Commit Graph

376516 Commits

Author SHA1 Message Date
Adam Czachorowski 2e1bb7940a [clangd] Add missing "override" to fix the build.
Follow-up to d4af86581e

Differential Revision: https://reviews.llvm.org/D94314
2021-01-08 17:24:47 +01:00
David Green e185b1dd7b [ConstProp] Constant propagation for get.active.lane.mask instrinsics
Similar to the Arm VCTP intrinsics, if the operands of an
active.lane.mask are both known, the constant lane mask can be
calculated. This can come up after unrolling the loops.

Differential Revision: https://reviews.llvm.org/D94103
2021-01-08 16:10:01 +00:00
Adam Czachorowski d4af86581e [clangd] Fix type printing in the presence of qualifiers
When printing QualType with qualifiers like "const", or pointing to an
elaborated type, we would print garbage like:
  std::const std::vector<int>&
with the initial std:: being calculated correctly, but inserted in the
wrong place and the second std:: not removed (due to elaborated type).

This affected, among others, ExtractFunction and ExpandAuto tweaks.

This change introduces a new callback to PrintingPolicy, which allows us
to influence the printing of namespace qualifiers. In the future, the
same callback can be used to improve handling of "using namespace"
directives as well.

Fixes:
  https://github.com/clangd/clangd/issues/640 (ExtractFunction)
  https://github.com/clangd/clangd/issues/264 (ExpandAuto)
  First point of https://github.com/clangd/clangd/issues/524

Differential Revision: https://reviews.llvm.org/D94259
2021-01-08 17:00:39 +01:00
Eugene Zhulenev 740950cf1c Revert "[mlir:JitRunner] Use custom shared library init/destroy functions if available"
This reverts commit 84dc9b451b.

Fix Windows breakage: http://lab.llvm.org:8011/#/builders/13/builds/3658/steps/6/logs/stdio

Differential Revision: https://reviews.llvm.org/D94309
2021-01-08 07:46:35 -08:00
Simon Pilgrim 80dee7965d [X86][SSE] Fold unpack(hop(),hop()) -> permute(hop())
UNPCKL/UNPCKH only uses one op from each hop, so we can merge the hops and then permute the result.
2021-01-08 15:22:17 +00:00
David Green a36a2864c0 [ARM][LV] Additional loop invariant reduction test. NFC 2021-01-08 15:15:08 +00:00
Eugene Zhulenev 84dc9b451b [mlir:JitRunner] Use custom shared library init/destroy functions if available
Use custom mlir runner init/destroy functions to safely init and destroy shared libraries loaded by the JitRunner.

Reviewed By: mehdi_amini

Differential Revision: https://reviews.llvm.org/D94270
2021-01-08 07:14:21 -08:00
Paul C. Anagnostopoulos d4f2fef746 [TableGen] Remove unused declaration that caused build failures. 2021-01-08 10:10:54 -05:00
Heejin Ahn 7be271537e [WebAssembly] Rename wasm_rethrow_in_catch intrinsic/builtin
`wasm_rethrow_in_catch` intrinsic and builtin are used in order to
rethrow an exception when the exception is caught but there is no
matching clause within the current `catch`. For example,
```
try {
  foo();
} catch (int n) {
  ...
}
```
If the caught exception does not correspond to C++ `int` type, it should
be rethrown. These intrinsic/builtin were renamed `rethrow_in_catch`
because at the time I thought there would be another intrinsic for C++'s
`throw` keyword, which rethrows an exception. It turned out that `throw`
keyword doesn't require wasm's `rethrow` instruction, so we rename
`rethrow_in_catch` to just `rethrow` here.

Reviewed By: dschuff, tlively

Differential Revision: https://reviews.llvm.org/D94038
2021-01-08 06:55:04 -08:00
David Green 1ae762469f [ARM] Update and regenerate test checks. NFC 2021-01-08 14:54:16 +00:00
Paul C. Anagnostopoulos 6e2b6351d2 [TableGen] Add the assert statement, step 1
Differential Revision: https://reviews.llvm.org/D93911

This first step adds the assert statement and supports it at top level
and in record definitions. Later steps will support it in class
definitions and multiclasses.
2021-01-08 09:47:51 -05:00
Kiran Chandramohan 268ff38a71 [MLIR][OpenMP] Attribute to include WsLoop upperbound
This patch adds an attribute `inclusive` which if present causes
the upperbound to be included in the loop iteration interval.

