lkml.org 
[lkml]   [2008]   [Dec]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [RFC][PATCH 4/6][v3] Define siginfo_from_ancestor_ns()
Eric W. Biederman [ebiederm@xmission.com] wrote:
| Oleg Nesterov <oleg@redhat.com> writes:
|
| >> I was going through the ->si_pid assignments to try and fix them at
| >> source (like the mqueue patch I sent last week).
| >
| > OK.
|
| Note. When a signal goes to a process group (or similar) we can't fix
| si_pid at the source. We have to fix it when only a single destination
| process is known. It doesn't mean that fixing it at the source
| is hopeless but...

Right. Most calls to kill_pgrp_info() and __kill_pgrp_info() use
SEND_SIG_PRIV or SEND_SIG_NOINFO, so their ->si_pid is set.
The only place seems to be the sys_kill()/sys_tkill().

|
| >> The two cases that don't fit the model are sys_kill() and sys_tkill().
| >> For that I was hoping we could use siginfo_from_user() again. i.e
| >>
| >> if (siginfo_from_user())
| >> masquerade_si_pid()
| >>
| >> in the default: case of send_signal(). To be safe, masquerade_si_pid()
| >> could do it only iff si_code is either SI_USER or SI_TKILL.
| >>
| >> IOW, with some tweaks, I am trying to see if we can use siginfo_from_user()
| >> in place of the SIG_FROM_USER.
| >
| > sys_rt_sigqueueinfo().
| >
| > But, perhaps we can just ignore the problems with sigqueueinfo() (and
| > document them).

Yes, thats one reason I was thinking of checking si_code == SI_USER or
SI_TKILL, but rt_sigqueueinfo() could still use those values too.

In fact it could use SI_ASYNCIO :-) and mess with sig_from_user() (i.e
if rt_sigqueueinfo() sets si_code to SI_ASYNCIO and sends SIGKILL/SIGSTOP
to a descendant cinit, SIGKILL/SIGSTOP will be ignored).

We could/should warn if rt_sigqueueinfo() sets si_code to SI_ASYNCIO.

One more reason to make SI_ASYNCIO a kernel signal I guess.

|
| Yes. I don't think si_pid is valid in that case anyway. It is the
| kernel signals where si_pid is a reliable field that are important.
|
| Eric

Sukadev


\
 
 \ /
  Last update: 2008-12-23 05:51    [W:0.041 / U:0.272 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site