Store the arguments of CXXConstructExpr in a trailing array. This is very
similar to the CallExpr case in D55771, with the exception that there is
only one derived class (CXXTemporaryObjectExpr) and that we compute the
offset to the trailing array instead of storing it.
This saves one pointer per CXXConstructExpr and CXXTemporaryObjectExpr.
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D56022
llvm-svn: 350003
This patch is a different approach to landing the reverted r349701.
It is expected to have the same object (memory region) treated as if it has
different types in different program points. The correct behavior for
RegionStore when an object is stored as an object of type T1 but loaded as
an object of type T2 is to store the object as if it has type T1 but cast it
to T2 during load.
Note that the cast here is some sort of a "reinterpret_cast" (even in C). For
instance, if you store an integer and load a float, you won't get your integer
represented as a float; instead, you will get garbage.
Admit that we cannot perform the cast and return an unknown value.
Differential Revision: https://reviews.llvm.org/D55875
rdar://problem/45062567
llvm-svn: 349984
This fixes compiler crash when we attempted to compile this code:
extern __device__ int data;
__device__ int data = 1;
Differential Revision: https://reviews.llvm.org/D56033
llvm-svn: 349981
It is faster to directly call the ObjC runtime for methods such as retain/release instead of sending a message to those functions.
Differential Revision: https://reviews.llvm.org/D55869
Reviewed By: rjmccall
llvm-svn: 349952
removed
Stat cache chaining was implemented for a StatListener in the PTH writer so that
it could write out the stat information to PTH. r348266 removed support for PTH,
and it doesn't seem like there are other uses of stat cache chaining. We can
remove the chaining support.
Differential Revision: https://reviews.llvm.org/D55455
llvm-svn: 349942
Since CallExpr::setNumArgs has been removed, it is now possible to store the
callee expression and the argument expressions of CallExpr in a trailing array.
This saves one pointer per CallExpr, CXXOperatorCallExpr, CXXMemberCallExpr,
CUDAKernelCallExpr and UserDefinedLiteral.
Given that CallExpr is used as a base of the above classes we cannot use
llvm::TrailingObjects. Instead we store the offset in bytes from the this pointer
to the start of the trailing objects and manually do the casts + arithmetic.
Some notes:
1.) I did not try to fit the number of arguments in the bit-fields of Stmt.
This leaves some space for future additions and avoid the discussion about
whether x bits are sufficient to hold the number of arguments.
2.) It would be perfectly possible to recompute the offset to the trailing
objects before accessing the trailing objects. However the trailing objects
are frequently accessed and benchmarks show that it is slightly faster to
just load the offset from the bit-fields. Additionally, because of 1),
we have plenty of space in the bit-fields of Stmt.
Differential Revision: https://reviews.llvm.org/D55771
Reviewed By: rjmccall
llvm-svn: 349910
All of the other constructors already take a reference to the AST context.
This avoids calling Decl::getASTContext in most cases. Additionally move
the definition of the constructor from Expr.h to Expr.cpp since it is calling
DeclRefExpr::computeDependence. NFC.
llvm-svn: 349901
functions that are unavailable on Darwin are explicitly called or called
from deleting destructors.
rdar://problem/40736230
Differential Revision: https://reviews.llvm.org/D47757
llvm-svn: 349890
The fix done in D55465 did not previously apply when the function was inlined.
rdar://46889541
Differential Revision: https://reviews.llvm.org/D55976
llvm-svn: 349876
Previously, we were not printing a note at all if at least one of the parameters was not annotated.
rdar://46888422
Differential Revision: https://reviews.llvm.org/D55972
llvm-svn: 349875
GCC does not mangle it when it is not explicit in the source. The
mangler as currently written cannot differentiate between explicit and
implicit calling conventions, so we can't match GCC. Explicit thiscall
conventions are rare, so mangle as if the convention was implicit to be
as ABI compatible as possible.
Also fixes some tests using %itanium_abi_triple in some configurations
as a side effect.
Fixes PR40107.
llvm-svn: 349872
If an -analyzer-config is passed through -Xanalyzer, it is not found while
looking for -Xclang.
Additionally, don't emit -analyzer-config-compatibility-mode for *every*
-analyzer-config flag we encounter; one is enough.
https://reviews.llvm.org/D55823
rdar://problem/46504165
llvm-svn: 349866
If it ends with "Retain" like CFRetain and returns a CFTypeRef like CFRetain,
then it is not necessarily a CFRetain. But it is indeed true that these two
return something retained.
Differential Revision: https://reviews.llvm.org/D55907
rdar://problem/39390714
llvm-svn: 349862
Fixes assertion
> Assertion failed: (isa<X>(Val) && "cast<Ty>() argument of incompatible type!"), function cast, file llvm/Support/Casting.h, line 255.
It was triggered by trying to cast `FunctionDecl` to `CXXMethodDecl` as
`CGF.CurCodeDecl` in `CallBaseDtor::Emit`. It was happening because
cleanups were emitted in `ScalarExprEmitter::VisitExprWithCleanups`
after destroying `InlinedInheritingConstructorScope`, so
`CodeGenFunction.CurCodeDecl` didn't correspond to expected cleanup decl.
Fix the assertion by emitting cleanups before leaving
`InlinedInheritingConstructorScope` and changing `CurCodeDecl`.
Test cases based on a patch by Shoaib Meenai.
Fixes PR36748.
rdar://problem/45805151
Reviewers: rsmith, rjmccall
Reviewed By: rjmccall
Subscribers: jkorous, dexonsmith, cfe-commits, smeenai, compnerd
Differential Revision: https://reviews.llvm.org/D55543
llvm-svn: 349848
Namespaces are introduced by adding an "identifier." before a
push/pop directive. Pop directives with namespaces can only pop a
attribute group that was pushed with the same namespace. Push and pop
directives that don't opt into namespaces have the same semantics.
This is necessary to prevent a pitfall of using multiple #pragma
clang attribute directives spread out in a large file, particularly
when macros are involved. It isn't easy to see which pop corripsonds
to which push, so its easy to inadvertently pop the wrong group.
