netdev - Re: [PATCH net] net: sch_red: Fix race between timer and red

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]

Message-ID: <CAGK4HS8SoXAGM-zgbofQp+Cy19Z0doQafvYceePt_kM9CEchYQ@mail.gmail.com>
Date:	Fri, 8 Nov 2013 13:59:19 -0800
From:	Vijay Subramanian <subramanian.vijay@...il.com>
To:	David Miller <davem@...emloft.net>
Cc:	netdev <netdev@...r.kernel.org>,
	Stephen Hemminger <shemminger@...tta.com>,
	Eric Dumazet <eric.dumazet@...il.com>
Subject: Re: [PATCH net] net: sch_red: Fix race between timer and red_destroy()

> del_timer_sync() waits for all instances of the timer handler which
> are running to complete, and then it deletes the timer from any of
> the scheduled timer lists.
>
> Your patch is making the mod_timer() in the timer handler conditional,
> and this really shouldn't be necessary.
>
> The whole point of del_timer_sync() is to address these kinds of
> situations.

Thanks for the explanation Dave. I am trying to reconcile your
explanation with the text above del_timer_sync()
vim +1013 kernel/timer.c

" * This function only differs from del_timer() on SMP: besides deactivating
 * the timer it also makes sure the handler has finished executing on other
 * CPUs.
 *
 * Synchronization rules: Callers must prevent restarting of the timer,
 * otherwise this function is meaningless.  ..."

This seems to suggest simply calling del_timer_sync() is not enough.
In fact, for the same reason  I was recently pointed towards
commit 980c478ddbb72 (sch_sfq: use del_timer_sync() in sfq_destroy() )
and I based the patch on how q->perturb_period is reset in
sfq_destroy() and used in sfq_perturbation().

Based on your explanation , I think we do not need (in sched/sch_sfq.c)
q->perturb_period = 0;  in sfq_destroy() and just del_timer_sync()
should be enough.

Thanks,
Vijay
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html