| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Now that MachineBasicBlock::reverse_instr_iterator knows when it's at
the end (since r281168 and r281170), implement
MachineBasicBlock::reverse_iterator directly on top of an
ilist::reverse_iterator by adding an IsReverse template parameter to
MachineInstrBundleIterator. This replaces another hard-to-reason-about
use of std::reverse_iterator on list iterators, matching the changes for
ilist::reverse_iterator from r280032 (see the "out of scope" section at
the end of that commit message). MachineBasicBlock::reverse_iterator
now has a handle to the current node and has obvious invalidation
semantics.
r280032 has a more detailed explanation of how list-style reverse
iterators (invalidated when the pointed-at node is deleted) are
different from vector-style reverse iterators like std::reverse_iterator
(invalidated on every operation). A great motivating example is this
commit's changes to lib/CodeGen/DeadMachineInstructionElim.cpp.
Note: If your out-of-tree backend deletes instructions while iterating
on a MachineBasicBlock::reverse_iterator or converts between
MachineBasicBlock::iterator and MachineBasicBlock::reverse_iterator,
you'll need to update your code in similar ways to r280032. The
following table might help:
[Old] ==> [New]
delete &*RI, RE = end() delete &*RI++
RI->erase(), RE = end() RI++->erase()
reverse_iterator(I) std::prev(I).getReverse()
reverse_iterator(I) ++I.getReverse()
--reverse_iterator(I) I.getReverse()
reverse_iterator(std::next(I)) I.getReverse()
RI.base() std::prev(RI).getReverse()
RI.base() ++RI.getReverse()
--RI.base() RI.getReverse()
std::next(RI).base() RI.getReverse()
(For more details, have a look at r280032.)
llvm-svn: 281172
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a prep commit before fixing MachineBasicBlock::reverse_iterator
invalidation semantics, ala r281167 for ilist::reverse_iterator. This
changes MachineBasicBlock::Instructions to track which node is the
sentinel regardless of LLVM_ENABLE_ABI_BREAKING_CHECKS.
There's almost no functionality change (aside from ABI). However, in
the rare configuration:
#if !defined(NDEBUG) && !defined(LLVM_ENABLE_ABI_BREAKING_CHECKS)
the isKnownSentinel() assertions in ilist_iterator<>::operator* suddenly
have teeth for MachineInstr. If these assertions start firing for your
out-of-tree backend, have a look at the suggestions in the commit
message for r279314, and at some of the commits leading up to it that
avoid dereferencing the end() iterator.
llvm-svn: 281168
|
|
|
|
|
|
|
|
| |
class. Fix memory operand size, remove unnecessary pattern.
Differential Revision: http://reviews.llvm.org/D24443
llvm-svn: 281164
|
|
|
|
|
|
|
|
| |
This should *actually* fix PR30244. This cranks up the workaround for PR30188 so that we never sink loads or stores of allocas.
The idea is that these should be removed by SROA/Mem2Reg, and any movement of them may well confuse SROA or just cause unwanted code churn. It's not ideal that the midend should be crippled like this, but that unwanted churn can really cause significant regressions in important workloads (tsan).
llvm-svn: 281162
|
|
|
|
|
|
|
|
|
|
|
|
| |
during sinking
Exposed by PR30244, we will split a block currently if we think we can sink at least one instruction. However this isn't right - the reason we split predecessors is so that we can sink instructions that otherwise couldn't be sunk because it isn't safe to do so - stores, for example.
So, change the heuristic to only split if it thinks it can sink at least one non-speculatable instruction.
Should fix PR30244.
llvm-svn: 281160
|
|
|
|
|
|
| |
commutable before calling findCommutedOpIndices for every operand. Also make sure the operand is a register before each call to save some work on commutable instructions that might have an operand.
llvm-svn: 281158
|
|
|
|
| |
llvm-svn: 281156
|
|
|
|
| |
llvm-svn: 281154
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
This will let e.g. the load/store vectorizer propagate this metadata
appropriately.
Reviewers: arsenm
Subscribers: tra, jholewinski, hfinkel, mzolotukhin
Differential Revision: https://reviews.llvm.org/D23479
llvm-svn: 281153
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
With this change (plus some changes to prevent !invariant from being
clobbered within llvm), clang will be able to model the __ldg CUDA
builtin as an invariant load, rather than as a target-specific llvm
intrinsic. This will let the optimizer play with these loads --
specifically, we should be able to vectorize them in the load-store
vectorizer.
Reviewers: tra
Subscribers: jholewinski, hfinkel, llvm-commits, chandlerc
Differential Revision: https://reviews.llvm.org/D23477
llvm-svn: 281152
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
An IR load can be invariant, dereferenceable, neither, or both. But
currently, MI's notion of invariance is IR-invariant &&
IR-dereferenceable.
