netdev - Re: [patch net-next 01/16] net: add dev arg to ndo_neigh

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:	Tue, 5 Jul 2016 14:05:22 +0200
From:	Jiri Pirko <jiri@...nulli.us>
To:	Yuval Mintz <Yuval.Mintz@...gic.com>
Cc:	netdev <netdev@...r.kernel.org>,
	David Miller <davem@...emloft.net>,
	"idosch@...lanox.com" <idosch@...lanox.com>,
	"yotamg@...lanox.com" <yotamg@...lanox.com>,
	"eladr@...lanox.com" <eladr@...lanox.com>,
	"nogahf@...lanox.com" <nogahf@...lanox.com>,
	"ogerlitz@...lanox.com" <ogerlitz@...lanox.com>,
	"sfeldma@...il.com" <sfeldma@...il.com>,
	"roopa@...ulusnetworks.com" <roopa@...ulusnetworks.com>,
	"andy@...yhouse.net" <andy@...yhouse.net>,
	"dsa@...ulusnetworks.com" <dsa@...ulusnetworks.com>,
	"tgraf@...g.ch" <tgraf@...g.ch>,
	"jhs@...atatu.com" <jhs@...atatu.com>,
	"linville@...driver.com" <linville@...driver.com>,
	"ivecera@...hat.com" <ivecera@...hat.com>
Subject: Re: [patch net-next 01/16] net: add dev arg to
 ndo_neigh_construct/destroy

Tue, Jul 05, 2016 at 01:03:36PM CEST, Yuval.Mintz@...gic.com wrote:
>> As the following patch will allow upper devices to follow the call down lower
>> devices, we need to add dev here and not rely on n->dev.
>
>How does all of this relates to neigh_alloc()? Is it being called only for the
>lower devices?

This is called for all neighs.


>I'm asking as it appears that __neigh_create() would allocates memory
>based on some dev's neigh_priv_len, then propagate it via
>ndo_neigh_construct() [which you change in patch 2] - looks like the current
>implementers of this function assume the private entry is always theirs. E.g.,

neigh_priv_len does not take stacked devices into account.


>
>> -static int clip_constructor(struct neighbour *neigh)
>> +static int clip_constructor(struct net_device *dev, struct neighbour
>> +*neigh)
>>  {
>>  	struct atmarp_entry *entry = neighbour_priv(neigh);
>> 
>
>[Writing with absolute zero knowledge of this; perhaps this is all bollocks]

Better to study the code first before you comment, next time.