Re: [PATCH] net: Fix typos

[Bjorn Helgaas]

On Mon, Jul 28, 2025 at 02:19:41PM +0100, Edward Cree wrote:
> On 23/07/2025 21:15, Bjorn Helgaas wrote:
> > From: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
> > 
> > Fix typos in comments and error messages.

> > +++ b/drivers/net/ethernet/sfc/mcdi_pcol.h
> 
> mcdi_pcol.h is automatically generated from an external source (it comes
>  from the firmware development team), these fixes will likely be
>  overwritten next time we pull in updates.  I will try to get them fed
>  back into the upstream sources.

Thanks!  Is there any way to include a note in the generated file
about the fact that it is generated, and maybe a pointer for where to
send corrections?  It looks like others have made the same mistake
before, but maybe those fixes got fed back upstream behind the scenes:

  982a2b5ffdbb ("sfc: fix repeated words in comments")
  ac6bef064f71 ("sfc: Fix spelling mistake "writting" -> "writing"")
  0f5e15583060 ("scripts/spelling.txt: add "an one" pattern and fix typo instances")

> > +++ b/drivers/net/ethernet/sfc/tc_encap_actions.c
> > @@ -442,7 +442,7 @@ static void efx_tc_update_encap(struct efx_nic *efx,
> >  			rule = container_of(acts, struct efx_tc_flow_rule, acts);
> >  			if (rule->fallback)
> >  				fallback = rule->fallback;
> > -			else /* fallback fallback: deliver to PF */
> > +			else /* fallback: deliver to PF */
> >  				fallback = &efx->tc->facts.pf;
> >  			rc = efx_mae_update_rule(efx, fallback->fw_id,
> >  						 rule->fw_id);
> 
> This wording was intentional, not a type: delivery to the PF is the
>  second-layer fallback when there is no fallback action, which makes it
>  the fallback to the fallback.
> I will post a partial revert to change this line back.

Oof, sorry about that!  Adding something about this being a
second-layer fallback, or a default or last resort when there is no
fallback, might help avoid confusion.

Bjorn