Differential revision: https://reviews.llvm.org/D55628
llvm-svn: 349845
use the pointer to the class as the result type of the message
Prior to this commit, messages to self in class methods were treated as instance
methods to a Class value. When these methods returned instancetype the compiler
only saw id through the instancetype, and not the Interface *. This caused
problems when that return value was a receiver in a message send, as the
compiler couldn't select the right method declaration and had to rely on a
selection from the global method pool.
This commit modifies the semantics of such message sends and uses class messages
that are dispatched to the interface that corresponds to the class that contains
the class method. This ensures that instancetypes are correctly interpreted by
the compiler. This change is safe under ARC (as self can't be reassigned),
however, it also applies to MRR code as we are assuming that the user isn't
doing anything unreasonable.
rdar://20940997
Differential Revision: https://reviews.llvm.org/D36790
llvm-svn: 349841
The -c flag causes a .o file to appear every time we run a test.
Remove it.
Differential Revision: https://reviews.llvm.org/D55823
rdar://problem/46504165
llvm-svn: 349835
Buildbots can't find the linker, which we don't really need in our tests.
Differential Revision: https://reviews.llvm.org/D55823
rdar://problem/46504165
llvm-svn: 349828
__builtin_cpu_supports and __builtin_cpu_is use information in __cpu_model to decide cpu features. Before this change, __cpu_model was not declared as dso local. The generated code looks up the address in GOT when reading __cpu_model. This makes it impossible to use these functions in ifunc, because at that time GOT entries have not been relocated. This change makes it dso local.
Differential Revision: https://reviews.llvm.org/D53850
llvm-svn: 349825
Since r348038 we emit an error every time an -analyzer-config option is not
found. The driver, however, suppresses this error with another flag,
-analyzer-config-compatibility-mode, so backwards compatibility is maintained,
while analyzer developers still enjoy the new typo-free experience.
The backwards compatibility turns out to be still broken when the -analyze
action is not specified; it is still possible to specify -analyzer-config
in that case. This should be fixed now.
Patch by Kristóf Umann!
Differential Revision: https://reviews.llvm.org/D55823
rdar://problem/46504165
llvm-svn: 349824
Instead of generating llvm.mem.parallel_loop_access metadata, generate
llvm.access.group on instructions and llvm.loop.parallel_accesses on
loops. There is one access group per generated loop.
This is clang part of D52116/r349725.
Differential Revision: https://reviews.llvm.org/D52117
llvm-svn: 349823
This adds anchors to all of the documented checks so that you can directly link to a check by a stable name. This is useful because the SARIF file format has a field for specifying a URI to documentation for a rule and some viewers, like CodeSonar, make use of this information. These links are then exposed through the SARIF exporter.
llvm-svn: 349812
When checking that the array access is not out-of-bounds in CheckArrayAccess
it is possible that the type of the base expression after IgnoreParenCasts is
incomplete, even though the type of the base expression before IgnoreParenCasts
is complete. In this case we have no information about whether the array access
is out-of-bounds and we should just bail-out instead. This fixes PR39746 which
was caused by trying to obtain the size of an incomplete type.
Differential Revision: https://reviews.llvm.org/D55862
Reviewed By: efriedma
llvm-svn: 349811
This reverts commit r349701.
The patch was incorrect. The whole point of CastRetrievedVal()
is to handle the case in which the type from which the cast is made
(i.e., the "type" of value `V`) has nothing to do with the type of
the region it was loaded from (i.e., `R->getValueType()`).
Differential Revision: https://reviews.llvm.org/D55875
rdar://problem/45062567
llvm-svn: 349798
Calls to this function are deleted in the ARC optimizer. However when the ARC
optimizer was updated to use intrinsics instead of functions (r349534), the corresponding
clang change (r349535) to use intrinsics missed this one so it wasn't being deleted.
llvm-svn: 349782
This adds assembly-level tests to verify that the high-level
intrinsics generate the instructions they're supposed to.
These tests would have caught the codegen bugs I just fixed.
llvm-svn: 349753
Replace multiple comparisons of getOS() value with FreeBSD, NetBSD,
OpenBSD and DragonFly with matching isOS*BSD() methods. This should
improve the consistency of coding style without changing the behavior.
Direct getOS() comparisons were left whenever used in switch or switch-
like context.
Differential Revision: https://reviews.llvm.org/D55916
llvm-svn: 349752
The following two bugs in SystemZ high-level vector intrinsics are
fixes by this patch:
- The float case of vec_insert_and_zero should generate a VLLEZF
pattern, but currently erroneously generates VLLEZLF.
- The float and double versions of vec_orc erroneously generate
and-with-complement instead of or-with-complement.
The patch also fixes a couple of typos in the associated test.
llvm-svn: 349751
It is expected to have the same object (memory region) treated as if it has
different types in different program points. The correct behavior for
RegionStore when an object is stored as an object of type T1 but loaded as
an object of type T2 is to store the object as if it has type T1 but cast it
to T2 during load.
Note that the cast here is some sort of a "reinterpret_cast" (even in C). For
instance, if you store a float and load an integer, you won't have your float
rounded to an integer; instead, you will have garbage.
Admit that we cannot perform the cast as long as types we're dealing with are
non-trivial (neither integers, nor pointers).
Of course, if the cast is not necessary (eg, T1 == T2), we can still load the
value just fine.
Differential Revision: https://reviews.llvm.org/D55875
rdar://problem/45062567
llvm-svn: 349701
Static Analyzer processes the program function-by-function, sometimes diving
into other functions ("inlining" them). When an object is returned from an
inlined function, Return Value Optimization is modeled, and the returned object
is constructed at its return location directly.
When an object is returned from the function from which the analysis has started
(the top stack frame of the analysis), the return location is unknown. Model it
with a SymbolicRegion based on a conjured symbol that is specifically tagged for
that purpose, because this is generally the correct way to symbolicate
unknown locations in Static Analyzer.
