d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fix tty_wait_background() to comply with standards.

Browse Source 
It turns out my handling of SIGTTOU and SIGTTIN didn't entirely comply
to the standards. It is true that in the SIGTTOU case we should not
return EIO when the signal is ignored/blocked, but in the SIGTTIN case
we must.

See also:	POSIX issue 7 section 11.1.4
This commit is contained in:
Ed Schouten 2009-04-08 15:56:50 +00:00
parent de24303241
commit e3b0b98073
1 changed files with 16 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
sys/kern/tty.c
View File

@ -371,23 +371,31 @@ tty_wait_background(struct tty *tp, struct thread *td, int sig)
* exit
* - the signal to send to the process isn't masked
*/
if (!tty_is_ctty(tp, p) ||
p->p_pgrp == tp->t_pgrp || p->p_flag & P_PPWAIT ||
SIGISMEMBER(p->p_sigacts->ps_sigignore, sig) ||
SIGISMEMBER(td->td_sigmask, sig)) {
if (!tty_is_ctty(tp, p) || p->p_pgrp == tp->t_pgrp) {
/* Allow the action to happen. */
PROC_UNLOCK(p);
return (0);
}
if (SIGISMEMBER(p->p_sigacts->ps_sigignore, sig) ||
SIGISMEMBER(td->td_sigmask, sig)) {
/* Only allow them in write()/ioctl(). */
PROC_UNLOCK(p);
return (sig == SIGTTOU ? 0 : EIO);
}
pg = p->p_pgrp;
if (p->p_flag & P_PPWAIT || pg->pg_jobc == 0) {
/* Don't allow the action to happen. */
PROC_UNLOCK(p);
return (EIO);
}
PROC_UNLOCK(p);
/*
* Send the signal and sleep until we're the new
* foreground process group.
*/
pg = p->p_pgrp;
PROC_UNLOCK(p);
if (pg->pg_jobc == 0)
return (EIO);
PGRP_LOCK(pg);
pgsignal(pg, sig, 1);
PGRP_UNLOCK(pg);