Table of PCIs, chipsets, etc. on b43 wiki page

[Rafał Miłecki]

W dniu 5 lutego 2011 18:35 użytkownik Rafał Miłecki <zajec5 at gmail.com> napisał:
> W dniu 5 lutego 2011 16:56 użytkownik Larry Finger
> <Larry.Finger at lwfinger.net> napisał:
>> On 02/04/2011 11:56 PM, Gábor Stefanik wrote:
>>> 2011/2/5 Larry Finger <Larry.Finger at lwfinger.net>:
>>>> On 02/04/2011 02:27 PM, Rafał Miłecki wrote:
>>>>> W dniu 4 lutego 2011 20:37 użytkownik Larry Finger
>>>>> <Larry.Finger at lwfinger.net> napisał:
>>>>>> On 02/04/2011 12:56 PM, Rafał Miłecki wrote:
>>>>>>> I'd like to propose little (nothing big) reogranisation of b43 hardware table:
>>>>>>> http://wireless.kernel.org/Zajec?action=recall&rev=4
>>>>>>>
>>>>>>> Changed columns support/supported and driver/alternatives.
>>>>>>>
>>>>>>> Are you fine with that? Can we use colored one?
>>>>>>
>>>>>> I like the colored one.
>>>>>>
>>>>>> Some small changes:
>>>>>>
>>>>>> I would eliminate the second 14e4:4312 line and move the wl entry to the first
>>>>>> one. I don't know if wl supports 5 GHz on that chip. I don't have mine any longer.
>>>>>
>>>>> Do you know anything about that duplicated ID? Is that true at all?
>>>>> Why we don't support device with that duplicated ID? Why do you want
>>>>> to remove it?
>>>>
>>>> I'm not sure that there really is an 802.11b/g device with ID 14e4:4312. There
>>>> are two flavors of BCM4312, one that is 802.11a/b/g (14e4:4312), and an
>>>> 802.11b/g device with ID 14e4:4315, which is the LP PHY. I suspect that the
>>>> confusion over the two BCM4312 items led to the duplicate entry in this table.
>>>
>>> The device with PCI ID 0x4312 is called BCM4311 (dual-band) - check
>>> Broadcom's official brochure on BCM4311 (has a single-band and a
>>> dual-band version). Same for 0x4319 - it's actually BCM4318 dual-band
>>> (for example, the BCM94318MPAGH has this ID). The only BCM4312 is
>>> 0x4315 (also, there exists no "BCM4315", and while BCM4310 might
>>> indeed be real, it has nothing to do with PCI ID 0x4315). Apparently
>>> the correlation between PCI ID and chip ID only exists for pure G-PHY
>>> cards.
>>>
>> Check Comment #6 in the thread at
>> http://forums.opensuse.org/forums/english/get-technical-help-here/wireless/453614-wireless-card-dormant-until-wired-connection-detected.html.
>> A BCM4312 with ID 14e4:4312.
>
> It's like from PCI database. Please check comment #17:

I meant "lie" of course. I also Googled for:
"14e4:4312" "Broadcom 4312 wlan found"
and then for:
"14e4:4312" "Broadcom 4311 wlan found"
the first one won't give you any real, non-false, results.

According to http://pci-ids.ucw.cz/read/PC/14e4/4312 john.s.gustafsson
submitted name for this PCI ID containing "BCM4312". He probably
though PCI ID is in direct relation with Broadcom chipset used on
card.

john.s.gustafsson referred to link:
https://lists.berlios.de/pipermail/bcm43xx-dev/2007-March/004312.html
that does not work. I searched mails from March 2007, but didn't find
anything related.

We really need to treat this as mistake and fix table and pci database.

-- 
Rafał