On Fri, Apr 19, 2024 at 09:59:26AM -0700, Shradha Gupta wrote:
> On Thu, Apr 18, 2024 at 08:42:59PM +0200, Andrew Lunn wrote:
> > > >From an RDMA perspective this is all available from other APIs already
> > > at least and I wouldn't want to see new sysfs unless there is a netdev
> > > justification.
> >
> > It is unlikely there is a netdev justification. Configuration happens
> > via netlink, not sysfs.
> >
> > Andrew
>
> Thanks. Sure, it makes sense to make the generic attribute configurable
> through the netdevice ops or netlink implementation. I will keep that in
> mind while adding the next set of configuration attributes for the driver.
> These attributes(from the patch) however, are hardware specific(that show
> the maximum supported values by the hardware in most cases).
ndev->max_mtu = gc->adapter_mtu - ETH_HLEN;
ndev->min_mtu = ETH_MIN_MTU;
This does not appear to be specific to your device. This is very
generic. We already have /sys/class/net/eth42/mtu, why not add
/sys/class/net/eth42/max_mtu and /sys/class/net/eth42/min_mtu for
every driver?
Are these values really hardware specific? Are they really unique to
your hardware? I have to wounder because you clearly did not think
much about MTU, and how it is actually generic...
Andrew
