch for SSP support in LLVM. There is no functional change. llvm.stackprotectorcheck is not needed, because SelectionDAG isn't actually lowering it in SelectBasicBlock; rather, it adds check code in FinishBasicBlock, ignoring the position where the intrinsic is inserted (See FindSplitPointForStackProtector()). llvm-svn: 265851
2016-04-09 05:26:31 +08:00
}
// Currently only support "standard" __stack_chk_guard.
// TODO: add LOAD_STACK_GUARD support.
[PPC, SSP] Support PowerPC Linux stack protection. llvm-svn: 266809
2016-04-20 04:14:52 +08:00
Value *TargetLoweringBase::getSDagStackGuard(const Module &M) const {
[AArch64] Don't crash trying to resolve __stack_chk_guard. In certain cases, the compiler might try to merge __stack_chk_guard with another global variable. (Or someone could theoretically define __stack_chk_guard as an alias.) In that case, make sure we don't crash. Differential Revision: https://reviews.llvm.org/D45746 llvm-svn: 330495
2018-04-21 08:07:46 +08:00
return M.getNamedValue("__stack_chk_guard");
[SSP] Remove llvm.stackprotectorcheck. This is a cleanup patch for SSP support in LLVM. There is no functional change. llvm.stackprotectorcheck is not needed, because SelectionDAG isn't actually lowering it in SelectBasicBlock; rather, it adds check code in FinishBasicBlock, ignoring the position where the intrinsic is inserted (See FindSplitPointForStackProtector()). llvm-svn: 265851
2016-04-09 05:26:31 +08:00
}
[stack-protection] Add support for MSVC buffer security check Summary: This patch is adding support for the MSVC buffer security check implementation The buffer security check is turned on with the '/GS' compiler switch. * https://msdn.microsoft.com/en-us/library/8dbf701c.aspx * To be added to clang here: http://reviews.llvm.org/D20347 Some overview of buffer security check feature and implementation: * https://msdn.microsoft.com/en-us/library/aa290051(VS.71).aspx * http://www.ksyash.com/2011/01/buffer-overflow-protection-3/ * http://blog.osom.info/2012/02/understanding-vs-c-compilers-buffer.html For the following example: ``` int example(int offset, int index) { char buffer[10]; memset(buffer, 0xCC, index); return buffer[index]; } ``` The MSVC compiler is adding these instructions to perform stack integrity check: ``` push ebp mov ebp,esp sub esp,50h [1] mov eax,dword ptr [__security_cookie (01068024h)] [2] xor eax,ebp [3] mov dword ptr [ebp-4],eax push ebx push esi push edi mov eax,dword ptr [index] push eax push 0CCh lea ecx,[buffer] push ecx call _memset (010610B9h) add esp,0Ch mov eax,dword ptr [index] movsx eax,byte ptr buffer[eax] pop edi pop esi pop ebx [4] mov ecx,dword ptr [ebp-4] [5] xor ecx,ebp [6] call @__security_check_cookie@4 (01061276h) mov esp,ebp pop ebp ret ``` The instrumentation above is: * [1] is loading the global security canary, * [3] is storing the local computed ([2]) canary to the guard slot, * [4] is loading the guard slot and ([5]) re-compute the global canary, * [6] is validating the resulting canary with the '__security_check_cookie' and performs error handling. Overview of the current stack-protection implementation: * lib/CodeGen/StackProtector.cpp * There is a default stack-protection implementation applied on intermediate representation. * The target can overload 'getIRStackGuard' method if it has a standard location for the stack protector cookie. * An intrinsic 'Intrinsic::stackprotector' is added to the prologue. It will be expanded by the instruction selection pass (DAG or Fast). * Basic Blocks are added to every instrumented function to receive the code for handling stack guard validation and errors handling. * Guard manipulation and comparison are added directly to the intermediate representation. * lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp * lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp * There is an implementation that adds instrumentation during instruction selection (for better handling of sibbling calls). * see long comment above 'class StackProtectorDescriptor' declaration. * The target needs to override 'getSDagStackGuard' to activate SDAG stack protection generation. (note: getIRStackGuard MUST be nullptr). * 'getSDagStackGuard' returns the appropriate stack guard (security cookie) * The code is generated by 'SelectionDAGBuilder.cpp' and 'SelectionDAGISel.cpp'. * include/llvm/Target/TargetLowering.h * Contains function to retrieve the default Guard 'Value'; should be overriden by each target to select which implementation is used and provide Guard 'Value'. * lib/Target/X86/X86ISelLowering.cpp * Contains the x86 specialisation; Guard 'Value' used by the SelectionDAG algorithm. Function-based Instrumentation: * The MSVC doesn't inline the stack guard comparison in every function. Instead, a call to '__security_check_cookie' is added to the epilogue before every return