Re: [PATCH] soc: imx8m: Fix match data lookup for soc device

[Frank Li]

On Thu, May 28, 2026 at 08:16:44PM +0200, Thorsten Leemhuis wrote:
> [+Alexander +Francesco]
>
> On 5/28/26 18:47, Frank Li wrote:
> > On Thu, May 28, 2026 at 12:55:34PM +0200, Thorsten Leemhuis wrote:
> >> On 5/7/26 17:46, Frank Li wrote:
> >>>
> >>> On Mon, 27 Apr 2026 09:01:48 +0800, Peng Fan (OSS) wrote:
> >>>> The i.MX8M soc device is registered via platform_device_register_simple(),
> >>>> so it is not associated with a Device Tree node and the imx8m_soc_driver
> >>>> has no of_match_table.
> >>>>
> >>>> As a result, device_get_match_data() always returns NULL when probing
> >>>> the soc device.
> >>>>
> >>>> [...]
> >>>
> >>> Applied, thanks!
> >>>
> >>> [1/1] soc: imx8m: Fix match data lookup for soc device
> >>>       commit: 2c6821657ce3b3c85f92719ea81ec9f9ff27df11
> >>
> >> Just wondering: are you planning to sent this towards mainline this
> >> cycle? The regression this fixes was introduced this cycle, so I guess
> >> it's something that should be merged before -rc6, as that's when Linus
> >> wants all known regression ideally fixed. But from the name of the
> >> branch it sounds more like this is scheduled to be merged in the next
> >> merge window (but I might easily mistaken there -- if I am: sorry!)
> >
> > Thanks reminders. I intent to send for 7.1. But second look, not sure any
> > big impact although it is regression.
> >
> > commmit message have not show what impact.
>
> Yup, there ideally should have been Link: tags to the problem reports,
> as explained by our docs. At least if that patch fixes the problems
> linked to below (looks like it, but I'm not totally sure):
>
> > I checked code, look like only
> > missed some soc informaiton. Let me know if I am wrong.
>
> Alexander reported problems here:
> https://lore.kernel.org/lkml/6593091.DvuYhMxLoT@steina-w/
> That mentions a warning and problems with soc_id and serial_number
>
> Francesco jumped in a little later:
> https://lore.kernel.org/lkml/20260427064704.GA17710@francesco-nb/
> This mentions "CAAM is not working"

Thank you point it. I just sent fixes pull-request to soc maintainer.

Frank

>
> Not sure what the impact is, but from the outside it looks a bit like
> "better fix this now to ensure those problems never enter a proper release"
>
> Ciao, Thorsten