Messages in this thread | | | Date | Sun, 28 Jun 2015 21:27:07 -0400 | Subject | Re: [PATCH v2 2/2] tty: check tcsetpgrp p is a process group | From | Patrick Donnelly <> |
| |
On Sun, Jun 28, 2015 at 12:07 PM, Peter Hurley <peter@hurleysoftware.com> wrote: > On 06/27/2015 08:51 PM, Patrick Donnelly wrote: >> This fixes a bug where a process can set the foreground process group to its >> pid even if its pid is not a valid pgrp. >> >> Signed-off-by: Patrick Donnelly <batrick@batbytes.com> >> --- >> drivers/tty/tty_io.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c >> index fbb55db..01b4769 100644 >> --- a/drivers/tty/tty_io.c >> +++ b/drivers/tty/tty_io.c >> @@ -2579,6 +2579,9 @@ static int tiocspgrp(struct tty_struct *tty, struct tty_struct *real_tty, pid_t >> retval = -ESRCH; >> if (!pgrp) >> goto out_unlock; >> + retval = -EINVAL; >> + if (!pid_task(pgrp, PIDTYPE_PGID)) >> + goto out_unlock; > > This change implies that the sequence in session_of_pgrp() that specifically > checks for pid_task(pgrp, PIDTYPE_PGID) == NULL is not doing anything > useful. However, that hypothesis is directly contradicted by the > comment above session_of_pgrp() > > "* This checks not only the pgrp, but falls back on the pid if no > * satisfactory pgrp is found. I dunno - gdb doesn't work correctly > * without this..." > > Regards, > Peter Hurley > >> retval = -EPERM; >> if (session_of_pgrp(pgrp) != task_session(current)) >> goto out_unlock; >>
Ah, missed that. Good catch! I guess this patch is no good since it was already accounted for and it breaks gdb.
-- Patrick Donnelly
| |