instructions. * To support function-based instrumentation, this patch is * adding a function to get the function-based check (llvm 'Value', see include/llvm/Target/TargetLowering.h), * If provided, the stack protection instrumentation won't be inlined and a call to that function will be added to the prologue. * modifying (SelectionDAGISel.cpp) do avoid producing basic blocks used for inline instrumentation, * generating the function-based instrumentation during the ISEL pass (SelectionDAGBuilder.cpp), * if FastISEL (not SelectionDAG), using the fallback which rely on the same function-based implemented over intermediate representation (StackProtector.cpp). Modifications * adding support for MSVC (lib/Target/X86/X86ISelLowering.cpp) * adding support function-based instrumentation (lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp, .h) Results * IR generated instrumentation: ``` clang-cl /GS test.cc /Od /c -mllvm -print-isel-input ``` ``` *** Final LLVM Code input to ISel *** ; Function Attrs: nounwind sspstrong define i32 @"\01?example@@YAHHH@Z"(i32 %offset, i32 %index) #0 { entry: %StackGuardSlot = alloca i8* <<<-- Allocated guard slot %0 = call i8* @llvm.stackguard() <<<-- Loading Stack Guard value call void @llvm.stackprotector(i8* %0, i8** %StackGuardSlot) <<<-- Prologue intrinsic call (store to Guard slot) %index.addr = alloca i32, align 4 %offset.addr = alloca i32, align 4 %buffer = alloca [10 x i8], align 1 store i32 %index, i32* %index.addr, align 4 store i32 %offset, i32* %offset.addr, align 4 %arraydecay = getelementptr inbounds [10 x i8], [10 x i8]* %buffer, i32 0, i32 0 %1 = load i32, i32* %index.addr, align 4 call void @llvm.memset.p0i8.i32(i8* %arraydecay, i8 -52, i32 %1, i32 1, i1 false) %2 = load i32, i32* %index.addr, align 4 %arrayidx = getelementptr inbounds [10 x i8], [10 x i8]* %buffer, i32 0, i32 %2 %3 = load i8, i8* %arrayidx, align 1 %conv = sext i8 %3 to i32 %4 = load volatile i8*, i8** %StackGuardSlot <<<-- Loading Guard slot call void @__security_check_cookie(i8* %4) <<<-- Epilogue function-based check ret i32 %conv } ``` * SelectionDAG generated instrumentation: ``` clang-cl /GS test.cc /O1 /c /FA ``` ``` "?example@@YAHHH@Z": # @"\01?example@@YAHHH@Z" # BB#0: # %entry pushl %esi subl $16, %esp movl ___security_cookie, %eax <<<-- Loading Stack Guard value movl 28(%esp), %esi movl %eax, 12(%esp) <<<-- Store to Guard slot leal 2(%esp), %eax pushl %esi pushl $204 pushl %eax calll _memset addl $12, %esp movsbl 2(%esp,%esi), %esi movl 12(%esp), %ecx <<<-- Loading Guard slot calll @__security_check_cookie@4 <<<-- Epilogue function-based check movl %esi, %eax addl $16, %esp popl %esi retl ``` Reviewers: kcc, pcc, eugenis, rnk Subscribers: majnemer, llvm-commits, hans, thakis, rnk Differential Revision: http://reviews.llvm.org/D20346 llvm-svn: 272053
2016-06-08 04:15:35 +08:00
[opaque pointer types] Pass function types to CallInst creation. This cleans up all CallInst creation in LLVM to explicitly pass a function type rather than deriving it from the pointer's element-type. Differential Revision: https://reviews.llvm.org/D57170 llvm-svn: 352909
2019-02-02 04:43:25 +08:00
Function *TargetLoweringBase::getSSPStackGuardCheck(const Module &M) const {
[stack-protection] Add support for MSVC buffer security check Summary: This patch is adding support for the MSVC buffer security check implementation The buffer security check is turned on with the '/GS' compiler switch. * https://msdn.microsoft.com/en-us/library/8dbf701c.aspx * To be added to clang here: http://reviews.llvm.org/D20347 Some overview of buffer security check feature and implementation: * https://msdn.microsoft.com/en-us/library/aa290051(VS.71).aspx * http://www.ksyash.com/2011/01/buffer-overflow-protection-3/ * http://blog.osom.info/2012/02/understanding-vs-c-compilers-buffer.html For the following example: ``` int example(int offset, int index) { char buffer[10]; memset(buffer, 0xCC, index); return buffer[index]; } ``` The MSVC compiler is adding these instructions to perform stack integrity check: ``` push ebp mov ebp,esp sub esp,50h [1] mov eax,dword ptr [__security_cookie (01068024h)] [2] xor eax,ebp [3] mov dword ptr [ebp-4],eax push ebx push esi push edi mov eax,dword ptr [index] push eax push 0CCh lea ecx,[buffer] push ecx call _memset (010610B9h) add esp,0Ch mov eax,dword ptr [index] movsx eax,byte ptr buffer[eax] pop edi pop esi pop ebx [4] mov ecx,dword ptr [ebp-4] [5] xor ecx,ebp [6] call @__security_check_cookie@4 (01061276h) mov esp,ebp pop ebp ret ``` The instrumentation above is: * [1] is loading the global security canary, * [3] is storing the local computed ([2]) canary to the guard slot, * [4] is loading the guard slot and ([5]) re-compute the global canary, * [6] is validating the resulting canary with the '__security_check_cookie' and performs error handling. Overview of