Anton Blanchard wrote:
>
>
> > Use pci_resource_end, avoid the unnecessary addition.
>
> Good point.
>
> > Why is the dev->base_addr assignment not conditional on USE_IO as well?
>
> OK, to be consistent it should not be.
>
> > Or be more specific,
> > 1) What are the semantics of mem_{start,end} versus base_addr? And,
> > 2) Why does ifconfig truncate a valid 32-bit address, when
> > dev->base_addr equals something like 0xF1234567? I noticed this but
> > never got around to looking into the reason.
>
> I do not know the intended semantics, but someone decided that a short was
> enough to store the base address:
>
> struct ifmap
> {
> unsigned long mem_start;
> unsigned long mem_end;
> unsigned short base_addr;
> unsigned char irq;
> unsigned char dma;
> unsigned char port;
> /* 3 bytes spare */
> };
>
> So we have to use mem_start/end for this.
Well crap. All of the drivers use base_addr regardless of PIO or MMIO,
and none use mem_start/end AFAICS. It should be no trouble to simply
assign mem_start/end for reporting purposes and ignore them otherwise,
but having 'base_addr' be a short here is a real PITA, since it doesn't
reflect real-life usage.
IMHO we should ditch mem_start/end, OR ditch base_addr...
Jeff
--
Jeff Garzik | Liberty is always dangerous, but
Building 1024 | it is the safest thing we have.
MandrakeSoft, Inc. | -- Harry Emerson Fosdick
|