Hello,
> > Perhaps it makes sense to make this a dev_dbg() over dev_info(), especially
> > since it appears that this information is of more use to the developer (who
> > most likely has the suitable log level set anyway), and given that there is
> > no way to reliably detect a presence in a slot on some platforms, this
> > might otherwise, add to the other messages that normal users don't pay
> > attention to usually - if this is not to be treated as an error.
> >
>
> No, this is clearly not a debug message. As I quoted above, the link up
> can fail due to an issue with PHY also. In that case, user has to see
> the log to debug/report the issue.
Apologies! I missed that. Thank you!
> So, either dev_info() or dev_err().
So, there is nothing to do here, then. This stays as dev_err() as per the
change from:
14c4ad125cf9 ("PCI: dwc: Log link speed and width if it comes up")
That said, perhaps adding a comment explaining why this is an error
might help future reference, as the linked commit didn't justify the
change. There also will be redundant messages printed, as per:
https://elixir.bootlin.com/linux/v6.0-rc5/source/drivers/pci/controller/dwc/pcie-fu740.c#L207
Which the linked commit didn't take into account, I suppose.
Krzysztof