the current stack-protection implementation: * lib/CodeGen/StackProtector.cpp * There is a default stack-protection implementation applied on intermediate representation. * The target can overload 'getIRStackGuard' method if it has a standard location for the stack protector cookie. * An intrinsic 'Intrinsic::stackprotector' is added to the prologue. It will be expanded by the instruction selection pass (DAG or Fast). * Basic Blocks are added to every instrumented function to receive the code for handling stack guard validation and errors handling. * Guard manipulation and comparison are added directly to the intermediate representation. * lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp * lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp * There is an implementation that adds instrumentation during instruction selection (for better handling of sibbling calls). * see long comment above 'class StackProtectorDescriptor' declaration. * The target needs to override 'getSDagStackGuard' to activate SDAG stack protection generation. (note: getIRStackGuard MUST be nullptr). * 'getSDagStackGuard' returns the appropriate stack guard (security cookie) * The code is generated by 'SelectionDAGBuilder.cpp' and 'SelectionDAGISel.cpp'. * include/llvm/Target/TargetLowering.h * Contains function to retrieve the default Guard 'Value'; should be overriden by each target to select which implementation is used and provide Guard 'Value'. * lib/Target/X86/X86ISelLowering.cpp * Contains the x86 specialisation; Guard 'Value' used by the SelectionDAG algorithm. Function-based Instrumentation: * The MSVC doesn't inline the stack guard comparison in every function. Instead, a call to '__security_check_cookie' is added to the epilogue before every return instructions. * To support function-based instrumentation, this patch is * adding a function to get the function-based check (llvm 'Value', see include/llvm/Target/TargetLowering.h), * If provided, the stack protection instrumentation won't be inlined and a call to that function will be added to the prologue. * modifying (SelectionDAGISel.cpp) do avoid producing basic blocks used for inline instrumentation, * generating the function-based instrumentation during the ISEL pass (SelectionDAGBuilder.cpp), * if FastISEL (not SelectionDAG), using the fallback which rely on the same function-based implemented over intermediate representation (StackProtector.cpp). Modifications * adding support for MSVC (lib/Target/X86/X86ISelLowering.cpp) * adding support function-based instrumentation (lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp, .h) Results * IR generated instrumentation: ``` clang-cl /GS test.cc /Od /c -mllvm -print-isel-input ``` ``` *** Final LLVM Code input to ISel *** ; Function Attrs: nounwind sspstrong define i32 @"\01?example@@YAHHH@Z"(i32 %offset, i32 %index) #0 { entry: %StackGuardSlot = alloca i8* <<<-- Allocated guard slot %0 = call i8* @llvm.stackguard() <<<-- Loading Stack Guard value call void @llvm.stackprotector(i8* %0, i8** %StackGuardSlot) <<<-- Prologue intrinsic call (store to Guard slot) %index.addr = alloca i32, align 4 %offset.addr = alloca i32, align 4 %buffer = alloca [10 x i8], align 1 store i32 %index, i32* %index.addr, align 4 store i32 %offset, i32* %offset.addr, align 4 %arraydecay = getelementptr inbounds [10 x i8], [10 x i8]* %buffer, i32 0, i32 0 %1 = load i32, i32* %index.addr, align 4 call void @llvm.memset.p0i8.i32(i8* %arraydecay, i8 -52, i32 %1, i32 1, i1 false) %2 = load i32, i32* %index.addr, align 4 %arrayidx = getelementptr inbounds [10 x i8], [10 x i8]* %buffer, i32 0, i32 %2 %3 = load i8, i8* %arrayidx, align 1 %conv = sext i8 %3 to i32 %4 = load volatile i8*, i8** %StackGuardSlot <<<-- Loading Guard slot call void @__security_check_cookie(i8* %4) <<<-- Epilogue function-based check ret i32 %conv } ``` * SelectionDAG generated instrumentation: ``` clang-cl /GS test.cc /O1 /c /FA ``` ``` "?example@@YAHHH@Z": # @"\01?example@@YAHHH@Z" # BB#0: # %entry pushl %esi subl $16, %esp movl ___security_cookie, %eax <<<-- Loading Stack Guard value movl 28(%esp), %esi movl %eax, 12(%esp) <<<-- Store to Guard slot leal 2(%esp), %eax pushl %esi pushl $204 pushl %eax calll _memset addl $12, %esp movsbl 2(%esp,%esi), %esi movl 12(%esp), %ecx <<<-- Loading Guard slot calll @__security_check_cookie@4 <<<-- Epilogue function-based check movl %esi, %eax addl $16, %esp popl %esi retl ``` Reviewers: kcc, pcc, eugenis, rnk Subscribers: majnemer, llvm-commits, hans, thakis, rnk Differential Revision: http://reviews.llvm.org/D20346 llvm-svn: 272053
2016-06-08 04:15:35 +08:00
return nullptr;
}