Reviewed By: ftynse
Differential Revision: https://reviews.llvm.org/D94235
2021-01-08 14:42:18 +00:00
Simon Pilgrim 4a582d766a [X86][SSE] Add vphaddd/vphsubd unpack(hop(),hop()) tests 2021-01-08 14:39:37 +00:00
Alexey Bataev 0e57084d0e [SLP][NFC]Add a test for reused shrink check, NFC. 2021-01-08 06:23:23 -08:00
Xiangling Liao e97071d795 [NFC] Renaming PackStack to AlignPackStack
This patch renames PackStack and related variable names to also contain align across Clang.
As it is right now, Clang already uses one stack to record the information from both #pragma
align and #pragma pack. Leaving it as PackStack is confusing, and could cause people to
ignore #pragma align when developing code that interacts with PackStack.

Differential Revision: https://reviews.llvm.org/D93901
2021-01-08 09:15:11 -05:00
Simon Pilgrim 7b9f541c1e [X86][SSE] Add tests for unpack(hop(),hop())
We should be able to convert these to permute(hop()) as we only ever use one of the ops from each hop.
2021-01-08 14:11:37 +00:00
KareemErgawy-TomTom 25d882e758 [MLIR][SPIRV] Add `UsableInSpecConstantOp` trait.
Instead of checking explicitly checking for whether an op is usalbe
inside a `SpecConstantOperationOP`, this commit adds a new trait to
filter such ops.

Reviewed By: antiagainst

Differential Revision: https://reviews.llvm.org/D94288
2021-01-08 15:07:40 +01:00
Kazushi (Jam) Marukawa 5ead757f1d [VE] Support pack_f32p and pack_f32a intrinsic instructions
Support pack_f32p and pack_f32a intrinsic instructions and regression tests.

Reviewed By: simoll

Differential Revision: https://reviews.llvm.org/D94296
2021-01-08 22:59:11 +09:00
Heejin Ahn d012430eee [WebAssembly] Change label numbers to variables in test
cfg-stackify-eh.ll contains many `CHECK` lines specifying label / catch
comments with numbers. These numbers are subject to change every time
any block/loop/try is added in the middle in existing functions or any
other function is added in the middle of the file, generating a large
number of lines in diffs. This change converts them to variables so they
can be more resistent to future changes.

Reviewed By: dschuff, tlively

Differential Revision: https://reviews.llvm.org/D94037
2021-01-08 05:49:59 -08:00
Anastasia Stulova 0ef2b68ff0 [OpenCL] Documentation for experimental C++ libs
Started a new doc section about the OpenCL experimental
features and described ongoing work on C++ libraries
e.g. type traits.

Tags: #clang

Differential Revision: https://reviews.llvm.org/D94188
2021-01-08 13:45:59 +00:00
Alexander Belyaev bcbdeafa9c Revert "[SLP]Need shrink the load vector after reordering."
This reverts commit 4284afdf94.

This changes computed values in fused_batchnorm_test_cpu.

Not equal to tolerance rtol=1e-06, atol=0.001
Mismatched value: a is different from b.
not close where = (array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1]), array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1]), array([0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1,
       1, 1]), array([0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5, 0, 1, 2, 3, 4, 5, 0, 1, 2, 3,
       4, 5]))
not close lhs = [-0.6636615  -0.9804948  -1.148275   -0.68193716 -0.8572368  -0.65046215
 -0.6993756  -1.2244141  -1.0938729  -0.50369143 -0.51830524 -0.738452
 -0.7214286  -0.48115745 -0.9380924  -0.9341769  -0.5916775  -1.2896856
 -0.7264182  -0.9746917  -0.783249   -0.7659018  -0.86214024 -0.47784212]
not close rhs = [ 0.44102234  0.12418899 -0.04359123  0.42274666  0.24744703  0.45422167
  0.40530816 -0.11973029  0.01081094  0.6009924   0.5863786   0.3662318
  0.38325527  0.62352633  0.1665914   0.1705069   0.5130063  -0.18500176
  0.37826565  0.12999213  0.3214348   0.338782    0.24254355  0.62684166]
not close dif = [1.1046839 1.1046838 1.1046838 1.1046839 1.1046839 1.1046839 1.1046838
 1.1046839 1.1046839 1.1046839 1.1046839 1.1046839 1.1046839 1.1046838
 1.1046839 1.1046839 1.1046839 1.1046839 1.1046839 1.1046839 1.1046839
 1.1046839 1.1046838 1.1046838]
not close tol = [0.00100044 0.00100012 0.00100004 0.00100042 0.00100025 0.00100045
 0.00100041 0.00100012 0.00100001 0.0010006  0.00100059 0.00100037
 0.00100038 0.00100062 0.00100017 0.00100017 0.00100051 0.00100019
 0.00100038 0.00100013 0.00100032 0.00100034 0.00100024 0.00100063]
2021-01-08 14:42:26 +01:00
LLVM GN Syncbot 657db0c6d4 [gn build] Port 9c4b2225b2 2021-01-08 13:30:54 +00:00
Simon Moll 611d3c63f3 [VP] ISD helper functions [VE] isel for vp_add, vp_and
This implements vp_add, vp_and for the VE target by lowering them to the
VVP_* layer. We also add helper functions for VP SDNodes (isVPSDNode,
getVPMaskIdx, getVPExplicitVectorLengthIdx).

