linux-kernel - Re: [PATCH RT 1/5] allow preemption in add_timer

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] [day] [month] [year] [list]

Date:	Fri, 14 Feb 2014 13:34:27 +0100
From:	Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To:	Nicholas Mc Guire <der.herr@...r.at>
Cc:	linux-rt-users@...r.kernel.org,
	LKML <linux-kernel@...r.kernel.org>,
	Steven Rostedt <rostedt@...dmis.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Carsten Emde <C.Emde@...dl.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Andreas Platschek <platschek@....tuwien.ac.at>
Subject: Re: [PATCH RT 1/5] allow preemption in add_timer_randomness

* Nicholas Mc Guire | 2014-02-10 16:37:28 [+0100]:

>allow preemption in add_timer_randomness
>
>This patch replaced the preempt_disable by migrate_disable in 
>add_timer_randomness.
>
>Does this really need even migration protection at all ? 
>if one would even drop migration protection - what would happen ?
>
>        /* if over the trickle threshold, use only 1 in 4096 samples */
>        if (input_pool.entropy_count > trickle_thresh &&
>            ((__this_cpu_inc_return(trickle_count) - 1) & 0xfff)) {
>                return;
>        }
>
>trickle_thresh and input_pool.entropy_count are global, so the only thing
>that would happen if this got migrated would be that the 1/4096 could be
>a bit less precise locally (the probability of being migrted here is not 
>very high this is a window of a few instructions at best) 
>
>If we got migrated away - so what ? that only would mean that it would be
>checking the trickle_count on the "wrong" cpu - in sum the cpus would though
>still not contribute more bits, the one countdown wouuld speed up only as
>much as some other countdown slowed down.
>
>In any case, even for precise 1/4096 when over threshhold we would not need
>more than a migration protection.
>
>patch is against 3.12.10-rt15
>
>Signed-off-by: Nicholas Mc Guire <der.herr@...r.at>

Ingo added this preempt_disable() in v2.6.9-rc4 because:
|A certain codepath in the random driver relied on vt_ioctl() being under
|the BKL and implicitly disabling preemption.  The code wasn't buggy
|upstream but it's slighly unrobust so I think we want the fix upstream too,
|independently of the remove-bkl patch.

later in -RT he made the preempt disable region smaller. Looking at the
code path I'm not sure if we need that preempt_disable() here or
upstream at all. It looks like a global lock which does not protect
much.

Applied.

Sebastian
--
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/