netdev - Re: [PATCH net] net: Fix icmp host relookup triggering ip_rt

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: <ZzLWcxskwi9e_bPf@gondor.apana.org.au>
Date: Tue, 12 Nov 2024 12:15:47 +0800
From: Herbert Xu <herbert@...dor.apana.org.au>
To: "dongchenchen (A)" <dongchenchen2@...wei.com>,
	Steffen Klassert <steffen.klassert@...unet.com>
Cc: davem@...emloft.net, dsahern@...nel.org, edumazet@...gle.com,
	pabeni@...hat.com, horms@...nel.org, netdev@...r.kernel.org,
	yuehaibing@...wei.com, zhangchangzhong@...wei.com
Subject: Re: [PATCH net] net: Fix icmp host relookup triggering ip_rt_bug

On Tue, Nov 12, 2024 at 11:42:47AM +0800, dongchenchen (A) wrote:
>
> If skb_in is outbound, fl4_dec.saddr is not nolocal. It may be no input
> route from B to A for
> 
> first communication.

You're right.  So the problem here is that for the case of A
being local, we should not be taking the ip_route_input code
path (this is intended for forwarded packets).

In fact if A is local, and we're sending an ICMP message to A,
then perhaps we could skip the IPsec lookup completely and just
do normal routing?

Steffen, what do you think?

So I think it boils down to two choices:

1) We could simply drop IPsec processing if we notice that A
(fl.fl4_dst) is local;
2) Or we could take the __ip_route_output_key code path and
continue to do the xfrm lookup when A is local.

Thanks,
-- 
Email: Herbert Xu <herbert@...dor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt