Install a cache of DBG_INSTR_REF -> ValueIDNum resolutions, for scenarios
where the value has to be reconstructed from several DBG_PHIs. Whenever
this happens, it's because branch folding + tail duplication has messed
with the SSA form of the program, and we have to solve a mini SSA problem
to find the variable value. This is always called twice, so it makes sense
to cache the value.
This gives a ~0.5% geomean compile-time-performance improvement on CTMark.
Differential Revision: https://reviews.llvm.org/D118455
Adds new optimization remarks when vectorization fails.
More specifically, new remarks are added for following 4 cases:
- Backward dependency
- Backward dependency that prevents Store-to-load forwarding
- Forward dependency that prevents Store-to-load forwarding
- Unknown dependency
It is important to note that only one of the sources
of failures (to vectorize) is reported by the remarks.
This source of failure may not be first in program order.
A regression test has been added to test the following cases:
a) Loop can be vectorized: No optimization remark is emitted
b) Loop can not be vectorized: In this case an optimization
remark will be emitted for one source of failure.
Reviewed By: sdesmalen, david-arm
Differential Revision: https://reviews.llvm.org/D108371
None of the external users actual touch these (they're purely used internally down the recursive call) - its trivial to add another wrapper if anything ever does want to track known elements.
We experienced some deadlocks when we used multiple threads for logging
using `scan-builds` intercept-build tool when we used multiple threads by
e.g. logging `make -j16`
```
(gdb) bt
#0 0x00007f2bb3aff110 in __lll_lock_wait () from /lib/x86_64-linux-gnu/libpthread.so.0
#1 0x00007f2bb3af70a3 in pthread_mutex_lock () from /lib/x86_64-linux-gnu/libpthread.so.0
#2 0x00007f2bb3d152e4 in ?? ()
#3 0x00007ffcc5f0cc80 in ?? ()
#4 0x00007f2bb3d2bf5b in ?? () from /lib64/ld-linux-x86-64.so.2
#5 0x00007f2bb3b5da27 in ?? () from /lib/x86_64-linux-gnu/libc.so.6
#6 0x00007f2bb3b5dbe0 in exit () from /lib/x86_64-linux-gnu/libc.so.6
#7 0x00007f2bb3d144ee in ?? ()
#8 0x746e692f706d742f in ?? ()
#9 0x692d747065637265 in ?? ()
#10 0x2f653631326b3034 in ?? ()
#11 0x646d632e35353532 in ?? ()
#12 0x0000000000000000 in ?? ()
```
I think the gcc's exit call caused the injected `libear.so` to be unloaded
by the `ld`, which in turn called the `void on_unload() __attribute__((destructor))`.
That tried to acquire an already locked mutex which was left locked in the
`bear_report_call()` call, that probably encountered some error and
returned early when it forgot to unlock the mutex.
All of these are speculation since from the backtrace I could not verify
if frames 2 and 3 are in fact corresponding to the `libear.so` module.
But I think it's a fairly safe bet.
So, hereby I'm releasing the held mutex on *all paths*, even if some failure
happens.
PS: I would use lock_guards, but it's C.
Reviewed-by: NoQ
Differential Revision: https://reviews.llvm.org/D118439
Was reverted in 1c1b670a73 as it broke all non-x86 bots. Original commit
message:
[DebugInfo][InstrRef] Add a max-stack-slots-to-track cut-out
In certain circumstances with things like autogenerated code and asan, you
can end up with thousands of Values live at the same time, causing a large
working set and a lot of information spilled to the stack. Unfortunately
InstrRefBasedLDV doesn't cope well with this and consumes a lot of memory
when there are many many stack slots. See the reproducer in D116821.
It seems very unlikely that a developer would be able to reason about
hundreds of live named local variables at the same time, so a huge working
set and many stack slots is an indicator that we're likely analysing
autogenerated or instrumented code. In those cases: gracefully degrade by
setting an upper bound on the amount of stack slots to track. This limits
peak memory consumption, at the cost of dropping some variable locations,
but in a rare scenario where it's unlikely someone is actually going to
use them.