Fixes leak false positives when an object is returned from top frame in C++17:
objects that are put into a SymbolicRegion-based memory region automatically
"escape" and no longer get reported as leaks. This only applies to C++17 return
values with destructors, because it produces a redundant CXXBindTemporaryExpr
in the call site, which confuses our liveness analysis. The actual fix
for liveness analysis is still pending, but it is no longer causing problems.
Additionally, re-enable temporary destructor tests in C++17.
Differential Revision: https://reviews.llvm.org/D55804
rdar://problem/46217550
llvm-svn: 349696
It turns out that it's not all that uncommon to have a C++ override of, say,
memcpy that receives a structure (or two) by reference (or by value, if it's
being copied from) and copies memory from it (or into it, if it's passed
by reference). In this case the argument will be of structure type (recall that
expressions of reference type do not exist: instead, C++ classifies expressions
into prvalues and lvalues and xvalues).
In this scenario we crash because we are trying to assume that, say,
a memory region is equal to an empty CompoundValue (the non-lazy one; this is
what makeZeroVal() return for compound types and it represents prvalue of
an object that is initialized with an empty initializer list).
Add defensive checks.
Differential Revision: https://reviews.llvm.org/D55873
rdar://problem/45366551
llvm-svn: 349682
Looks like these were in place to make these types move-only. That's
generally not a feature that the type should prescribe (unless it's an
inherent limitation) - instead leaving it up to the users of a type.
llvm-svn: 349669
Need to mark the loop as started when the initialization statement is
found. It is required to prevent possible incorrect loop iteraton
variable detection during template instantiation and fix the compiler
crash during the codegen.
llvm-svn: 349657
Summary:
The clang used to pick up the qualifiers of the lamba's call operator
(which is always const) and fail to show non-const methods of 'this' in
completion results.
Reviewers: kadircet
Reviewed By: kadircet
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D55885
llvm-svn: 349655
NetBSD intends to support only reentrant interfaces in interceptors.
When -lpthread is used without _REENTRANT defined, things are
not guaranteed to work.
This is especially important for <stdio.h> and sanitization of
interfaces around FILE. Some APIs have alternative modes depending
on the _REENTRANT definition, and NetBSD intends to support sanitization
of the _REENTRANT ones.
Differential Revision: https://reviews.llvm.org/D55654
llvm-svn: 349650
Add a simple method to query whether any sanitizer was enabled,
via SanitizerArgs. This will be used in the NetBSD driver to pass
additional definitions that are required by all sanitizers.
Differential Revision: https://reviews.llvm.org/D55832
llvm-svn: 349649
Avoid passing -faddrsig by default on NetBSD. This platform is still
using old GNU binutils that crashes on executables containing those
sections.
Differential Revision: https://reviews.llvm.org/D55828
llvm-svn: 349647
Without this patch, clang doesn't complain that X needs explicit data
sharing attributes in the following:
```
#pragma omp target teams default(none)
{
#pragma omp parallel num_threads(X)
;
}
```
However, clang does produce that complaint after the braces are
removed. With this patch, clang complains in both cases.
Reviewed By: ABataev
Differential Revision: https://reviews.llvm.org/D55861
llvm-svn: 349635
Sibling patch to D55855, this emits UADD_SAT/USUB_SAT generic intrinsics for the SSE saturated math intrinsics instead of expanding to a IR code sequence that could be difficult to reassemble.
Differential Revision: https://reviews.llvm.org/D55879
llvm-svn: 349631
urllib2 as been renamed into urllib and the library layout has changed.
Workaround that in a consistent manner.
Differential Revision: https://reviews.llvm.org/D55199
llvm-svn: 349627
Summary:
Some ASM input constraints (e.g., "i" and "n") require immediate values. At O0,
very few code transformations are performed. So if we cannot resolve to an
immediate when emitting the ASM input we shouldn't delay its processing.
Reviewers: rsmith, efriedma
Reviewed By: efriedma
Subscribers: rehana, efriedma, craig.topper, jyknight, cfe-commits
Differential Revision: https://reviews.llvm.org/D55616
llvm-svn: 349561
A map clause with the close map-type-modifier is a hint to
prefer that the variables are mapped using a copy into faster
memory.
Patch by Ahsan Saghir (saghir)
Differential Revision: https://reviews.llvm.org/D55719
llvm-svn: 349551
This updates the FunctionProtoType visualizer to use the proper bits for determining parameter information and the DeclarationName visualizer to use the detail namespace. It also adds support for viewing newer special declaration names (like deduction guides).
Patch with help of Bruno Ricci.
llvm-svn: 349547
The special lowering for __builtin_mul_overflow introduced in r320902
fixed an ICE seen when passing mixed-sign operands to the builtin.
This patch extends the special lowering to cover mixed-width, mixed-sign
operands. In a few common scenarios, calls to muloti4 will no longer be
emitted.
This should address the latest comments in PR34920 and work around the
link failure seen in:
https://bugzilla.redhat.com/show_bug.cgi?id=1657544
Testing:
- check-clang
- A/B output comparison with: https://gist.github.com/vedantk/3eb9c88f82e5c32f2e590555b4af5081
Differential Revision: https://reviews.llvm.org/D55843
llvm-svn: 349542
buffer.
Seems to me, nvlink has a bug with the proper support of the weakly
linked symbols. It does not allow to define several shared memory buffer
with the different sizes even with the weak linkage. Instead we always
use 128 bytes buffer to prevent nvlink from the error message emission.
llvm-svn: 349540
NFC for targets other than PS4.
Respect -nostdlib and -nodefaultlibs when enabling asan or ubsan.
Differential Revision: https://reviews.llvm.org/D55712
llvm-svn: 349508
This is exactly a "CreateBitCast", so refactor this to get rid of a
'new'.
Note that this slightly changes the test, as the Builder is now
seemingly smart enough to fold one of the bitcasts into the annotation
call.