Reviewed By: kaz7

Differential Revision: https://reviews.llvm.org/D93766
2021-01-08 14:29:45 +01:00
Raphael Isemann b0dc54e08a [lldb][NFC] Refactor setup code for Clang diagnostics 2021-01-08 14:26:04 +01:00
Shivam Gupta 195ffcd890 [MLIR][NFC] remove unnecessary includes form tablegen command
With [[ https://reviews.llvm.org/D77156 | D77156 ]] includes are not needed here.

Reviewed By: ftynse

Differential Revision: https://reviews.llvm.org/D94216
2021-01-08 14:25:52 +01:00
Mark Murray 7d4a8bc417 [AArch64] Add +flagm archictecture option, allowing the v8.4a flag modification extension.
Differential Revision: https://reviews.llvm.org/D94081
2021-01-08 13:21:12 +00:00
Mark Murray af7cce2fa4 [AArch64] Add +pauth archictecture option, allowing the v8.3a pointer authentication extension.
Differential Revision: https://reviews.llvm.org/D94083
2021-01-08 13:21:11 +00:00
Alexander Belyaev 9c4b2225b2 Revert "Revert "Revert "Revert "Revert "[analyzer] NFC: Move path diagnostic consumer implementations to libAnalysis."""""
This reverts commit 6b0ee02747.

Circular dependency again.
2021-01-08 14:17:18 +01:00
Sanjay Patel 267ff7901c [SLP] limit verifyFunction to debug build (PR48689)
As noted in PR48689, the verifier may have some kind
of exponential behavior that should be addressed
separately. For now, only run it in debug mode to
prevent problems for release+asserts.
That limit is what we had before D80401, and I'm
not sure if there was a reason to change it in that
patch.
2021-01-08 08:10:17 -05:00
Haojian Wu ed3b1f9061 [clangd] go-to-implementation on a base class jumps to all subclasses.
Differential Revision: https://reviews.llvm.org/D92749
2021-01-08 13:50:57 +01:00
Alex Richardson 32733c347d [libc++] Add missing XFAIL to tests that need __atomic_* libcalls
FreeBSD did not provide the __atomic_* functions as part of the base
system until recently. They were added to libgcc_s in SVN revision r364753
(August 2020), so check for availability of 'non-lockfree-atomics' so that
these tests do not fail unexpectedly on older versions of FreeBSD.

This also removes the #ifndef __APPLE__ from atomic_helpers.h that was used
to work around lack of atomic runtime functions on older Apple platforms
and replaces it with XFAIL: !non-lockfree-atomics.

Reviewed By: #libc, ldionne
Differential Revision: https://reviews.llvm.org/D88818
2021-01-08 12:48:22 +00:00
Alex Richardson 00530dee5d [compiler-rt] Implement __atomic_is_lock_free
This function is called by the __atomic_is_lock_free() builtin if the value
cannot be resolved to true at compile time. Lack of this function is
causing the non-lockfree atomics tests in libc++ to not be run (see D91911)

This function is also added in D85044, but that review also adds support
for using lock-free atomics in more cases, whereas this is a minimal change
that just adds __atomic_is_lock_free() for the implementation of atomic.c.

Reviewed By: ldionne

Differential Revision: https://reviews.llvm.org/D92302
2021-01-08 12:48:22 +00:00
Kazushi (Jam) Marukawa 99c84af1a2 [VE][NFC] Clean ISel patterns for LSV and LVS
Clean ISel patterns for LSV and LVS before upstream more hand-written
ISel patterns.