Add support to optionally limit the size of jump tables. Many high-performance processors have a dedicated branch predictor for indirect branches, commonly used with jump tables. As sophisticated as such branch predictors are, they tend to have well defined limits beyond which their effectiveness is hampered or even nullified. One such limit is the number of possible destinations for a given indirect branches that such branch predictors can handle. This patch considers a limit that a target may set to the number of destination addresses in a jump table. Patch by: Evandro Menezes <e.menezes@samsung.com>, Aditya Kumar <aditya.k7@samsung.com>, Sebastian Pop <s.pop@samsung.com>. Differential revision: https://reviews.llvm.org/D21940 llvm-svn: 282412
2016-09-26 23:32:33 +08:00
Add option to specify minimum number of entries for jump tables Add an option to allow easier experimentation by target maintainers with the minimum number of entries to create jump tables. Also clarify the name of the other existing option governing the creation of jump tables. Differential revision: https://reviews.llvm.org/D25883 llvm-svn: 285104
2016-10-26 03:53:51 +08:00
unsigned TargetLoweringBase::getMinimumJumpTableEntries() const {
return MinimumJumpTableEntries;
}
void TargetLoweringBase::setMinimumJumpTableEntries(unsigned Val) {
MinimumJumpTableEntries = Val;
}
Revert r372893 "[CodeGen] Replace -max-jump-table-size with -max-jump-table-targets" This caused severe compile-time regressions, see PR43455. > Modern processors predict the targets of an indirect branch regardless of > the size of any jump table used to glean its target address. Moreover, > branch predictors typically use resources limited by the number of actual > targets that occur at run time. > > This patch changes the semantics of the option `-max-jump-table-size` to limit > the number of different targets instead of the number of entries in a jump > table. Thus, it is now renamed to `-max-jump-table-targets`. > > Before, when `-max-jump-table-size` was specified, it could happen that > cluster jump tables could have targets used repeatedly, but each one was > counted and typically resulted in tables with the same number of entries. > With this patch, when specifying `-max-jump-table-targets`, tables may have > different lengths, since the number of unique targets is counted towards the > limit, but the number of unique targets in tables is the same, but for the > last one containing the balance of targets. > > Differential revision: https://reviews.llvm.org/D60295 llvm-svn: 373060
2019-09-27 17:54:26 +08:00
unsigned TargetLoweringBase::getMinimumJumpTableDensity(bool OptForSize) const {
return OptForSize ? OptsizeJumpTableDensity : JumpTableDensity;
[InlineCost] Improve the cost heuristic for Switch Summary: The motivation example is like below which has 13 cases but only 2 distinct targets ``` lor.lhs.false2: ; preds = %if.then switch i32 %Status, label %if.then27 [ i32 -7012, label %if.end35 i32 -10008, label %if.end35 i32 -10016, label %if.end35 i32 15000, label %if.end35 i32 14013, label %if.end35 i32 10114, label %if.end35 i32 10107, label %if.end35 i32 10105, label %if.end35 i32 10013, label %if.end35 i32 10011, label %if.end35 i32 7008, label %if.end35 i32 7007, label %if.end35 i32 5002, label %if.end35 ] ``` which is compiled into a balanced binary tree like this on AArch64 (similar on X86) ``` .LBB853_9: // %lor.lhs.false2 mov w8, #10012 cmp w19, w8 b.gt .LBB853_14 // BB#10: // %lor.lhs.false2 mov w8, #5001 cmp w19, w8 b.gt .LBB853_18 // BB#11: // %lor.lhs.false2 mov w8, #-10016 cmp w19, w8 b.eq .LBB853_23 // BB#12: // %lor.lhs.false2 mov w8, #-10008 cmp w19, w8 b.eq .LBB853_23 // BB#13: // %lor.lhs.false2 mov w8, #-7012 cmp w19, w8 b.eq .LBB853_23 b .LBB853_3 .LBB853_14: // %lor.lhs.false2 mov w8, #14012 cmp w19, w8 b.gt .LBB853_21 // BB#15: // %lor.lhs.false2 mov w8, #-10105 add w8, w19, w8 cmp w8, #9 // =9 b.hi .LBB853_17 // BB#16: // %lor.lhs.false2 orr w9, wzr, #0x1 lsl w8, w9, w8 mov w9, #517 and w8, w8, w9 cbnz w8, .LBB853_23 .LBB853_17: // %lor.lhs.false2 mov w8, #10013 cmp w19, w8 b.eq .LBB853_23 b .LBB853_3 .LBB853_18: // %lor.lhs.false2 mov w8, #-7007 add w8, w19, w8 cmp w8, #2 // =2 b.lo .LBB853_23 // BB#19: // %lor.lhs.false2 mov w8, #5002 cmp w19, w8 b.eq .LBB853_23 // BB#20: // %lor.lhs.false2 mov w8, #10011 cmp w19, w8 b.eq .LBB853_23 b .LBB853_3 .LBB853_21: // %lor.lhs.false2 mov w8, #14013 cmp w19, w8 b.eq .LBB853_23 // BB#22: // %lor.lhs.false2 mov w8, #15000 cmp w19, w8 b.ne .LBB853_3 ``` However, the inline cost model estimates the cost to be linear with the number of distinct targets and the cost of the above switch is just 2 InstrCosts. The function containing this switch is then inlined about 900 times. This change use the general way of switch lowering for the inline heuristic. It etimate the number of case clusters with the suitability check for a jump table or bit test. Considering the binary search tree built for the clusters, this change modifies the model to be linear with the size of the balanced binary tree. The model is off by default for now : -inline-generic-switch-cost=false This change was originally proposed by Haicheng in D29870. Reviewers: hans, bmakam, chandlerc, eraman, haicheng, mcrosier Reviewed By: hans Subscribers: joerg, aemerson, llvm-commits, rengolin Differential Revision: https://reviews.llvm.org/D31085 llvm-svn: 301649