In terms of the patch, this adds a cl::opt for max number of stack slots to
track, and has the stack-slot-numbering code optionally return None. That
then filters through a number of code paths, which can then chose to not
track a spill / restore if it touches an untracked spill slot. The added
test checks that we drop variable locations that are on the stack, if we
set the limit to zero.
Differential Revision: https://reviews.llvm.org/D118601
This revision avoids incorrect hoisting of alloca'd buffers across an AutomaticAllocationScope boundary.
In the more general case, we will probably need a ParallelScope-like interface.
Differential Revision: https://reviews.llvm.org/D118768
This is to fix build errors "Cannot open include file:
'sanitizer/asan_interface.h'" when building LLVM with MSVC and
LLVM_USE_SANITIZER=Address.
asan_interface.h is not available in MSVC's search path, instead it is
located under %VCToolsInstallDir%/crt/src/sanitizer.
This is an alternate solution to https://reviews.llvm.org/D118159, to
avoid adding all internal crt sources to the header search paths.
Tested with visual studio 2019 v16.9.6 and visual studio 2022 v17.0.5
Reviewed By: aaron.ballman, rnk
Differential Revision: https://reviews.llvm.org/D118624
By fully qualifying the use of any types and functions from the mlir namespace, users are not required to add using namespace mlir; into the C++ file including the Tablegen output.
Differential Revision: https://reviews.llvm.org/D118767
This reverts commit 9d6a615973.
Exit Code: 1
Command Output (stderr):
--
/scratch/buildbot/bothome/clang-ve-ninja/llvm-project/clang/test/Analysis/analyze-function-guide.cpp:53:21: error: CHECK-EMPTY-NOT: excluded string found in input // CHECK-EMPTY-NOT: Every top-level function was skipped.
^
<stdin>:1:1: note: found here
Every top-level function was skipped.
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Input file: <stdin>
Check file: /scratch/buildbot/bothome/clang-ve-ninja/llvm-project/clang/test/Analysis/analyze-function-guide.cpp
-dump-input=help explains the following input dump.
Input was:
<<<<<<
1: Every top-level function was skipped.
not:53 !~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ error: no match expected
2: Pass the -analyzer-display-progress for tracking which functions are analyzed.
>>>>>>
Apparently GCC 5.4 (a supported compiler) has a bug where it will
use the "MachineInstr &MI" defined by the range-based for loop
to evaluate the for loop expression. Pick a different variable
name to avoid this.
Sometimes when I pass the mentioned option I forget about passing the
parameter list for c++ sources.
It would be also useful newcomers to learn about this.
This patch introduces some logic checking common misuses involving
`-analyze-function`.
Reviewed-By: martong
Differential Revision: https://reviews.llvm.org/D118690
This patch adds custom lowering support for ISD::SDIV and ISD::UDIV
when SVE is enabled, regardless of the minimum SVE vector length. We do
this because NEON simply does not have vector integer divide support, so
we want to take advantage of these instructions in SVE.
As part of this patch I've also simplified LowerToPredicatedOp to avoid
re-asking the same question about whether we should be using SVE for
fixed length vectors. Once we've made the decision to call
LowerToPredicatedOp, then we should simply assert we should be using SVE.
I've updated the 128-bit min SVE vector bits tests here:
CodeGen/AArch64/sve-fixed-length-int-div.ll
CodeGen/AArch64/sve-fixed-length-int-rem.ll
Differential Revision: https://reviews.llvm.org/D117871
After adding another value kind in 8a12cae862, Value * pointers do not
have enough available empty bits to store the kind (e.g. on ARM)
To address this, the patch replaces the PointerIntPair with separate
value and kind fields.
I was looking around and noticed that builtins for iossim, tvossim
and watchossim was missing arm64 builds, while apple's clang
toolchain ship with these. After a bit of searching around it just
seems like these are not listed correctly in CMake to be enabled.
I enabled just arm64 since I saw that Apple clang didn't include
arm64e.
Reviewed By: t.p.northover
Differential Revision: https://reviews.llvm.org/D118759
Fixes https://github.com/llvm/llvm-project/issues/31592.
This commits enables lexing of digraphs in C++11 and onwards.
Enabling them in C++03 is error-prone, as it would unconditionally treat sequences like "<:" as digraphs, even if they are followed by a single colon, e.g. "<::" would be treated as "[:" instead of "<" followed by "::". Lexing in C++11 doesn't have this problem as it looks ahead the following token.
