On Mon, Jan 13, 2025 at 02:12:01PM +0100, Arnd Bergmann wrote: > On Mon, Jan 13, 2025, at 13:20, Andy Shevchenko wrote: > > On Mon, Jan 13, 2025 at 11:36:08AM +0100, Arnd Bergmann wrote: > >> On Mon, Jan 13, 2025, at 10:59, Andy Shevchenko wrote: > >> > On Thu, Jan 09, 2025 at 01:40:14PM +0100, Arnd Bergmann wrote: > >> >> On Thu, Jan 9, 2025, at 13:08, WangYuli wrote: > >> >> > >> >> Can you explain why this isn't done as part of > >> >> drivers/tty/serial/8250/8250_pnp.c? > >> > > >> > I assume you meant 8250_platform.c. PNP is for devices went through legacy PNP > >> > enumeration, most likely having IOPORT instead of IOMEM. > >> > >> No, I meant the 8250_pnp.c file. > > > > I am puzzled then. How should it work? PNP ID != ACPI HID that's provided in > > this patch commit message. On top of that, PNP driver uses _legacy_ PMP bus > > and infrastructure. > > I guess I don't understand enough about ACPI then, I always > assumed these were the same identifiers. I do see that > CONFIG_ACPI force-enables CONFIG_PNP, and I see the examples in > https://uefi.org/htmlspecs/ACPI_Spec_6_4_html/06_Device_Configuration/ evice_Configuration.html > refer to _HID values in the form of "PNP####". Check the implementation of the pnp_add_id(), for example. Besides that PNP device has only 24-bit DMA mask by default, which is due to legacy DMA controllers. In any case the differences between PNP IDs and ACPI IDs are described in the specification. -- With Best Regards, Andy Shevchenko