2017-04-29 00:04:03 +08:00
}
Revert r372893 "[CodeGen] Replace -max-jump-table-size with -max-jump-table-targets" This caused severe compile-time regressions, see PR43455. > Modern processors predict the targets of an indirect branch regardless of > the size of any jump table used to glean its target address. Moreover, > branch predictors typically use resources limited by the number of actual > targets that occur at run time. > > This patch changes the semantics of the option `-max-jump-table-size` to limit > the number of different targets instead of the number of entries in a jump > table. Thus, it is now renamed to `-max-jump-table-targets`. > > Before, when `-max-jump-table-size` was specified, it could happen that > cluster jump tables could have targets used repeatedly, but each one was > counted and typically resulted in tables with the same number of entries. > With this patch, when specifying `-max-jump-table-targets`, tables may have > different lengths, since the number of unique targets is counted towards the > limit, but the number of unique targets in tables is the same, but for the > last one containing the balance of targets. > > Differential revision: https://reviews.llvm.org/D60295 llvm-svn: 373060
2019-09-27 17:54:26 +08:00
unsigned TargetLoweringBase::getMaximumJumpTableSize() const {
return MaximumJumpTableSize;
Add support to optionally limit the size of jump tables. Many high-performance processors have a dedicated branch predictor for indirect branches, commonly used with jump tables. As sophisticated as such branch predictors are, they tend to have well defined limits beyond which their effectiveness is hampered or even nullified. One such limit is the number of possible destinations for a given indirect branches that such branch predictors can handle. This patch considers a limit that a target may set to the number of destination addresses in a jump table. Patch by: Evandro Menezes <e.menezes@samsung.com>, Aditya Kumar <aditya.k7@samsung.com>, Sebastian Pop <s.pop@samsung.com>. Differential revision: https://reviews.llvm.org/D21940 llvm-svn: 282412
2016-09-26 23:32:33 +08:00
}
Revert r372893 "[CodeGen] Replace -max-jump-table-size with -max-jump-table-targets" This caused severe compile-time regressions, see PR43455. > Modern processors predict the targets of an indirect branch regardless of > the size of any jump table used to glean its target address. Moreover, > branch predictors typically use resources limited by the number of actual > targets that occur at run time. > > This patch changes the semantics of the option `-max-jump-table-size` to limit > the number of different targets instead of the number of entries in a jump > table. Thus, it is now renamed to `-max-jump-table-targets`. > > Before, when `-max-jump-table-size` was specified, it could happen that > cluster jump tables could have targets used repeatedly, but each one was > counted and typically resulted in tables with the same number of entries. > With this patch, when specifying `-max-jump-table-targets`, tables may have > different lengths, since the number of unique targets is counted towards the > limit, but the number of unique targets in tables is the same, but for the > last one containing the balance of targets. > > Differential revision: https://reviews.llvm.org/D60295 llvm-svn: 373060
2019-09-27 17:54:26 +08:00
void TargetLoweringBase::setMaximumJumpTableSize(unsigned Val) {
MaximumJumpTableSize = Val;
Add support to optionally limit the size of jump tables. Many high-performance processors have a dedicated branch predictor for indirect branches, commonly used with jump tables. As sophisticated as such branch predictors are, they tend to have well defined limits beyond which their effectiveness is hampered or even nullified. One such limit is the number of possible destinations for a given indirect branches that such branch predictors can handle. This patch considers a limit that a target may set to the number of destination addresses in a jump table. Patch by: Evandro Menezes <e.menezes@samsung.com>, Aditya Kumar <aditya.k7@samsung.com>, Sebastian Pop <s.pop@samsung.com>. Differential revision: https://reviews.llvm.org/D21940 llvm-svn: 282412
2016-09-26 23:32:33 +08:00
}
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
[TargetLowering] Move TargetLoweringBase::isJumpTableRelative() implementation into TargetLoweringBase.cpp. NFC. This will help with reducing header dependencies in TargetLowering.h in a future patch.