Change-Id: I1733fb1fdf91f5c9d88651067130b9a4e7b5ab67
llvm-svn: 349506
In Python3, dict.items, dict.keys, dict.values, zip, map and filter no longer return lists, they create generator instead.
The portability patch consists in forcing an extra `list` call if the result is actually used as a list.
`map` are replaced by list comprehension and `filter` by filtered list comprehension.
Differential Revision: https://reviews.llvm.org/D55197
llvm-svn: 349501
Only explicitly look through integer and floating-point promotion where the result type is actually a promotion, which is not always the case for bit-fields in C.
Patch by Bevin Hansson.
llvm-svn: 349497
Summary:
This supposes to be a non-functional change. We have two code paths when
traversing lambda expressions:
1) traverse the function proto typeloc when parameters and return type
are explicit;
2) otherwise fallback to traverse parameter decls and return type loc
individually;
This patch unifies the code path to always traverse parameters and
return type, rather than relying on traversing the full type-loc.
Reviewers: ilya-biryukov
Subscribers: arphaman, cfe-commits
Differential Revision: https://reviews.llvm.org/D55820
llvm-svn: 349494
In Python2, division between integer yields an integer, while it yields a float in Python3.
Use a combination of from __future__ import division and // operator to get a portable behavior.
Differential Revision: https://reviews.llvm.org/D55204
llvm-svn: 349455
Using from __future__ import print_function it is possible to have a compatible behavior of `print(...)` across Python version.
Differential Revision: https://reviews.llvm.org/D55213
llvm-svn: 349454
For targets where SEH exceptions are used by default (on MinGW,
only x86_64 so far), -munwind-tables are added automatically. If
-fseh-exeptions is enabled on a target where SEH exeptions are
availble but not enabled by default yet (aarch64), we need to
pass -munwind-tables if -fseh-exceptions was specified.
Differential Revision: https://reviews.llvm.org/D55749
llvm-svn: 349452
Replace `xrange(...)` by either `range(...)` or `list(range(...))` depending on the context.
Differential Revision: https://reviews.llvm.org/D55193
llvm-svn: 349448
Summary:
Add an option to initialize automatic variables with either a pattern or with
zeroes. The default is still that automatic variables are uninitialized. Also
add attributes to request uninitialized on a per-variable basis, mainly to disable
initialization of large stack arrays when deemed too expensive.
This isn't meant to change the semantics of C and C++. Rather, it's meant to be
a last-resort when programmers inadvertently have some undefined behavior in
their code. This patch aims to make undefined behavior hurt less, which
security-minded people will be very happy about. Notably, this means that
there's no inadvertent information leak when:
- The compiler re-uses stack slots, and a value is used uninitialized.
- The compiler re-uses a register, and a value is used uninitialized.
- Stack structs / arrays / unions with padding are copied.
This patch only addresses stack and register information leaks. There's many
more infoleaks that we could address, and much more undefined behavior that
could be tamed. Let's keep this patch focused, and I'm happy to address related
issues elsewhere.
To keep the patch simple, only some `undef` is removed for now, see
`replaceUndef`. The padding-related infoleaks are therefore not all gone yet.
This will be addressed in a follow-up, mainly because addressing padding-related
leaks should be a stand-alone option which is implied by variable
initialization.
There are three options when it comes to automatic variable initialization:
0. Uninitialized
This is C and C++'s default. It's not changing. Depending on code
generation, a programmer who runs into undefined behavior by using an
uninialized automatic variable may observe any previous value (including
program secrets), or any value which the compiler saw fit to materialize on
the stack or in a register (this could be to synthesize an immediate, to
refer to code or data locations, to generate cookies, etc).
1. Pattern initialization
This is the recommended initialization approach. Pattern initialization's
goal is to initialize automatic variables with values which will likely
transform logic bugs into crashes down the line, are easily recognizable in
a crash dump, without being values which programmers can rely on for useful
program semantics. At the same time, pattern initialization tries to
generate code which will optimize well. You'll find the following details in
`patternFor`:
- Integers are initialized with repeated 0xAA bytes (infinite scream).
- Vectors of integers are also initialized with infinite scream.
- Pointers are initialized with infinite scream on 64-bit platforms because
it's an unmappable pointer value on architectures I'm aware of. Pointers
are initialize to 0x000000AA (small scream) on 32-bit platforms because
32-bit platforms don't consistently offer unmappable pages. When they do
it's usually the zero page. As people try this out, I expect that we'll
want to allow different platforms to customize this, let's do so later.
- Vectors of pointers are initialized the same way pointers are.
- Floating point values and vectors are initialized with a negative quiet
NaN with repeated 0xFF payload (e.g. 0xffffffff and 0xffffffffffffffff).
NaNs are nice (here, anways) because they propagate on arithmetic, making
it more likely that entire computations become NaN when a single
uninitialized value sneaks in.
- Arrays are initialized to their homogeneous elements' initialization
value, repeated. Stack-based Variable-Length Arrays (VLAs) are
runtime-initialized to the allocated size (no effort is made for negative
size, but zero-sized VLAs are untouched even if technically undefined).
- Structs are initialized to their heterogeneous element's initialization
values. Zero-size structs are initialized as 0xAA since they're allocated
a single byte.
- Unions are initialized using the initialization for the largest member of
the union.
Expect the values used for pattern initialization to change over time, as we
refine heuristics (both for performance and security). The goal is truly to
avoid injecting semantics into undefined behavior, and we should be
comfortable changing these values when there's a worthwhile point in doing
so.
Why so much infinite scream? Repeated byte patterns tend to be easy to
synthesize on most architectures, and otherwise memset is usually very
efficient. For values which aren't entirely repeated byte patterns, LLVM
will often generate code which does memset + a few stores.
2. Zero initialization
Zero initialize all values. This has the unfortunate side-effect of
providing semantics to otherwise undefined behavior, programs therefore
might start to rely on this behavior, and that's sad. However, some
programmers believe that pattern initialization is too expensive for them,
and data might show that they're right. The only way to make these
programmers wrong is to offer zero-initialization as an option, figure out
where they are right, and optimize the compiler into submission. Until the
compiler provides acceptable performance for all security-minded code, zero
initialization is a useful (if blunt) tool.
