[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <X/dgphCCXr9KHY7h@lunn.ch>
Date: Thu, 7 Jan 2021 20:27:34 +0100
From: Andrew Lunn <andrew@...n.ch>
To: Oliver Neukum <oneukum@...e.com>
Cc: netdev@...r.kernel.org, davem@...emloft.net, roland@...nel.org
Subject: Re: [PATCH 2/3] usbnet: add method for reporting speed without MDIO
On Thu, Jan 07, 2021 at 12:35:17PM +0100, Oliver Neukum wrote:
Hi Oliver
> +++ b/include/linux/usb/usbnet.h
> @@ -53,6 +53,8 @@ struct usbnet {
> u32 hard_mtu; /* count any extra framing */
> size_t rx_urb_size; /* size for rx urbs */
> struct mii_if_info mii;
> + long rxspeed; /* if MII is not used */
> + long txspeed; /* if MII is not used */
How generic is this really? I don't know USB networking, so i cannot
say for myself.
I'm wondering if these should be added to cdc_ncm_ctx, and
usbnet_get_link_ksettings_ncm function should be added?
Having said that, USB_CDC_NOTIFY_SPEED_CHANGE seems like it could also
use this. Should the patch set also handle that notification and set
usbnet->rxspeed and usbnet->txspeed? These would then be used in
multiple places and that would justify it being in struct usbnet.
Andrew
Powered by blists - more mailing lists