On 2/4/2013 4:32 PM, Charles Sprickman wrote:
> On Feb 4, 2013, at 5:00 PM, Ian Lepore wrote:
>
>> On Mon, 2013-02-04 at 16:31 -0500, Charles Sprickman wrote:
>>> On Feb 4, 2013, at 4:13 PM, Ian Lepore wrote:
>>>
>>>> On Mon, 2013-02-04 at 14:58 -0600, Karl Denninger wrote:
>>>>> On 2/4/2013 2:06 PM, Ian Lepore wrote:
>>>>>> On Mon, 2013-02-04 at 12:57 -0600, Karl Denninger wrote:
>>>>>>> ... and plug it into FreeBSD 9.1-Stable with the rev ID FreeBSD
>>>>>>> 9.1-STABLE #16 r244942
>>>>>>>
>>>>>>> and it returns....
>>>>>>>
>>>>>>> ugen4.4: <vendor 0x0409> at usbus4
>>>>>>> uhub6: <vendor 0x0409 product 0x0050, class 9/0, rev 2.00/1.00, addr 4>
>>>>>>> on usbus4
>>>>>>> uhub_attach: port 1 power on failed, USB_ERR_STALLED
>>>>>>> uhub_attach: port 2 power on failed, USB_ERR_STALLED
>>>>>>> uhub_attach: port 3 power on failed, USB_ERR_STALLED
>>>>>>> uhub_attach: port 4 power on failed, USB_ERR_STALLED
>>>>>>> uhub_attach: port 5 power on failed, USB_ERR_STALLED
>>>>>>> uhub_attach: port 6 power on failed, USB_ERR_STALLED
>>>>>>> uhub_attach: port 7 power on failed, USB_ERR_STALLED
>>>>>>> uhub6: 7 ports with 7 removable, self powered
>>>>>>>
>>>>>>> Yuck.
>>>>>>>
>>>>>>> The last time it was working was on a FreeBSD 7 box (yeah, I know,
>>>>>>> rather old) but I never had problems there.  And it appears that all of
>>>>>>> the device declarations that I used to have to put in the kernel as
>>>>>>> non-standard stuff are now in GENERIC, so I would expect it to work.
>>>>>>>
>>>>>>> Ideas as to what may have gotten hosed up here?
>>>>>>>
>>>>>> Those messages all seem to be related to a hub. Vendor ID 0x0409 is NEC.
>>>>>>
>>>>>> FTDI's vendor ID is 0x0403, and FTDI stuff works fine in FreeBSD 9 and
>>>>>> 10; I use it all the time.  Sometimes aftermarket vendors who use FTDI's
>>>>>> parts program different vendor/product info and IDs have to be added to
>>>>>> code to recognize them, that's the only trouble one usually encounters.
>>>>>>
>>>>>> -- Ian
>>>>> Well, that sorta kinda worked. 
>>>>>
>>>>> Except that it still is identifying it as a hub too, and the two collide
>>>>> and crash the stack.
>>>>>
>>>>> But I can't find anything that is looking at the PID (0x0050) or the
>>>>> definition (HUB_0050) anywhere in the code. 
>>>>>
>>>>> I'll go pull the NEC defs and set up something else instead of simply
>>>>> adding it to the FTDI probe list.
>>>>>
>>>> It seems to me you have a problem with a hub (perhaps the root hub or a
>>>> motherboard hub if you don't have an external one) and this has nothing
>>>> to do with the ftdi device at all.
>>> I assume we're talking about a multi-port usb to serial adapter, correct?
>>>
>>> If so, they generally do have a hub included in the device.
>>>
>>> Example:
>>>
>>> ugen1.3: <vendor 0x0409> at usbus1
>>> uhub4: <vendor 0x0409 product 0x0050, class 9/0, rev 2.00/1.00, addr 3> on 
>>> usbus1
>>> uhub4: 7 ports with 7 removable, self powered
>>>
>>> Then the individual ports look like this:
>>>
>>> ugen1.4: <FTDI> at usbus1
>>> uftdi0: <FT232R USB UART> on usbus1
>>> ugen1.5: <FTDI> at usbus1
>>> uftdi1: <FT232R USB UART> on usbus1
>>> (etc.)
>>>
>>> We use these for serial console ports, they're (relatively) cheap and have 
>>> generally been well supported.
>>>
>>> The above info is from an 8.3 box.
>>>
>>> Just wanted to clarify that there is likely a hub in the serial box Karl is 
>>> working with…
>>>
>>> Charles
>> Oh, interesting.  The biggest ftdi dongle I have is 4 ports, using the
>> ftdi 4232 chip.  I guess to get more ports than that, folks are now
>> using an internal hub and multiple ftdi chips.
> These multiport things have been around for a long time.  Someone at ISC 
> recommended them when we were looking to replace some unsupported RocketPort 
> cards.  Not affiliated with this place, but it's the largest collection of 
> USB to serial stuff I've ever seen (and they document for the most part what 
> chips are involved):
>
> http://usbgear.com/USB-Serial.html
>
> Our first 16 ports are on one of these:
>
> http://usbgear.com/computer_cable_details.cfm?sku=USB-16COM-RM&cats=199&catid=493%2C494%2C474%2C199%2C461%2C106%2C1009%2C601
> (the tx/rx blinky lights are handy in troubleshooting)
>
> Then the rest on this cheaper model:
>
> http://usbgear.com/computer_cable_details.cfm?sku=USBG-8COM-M&cats=199&catid=494%2C199%2C474%2C2345%2C1009
>
>> So for some reason there's a problem with the hub, and that's probably
>> preventing it from getting as far as seeing the ftdi parts that are
>> downstream of that.
> My dmesg snippet is from the latter box.  Note that the vendor and product ID 
> are the same as Karl's.  Perhaps there is a regression, as I am running 8.3 
> and have had no issues there (previously it was on a 4.11 box).
>
> Charles
>
That's the EXACT box.

I've used them for a VERY long time on FreeBSD and they have always
worked perfectly well, but never on 9.x until now -- and it doesn't work
on 9.x.

Had to run out for a while -- continuing testing.

-- 
-- Karl Denninger
/The Market Ticker ®/ <http://market-ticker.org>
Cuda Systems LLC
_______________________________________________
[email protected] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "[email protected]"

Reply via email to