I've been asked for a fourth initialization option: user-provided byte value.
This might be useful, and can easily be added later.
Why is an out-of band initialization mecanism desired? We could instead use
-Wuninitialized! Indeed we could, but then we're forcing the programmer to
provide semantics for something which doesn't actually have any (it's
uninitialized!). It's then unclear whether `int derp = 0;` lends meaning to `0`,
or whether it's just there to shut that warning up. It's also way easier to use
a compiler flag than it is to manually and intelligently initialize all values
in a program.
Why not just rely on static analysis? Because it cannot reason about all dynamic
code paths effectively, and it has false positives. It's a great tool, could get
even better, but it's simply incapable of catching all uses of uninitialized
values.
Why not just rely on memory sanitizer? Because it's not universally available,
has a 3x performance cost, and shouldn't be deployed in production. Again, it's
a great tool, it'll find the dynamic uses of uninitialized variables that your
test coverage hits, but it won't find the ones that you encounter in production.
What's the performance like? Not too bad! Previous publications [0] have cited
2.7 to 4.5% averages. We've commmitted a few patches over the last few months to
address specific regressions, both in code size and performance. In all cases,
the optimizations are generally useful, but variable initialization benefits
from them a lot more than regular code does. We've got a handful of other
optimizations in mind, but the code is in good enough shape and has found enough
latent issues that it's a good time to get the change reviewed, checked in, and
have others kick the tires. We'll continue reducing overheads as we try this out
on diverse codebases.
Is it a good idea? Security-minded folks think so, and apparently so does the
Microsoft Visual Studio team [1] who say "Between 2017 and mid 2018, this
feature would have killed 49 MSRC cases that involved uninitialized struct data
leaking across a trust boundary. It would have also mitigated a number of bugs
involving uninitialized struct data being used directly.". They seem to use pure
zero initialization, and claim to have taken the overheads down to within noise.
Don't just trust Microsoft though, here's another relevant person asking for
this [2]. It's been proposed for GCC [3] and LLVM [4] before.
What are the caveats? A few!
- Variables declared in unreachable code, and used later, aren't initialized.
This goto, Duff's device, other objectionable uses of switch. This should
instead be a hard-error in any serious codebase.
- Volatile stack variables are still weird. That's pre-existing, it's really
the language's fault and this patch keeps it weird. We should deprecate
volatile [5].
- As noted above, padding isn't fully handled yet.
I don't think these caveats make the patch untenable because they can be
addressed separately.
Should this be on by default? Maybe, in some circumstances. It's a conversation
we can have when we've tried it out sufficiently, and we're confident that we've
eliminated enough of the overheads that most codebases would want to opt-in.
Let's keep our precious undefined behavior until that point in time.
How do I use it:
1. On the command-line:
-ftrivial-auto-var-init=uninitialized (the default)
-ftrivial-auto-var-init=pattern
-ftrivial-auto-var-init=zero -enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
2. Using an attribute:
int dont_initialize_me __attribute((uninitialized));
[0]: https://users.elis.ugent.be/~jsartor/researchDocs/OOPSLA2011Zero-submit.pdf
[1]: https://twitter.com/JosephBialek/status/1062774315098112001
[2]: https://outflux.net/slides/2018/lss/danger.pdf
[3]: https://gcc.gnu.org/ml/gcc-patches/2014-06/msg00615.html
[4]: 776a0955ef
[5]: http://wg21.link/p1152
I've also posted an RFC to cfe-dev: http://lists.llvm.org/pipermail/cfe-dev/2018-November/060172.html
<rdar://problem/39131435>
Reviewers: pcc, kcc, rsmith
Subscribers: JDevlieghere, jkorous, dexonsmith, cfe-commits
Differential Revision: https://reviews.llvm.org/D54604
llvm-svn: 349442
Now that MSVC compatibility versions are stored as a four digit number
(1912) instead of a two digit number (19), we need to adjust how we
handle this attribute.
Also add a new test that was intended to be part of r349414.
llvm-svn: 349415
Summary:
The msvc exception specifier for noexcept function types has changed
from the prior default of "Z" to "_E" if the function cannot throw when
compiling with /std:C++17.
Patch by Zachary Henkel!
Reviewers: zturner, rnk
Reviewed By: rnk
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D55685
llvm-svn: 349414
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
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
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
Accidentally commited earlier with the same commit title, but really it
should've been
"Revert rC349283 '[analyzer][MallocChecker] Improve warning messages on double-delete errors'"
llvm-svn: 349344
Improve the description of these command line options
by providing specific heuristic information, as outlined
for the ssp function attribute(s) in LLVM's documentation.
Also rewords -fstack-protector-all for affinity.
Differential Revision: https://reviews.llvm.org/D55428
llvm-svn: 349335
This checker warns you when you re-use an object after moving it.
Mostly developed by Peter Szecsi!
Differential Revision: https://reviews.llvm.org/D38675
llvm-svn: 349328
Re-using a moved-from local variable is most likely a bug because there's
rarely a good motivation for not introducing a separate variable instead.
We plan to keep emitting such warnings by default.
Introduce a flag that allows disabling warnings on local variables that are
not of a known move-unsafe type. If it doesn't work out as we expected,
we'll just flip the flag.
We still warn on move-unsafe objects and unsafe operations on known move-safe
objects.
Differential Revision: https://reviews.llvm.org/D55730
llvm-svn: 349327
This re-applies commit r349226 that was reverted in r349233 due to failures
on clang-x64-windows-msvc.
Specify enum type as unsigned for use in bit field. Otherwise overflows
may cause UB.
Differential Revision: https://reviews.llvm.org/D55388
llvm-svn: 349326
StaticAnalyzer uses the CFG-based RelaxedLiveVariables analysis in order to,
in particular, figure out values of which expressions are still needed.
