linux-kernel - Re: [PATCH] clk: at91: clk-generated: Limit the requested rate to our range

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]

Message-ID: <25631072-d9a3-0d84-fd47-4d2414f079f6@microchip.com>
Date:   Fri, 9 Jul 2021 09:59:57 +0000
From:   <Codrin.Ciubotariu@...rochip.com>
To:     <Nicolas.Ferre@...rochip.com>, <linux-clk@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>
CC:     <mturquette@...libre.com>, <sboyd@...nel.org>,
        <alexandre.belloni@...tlin.com>, <Ludovic.Desroches@...rochip.com>,
        <Claudiu.Beznea@...rochip.com>
Subject: Re: [PATCH] clk: at91: clk-generated: Limit the requested rate to our
 range

On 09.07.2021 12:21, Nicolas Ferre wrote:
> On 07/07/2021 at 15:12, Codrin Ciubotariu wrote:
>> On clk_generated_determine_rate(), the requested rate could be outside
>> of clk's range. Limit the rate to the clock's range to not return an
>> error.
> 
> Isn't it saner for the user to return an error code instead of 
> automatically restrain the dynamics requested without notice?
> 
> Can you elaborate the use case where returning an error is not convenient?

The way I see it, if the user requests a rate that is out of clock's 
range, the driver's determine_rate() should return min/max, not an 
error. That is actually the closest rate supported by the clock, which 
is what determine_rate() should accomplish. The user has no clk API to 
get clock's range, so there is no way to call clk_round_rate() only for 
values within our range.

The use cause is with sam9x60's I2S driver, which has to try different 
rates to get the closest one to what it needs. There is no 'perfect' 
rate, because there is no AUDIO PLL and we have to try different values 
for our internal dividers to find the closest one.

https://elixir.bootlin.com/linux/latest/source/sound/soc/atmel/mchp-i2s-mcc.c#L416

Best regards,
Codrin