This patch makes vector spills valid for tail predication when all loads
from the same stack slot are within the loop
Differential Revision: https://reviews.llvm.org/D105443
This patch implements the `__popcntb` XL compatibility builtin for 32bit in the frontend and backend. This patch also updates tests for `__popcntb` and other XL Compat sync related builtins.
Reviewed By: #powerpc, nemanjai, amyk
Differential Revision: https://reviews.llvm.org/D105360
The new GET_CURRENT_PC() can lead to spurious top inlined internal frames.
Here are 2 examples from bots, in both cases the malloc is supposed to be
the top frame (#0):
WARNING: ThreadSanitizer: signal-unsafe call inside of a signal
#0 __sanitizer::StackTrace::GetNextInstructionPc(unsigned long)
#1 malloc
Location is heap block of size 99 at 0xbe3800003800 allocated by thread T1:
#0 __sanitizer::StackTrace::GetNextInstructionPc(unsigned long)
#1 malloc
Let's strip these internal top frames from reports.
With other code changes I also observed some top frames
from __tsan::ScopedInterceptor, proactively remove these as well.
Differential Revision: https://reviews.llvm.org/D106081
This is split from D105216, it handles only a subset of the cases in that patch.
Specifically, the issue being fixed is that the code incorrectly assumed that (Start-Stide) < End implied that the backedge was taken at least once. This is not true when e.g. Start = 4, Stride = 2, and End = 3. Note that we often do produce the right backedge taken count despite the flawed reasoning.
The fix chosen here is to use an alternate form of uceil (ceiling of unsigned divide) lowering which is safe when max(RHS,Start) > Start - Stride. (Note that signedness of both max expression and comparison depend on the signedness of the comparison being analyzed, and that overflow in the Start - Stride expression is allowed.) Note that this is weaker than proving the backedge is taken because it allows start - stride < end < start. Some cases which can't be proven safe are sent down the generic path, and we do end up generating less optimal expressions in a few cases.
Credit for coming up with the approach goes entirely to Eli. I just split it off, tweaked the comments a bit, and did some additional testing.
Differential Revision: https://reviews.llvm.org/D105942
Details:
Switch all #includes to use <> because that is consistent with what happens in the cmake checks.
Otherwise, we could be in the situation where cmake checks see that headers exist at <perfmon/...>
but in llvm-exegesis code, we use "perfmon/...", which may not exist.
Related PR/revisions: D84076, PR51017+D105615
Differential Revision: https://reviews.llvm.org/D105861
This adds support for the lld-only `--thinlto-cache-policy` option, as well as
implementations for ld64's `-cache_path_lto`, `-prune_interval_lto`,
`-prune_after_lto`, and `-max_relative_cache_size_lto`.
Test is adapted from lld/test/ELF/lto/cache.ll
Differential Revision: https://reviews.llvm.org/D105922
The documentation and help messages have recommended the double-dash forms for
quite a while. Remove one-dash long options which are not recognized by GNU
style `getopt_long`.
`-arch` is kept as it is in the manpage of classic nm
https://keith.github.io/xcode-man-pages/nm.1.html
Note: the dyldinfo related options don't have a test.
Reviewed By: jhenderson
Differential Revision: https://reviews.llvm.org/D105948
Arbitrary shifts have some complications, but shift by invariants
(viz. tensor index exp only at left hand side) can be easily
handled with the conjunctive rule.
Reviewed By: gussmith23
Differential Revision: https://reviews.llvm.org/D106002
This patch makes sure that the base name of the temporary unparsed files
(generated by the `flang` bash script) are randomised and unique to a
particular invocation of the script. Otherwise, we cannot reliably run
the script in parallel.
Differential Revision: https://reviews.llvm.org/D106052
With the migration from linalg.copy to memref.copy, this pass
(which was there solely to handle the linalg.copy op) is no
longer required for the end-to-end path for sparse compilation.
Reviewed By: ftynse
Differential Revision: https://reviews.llvm.org/D106073
As we automate more and more things in the library, it becomes useful for
contributors to have a single target for running all the automation as
part of their workflow. This commit adds a new `libcxx-generate-files`
target that should re-generate all the auto-generated files in the library.
As a fly-by, I also revamped the documentation on Contributing to account
for this new target and present it as a bullet list of things to check
before committing. I also added a few things that are often overlooked
to that list, such as updating the synopsis and the status files.
Differential Revision: https://reviews.llvm.org/D106067
This implements the elementtype attribute specified in D105407. It
just adds the attribute and the specified verifier rules, but
doesn't yet make use of it anywhere.
Differential Revision: https://reviews.llvm.org/D106008
This adds an elementtype(<ty>) attribute, which can be used to
attach an element type to a pointer typed argument. It is similar
to byval/byref in purpose, but unlike those does not carry any
specific semantics by itself. However, certain intrinsics may
require it and interpret it in specific ways.
The in-tree use cases for this that I'm currently aware of are:
call ptr @llvm.preserve.array.access.index.p0.p0(ptr elementtype(%ty) %base, i32 %dim, i32 %index)
call ptr @llvm.preserve.struct.access.index.p0.p0(ptr elementtype(%ty) %base, i32 %gep_index, i32 %di_index)
call token @llvm.experimental.gc.statepoint.p0(i64 0, i32 0, ptr elementtype(void ()) @foo, i32 0, i32 0, i32 0, i32 0, ptr addrspace(1) %obj)
Notably, the gc.statepoint case needs a function as element type,
in which case the workaround of adding a separate %ty undef
argument would not work, as arguments cannot be unsized.
Differential Revision: https://reviews.llvm.org/D105407
Fixes some regressions with -fstrict-vtable-pointers in llvm-test-suite.
Reviewed By: lebedev.ri
Differential Revision: https://reviews.llvm.org/D106017
This change enables vectorization of multiple exit loops when the exit count is statically computable. That requirement - shared with the rest of LV - in turn requires each exit to be analyzeable and to dominate the latch.
The majority of work to support this was done in a set of previous patches. In particular,, 72314466 avoids having multiple edges from the middle block to the exits, and 4b33b2387 which added support for non-latch single exit and multiple exits with a single exiting block. As a result, this change is basically just removing a bailout and adjusting some tests now that the prerequisite work is done and has stuck in tree for a bit.
Differential Revision: https://reviews.llvm.org/D105817
Continuing on from D105780, this should be the last major bit of
attribute cleanup. Currently, LLParser implements attribute parsing
for functions, parameters and returns separately, enumerating all
supported (and unsupported) attributes each time. This patch
extracts the common parsing logic, and performs a check afterwards
whether the attribute is valid in the given position. Parameters
and returns are handled together, while function attributes need
slightly different logic to support attribute groups.
Differential Revision: https://reviews.llvm.org/D105938
We obtain the current PC is all interceptors and collectively
common interceptor code contributes to overall slowdown
(in particular cheaper str/mem* functions).
The current way to obtain the current PC involves:
4493e1: e8 3a f3 fe ff callq 438720 <_ZN11__sanitizer10StackTrace12GetCurrentPcEv>
4493e9: 48 89 c6 mov %rax,%rsi
and the called function is:
uptr StackTrace::GetCurrentPc() {
438720: 48 8b 04 24 mov (%rsp),%rax
438724: c3 retq
The new way uses address of a local label and involves just:
44a888: 48 8d 35 fa ff ff ff lea -0x6(%rip),%rsi
I am not switching all uses of StackTrace::GetCurrentPc to GET_CURRENT_PC
because it may lead some differences in produced reports and break tests.
The difference comes from the fact that currently we have PC pointing
to the CALL instruction, but the new way does not yield any code on its own
so the PC points to a random instruction in the function and symbolizing
that instruction can produce additional inlined frames (if the random
instruction happen to relate to some inlined function).
Reviewed By: melver
Differential Revision: https://reviews.llvm.org/D106046
This patch is in a series of patches to provide builtins for compatibility
with the XL compiler. This patch adds the builtins and emit target independent
code for rotate related operations.
Reviewed By: nemanjai, #powerpc
Differential revision: https://reviews.llvm.org/D104744
Similar to the folds performed in InstCombinerImpl::foldSelectOpOp, this attempts to push a select further up to help merge a pair of binops.
I'm primarily interested in select(cond,add(x,y),add(x,z)) folds to help expose pointer math (see https://bugs.llvm.org/show_bug.cgi?id=51069 etc.) but I've tried to use the more generic isBinOp().
Differential Revision: https://reviews.llvm.org/D106058
../../git/llvm-project/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp:2395:17: warning: 'clang::ento::ProgramStateRef {anonymous}::RangeConstraintManager::setRange(clang::ento::ProgramStateRef, {anonymous}::EquivalenceClass, clang::ento::RangeSet)' defined but not used [-Wunused-function]
../../git/llvm-project/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp:2384:10: warning: 'clang::ento::RangeSet {anonymous}::RangeConstraintManager::getRange(clang::ento::ProgramStateRef, {anonymous}::EquivalenceClass)' defined but not used [-Wunused-function]
Differential Revision: https://reviews.llvm.org/D106063
LLVM changed to not emit L... labels for things marked "do_not_dead_strip"
because the linker can sometimes drop the flag if there's no proper symbol.
This Clang test checked for the old behaviour, but doesn't actually care about
that bit.
This breaks out some (more) common llvm-specific
variables. Controlling the subprojects and target architectures, along
with clues about restricting build parallelism when linking. 'more
common' is somewhat subjective, of course.
Differential Revision: https://reviews.llvm.org/D105822
We have a DAG combine for recognizing the sequence of nodes that make up
an MVE VQDMULH, but only currently handles specifically legal types.
This patch expands that to other power-2 vector types. For smaller than
legal types this means any_extending the type and casting it to a legal
type, using a VQDMULH where we only use some of the lanes. The result is
sign extended back to the original type, to properly set the invalid
lanes. Larger than legal types are split into chunks with extracts and
concat back together.
Differential Revision: https://reviews.llvm.org/D105814
The linker can sometimes drop the do_not_dead_strip if it can't associate the
atom with a symbol (the other place to specify no dead-stripping in MachO
files).
`SinkCommonCodeFromPredecessors()` doesn't itself ensure that duplicate PHI nodes aren't created.
I suppose, we could teach it to do that on-the-fly (& account for the already-existing PHI nodes,
& adjust costmodel), the diff will be bigger than this.
The alternative is to schedule a new EarlyCSE pass invocation somewhere later in the pipeline.
Clearly, we don't have any EarlyCSE runs in module optimization passline, so this pattern isn't cleaned up...
That would perhaps better, but it will again have some compile time impact.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D106010
The underlying getMinVectorRegisterBitWidth() methods are const, but it was missed in a couple of TargetTransformInfo wrappers.
Noticed while working on D103925
The sort function for emitting an OptRemark was not deterministic,
which caused scalable-call.ll to fail on some buildbots. This patch
fixes that.
This patch also fixes an issue where `Instruction::comesBefore()`
is called when two Instructions are in different basic blocks,
which would otherwise cause an assertion failure.
Instead of relying on adhoc bounds calculations, use a projection-based
implementation. This simplifies the implementation and finds more static
constant sizes than previously/
Differential Revision: https://reviews.llvm.org/D106054
Fangrui Song