Reviewed By: simoll

Differential Revision: https://reviews.llvm.org/D94291
2021-01-08 21:47:33 +09:00
Haojian Wu c909512fdb [clangd] Cleanup a remaining Optional<SymbolID> usage, NFC. 2021-01-08 13:44:20 +01:00
Haojian Wu b83b7d0873 [clangd] NFC, avoid potential ODR violation. 2021-01-08 13:29:11 +01:00
Nicholas Guy ed23229a64 [AArch64] Fix crash caused by invalid vector element type
Fixes a crash caused by D91255, when LLVMTy is null when
calling changeExtendedVectorElementType.

Differential Revision: https://reviews.llvm.org/D94234
2021-01-08 12:02:54 +00:00
David Sherwood 38d18d9353 [SVE] Add support to vectorize_width loop pragma for scalable vectors
This patch adds support for two new variants of the vectorize_width
pragma:

1. vectorize_width(X[, fixed|scalable]) where an optional second
parameter is passed to the vectorize_width pragma, which indicates if
the user wishes to use fixed width or scalable vectorization. For
example the user can now write something like:

  #pragma clang loop vectorize_width(4, fixed)
or
  #pragma clang loop vectorize_width(4, scalable)

In the absence of a second parameter it is assumed the user wants
fixed width vectorization, in order to maintain compatibility with
existing code.
2. vectorize_width(fixed|scalable) where the width is left unspecified,
but the user hints what type of vectorization they prefer, either
fixed width or scalable.

I have implemented this by making use of the LLVM loop hint attribute:

  llvm.loop.vectorize.scalable.enable

Tests were added to

  clang/test/CodeGenCXX/pragma-loop.cpp

for both the 'fixed' and 'scalable' optional parameter.

See this thread for context: http://lists.llvm.org/pipermail/cfe-dev/2020-November/067262.html

Differential Revision: https://reviews.llvm.org/D89031
2021-01-08 11:37:27 +00:00
Cullen Rhodes 1e7efd397a [LV] Legalize scalable VF hints
In the following loop:

  void foo(int *a, int *b, int N) {
    for (int i=0; i<N; ++i)
      a[i + 4] = a[i] + b[i];
  }

The loop dependence constrains the VF to a maximum of (4, fixed), which
would mean using <4 x i32> as the vector type in vectorization.
Extending this to scalable vectorization, a VF of (4, scalable) implies
a vector type of <vscale x 4 x i32>. To determine if this is legal
vscale must be taken into account. For this example, unless
max(vscale)=1, it's unsafe to vectorize.

For SVE, the number of bits in an SVE register is architecturally
defined to be a multiple of 128 bits with a maximum of 2048 bits, thus
the maximum vscale is 16. In the loop above it is therefore unfeasible
to vectorize with SVE. However, in this loop:

  void foo(int *a, int *b, int N) {
    #pragma clang loop vectorize_width(X, scalable)
    for (int i=0; i<N; ++i)
      a[i + 32] = a[i] + b[i];
  }

As long as max(vscale) multiplied by the number of lanes 'X' doesn't
exceed the dependence distance, it is safe to vectorize. For SVE a VF of
(2, scalable) is within this constraint, since a vector of <16 x 2 x 32>
will have no dependencies between lanes. For any number of lanes larger
than this it would be unsafe to vectorize.

This patch extends 'computeFeasibleMaxVF' to legalize scalable VFs
specified as loop hints, implementing the following behaviour:
  * If the backend does not support scalable vectors, ignore the hint.
  * If scalable vectorization is unfeasible given the loop
    dependence, like in the first example above for SVE, then use a
    fixed VF.
  * Accept scalable VFs if it's safe to do so.
  * Otherwise, clamp scalable VFs that exceed the maximum safe VF.

Reviewed By: sdesmalen, fhahn, david-arm

Differential Revision: https://reviews.llvm.org/D91718
2021-01-08 10:49:44 +00:00
Simon Moll eeba70a463 [VE] Expand single-element BUILD_VECTOR to INSERT_VECTOR_ELT
We do this mostly to be able to test the insert_vector_elt isel
patterns. As long as we don't, most single element insertions show up as
`BUILD_VECTOR` in the backend.

Reviewed By: kaz7

Differential Revision: https://reviews.llvm.org/D93759
2021-01-08 11:48:01 +01:00
Simon Moll d1b606f897 [VE] Extract & insert vector element isel
Isel and tests for extract_vector_elt and insert_vector_elt.

