Reid Kleckner [Tue, 26 Mar 2013 18:30:28 +0000 (18:30 +0000)]
[ms-cxxabi] Give the MS inheritance attributes a base class
Required making a handful of changes to the table generator. Also adds
an unspecified inheritance attribute. This opens the path for us to
apply these attributes to C++ records implicitly.
Michael Liao [Tue, 26 Mar 2013 17:52:08 +0000 (17:52 +0000)]
Add PRFCHW intrinsic support
- Add head 'prfchwintrin.h' to define '_m_prefetchw' which is mapped to
LLVM/clang prefetch builtin
- Add option '-mprfchw' to enable PRFCHW feature and pre-define '__PRFCHW__'
macro
[Preprocessor/Modules] Separate the macro directives kinds into their own MacroDirective's subclasses.
For each macro directive (define, undefine, visibility) have a separate object that gets chained
to the macro directive history. This has several benefits:
-No need to mutate a MacroDirective when there is a undefine/visibility directive. Stuff like
PPMutationListener become unnecessary.
-No need to keep extra source locations for the undef/visibility locations for the define directive object
(which is the majority of the directives)
-Much easier to hide/unhide a section in the macro directive history.
-Easier to track the effects of the directives across different submodules.
Reid Kleckner [Tue, 26 Mar 2013 16:56:59 +0000 (16:56 +0000)]
[ms-cxxabi] Mangle vector types
Summary:
The only vector types a user can pass from MSVC code to clang code are
the ones from *mmintrin.h, so we only have to match the MSVC mangling
for these types. MSVC mangles the __m128 family of types as tag types,
which we match. For other vector types, we emit a unique tag type
mangling that won't match anything produced by MSVC.
Evgeniy Stepanov [Tue, 26 Mar 2013 13:44:29 +0000 (13:44 +0000)]
Fix uninitialized read of CalleeWithThisReturn.
CalleeWithThisReturn can be left initialized if HasThisReturn() is false.
This change reverses the order of checks in EmitFunctionEpilog such that
CalleeWithThisReturn is only examined when it has a meaningful value.
Chandler Carruth [Tue, 26 Mar 2013 03:45:48 +0000 (03:45 +0000)]
Manually specify the link dependencies. Turns out that all the work on
LLVMBuild stuff didn't actually provide a single place for dependencies,
it just added a third place.
Richard Smith [Tue, 26 Mar 2013 01:15:19 +0000 (01:15 +0000)]
Implement special-case name lookup for inheriting constructors: member
using-declarations with names which look constructor-like are interpreted as
constructor names.
Nick Lewycky [Mon, 25 Mar 2013 21:24:30 +0000 (21:24 +0000)]
Try harder to be signal-safe inside our signal handler. The most prominent behavioural
difference is that we no longer clean the token before emitting it. This fixes a bug where
clang hangs in the middle of crashing because the crash handler calls malloc from inside
a crash that happened inside of free.
Nadav Rotem [Sat, 23 Mar 2013 06:43:35 +0000 (06:43 +0000)]
Make clang to mark static stack allocations with lifetime markers to enable a more aggressive stack coloring.
Patch by John McCall with help by Shuxin Yang.
rdar://13115369
Bob Wilson [Sat, 23 Mar 2013 05:17:59 +0000 (05:17 +0000)]
Revert svn r176894 and r177658.
Changing -ccc-install-dir to affect cc1's resource-dir setting broke our
internal LNT tests. After discussing the situation with Jim, we've decided to
pursue an alternate approach. We really want the resource-dir to be located
relative to clang, even when using -ccc-install-dir, but we're going to
add a fallback setting for the libc++ headers if they don't exist alongside
the compiler.
John McCall [Sat, 23 Mar 2013 02:35:54 +0000 (02:35 +0000)]
Under ARC, when we're passing the address of a strong variable
to an out-parameter using the indirect-writeback conversion,
and we copied the current value of the variable to the temporary,
make sure that we register an intrinsic use of that value with
the optimizer so that the value won't get released until we have
a chance to retain it.
Richard Trieu [Sat, 23 Mar 2013 01:38:36 +0000 (01:38 +0000)]
Strip off local qualifiers when converting from RecordType to
TemplateSpecializationType during template type diffing. This allows the
correct printing of diffing qualifiers on templates.
