Re: [RFC] [PATCH] PNP: request ioport and iomem resources used by active devices

[Bjorn Helgaas <bjorn.helgaas@hp.com>]

On Wednesday 07 November 2007 06:22:48 am Maciej W. Rozycki wrote:
> On Mon, 29 Oct 2007, Bjorn Helgaas wrote:
> 
> >     0000-001f : dma1		<-- built-in resource includes 2 controllers
> >       0000-000f : 00:02		<-- PNP reports only one DMA controller
> [...]
> >     0060-006f : keyboard	<-- built-in resource groups several things
> >       0060-0060 : 00:04		<-- PNP reports 8042 controller data register
> >       0061-0061 : 00:03		<-- PNP reports AT-style speaker
> >       0064-0064 : 00:04		<-- PNP reports 8042 controller status register
> 
>  Note that the built-in resources are requested according to how the PC/AT 
> architecture defined address decoding for the relevant subsystems.  In 
> particular, there was always a single DMA controller at 0x0000 and the 
> 8042 controller was decoded at 0x0060.  No full decoding was done and 
> unused locations aliased to the corresponding ones with the don't-care 
> bits set to the other value.  I do not recall all the details of the port 
> B (named after the original from the 8255 as used in the PC/XT) at 0x0061 
> as implemented in the PC/AT and I do not have a reference handy, but it is 
> a bunch of GPIO bits, some being r/o and some r/w, possibly with 
> side-effects.  I believe it was a set of additional latches sub-decoded 
> from the 8042 controller range for compatibility with how the 8255 was 
> wired and programmed.
> 
>  Anyway, PNP may provide a more accurate picture of the address ranges in 
> a particular system... if it gets it right!  Which is a lost hope, I 
> suppose, as even the above is questionable -- "AT-style speaker" is a 
> misnomer as most of the GPIO bits involved deal with the handling of 
> system error interrupts normally routed to NMI.

Thanks for the background and the comments.

There will undoubtedly be errors in what the firmware reports.  But if
firmware tells us about some non PC/AT device, we should believe it at
least enough to avoid placing some other device on top of it.

If we have PNPBIOS or ACPI, the built-in stuff reserved in
request_standard_resources() should be superfluous.  In theory.
But I doubt there's any benefit in skipping request_standard_resources().
Possibly it could be done *after* the PNP reservations, so the
hierarchy would make more sense, but I don't know if even that
is worth the trouble.

>  The patch itself is useful and makes sense though.  Have you passed it 
> through `scripts/checkpatch.pl'?

Yes.  The patch was added to -mm on Nov 1 as
pnp-request-ioport-and-iomem-resources-used-by-active-devices.patch.

Bjorn
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/