RE: [PATCH net-next v1] mlxbf_gige: emit messages during open and probe failures

[David Thompson]

> -----Original Message-----
> From: Simon Horman <horms@xxxxxxxxxx>
> Sent: Monday, June 16, 2025 3:18 PM
> To: Alexander Lobakin <aleksander.lobakin@xxxxxxxxx>
> Cc: David Thompson <davthompson@xxxxxxxxxx>; andrew+netdev@xxxxxxx;
> davem@xxxxxxxxxxxxx; edumazet@xxxxxxxxxx; kuba@xxxxxxxxxx;
> pabeni@xxxxxxxxxx; Asmaa Mnebhi <asmaa@xxxxxxxxxx>; u.kleine-
> koenig@xxxxxxxxxxxx; netdev@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH net-next v1] mlxbf_gige: emit messages during open and
> probe failures
> 
> On Mon, Jun 16, 2025 at 04:06:49PM +0200, Alexander Lobakin wrote:
> > From: Simon Horman <horms@xxxxxxxxxx>
> > Date: Mon, 16 Jun 2025 14:57:10 +0100
> >
> > > On Fri, Jun 13, 2025 at 05:42:28PM +0000, David Thompson wrote:
> > >> The open() and probe() functions of the mlxbf_gige driver check for
> > >> errors during initialization, but do not provide details regarding
> > >> the errors. The mlxbf_gige driver should provide error details in
> > >> the kernel log, noting what step of initialization failed.
> > >>
> > >> Signed-off-by: David Thompson <davthompson@xxxxxxxxxx>
> > >
> > > Hi David,
> > >
> > > I do have some reservations about the value of printing out raw err
> > > values. But I also see that the logging added by this patch is
> > > consistent with existing code in this driver.
> > > So in that context I agree this is appropriate.
> > >
> > > Reviewed-by: Simon Horman <horms@xxxxxxxxxx>
> >
> > I still think it's better to encourage people to use %pe for printing
> > error codes. The already existing messages could be improved later,
> > but then at least no new places would sneak in.
> 
> Thanks, I agree that is reasonable.
> And as a bonus the patch-set could update existing messages.
> 
> David, could you consider making this so?

Sorry for late response.

Yes, will look into updating this patch to use %pe

Thanks for the feedback.

Dave