The relevant excerpt from Lexer::LexTokenInternal:
```
// C++0x [lex.pptoken]p3:
// Otherwise, if the next three characters are <:: and the subsequent
// character is neither : nor >, the < is treated as a preprocessor
// token by itself and not as the first character of the alternative
// token <:.
```
Also, note that both clang and gcc turn on digraphs by default (-fdigraphs), so clang-format should match this behaviour.
Reviewed By: MyDeveloperDay, HazardyKnusperkeks, owenpan
Differential Revision: https://reviews.llvm.org/D118706
This patch extends the available-value logic to detect loads
of pointer-selects that can be replaced by a value select.
For example, consider the code below:
loop:
%sel.phi = phi i32* [ %start, %ph ], [ %sel, %ph ]
%l = load %ptr
%l.sel = load %sel.phi
%sel = select cond, %ptr, %sel.phi
...
exit:
%res = load %sel
use(%res)
The load of the pointer phi can be replaced by a load of the start value
outside the loop and a new phi/select chain based on the loaded values,
as illustrated below
%l.start = load %start
loop:
sel.phi.prom = phi i32 [ %l.start, %ph ], [ %sel.prom, %ph ]
%l = load %ptr
%sel.prom = select cond, %l, %sel.phi.prom
...
exit:
use(%sel.prom)
This is a first step towards alllowing vectorizing loops using common libc++
library functions, like std::min_element (https://clang.godbolt.org/z/6czGzzqbs)
#include <vector>
#include <algorithm>
int foo(const std::vector<int> &V) {
return *std::min_element(V.begin(), V.end());
}
Reviewed By: reames
Differential Revision: https://reviews.llvm.org/D118143
Use type inference when building the TransferWriteOp in the TransferWritePermutationLowering. Previously, the result type has been set to Type() which triggers an assertion if the pattern is used with tensors instead of memrefs.
Reviewed By: springerm
Differential Revision: https://reviews.llvm.org/D118758
Whilst adding legal types <-> register classes for Streaming SVE in
D118561 I noticed the hasSVE predication block set operation actions for
opcodes that may not be legal in Streaming SVE. Move these operations to
the later hasSVE block which has loops over the same types.
Reviewed By: sdesmalen
Differential Revision: https://reviews.llvm.org/D118560
Previously the options category given to cl::HideUnrelatedOptions was
local to llvm-reduce.cpp and as a result only options declared in that
file were visible in the -help options listing. This was a bit
unfortunate since there were several useful options declared in other
files. This patch addresses that.
Differential Revision: https://reviews.llvm.org/D118682
This patch allows:
- fir.box type to be a member of tuple<> or fir.type<> types,
- tuple<> type to be a member of tuple<> type.
When a fir.box types are nested in tuple<> or fir.type<>, it is translated
to the struct type of a Fortran runtime descriptor, and not a
pointer to a descriptor. This is because the fir.box is owned by the tuple
or fir.type.
FIR type translation was also flattening nested tuple while lowering to LLVM
dialect types. There does not seem to be a deep reason for doing that
and doing it causes issues in fir.coordinate_of generated on such tuple
(a fir.coordinate_of getting tuple<B, C> in tuple<A, tuple<B, C>>
ended-up lowered to an LLVM GEP getting B).
Differential Revision: https://reviews.llvm.org/D118701
The new LEGALAVL node annotates that the AVL refers to packs of 64bit.
We use a two-stage lowering approach with LEGALAVL:
First, standard SDNodes are translated into illegal VVP layer nodes.
Regardless of source (VP or standard), all VVP nodes have a mask and AVL
parameter. The AVL parameter refers to the element position (just as in
VP intrinsics).
Second, we legalize the AVL usage in VVP layer nodes. If the element
size is < 64bit, the EVL parameter has to be adjusted to refer to packs
of 64bits. We wrap the legalized AVL in a LEGALAVL node to track this.
Reviewed By: kaz7
Differential Revision: https://reviews.llvm.org/D118321
Use the same mnemonics in the tests that are used in the AtomicLoadOp
pattern ($rd, $rr) but use RR1 instead of $operand. This matches similar
tests in load8.ll.
