[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <46528573-b584-4e37-82c9-f8fc4e0428d9@lunn.ch>
Date: Wed, 3 Jan 2024 14:54:29 +0100
From: Andrew Lunn <andrew@...n.ch>
To: Yajun Deng <yajun.deng@...ux.dev>
Cc: davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
pabeni@...hat.com, olteanv@...il.com, hkallweit1@...il.com,
linux@...linux.org.uk, przemyslaw.kitszel@...el.com,
rmk+kernel@...linux.org.uk, kabel@...nel.org,
netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-phy@...ts.infradead.org
Subject: Re: [PATCH net-next v2 0/2] net: phy: Use is_phy_driver() and
is_phy_device()
On Wed, Jan 03, 2024 at 10:53:32AM +0800, Yajun Deng wrote:
> There is only one flag that can be set in struct mdio_driver_common and
> mdio_device. We can compare the probe of the driver or the type of the
> device to implement it. Hence, these flags in struct mdio_driver_common
> and mdio_device can be removed.
>
> Introduce is_phy_driver() and is_phy_device(). Use them test the driver
> or device.
I'm still not convinced this is useful. Please expand your commit
message. One things which might convince me this is useful is if the
PHY drivers can make there struct phy_driver structures const.
Also, please break this patch series up. You should be able to add the
helper is_phy_driver() and make use of it, without changing
common. You should be able to add is_phy_device() without changing
common. So do these little steps first. The current code is hard to
review because these changes are all mixed in with everything
else. Once you have done the preparation steps, you can then do the
mass change.
Andrew
---
pw-bot: cr
Powered by blists - more mailing lists