On Wed, 2016-05-04 at 14:20 +0300, Andy Shevchenko wrote:

quoted

To move away from a valid/standard PCI probe routine into a new
special
LPSS/PCI shim (which the hardware doesn't actually mandate) I do
think
you should to setup the dependency CONFIG_8250_PCI =>
CONFIG_8250_LPSS.

No, this is what we try avoiding

Fine.

Could you then select CONFIG_SERIAL_8250_LPSS when
CONFIG_X86_INTEL_QUARK is true - since it will be a dependency.

If user selects CONFIG_SERIAL_8250_PCI, the CONFIG_SERIAL_8250_LPSS
will
be selected as well since it has same dependencies.

I still think the change is not an obvious one i.e. LPSS (as an ACPI
enumeration concept) is not a requirement to enumerate on Quark X1000.

So I understand why you want to separate out the code from 8250_pci -
however I think the *minimum* here should be a descriptive comment in
kconfig listing which PCI-enumerated SoCs now require the 8250_LPSS
work-around if just selecting 8250_LPSS isn't possible.

So how about listing out those SoCs - something like

"Selecting this option will enable handling of the extra features 
 present on the UART found on Intel Braswell SoC and various  other
 Intel platforms."

=>

"Selecting this option will enable handling of the extra features 
 present on the UART found on
 - Intel Braswell SoC
 - Intel Quark x1000 SoC
 - etc
"
If you make those changes - please add.

Reviewed-by: Bryan O'Donoghue <redacted>