Now, that we have funnel shift intrinsics, it should be safe to convert this form of rotate to it.
In the worst case (a target that doesn't have rotate instructions), we will expand this into a
branch-less sequence of ALU ops (neg/and/and/lshr/shl/or) in the backend, so it's still very
likely to be a perf improvement over the original code.
The motivating source code pattern for this is shown in:
https://bugs.llvm.org/show_bug.cgi?id=34924
Background:
I looked at several different options before deciding where to try this - instcombine, simplifycfg,
CGP - because it doesn't fit cleanly anywhere AFAIK.
The backend (CGP, SDAG, GlobalIsel?) is too late for what we're trying to accomplish. We want to
have the IR converted before we reach things like vectorization because the reduced code can make a
loop much simpler to transform.
Technically, this could be included in instcombine, but it's a large pattern match that includes
control-flow, so it just felt wrong to stuff into there (although I have a draft of that patch).
Similarly, this could be part of simplifycfg, but all of this pattern matching is a stretch.
So we're left with our relatively new dumping ground for homeless transforms: aggressive-instcombine.
This only runs at -O3, but that seems like a reasonable limitation given that source code has many
options to avoid this pattern (including the recently added clang intrinsics for rotates).
I'm including a PhaseOrdering test because we require the teamwork of 3 passes (aggressive-instcombine,
instcombine, simplifycfg) to get this into the minimal IR form that we want. That test shows a bug
with the new pass manager that's independent of this change (but it will be masked if we canonicalize
harder to funnel shift intrinsics in instcombine).
Differential Revision: https://reviews.llvm.org/D55604
llvm-svn: 349396
The test test/Driver/darwin-sdk-version.c from r349380 checks if the macOS
deployment target can be correctly inferred from the SDK version. When the
SDK version is > host version, the driver will pick the host version, so
the old test failed on macOS < 10.14. This commit makes this test more
resilient by using an older SDK version.
llvm-svn: 349393
The assertion type is always supposed to be a scalar type. So if the result VT of the assertion is a vector, we need to get the scalar VT before we can compare them.
Similarly for the assert above it.
I don't have a test case because I don't know of any place we violate this today. A coworker found this while trying to use r347287 on the 6.0 branch without also having r336868
llvm-svn: 349390
The problem is shown specifically for a case with vector multiply here:
https://bugs.llvm.org/show_bug.cgi?id=40032
...and this might mask the original backend bug for ARM shown in:
https://bugs.llvm.org/show_bug.cgi?id=39967
As the test diffs here show, we were (and probably still aren't) doing
these kinds of transforms in a principled way. We are producing more or
equal wide instructions than we started with in some cases, so we still
need to restrict/correct other transforms from overstepping.
If there are perf regressions from this change, we can either carve out
exceptions to the general IR rules, or improve the backend to do these
transforms when we know the transform is profitable. That's probably
similar to a change like D55448.
Differential Revision: https://reviews.llvm.org/D55744
llvm-svn: 349389
The code here wants the output section offset of the instruction
requiring the errata patch, not the virtual address. Without this
change we can end up placing a patch out of range if the virtual
address of the code section is large enough.
Differential Revision: https://reviews.llvm.org/D55732
llvm-svn: 349386
This allows a TEST to be used and can be combined with any AND that may already exist as an input to the shift.
This was already done in EmitTest, but was easily tricked by multiple uses because the setcc might be used by multiple instructions. Once the SETCC and users are legalized then we can look for the shift to be used by a single CMP, but the CMP itself can have multiple users.
This appears to fix the case in PR39968.
llvm-svn: 349385
Previously the code that parsed debug info to create lldb's Symbol
objects such as Variable, Type, Function, etc was tightly coupled
to the AST reconstruction code. This made it difficult / impossible
to implement functions such as ParseDeclsForContext() that were only
supposed to be operating on clang AST's. By splitting these apart,
the logic becomes much cleaner and we have a clear separation of
responsibilities.
llvm-svn: 349383
is not specified
The -target option allows the user to specify the build target using LLVM
triple. The triple includes the arch, and so the -arch option is redundant.
This should work just as well without the -arch. However, the driver has a bug
in which it doesn't target the "Cyclone" CPU for darwin if -target is used
without -arch. This commit fixes this issue.
rdar://46743182
Differential Revision: https://reviews.llvm.org/D55731
llvm-svn: 349382
On Darwin, using '-arch x86_64h' would always override the option passed
through '-march'.
This patch allows users to use '-march' with x86_64h, while keeping the
default to 'core-avx2'
Differential Revision: https://reviews.llvm.org/D55775
llvm-svn: 349381
pass in the -target-sdk-version to the compiler and backend
This commit adds support for reading the SDKSettings.json file in the Darwin
driver. This file is used by the driver to determine the SDK's version, and it
uses that information to pass it down to the compiler using the new
-target-sdk-version= option. This option is then used to set the appropriate
SDK Version module metadata introduced in r349119.
Note: I had to adjust the two ast tests as the SDKROOT environment variable
on macOS caused SDK version to be picked up for the compilation of source file
but not the AST.
rdar://45774000
Differential Revision: https://reviews.llvm.org/D55673
llvm-svn: 349380
Add a target_info definition for NetBSD. The definition is based
on the one used by FreeBSD, with libcxxrt replaced by libc++abi,
and using llvm-libunwind since we need to use its unwinder
implementation to build anyway.
Additionally, XFAIL the 30 tests that fail because of non-implemented
locale features. According to the manual, NetBSD implements only
LC_CTYPE part of locale handling. However, there is a locale database
in the system and locale specifications are validated against it,
so it makes sense to list the common locales as supported.
If I'm counting correctly, this change enables additional 43 passing
tests.