Jordan Rose [Sat, 23 Mar 2013 01:21:33 +0000 (01:21 +0000)]
[analyzer] Teach constraint managers about unsigned comparisons.
In C, comparisons between signed and unsigned numbers are always done in
unsigned-space. Thus, we should know that "i >= 0U" is always true, even
if 'i' is signed. Similarly, "u >= 0" is also always true, even though '0'
is signed.
Part of <rdar://problem/13239003> (false positives related to std::vector)
Jordan Rose [Sat, 23 Mar 2013 01:21:29 +0000 (01:21 +0000)]
[analyzer] Loc-Loc operations (subtraction or comparison) produce a NonLoc.
For two concrete locations, we were producing another concrete location and
then casting it to an integer. We should just create a nonloc::ConcreteInt
to begin with.
Jordan Rose [Sat, 23 Mar 2013 01:21:26 +0000 (01:21 +0000)]
[analyzer] CmpRuns.py: Accept single files as input.
This allows us to compare two direct invocations of the analyzer on a
single source file without having to wrap the output plists in their
own directories.
Jordan Rose [Sat, 23 Mar 2013 01:21:16 +0000 (01:21 +0000)]
[analyzer] Translate "a != b" to "(b - a) != 0" in the constraint manager.
Canonicalizing these two forms allows us to better model containers like
std::vector, which use "m_start != m_finish" to implement empty() but
"m_finish - m_start" to implement size(). The analyzer should have a
consistent interpretation of these two symbolic expressions, even though
it's not properly reasoning about either one yet.
The other unfortunate thing is that while the size() expression will only
ever be written "m_finish - m_start", the comparison may be written
"m_finish == m_start" or "m_start == m_finish". Right now the analyzer does
not attempt to canonicalize those two expressions, since it doesn't know
which length expression to pick. Doing this correctly will probably require
implementing unary minus as a new SymExpr kind (<rdar://problem/12351075>).
For now, the analyzer inverts the order of arguments in the comparison to
build the subtraction, on the assumption that "begin() != end()" is
written more often than "end() != begin()". This is purely speculation.
Jordan Rose [Sat, 23 Mar 2013 01:21:05 +0000 (01:21 +0000)]
[analyzer] Use SymExprs to represent '<loc> - <loc>' and '<loc> == <loc>'.
We just treat this as opaque symbols, but even that allows us to handle
simple cases where the same condition is tested twice. This is very common
in the STL, which means that any project using the STL gets spurious errors.
documentation parsing: when providing code completion comment
for a getter used in property-dot syntax, if geter has its own
comment use it. // rdar://12791315
Richard Smith [Sat, 23 Mar 2013 00:30:08 +0000 (00:30 +0000)]
If a .syms file is available alongside a sanitizer runtime, pass it to the
linker via --dynamic-list instead of using --export-dynamic. This reduces the
size of the dynamic symbol table, and thus of the binary (in some cases by up
to ~30%).
Bill Wendling [Fri, 22 Mar 2013 21:33:46 +0000 (21:33 +0000)]
Use RequireCompleteType() instead of isIncompleteType().
isIncompleteType() returns true or false for template types depending on whether
the type is instantiated yet. In this context, that's arbitrary. The better way
to check for a complete type is RequireCompleteType().
Jordan Rose [Fri, 22 Mar 2013 21:15:33 +0000 (21:15 +0000)]
Revert "[analyzer] Break cycles (optionally) when trimming an ExplodedGraph."
The algorithm used here was ridiculously slow when a potential back-edge
pointed to a node that already had a lot of successors. The previous commit
makes this feature unnecessary anyway.
Jordan Rose [Fri, 22 Mar 2013 21:15:28 +0000 (21:15 +0000)]
[analyzer] Use a forward BFS instead of a reverse BFS to find shortest paths.
For a given bug equivalence class, we'd like to emit the report with the
shortest path. So far to do this we've been trimming the ExplodedGraph to
only contain relevant nodes, then doing a reverse BFS (starting at all the
error nodes) to find the shortest paths from the root. However, this is
fairly expensive when we are suppressing many bug reports in the same
equivalence class.
r177468-9 tried to solve this problem by breaking cycles during graph
trimming, then updating the BFS priorities after each suppressed report
instead of recomputing the whole thing. However, breaking cycles is not
a cheap operation because an analysis graph minus cycles is still a DAG,
not a tree.
