[PATCH v1] netdevice: clean old FIXME that it is not worthed

[PATCH v1] netdevice: clean old FIXME that it is not worthed

[Vincenzo Palazzo]

Alternative patch that removes an old FIXME because it currently
the change is worthed as some comments in the patch point out
(https://lore.kernel.org/all/20230304080650.74e8d396@hermes.local/#t)

Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
---
 include/linux/netdevice.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 6a14b7b11766..82af7eb62075 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -2033,7 +2033,6 @@ struct net_device {
 	struct dev_ifalias	__rcu *ifalias;
 	/*
 	 *	I/O specific fields
-	 *	FIXME: Merge these and struct ifmap into one
 	 */
 	unsigned long		mem_end;
 	unsigned long		mem_start;
-- 
2.39.2

Re: [PATCH v1] netdevice: clean old FIXME that it is not worthed

[Stephen Hemminger]

On Sat,  4 Mar 2023 20:44:33 +0100
Vincenzo Palazzo <vincenzopalazzodev@gmail.com> wrote:

> Alternative patch that removes an old FIXME because it currently
> the change is worthed as some comments in the patch point out
> (https://lore.kernel.org/all/20230304080650.74e8d396@hermes.local/#t)
> 
> Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
> ---
>  include/linux/netdevice.h | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
> index 6a14b7b11766..82af7eb62075 100644
> --- a/include/linux/netdevice.h
> +++ b/include/linux/netdevice.h
> @@ -2033,7 +2033,6 @@ struct net_device {
>  	struct dev_ifalias	__rcu *ifalias;
>  	/*
>  	 *	I/O specific fields
> -	 *	FIXME: Merge these and struct ifmap into one
>  	 */
>  	unsigned long		mem_end;
>  	unsigned long		mem_start;

These fields actually are only used by old hardware devices that
pre-date buses with auto discovery. I.e ISA bus not PCI.

Since ISA bus support is gone, either these devices should have
been removed as well or they really aren't using those fields..

If someone wanted to clean this stuff out, start by seeing
if any of those devices still live. For example, the E1000e has
a couple of variants and dropping support for the non-PCI variant
would be ok.

All of arcnet could/should go away? Maybe move to staging?

The wan devices might also have been ISA only devices that are now
unusable.

Re: [PATCH v1] netdevice: clean old FIXME that it is not worthed

[Vincenzo Palazzo]

On Sat Mar 4, 2023 at 9:17 PM CET, Stephen Hemminger wrote:
> On Sat,  4 Mar 2023 20:44:33 +0100
> Vincenzo Palazzo <vincenzopalazzodev@gmail.com> wrote:
>
> > Alternative patch that removes an old FIXME because it currently
> > the change is worthed as some comments in the patch point out
> > (https://lore.kernel.org/all/20230304080650.74e8d396@hermes.local/#t)
> > 
> > Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
> > ---
> >  include/linux/netdevice.h | 1 -
> >  1 file changed, 1 deletion(-)
> > 
> > diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
> > index 6a14b7b11766..82af7eb62075 100644
> > --- a/include/linux/netdevice.h
> > +++ b/include/linux/netdevice.h
> > @@ -2033,7 +2033,6 @@ struct net_device {
> >  	struct dev_ifalias	__rcu *ifalias;
> >  	/*
> >  	 *	I/O specific fields
> > -	 *	FIXME: Merge these and struct ifmap into one
> >  	 */
> >  	unsigned long		mem_end;
> >  	unsigned long		mem_start;
>
> These fields actually are only used by old hardware devices that
> pre-date buses with auto discovery. I.e ISA bus not PCI.
>
> Since ISA bus support is gone, either these devices should have
> been removed as well or they really aren't using those fields..
>
> If someone wanted to clean this stuff out, start by seeing
> if any of those devices still live. For example, the E1000e has
> a couple of variants and dropping support for the non-PCI variant
> would be ok.
>
> All of arcnet could/should go away? Maybe move to staging?
>
> The wan devices might also have been ISA only devices that are now
> unusable.

Well I could do this, also because from the patch [1] that I made
looks like there are not a lot of driver that are using these fields.

However, what is the right process for this kind of clean up work?

[1] https://lore.kernel.org/all/20230304122432.265902-1-vincenzopalazzodev@gmail.com/