Commit Graph

78 Commits

Author SHA1 Message Date
Nadav Rotem
c05d30601c Add a new optimization pass: Stack Coloring, that merges disjoint static allocations (allocas). Allocas are known to be
disjoint if they are marked by disjoint lifetime markers (@llvm.lifetime.XXX intrinsics).



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163299 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-06 09:17:37 +00:00
Roman Divacky
5932429765 Stop casting away const qualifier needlessly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@163258 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-05 22:26:57 +00:00
Jakob Stoklund Olesen
228e3f5fac Add a verification pass after ExpandISelPseudos.
This pass often has weird CFG hacks and hand-written MI building code
that can go wrong in many ways.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@162224 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-20 20:52:08 +00:00
Jakob Stoklund Olesen
dcc4436cdd Add an experimental -early-live-intervals option.
This option runs LiveIntervals before TwoAddressInstructionPass which
will eventually learn to exploit and update the analysis.

Eventually, LiveIntervals will run before PHIElimination, and we can get
rid of LiveVariables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@161270 91177308-0d34-0410-b5e6-96231b3b80d8
2012-08-03 22:12:54 +00:00
Jakob Stoklund Olesen
f86c00f1f8 Make sure -print-machineinstrs applies to the first pass as well.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159720 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-04 19:28:27 +00:00
Jakob Stoklund Olesen
33242fd3ed Add an experimental early if-conversion pass, off by default.
This pass performs if-conversion on SSA form machine code by
speculatively executing both sides of the branch and using a cmov
instruction to select the result. This can help lower the number of
branch mispredictions on architectures like x86 that don't have
predicable instructions.

The current implementation is very aggressive, and causes regressions on
mosts tests. It needs good heuristics that have yet to be implemented.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159694 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-04 00:09:54 +00:00
Chandler Carruth
6068c48498 All glory to address sanitizer. ;]
It appears to have caught a use-after-free introduced as by r159567
and/or friends which call 'addPass' from many more places. The bug in
'addPass' doesn't appear to be new, and was spotted by inspection when
ASan shown a bright light of a stacktrace at these functions.

Hopefully this will fix the ASan failure -- I have no test case other
than running an ASan-built clang over the test suite.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159614 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-02 22:56:41 +00:00
Bob Wilson
30a507a1f5 Extend TargetPassConfig to allow running only a subset of the normal passes.
This is still a work in progress but I believe it is currently good enough
to fix PR13122 "Need unit test driver for codegen IR passes".  For example,
you can run llc with -stop-after=loop-reduce to have it dump out the IR after
running LSR.  Serializing machine-level IR is not yet supported but we have
some patches in progress for that.

The plan is to serialize the IR to a YAML file, containing separate sections
for the LLVM IR, machine-level IR, and whatever other info is needed.  Chad
suggested that we stash the stop-after pass in the YAML file and use that
instead of the start-after option to figure out where to restart the
compilation.  I think that's a great idea, but since it's not implemented yet
I put the -start-after option into this patch for testing purposes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159570 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-02 19:48:45 +00:00
Bob Wilson
6b2bb15bf7 Move assertion with TargetPassConfig's Initialized flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159569 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-02 19:48:39 +00:00
Bob Wilson
3fb99a7368 Consistently use AnalysisID types in TargetPassConfig.
This makes it possible to just use a zero value to represent "no pass", so
the phony NoPassID global variable is no longer needed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159568 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-02 19:48:37 +00:00
Bob Wilson
564fbf6aff Add all codegen passes to the PassManager via TargetPassConfig.
This is a preliminary step toward having TargetPassConfig be able to
start and stop the compilation at specified passes for unit testing
and debugging.  No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159567 91177308-0d34-0410-b5e6-96231b3b80d8
2012-07-02 19:48:31 +00:00
Jakob Stoklund Olesen
34f5a2b596 Allow targets to inject passes before the virtual register rewriter.
Such passes can be used to tweak the register assignments in a
target-dependent way, for example to avoid write-after-write
dependencies.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159209 91177308-0d34-0410-b5e6-96231b3b80d8
2012-06-26 17:09:29 +00:00
Jakob Stoklund Olesen
5984d2b31f Run ProcessImplicitDefs on SSA form where it can be much simpler.
Implicitly defined virtual registers can simply have the <undef> bit set
on all uses, and copies can be turned into implicit defs recursively.

