RFC: DSI panel lane frequency

Linus Walleij linus.walleij at linaro.org
Wed Oct 17 14:07:48 UTC 2018


On Wed, Oct 17, 2018 at 9:22 AM Andrzej Hajda <a.hajda at samsung.com> wrote:
> On 17.10.2018 08:56, Linus Walleij wrote:
> > On Tue, Oct 16, 2018 at 4:45 PM Andrzej Hajda <a.hajda at samsung.com> wrote:
> >
> >> Yes, it is missing part, please add units (Hz ?) and  some optional
> >> propositions:
> >> 1. maybe hs_speed instead of hs_frequency - shorter.
>
> Now I see some disadvantage of hs_speed - it suggests Mbps, not Hz,
> maybe hs_rate ?
> Up to you.

Makes sense. Fixed and sent out.

> > So if we get frequencies above 4.3GHz we're gonna need to
> > refactor the whole kernel anyway. DSI speed is gonna be
> > the least of our problems.
>
> We could just use u64 to avoid the issue.

I opted not to do that for this reason:

u64 foo = 1;
unsigned long bar = foo;

This will give compiler warnings on some 32bit architectures
and you have to explicitly:

unsigned long bar = (unsigned long)foo;

Which is gonna look really bad. Since this will in 10 cases out
of 10 be backed by the clk subsystem and <linux/clk.h> I think
we should not use any other type than what clk is using.
When the change, we change.

Yours,
Linus Walleij


More information about the dri-devel mailing list