This patch splits up the notions of invariance and dereferenceability at
the MI level. It's NFC, so adds some probably-unnecessary
"is-dereferenceable" checks, which we can remove later if desired.
Reviewers: chandlerc, tstellarAMD
Subscribers: jholewinski, arsenm, nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D23371
llvm-svn: 281151
|
|
|
|
|
|
|
|
| |
argument.
rdar://28233388
llvm-svn: 281147
|
|
|
|
|
|
|
|
|
| |
This generates invalid IR: the only users of swifterror can be call
arguments, loads, and stores.
rdar://28242257
llvm-svn: 281144
|
|
|
|
| |
llvm-svn: 281143
|
|
|
|
|
|
|
|
|
| |
foldICmpIntrinsicWithConstant ; NFC
1. Rename variables to be consistent with related/preceding code (may want to reorganize).
2. Fix comments/formatting.
llvm-svn: 281140
|
|
|
|
|
|
|
|
|
|
| |
Everything under foldICmpInstWithConstant() should now be working for
splat vectors via m_APInt matchers. Ie, I've removed all of the FIXMEs
that I added while cleaning that section up. Note that not all of the
associated FIXMEs in the regression tests are gone though, because some
of the tests require earlier folds that are still scalar-only.
llvm-svn: 281139
|
|
|
|
|
|
| |
rdar://28190687
llvm-svn: 281138
|
|
|
|
|
|
| |
Differential revision: https://reviews.llvm.org/D24295
llvm-svn: 281137
|
|
|
|
| |
llvm-svn: 281131
|
|
|
|
| |
llvm-svn: 281130
|
|
|
|
| |
llvm-svn: 281128
|
|
|
|
|
|
|
| |
These defaulted to Write32Bit. I don't think this actually matters
since these don't exist during scheduling.
llvm-svn: 281127
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
Could be useful for comparison when we suspect that alloca was skipped
because of this.
Reviewers: eugenis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D24437
llvm-svn: 281126
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
isDereferenceableInvariantLoad. NFC
Summary:
I want to separate out the notions of invariance and dereferenceability
at the MI level, so that they correspond to the equivalent concepts at
the IR level. (Currently an MI load is MI-invariant iff it's
IR-invariant and IR-dereferenceable.)
First step is renaming this function.
Reviewers: chandlerc
Subscribers: MatzeB, jfb, llvm-commits
Differential Revision: https://reviews.llvm.org/D23370
llvm-svn: 281125
|
|
|
|
| |
llvm-svn: 281124
|
|
|
|
| |
llvm-svn: 281122
|
|
|
|
|
|
|
|
|
|
| |
If the literal is being folded into src0, it doesn't matter
if it's an SGPR because it's being replaced with the literal.
Also fixes initially selecting 32-bit versions of some instructions
which also confused commuting.
llvm-svn: 281117
|
|
|
|
|
|
|
|
|
| |
This would create a bitcast use which fails the verifier: swifterror values may
only be used by loads, stores, and as function arguments.
rdar://28233244
llvm-svn: 281114
|
|
|
|
|
|
|
|
|
| |
This extends the optimization in r280832 to also work for 64-bit. The only
quirk is that we can't do this for 64-bit Windows (yet).
Differential Revision: https://reviews.llvm.org/D24423
llvm-svn: 281113
|
|
|
|
| |
llvm-svn: 281112
|
|
|
|
| |
llvm-svn: 281111
|
|
|
|
|
|
|
| |
If the unaligned access has a dynamic offset, it may be odd which
would make the adjusted alignment incorrect to use.
llvm-svn: 281110
|
|
|
|
|
|
| |
vectors
llvm-svn: 281107
|
|
|
|
| |
llvm-svn: 281106
|
|
|
|
|
|
| |
Use getConstVector helper to correctly create v2i64/v4i64 constants on 32-bit targets
llvm-svn: 281105
|
|
|
|
|
|
|
| |
When p0 was added as an explicit operand to the duplex subinstructions,
the disassembler was not updated to reflect this.
llvm-svn: 281104
|
|
|
|
|
|
|
|
| |
ISel makes assumption about the order of phi nodes.
rdar://28190150
llvm-svn: 281095
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
Previously these only worked via NVPTX-specific intrinsics.
This change will allow us to convert these target-specific intrinsics
into the general LLVM versions, allowing existing LLVM passes to reason
about their behavior.
It also gets us some minor codegen improvements as-is, from situations
where we canonicalize code into one of these llvm intrinsics.