Physical registers are a bit trickier. We handle the common case where a
physreg def is used by a nearby instruction in the same basic block. For
more complicated cases, just leave the IMPLICIT_DEF instruction in.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@159149 91177308-0d34-0410-b5e6-96231b3b80d8
2012-06-25 18:12:18 +00:00
Jakob Stoklund Olesen
05ec712e7f Reintroduce VirtRegRewriter.
OK, not really. We don't want to reintroduce the old rewriter hacks.

This patch extracts virtual register rewriting as a separate pass that
runs after the register allocator. This is possible now that
CodeGen/Passes.cpp can configure the full optimizing register allocator
pipeline.

The rewriter pass uses register assignments in VirtRegMap to rewrite
virtual registers to physical registers, and it inserts kill flags based
on live intervals.

These finalization steps are the same for the optimizing register
allocators: RABasic, RAGreedy, and PBQP.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158244 91177308-0d34-0410-b5e6-96231b3b80d8
2012-06-08 23:44:45 +00:00
Bob Wilson
6e1b812850 Add an insertPass API to TargetPassConfig. <rdar://problem/11498613>
Besides adding the new insertPass function, this patch uses it to
enhance the existing -print-machineinstrs so that the MachineInstrs
after a specific pass can be printed.

Patch by Bin Zeng!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157655 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-30 00:17:12 +00:00
Bill Wendling
7c4ce30ea6 Change the PassManager from a reference to a pointer.
The TargetPassManager's default constructor wants to initialize the PassManager
to 'null'. But it's illegal to bind a null reference to a null l-value. Make the
ivar a pointer instead.
PR12468


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155902 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01 08:27:43 +00:00
Chandler Carruth
9e67db4af1 Flip the new block-placement pass to be on by default.
This is mostly to test the waters. I'd like to get results from FNT
build bots and other bots running on non-x86 platforms.

This feature has been pretty heavily tested over the last few months by
me, and it fixes several of the execution time regressions caused by the
inlining work by preventing inlining decisions from radically impacting
block layout.

I've seen very large improvements in yacr2 and ackermann benchmarks,
along with the expected noise across all of the benchmark suite whenever
code layout changes. I've analyzed all of the regressions and fixed
them, or found them to be impossible to fix. See my email to llvmdev for
more details.

I'd like for this to be in 3.1 as it complements the inliner changes,
but if any failures are showing up or anyone has concerns, it is just
a flag flip and so can be easily turned off.

I'm switching it on tonight to try and get at least one run through
various folks' performance suites in case SPEC or something else has
serious issues with it. I'll watch bots and revert if anything shows up.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@154816 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-16 13:49:17 +00:00
Jakob Stoklund Olesen
8b4c502098 Enable machine code verification in the entire code generator.
Some targets still mess up the liveness information, but that isn't
verified after MRI->invalidateLiveness().

The verifier can still check other useful things like register classes
and CFG, so it should be enabled after all passes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153615 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-28 23:54:28 +00:00
Jakob Stoklund Olesen
78811669d5 Enable machine code verification after PreSched2 passes.
The late scheduler depends on accurate liveness information if it is
breaking anti-dependencies, so we should be able to verify it.

Relax the terminator checking in the machine code verifier so it can
handle the basic blocks created by if conversion.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153614 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-28 23:31:15 +00:00
Jakob Stoklund Olesen
2ef5bf6b34 Also verify after ExpandPostRAPseudos.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153599 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-28 20:49:30 +00:00
Jakob Stoklund Olesen
663ee20cc4 Enable machine code verification after the late machine optimization passes.
Branch folding invalidates liveness and disables liveness verification
on some targets.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@153597 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-28 20:47:37 +00:00
Andrew Trick
17d35e57a5 misched: implemented a framework for top-down or bottom-up scheduling.
New flags: -misched-topdown, -misched-bottomup. They can be used with
the default scheduler or with -misched=shuffle. Without either
topdown/bottomup flag -misched=shuffle now alternates scheduling
direction.

