From mboxrd@z Thu Jan  1 00:00:00 1970
From: Eric Dumazet <eric.dumazet@gmail.com>
Subject: Re: [PATCH net-next 4/6] inet: frags: use rhashtables for reassembly
 units
Date: Fri, 30 Mar 2018 06:30:42 -0700
Message-ID: <bc46ffdc-e3c4-93be-39ab-76c2634f2c6e@gmail.com>
References: <20180330052241.206667-2-edumazet@google.com>
 <20180330052241.206667-5-edumazet@google.com>
 <20180330071814.GA27025@gondor.apana.org.au>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
Cc: "David S . Miller" <davem@davemloft.net>,
        netdev <netdev@vger.kernel.org>, Florian Westphal <fw@strlen.de>,
        Thomas Graf <tgraf@suug.ch>,
        Jesper Dangaard Brouer <brouer@redhat.com>,
        Alexander Aring <alex.aring@gmail.com>,
        Stefan Schmidt <stefan@osg.samsung.com>,
        Eric Dumazet <eric.dumazet@gmail.com>,
        Nikolay Aleksandrov <nikolay@redhat.com>
To: Herbert Xu <herbert@gondor.apana.org.au>,
        Eric Dumazet <edumazet@google.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-pg0-f51.google.com ([74.125.83.51]:38598 "EHLO
        mail-pg0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751223AbeC3NbL (ORCPT
        <rfc822;netdev@vger.kernel.org>); Fri, 30 Mar 2018 09:31:11 -0400
Received: by mail-pg0-f51.google.com with SMTP id a15so5070468pgn.5
        for <netdev@vger.kernel.org>; Fri, 30 Mar 2018 06:31:11 -0700 (PDT)
In-Reply-To: <20180330071814.GA27025@gondor.apana.org.au>
Content-Language: en-US
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>



On 03/30/2018 12:18 AM, Herbert Xu wrote:

> Instead of using the walk interface, how about
> rhashtable_free_and_destroy?

Oh, I was not aware of this beast ;)

I guess I will need to add a cond_resched() in it, right ?

My rough guess is that freeing 30 million frags would take few seconds ;)

> 
>>  void inet_frag_kill(struct inet_frag_queue *fq)
>>  {
>>  	if (del_timer(&fq->timer))
>>  		refcount_dec(&fq->refcnt);
>>  
>>  	if (!(fq->flags & INET_FRAG_COMPLETE)) {
>> -		fq_unlink(fq);
>> +		struct netns_frags *nf = fq->net;
>> +
>> +		fq->flags |= INET_FRAG_COMPLETE;
>> +		rhashtable_remove_fast(&nf->rhashtable, &fq->node, nf->f->rhash_params);
>>  		refcount_dec(&fq->refcnt);
>>  	}
>>  }
> 
> This means that the hash won't inline properly.  Don't know big
> of an issue it is to you.  But you could fix it by doing the same
> hack as rhashtable by making inet_frag_kill an inline function and
> take the rhash_params as an explicit argument.

I will investigate this, thanks !