On Wed, Nov 23, 2005 at 05:24:05PM +0300, Oleg Nesterov wrote: > Use symbolic names instead of hardcoded constants. I'm more than happy with that. Seems like the original code in check_kill_permission() was changed in a similar way after I made that _as_uid() function. I don't think that I am a person with any say in the signal.c code,but in case of doubt: Acked-by: Harald Welte <[email protected]> > Signed-off-by: Oleg Nesterov <[email protected]> > > --- 2.6.15-rc2/kernel/signal.c~1_KUID 2005-11-22 19:35:52.000000000 +0300 > +++ 2.6.15-rc2/kernel/signal.c 2005-11-23 19:17:35.000000000 +0300 > @@ -1163,8 +1163,7 @@ int kill_proc_info_as_uid(int sig, struc > ret = -ESRCH; > goto out_unlock; > } > - if ((!info || ((unsigned long)info != 1 && > - (unsigned long)info != 2 && SI_FROMUSER(info))) > + if ((info == SEND_SIG_NOINFO || (!is_si_special(info) && SI_FROMUSER(info))) > && (euid != p->suid) && (euid != p->uid) > && (uid != p->suid) && (uid != p->uid)) { > ret = -EPERM; -- - Harald Welte <[email protected]> http://gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)
Attachment:
pgpJWinynMJ7C.pgp
Description: PGP signature
- References:
- [PATCH] kill_proc_info_as_uid: don't use hardcoded constants
- From: Oleg Nesterov <[email protected]>
- [PATCH] kill_proc_info_as_uid: don't use hardcoded constants
- Prev by Date: Re: [PATCH] Fix USB suspend/resume crasher
- Next by Date: Re: Linux 2.6.15-rc2
- Previous by thread: [PATCH] kill_proc_info_as_uid: don't use hardcoded constants
- Next by thread: [PATCH] do_coredump() should reset group_stop_count earlier
- Index(es):
 |