Summary:
There's little of interest that can be done to an already-erased instruction.
You can't inspect it, write it to a debug log, etc. It ought to be notification
that we're about to erase it. Rename the function to clarify the timing of the
event and reflect current usage.
Also fixed one case where we were trying to print an erased instruction.
Reviewers: aditya_nandakumar
Reviewed By: aditya_nandakumar
Subscribers: rovka, kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D55611
llvm-svn: 348976
MULX has somewhat improved register allocation constraints compared to the legacy MUL instruction. Both output registers are encoded instead of fixed to EAX/EDX, but EDX is used as input. It also doesn't touch flags. Unfortunately, the encoding is longer.
Prefering it whenever BMI2 is enabled is probably not optimal. Choosing it should somehow be a function of register allocation constraints like converting adds to three address. gcc and icc definitely don't pick MULX by default. Not sure what if any rules they have for using it.
Differential Revision: https://reviews.llvm.org/D55565
llvm-svn: 348975
A future patch may stop using MULX by default so use MIR to ensure we're always testing MULX.
Add the 32-bit case that we couldn't do in the 64-bit mode IR test due to it being promoted to a 64-bit mul.
llvm-svn: 348972
Updated the annotate-kernel-features pass to support the propagation of uniform-work-group attribute from the kernel to the called functions. Once this pass is run, all kernels, even the ones which initially did not have the attribute, will be able to indicate weather or not they have uniform work group size depending on the value of the attribute.
Differential Revision: https://reviews.llvm.org/D50200
llvm-svn: 348971
Use the replacement execute once threading support in LLVM for PPC64le. It
seems that GCC does not define `__ppc__` and so we would actually call out to
the C++ runtime there which is not what the current code intended. Check both
`__ppc__` and `__PPC__`. This avoids the need for checking the endianness.
Thanks to nemanjai for the hint about GCC's behaviour and the fact that the
reviewed condition could be simplified.
Original patch by Sarvesh Tamba!
llvm-svn: 348970
The __builtin_unpredictable implementation is confused by any implicit
casts, which happen in C++. This patch strips those off so that
if/switch statements now work with it in C++.
Change-Id: I73c3bf4f1775cd906703880944f4fcdc29fffb0a
llvm-svn: 348969
Doesn't handle varargs and other fun things, but it's a start. (also
doesn't print these strictly as valid C++ when it's a pointer to
function, it'll print as "void(int)*" instead of "void (*)(int)")
llvm-svn: 348965
Continue to present HSA metadata as YAML in ASM and when output by tools
(e.g. llvm-readobj), but encode it in Messagepack in the code object.
Differential Revision: https://reviews.llvm.org/D48179
llvm-svn: 348963
This lays the foundation for dumping types not referenced by DW_AT_type
attributes (in the near-term, that'll be DW_AT_containing_type for a
DW_TAG_ptr_to_member_type - in the future, potentially dumping the
pretty printed name next to the DW_TAG for the type, rather than only
when the type is referenced from elsewhere)
llvm-svn: 348961
I'm hoping we can just replace SETCC_CARRY with SBB. This is another step towards that.
I've explicitly used zero as the input to the setcc to avoid a false dependency that we've had with the SETCC_CARRY. I changed one of the patterns that used NEG to instead use an explicit compare with 0 on the LHS. We needed the zero anyway to avoid the false dependency. The negate would clobber its input register. By using a CMP we can avoid that which could be useful.
Differential Revision: https://reviews.llvm.org/D55414
llvm-svn: 348959
Indices for getelementptr can be signed so we should use
getMinSignedBits instead of getActiveBits here. The function later calls
getSExtValue to get the int64_t value, which also checks
getMinSignedBits.
This fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=11647.
Reviewers: mssimpso, efriedma, davide
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D55536
llvm-svn: 348957
This patch introduces a generic function to determine whether a given vector type is known to be a splat value for the specified demanded elements, recursing up the DAG looking for BUILD_VECTOR or VECTOR_SHUFFLE splat patterns.
It also keeps track of the elements that are known to be UNDEF - it returns true if all the demanded elements are UNDEF (as this may be useful under some circumstances), so this needs to be handled by the caller.
A wrapper variant is also provided that doesn't take the DemandedElts or UndefElts arguments for cases where we just want to know if the SDValue is a splat or not (with/without UNDEFS).
I had hoped to completely remove the X86 local version of this function, but I'm seeing some regressions in shift/rotate codegen that will take a little longer to fix and I hope to get this in sooner so I can continue work on PR38243 which needs more capable splat detection.