Differential Revision: https://reviews.llvm.org/D117991
This patch fixes the atomicrmw result value to be the value before the
operation instead of the value after the operation. This was a bug, left
as a FIXME in the code (see https://reviews.llvm.org/D97127).
From the LangRef:
> The contents of memory at the location specified by the <pointer>
> operand are atomically read, modified, and written back. The original
> value at the location is returned.
Doing this expansion early allows the register allocator to arrange
registers in such a way that commutable operations are simply swapped
around as needed, which results in shorter code while still being
correct.
Differential Revision: https://reviews.llvm.org/D117725
Lower the PAUSE statement to a runtime call.
This patch is part of the upstreaming effort from fir-dev branch.
Reviewed By: kiranchandramohan, schweitz
Differential Revision: https://reviews.llvm.org/D118699
Co-authored-by: Eric Schweitz <eschweitz@nvidia.com>
Co-authored-by: Jean Perier <jperier@nvidia.com>
This updates all the non-runtime project release notes to use the
version number from CMake instead of the hard-coded version numbers
in conf.py.
It also hides warnings about pre-releases when the git suffix
is dropped from the LLVM version in CMake.
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D112181
Symbol versioning can prevent unintented install-time conflicts
between different llvm versions. Users may need to override this
for particular products (e.g. Julia), but this requires carrying
a source code patch. This patch moves this ability to a
configuration option. NFC for existing usage.
Differential Revision: https://reviews.llvm.org/D118672
This diff adds support for ADRP+ADD optimization for AArch64 described in
d2ca58c54b
i.e. under appropriate constraints
ADRP x0, symbol
ADD x0, x0, :lo12: symbol
can be turned into
NOP
ADR x0, symbol
Test plan: make check-all
Differential revision: https://reviews.llvm.org/D117614
Based on the output of include-what-you-use.
This is a big chunk of changes. It is very likely to break downstream code
unless they took a lot of care in avoiding hidden ehader dependencies, something
the LLVM codebase doesn't do that well :-/
I've tried to summarize the biggest change below:
- llvm/include/llvm-c/Core.h: no longer includes llvm-c/ErrorHandling.h
- llvm/IR/DIBuilder.h no longer includes llvm/IR/DebugInfo.h
- llvm/IR/IRBuilder.h no longer includes llvm/IR/IntrinsicInst.h
- llvm/IR/LLVMRemarkStreamer.h no longer includes llvm/Support/ToolOutputFile.h
- llvm/IR/LegacyPassManager.h no longer include llvm/Pass.h
- llvm/IR/Type.h no longer includes llvm/ADT/SmallPtrSet.h
- llvm/IR/PassManager.h no longer includes llvm/Pass.h nor llvm/Support/Debug.h
And the usual count of preprocessed lines:
$ clang++ -E -Iinclude -I../llvm/include ../llvm/lib/IR/*.cpp -std=c++14 -fno-rtti -fno-exceptions | wc -l
before: 6400831
after: 6189948
200k lines less to process is no that bad ;-)
Discourse thread on the topic: https://llvm.discourse.group/t/include-what-you-use-include-cleanup
Differential Revision: https://reviews.llvm.org/D118652
Previously, all children would be checked to see if any were an
explicit Register. If anywhere no commutable patterns would be
generated. This patch loosens the restriction to only check the
children that are being commuted.
Digging back through history, this code predates the existence of
commutable intrinsics and commutable SDNodes with more than 2
operands. At that time the loop would count the number of children that
weren't registers and if that was equal to 2 it would allow commuting.
I don't think this loop was re-considered when commutable
intrinsics were added or when we allowed SDNodes with more than 2
operands.
This important for RISCV were our isel patterns have a V0 mask
operand after the commutable operands on some RISCVISD opcodes.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D117955
Some FormatElement subclasses contain `std::vector`. Since these use
BumpPtrAllocator, they need to be converted to trailing objects.
However, this is not a trivial fix so I will leave it as a FIXME and use
a workaround.
The commit adds a unit test that uses the facilities of libLLVMCore
without adding it to link components. This causes failures with
the shared libraries builds.
This patch just adds the missing library to the link step.
Jeremy Morse