Summary:
It turns out that the previous code construct was not optimizing the allocation
and deallocation of batches. The class id was read as a class member (even
though a precomputed one) and nothing else was optimized. By changing the
construct this way, the compiler actually optimizes most of the allocation and
deallocation away to only work with a single class id, which not only saves some
CPU but also some code footprint.
Reviewers: alekseyshl, dvyukov
Reviewed By: dvyukov
Subscribers: dvyukov, delcypher, llvm-commits, #sanitizers
Differential Revision: https://reviews.llvm.org/D46961
llvm-svn: 332502
We need to clean up the DAG floating-point undef logic.
This process is similar to how we handled integer undef
logic in D43141.
And as we did there, I'm trying to reduce the patch by
changing tests that would probably become meaningless
once we make those fixes.
llvm-svn: 332501
We need to clean up the DAG floating-point undef logic.
This process is similar to how we handled integer undef
logic in D43141.
And as we did there, I'm trying to reduce the patch by
changing tests that would probably become meaningless
once we make those fixes.
llvm-svn: 332500
We need to clean up the DAG floating-point undef logic.
This process is similar to how we handled integer undef
logic in D43141.
And as we did there, I'm trying to reduce the patch by
changing tests that would probably become meaningless
once we make those fixes.
llvm-svn: 332499
As i64 types are not legal on 32-bit targets, insert these into a suitable zero vector and use the packed vXi64<->FP conversion instructions instead.
Fixes PR3163.
Differential Revision: https://reviews.llvm.org/D43441
llvm-svn: 332498
In post-commit review for r332416, Paul Robinson pointed out that the
test for -debugify-each is not checking what it needs to.
This commit tightens up the test.
llvm-svn: 332497
Summary:
Before this patch, signal handling wasn't signal safe. This leads to real-world
crashes. It used ManagedStatic inside of signals, this can allocate and can lead
to unexpected state when a signal occurs during llvm_shutdown (because
llvm_shutdown destroys the ManagedStatic). It also used cl::opt without custom
backing storage. Some de-allocation was performed as well. Acquiring a lock in a
signal handler is also a great way to deadlock.
We can't just disable signals on llvm_shutdown because the signals might do
useful work during that shutdown. We also can't just disable llvm_shutdown for
programs (instead of library uses of clang) because we'd have to then mark the
pointers as not leaked and make sure all the ManagedStatic uses are OK to leak
and remain so.
Move all of the code to lock-free datastructures instead, and avoid having any
of them in an inconsistent state. I'm not trying to be fancy, I'm not using any
explicit memory order because this code isn't hot. The only purpose of the
atomics is to guarantee that a signal firing on the same or a different thread
doesn't see an inconsistent state and crash. In some cases we might miss some
state (for example, we might fail to delete a temporary file), but that's fine.
Note that I haven't touched any of the backtrace support despite it not
technically being totally signal-safe. When that code is called we know
something bad is up and we don't expect to continue execution, so calling
something that e.g. sets errno is the least of our problems.
A similar patch should be applied to lib/Support/Windows/Signals.inc, but that
can be done separately.
Fix r332428 which I reverted in r332429. I originally used double-wide CAS
because I was lazy, but some platforms use a runtime function for that which
thankfully failed to link (it would have been bad for signal handlers
otherwise). I use a separate flag to guard the data instead.
<rdar://problem/28010281>
Reviewers: dexonsmith
Subscribers: steven_wu, llvm-commits
llvm-svn: 332496
We already know where the CUDA SDK is, so there is no point in
letting Clang search for it again and possibly finding no or
a different installation.
--cuda-path is supported since the beginning of CUDA support in
Clang, so making this required doesn't impose additional restrictions.
Differential Revision: https://reviews.llvm.org/D46930
llvm-svn: 332495
Move all logic related to selecting the bitcode compiler and linker
into a new file and dynamically test required compiler flags. This
also adds -fcuda-rdc for Clang trunk as previously attempted in D44992
which fixes the build.
