Message-ID: <3897d6e8-1191-b42b-9553-c2720f3a92eb@linux.ibm.com>
Date:   Tue, 11 Jan 2022 16:46:38 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Thunderbird/91.4.1
Subject: Re: [PATCH net 1/3] net/smc: Resolve the race between link group
 access and termination
Content-Language: en-US
To:     Wen Gu <guwen@linux.alibaba.com>, davem@davemloft.net,
        kuba@kernel.org
Cc:     linux-s390@vger.kernel.org, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <1641806784-93141-1-git-send-email-guwen@linux.alibaba.com>
 <1641806784-93141-2-git-send-email-guwen@linux.alibaba.com>
 <8b720956-c8fe-0fe2-b019-70518d5c60c8@linux.ibm.com>
 <ee973642-6bae-e748-cea9-ed18bca461f0@linux.alibaba.com>
From:   Karsten Graul <kgraul@linux.ibm.com>
Organization: IBM Deutschland Research & Development GmbH
In-Reply-To: <ee973642-6bae-e748-cea9-ed18bca461f0@linux.alibaba.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 11/01/2022 16:36, Wen Gu wrote:
> Thanks for your review.
> 
> On 2022/1/11 4:23 pm, Karsten Graul wrote:
>> On 10/01/2022 10:26, Wen Gu wrote:
>>> We encountered some crashes caused by the race between the access
>>> and the termination of link groups.
>>>
>>
>> These waiters (seaparate ones for smcd and smcr) are used to wait for all lgrs
>> to be deleted when a module unload or reboot was triggered, so it must only be
>> woken up when the lgr is actually freed.
> 
> Thanks for your reminding, I will move the wake-up code to __smc_lgr_free().
> 
> And maybe the vlan put and device put of smcd are also need to be moved
> to __smc_lgr_free()?, because it also seems to be more suitable to put these
> resources when lgr is actually freed. What do you think?

Keep the calls to smc_ism_put_vlan() and put_device() in smc_lgr_free(),
thats okay for SMC-D.