2020-05-22 20:18:06 +08:00
bool TargetLoweringBase::isJumpTableRelative() const {
return getTargetMachine().isPositionIndependent();
}
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
//===----------------------------------------------------------------------===//
// Reciprocal Estimates
//===----------------------------------------------------------------------===//
/// Get the reciprocal estimate attribute string for a function that will
/// override the target defaults.
static StringRef getRecipEstimateForFunc(MachineFunction &MF) {
MachineFunction: Return reference from getFunction(); NFC The Function can never be nullptr so we can return a reference. llvm-svn: 320884
2017-12-16 06:22:58 +08:00
const Function &F = MF.getFunction();
return F.getFnAttribute("reciprocal-estimates").getValueAsString();
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
}
/// Construct a string for the given reciprocal operation of the given type.
/// This string should match the corresponding option to the front-end's
/// "-mrecip" flag assuming those strings have been passed through in an
/// attribute string. For example, "vec-divf" for a division of a vXf32.
static std::string getReciprocalOpName(bool IsSqrt, EVT VT) {
std::string Name = VT.isVector() ? "vec-" : "";
Name += IsSqrt ? "sqrt" : "div";
// TODO: Handle "half" or other float types?
if (VT.getScalarType() == MVT::f64) {
Name += "d";
} else {
assert(VT.getScalarType() == MVT::f32 &&
"Unexpected FP type for reciprocal estimate");
Name += "f";
}
return Name;
}
/// Return the character position and value (a single numeric character) of a
/// customized refinement operation in the input string if it exists. Return
/// false if there is no customized refinement step count.
static bool parseRefinementStep(StringRef In, size_t &Position,
uint8_t &Value) {
const char RefStepToken = ':';
Position = In.find(RefStepToken);
if (Position == StringRef::npos)
return false;
StringRef RefStepString = In.substr(Position + 1);
// Allow exactly one numeric character for the additional refinement
// step parameter.
if (RefStepString.size() == 1) {
char RefStepChar = RefStepString[0];
if (RefStepChar >= '0' && RefStepChar <= '9') {
Value = RefStepChar - '0';
return true;
}
}
report_fatal_error("Invalid refinement step for -recip.");
}
/// For the input attribute string, return one of the ReciprocalEstimate enum
/// status values (enabled, disabled, or not specified) for this operation on
/// the specified data type.
static int getOpEnabled(bool IsSqrt, EVT VT, StringRef Override) {
if (Override.empty())
return TargetLoweringBase::ReciprocalEstimate::Unspecified;
SmallVector<StringRef, 4> OverrideVector;
[TargetLowering] Use StringRef::split instead of SplitString. NFC SplitString splits based on a list of delimeters, but we're only using one delimeter so we should use the simpler split. llvm-svn: 331613
2018-05-07 09:32:18 +08:00
Override.split(OverrideVector, ',');
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
unsigned NumArgs = OverrideVector.size();
// Check if "all", "none", or "default" was specified.
if (NumArgs == 1) {
// Look for an optional setting of the number of refinement steps needed
// for this type of reciprocal operation.
size_t RefPos;
uint8_t RefSteps;
if (parseRefinementStep(Override, RefPos, RefSteps)) {
// Split the string for further processing.
Override = Override.substr(0, RefPos);
}
// All reciprocal types are enabled.
if (Override == "all")
return TargetLoweringBase::ReciprocalEstimate::Enabled;
// All reciprocal types are disabled.
if (Override == "none")
return TargetLoweringBase::ReciprocalEstimate::Disabled;
// Target defaults for enablement are used.
if (Override == "default")
return TargetLoweringBase::ReciprocalEstimate::Unspecified;
}
// The attribute string may omit the size suffix ('f'/'d').
std::string VTName = getReciprocalOpName(IsSqrt, VT);
std::string VTNameNoSize = VTName;
fix variable names; NFCI Because we're just 'or-ing' these 2 variables later in the code, I don't think there's a logical bug here, but of course the string with "no size" is the one that should have the size suffix stripped off. llvm-svn: 284826
2016-10-21 22:58:30 +08:00
VTNameNoSize.pop_back();
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
static const char DisabledPrefix = '!';
for (StringRef RecipType : OverrideVector) {
size_t RefPos;
uint8_t RefSteps;
if (parseRefinementStep(RecipType, RefPos, RefSteps))
RecipType = RecipType.substr(0, RefPos);
// Ignore the disablement token for string matching.
bool IsDisabled = RecipType[0] == DisabledPrefix;
if (IsDisabled)
RecipType = RecipType.substr(1);
if (RecipType.equals(VTName) || RecipType.equals(VTNameNoSize))
return IsDisabled ? TargetLoweringBase::ReciprocalEstimate::Disabled
: TargetLoweringBase::ReciprocalEstimate::Enabled;
}
return TargetLoweringBase::ReciprocalEstimate::Unspecified;
}
/// For the input attribute string, return the customized refinement step count
/// for this operation on the specified data type. If the step count does not
/// exist, return the ReciprocalEstimate enum value for unspecified.