As a result this change also enables building the library by default
if all prerequisites are met.
Differential Revision: https://reviews.llvm.org/D46901
llvm-svn: 332494
Summary: This change will help us turn the DispatchUnit into its own stage.
Reviewers: andreadb, RKSimon, courbet
Reviewed By: andreadb, courbet
Subscribers: mgorny, tschuett, gbedwell, llvm-commits
Differential Revision: https://reviews.llvm.org/D46916
llvm-svn: 332493
Clang often tries to create implicit module import for error recovery,
which does a great job helping out with diagnostics. However, sometimes
clang does not have enough information given that it's using an invalid
context to move on. Be more strict in those cases to avoid crashes.
We hit crash on invalids because of this but unfortunately there are no
testcases and I couldn't manage to create one. The crashtrace however
indicates pretty clear why it's happening.
rdar://problem/39313933
llvm-svn: 332491
As part of merging stores we check that fusing the nodes does not
cause a cycle due to one candidate store being indirectly dependent on
another store (this may happen via chained memory copies). This is
done by searching if a store is a predecessor to another store's
value.
Prune the search at the candidate search's root node which is a
predecessor to all candidate stores. This reduces the
size of the subgraph searched in large basic blocks.
Reviewers: jyknight
Subscribers: llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D46955
llvm-svn: 332490
An assertion was not prepared to be passed a nullptr because the
out-of-quota limit was exceeded. Bail-out before the assertion
since the assertion does not apply on out-of-quote.
This fixes llvm.org/PR37477.
llvm-svn: 332488
As far as I can tell from revision history, there's no good reason to call
these files .so instead of .dll in Windows, so use the normal extension.
Also change PipSquak from SHARED to MODULE -- it's never passed to
target_link_libraries() and only loaded via dlopen(), so MODULE is more
appropriate. This makes it possible to delete a workaround for SHARED ldflags
being not quite right as well.
No intended behavior change.
https://reviews.llvm.org/D46898
llvm-svn: 332487
For regular SVE vector operands, this patch introduces a more
sensible diagnostic when the vector has a wrong suffix (e.g. z0.s vs z0.b).
For example:
add z0.s, z1.s, z2.b -> invalid element width
^_____^
mismatch
For the vector-with-shift/extend (e.g. z0.s, uxtw #2) this patch takes
a slightly different approach and instead returns a 'invalid operand'
if the element size is not as expected. This is because the diagnostics
are more specificied to suggest using the right shift/extend suffix. This
is a trade-off not to introduce more operand classes and still provide
useful diagnostics for LD1 and PRF instructions.
For example:
ld1w z1.s, p0/z, [x0, z0.s] -> invalid shift/extend specified, expected 'z[0..31].s, (uxtw|sxtw)'
ld1w z1.d, p0/z, [x0, z0.s] -> invalid operand
^________________^
mismatch
For gather prefetches, both 'z0.s' and 'z0.d' would be allowed:
prfw #0, p0, [x0, z0.s] -> invalid shift/extend specified, expected 'z[0..31].s, (uxtw|sxtw) #2'
prfw #0, p0, [x0, z0.d] -> invalid shift/extend specified, expected 'z[0..31].d, (lsl|uxtw|sxtw) #2'
Without this change, the diagnostic would unnecessarily suggest a
different element size:
prfw #0, p0, [x0, z0.s] -> invalid shift/extend specified, expected 'z[0..31].d, (lsl|uxtw|sxtw) #2'
Reviewers: SjoerdMeijer, aemerson, fhahn, samparker, javed.absar
Reviewed By: SjoerdMeijer
Differential Revision: https://reviews.llvm.org/D46688
llvm-svn: 332483
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
The canonicalization was restricted to shuffle masks with
a 1-to-1 mapping to the constant vector, but that disqualifies
the common splat pattern. This is part of solving PR37463:
https://bugs.llvm.org/show_bug.cgi?id=37463
llvm-svn: 332479
Summary:
For the 32-bit TransferBatch:
- `SetFromArray` callers have bounds `count`, so relax the `CHECK` to `DCHECK`;
- same for `Add`;
- mark `CopyToArray` as `const`;
For the 32-bit Primary:
- `{Dea,A}llocateBatch` are only called from places that check `class_id`,
relax the `CHECK` to `DCHECK`;
- same for `AllocateRegion`;
- remove `GetRegionBeginBySizeClass` that is not used;
- use a local variable for the random shuffle state, so that the compiler can
use a register instead of reading and writing to the `SizeClassInfo` at every
iteration;
For the 32-bit local cache:
- pass the count to drain instead of doing a `Min` everytime which is at times
superfluous.