Reviewers: majnemer
Subscribers: llvm-commits, jholewinski, tra
Differential Revision: https://reviews.llvm.org/D24300
llvm-svn: 281092
|
|
|
|
|
|
|
|
|
| |
Move the target specific setup into the target specific lowering setup. As
pointed out by Anton, the initial change was moving this too high up the stack
resulting in a violation of the layering (the target generic code path setup
target specific bits). Sink this into the ARM specific setup. NFC.
llvm-svn: 281088
|
|
|
|
|
|
|
|
|
|
|
| |
SmallVectors are convenient, but they don't cover every use case.
In particular, they are fairly large (3 pointers + one element) and
there is no way to take ownership of the buffer to put it somewhere
else. This patch then adds a lower lever interface that works with
any buffer.
llvm-svn: 281082
|
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D23700
llvm-svn: 281081
|
|
|
|
|
|
|
|
|
|
| |
Reviewers: arsenm
Subscribers: arsenm, wdng, nhaehnle, llvm-commits
Differential Revision: https://reviews.llvm.org/D24405
llvm-svn: 281080
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We have various command line options that print the type of a
stream, the size of a stream, etc but nowhere that it can all be
viewed together.
Since a previous patch introduced the ability to dump the bytes
of a stream, this seems like a good place to present a full view
of the stream's properties including its size, what kind of data
it represents, and the blocks it occupies. So I added the
ability to print that information to the -stream-data command
line option.
llvm-svn: 281077
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Summary:
Widening load in GVN is too early because it will block other optimizations like PRE, LICM.
https://llvm.org/bugs/show_bug.cgi?id=29110
The SPECCPU2006 benchmark impact of this patch:
Reference: o2_nopatch
(1): o2_patched
Benchmark Base:Reference (1)
-------------------------------------------------------
spec/2006/fp/C++/444.namd 25.2 -0.08%
spec/2006/fp/C++/447.dealII 45.92 +1.05%
spec/2006/fp/C++/450.soplex 41.7 -0.26%
spec/2006/fp/C++/453.povray 35.65 +1.68%
spec/2006/fp/C/433.milc 23.79 +0.42%
spec/2006/fp/C/470.lbm 41.88 -1.12%
spec/2006/fp/C/482.sphinx3 47.94 +1.67%
spec/2006/int/C++/471.omnetpp 22.46 -0.36%
spec/2006/int/C++/473.astar 21.19 +0.24%
spec/2006/int/C++/483.xalancbmk 36.09 -0.11%
spec/2006/int/C/400.perlbench 33.28 +1.35%
spec/2006/int/C/401.bzip2 22.76 -0.04%
spec/2006/int/C/403.gcc 32.36 +0.12%
spec/2006/int/C/429.mcf 41.04 -0.41%
spec/2006/int/C/445.gobmk 26.94 +0.04%
spec/2006/int/C/456.hmmer 24.5 -0.20%
spec/2006/int/C/458.sjeng 28 -0.46%
spec/2006/int/C/462.libquantum 55.25 +0.27%
spec/2006/int/C/464.h264ref 45.87 +0.72%
geometric mean +0.23%
For most benchmarks, it's a wash, but we do see stable improvements on some benchmarks, e.g. 447,453,482,400.
Reviewers: davidxl, hfinkel, dberlin, sanjoy, reames
Subscribers: gberry, junbuml
Differential Revision: https://reviews.llvm.org/D24096
llvm-svn: 281074
|
|
|
|
| |
llvm-svn: 281073
|
|
|
|
| |
llvm-svn: 281069
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This simplifies a lot of code, and will actually be necessary for
an upcoming patch to serialize TPI record hash values.
The idea before was that visitors should be examining records, not
modifying them. But this is no longer true with a visitor that
constructs a CVRecord from Yaml. To handle this until now, we
were doing some fixups on CVRecord objects at a higher level, but
the code is really awkward, and it makes sense to just have the
visitor write the bytes into the CVRecord. In doing so I uncovered
a few bugs related to `Data` and `RawData` and fixed those.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D24362
llvm-svn: 281067
|
|
|
|
| |
llvm-svn: 281066
|
|
|
|
|
|
|
|
|
|
| |
This writes the full sequence of type records described in
Yaml to the TPI stream of the PDB file.
Reviewed By: rnk
Differential Revision: https://reviews.llvm.org/D24316
llvm-svn: 281063
|
|
|
|
|
|
|
|
|
| |
This can happen when the frontend knows the debug info will be emitted
somewhere else. Usually this happens for dynamic classes with out of
line constructors or key functions, but it can also happen when modules
are enabled.
llvm-svn: 281060
|