[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d1b76c8e-a9b3-b284-9a79-752afa73bb90@gmail.com>
Date: Thu, 19 Dec 2019 14:05:09 -0800
From: Florian Fainelli <f.fainelli@...il.com>
To: Michael Grzeschik <m.grzeschik@...gutronix.de>, andrew@...n.ch
Cc: netdev@...r.kernel.org, davem@...emloft.net, kernel@...gutronix.de
Subject: Re: [PATCH] mdio-bitbang: add support for lowlevel mdio read/write
On 12/18/19 8:29 AM, Michael Grzeschik wrote:
> Some phys support special opcode handling when communicating via mdio.
> This patch introduces mdio_ll_read/write which makes it possible to set
> the opcode. It implements these functions in the gpio-bitbang driver,
> which is capable of setting the opcode on read and write.
>
> Signed-off-by: Michael Grzeschik <m.grzeschik@...gutronix.de>
> ---
> Hi Andrew,
>
> I worked on your suggestion moving the proprietary call to
> mdio-ksz88x3.c which does not seem to work out very well.
> I still end up having an MII_ADDR_SMI???? define in linux/phy.h.
>
> Instead of having to support this special case in one extra file
> what do you think of adding mdiobus_lowlevel_write/read to mdio_bus.
> This way it would be possible to add the opcode directly as user.
>
> Other controllers which have the possibility to set the op code in hardware
> will also profit from that and can implement these functions.
I am not sure it makes sense for the entire struct mii_bus to gain two
new pointers, when you could just exported the mdiobb_ll_read() and
mdiobb_ll_write() to the kernel modules that needs those, and wrap them
however you need them to implement the mdiobus->read() and
mdiobus->write() operations?
--
Florian
Powered by blists - more mailing lists