This fix changes the algorithm to do a single forward BFS (starting from the
root) and to use that to choose the report with the shortest path by looking
at the error nodes with the lowest BFS priorities. This was Anna's idea, and
has the added advantage of requiring no update step: we can just pick the
error node with the next lowest priority to produce the next bug report.
[PCH/Modules] De/Serialize MacroInfos separately than MacroDirectives.
-Serialize the macro directives history into its own section
-Get rid of the macro updates section
-When de/serializing an identifier from a module, associate only one macro per
submodule that defined+exported it.
Reid Kleckner [Fri, 22 Mar 2013 19:02:54 +0000 (19:02 +0000)]
[ms-cxxabi] Implement member data pointers for non-dynamic classes
Summary:
For non-dynamic classes (no virtual bases), member data pointers are
simple offsets from the base of the record. Dynamic classes use an
aggregate for member data pointers and are therefore currently
unsupported.
Unlike Itanium, the ms ABI uses 0 to represent null for polymorphic
classes. Non-polymorphic classes use -1 like Itanium, since 0 is a
valid field offset.
Douglas Gregor [Fri, 22 Mar 2013 18:50:14 +0000 (18:50 +0000)]
<rdar://problem/13479539> Simplify ModuleManager/GlobalModuleIndex interaction to eliminate a pile of extraneous stats().
The refactoring in r177367 introduced a serious performance bug where
the "lazy" resolution of module file names in the global module index
to actual module file entries in the module manager would perform
repeated negative stats(). The new interaction requires the module
manager to inform the global module index when a module file has been
loaded, eliminating the extraneous stat()s and a bunch of bookkeeping
on both sides.
documentation parsing. Provide code completion comment
for self.GetterName where GetterName is the getter method
for a property with name different from the property name
(declared via a property getter attribute) // rdar://12791315
Daniel Jasper [Fri, 22 Mar 2013 10:01:29 +0000 (10:01 +0000)]
Make clang-format understand more line comments.
Apparently one needs to set LangOptions.LineComment.
Before "//* */" got reformatted to "/ /* */" as the lexer was returning
the token sequence (slash, comment). This could also lead to weird other
stuff, e.g. for people that like to using comments like:
//****************
John McCall [Fri, 22 Mar 2013 02:10:40 +0000 (02:10 +0000)]
Fix a crash-on-valid where a block capture copy expression was
picking up cleanups from earlier in the statement. Also fix a
crash-on-invalid where a reference to an invalid decl from an
enclosing scope was causing an expression to fail to build, but
only *after* a cleanup was registered from that statement,
causing an assertion downstream.
Richard Smith [Fri, 22 Mar 2013 00:47:07 +0000 (00:47 +0000)]
ubsan: Pass floating-point arguments to the runtime by value if they fit the
value argument. If not, be sure we don't accidentally use a dynamic alloca.
Douglas Gregor [Fri, 22 Mar 2013 00:10:49 +0000 (00:10 +0000)]
<rdar://problem/13479214> Make Clang's <stddef.h> robust against system headers defining size_t/ptrdiff_t/wchar_t.
Clang's <stddef.h> provides definitions for the C standard library
types size_t, ptrdiff_t, and wchar_t. However, the system's C standard
library headers tend to provide the same typedefs, and the two
generally avoid each other using the macros
_SIZE_T/_PTRDIFF_T/_WCHAR_T. With modules, however, we need to see
*all* of the places where these types are defined, so provide the
typedefs (ignoring the macros) when modules are enabled.
Douglas Gregor [Thu, 21 Mar 2013 21:48:48 +0000 (21:48 +0000)]
<rdar://problem/13477190> Give the Clang module cache directory some structure, so it's easier to find.
We now put the Clang module cache in
<system-temp-directory>/org.llvm.clang/ModuleCache. Perhaps some day
there will be other caches under <system-temp-directory>/org.llvm.clang>.
Objective-C: Tighten the rules when warning
is issused for on overriding 'readwrite'
property which is not auto-synthesized.
Buttom line is that if hueristics determine
that there will be a user implemented setter,
no warning will be issued. // rdar://13388503