linux-kernel - Re: [PATCH 1/2] scsi: ufs: Add Multi-Circular Queue support

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 26 Jul 2022 14:48:45 +0800
From:   Can Guo <quic_cang@...cinc.com>
To:     Bart Van Assche <bvanassche@....org>,
        Avri Altman <Avri.Altman@....com>,
        "stanley.chu@...iatek.com" <stanley.chu@...iatek.com>,
        "adrian.hunter@...el.com" <adrian.hunter@...el.com>,
        "alim.akhtar@...sung.com" <alim.akhtar@...sung.com>,
        "beanhuo@...ron.com" <beanhuo@...ron.com>,
        "quic_asutoshd@...cinc.com" <quic_asutoshd@...cinc.com>,
        "quic_nguyenb@...cinc.com" <quic_nguyenb@...cinc.com>,
        "quic_ziqichen@...cinc.com" <quic_ziqichen@...cinc.com>,
        "linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>,
        "kernel-team@...roid.com" <kernel-team@...roid.com>
CC:     "James E.J. Bottomley" <jejb@...ux.ibm.com>,
        "Martin K. Petersen" <martin.petersen@...cle.com>,
        Daejun Park <daejun7.park@...sung.com>,
        Jinyoung Choi <j-young.choi@...sung.com>,
        Kiwoong Kim <kwmad.kim@...sung.com>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] scsi: ufs: Add Multi-Circular Queue support

Hi Bart,

On 7/23/2022 1:58 AM, Bart Van Assche wrote:
> On 7/22/22 00:31, Avri Altman wrote:
>>> +#define UFSHCD_MCQ_IO_QUEUE_OFFSET 1
>> Maybe add a comment above: "queue 0 is reserved for query commands" 
>> or something
>> That is if the query commands don't use the  legacy doorbell
>
> Is it essential to reserve a queue for device management commands? 
> Wouldn't it be better to have one additional queue for submitting I/O 
> commands rather than reserving a queue for device management commands?


Since this is just RFC change, we are trying to make the whole thing 
work with minimal efforts.

So we found that having a reserved queue (with only one active command) 
for device management

requires much less changes in ufshcd.c, because current device 
management commands anyways come

one by one (we have a mutex lock dev_cmd.lock held in 
exec_dev_command()) and it is easy to handle/assign

the task tag for device management command by just reading sq_tp_slot. 
If you think this needs to be improved,

can you please elaborate your idea? Thanks.


Regards,

Can Guo.

>
> Thanks,
>
> Bart.
>