LiveIntervals update is unimplemented with bottom-up scheduling, so
only -misched-topdown currently works.

Capped the ScheduleDAG hierarchy with a concrete ScheduleDAGMI class.
ScheduleDAGMI is aware of the top and bottom of the unscheduled zone
within the current region. Scheduling policy can be plugged into
the ScheduleDAGMI driver by implementing MachineSchedStrategy.
ConvergingScheduler is now the default scheduling algorithm.
It exercises the new driver but still does no reordering.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@152700 91177308-0d34-0410-b5e6-96231b3b80d8
2012-03-14 04:00:41 +00:00
Evan Cheng
01b623c8c2 Fix machine-cp by having it to check sub-register indicies. e.g.
ecx = mov eax
al  = mov ch
The second copy is not a nop because the sub-indices of ecx,ch is not the
same of that of eax/al.

Re-enabled machine-cp.
PR11940


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@151002 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-20 23:28:17 +00:00
Benjamin Kramer
1a7ea9dd04 Disable machine copy propagation for now. It's known to be buggy (PR11940) and introduces subtle miscompiles in many places.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150703 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-16 17:29:50 +00:00
Andrew Trick
900d7b78e6 Revert r150565 again. Appears to be a stage2 failure with dragonegg.
I'll put MachineLICM back before PEI. All my arm/x86 benchmarks look good, but buildbots don't like it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150568 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-15 07:57:03 +00:00
Andrew Trick
788066a1f4 Reapply r150565 with the typo fix properly merged.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150567 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-15 05:43:27 +00:00
Andrew Trick
40bb92a5cc reverting r150565. Premature push.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150566 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-15 05:22:12 +00:00
Andrew Trick
352aa4f67f Move PostRAMachineLICM into MachineLateOptimization. It now runs after PEI!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150565 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-15 05:13:47 +00:00
Andrew Trick
79bf288584 Allow CodeGen (llc) command line options to work as expected.
The llc command line options for enabling/disabling passes are local to CodeGen/Passes.cpp. This patch associates those options with standard pass IDs so they work regardless of how the target configures the passes.

A target has two ways of overriding standard passes:
1) Redefine the pass pipeline (override TargetPassConfig::add%Stage)
2) Replace or suppress individiual passes with TargetPassConfig::substitutePass.

In both cases, the command line options associated with the pass override the target default.

For example, say a target wants to disable machine instruction scheduling by default:

- The target calls disablePass(MachineSchedulerID) but otherwise does not override any TargetPassConfig methods.

- Without any llc options, no scheduler is run.

- With -enable-misched, the standard machine scheduler is run and honors the -misched=... flag to select the scheduler variant, which may be used for performance evaluation or testing.

Sorry overridePass is ugly. I haven't thought of a better way without replacing the cl::opt framework. I hope to do that one day...

I haven't figured out why CodeGen uses char& for pass IDs. AnalysisID is much easier to use and less bug prone. I'm using it wherever I can for internal implementation. Maybe later we can change the global pass ID definitions as well.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150563 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-15 03:21:51 +00:00
Andrew Trick
5e108eeeef Added TargetPassConfig::disablePass/substitutePass as a general mechanism to override specific passes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150562 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-15 03:21:47 +00:00
Andrew Trick
746f24b732 Add TargetPassConfig hooks for scheduling/bundling.
In case the MachineScheduling pass I'm working on doesn't work well
for another target, they can completely override it. This also adds a
hook immediately after the RegAlloc pass to cleanup immediately after
vregs go away. We may want to fold it into the postRA hook later.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150298 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-11 07:11:32 +00:00
Andrew Trick
06efdd238e comment grammar
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150233 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-10 07:08:25 +00:00
Andrew Trick
8dd26253f5 RegAlloc superpass: includes phi elimination, coalescing, and scheduling.
Creates a configurable regalloc pipeline.

