>> >The pcibios_find_class() function quite often does NOT WORK!
>> >
>> >I tried to use it once within the IDE driver, but had trouble
>> >with buggy BIOSs.
>>
>> Hmmm ... that looks bad. The problem is it seems to run ok
>> with every other damn PCI device. Is there anywhere a test
>> program to check if my PCI-BIOS is buggy?
>
>Yes in Kernel 2.0.30 there were introduced some functions for accessing
>the configuration information without using the pci-bios.
>They are named pci_direct_find_class() ..... (see bios32.c)
>You can try this. If it works with this function, but not with your BIOS,
>it is buggy.
As it seems, it reads the pci_devices info, this is, directly
from the kernel data structures. This seems to be ok, as the
/proc/pci shows something that seems plausible :) , and it gets
it from there.
This is what makes me suspect that the PCI BIOS looks buggy
... grrrr
>Also you can read PCI-Configurationspace from user space. For example
>scanpci from XFree can be used. If your chipset has a type1 configuration
>space access mechanism you also can use the program pciprobe from
>the devicedriver for my framegrabber mv1000drv-0.37.tgz. This does
>a dump of all 256 bytes from the config space of all devices. So
>you can look at the output and see if the configspace of your device is ok.
Hmmm ... never had thought'bout it. Thanks :) will try it.
-- Linux USB is in development -- http://www2.teknoland.es/~inaky
Inaky Perez Gonzalez -- PGP pubkey fingerprint -
inaky@peloncho.fis.ucm.es 2:341/5.31 -- 8E 34 3A 62 64 99 E2 44 -
http://peloncho.fis.ucm.es/~inaky -- AD 7B 30 D9 DD FF 3E 4C -
-----------------------------------------------------------------
The loneliness of the long distance runner .....