Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <735b9c21-6a8f-0f28-d11d-bd9bbd78986b@marvell.com>
Date:   Thu, 5 May 2022 09:11:02 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:100.0) Gecko/20100101
 Thunderbird/100.0
Subject: Re: [EXT] Re: [PATCH 0/5] net: atlantic: more fuzzing fixes
Content-Language: en-US
To:     Grant Grundler <grundler@chromium.org>,
        Dmitrii Bezrukov <dbezrukov@marvell.com>
CC:     Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
        netdev <netdev@vger.kernel.org>,
        "David S . Miller" <davem@davemloft.net>,
        LKML <linux-kernel@vger.kernel.org>,
        Aashay Shringarpure <aashay@google.com>,
        "Yi Chou" <yich@google.com>,
        Shervin Oloumi <enlightened@google.com>
References: <20220418231746.2464800-1-grundler@chromium.org>
 <CANEJEGtaFCRhVBaVtHrQiJvwsuBk3f_4RNTg87CWERHt+453KA@mail.gmail.com>
 <23cbe4be-7ced-62da-8fdb-366b726fe10f@marvell.com>
 <CANEJEGtVFE8awJz3j9j7T2BseJ5qMd_7er7WbdPQNgrdz9F5dg@mail.gmail.com>
 <BY3PR18MB4578949E822F4787E95A126CB4C09@BY3PR18MB4578.namprd18.prod.outlook.com>
 <CANEJEGvsfnry=tFOyx+cTRHJyTo2-TNOe1u4AWV+J=amrFyZpw@mail.gmail.com>
 <BY3PR18MB4578158E656F2508B43B21F6B4C39@BY3PR18MB4578.namprd18.prod.outlook.com>
 <CANEJEGuVwMa9ufwBM817dnbUxBghM0mcsPvrwx1vAWLoZ+CLaA@mail.gmail.com>
From:   Igor Russkikh <irusskikh@marvell.com>
In-Reply-To: <CANEJEGuVwMa9ufwBM817dnbUxBghM0mcsPvrwx1vAWLoZ+CLaA@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
Precedence: bulk


Hi Grant and Dmitrii,

>> So to close session I guess need to set is_rsc_completed to true when 
>> number of frags is going to exceed value MAX_SKB_FRAGS, then packet will 
>> be built and submitted to stack.
>> But of course need to check that there will not be any other corner cases 
>> with this new change.
> 
> Ok. Sounds like I should post a v2 then and just drop 1/5 and 5/5
> patches.  Will post that tomorrow.

I think the part with check `hw_head_ >= ring->size` still can be used safely (patch 5).

For patch 1 - I agree it may make things worse, so either drop or think on how to interpret invalid `next` and stop LRO session.

Thanks,
   Igor