When the expression becomes "dead", it is garbage-collected during
the dead binding scan.
Expressions that constitute branches/bodies of control flow statements,
eg. `E1' in `if (C1) E1;' but not `E2' in `if (C2) { E2; }', were kept alive
for too long. This caused false positives in MoveChecker because it relies
on cleaning up loop-local variables when they go out of scope, but some of those
live-for-too-long expressions were keeping a reference to those variables.
Fix liveness analysis to correctly mark these expressions as dead.
Add a debug checker, debug.DumpLiveStmts, in order to test expressions liveness.
Differential Revision: https://reviews.llvm.org/D55566
llvm-svn: 349320
Summary:
The pattern is problematic with C++ exceptions, and not as widespread as
scoped locks, but it's still used by some, for example Chromium.
We are a bit stricter here at join points, patterns that are allowed for
scoped locks aren't allowed here. That could still be changed in the
future, but I'd argue we should only relax this if people ask for it.
Fixes PR36162.
Reviewers: aaron.ballman, delesley, pwnall
Reviewed By: delesley, pwnall
Subscribers: pwnall, cfe-commits
Differential Revision: https://reviews.llvm.org/D52578
llvm-svn: 349300
This patch merely reorganizes some things, and features no functional change.
In detail:
* Provided documentation, or moved existing documentation in more obvious
places.
* Added dividers. (the //===----------===// thing).
* Moved getAllocationFamily, printAllocDeallocName, printExpectedAllocName and
printExpectedDeallocName in the global namespace on top of the file where
AllocationFamily is declared, as they are very strongly related.
* Moved isReleased and MallocUpdateRefState near RefState's definition for the
same reason.
* Realloc modeling was very poor in terms of variable and structure naming, as
well as documentation, so I renamed some of them and added much needed docs.
* Moved function IdentifierInfos to a separate struct, and moved isMemFunction,
isCMemFunction adn isStandardNewDelete inside it. This makes the patch affect
quite a lot of lines, should I extract it to a separate one?
* Moved MallocBugVisitor out of MallocChecker.
* Preferred switches to long else-if branches in some places.
* Neatly organized some RUN: lines.
Differential Revision: https://reviews.llvm.org/D54823
llvm-svn: 349281
Now that CheckerRegistry lies in Frontend, we can finally eliminate
ClangCheckerRegistry. Fortunately, this also provides us with a
DiagnosticsEngine, so I went ahead and removed some parameters from it's
methods.
Differential Revision: https://reviews.llvm.org/D54437
llvm-svn: 349280
ClangCheckerRegistry is a very non-obvious, poorly documented, weird concept.
It derives from CheckerRegistry, and is placed in lib/StaticAnalyzer/Frontend,
whereas it's base is located in lib/StaticAnalyzer/Core. It was, from what I can
imagine, used to circumvent the problem that the registry functions of the
checkers are located in the clangStaticAnalyzerCheckers library, but that
library depends on clangStaticAnalyzerCore. However, clangStaticAnalyzerFrontend
depends on both of those libraries.
One can make the observation however, that CheckerRegistry has no place in Core,
it isn't used there at all! The only place where it is used is Frontend, which
is where it ultimately belongs.
This move implies that since
include/clang/StaticAnalyzer/Checkers/ClangCheckers.h only contained a single function:
class CheckerRegistry;
void registerBuiltinCheckers(CheckerRegistry ®istry);
it had to re purposed, as CheckerRegistry is no longer available to
clangStaticAnalyzerCheckers. It was renamed to BuiltinCheckerRegistration.h,
which actually describes it a lot better -- it does not contain the registration
functions for checkers, but only those generated by the tblgen files.
Differential Revision: https://reviews.llvm.org/D54436
llvm-svn: 349275
Renaming collectCheckers to getEnabledCheckers
Changing the functionality to acquire all enabled checkers, rather then collect
checkers for a specific CheckerOptInfo (for example, collecting all checkers for
{ "core", true }, which meant enabling all checkers from the core package, which
was an unnecessary complication).
Removing CheckerOptInfo, instead of storing whether the option was claimed via a
field, we handle errors immediately, as getEnabledCheckers can now access a
DiagnosticsEngine. Realize that the remaining information it stored is directly
accessible through AnalyzerOptions.CheckerControlList.
Fix a test with -analyzer-disable-checker -verify accidentally left in.
llvm-svn: 349274
Frontend headers have undefined reference on the symbol `clang::PCHContainerOperations::PCHContainerOperations()` through some shared_ptr usage. Any dependents will get the undefined reference which can only be resolved by explicit dependency on clangSerialization (due to -z defs).
llvm-svn: 349259
This matches what GCC does in these situations.
This fixes compiling Qt in debug mode. In release mode, references to
the vtable of this particular class ends up optimized away, but in debug
mode, the compiler creates references to the vtable, which is expected
to be dllexported from a different DLL. Make sure the dllexported
version actually ends up emitted.
Differential Revision: https://reviews.llvm.org/D55698
llvm-svn: 349256
Right now they report to have one parameter with null decl,
because initializing an ArrayRef of pointers with a nullptr
yields an ArrayRef to an array of one null pointer.
Fixes a crash in the OSObject section of RetainCountChecker.
Differential Revision: https://reviews.llvm.org/D55671
llvm-svn: 349229
The checker wasn't prepared to see the dealloc message sent to the class itself
rather than to an instance, as if it was +dealloc.
Additionally, it wasn't prepared for pure-unknown or undefined self values.
The new guard covers that as well, but it is annoying to test because
both kinds of values shouldn't really appear and we generally want to
get rid of all of them (by modeling unknown values with symbols and
by warning on use of undefined values before they are used).
The CHECK: directive for FileCheck at the end of the test looks useless,
so i removed it.
Differential Revision: https://reviews.llvm.org/D55680
llvm-svn: 349228
Use trackExpressionValue() (previously known as trackNullOrUndefValue())
to track index value in the report, so that the user knew
what Static Analyzer thinks the index is.