Ensure specific llc options do what they say and nothing more: -reglloc=... has no effect other than selecting the allocator pass itself. This patch introduces a new umbrella flag, "-optimize-regalloc", to enable/disable the optimizing regalloc "superpass". This allows for example testing coalscing and scheduling under -O0 or vice-versa.

When a CodeGen pass requires the MachineFunction to have a particular property, we need to explicitly define that property so it can be directly queried rather than naming a specific Pass. For example, to check for SSA, use MRI->isSSA, not addRequired<PHIElimination>.

CodeGen transformation passes are never "required" as an analysis

ProcessImplicitDefs does not require LiveVariables.

We have a plan to massively simplify some of the early passes within the regalloc superpass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150226 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-10 04:10:36 +00:00
Andrew Trick
f7b96311d2 Improve TargetPassConfig. No intended functionality.
Split CodeGen into stages.
Distinguish between optimization and correctness.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150122 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-09 00:40:55 +00:00
Andrew Trick
1dd8c8560d Codegen pass definition cleanup. No functionality.
Moving toward a uniform style of pass definition to allow easier target configuration.
Globally declare Pass ID.
Globally declare pass initializer.
Use INITIALIZE_PASS consistently.
Add a call to the initializer from CodeGen.cpp.
Remove redundant "createPass" functions and "getPassName" methods.

While cleaning up declarations, cleaned up comments (sorry for large diff).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150100 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:23:13 +00:00
Andrew Trick
9d41bd5c78 Move pass configuration out of pass constructors: MachineLICM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150099 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:23:03 +00:00
Andrew Trick
32986572bf Move pass configuration out of pass constructors: StackSlotColoring.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150097 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:22:57 +00:00
Andrew Trick
c7d081b594 Move pass configuration out of pass constructors: PostRAScheduler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150096 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:22:53 +00:00
Andrew Trick
61f1e3db43 Move pass configuration out of pass constructors: BranchFolderPass
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150095 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:22:48 +00:00
Andrew Trick
ffea03f216 Added TargetPassConfig::setOpt
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150093 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:22:39 +00:00
Andrew Trick
ebe18ef5c2 Added Pass::createPass(ID) to handle pass configuration by ID
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150092 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:22:34 +00:00
Andrew Trick
d2a7bedbc9 Move pass configuration out of pass constructors: TailDuplicate::PreRegAlloc
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@150091 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-08 21:22:30 +00:00
Andrew Trick
061efcfb3e TargetPassConfig: confine the MC configuration to TargetMachine.
Passes prior to instructon selection are now split into separate configurable stages.
Header dependencies are simplified.
The bulk of this diff is simply removal of the silly DisableVerify flags.

Sorry for the target header churn. Attempting to stabilize them.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@149754 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-04 02:56:59 +00:00
Andrew Trick
d542265401 Move TargetPassConfig implementation into Passes.cpp
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@149753 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-04 02:56:48 +00:00
Andrew Trick
7461334084 Make TargetPassConfig an ImmutablePass so CodeGenPasses can query options
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@149752 91177308-0d34-0410-b5e6-96231b3b80d8
2012-02-04 02:56:45 +00:00
Jakob Stoklund Olesen
799c1ede71 Delete the linear scan register allocator.
RegAllocGreedy has been the default for six months now.

Deleting RegAllocLinearScan makes it possible to also delete
VirtRegRewriter and clean up the spiller code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@144475 91177308-0d34-0410-b5e6-96231b3b80d8
2011-11-12 22:39:45 +00:00
Jakob Stoklund Olesen
141f529baf Update comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130582 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 03:13:08 +00:00
Jakob Stoklund Olesen
5aa3211e2d Use a greedy algorithm for allocating registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@130568 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-30 01:37:54 +00:00
Jakob Stoklund Olesen
430721cff8 Force the greedy register allocator to be linked alongside linear scan.
This means that the new register allocator can be used with 'clang -mllvm -regalloc=greedy'.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129764 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-19 17:17:58 +00:00
Jakob Stoklund Olesen
8b89c64b5c Use the fast register allocator by default for -O0 builds.
This affects both llvm-gcc and clang.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@105372 91177308-0d34-0410-b5e6-96231b3b80d8
2010-06-03 00:39:06 +00:00