linux-kernel - Re: [PATCH V2 24/30] serial: qcom: Migrate to dev_pm_opp_set

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:   Fri, 1 Jul 2022 12:41:23 +0200
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     Viresh Kumar <viresh.kumar@...aro.org>
Cc:     Andy Gross <agross@...nel.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        Jiri Slaby <jirislaby@...nel.org>, linux-pm@...r.kernel.org,
        Vincent Guittot <vincent.guittot@...aro.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>,
        Stephen Boyd <sboyd@...nel.org>, Nishanth Menon <nm@...com>,
        linux-arm-msm@...r.kernel.org, linux-serial@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH V2 24/30] serial: qcom: Migrate to dev_pm_opp_set_config()

On Fri, Jul 01, 2022 at 03:59:26PM +0530, Viresh Kumar wrote:
> On 01-07-22, 12:18, Greg Kroah-Hartman wrote:
> > On Fri, Jul 01, 2022 at 03:31:00PM +0530, Viresh Kumar wrote:
> > Still crazy, but a bit better.
> 
> :)
> 
> > Why do you need the clk_count?  A null terminated list is better,
> 
> Because I am not a big fan of the null terminated lists :)
> 
> I had to chase a bug once where someone removed that NULL at the end
> and it was a nightmare to understand what's going on.

But that's the "normal" way the kernel does things.  Trying to keep a
count in sync with a list is a pain, and just gets harder and harder
over time.  Make it a null-terminated list so that the cpu makes this
always work and prevents errors.

> > as the
> > compiler can do it for you and you do not have to keep things in sync
> > like you are expecting people to be forced to do now.
> 
> I am not sure I understand what the compiler can do for us here.
> 
> The users will be required to do this here, isn't it ?
> 
>         const char *clks[] = { "core", NULL };
>         struct dev_pm_opp_config opp_config = {
>                .clk_names = clks,
>         };
> 

The "in sync" is the count issue.  Don't force humans to count up the
number of items in a list please.

> > The above is much more complex than a simple function call to make.
> > Remember to make it very simple for driver authors, and more
> > importantly, reviewers.
> 
> Hmm.
> 
> > Thanks, and drop the count field please.
> 
> There is one case at least [1] where we actually have to pass NULL in
> the clk name. This is basically to allow the same code to run on
> different devices, one where an OPP table is present and one where it
> isn't. We don't want to do clk_set_rate() for the second case but just
> use dev_pm_opp_set_rate() (which does a lot of stuff apart from just
> clk).

That feels completely wrong, don't have NULL for a name, make a fake name
or something.  Don't make all users in the kernel have a horrible
interface just for one piece of broken hardware out there.

Worst case, name it "".

thanks,

greg k-h