The intrinsic has an immarg so its gets created with a TargetConstant
instead of a Constant after r372338. The isel pattern was only
updated for the register form, but not the memory form.
llvm-svn: 372457
APInt comparison require both to have the same bitwidth. Since only the value
is needed, use the compare function APInt::isSameValue instead.
llvm-svn: 372454
-Wtautological-overlap-compare and self-comparison from -Wtautological-compare
relay on detecting the same operand in different locations. Previously, each
warning had it's own operand checker. Now, both are merged together into
one function that each can call. The function also now looks through member
access and array accesses.
Differential Revision: https://reviews.llvm.org/D66045
llvm-svn: 372453
I believe all of the uniform/divergent pattern predicates are
redundant and can be removed. The uniformity bit already influences
the register class, and nothhing has broken when I've removed this and
others.
llvm-svn: 372450
Allow this warning to detect a larger number of constant values, including
negative numbers, and handle non-int types better.
Differential Revision: https://reviews.llvm.org/D66044
llvm-svn: 372448
ELF files generated for X86_64 targets may contain 64-bit PC-relative
relocations. For instance, an exception handler table entry contains the start
of exception-throwing frame relative to the start of exception handler. As these
two labels belong to different sections, their difference and so the relocation
is 64-bit.
An attempt to parse such file, i.e. in DWARFContext::create, results in "failed
to compute relocation" error.
This fix adds support for such relocations to RelocationResolver.cpp.
Reviewed By: MaskRay
Differential Revision: https://reviews.llvm.org/D67779
Patch by Oleg Pliss (Oleg.Pliss@azul.com)
llvm-svn: 372447
Summary:
Apple documentation states that:
"If two objects are equal, they must have the same hash value. This last
point is particularly important if you define isEqual: in a subclass and
intend to put instances of that subclass into a collection. Make sure
you also define hash in your subclass."
https://developer.apple.com/documentation/objectivec/1418956-nsobject/1418795-isequal?language=objc
In many or all versions of libobjc, -[NSObject isEqual:] is a pointer
equality check and -[NSObject hash] returns the messaged object's
pointer. A relatively common form of developer error is for a developer to
override -isEqual: in a subclass without overriding -hash to ensure that
hashes are equal for objects that are equal.
It is assumed that an override of -isEqual: is a strong signal for
changing the object's equality operator to something other than pointer
equality which implies that a missing override of -hash could result in
distinct objects being equal but having distinct hashes because they are
independent instances. This added check flags classes that override
-isEqual: but inherit NSObject's implementation of -hash to warn of the
potential for unexpected behavior.
The proper implementation of -hash is the responsibility of the
developer and the check will only verify that the developer made an
effort to properly implement -hash. Developers can set up unit tests
to verify that their implementation of -hash is appropriate.
Test Notes:
Ran check-clang-tools.
Reviewers: aaron.ballman, benhamilton
Reviewed By: aaron.ballman
Subscribers: Eugene.Zelenko, mgorny, xazax.hun, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67737
llvm-svn: 372445
We currently always set the HasCalls on MFI during translation and legalization if
we're handling a call or legalizing to a libcall. However, if that call is later
optimized to a tail call then we don't need the flag. The flag being set to true
causes frame lowering to always save and restore FP/LR, which adds unnecessary code.
This change does the same thing as SelectionDAG and ports over some code that scans
instructions after selection, using TargetInstrInfo to determine if target opcodes
are known calls.
Code size geomean improvements on CTMark:
-O0 : 0.1%
-Os : 0.3%
Differential Revision: https://reviews.llvm.org/D67868
llvm-svn: 372443
When swig fails, all the errors are squished onto one line with \n
quoting. It's very hard to read. This will print them out in a more
reasonable format.
Patch by: Lawrence D'Anna
Differential revision: https://reviews.llvm.org/D67790
llvm-svn: 372442
dotest -f does not work on Python3.
The name types.UnboundMethodType was an alias for types.MethodType in
2.7, but it does not exist in python3. MethodType works in both.
Also the actual type returned from SomeClass.some_method in python3
will be types.Function, not MethodType.
Patch by: Lawrence D'Anna
Differential revision: https://reviews.llvm.org/D67791
llvm-svn: 372441
Summary:
The CallAnalyzer::visitSwitchInst has an early exit when the estimated
lower bound of the switch cost will put the overall cost of the inline
above the threshold. However, this code is not correctly estimating the
lower bound for switches that can be transformed into bit tests, leading
to unnecessary lost inlines, and also differing behavior with
optimization remarks enabled.
First, the early exit is controlled by whether ComputeFullInlineCost is
enabled or not, and that in turn is disabled by default but enabled when
enabling -pass-remarks=missed. This by itself wouldn't lead to a
problem, except that as described below, the lower bound can be above
the real lower bound, so we can sometimes get different inline decisions
with inline remarks enabled, which is problematic.
The early exit was added in along with a new switch cost model in D31085.
The reason why this early exit was added is due to a concern one reviewer
raised about compile time for large switches:
https://reviews.llvm.org/D31085?id=94559#inline-276200
However, the code just below there calls
getEstimatedNumberOfCaseClusters, which in turn immediately calls
BasicTTIImpl getEstimatedNumberOfCaseClusters, which in the worst case
does a linear scan of the cases to get the high and low values. The
bit test handling in particular is guarded by whether the number of
cases fits into the max bit width. There is no suggestion that anyone
measured a compile time issue, it appears to be theoretical.
The problem is that the reviewer's comment about the lower bound
calculation is incorrect, specifically in the case of a switch that can
be lowered to a bit test. This isn't followed up on the comment
thread, but the author does add a FIXME to that effect above the early
exit added when they subsequently revised the patch.
