[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <49783679.8010005@wincor-nixdorf.com>
Date: Thu, 22 Jan 2009 10:03:53 +0100
From: Niels de Vos <niels.devos@...cor-nixdorf.com>
To: Tosoni <jp.tosoni@...sys.fr>
CC: 'Michael Bramer' <michael@...support.de>,
'Alan Cox' <alan@...rguk.ukuu.org.uk>,
'Paulius Zaleckas' <paulius.zaleckas@...tonika.lt>,
linux-kernel@...r.kernel.org, linux-serial@...r.kernel.org
Subject: Re: [PATCH] drivers/serial/8250_pci.c (add support for '8-port RS-232
MIC-3620 from advantech'
Tosoni wrote:
>> On Wed, Jan 21, 2009 at 12:32:15PM +0000, Alan Cox wrote:
>>>>>> static struct pci_device_id serial_pci_tbl[] = {
>>>>>> + { PCI_VENDOR_ID_ADVANTECH,
>> PCI_DEVICE_ID_ADVANTECH_PCI3620,
>>>>>> + 0x3620, PCI_ANY_ID, 0, 0,
>>>> Why not use PCI_VENDOR_ID_ADVANTECH as PCI_SUBVENDOR_ID too?
>>> The Advantech vendor id is not 0x3620. This confused me as
>> well which is
>>> why I asked for an lspci. Advantech has stuck the device id in the
>>> subvendor bits and '1' in the subdevice (so it should be 1 not
>>> PCI_ANY_ID).
>> is this better?
>>
>> + { PCI_VENDOR_ID_ADVANTECH,
>> PCI_DEVICE_ID_ADVANTECH_PCI3620,
>> + PCI_DEVICE_ID_ADVANTECH_PCI3620, 1, 0, 0,
>
> Since the name describes a device id where it should be a (sub)vendor id,
> I would suggest that you add a line of comment to explain the case.
> So that no one will be tempted to change it back to PCI_VENDOR_ID_ADVANTECH
> in the future.
Definitely!
Niels
Download attachment "signature.asc" of type "application/pgp-signature" (190 bytes)
Powered by blists - more mailing lists