Reviewed By: kaz7

Differential Revision: https://reviews.llvm.org/D93687
2021-01-08 11:46:59 +01:00
Christian Sigg 2fe625e9ab Fix two pessimizing moves.
See https://en.cppreference.com/w/cpp/language/return#Automatic_move_from_local_variables_and_parameters

Reviewed By: rriddle

Differential Revision: https://reviews.llvm.org/D94218
2021-01-08 11:44:29 +01:00
Qiu Chaofan 6175fcf01f [NFC] Update some PPC tests marked as auto-generated
Update CodeGen regression tests with marker at first line telling it's
auto-generated by the script, under PowerPC directory. For some reason,
these tests are generated but manually written, which makes things
unclear when someone's change affecting them.

However, some tests only show simple change after re-generated, like
extra blank lines, disappearing '.localentry', etc. Besides, some tests
are generated but added checks for debug output. This commit doesn't try
updating them.
2021-01-08 17:59:13 +08:00
Jan Svoboda 8e3e148c88 Revert "[clang][cli] Port DiagnosticOpts to new option parsing system"
This reverts commit 8e3230ff
2021-01-08 10:53:12 +01:00
David Green 72fb5ba079 [LV] Don't sink into replication regions
The new test case here contains a first order recurrences and an
instruction that is replicated. The first order recurrence forces an
instruction to be sunk _into_, as opposed to after the replication
region. That causes several things to go wrong including registering
vector instructions multiple times and failing to create dominance
relations correctly.

Instead we should be sinking to after the replication region, which is
what this patch makes sure happens.

Differential Revision: https://reviews.llvm.org/D93629
2021-01-08 09:50:10 +00:00
Kazushi (Jam) Marukawa 12167632bc [VE] Add SVOB intrinsic instruction
Add SVOB intrinsic instruction and a regression test.

Reviewed By: simoll

Differential Revision: https://reviews.llvm.org/D94279
2021-01-08 18:49:17 +09:00
Jan Svoboda 8e3230ffa3 [clang][cli] Port DiagnosticOpts to new option parsing system
This patch introduces additional infrastructure necessary to accommodate DiagnosticOptions.

DiagnosticOptions are unique in that they are parsed by the same function in cc1 AND in the Clang driver. The call to the parsing function from the driver occurs early on in the compilation process, where no proper DiagnosticEngine exists, because the diagnostic options (passed through command line) are not known yet.

To preserve the current behavior, we need to be able to selectively parse:
* all options (for -cc1),
* only diagnostic options (for driver).

This patch achieves that in the following way:
* new MacroPrefix field is added to the Option TableGen class,
* new IsDiag TableGen mixin sets MacroPrefix to "DIAG_",
* TableGen backend serializes option records into a macro with the prefix,
* CompilerInvocation parse/generate methods define the [DIAG_]OPTION_WITH_MARSHALLING macros to handle diagnostic options separately.

Depends on D93700, D93701 & D93702.

Reviewed By: dexonsmith

Differential Revision: https://reviews.llvm.org/D84673
2021-01-08 10:44:22 +01:00
Sjoerd Meijer 8af859d514 [MachineLoop] New helper isLoopInvariant()
This factors out code from MachineLICM that determines whether an instruction
is loop-invariant, which is a generally useful function. Thus this allows to
use that helper elsewhere too.

Differential Revision: https://reviews.llvm.org/D94082
2021-01-08 09:04:56 +00:00
David Sherwood d1bf26fd94 [AArch64][SVE] Add lowering for llvm abs intrinsic
Add functionality to permit lowering of the abs and neg intrinsics
using the passthru variants.

Differential Revision: https://reviews.llvm.org/D94160
2021-01-08 08:55:25 +00:00
Martin Storsjö 7a91dad9e5 [llvm-readobj] [ARMWinEH] Clearly print an invalid case of packed unwind info as such
As the actual windows unwinder doesn't support this case, don't
pretend that it is supported when dumping the generated unwind info
either, even if it would be possible to interpret it as something
sensible.

This should reduce the risk of us emitting such a case in code
(although it's unlikely as long as the unwind info is generated
through the SEH opcodes, as the opcodes can't describe this case).

Differential Revision: https://reviews.llvm.org/D91529
2021-01-08 10:04:44 +02:00
Wang, Pengfei c102b9697b [X86] Correct the comments about comparison intrinsics. NFCI. 2021-01-08 15:36:15 +08:00