Reviewers: alekseyshl
Reviewed By: alekseyshl
Subscribers: kubamracek, delcypher, #sanitizers, llvm-commits
Differential Revision: https://reviews.llvm.org/D46657
llvm-svn: 332478
functions.
If the combined construct is specified in the declare target function
and the device code is emitted, the compiler crashes because of the
incorrectly chosen captured stmt. We should choose the innermost
captured statement, not the outermost.
llvm-svn: 332477
The module ID numbering typically starts at 0 (in both the new and old
LTO APIs, used by linkers). Make llvm-lto consistent with that.
Split out of D46699.
llvm-svn: 332476
There shouldn't be any tests that run the entire optimizer here,
but the last test in this file is definitely going to break with
a change in LLVM IR canonicalization. Change that part to check
the unoptimized IR because that's the real intent of this file.
llvm-svn: 332473
Summary: This patch aims to remove the usage of old C-styled isl functions (in this case `isl_set_n_basic_set()`) in favor of new C++ isl interface based methods in `ScopInfo.cpp`.
Patch by Sahil Yerawar
Differential Revision: https://reviews.llvm.org/D46935
llvm-svn: 332471
Add support for __declspec(code_seg("segname"))
This patch is built on the existing support for #pragma code_seg. The code_seg
declspec is allowed on functions and classes. The attribute enables the
placement of code into separate named segments, including compiler-generated
members and template instantiations.
For more information, please see the following:
https://msdn.microsoft.com/en-us/library/dn636922.aspx
A new CodeSeg attribute is used instead of adding a new spelling to the existing
Section attribute since they don’t apply to the same Subjects. Section
attributes are also added for the code_seg declspec since they are used for
#pragma code_seg. No CodeSeg attributes are added to the AST.
The patch is written to match with the Microsoft compiler’s behavior even where
that behavior is a little complicated (see https://reviews.llvm.org/D22931, the
Microsoft feedback page is no longer available since MS has removed the page).
That code is in getImplicitSectionAttrFromClass routine.
Diagnostics messages are added to match with the Microsoft compiler for code-seg
attribute mismatches on base and derived classes and virtual overrides.
Differential Revision: https://reviews.llvm.org/D43352
llvm-svn: 332470
The parsing that is done for code completion is a special case that will
discard any generated diagnostics, so avoid running plugins for this
case in the first place to avoid performance penalties due to the
plugins.
A scenario for this is for example libclang with extra plugins like tidy.
Patch by Nikolai Kosjar
Differential Revision: https://reviews.llvm.org/D46050
llvm-svn: 332469
If the orphaned directive is executed in SPMD mode, we need to emit the
check for the SPMD mode and run the orphaned parallel directive in
sequential mode.
llvm-svn: 332467
Summary:
A recent patch ([[ https://reviews.llvm.org/rL331587 | rL331587 ]]) to Capture Tracking taught it that the `launder_invariant_group` intrinsic captures its argument only by returning it. Unfortunately, BasicAA still considered every call instruction as a possible escape source and hence concluded that the result of a `launder_invariant_group` call cannot alias any local non-escaping value. This led to [[ https://bugs.llvm.org/show_bug.cgi?id=37458 | bug 37458 ]].
