On Tue, Jun 22, 2021 at 1:29 PM Geert Uytterhoeven [off-list ref] wrote:

On Tue, Jun 15, 2021 at 6:21 PM Rob Herring [off-list ref] wrote:

quoted

On Tue, Jun 15, 2021 at 10:06 AM Alex Nemirovsky [off-list ref] wrote:

quoted

It sounds like you are find with patch 2/3 from a DT point of view.  Could we review the rest from the DT point of view
to get either feedback for changes or ACK these, so we can unblock this series?

Can't say I've seen it as I only see replies in my mail. Did this
originally go to the DT list? If not, it's never in my queue[1].

None of the patches in this series ended up on a mailing list or on
lore, so no one could comment on them, except for people CCed directly.

Right, in fact every email I see from @cortina-access.com at
https://lore.kernel.org/lkml/?q=cortina-access.com is a reply, but the
original mails are all missing, not just this series.

It appears that there is a problem between the mail servers of cortina-access
and kernel.org. Alex and Jason, please double-check if you find
anything in

https://www.kernel.org/doc/html/latest/process/submitting-patches.html
https://www.kernel.org/doc/html/latest/process/email-clients.html

that may have caused this on your end, or otherwise contact
postmaster@vger.kernel.org for help.

The driver has been accepted in tty-next, commit b61c8bf4694b5115
("tty: serial: Add UART driver for Cortina-Access platform").  From a
quick glance, it could have used some review.

The driver is using the compatible value "cortina-access,serial", so I
guess the binding patch added "cortina-access" to vendor-prefixes.yaml.

Ah right. That patch has not made it in because of the issue, but right
away I can tell that the binding is not great because the compatible
string is way too generic: it implies that cortina-access only uses a single
serial port type in every past, present or future SoC, which is clearly
not the case.

It's obviously up to Greg to decide whether to revert the patch, but
I'd suggest we don't treat the driver as conforming to the binding until
it has been reviewed, and assume that changes will be made.

The SERIAL_CORTINA_ACCESS symbol doesn't depend on anything, so
it will show up on everyone's oldconfig radar soon, regardless of
building a kernel for a Cortina Access system or not.
I wanted to change it to something like:

     config SERIAL_CORTINA_ACCESS
            tristate "Cortina-Access serial port support"
    +       depends on FIXME || COMPILE_TEST
            select SERIAL_CORE
            help
              This driver is for Cortina-Access SoC's UART. If you
have a machine
              based on the Cortina-Access SoC and wish to use the serial port,
              say 'Y' here. Otherwise, say 'N'.

but given there is no evidence of patches to add support for the
CAXXXX line of SoCs, there's no symbol to depend on...

Assuming this is a SoC based on an Arm CPU core, it's clear that
nothing from cortina-access.com ever made it through the
linux-arm-kernel@lists.infradead.org mailing list either, if anything
was sent at all.

Alex, please try to contact me off-list about merging the full SoC
support, I should be able to help you come up with a plan for
submitting the rest as soon as the email troubles are resolved.

For new Arm based SoC platforms, arch/{arm,arm64} patches
should go through the soc tree and be picked up by Olof (on Cc)
or me. The exact process is not well documented, but please
contact us by email, or on irc://irc.libera.chat/#armlinux if you
have questions.

       Arnd