netdev - Re: [PATCH 00/10] pull request for net-next: batman-adv 2019-06-27 v2

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: <5314018.Pl9QOMrv0R@sven-edge>
Date:   Fri, 28 Jun 2019 19:01:41 +0200
From:   Sven Eckelmann <sven@...fation.org>
To:     b.a.t.m.a.n@...ts.open-mesh.org
Cc:     David Miller <davem@...emloft.net>, sw@...onwunderlich.de,
        netdev@...r.kernel.org,
        Linus Lüssing <linus.luessing@...3.blue>
Subject: Re: [PATCH 00/10] pull request for net-next: batman-adv 2019-06-27 v2

On Friday, 28 June 2019 18:49:05 CEST David Miller wrote:
[...]
> I think that when you have the read_lock held, RCU is not necessary in order
> to use __in6_dev_get() but I may be mistaken.  Just FYI...

Problem is that the read_lock() can only be used after the __in6_dev_get() 
finished sucessfully. Because the read_lock's lock is stored in the inet6_dev 
object which was retrieved via __in6_dev_get. And the __in6_dev_get kerneldoc 
states [1] that you either have to hold RTNL or RCU (see also the 
rcu_dereference_rtnl call inside this function).

So we can only drop the rcu_read_lock when RTNL lock is held . I would guess 
now that this is not the case here - Linus' might want to correct me.

Kind regards,
	Sven

[1] https://lxr.missinglinkelectronics.com/linux/include/net/addrconf.h#L335
Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)