| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
| |
llvm-svn: 207170
|
|
|
|
| |
llvm-svn: 207169
|
|
|
|
|
|
|
| |
Fixes compiler complaint about:
src/Unwind/UnwindCursor.hpp:366:25: error: 'libunwind::AbstractUnwindCursor' has virtual functions but non-virtual destructor [-Werror,-Wnon-virtual-dtor]
llvm-svn: 206942
|
|
|
|
|
|
|
| |
The was working because, given __APPLE__, _LIBUNWIND_BUILD_SJLJ_APIS was set to
__arm__, but other ARM targets not using SJ/LJ will fail to compile.
llvm-svn: 206941
|
|
|
|
|
|
|
| |
Darwin and the BSDs define a tagged struct dl_info typedef'd as Dl_info. glibc
and bionic typedef an anonymous struct as Dl_info.
llvm-svn: 206939
|
|
|
|
|
|
| |
... so that we can use newlib as the c library.
llvm-svn: 206427
|
|
|
|
| |
llvm-svn: 206122
|
|
|
|
| |
llvm-svn: 206121
|
|
|
|
| |
llvm-svn: 206108
|
|
|
|
|
|
| |
in the wild on SH3.
llvm-svn: 205756
|
|
|
|
| |
llvm-svn: 205573
|
|
|
|
| |
llvm-svn: 204292
|
|
|
|
|
|
|
|
|
|
|
| |
opcode is VAX. A function call pushes the number of arguments given onto
the stack and "ret" will pop it automatically. The FDE of the caller
contains the amount of stack space used for arguments (and possibly
extra padding), so unwinding has to compensate for this when "returning"
from a function. This is exactly the case when step() is done. The
existing handling in unw_set_reg no longer makes sense.
llvm-svn: 204290
|
|
|
|
|
|
|
| |
programs. Initialize lsdaEncoding to DW_EH_PE_omit and check for that
value to decide whether a value should be decoded.
llvm-svn: 203626
|
|
|
|
| |
llvm-svn: 201533
|
|
|
|
|
| |
review: http://llvm-reviews.chandlerc.com/D2718
llvm-svn: 201208
|
|
|
|
|
|
| |
pointers. Specifically, libc++abi would crash when you tried it.
llvm-svn: 200904
|
|
|
|
| |
llvm-svn: 200864
|
|
|
|
| |
llvm-svn: 199941
|
|
|
|
| |
llvm-svn: 199915
|
|
|
|
| |
llvm-svn: 199860
|
|
|
|
| |
llvm-svn: 199859
|
|
|
|
| |
llvm-svn: 198643
|
|
|
|
|
|
| |
unwinding.
llvm-svn: 197591
|
|
|
|
|
|
|
| |
This is in preparation for landing an implementation of unw_getcontext
on a system where it's mangled 'unw_getcontext', not '_unw_getcontext'.
llvm-svn: 197523
|
|
|
|
|
|
| |
parameter packs, which exposed a logic bug causing an empty vector<string> to be accessed with back(). In addition to fixing the bug, I've inserted numerous preemptive checks for similar bugs in the hopes that if another bug is uncovered, the bug results in an invalid mangled string instead of a demangler crash. Test suite updated with string that was causing the crash.
llvm-svn: 197063
|
|
|
|
|
|
| |
These files require preprocessing, so their name should end with a capital S.
llvm-svn: 196555
|
|
|
|
|
|
| |
These files require preprocessing, so their name should end with a capital S.
llvm-svn: 196554
|
|
|
|
|
|
|
| |
According to kledzik, this constructor isn't live code, so no functionality
change.
llvm-svn: 196550
|
|
|
|
| |
llvm-svn: 196436
|
|
|
|
|
|
| |
Align to 8 spaces instead of an inconsistent 9.
llvm-svn: 195974
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
type_info* will work for typeids from the same compiled file but fail
for typeids from a DLL and an executable. Among other things, exceptions
are not caught by handlers since can_catch() returns false.
Defining _LIBCXX_DYNAMIC_FALLBACK does not help since can_catch() calls
is_equal() with use_strcmp=false so the string names are not compared.
This patch compares typeids first (cheap) and only they are different
calls strcmp.
llvm-svn: 195502
|
|
|
|
|
|
|
|
| |
void syslog(int facility_priority, const char* format, ...);
This is indeed how it is used in private_typeinfo.cpp.
llvm-svn: 195452
|
|
|
|
|
|
| |
that it gets into the binary.
llvm-svn: 194208
|
|
|
|
| |
llvm-svn: 193840
|
|
|
|
| |
llvm-svn: 193169
|
|
|
|
|
|
| |
parameter to match other implementations
llvm-svn: 192248
|
|
|
|
| |
llvm-svn: 192245
|
|
|
|
|
|
| |
architectures
llvm-svn: 192241
|
|
|
|
| |
llvm-svn: 192180
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
realized, it is not complete. It relies on some _Unwind_* functions to be
supplied by the OS. That means it cannot be ported to platforms that don’t
already have an unwinder.
Years ago Apple wrote its own unwinder for MacOSX and iOS. To make libcxxabi
complete, Apple has decided the source code for its unwinder can be contributed
to the open source LLVM libcxxabi project, with a dual licensed under LLVM
and MIT license.
So, I’ve spent some time cleaning up the sources to make them conform with
LLVM style and to conditionalize the sources in a way that should make it
easier to port to other platforms. The sources are in a separate "Unwind"
directory under "src" in libcxxabi.
Background:
Most architectures now use "zero cost" exceptions for C++. The zero cost means
there are no extra instructions executed if no exceptions are thrown. But if
an exception is thrown, the runtime must consult side tables and figure out how
to restore registers and "unwind" from the current stack frame to the catch
clause. That ability to modify the stack frames and cause the thread to resume
in a catch clause with all registers restored properly is the main purpose
of the unwinder.
This unwinder has two levels of API. The high level APIs are the _Unwind_*
functions which the cxa_* exception functions in libcxxabi require. The low
level APIs are the unw_* functions which are an interface defined by the the
old HP libunwind project (which shares no code with this unwinder).
llvm-svn: 192136
|
|
|
|
| |
llvm-svn: 190479
|
|
|
|
| |
llvm-svn: 189603
|
|
|
|
|
|
| |
fixes http://llvm.org/bugs/show_bug.cgi?id=16996.
llvm-svn: 189194
|
|
|
|
| |
llvm-svn: 187255
|
|
|
|
|
|
| |
case that led me to yet another closely related test case that the current design could not handle. I've now changed the way forward references are handled completely. It wasn't that much code to change. The demangler, when confronted with a forward reference to a template parameter, now parses things twice. During the second parse, all forward references are remembered from the first parse. Test suite updated with new case.
llvm-svn: 184672
|
|
|
|
|
|
| |
I believe this fix is a better fix than what I committed in r184656 yesterday. I've basically moved the checking for '`' from the start of the demangling process to the end of it. In the process I discovered that one of the test cases no longer demangled to the expected string. After further investigation I believe this case to not be a valid mangled string, and so I moved the test case to the 'invalid cases'. The reason I believe it is invalid is that it should use T_ instead of T0_ to index the template parameter.
llvm-svn: 184668
|
|
|
|
| |
llvm-svn: 184656
|
|
|
|
|
|
| |
random strings at the demangler. I succeeded in crashing it twice more and those crashers have been fixed and the test suite updated with the crash cases.
llvm-svn: 184562
|
|
|
|
| |
llvm-svn: 184477
|