This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Add tests that backtrace and backtrace_symbols produce correctresults
On Sun, 20 Jan 2013, Andreas Schwab wrote:
> "Joseph S. Myers" <joseph@codesourcery.com> writes:
>
> > I'd expect there to be unwind information for read as it's a cancellation
> > point; the tests themselves are explicitly built with unwind information.
>
> This is not a multithreaded program, and read is invisible because it
> does not create a frame in that case.
There must surely be a call frame, where the instructions executed include
a syscall instruction and a subsequent return, and the return address is
in the caller of read, and the address of the instruction interrupted by
the signal should be within read.
For many architectures it's been considered generally useful to annotate
all .S sources, including syscalls, with unwind information whether or not
cancellation may occur there. But if it's not considered desirable to
have unwind information for all syscalls on some architecture for whatever
reason, then making the testcase multithreaded would seem a reasonable
workaround, if it does work.
--
Joseph S. Myers
joseph@codesourcery.com