linux-kernel - Re: [RESEND PATCH 2/3] clk: ingenic: Mark critical clocks in Ingenic SoCs

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] [day] [month] [year] [list]

Message-ID: <YmLPotnIBStYpapa@localhost>
Date:   Fri, 22 Apr 2022 16:54:10 +0100
From:   Aidan MacDonald <aidanmacdonald.0x0@...il.com>
To:     Stephen Boyd <sboyd@...nel.org>
Cc:     linux-clk@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-mips@...r.kernel.org, mturquette@...libre.com,
        paul@...pouillou.net, paulburton@...nel.org,
        tsbogend@...ha.franken.de
Subject: Re: [RESEND PATCH 2/3] clk: ingenic: Mark critical clocks in Ingenic
 SoCs

On Thu, Apr 21, 2022 at 07:33:57PM -0700, Stephen Boyd wrote:
> Quoting Aidan MacDonald (2022-04-11 03:14:40)
> > Consider the CPU, L2 cache, and memory as critical to ensure they
> > are not disabled.
> > 
> > Signed-off-by: Aidan MacDonald <aidanmacdonald.0x0@...il.com>
> > Reviewed-by: Paul Cercueil <paul@...pouillou.net>
> > ---
> 
> General comment, please add a comment around CLK_IS_CRITICAL usage if it
> isn't very clear why such a clk shouldn't be turned off. Second, is
> there any point in describing these clks in the kernel and using memory
> to do that if they're just going to always be on? Wouldn't a dummy clk
> returned from clk_get() work just as well if anything is grabbing a
> reference with clk_get()?

I'd guess they're there to keep track of which PLLs are in use, at least
for SoCs that have more than one PLL. Using a dummy clock sounds like a
bad idea since it won't represent that, and besides the clock configuration
is something that can change at runtime so hardcoding it would be foolish.

I'll send a v2 with explanatory comments around CLK_IS_CRITICAL.

Regards,
Aidan