As a result, we were incorrectly early exiting and not inlining
functions with switch statements that would be lowered to bit tests in
cases where we were nearing the threshold. Combined with the fact that
this early exit was skipped with opt remarks enabled, this caused
different inlining decisions to be made when -pass-remarks=missed is
enabled to debug the missing inline.
Remove the early exit for the above reasons.
I also copied over an existing AArch64 inlining test to X86, and
adjusted the threshold so that the bit test inline only occurs with the
fix in this patch.
Reviewers: davidxl
Subscribers: eraman, kristof.beyls, haicheng, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67716
llvm-svn: 372440
of the profile for profile in ExtBinary format.
Sometimes we want to limit the size of the profile by stripping some functions
with low sample count or by stripping some function names with small text size
from profile symbol list. That requires the profile reader to have the
interfaces returning the size of a section or the size of total profile. The
patch add those interfaces.
At the same time, add some dump facility to show the size of each section.
llvm-svn: 372439
The recently announced IBM z15 processor implements the architecture
already supported as "arch13" in LLVM. This patch adds support for
"z15" as an alternate architecture name for arch13.
Corrsponding LLVM support was committed as rev. 372435.
llvm-svn: 372436
The recently announced IBM z15 processor implements the architecture
already supported as "arch13" in LLVM. This patch adds support for
"z15" as an alternate architecture name for arch13.
The patch also uses z15 in a number of places where we used arch13
as long as the official name was not yet announced.
llvm-svn: 372435
Adds a custom sidebar to LLVM docs. Sidebar includes links to How to submit a bug and FAQ topics, as well as a Show Source link and search box.
llvm-svn: 372432
The static analyzer is warning about potential null dereference, but we can use cast<ConstantInt> directly and if not assert will fire for us.
llvm-svn: 372429
Move the definition of Elf_Addr typedef to the only place it is used, to avoid:
```
llvm-project/libunwind/src/AddressSpace.hpp:501:28: warning: unused typedef 'Elf_Addr' [-Wunused-local-typedef]
```
when compiling for Android with _LIBUNWIND_ARM_EHABI defined and
_LIBUNWIND_SUPPORT_DWARF_UNWIND not defined.
Patch by Joel Klinghed!
llvm-svn: 372427
If one reverts D66398 then the TestDataFormatterStdList does fail - as the C++
formatters are initialized in the opposite order. But the current state of
trunk does not mind the order for C++ formatters.
It is using now a single std::vector as suggested by Pavel Labath.
Differential Revision: https://reviews.llvm.org/D66654
llvm-svn: 372424
RebuildAtomicExpr was skipping doing semantic analysis which broke in
the cases where the expressions were not dependent. This resulted in the
ImplicitCastExpr from an array to a pointer being lost, causing a crash
in IR CodeGen.
Differential Revision: https://reviews.llvm.org/D67854
llvm-svn: 372422
Summary:
Allow users to pass the path to a .dSYM directory to llvm-objdump's -dsym
flag rather than requiring users to find the DWARF DSYM Mach-O within the
bundle structure by hand.
rdar://46873333
Reviewers: pete, lhames, friss, aprantl
Reviewed By: pete, aprantl
Subscribers: MaskRay, aprantl, rupprecht, seiya, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67780
llvm-svn: 372421
We always(and only) check the NLP flag after calling
classifyGlobalReference to see whether it is accessed
indirectly.
Refactor to code to use isGVIndirectSym instead.
llvm-svn: 372417
Summary:
Adds two new combinators and corresponding tests to the RangeSelector library.
* `ifBound` -- conditional evaluation of range-selectors, based on whether a
given node id is bound in the match.
* `elseBranch` -- selects the source range of the else and its statement.
Reviewers: gribozavr
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D67621
llvm-svn: 372410
Summary:
After the switch in SimplifyDemandedBits, it tries to create a
constant when possible. If the original node is a TargetConstant
the default in the switch will call computeKnownBits on the
TargetConstant which will succeed. This results in the
TargetConstant becoming a Constant. But TargetConstant exists to
avoid being changed.
I've fixed the two cases that relied on this in tree by explicitly
making the nodes constant instead of target constant. The Sparc
case is an old bug. The Mips case was recently introduced now that
ImmArg on intrinsics gets turned into a TargetConstant when the
SelectionDAG is created. I've removed the ImmArg since it lowers
to generic code.
Reviewers: arsenm, RKSimon, spatel
Subscribers: jyknight, sdardis, wdng, arichardson, hiraditya, fedor.sergeev, jrtc27, atanasyan, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D67802
llvm-svn: 372409
This patch adds the patterns to select the dot product instructions.
Tested on aarch64-linux with make check-all.
Differential Revision: https://reviews.llvm.org/D67645
llvm-svn: 372408
If unwind info is not available at the current IP, unw_get_proc_info should
return a zero-filled structure rather than the info of the previous IP.
This change also makes unw_get_proc_info return UNW_ENOINFO instead of
UNW_ESUCCESS.
Patch by Amanieu d'Antras!
llvm-svn: 372407
According to the documentation method returns predecessor
if the given loop's header has exactly one unique predecessor
outside the loop. Otherwise return null.
In reality it asserts if there is no predecessor outside of
the loop.
The testcase has the loop where predecessors outside of the
loop were not identified as analyzeBranch() was unable to
process the mask branch and returned true. That is also not
correct to assert for the truly dead loops.
Differential Revision: https://reviews.llvm.org/D67634
llvm-svn: 372405
Craig Topper