Differential Revision: https://reviews.llvm.org/D55426
llvm-svn: 348953
If a module has function references, but no functions
themselves, we may end up never calling runOnMachineFunction
and therefore would never initialize nvptxSubtarget field
which would eventually cause a crash.
Instead of relying on nvptxSubtarget being initialized by
one of the methods, retrieve subtarget info directly.
Differential Revision: https://reviews.llvm.org/D55580
llvm-svn: 348952
CallGraph previously would just show the normal name of a function,
which gets really confusing when using it on large C++ projects. This
patch switches the printName call to a printQualifiedName, so that the
namespaces are included.
Change-Id: Ie086d863f6b2251be92109ea1b0946825b28b49a
llvm-svn: 348950
This extends the code that handles 16-bit add promotion to form LEA to also allow 8-bit adds.
That allows us to combine add ops with register moves and save some instructions. This is
another step towards allowing add truncation in generic DAGCombiner (see D54640).
Differential Revision: https://reviews.llvm.org/D55494
llvm-svn: 348946
Version.inc.in processing has a potentially interesting part which I've punted
on for now (LLD_REVISION and LLD_REPOSITORY are set to empty strings for now).
lld now builds in the gn build. But no symlinks to it are created yet, so it
can't be meaningfully run yet.
Differential Revision: https://reviews.llvm.org/D55593
llvm-svn: 348945
When multiple loop transformation are defined in a loop's metadata, their order of execution is defined by the order of their respective passes in the pass pipeline. For instance, e.g.
#pragma clang loop unroll_and_jam(enable)
#pragma clang loop distribute(enable)
is the same as
#pragma clang loop distribute(enable)
#pragma clang loop unroll_and_jam(enable)
and will try to loop-distribute before Unroll-And-Jam because the LoopDistribute pass is scheduled after UnrollAndJam pass. UnrollAndJamPass only supports one inner loop, i.e. it will necessarily fail after loop distribution. It is not possible to specify another execution order. Also,t the order of passes in the pipeline is subject to change between versions of LLVM, optimization options and which pass manager is used.
This patch adds 'followup' attributes to various loop transformation passes. These attributes define which attributes the resulting loop of a transformation should have. For instance,
!0 = !{!0, !1, !2}
!1 = !{!"llvm.loop.unroll_and_jam.enable"}
!2 = !{!"llvm.loop.unroll_and_jam.followup_inner", !3}
!3 = !{!"llvm.loop.distribute.enable"}
defines a loop ID (!0) to be unrolled-and-jammed (!1) and then the attribute !3 to be added to the jammed inner loop, which contains the instruction to distribute the inner loop.
Currently, in both pass managers, pass execution is in a fixed order and UnrollAndJamPass will not execute again after LoopDistribute. We hope to fix this in the future by allowing pass managers to run passes until a fixpoint is reached, use Polly to perform these transformations, or add a loop transformation pass which takes the order issue into account.
For mandatory/forced transformations (e.g. by having been declared by #pragma omp simd), the user must be notified when a transformation could not be performed. It is not possible that the responsible pass emits such a warning because the transformation might be 'hidden' in a followup attribute when it is executed, or it is not present in the pipeline at all. For this reason, this patche introduces a WarnMissedTransformations pass, to warn about orphaned transformations.
Since this changes the user-visible diagnostic message when a transformation is applied, two test cases in the clang repository need to be updated.
To ensure that no other transformation is executed before the intended one, the attribute `llvm.loop.disable_nonforced` can be added which should disable transformation heuristics before the intended transformation is applied. E.g. it would be surprising if a loop is distributed before a #pragma unroll_and_jam is applied.
With more supported code transformations (loop fusion, interchange, stripmining, offloading, etc.), transformations can be used as building blocks for more complex transformations (e.g. stripmining+stripmining+interchange -> tiling).
Reviewed By: hfinkel, dmgreen
Differential Revision: https://reviews.llvm.org/D49281
Differential Revision: https://reviews.llvm.org/D55288
llvm-svn: 348944
Clang's CallGraph analysis doesn't use the RecursiveASTVisitor's setting
togo into template instantiations. The result is that anything wanting
to do call graph analysis ends up missing any template function calls.
Change-Id: Ib4af44ed59f15d43f37af91622a203146a3c3189
llvm-svn: 348942
Previously CreateParameterDeclaration was always using the translation
unit DeclContext. We would later go and add parameters to the
FunctionDecl, but internally clang makes a copy when you do this, and
we'd end up with ParmVarDecl's at the global scope as well as in the
function scope.
