Summary: A count profile may affect tail duplication's heuristic causing a block to be duplicated in only a part of its predecessors. This is not allowed in the Machine Block Placement pass where an assert will go off. I'm removing the assert and making the optimization bail out when such case happens.
Reviewers: wenlei, davidxl, Carrot
Reviewed By: Carrot
Subscribers: hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77748
Summary:
If we don't read the flag, we don't need to set it. Then we also don't
need to filter it out to get the pointer.
This should have a (small) positive impact on code size and performance.
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D76259
Summary:
This revision performs a few refactorings on the main docs folder. Namely it:
* Adds a new Rationale/ folder to contain various rationale documents
* Moves several "getting started" documents to the Tutorials/ folder
* Cleans up the titles of various documents
Differential Revision: https://reviews.llvm.org/D77934
Selection would fail after the post legalize combiner put an illegal
zextload back together.
The base combiner has parameter to only allow legal operations, but
they appear to not be used. I also don't see a nice way to remove a
single entry from all_combines, so just hack around this.
Summary:
Fuchsia's gcc uses this, which in turn prevents us to compile successfully
due to a few `memset`'ing some non-trivial classes in some `init`.
Change those `memset` to members initialization.
Reviewers: pcc, hctim
Subscribers: #sanitizers, llvm-commits
Tags: #sanitizers
Differential Revision: https://reviews.llvm.org/D77902
OperatioFolder::tryToFold performs both true folding and in a few
instances in-place updates through op rewrites. In the latter case, we
should still be applying the supplied pattern rewrites in the same
iteration; however this wasn't the case since tryToFold returned
success() for both true folding and in-place updates, and the patterns
for the in-place updated ops were being applied only in the next
iteration of the driver's outer loop. This fix would make it converge
faster.
Differential Revision: https://reviews.llvm.org/D77485
As proposed in D77881, we'll have the related widening operation,
so this name becomes too vague.
While here, change the function signature to take an 'int' rather
than 'size_t' for the scaling factor, add an assert for overflow of
32-bits, and improve the documentation comments.
- Move Adapters array to the stack, we know the size precisely
- Parse format string on demand into a SmallVector. In theory this could
lead to parsing it multiple times, but I couldn't find a single instance
of that in LLVM.
- Make more of the implementation details private.
Default visibility for classes is private, so the private: at the top of
various class definitions is redundant.
Reviewers: gilr, rengolin, Ayal, hsaito
Reviewed By: Ayal
Differential Revision: https://reviews.llvm.org/D77810
For targets where char is unsigned (like PowerPC), something like
char c = fgetc(...) will never produce a char that will compare
equal to EOF so this loop does not terminate.
Change the type to int (which appears to be the POSIX return type
for fgetc).
This allows the test case to terminate normally on PPC.
...onization
Summary: In previous patch, in order to optimize performance, we only synchronize once
for each target region. The syncrhonization is via stream synchronization.
However, in the extreme situation, the performce might be bad. Consider the
following case: There is a task that requires transferring huge amount of data
(call many times of data transferring function). It is scheduled to the first
stream. And then we have 255 very light tasks scheduled to the remaining 255
streams (by default we have 256 streams). They can be finished before we do
synchronization at the end of the first task. Next, we get another very huge
task. It will be scheduled again to the first stream. Now the first task
finishes its kernel launch and call stream synchronization. Right now, the
stream already contains two kernels, and the synchronization will wait until the
two kernels finish instead of just the first one for the first task.
In this patch, we introduce stream pool. After each synchronization, the stream
will be returned back to the pool to make sure that for each synchronization,
only expected operations are waited.
Reviewers: jdoerfert
Reviewed By: jdoerfert
Subscribers: gregrodgers, yaxunl, lildmh, guansong, openmp-commits
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D77412
Buildbots say:
[126/127] Running lint check for sanitizer sources...
FAILED: projects/compiler-rt/lib/CMakeFiles/SanitizerLintCheck
cd /home/buildbots/ppc64be-clang-multistage-test/clang-ppc64be-multistage/stage1/projects/compiler-rt/lib && env LLVM_CHECKOUT=/home/buildbots/ppc64be-clang-multistage-test/clang-ppc64be-multistage/llvm/llvm SILENT=1 TMPDIR= PYTHON_EXECUTABLE=/usr/bin/python COMPILER_RT=/home/buildbots/ppc64be-clang-multistage-test/clang-ppc64be-multistage/llvm/compiler-rt /home/buildbots/ppc64be-clang-multistage-test/clang-ppc64be-multistage/llvm/compiler-rt/lib/sanitizer_common/scripts/check_lint.sh
/home/buildbots/ppc64be-clang-multistage-test/clang-ppc64be-multistage/llvm/compiler-rt/test/tsan/fiber_cleanup.cpp:71: Could not find a newline character at the end of the file. [whitespace/ending_newline] [5]
ninja: build stopped: subcommand failed.
Somehow this check is not part of 'ninja check-tsan'.
When creating and destroying fibers in tsan a thread state is created and destroyed. Currently, a memory mapping is leaked with each fiber (in __tsan_destroy_fiber). This causes applications with many short running fibers to crash or hang because of linux vm.max_map_count.
The root of this is that ThreadState holds a pointer to ThreadSignalContext for handling signals. The initialization and destruction of it is tied to platform specific events in tsan_interceptors_posix and missed when destroying a fiber (specifically, SigCtx is used to lazily create the ThreadSignalContext in tsan_interceptors_posix). This patch cleans up the memory by makinh the ThreadState create and destroy the ThreadSignalContext.
The relevant code causing the leak with fibers is the fiber destruction:
void FiberDestroy(ThreadState *thr, uptr pc, ThreadState *fiber) {
FiberSwitchImpl(thr, fiber);
ThreadFinish(fiber);
FiberSwitchImpl(fiber, thr);
internal_free(fiber);
}
Author: Florian
Reviewed-in: https://reviews.llvm.org/D76073
Summary: ClassID is a bit janky right now as it involves passing a magic pointer around. This revision hides the internal implementation mechanism within a new class TypeID. This class is a value-typed wrapper around the original ClassID implementation.
Differential Revision: https://reviews.llvm.org/D77768
This is the same as what was done to the CallLoweringInfo in
TargetLowering.h in r309159.
This is just a step on the way to replacing this with CallBase.
Summary: This hook allows for passes to specify the command line argument without the need for registration. More concretely this will allow for generating pass crash reproducers without needing to have the passes registered. This should remove the need for production tools to register passes, leaving that solely to development tools like mlir-opt.
Differential Revision: https://reviews.llvm.org/D77907
Summary: We allow non-relaxable instructions emitted into relaxable Fragment when we prefix padding branch. So we need to check if the instruction need relaxation before relaxing it. Without this patch, it currently triggers a `report_fatal_error` in `llvm::MCAsmBackend::relaxInstruction` when we prefix padding branch along with `--mc-relax-all`.
Reviewers: LuoYuanke, reames, MaskRay
Reviewed By: MaskRay
Subscribers: MaskRay, hiraditya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D77851
I only left it at the interface to ParseConstraints since that
needs updates to other callers in different files. I'll do that
as a follow up.
Differential Revision: https://reviews.llvm.org/D77892
There was another issue introduced by this commit that the OP
initially missed. Namely, for functions that are free to use
R2 as a callee-saved register, we emit a TOC expression based
on the address of the GEP label without emitting the GEP label.
Since we only emit such expressions for the large code model, this
issue only surfaced there.
I have confirmed that with this fix, the kernel build is successful
with target "all".
Matt Arsenault