Re: [RFC PATCH net-next v4 01/28] net: mdio: ipq8064: clean whitespaces in define

On Sun, May 16, 2021 at 10:49:59AM +0100, Jonathan McDowell wrote:

On Sat, May 15, 2021 at 07:30:26PM +0200, Ansuel Smith wrote:

quoted

On Sat, May 15, 2021 at 06:00:46PM +0100, Jonathan McDowell wrote:

quoted

On Sat, May 08, 2021 at 08:05:58PM +0200, Ansuel Smith wrote:

quoted

On Sat, May 08, 2021 at 08:02:33PM +0200, Andrew Lunn wrote:

quoted

On Sat, May 08, 2021 at 02:28:51AM +0200, Ansuel Smith wrote:

quoted

Fix mixed whitespace and tab for define spacing.

Please add a patch [0/28] which describes the big picture of what
these changes are doing.

Also, this series is getting big. You might want to split it into two,
One containing the cleanup, and the second adding support for the new
switch.

	Andrew

There is a 0/28. With all the changes. Could be that I messed the cc?
I agree think it's better to split this for the mdio part, the cleanup
and the changes needed for the internal phy.

FWIW I didn't see the 0/28 mail either. I tried these out on my RB3011
today. I currently use the GPIO MDIO driver because I saw issues with
the IPQ8064 driver in the past, and sticking with the GPIO driver I see
both QCA8337 devices and traffic flows as expected, so no obvious
regressions from your changes.

I also tried switching to the IPQ8064 MDIO driver for my first device
(which is on the MDIO0 bus), but it's still not happy:

qca8k 37000000.mdio-mii:10: Switch id detected 0 but expected 13

Can you try the v6 version of this series?

FWIW I tried v6 without altering my DT at all (so still using the GPIO
MDIO driver, and not switching to use the alternate PHY support) and got
an oops due to a NULL pointer dereference, apparently in the mdio_bus
locking code. I'm back porting to 5.10.37 (because I track LTS on the
device) so I might be missing something, but the v4 version I tried
previously worked ok.

I dropped patches 20-25 of the series (i.e. the piece that adds the
internal phy/mdio support) and tried again and that works fine, so it
does look like I either managed to mismerge them somehow (and those
pieces weren't the ones with conflicts) or there's a problem (possibly
only when the DT hasn't been updated to use the internal bus?).

J.

-- 
   101 things you can't have too   |  .''`.  Debian GNU/Linux Developer
     much of : 45 - Toblerone.     | : :' :  Happy to accept PGP signed
                                   | `. `'   or encrypted mail - RSA
                                   |   `-    key on the keyservers.