static int getOpRefinementSteps(bool IsSqrt, EVT VT, StringRef Override) {
if (Override.empty())
return TargetLoweringBase::ReciprocalEstimate::Unspecified;
SmallVector<StringRef, 4> OverrideVector;
[TargetLowering] Use StringRef::split instead of SplitString. NFC SplitString splits based on a list of delimeters, but we're only using one delimeter so we should use the simpler split. llvm-svn: 331613
2018-05-07 09:32:18 +08:00
Override.split(OverrideVector, ',');
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
unsigned NumArgs = OverrideVector.size();
// Check if "all", "default", or "none" was specified.
if (NumArgs == 1) {
// Look for an optional setting of the number of refinement steps needed
// for this type of reciprocal operation.
size_t RefPos;
uint8_t RefSteps;
if (!parseRefinementStep(Override, RefPos, RefSteps))
return TargetLoweringBase::ReciprocalEstimate::Unspecified;
// Split the string for further processing.
Override = Override.substr(0, RefPos);
assert(Override != "none" &&
"Disabled reciprocals, but specifed refinement steps?");
// If this is a general override, return the specified number of steps.
if (Override == "all" || Override == "default")
return RefSteps;
}
// The attribute string may omit the size suffix ('f'/'d').
std::string VTName = getReciprocalOpName(IsSqrt, VT);
std::string VTNameNoSize = VTName;
fix variable names; NFCI Because we're just 'or-ing' these 2 variables later in the code, I don't think there's a logical bug here, but of course the string with "no size" is the one that should have the size suffix stripped off. llvm-svn: 284826
2016-10-21 22:58:30 +08:00
VTNameNoSize.pop_back();
[Target] remove TargetRecip class; 2nd try This is a retry of r284495 which was reverted at r284513 due to use-after-scope bugs caused by faulty usage of StringRef. This version also renames a pair of functions: getRecipEstimateDivEnabled() getRecipEstimateSqrtEnabled() as suggested by Eric Christopher. original commit msg: [Target] remove TargetRecip class; move reciprocal estimate isel functionality to TargetLowering This is a follow-up to https://reviews.llvm.org/D24816 - where we changed reciprocal estimates to be function attributes rather than TargetOptions. This patch is intended to be a structural, but not functional change. By moving all of the TargetRecip functionality into TargetLowering, we can remove all of the reciprocal estimate state, shield the callers from the string format implementation, and simplify/localize the logic needed for a target to enable this. If a function has a "reciprocal-estimates" attribute, those settings may override the target's default reciprocal preferences for whatever operation and data type we're trying to optimize. If there's no attribute string or specific setting for the op/type pair, just use the target default settings. As noted earlier, a better solution would be to move the reciprocal estimate settings to IR instructions and SDNodes rather than function attributes, but that's a multi-step job that requires infrastructure improvements. I intend to work on that, but it's not clear how long it will take to get all the pieces in place. Differential Revision: https://reviews.llvm.org/D25440 llvm-svn: 284746
2016-10-21 00:55:45 +08:00
for (StringRef RecipType : OverrideVector) {
size_t RefPos;
uint8_t RefSteps;
if (!parseRefinementStep(RecipType, RefPos, RefSteps))
continue;
RecipType = RecipType.substr(0, RefPos);
if (RecipType.equals(VTName) || RecipType.equals(VTNameNoSize))
return RefSteps;
}
return TargetLoweringBase::ReciprocalEstimate::Unspecified;
}
int TargetLoweringBase::getRecipEstimateSqrtEnabled(EVT VT,
MachineFunction &MF) const {
return getOpEnabled(true, VT, getRecipEstimateForFunc(MF));
}
int TargetLoweringBase::getRecipEstimateDivEnabled(EVT VT,
MachineFunction &MF) const {
return getOpEnabled(false, VT, getRecipEstimateForFunc(MF));
}
int TargetLoweringBase::getSqrtRefinementSteps(EVT VT,
MachineFunction &MF) const {
return getOpRefinementSteps(true, VT, getRecipEstimateForFunc(MF));
}
int TargetLoweringBase::getDivRefinementSteps(EVT VT,
MachineFunction &MF) const {
return getOpRefinementSteps(false, VT, getRecipEstimateForFunc(MF));
}
TargetLowering: Add finalizeLowering() function; NFC Adds a new method finalizeLowering to TargetLoweringBase. This is in preparation for an upcoming commit. This function is meant for target specific adjustments to MachineFrameInfo or register reservations. Move the freezeRegisters() and the hasCopyImplyingStackAdjustment() handling into the new function to prove the concept. As an added bonus GlobalISel no longer missed the hasCopyImplyingStackAdjustment() handling with this. Differential Revision: https://reviews.llvm.org/D32621 llvm-svn: 301679
2017-04-29 04:25:05 +08:00
void TargetLoweringBase::finalizeLowering(MachineFunction &MF) const {
MF.getRegInfo().freezeReservedRegs(MF);
}
GlobalISel: Preserve load/store metadata in IRTranslator This was dropping the invariant metadata on dead argument loads, so they weren't deleted. Atomics still need to be fixed the same way. Also, apparently store was never preserving dereferencable which should also be fixed.
