* [PATCH] Fix signal.h
@ 2004-11-20 17:09 Jakub Jelinek
2004-11-20 18:34 ` Jakub Jelinek
0 siblings, 1 reply; 3+ messages in thread
From: Jakub Jelinek @ 2004-11-20 17:09 UTC (permalink / raw)
To: Ulrich Drepper; +Cc: Glibc hackers
Hi!
I believe the nonnull attributes for __sighandler_t arguments are wrong.
SIG_DFL (which is ((__sighandler_t) 0) == NULL) is a valid argument
for these functions.
2004-11-20 Jakub Jelinek <jakub@redhat.com>
* signal/signal.h (__sysv_signal, sysv_signal, signal, bsd_signal,
ssignal): Remove __nonnull attribute.
--- libc/signal/signal.h.jj 2004-11-20 13:01:00.000000000 +0100
+++ libc/signal/signal.h 2004-11-20 18:02:54.613455520 +0100
@@ -78,10 +78,10 @@ typedef void (*__sighandler_t) (int);
the additional function `sysv_signal' when X/Open compatibility is
requested. */
extern __sighandler_t __sysv_signal (int __sig, __sighandler_t __handler)
- __THROW __nonnull ((2));
+ __THROW;
#ifdef __USE_GNU
extern __sighandler_t sysv_signal (int __sig, __sighandler_t __handler)
- __THROW __nonnull ((2));
+ __THROW;
#endif
/* Set the handler for the signal SIG to HANDLER, returning the old
@@ -90,13 +90,13 @@ extern __sighandler_t sysv_signal (int _
__BEGIN_NAMESPACE_STD
#ifdef __USE_BSD
extern __sighandler_t signal (int __sig, __sighandler_t __handler)
- __THROW __nonnull ((2));
+ __THROW;
#else
/* Make sure the used `signal' implementation is the SVID version. */
# ifdef __REDIRECT_NTH
extern __sighandler_t __REDIRECT_NTH (signal,
(int __sig, __sighandler_t __handler),
- __sysv_signal) __nonnull ((2));
+ __sysv_signal);
# else
# define signal __sysv_signal
# endif
@@ -107,7 +107,7 @@ __END_NAMESPACE_STD
/* The X/Open definition of `signal' conflicts with the BSD version.
So they defined another function `bsd_signal'. */
extern __sighandler_t bsd_signal (int __sig, __sighandler_t __handler)
- __THROW __nonnull ((2));
+ __THROW;
#endif
/* Send signal SIG to process number PID. If PID is zero,
@@ -132,7 +132,7 @@ __END_NAMESPACE_STD
#ifdef __USE_SVID
/* SVID names for the same things. */
extern __sighandler_t ssignal (int __sig, __sighandler_t __handler)
- __THROW __nonnull ((2));
+ __THROW;
extern int gsignal (int __sig) __THROW;
#endif /* Use SVID. */
Jakub
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] Fix signal.h
2004-11-20 17:09 [PATCH] Fix signal.h Jakub Jelinek
@ 2004-11-20 18:34 ` Jakub Jelinek
2004-11-20 18:43 ` Ulrich Drepper
0 siblings, 1 reply; 3+ messages in thread
From: Jakub Jelinek @ 2004-11-20 18:34 UTC (permalink / raw)
To: Ulrich Drepper; +Cc: Glibc hackers
On Sat, Nov 20, 2004 at 06:06:23PM +0100, Jakub Jelinek wrote:
> I believe the nonnull attributes for __sighandler_t arguments are wrong.
> SIG_DFL (which is ((__sighandler_t) 0) == NULL) is a valid argument
> for these functions.
BTW, is the nonnull attribute for first sigwait/sigwaitinfo/sigtimedwait
correct? I think so from reading the standards, but in that case I wonder
why we test it, e.g.:
static int
do_sigwaitinfo (const sigset_t *set, siginfo_t *info)
{
#ifdef SIGCANCEL
sigset_t tmpset;
if (set != NULL
...
Jakub
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] Fix signal.h
2004-11-20 18:34 ` Jakub Jelinek
@ 2004-11-20 18:43 ` Ulrich Drepper
0 siblings, 0 replies; 3+ messages in thread
From: Ulrich Drepper @ 2004-11-20 18:43 UTC (permalink / raw)
To: Jakub Jelinek; +Cc: Glibc hackers
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Jakub Jelinek wrote:
> I think so from reading the standards, but in that case I wonder
> why we test it, e.g.:
Should be non-NULL. I added the test in the userlevel code because I
wanted to provide the same error handling as if there would be no
userlevel part to the code. I.e., the kernel will return EINVAL or
EFAULT (don't know which).
- --
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA â–
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
iD8DBQFBn5BH2ijCOnn/RHQRAr3mAJ4nEArXNsoyoG2SucCIEAD13Yy9HACgmhzq
FI4HSCSh3PkEEoHeRtRSgW8=
=ljnI
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2004-11-20 18:43 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-11-20 17:09 [PATCH] Fix signal.h Jakub Jelinek
2004-11-20 18:34 ` Jakub Jelinek
2004-11-20 18:43 ` Ulrich Drepper
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).