Differential Revision: https://reviews.llvm.org/D55767
llvm-svn: 349379
Remove the two test cases for \xDA and \xFA with UTF-8 locale, as both
characters alone are invalid in UTF-8 (short sequences). Upon removing
them, the test passes on Linux again (and also on NetBSD, after adding
appropriate locale configuration).
Differential Revision: https://reviews.llvm.org/D55746
llvm-svn: 349378
Summary: It looks like this support was added to match GNU AS, but only tests
.float and not .double. I asked RedHat folks to confirm that 0x7fffffffffffffff
was indeed the right value for NaN.
Same for infinity, but it only has positive / negative encodings.
Reviewers: scanon, rjmccall
Subscribers: jkorous, dexonsmith, llvm-commits
Differential Revision: https://reviews.llvm.org/D55531
llvm-svn: 349376
This is an initial patch to add the necessary support for a DemandedElts argument to SimplifyDemandedBits, more closely matching computeKnownBits and to help improve vector codegen.
I've added only a small amount of the changes necessary to get at least one test to update - a lot more can be done but I'd like to add these methodically with proper test coverage, at the same time the hope is to slowly move some/all of SimplifyDemandedVectorElts into SimplifyDemandedBits as well.
Differential Revision: https://reviews.llvm.org/D55768
llvm-svn: 349374
The --repl option was incorrectly defined as "Separate" (option and
value separated by a space). This resulted in the option not being
picked up when no value was specified.
This patch fixes the driver so that `--repl` is recognized again. I
split the option into two:
- A flag: `--repl` and `-r` which take no arguments.
- A joined option: `--repl=<flags>` and `-r=<flags>` that forward its
values to the repl.
This should match the driver's old behavior.
llvm-svn: 349371
If a saturating add/sub has one constant operand, then we can
determine the possible range of outputs it can produce, and simplify
an icmp comparison based on that.
The implementation is based on a similar existing mechanism for
simplifying binary operator + icmps.
Differential Revision: https://reviews.llvm.org/D55735
llvm-svn: 349369
This is a little dangerous since the crashlog files aren't 100%
unambiguous, but the risk is mitigated by using a non-greedy +?
pattern.
rdar://problem/38478511
Differential Revision: https://reviews.llvm.org/D55608
llvm-svn: 349367
Often users have a crash log an d a .dSYM bundle, but not the original
application binary. It turns out that for crash symbolication, we can
safely fall back to using the binary inside the .dSYM bundle.
Differential Revision: https://reviews.llvm.org/D55607
llvm-svn: 349366
We keep a few iterators into the basic block we're selecting while
performing FastISel. Usually this is fine, but occasionally code wants
to remove already-emitted instructions. When this happens we have to be
careful to update those iterators so they're not pointint at dangling
memory.
llvm-svn: 349365
The first one allows us to add an enumerator to an enum if we
already have an APSInt, since ultimately the implementation just
constructs one anyway. The second is just a general utility
function to covert a CompilerType to a clang::TagDecl.
llvm-svn: 349360
This is a re-application of r345525, which had been reverted by fear of
a regression.
Reviewed as https://reviews.llvm.org/D53994.
Thanks to Denis Yaroshevskiy for the patch.
llvm-svn: 349358
These features (fairly) recently got split out into their own feature, so we
should make CodeGen use them when available. The main change here is that the
check used to be based on the triple, but now it's based on CPU features.
llvm-svn: 349355
Class InstrBuilder wrongly assumed that llvm targets were always able to return
a non-null pointer when createMCInstrAnalysis() was called on them.
This was causing crashes when simulating executions for targets that don't
provide an MCInstrAnalysis object.
This patch fixes the issue by making MCInstrAnalysis optional.
llvm-svn: 349352
Summary:
There are certain cases when normal C/C++ lookup (localUncachedLookup)
does not find AST nodes. E.g.:
Example 1:
template <class T>
struct X {
friend void foo(); // this is never found in the DC of the TU.
};
Example 2:
// The fwd decl to Foo is not found in the lookupPtr of the DC of the
// translation unit decl.
struct A { struct Foo *p; };
In these cases we create a new node instead of returning with the old one.
To fix it we create a new lookup table which holds every node and we are
not interested in any C++ specific visibility considerations.
Simply, we must know if there is an existing Decl in a given DC.
Reviewers: a_sidorin, a.sidorin
Subscribers: mgorny, rnkovacs, dkrupp, Szelethus, cfe-commits
Differential Revision: https://reviews.llvm.org/D53708
llvm-svn: 349351
Summary:
The crux of the issue that is being fixed is that lookup could not find
previous decls of a friend class. The solution involves making the
friend declarations visible in their decl context (i.e. adding them to
the lookup table).
Also, we simplify `VisitRecordDecl` greatly.
This fix involves two other repairs (without these the unittests fail):
(1) We could not handle the addition of injected class types properly
when a redecl chain was involved, now this is fixed.
(2) DeclContext::removeDecl failed if the lookup table in Vector form
did not contain the to be removed element. This caused troubles in
ASTImporter::ImportDeclContext. This is also fixed.
Reviewers: a_sidorin, balazske, a.sidorin
Subscribers: rnkovacs, dkrupp, Szelethus, cfe-commits
Differential Revision: https://reviews.llvm.org/D53655
llvm-svn: 349349
Summary:
There was a chance that multiple clangd instances could try to write
same shard, in which case we would get a malformed file most likely. This patch
changes the writing mechanism to first write to a temporary file and then rename
it to fit real destination. Which is guaranteed to be atomic by POSIX.
Reviewers: ilya-biryukov
Subscribers: ioeric, MaskRay, jkorous, arphaman, jfb, cfe-commits
Differential Revision: https://reviews.llvm.org/D55417
llvm-svn: 349348