2020-01-13 03:10:42 +08:00
MachineMemOperand::Flags
TargetLoweringBase::getLoadMemOperandFlags(const LoadInst &LI,
const DataLayout &DL) const {
MachineMemOperand::Flags Flags = MachineMemOperand::MOLoad;
if (LI.isVolatile())
Flags |= MachineMemOperand::MOVolatile;
if (LI.hasMetadata(LLVMContext::MD_nontemporal))
Flags |= MachineMemOperand::MONonTemporal;
if (LI.hasMetadata(LLVMContext::MD_invariant_load))
Flags |= MachineMemOperand::MOInvariant;
if (isDereferenceablePointer(LI.getPointerOperand(), LI.getType(), DL))
Flags |= MachineMemOperand::MODereferenceable;
Flags |= getTargetMMOFlags(LI);
return Flags;
}
MachineMemOperand::Flags
TargetLoweringBase::getStoreMemOperandFlags(const StoreInst &SI,
const DataLayout &DL) const {
MachineMemOperand::Flags Flags = MachineMemOperand::MOStore;
if (SI.isVolatile())
Flags |= MachineMemOperand::MOVolatile;
if (SI.hasMetadata(LLVMContext::MD_nontemporal))
Flags |= MachineMemOperand::MONonTemporal;
// FIXME: Not preserving dereferenceable
Flags |= getTargetMMOFlags(SI);
return Flags;
}
GlobalISel: Apply target MMO flags to atomics Unify MMO flag handling with SelectionDAG like with loads and stores.
2020-01-13 04:54:09 +08:00
MachineMemOperand::Flags
TargetLoweringBase::getAtomicMemOperandFlags(const Instruction &AI,
const DataLayout &DL) const {
auto Flags = MachineMemOperand::MOLoad | MachineMemOperand::MOStore;
if (const AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(&AI)) {
if (RMW->isVolatile())
Flags |= MachineMemOperand::MOVolatile;
} else if (const AtomicCmpXchgInst *CmpX = dyn_cast<AtomicCmpXchgInst>(&AI)) {
if (CmpX->isVolatile())
Flags |= MachineMemOperand::MOVolatile;
} else
llvm_unreachable("not an atomic instruction");
// FIXME: Not preserving dereferenceable
Flags |= getTargetMMOFlags(AI);
return Flags;
}
GlobalISel: Move Localizer::shouldLocalize(..) to TargetLowering Add a new target hook for shouldLocalize so that targets can customize the logic. https://reviews.llvm.org/D75207
2020-03-03 01:15:40 +08:00
//===----------------------------------------------------------------------===//
// GlobalISel Hooks
//===----------------------------------------------------------------------===//
bool TargetLoweringBase::shouldLocalize(const MachineInstr &MI,
const TargetTransformInfo *TTI) const {
auto &MF = *MI.getMF();
auto &MRI = MF.getRegInfo();
// Assuming a spill and reload of a value has a cost of 1 instruction each,
// this helper function computes the maximum number of uses we should consider
// for remat. E.g. on arm64 global addresses take 2 insts to materialize. We
// break even in terms of code size when the original MI has 2 users vs
// choosing to potentially spill. Any more than 2 users we we have a net code
// size increase. This doesn't take into account register pressure though.
auto maxUses = [](unsigned RematCost) {
// A cost of 1 means remats are basically free.
if (RematCost == 1)
return UINT_MAX;
if (RematCost == 2)
return 2U;
// Remat is too expensive, only sink if there's one user.
if (RematCost > 2)
return 1U;
llvm_unreachable("Unexpected remat cost");
};
// Helper to walk through uses and terminate if we've reached a limit. Saves
// us spending time traversing uses if all we want to know is if it's >= min.
auto isUsesAtMost = [&](unsigned Reg, unsigned MaxUses) {
unsigned NumUses = 0;
auto UI = MRI.use_instr_nodbg_begin(Reg), UE = MRI.use_instr_nodbg_end();
for (; UI != UE && NumUses < MaxUses; ++UI) {
NumUses++;
}
// If we haven't reached the end yet then there are more than MaxUses users.
return UI == UE;
};
switch (MI.getOpcode()) {
default:
return false;
// Constants-like instructions should be close to their users.
// We don't want long live-ranges for them.
case TargetOpcode::G_CONSTANT:
case TargetOpcode::G_FCONSTANT:
case TargetOpcode::G_FRAME_INDEX:
case TargetOpcode::G_INTTOPTR:
return true;
case TargetOpcode::G_GLOBAL_VALUE: {
unsigned RematCost = TTI->getGISelRematGlobalCost();
Register Reg = MI.getOperand(0).getReg();
unsigned MaxUses = maxUses(RematCost);
if (MaxUses == UINT_MAX)
return true; // Remats are "free" so always localize.
bool B = isUsesAtMost(Reg, MaxUses);
return B;
}
}
}