Re: [PATCH 1/2] serial: 8250: Allow port registration without UPF_BOOT_AUTOCONF

[Andy Shevchenko]

On Mon, Apr 29, 2019 at 9:27 AM Esben Haabendal <esben@xxxxxxxxxxxx> wrote:
> Andy Shevchenko <andy.shevchenko@xxxxxxxxx> writes:
> > On Sat, Apr 27, 2019 at 12:01 PM Esben Haabendal <esben@xxxxxxxxxxxx> wrote:
> >> Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> writes:
> >> > On Fri, Apr 26, 2019 at 06:54:05PM +0200, Esben Haabendal wrote:
> >> >> Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> writes:
> >> >> The reason for this patch is to be able to do exactly that (set port
> >> >> type and UPF_FIXED_TYPE) without having UPF_BOOT_AUTOCONF added.
> >> >>
> >> >> In the current serial8250_register_8250_port() there is:
> >> >>
> >> >>     uart->port.flags        = up->port.flags | UPF_BOOT_AUTOCONF;
> >> >>
> >> >> So, even though I set UPF_FIXED_TYPE, I get
> >> >> UPF_FIXED_TYPE|UPF_BOOT_AUTOCONF.
> >> >
> >> > Yes.
> >> >
> >> >> So I need this patch.
> >> >
> >> > Why? I don't see any problems to have these flags set.
> >>
> >> The problem with having UPF_BOOT_AUTOCONF is the call to
> >> serial8250_request_std_resource().  It calls request_mem_region(), which
> >> fails if the MFD driver already have requested the memory region for the
> >> MFD device.
> >
> > If it's MFD, why it requested the region for its child?
> > Isn't it a bug in MFD driver?
>
> It is a PCI driver, which calls pci_request_regions().  The PCI device
> carries a lot of different functions, which uses small slices of the PCI
> memory region(s).  With the resources being a tree structure, I don't
> think it is a bug when a parent driver requests the entire memory
> region.

If it's MFD driver, it's not its business to do something
child-related on child behalf.
In any case, Linux device resource model uses exclusive region
slicing. If you do like above, you call for a problems.

Btw, we have PCI MFD driver which enumerates 8250 (more precisely
8250_dw) w/o any issues.

> It would be nice if child drivers requesting memory would pass the
> parent memory resource.  Maybe 8250 driver could be changed to accept a
> struct resource pointer instead of a simple mapbase value, allowing to
> setup the resource with parent pointing to the MFD memory resource.

I don't see the problem in certain driver, I guess you are trying to
workaround existin Linux device resource model.

-- 
With Best Regards,
Andy Shevchenko