linux-kernel - Re: [PATCH v4 6/8] media: imx274: add helper function to fill a reg

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:   Wed, 27 Jun 2018 11:50:50 +0200
From:   Luca Ceresoli <luca@...aceresoli.net>
To:     Sakari Ailus <sakari.ailus@....fi>
Cc:     linux-media@...r.kernel.org,
        Sakari Ailus <sakari.ailus@...ux.intel.com>,
        Leon Luo <leonl@...pardimaging.com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v4 6/8] media: imx274: add helper function to fill a reg_8
 table chunk

Hi Sakari,

On 27/06/2018 11:30, Sakari Ailus wrote:
> On Wed, Jun 27, 2018 at 10:13:12AM +0200, Luca Ceresoli wrote:
>> Hi Sakari,
>>
>> On 26/06/2018 14:20, Sakari Ailus wrote:
>>> Hi Luca,
>>>
>>> On Mon, Jun 11, 2018 at 01:35:37PM +0200, Luca Ceresoli wrote:
>>>> Tables of struct reg_8 are used to simplify multi-byte register
>>>> assignment. However filling these snippets with values computed at
>>>> runtime is currently implemented by very similar functions doing the
>>>> needed shift & mask manipulation.
>>>>
>>>> Replace all those functions with a unique helper function to fill
>>>> reg_8 tables in a simple and clean way.
>>>
>>> What's the purpose of writing these registers as multiple I²C writes, when
>>> this can be done as a single write (i.e. the address followed by two or
>>> three octets of data)?
>>
>> Good point. The for loops applying the register values (the lines just
>> after those changed by my patch) defuse the regmap bulk write capability.
>>
>> I guess this could be improved not filling any table, but directly
>> calling regmap_bulk_write(), passing the u16 or u32 register value with
>> proper endianness. No tables, less code. This would replace the present
>> patch with a shorter and more effective one. Is it what you was suggesting?
> 
> Yes, please.

Ok, will do. I think this will cut many lines of code, wow!

Patch 8 is only marginally related to this change, so it can be reviewed
independently except for the few lines in imx274_apply_trimming() where
registers are written.

-- 
Luca