linux-kernel - Re: [RFC] ps command race fix

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 24 Jul 2006 18:20:00 -0700
From:	Andrew Morton <akpm@...l.org>
To:	KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>
Cc:	linux-kernel@...r.kernel.org,
	"Eric W. Biederman" <ebiederm@...ssion.com>
Subject: Re: [RFC] ps command race fix

On Fri, 14 Jul 2006 20:39:39 +0900
KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com> wrote:

> Hi, this is an experimental patch for the probelm
> 	- "ps command can miss some pid occationally"
> please comment 
> 
> 
> the problem itself is very rare case, but the result is sometimes terrible
> 
> for example, when a user does
> 
> alive=`ps | grep command | grep -v command | wc -l`
> 
> to check process is alive or not (I think this user should use kill-0 ;)
> 
> -Kame
> ==
> Now, prod_pid_readir() uses direct access to task and 
> indexing 'task list' as fallback.
> Of course, entries in this list can be removed randomly.
> 
> So, following can happen when using 'ps' command.
> ==
> 1. assume task_list as
> ....-(taskA)-(taskB)-(taskC)-(taskD)-(taskE)-(taskF)-(taskG)-...
> 
> 2. at getdents() iteration 'N', ps command's getdents() read entries before taskC.
> and remenbers "I read X entries".
> 
> ....-(taskA)-(taskB)-(taskC)-(taskD)-(taskE)-(taskF)-(taskG)-...
> ------(f_pos=X)---------^
> 
> getdents() remembers
> 	- "taskC is next candidate to be read"
> 	- "we already read X ents".
> 
> 3. consider taskA and taskC exits, before next getdents(N+1)
> 
> ....-(lost)-(taskB)-(lost)-(taskD)-(taskE)-(taskF)-(taskG)-...
> ------(f_pos=X)--------^
> 
> 4. at getdents(N+1), becasue getdents() cannot find taskC, it skips 'X'
>    ents in the list.
>    from head of the list.
> ....-(taskB)-(taskD)-(taskE)-(taskF)-(taskG)-..
> ------(f_pos=X)--------^
> 
> 5. in this case, taskD is skipped.
> ==
> 
> This patch changes indexing in the list to indexing in a table.
> Table is created only for storing valid tgid.(not pid)
> Tested on x86/ia64.
> 

It allocates a potentially-significant amount of memory per-task, until
that tasks exits (we could release it earlier, but the problem remains) and
it adds yet another global lock in the process exit path.

>  5 files changed, 138 insertions(+), 62 deletions(-)

And it adds complexity and code.

So I think we're still seeking a solution to this.

Options might be:

a) Pin the most-recently-visited task in some manner, so that it is
   still on the global task list when we return.  That's fairly simple to
   do (defer the release_task()) but it affects task lifetime and visibility
   in rare and worrisome ways.

b) Change proc_pid_readdir() so that it walks the pid_hash[] array
   instead of the task list.  Need to do something clever when traversing
   each bucket's list, but I'm not sure what ;) It's the same problem.

   Possibly what we could do here is to permit the task which is walking
   /proc to pin a particular `struct pid': take a ref on it then when we
   next start walking one of the pid_hash[] chains, we _know_ that the
   `struct pid' which we're looking for will still be there.  Even if it
   now refers to a departed process.

c) Nuke the pid_hash[], convert the whole thing to a radix-tree. 
   They're super-simple to traverse.  Not sure what we'd index it by
   though.

I guess b) is best.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/