Re: [PATCH v1] can: flexcan: Drop unused platform driver data

Hey Marc,

On Thu, May 28, 2026 at 03:26:29PM +0200, Uwe Kleine-König (The Capable Hub) wrote:
> On Thu, May 28, 2026 at 01:00:15PM +0200, Marc Kleine-Budde wrote:
> > On 28.05.2026 11:45:39, Uwe Kleine-König (The Capable Hub) wrote:
> > > No function in the flexcan driver ever accesses the .driver_data member
> > > of its .id_table array as device_get_match_data() returns NULL for
> > > non-of platform devices. So drop that unused assignment.
> > >
> > > While touching that array, drop the trailing comma after the terminator
> > > entry.
> > 
> > Does this mean that commit 5e6c3454b405 ("net: can: Use
> > device_get_match_data()") introduced a regression?
> 
> Yes, I'd claim that this commit introduces a null pointer exception when
> mcf_flexcan0 (defined in arch/m68k/coldfire/device.c) probes.

I wonder what your plan is here. Should I send a v2 that also drops the
unused variable that I missed in my v1, or do you want this NPE
addressed first?

Best regards
Uwe

-----BEGIN PGP SIGNATURE-----

iQEzBAABCgAdFiEEP4GsaTp6HlmJrf7Tj4D7WH0S/k4FAmopnn0ACgkQj4D7WH0S
/k6vGAf+P07h8wiZcTaqCGFDv3/2aYJax+UlwADLeBRQ17i1kGZEG57kJGO/DtG3
AJ3UwGF1EmwGKsZe0o1RY2aslrGi/H+9/YoB6DmZVyG+tox+0yvi4bM18XZOyllV
mklJFAB1Zm7Lbavlmq7U0WsaQ5JcNIOn5VvRTBuJsyGluJfMHr+F0n8qrzqEtMYY
XR/+Iaexegpe75VAXdLqPnlKPC6FzStBYkOWr1jLgIKJj5BrvXmlkAoL90+4yiJy
zv4gtclllsXd5q8/Btjclk5kPNVk1HYYVVVZDPRZ1VXeWg4zK5E8a8Gktu8SHbri
co1ZFXcx/TxF9w1oYwXWBdkD4jWZ3A==
=yx/5
-----END PGP SIGNATURE-----