VIA EPIA MII 6000, pcmcia_socket0: unable to apply power

Paul Webster pjwebster at gmail.com
Sun Aug 21 00:02:56 EDT 2005


On 8/21/05, Samuel Abels <newsgroups at debain.org> wrote:
> Also sorry for the late response, I knew I forgot to answer an email...
> 
> Am Donnerstag, den 18.08.2005, 07:57 +1000 schrieb Paul Webster:
> > agpgart: AGP aperture is 64M @ 0xe0000000
> > Kernel command line: root=/dev/hda1 ro reserve=0x0e000000,0x02000000
> > quiet splash
> > [...]
> > dmeg -> PCI: Failed to allocate mem resource #0:10000 at ea010000 for 0000:02:00.0
> >
> > This indicates that you need to have the following appended to your
> > kernel boot options:
> > reserve=0xea010000,0x10000
> 
> Thanks, but you probably missed the part in my initial posting where I
> said that I already tried this?

Have you try the exact values I suggested? 0xea010000 is a very
different location than 0x0e000000, which you were using.

For reference, the "reserve=start_location,size" boot option is for
reserving a contiguous "size" block of memory, starting at
"start_location". As far as I can tell, the range you were reserving
would have been consumed by the AGP (ie. they both had the same start
address). The start location and size of the AGP aperture is set by
BIOS. PCI memory space is allocated afterwards, usually from a
location somewhere after the end of the AGP block.

According to dmesg, you reserved from 0x0e000000 to 0x0fffffff
inclusive. If you look at the /proc/iomem you posted (either one),
you'll notice the big jump in memory addresses between "Kernel data"
and "ACPI Non-volatile Storage". In the middle of this region is where
you reserved memory.

Somebody correct me if I'm wrong, but that's my understanding of things tho.



More information about the linux-pcmcia mailing list