From mboxrd@z Thu Jan 1 00:00:00 1970 From: Patrick McHardy Subject: Re: [XFRM]: xfrm_user: fix selector family initialization Date: Thu, 10 Apr 2008 02:55:46 +0200 Message-ID: <47FD6592.609@trash.net> References: <47F4EB73.1040705@trash.net> <47FBC2AF.7050508@trash.net> <47FCE30F.8060400@trash.net> <20080409.150931.93235087.davem@davemloft.net> <47FD5B1C.3060005@miyazawa.org> <47FD5CBC.6040905@trash.net> <47FD6464.2000608@miyazawa.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Cc: David Miller , pupilla@hotmail.com, netdev@vger.kernel.org To: Kazunori MIYAZAWA Return-path: Received: from stinky.trash.net ([213.144.137.162]:64402 "EHLO stinky.trash.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753197AbYDJAzu (ORCPT ); Wed, 9 Apr 2008 20:55:50 -0400 In-Reply-To: <47FD6464.2000608@miyazawa.org> Sender: netdev-owner@vger.kernel.org List-ID: Kazunori MIYAZAWA wrote: > Patrick McHardy wrote: >> Kazunori MIYAZAWA wrote: >>> >>> Sorry, I missed the thread. >>> I have no idea to solve the regression. >> >> >> Wouldn't explicitly initializing the SA family for >> inter-family tunnels work? > > I guess it is same as the original Herbert's patch. > It needs to explicitly initialize the family for inter-family tunnel. Thanks for looking into this, I haven't seen Herbert's patch, but I'd assume comparing the SA family to the policy family would work for determining whether its an inter-family tunnel in case the family in explicitly initialized. The only thing that would need to be changed in the kernel is the check in xfrm_init_state().