This adapts LV to the new semantics of get.active.lane.mask as discussed in
D86147, which means that the LV now emits intrinsic get.active.lane.mask with
the loop tripcount instead of the backedge-taken count as its second argument.
The motivation for this is described in D86147.
Differential Revision: https://reviews.llvm.org/D86304
Some functions also include a `.Lfunc$local:` label due to
-fno-semantic-interposition
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D85888
The arm backend does not handle select/select_cc on vectors with scalar
conditions, preferring to expand them in codegenprepare instead. This
usually works except when optimizing for size, where the optsize check
would end up overruling the backend isSelectSupported check.
We could handle the selects in ISel too, but this seems like smaller
code than trying to splat the condition to all lanes.
Differential Revision: https://reviews.llvm.org/D86433
Without the fix gcc 7.4 warns with
../lib/Target/PowerPC/PPCAsmPrinter.cpp: In member function 'void {anonymous}::PPCAsmPrinter::EmitTlsCall(const llvm::MachineInstr*, llvm::MCSymbolRefExpr::VariantKind)':
../lib/Target/PowerPC/PPCAsmPrinter.cpp:525:53: warning: enumeral and non-enumeral type in conditional expression [-Wextra]
MCInstBuilder(Subtarget->isPPC64() ? Opcode : PPC::BL_TLS)
~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~
Also updates isConstOrConstSplatFP to allow the mul(A,-1) -> neg(A)
transformation when -1 is expressed as an ISD::SPLAT_VECTOR.
Differential Revision: https://reviews.llvm.org/D86415
Currently, when a program header type is unknown, we dont print anything:
```
ProgramHeader {
Type: (0x60000000)
```
With this patch the output will be:
```
ProgramHeader {
Type: Unknown (0x60000000)
```
It was discussed in D85526 and consistent with what we print for
'--sections' already, e.g.:
```
Section {
Name: .sec
Type: Unknown (0x7FFFFFFF)
}
```
Differential revision: https://reviews.llvm.org/D86213
This allows to get rid of "Invalid data was encountered while parsing the file"
error reported in cases when sh_size/sh_offset of sections are broken.
Differential revision: https://reviews.llvm.org/D86451
Fixes PR46575.
Bump statistics version to 6.
Without this patch, for a variable described with a location list the stat
'sum_all_variables(#bytes in parent scope covered by DW_AT_location)' is
calculated by summing all bytes covered by the location ranges in the list and
capping the result to the number of bytes in the parent scope. With the patch,
only bytes which overlap with the parent DIE scope address ranges contribute to
the stat. A new stat 'sum_all_variables(#bytes in any scope covered by
DW_AT_location)' has been added which displays the total bytes covered when
ignoring scopes.
In getCastInstrCost when the instruction is a truncate we were relying
upon the implicit TypeSize -> uint64_t cast when asking if a given type
has the same size as a legal integer. I've changed the code to only
ask the question if the type is fixed length.
I have also changed InstCombinerImpl::SimplifyDemandedUseBits to bail
out for now if the type is a scalable vector.
I've added the following new tests:
Analysis/CostModel/AArch64/sve-trunc.ll
Transforms/InstCombine/AArch64/sve-trunc.ll
for both of these fixes.
Differential revision: https://reviews.llvm.org/D86432
Currently we repeatedly check the same uses for read clobbers in some
cases. We can avoid unnecessary checks by keeping track of the memory
accesses we already found read clobbers for. To do so, we just add
memory access causing read-clobbers to a set. Note that marking all
visited accesses as read-clobbers would be to pessimistic, as that might
include accesses not on any path to the actual read clobber.
If we do not find any read-clobbers, we can add all visited instructions
to another set and use that to skip the same accesses in the next call.
Reviewed By: asbirlea
Differential Revision: https://reviews.llvm.org/D75025
Explicitly check that there is a local def prior to the given
instruction in getReachingLocalMIDef instead of just relying on
a nullptr return from getInstFromId.
Support -march=sapphirerapids for x86.
Compare with Icelake Server, it includes 14 more new features. They are
amxtile, amxint8, amxbf16, avx512bf16, avx512vp2intersect, cldemote,
enqcmd, movdir64b, movdiri, ptwrite, serialize, shstk, tsxldtrk, waitpkg.
Reviewed By: craig.topper
Differential Revision: https://reviews.llvm.org/D86503
For the Windows GNU platform, CMAKE_FIND_LIBRARY_PREFIXES is a list
containing an empty string, which ended up in a regex capturing group,
which is invalid in CMake's regex engine. With this change, we get the
following:
set(CMAKE_FIND_LIBRARY_PREFIXES "lib" "")
set(CMAKE_FIND_LIBRARY_SUFFIXES ".dll.a" ".a" ".lib")
get_system_libname(path/to/libz.dll.a zlib)
message("${zlib}")
outputs z, as expected.
Patch By: haampie
Differential Revision: https://reviews.llvm.org/D86434
If we use training algorithms that don't need partial rewards, we don't
need to worry about an ir2native model. In that case, training logs
won't contain a 'delta_size' feature either (since that's the partial
reward).
Differential Revision: https://reviews.llvm.org/D86481
On Windows, 'env' or 'printenv' may not exist.
Also switch back to 'env' which is specified by POSIX.1-2017. 'printenv' is not
standard (I picked it because 'printenv' exists on GnuWin32 but 'env' does not).
