Check if info.dli_sname is NULL,
If the image containing addr is found, but no nearest symbol was found, the dli_sname and dli_saddr fields are set to NULL.
Fallback to using file offsets if no symbol is found, like we do if the
object could not be opened.
This makes backtraces usable even if objects in the trace are stripped,
since the trace can be symbolized by post-processing it with a tool like
asan_symbolize.py.
Note that this is not currently compatible with SymbolizeCallback as
this overwrites the filename in the buffer. The behavior is unchanged
in that case.
Closes: #514
* Upstream Chromium local changes to symbolize.cc
Chromium has its own fork of symbolize.cc and symbolize.h, and it has
several not-yet-upstreamed changes, that are not specific to Chromium.
This patch upstreams such a changes.
Fixed google::FindSymbol reading past end of a section
3dae0a2d7d
Fix -INT_MIN integer overflow in itoa_r().
ac4d28e9cb
Add print_unsymbolized_stack_traces gn arg.
6a2726776f
Switch to standard integer types in base/.
9b6f42934e
* Add a build option to print unsymbolized traces
This PRINT_UNSYMBOLIZED_STACK_TRACES option to cmake, and
--enable-unsymbolized-traces option to autoconf.
ReadFromOffset in symbolize.cc used to call lseek() + read() to read
data from fd. However, the fd may be reused for multiple symbolize
requests from multiple threads, and causes a race around the fd read
offset.
This updates it to use pread() to resolve the race.
* added dbghelp symbolizer support for mingw and cygwin
* fixed compiler errors in case <stdint.h> is not available
* cmake: check whether SymFromAddr actually works
We previously had logic to compute the base address from program
headers as part of symbolization. The problem is that we need a correct
base address earlier in order to adjust a PC into the image's address
space, as these addresses can appear in unsymbolized output.
There was previously an assumption that only the mapping that
was lowest in the address space did not need to be adjusted. This
assumption is not guaranteed (for example, the kernel may choose to
map an ET_DYN lowest) and in fact turned out to be wrong in binaries
linked with lld because the first mapping is read-only.
The solution is to move the program header reading logic into the
code that reads /proc/self/maps.
There is a change in semantics for clients that install a callback
using the InstallSymbolizeOpenObjectFileCallback function. Any such
clients will need to return a correct base address from the callback
by reading program headers using code similar to that in the function
OpenObjectFileContainingPcAndGetStartAddress.
Previously we were using a module's "start address", i.e. the
address at which the module's executable region was mapped, as the
zero virtual address, i.e. the address from which the DSO's virtual
addresses are calculated. This works fine for DSOs created by the
bfd and gold linkers, which will emit a PT_LOAD directive into the
program header which loads the executable region at virtual address
(p_vaddr) and file offset (p_offset) 0.
However, the lld linker may place a read-only region before the
executable region, meaning that both p_vaddr and p_offset for the
executable region are non-zero. This means that any symbols resolved
by the symbolizer are resolved to an incorrect virtual address. To
correctly calculate the address corresponding to virtual address zero,
we need to take into account p_vaddr and p_offset.
Specifically, the calculation starts with the "base address", i.e. the
start address minus the file offset. To get from the base address to
virtual address zero, we first add p_offset. This gives us the mapped
address of the start of the segment, or in other words the mapped
address corresponding to the virtual address of the segment. (Note
that this is distinct from the start address, as p_offset is not
guaranteed to be page aligned.) We then subtract p_vaddr, which takes
us to virtual address zero.
their own implementation of OpenObjectFileContainingPcAndGetStartAddress.
GTTF: Make stack trace symbolization code recognize more symbols.
One example is __libc_start_main (easy to check), there may be more.
Include the correct elf header on OpenBSD and Android.
Fix Coverity NO_EFFECT defect. Comparing unsigned greater than or equal zero is always true.
git-svn-id: https://google-glog.googlecode.com/svn/trunk@140 eb4d4688-79bd-11dd-afb4-1d65580434c0
has execinfo.h (e.g., MacOSX 10.5). Though dladdr may not be async
signal safe, it's OK since glog's stacktrace doesn't depend on signals.
git-svn-id: https://google-glog.googlecode.com/svn/trunk@7 eb4d4688-79bd-11dd-afb4-1d65580434c0
