Cyril Plisko wrote:
> On Thu, Jan 29, 2009 at 9:15 PM, Garrett D'Amore <gdamore at sun.com> wrote:
>
>> Cyril Plisko wrote:
>>
>>> On Thu, Jan 29, 2009 at 7:37 PM, Jerry Gilliam
>>> <jg at jurassic.sfbay.sun.com> wrote:
>>>
>>>
>>>> - Output of 'prtconf -d'
>>>> * Actual PCI vendor ID and device ID on X86:
>>>> pci1028,214 (pciex8086,105e), instance #0
>>>> * and on SPARC:
>>>> network (pciex8086,105e), instance #0
>>>>
>>>>
>>>>
>>> I think that in many cases the next step would be to manually search
>>> in various tables what the actual device is. So why not do one more
>>> step and lookup these numbers in /usr/share/hwdata/pci.ids ? Perhaps,
>>> while combined with -v?
>>>
>>> I.e.
>>>
>>> - Output of 'prtconf -vd'
>>>
>>> * Actual PCI vendor ID and device ID on X86:
>>> pci1028,214 (pciex8086,105e [Intel 82571EB Gigabit Ethernet
>>> Controller]), instance #0
>>> * and on SPARC:
>>> network (pciex8086,105e [Intel 82571EB Gigabit Ethernet
>>> Controller]), instance #0
>>>
>>>
>> Seems like that overlaps with scanpci. If we could share the PCI id table
>> between these projects, then I'd be happy. But I'm loathe to create yet
>> another copy of the PCI ID mapping table that has to be maintained
>> separately (and can get out of sync with the others.)
>>
>>
>
> That's why I specifically mentioned /usr/share/hwdata/pci.ids. It is
> in this file where scanpci (libpciaccess to be more precise) gets all
> the info from.
>
>
Okay, I'll +1 the project -- either with the name mapping, or without.
(It'd be nice to have it, but it isn't, IMO, absolutely critical.)
-- Garrett