This fixes the issue. It's hard to say whether this will introduce
a behavioral change in name lookup, but I know there have been several
hacks introduced in previous years to deal with collisions between
various types of variables, so there's a chance that this patch could
obviate one of those hacks.
Differential Revision: https://reviews.llvm.org/D55571
llvm-svn: 348941
For SampleFDO, when a callsite doesn't appear in the profile, it will not be marked as cold callsite unless the option -profile-sample-accurate is specified.
But profile-sample-accurate doesn't cover function isFunctionColdInCallGraph which is used to decide whether a function should be put into text.unlikely section, so even if the user knows the profile is accurate and specifies profile-sample-accurate, those functions not appearing in the sample profile are still not be put into text.unlikely section right now.
The patch fixes that.
Differential Revision: https://reviews.llvm.org/D55567
llvm-svn: 348940
The Darwin targets use `int64_t` and `uint64_t` to define the `int_least64_t`
and `int_fast64_t` types. The underlying type is actually a `long long`. Match
the types to allow the printf specifiers to work properly and have the compiler
vended macros match the implementation on the target.
llvm-svn: 348939
Summary:
`memchr` and `memcmp` operate upon the character units of the object
representation; that is, the `size_t` parameter expresses the number of
character units. The constant folding implementation is updated in this
patch to account for multibyte element types in the arrays passed to
`memchr`/`memcmp` and, in the case of `memcmp`, to account for the
possibility that the arrays may have differing element types (even when
they are byte-sized).
Actual inspection of the object representation is not implemented.
Comparisons are done only between elements with the same object size;
that is, `memchr` will fail when inspecting at least one character unit
of a multibyte element. The integer types are assumed to have two's
complement representation with 0 for `false`, 1 for `true`, and no
padding bits.
`memcmp` on multibyte elements will only be able to fold in cases where
enough elements are equal for the answer to be 0.
Various tests are added to guard against incorrect folding for cases
that miscompile on some system or other prior to this patch. At the same
time, the unsigned 32-bit `wchar_t` testing in
`test/SemaCXX/constexpr-string.cpp` is restored.
Reviewers: rsmith, aaron.ballman, hfinkel
Reviewed By: rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D55510
llvm-svn: 348938
I've extended the load/store optimizer to be able to produce dwordx3
loads and stores, This change allows many more load/stores to be combined,
and results in much more optimal code for our hardware.
Differential Revision: https://reviews.llvm.org/D54042
llvm-svn: 348937
Move code into a separate function, and replace the if-else chain with
llvm::StringSwitch.
A slight behavioral change is that now I use the section flags
(SHF_TLS) instead of the section name to set the thread-specific
property. There is no explanation in the original commit introducing
this (r153537) as to why that was done this way, but the new behavior
should be more correct.
llvm-svn: 348936
Summary:
Added support for the -gline-directives-only option + fixed logic of the
debug info for CUDA devices. If optimization level is O0, then options
--[no-]cuda-noopt-device-debug do not affect the debug info level. If
the optimization level is >O0, debug info options are used +
--no-cuda-noopt-device-debug is used or no --cuda-noopt-device-debug is
used, the optimization level for the device code is kept and the
emission of the debug directives is used.
If the opt level is > O0, debug info is requested +
--cuda-noopt-device-debug option is used, the optimization is disabled
for the device code + required debug info is emitted.
Reviewers: tra, echristo
Subscribers: aprantl, guansong, JDevlieghere, cfe-commits
Differential Revision: https://reviews.llvm.org/D51554
llvm-svn: 348930
This library was breaking my -DBUILD_SHARED_LIBS=1 build. rC348915 seemed to miss this case.
As this seems an "obvious" fix, I am committing without pre-commit review as
per the LLVM developer policy.
llvm-svn: 348929
Instead of GetProgramHeaderCount+GetProgramHeaderByIndex, expose an
ArrayRef of all program headers, to enable range-based iteration.
Instead of GetSegmentDataByIndex, expose GetSegmentData, taking a
program header (reference).
This makes the code simpler by enabling range-based loops and also
allowed to remove some null checks, as it became locally obvious that
some pointers can never be null.
llvm-svn: 348928
Address spaces are cast into generic before invoking the constructor.
Added support for a trailing Qualifiers object in FunctionProtoType.
Differential Revision: https://reviews.llvm.org/D54862
llvm-svn: 348927
If either of the operand elements are zero then we know the result element is going to be zero (even if the other element is undef).
Differential Revision: https://reviews.llvm.org/D55558
llvm-svn: 348926
Daniel Sanders