Additionally, implement printState() to help debugging the checker later.
Differential Revision: https://reviews.llvm.org/D55458
llvm-svn: 349227
Calling operator*() or operator->() on a null STL smart pointer is
undefined behavior.
Smart pointers are specified to become null after being moved from.
So we can't warn on arbitrary method calls, but these two operators
definitely make no sense.
The new bug is fatal because it's an immediate UB,
unlike other use-after-move bugs.
The work on a more generic null smart pointer dereference checker
is still pending.
Differential Revision: https://reviews.llvm.org/D55388
llvm-svn: 349226
Summary:
GCC 5.1 began mangling these Windows calling conventions into function
types, since they can be used for overloading. They've always been
mangled in the MS ABI, but they are new to the Itanium mangler. Note
that the calling convention doesn't appear as part of the main
declaration, it only appears on function parameter types and other
types.
Fixes PR39860
Reviewers: rjmccall, efriedma
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D55672
llvm-svn: 349212
All of the symbols demangle on llvm-undname and demangler.com. This
address space qualifier is useful for when we want to use opencl C++ in
Windows mode. Additionally, C++ address-space using functions will now
be usable on windows.
Differential Revision: https://reviews.llvm.org/D55715
Change-Id: Ife4506613c3cce778a783456d62117fbf7d83c26
llvm-svn: 349209
This reverts commit 46efdf2ccc2a80aefebf8433dbf9c7c959f6e629.
Richard Smith commented just after I submitted this that this is the
wrong solution. Reverting so that I can fix differently.
llvm-svn: 349206
Core issue 1013 suggests that having an uninitialied std::nullptr_t be
UB is a bit foolish, since there is only a single valid value. This DR
reports that DR616 fixes it, which does so by making lvalue-to-rvalue
conversions from nullptr_t be equal to nullptr.
However, just implementing that results in warnings/etc in many places.
In order to fix all situations where nullptr_t would seem uninitialized,
this patch instead (as an otherwise transparent extension) default
initializes uninitialized VarDecls of nullptr_t.
Differential Revision: https://reviews.llvm.org/D53713
Change-Id: I84d72a9290054fa55341e8cbdac43c8e7f25b885
llvm-svn: 349201
Summary:
This patch adds `__builtin_launder`, which is required to implement `std::launder`. Additionally GCC provides `__builtin_launder`, so thing brings Clang in-line with GCC.
I'm not exactly sure what magic `__builtin_launder` requires, but based on previous discussions this patch applies a `@llvm.invariant.group.barrier`. As noted in previous discussions, this may not be enough to correctly handle vtables.
Reviewers: rnk, majnemer, rsmith
Reviewed By: rsmith
Subscribers: kristina, Romain-Geissler-1A, erichkeane, amharc, jroelofs, cfe-commits, Prazek
Differential Revision: https://reviews.llvm.org/D40218
llvm-svn: 349195
Inlined runtime with the current implementation of the interwarp copy
function leads to the undefined behavior because of the not quite
correct implementation of the barriers. Start using generic
__kmpc_barier function instead of the custom made barriers.
llvm-svn: 349192
Some C++ standard library classes provide additional guarantees about their
state after move. Suppress warnings on such classes until a more precise
behavior is implemented. Warnings for locals are not suppressed anyway
because it's still most likely a bug.
Differential Revision: https://reviews.llvm.org/D55307
llvm-svn: 349191
If a moved-from object is passed into a conservatively evaluated function
by pointer or by reference, we assume that the function may reset its state.
Make sure it doesn't apply to const pointers and const references. Add a test
that demonstrates that it does apply to rvalue references.
Additionally, make sure that the object is invalidated when its contents change
for reasons other than invalidation caused by evaluating a call conservatively.
In particular, when the object's fields are manipulated directly, we should
assume that some sort of reset may be happening.
Differential Revision: https://reviews.llvm.org/D55289
llvm-svn: 349190
Functional changes include:
* The run.files property is now an array instead of a mapping.
* fileLocation objects now have a fileIndex property specifying the array index into run.files.
* The resource.rules property is now an array instead of a mapping.
* The result object was given a ruleIndex property that is an index into the resource.rules array.
* rule objects now have their "id" field filled out in addition to the name field.
* Updated the schema and spec version numbers to 11-28.
llvm-svn: 349188
Implement options in clang to enable recording the driver command-line
in an ELF section.
Implement a new special named metadata, llvm.commandline, to support
frontends embedding their command-line options in IR/ASM/ELF.
This differs from the GCC implementation in some key ways:
* In GCC there is only one command-line possible per compilation-unit,
in LLVM it mirrors llvm.ident and multiple are allowed.
* In GCC individual options are separated by NULL bytes, in LLVM entire
command-lines are separated by NULL bytes. The advantage of the GCC
approach is to clearly delineate options in the face of embedded
spaces. The advantage of the LLVM approach is to support merging
multiple command-lines unambiguously, while handling embedded spaces
with escaping.
Differential Revision: https://reviews.llvm.org/D54487
Clang Differential Revision: https://reviews.llvm.org/D54489
llvm-svn: 349155
Move some diagnostics around between Diagnostic*Kinds.td files. Diagnostics
used in multiple places were moved to DiagnosticCommonKinds.td. Diagnostics
listed in the wrong place (ie, Sema diagnostics listed in
DiagnosticsParseKinds.td) were moved to the correct places. One diagnostic
split into two so that the diagnostic string is in the .td file instead of in
code. Cleaned up the diagnostic includes after all the changes.
llvm-svn: 349125
This reverts commit r349064.
This wasn't updating the right test. Causing (not the different line number
from the previous revert):
======================================================================
FAIL: test_diagnostic_warning (tests.cindex.test_diagnostics.TestDiagnostics)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/Users/buildslave/jenkins/workspace/clang-stage1-configure-RA/llvm/tools/clang/bindings/python/tests/cindex/test_diagnostics.py", line 18, in test_diagnostic_warning
self.assertEqual(len(tu.diagnostics), 2)
AssertionError: 1 != 2
llvm-svn: 349118
intrin.h had forward declarations for these and lzcntintrin.h had implementations that were only available with -mlzcnt or a -march that supported the lzcnt feature.
