> On Sun, Oct 25, 2026 at 03:25:28PM +0000, Chad Monroe wrote:
> > On Wednesday, October 21st, 2026 at 5:49 AM, Andrew Lunn <andrew@lunn.ch> wrote:
> > 
> > > On Wed, Oct 21, 2026 at 03:34:23PM +0530, SAYANTAN NANDY wrote:
> > > > From: Chad Monroe <chad@monroe.io>
> > > >
> > > > With flow stats enabled, PPE offload setup requests flow statistics
> > > > from the NPU. If the running NPU firmware does not support flow
> > > > counters, the request fails and the current setup path aborts PPE
> > > > offload.
> > > 
> > > Since this never worked, it is not a regression fix.
> > > 
> > > Why not keep it KISS and report the firmware is too old and suggest
> > > the firmware is updated?
> > > 
> > > Can the driver do the firmware upgrade?
> > > 
> > >     Andrew
> > > 
> > 
> > Hi Andrew,
> > 
> > The NPU firmware bin is downloaded at each boot and is present in
> > linux-firmware so no issue with keeping it up to date.
> > 
> > The main reason for this change is to behave better when flags mismatch.
> > Previously, loading a driver compiled with flow stats support enabled
> > would silently fail to offload if the firmware didn't have support.
> 
> Sayantan suggested this is a regression. It used to work, but does not
> now?
> 
> We should avoid regressions.

I do not think this is a regression. In this case, the default behaviour is
just fall-back to sw flowtable acceleration.

@sayantan: can you please confirm?

Regards,
Lorenzo

> 
>    Andrew