Re: [PATCH] blk-mq: use mutex_trylock to avoid lock inversion

[jianchao.wang]

On 06/28/2018 09:11 AM, jianchao.wang wrote:
> Hi Bart
> 
> On 06/26/2018 11:36 PM, Bart Van Assche wrote:
>> On 06/25/18 20:47, jianchao.wang wrote:
>>> V2 patch has been posted, would you please take a look at on them ?
>>> https://urldefense.proofpoint.com/v2/url?u=https-3A__marc.info_-3Fl-3Dlinux-2Dblock-26m-3D152965143412927-26w-3D2&d=DwICaQ&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=7WdAxUBeiTUTCy8v-7zXyr4qk7sx26ATvfo6QSTvZyQ&m=FPLHVLVF0V8ZCvwZJx7vHy9v01iAQogJ_IbvseGhElo&s=tQ0ZiK7VrV4j-6K_WEfiWAyxogOrC8hOm6coS9nAmYI&e=
>>
>> Hello Jianchao,
>>
>> Patches 1/3 and 2/3 of that series look fine to me but I don't like the approach of patch 3/3 so please do not expect a review from me for that patch.
>>
> 
> Would you please explain the defects of the patch 3/3 ?

We still need some way to serialize the blk_register/unregister_queue, blk_mq_sysfs_register/unregister and sysfs queue, mq, hctx entries' show and store method. So I didn't introduce another mutex as your suggestions.

> 
> Thanks
> Jianchao
>