This patch updates the relevant check for escape sources in BasicAA.
Reviewers: Prazek, kuhar, rsmith, hfinkel, sanjoy, xbolva00
Reviewed By: hfinkel, xbolva00
Subscribers: JDevlieghere, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D46900
llvm-svn: 332466
Since there is no perfect way bind the non-zero value with the default binding, this patch only considers the case where buffer's offset is zero and the char value is 0. And according to the value for overwriting, decide how to update the string length.
Reviewers: dcoughlin, NoQ, xazax.hun, a.sidorin, george.karpenkov
Reviewed By: NoQ
Differential Revision: https://reviews.llvm.org/D44934
llvm-svn: 332463
Revision 332390 introduced a FetchStage class in llvm-mca.
By design, FetchStage owns all the instructions in-flight in the OoO Backend.
Before this change, new instructions were added to a DenseMap indexed by
instruction id. The problem with using a DenseMap is that elements are not
ordered by key. This was causing a massive slow down in method
FetchStage::postExecute(), which searches for instructions retired that can be
deleted.
This patch replaces the DenseMap with a std::map ordered by instruction index.
At the end of every cycle, we search for the first instruction which is not
marked as "retired", and we remove all the previous instructions before it.
This works well because instructions are retired in-order.
Before this patch, a debug build of llvm-mca (on my Ryzen linux machine) took
~8.0 seconds to simulate 3000 iterations of a x86 dot-product (a `vmulps,
vpermilps, vaddps, vpermilps, vaddps` sequence). With this patch, it now takes
~0.8s to run all the 3000 iterations.
llvm-svn: 332461
Summary:
And add tests for the comment extraction code.
clangd will now show non-doxygen comments in completion for results
coming from Sema and Dynamic index.
Static index does not include the comments yet, I will enable it in
a separate commit after investigating which implications it has for
the size of the index.
Reviewers: sammccall, hokein, ioeric
Reviewed By: sammccall
Subscribers: klimek, MaskRay, jkorous, cfe-commits
Differential Revision: https://reviews.llvm.org/D46002
llvm-svn: 332460
Summary:
Previous implementation used to extract brief text from doxygen comments.
Brief text parsing slows down completion and is not suited for
non-doxygen comments.
This commit switches to providing comments that mimic the ones
originally written in the source code, doing minimal reindenting and
removing the comments markers to make the output more user-friendly.
It means we lose support for doxygen-specific features, e.g. extracting
brief text, but provide useful results for non-doxygen comments.
Switching the doxygen support back is an option, but I suggest to see
whether the current approach gives more useful results.
Reviewers: sammccall, hokein, ioeric
Reviewed By: sammccall
Subscribers: klimek, MaskRay, jkorous, cfe-commits
Differential Revision: https://reviews.llvm.org/D45999
llvm-svn: 332459
Summary:
The helper is used in clangd for documentation shown in code completion
and storing the docs in the symbols. See D45999.
This patch reuses the code of the Doxygen comment lexer, disabling the
bits that do command and html tag parsing.
The new helper works on all comments, including non-doxygen comments.
However, it does not understand or transform any doxygen directives,
i.e. cannot extract brief text, etc.
Reviewers: sammccall, hokein, ioeric
Reviewed By: ioeric
Subscribers: mgorny, cfe-commits
Differential Revision: https://reviews.llvm.org/D46000
llvm-svn: 332458
Summary:
This uses heuristics to identify private proto symbols. For example,
top-level symbols whose name contains "_" are considered private. These symbols
are not expected to be used by users.
Reviewers: ilya-biryukov, malaperle
Reviewed By: ilya-biryukov
Subscribers: sammccall, klimek, MaskRay, jkorous, cfe-commits
Differential Revision: https://reviews.llvm.org/D46751
llvm-svn: 332456
Kostya Kortchinsky