[PATCH 007/108] dm: pci: Allow disabling auto-config for a device

[Simon Glass]

Hi Bin,

On Sat, 8 Feb 2020 at 08:54, Bin Meng <bmeng.cn at gmail.com> wrote:
>
> Hi Simon,
>
> On Sat, Feb 8, 2020 at 11:51 PM Simon Glass <sjg at chromium.org> wrote:
> >
> > Hi Bin,
> >
> > On Sat, 8 Feb 2020 at 08:15, Bin Meng <bmeng.cn at gmail.com> wrote:
> > >
> > > Hi Simon,
> > >
> > > On Mon, Jan 27, 2020 at 1:08 PM Simon Glass <sjg at chromium.org> wrote:
> > > >
> > > > Add a means to avoid configuring a device when needed. Add an explanation
> > > > of why this is useful to the binding file.
> > > >
> > >
> > > Does disabling CONFIG_PCI_PNP not help?
> >
> > That is for all devices. Here we just want the UART to stay where it
> > is so that we can debug serial drivers, FSP-S, etc. If the UART moves
> > then any output after that hangs the machine until we reinit the
> > serial driver.
>
> OK, I think this is for debugging purpose, right? As you must have get
> Coral boot for some time, and this is only for retaining the console
> debug output?

Sort-of, but I've actually left it enabled. The problem is that if you
get an error or any debug output very early after relocation in
U-Boot, it hangs. I feel it is quite confusing for people to add a bit
of debugging and have it crash.

Admittedly there is a very small window where this happens (roughly
between arch_fsp_init_r() returning and initr_serial() being called.
But i found it pretty annoying.

What do you think?

Also I'd like to change it to suppress console output when it is not
allowed (and print a warning later), but haven't tried that yet.

>
> Reviewed-by: Bin Meng <bmeng.cn at gmail.com>
>
> Regards,
> Bin