Re: [PATCH] regulator: qcom-rpmh: Use PROBE_FORCE_SYNCHRONOUS

[Doug Anderson]

Hi,

On Fri, Mar 24, 2023 at 7:12 AM Andrew Halaney <ahalaney@xxxxxxxxxx> wrote:
>
> On Fri, Mar 24, 2023 at 12:18:53PM +0100, Marek Szyprowski wrote:
> > Hi,
> >
> > On 23.03.2023 23:08, Doug Anderson wrote:
> > > On Thu, Mar 23, 2023 at 3:05 PM Marek Szyprowski
> > > <m.szyprowski@xxxxxxxxxxx> wrote:
> > >> Restore synchronous probing for 'qcom,pm8150-rpmh-regulators' because
> > >> otherwise the UFSHC device is not properly initialized on QRB5165-RB5
> > >> board.
> > >>
> > >> Fixes: ed6962cc3e05 ("regulator: Set PROBE_PREFER_ASYNCHRONOUS for drivers between 4.14 and 4.19")
> > >> Signed-off-by: Marek Szyprowski <m.szyprowski@xxxxxxxxxxx>
> > >> ---
> > >>   drivers/regulator/qcom-rpmh-regulator.c | 2 +-
> > >>   1 file changed, 1 insertion(+), 1 deletion(-)
> > > I don't object to this patch landing temporarily, but can you provide
> > > any more details, please? On Qualcomm Chromebooks I'm not seeing any
> > > issues with RPMH regulators probing asynchronously, so I can only
> > > assume that there's a bug in the UFSHC driver that's being tickled.
> >
> > You are right. I've analyzed this case further and it turned out that it
> > was my fault. In short - 'rootwait' kernel cmdline parameter was missing
> > and root was specified as '/dev/sda7'.
> >
> > UFSHC driver properly retried probing after it cannot get its
> > regulators, but it happened at the same time when kernel tried to mount
> > rootfs. I was confused that this is really a regulator issue, because
> > the mentioned /dev/sda* devices were properly reported as available in
> > the system in the root mounting failure message, but adding the
> > 'rootwait' cmdline parameter fixed this problem. It would be safe to
> > revert this change. I'm really sorry for the false report and the noise.
> >
>
> It looks like this got applied, but reading your above message makes it
> seem like this patch is not necessary. Did I understand that correctly?
>
> If so we should see if Mark can drop / revert it?

Ah, sorry. I didn't reply with breadcrumbs. The revert is at:

https://lore.kernel.org/r/20230324063357.1.Ifdf3625a3c5c9467bd87bfcdf726c884ad220a35@changeid

-Doug