Re: [PATCH 2/2] timer: Mark this as incompatible with SYS_ARCH_TIMER

[Marius Dinu <m95d+git@psihoexpert.ro>]

On Sat, 2025-10-11 12.00.23 +-0600, Tom Rini wrote:
> On Thu, Aug 14, 2025 at 09:10:49PM +0300, Marius Dinu wrote:
> > On Thu, 2025-08-14 10.40.21 +-0600, Tom Rini wrote:
> > > On Thu, Aug 14, 2025 at 07:26:47PM +0300, Marius Dinu wrote:
> > > > On Wed, 2025-08-13 14.55.11 +-0600, Tom Rini wrote:
> > > > > The timers implemented under SYS_ARCH_TIMER are not part of the TIMER
> > > > > framework and thus cannot be enabled at the same time. Express this
> > > > > dependency in Kconfig and add a note to the help of SYS_ARCH_TIMER.
> > > > > 
> > > > > Reported-by: Marius Dinu <m95d+git@psihoexpert.ro>
> > > > > Signed-off-by: Tom Rini <trini@konsulko.com>
> > > > > ---
> > > > > Marius, I ended up reworking things a bit (and also found part 1/2 of
> > > > > this series) but please let me know if you feel this should retain your
> > > > > Author and such instead. Thanks!
> > > > 
> > > > No need for credits.
> > > 
> > > OK.
> > > 
> > > > I think you should keep the warning message in the Timer drivers menu.
> > > > It provides a quick explanation why the menu is empty.
> > > 
> > > I thought about it, and we don't do anything like that for other
> > > either/or choices, and it's now in the help text itself at least. The
> > > other thing however, is that a few platforms do allow for both, because
> > > they play some games with CFG_SYS_TIMER* defines. So I'm not entirely
> > > sure what to do here.
> > > 
> > > > Bugs are reported in this mailing list too?
> > > 
> > > Yes.
> > > 
> > > -- 
> > > Tom
> > 
> > With the original patch I only tried to reduce the number of compilation
> > failures. I encounter a lot. I'm no expert and I can only try to fix little
> > things such as this one.
> > 
> > I belive some people (like myself) would be confused by an empty menu and
> > would not immediately remember the SYS_ARCH_TIMER in a completely different
> > part of menuconfig.
> > 
> > 
> > I tried to find configs with both CONFIG_SYS_ARCH_TIMER and CONFIG_TIMER
> > enabled amd I couldn't find any. So it's probably OK to leave the patch as it
> > is and exceptions, if needed, can be added later.
> 
> I haven't forgotten about this, sorry. I'm just debating with myself
> what's best here, because another part of the proper solution is that we
> should get the generic ARMv8 timer on moved to the framework too, but
> it's more easily said than done, but I have asked some people to look in
> to it.
> 
> -- 
> Tom

No hurry. Take your time.

Marius