netdev - Re: [PATCH net-next] rhashtable: Make sure max

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]

Message-ID: <bb1475bc-a129-865f-be43-829a76e99dc9@gmail.com>
Date:   Thu, 27 Apr 2017 15:32:41 -0700
From:   Florian Fainelli <f.fainelli@...il.com>
To:     netdev@...r.kernel.org
Cc:     davem@...emloft.net, herbert@...dor.apana.org.au, fw@...len.de,
        tgraf@...g.ch
Subject: Re: [PATCH net-next] rhashtable: Make sure max_size is non zero

On 04/27/2017 03:28 PM, Florian Fainelli wrote:
> After commit 6d684e54690c ("rhashtable: Cap total number of
> entries to 2^31"), we would be hitting a panic() in net/core/rtnetlink.c
> during initialization. The call stack would look like this:
> 
> register_pernet_subsys()
>     ...
>     ops->init()
> 	rtnetlink_net_init()
> 	  netlink_kernel_create()
> 	     netlink_insert()
> 		__netlink_insert()
> 		   rhashtable_lookup_insert_key()
> 		      __rhashtable_insert_fast()
> 			rht_grow_above_max()
> 
> And here, we have rht_grow_above_max() return true, because ht->nelemts = 0 (legit)
> && ht->max_elems = 0 (looks bogus).
> 
> Eventually, we would be return -E2BIG from __rhashtable_insert_fast()
> and propagate this all the way back to the caller.
> 
> After commit 6d684e54690c what changed is that we would take the
> following condition:
> 
> if (ht->p.max_size < ht->max_elems / 2)
> 	ht->max_elems = ht->p.max_size * 2;
> 
> and since ht->p.max_size = 0, we would set ht->max_elems to 0 as well.
> 
> Fix this by taking this branch only when ht->p.max_size is non-zero
> 
> Fixes: Fixes: 6d684e54690c ("rhashtable: Cap total number of entries to 2^31")
> Signed-off-by: Florian Fainelli <florian.fainelli@...adcom.com>

Sent another version with the correct email address and marked this one
as superseded in patchwork, not that this email is not valid, but it's
all about consistency.

David pleas apply this one instead:
http://patchwork.ozlabs.org/patch/756172/

/me remembers to stop switching between machines.
-- 
Florian