For MS compatibility we should always have these builtins available regardless of X86 being the target or the CPU support the lzcnt instruction. The backends should be able to gracefully fallback to something support even if its just shifts and bit ops.
Unfortunately, gcc also implements 2 of the 3 function names here on X86 when lzcnt feature is enabled.
This patch adds builtins for these for MSVC compatibility and drops the forward declarations from intrin.h. To keep the gcc compatibility the two intrinsics that collided have been turned into macros that use the X86 specific builtins with the lzcnt feature check. These macros are only defined when _MSC_VER is not defined. Without them being macros we can get a redefinition error because -ms-extensions doesn't seem to set _MSC_VER but does make the MS builtins available.
Should fix PR40014
Differential Revision: https://reviews.llvm.org/D55677
llvm-svn: 349098
The host-side code can't (and should not) access the values that may
only exist on the device side. E.g. address of a __device__ function
does not exist on the host side as we don't generate the code for it there.
Differential Revision: https://reviews.llvm.org/D55663
llvm-svn: 349087
This updates the scan-build perl script to allow outputting to sarif in a more natural fashion by specifying -sarif as a command line argument, similar to how -plist is already supported.
llvm-svn: 349082
The DIFile used by the CU is special and distinct from the main source
file. Its directory part specifies what becomes the DW_AT_comp_dir
(the compilation directory), even if the source file was specified
with an absolute path.
To support the .dwo workflow, a valid DW_AT_comp_dir is necessary even
if source files were specified with an absolute path.
llvm-svn: 349065
Found the case in the clang codebase where the assertion fires.
To avoid crashing assertion-enabled builds before I re-add the missing
operation.
Will restore the assertion alongside the upcoming fix.
llvm-svn: 349061
Some versions of gcc, especially when invoked through ccache (-E), can have
trouble with raw string literals inside macros. This moves the string out of
the macro.
llvm-svn: 349059
Address spaces are cast into generic before invoking the constructor.
Added support for a trailing Qualifiers object in FunctionProtoType.
Note: This recommits the previously reverted patch,
but now it is commited together with a fix for lldb.
Differential Revision: https://reviews.llvm.org/D54862
llvm-svn: 349019
Summary:
private and internal: should not trigger ODR at all.
unnamed_addr: current ODR checking approach fail and rereport false violation if
a linker merges such globals
linkonce_odr, weak_odr: could cause similar problems and they are already not
instrumented for ELF.
Reviewers: eugenis, kcc
Subscribers: kubamracek, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D55621
llvm-svn: 349015
Summary:
This change adds a new AST matcher for block expressions.
Test Notes:
Ran the clang unit tests.
Reviewers: aaron.ballman
Reviewed By: aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D55546
llvm-svn: 349004
Statement memoization was removed in r348822 because it was noticed to cause
memory corruption. This was happening because a reference to an object
in a DenseMap was used after being invalidated by inserting a new key
into the map.
This test case crashes reliably under ASan (i.e., when Clang is built with
-DLLVM_USE_SANITIZER="Address") on at least some machines before r348822
and doesn't crash after it.
llvm-svn: 349000
The previous assertion was relatively easy to trigger, and likely will
be easy to trigger going forward. EmitDelegateCallArg is relatively
popular.
This cleanly diagnoses PR28299 while I work on a proper solution.
llvm-svn: 348991
__builtin_cpu_supports and __builtin_cpu_is use information in __cpu_model to decide cpu features. Before this change, __cpu_model was not declared as dso local. The generated code looks up the address in GOT when reading __cpu_model. This makes it impossible to use these functions in ifunc, because at that time GOT entries have not been relocated. This change makes it dso local.
Differential Revision: https://reviews.llvm.org/D53850
llvm-svn: 348978
Summary:
Currently the Clang AST doesn't store information about how the callee of a CallExpr was found. Specifically if it was found using ADL.
However, this information is invaluable to tooling. Consider a tool which renames usages of a function. If the originally CallExpr was formed using ADL, then the tooling may need to additionally qualify the replacement.
Without information about how the callee was found, the tooling is left scratching it's head. Additionally, we want to be able to match ADL calls as quickly as possible, which means avoiding computing the answer on the fly.
This patch changes `CallExpr` to store whether it's callee was found using ADL. It does not change the size of any AST nodes.
Reviewers: fowles, rsmith, klimek, shafik
Reviewed By: rsmith
Subscribers: aaron.ballman, riccibruno, calabrese, titus, cfe-commits
Differential Revision: https://reviews.llvm.org/D55534
llvm-svn: 348977
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
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
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
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
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
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
This is a more thorough fix of rC348911.
The story about -DBUILD_SHARED_LIBS=on build after rC348907 (Move PCHContainerOperations from Frontend to Serialization) is:
1. libclangSerialization.so defines PCHContainerReader dtor, ...
2. clangFrontend and clangTooling define classes inheriting from PCHContainerReader, thus their DSOs have undefined references on PCHContainerReader dtor
3. Components depending on either clangFrontend or clangTooling cannot be linked unless they have explicit dependency on clangSerialization due to the default linker option -z defs. The explicit dependency could be avoided if libclang{Frontend,Tooling}.so had these undefined references.
This patch adds the explicit dependency on clangSerialization to make them build.
llvm-svn: 348915
As reported in PR39946, these two implementations cause stack overflows
to occur when a type recursively contains itself. While this only
happens when an incomplete version of itself is used by membership (and
thus an otherwise invalid program), the crashes might be surprising.
The solution here is to replace the recursive implementation with one
that uses a std::vector as a queue. Old values are kept around to
prevent re-checking already checked types.
Change-Id: I582bb27147104763d7daefcfee39d91f408b9fa8
llvm-svn: 348899
Bruno Ricci