Re: [PATCH v2 6/6] serial: Consolidate BOTH_EMPTY use

[Andy Shevchenko]

On Thu, Jun 23, 2022 at 12:15 PM Andy Shevchenko
<andy.shevchenko@xxxxxxxxx> wrote:
> On Thu, Jun 23, 2022 at 10:24 AM Jiri Slaby <jirislaby@xxxxxxxxxx> wrote:
> > On 23. 06. 22, 10:11, Andy Shevchenko wrote:
> > >     * prom_putchar_wait() should be implemented using
> > >        read_poll_timeout_atomic(), incl. failure/timeout handling.
> > >
> > > Not sure since it is an early stage and scheduler might not work as
> > > expected. Conversions to iopoll.h macros bitten us a few times already.
> >
> > Except _atomic does not use scheduler :).
>
> Sorry for a bit misleading comment, but I chased it down, so this what
> I had in mind when commenting:
> be24c6a71ecf ("soc: qcom: rpmh-rsc: Don't use ktime for timeout in
> write_tcs_reg_sync()")

...and this one (specifically for early stages)

c4d936efa46d ("Revert "usb: early: convert to readl_poll_timeout_atomic()"")

> (Yes, it's about _atomic variant)
>
> Means we need to use those macros with care.

-- 
With Best Regards,
Andy Shevchenko