Reviewed By: zequanwu
Differential Revision: https://reviews.llvm.org/D86496
With FMF ( "nsz" and " reassoc") fold X/Sqrt(X) to Sqrt(X).
This is done after targets have the chance to produce a
reciprocal sqrt estimate sequence because that expansion
is probably more efficient than an expansion of a
non-reciprocal sqrt. That is also why we deferred doing
this transform in IR (D85709).
Differential Revision: https://reviews.llvm.org/D86403
PC-Relative addressing introduces a fair bit of complexity for correctly
eliminating TOC accesses. FastISel does not include any of that handling so we
miscompile code with -mcpu=pwr10 -O0 if it includes an external call that
FastISel does not handle followed by any of the following:
Floating point constant materialization
Materialization of a GlobalValue
Call that FastISel does handle
This patch switches to SDISel for any of the above.
Differential revision: https://reviews.llvm.org/D86343
This is preparation for making clang default to -mtune=generic when no -march is specified. This will allow the default tuning to be "generic" even though our default march is "pentium4" or "x86-64".
To avoid llc lit test regressions, if no mcpu is specified, I've defaulted tune to use i586 to match the old tuning settings of no CPU. Some tests explicitly used -mcpu=generic which I've removed so they instead get this default of architecture features from generic and tune from i586.
I updated one llvm-mca test to check a different CPU since generic has a scheduler model now
Differential Revision: https://reviews.llvm.org/D86312
SelectionDAG and GlobalISel take different failure paths for these and
end up producing different failure errors. Check both so the test
passes when the default is switched.
InstCombine currently has odd rules for folding insert-extract chains to shuffles,
so we miss collapsing seemingly simple cases as shown in the tests here.
But poison makes this not quite as easy as we might have guessed. Alive2 tests to
show the subtle difference (similar to the regression tests):
https://alive2.llvm.org/ce/z/hp4hv3 (this is ok)
https://alive2.llvm.org/ce/z/ehEWaN (poison leakage)
SLP tends to create these patterns (as shown in the SLP tests), and this could
help with solving PR16739.
Differential Revision: https://reviews.llvm.org/D86460
It's full featured now and we can use it for the runtimes build instead
of relying on an external libtool, which means the CMAKE_HOST_APPLE
restriction serves no purpose either now. Restrict llvm-lipo to Darwin
targets while I'm here, since it's only needed there.
Reviewed By: phosek
Differential Revision: https://reviews.llvm.org/D86367
{builtin,runtime}_register_target passes a TOOLCHAIN_TOOLS list, whereas
{builtin,runtime}_default_target does notl. The explicit TOOLCHAIN_TOOLS
list matches what LLVMExternalProjectUtils would have set anyway,
barring some target-specific adjustments, and those target-specific
adjustments seem valuable, so let's drop the explicit TOOLCHAIN_TOOLS
list and let LLVMExternalProjectUtils take care of it.
Reviewed By: phosek
Differential Revision: https://reviews.llvm.org/D86366
The -V option in cctools' libtool prints out the version number and
performs any specified operation. Add this option to LLVM's version.
cctools is more forgiving of invalid command lines when -V is specified,
but I think it's better to give errors instead of silently producing no
output.
Unfortunately, when -V is present, options that would otherwise be
required aren't anymore, so we need to perform some manual argument
validation.
Reviewed By: alexshap
Differential Revision: https://reviews.llvm.org/D86359
The "takeName" logic at the end of ScalarizerVisitor::finish
could end up renaming global variables when having simplified
and extractelement instruction to simply pick a single vector
element. If the input vector to the extractelement instruction
held pointers to global variables we ended up renaming the global
variable.
The patch make sure we only take the name of the replaced Op when
we have added new instructions that might need a useful name.
Reviewed By: lebedev.ri
Differential Revision: https://reviews.llvm.org/D86472
Current custom lowering of truncate vector handles a source of up to 128 bits, but that only uses one of the two shuffle vector operands. Extend it to use both operands to handle 256 bit sources.
Differential Revision: https://reviews.llvm.org/D68035
This reverts commit 2e43acfed89b1903de473f682c65878bdebc395a.
LLVMCoroutines (the library which contains Coroutines.h) depends on LLVMipo (the
library which contains SampleProfile.cpp). It is inappropriate for
SampleProfile.cpp to depent on Coroutines.h (circular dependency).
The test inverted dependencies as well:
llvm/test/Transforms/Coroutines/coro-inline.ll uses -sample-profile.
This interferes with GlobalISel's much better handling of the
situation.
This should really be disable for GlobalISel. However, the fallback
only re-runs the selection passes, and doesn't go back and rerun any
codegen IR passes. I haven't come up with a good solution to this
problem.
This is to initially handleg immAllOnesV, which should match
G_BUILD_VECTOR or G_BUILD_VECTOR_TRUNC. In the future, it could be
used for other patterns cases that map to multiple G_* instructions,
such as G_ADD and G_PTR_ADD.
D77152 tried to do this but got it wrong in the shift-by-zero case.
D86430 reverted the wrong code. Reimplement the optimization with
different code depending on whether the shift amount is known to be
non-zero (modulo bitwidth).
This improves code quality for fshl tests on AMDGPU, which only has an
fshr instruction.
Differential Revision: https://reviews.llvm.org/D86438
