This is the mail archive of the libc-hacker@sources.redhat.com mailing list for the glibc project.

Note that libc-hacker is a closed list. You may look at the archives of this list, but subscription and posting are not open.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: new syscall stub support for ia64 libc


>>>>> On Fri, 14 Nov 2003 12:29:28 -0800, Ulrich Drepper <drepper@redhat.com> said:

  >> Which tests specifically?

  Uli> Tons of them.  tst-mutex8, tst-cond7, ...  Don't have the
  Uli> complete list.  The interesting part is that they abort, not
  Uli> crash.

Related question: it seems to me tst-cancel6 cannot possibly succeed
with NPTL _unless_ HAVE_FORCED_UNWIND is true?  The reason is that the
only way the file lock will be released is through a cleanup handler:

# ifdef __EXCEPTIONS
#  define _IO_acquire_lock(_fp) \
  do {									      \
    _IO_FILE *_IO_acquire_lock_file					      \
	__attribute__((cleanup (_IO_acquire_lock_fct)))			      \
	= (_fp);							      \
    _IO_flockfile (_IO_acquire_lock_file);

# else
#  define _IO_acquire_lock(_fp) _IO_acquire_lock_needs_exceptions_enabled
# endif

So here, if __EXCEPTIONS is defined, but HAVE_FORCED_UNWIND is false,
the compilation will succeed, but at runtime, the cleanup handler
won't get invoked.  Perhaps #ifdef __EXCEPTIONS should be changed to
#if defined(__EXCEPTIONS) && defined(HAVE_FORCED_UNWIND) ?  (Yeah,
it's a bit weird to have one but not the other defined, but it seems
to be the case on all existing Debian/unstable systems, so it's
perhaps not so rare.)

	--david


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]