New uhso(4) device: Globetrotter HSUPA Modem Option N.V.

[Brandon Gooch]

On Mon, Feb 21, 2011 at 3:57 PM, Bjoern A. Zeeb <bz at freebsd.org> wrote:
> On Mon, 21 Feb 2011, Brandon Gooch wrote:
>
>> On Mon, Feb 21, 2011 at 1:28 PM, Brandon Gooch
>> <jamesbrandongooch at gmail.com> wrote:
>>>>
>>>> Yeah, I got that message.  Not connecting and not getting an serial
>>>> console is a bit different.  From the output above it looks like
>>>> random garbage on the serial port.
>>>>
>>>> Try connecting to all different serial ports (without uhsoctl running)
>>>> with minicom or cu and try typing AT<enter>. (The diagnostic port
>>>> doesn't work, so skip that one).
>>>
>>> I just encountered a panic when the driver attaches after plugging in
>>> the device.
>>>
>>> The panic stems from uhso_attach(), and seems due to my kernel having
>>> the VIMAGE option compiled in -- it doesn't panic on my non-VIMAGE
>>> kernel (which I need to rebuild to continue helping debug).
>>>
>>> I'm trying to get a textdump ATM...
>>
>> Looks like I exceeded the VNET if_indexlim in /usr/src/sys/net/if.c on
>> line 190:
>>
>> static VNET_DEFINE(int, if_indexlim) = 8;
>>
>> Maybe I'll bump it up and give it another go...
>>
>> Bjoern, is there any reason I shouldn't be able to increase the number
>> from 8 to say, 16?
>
> I am lacking context reading about serial ports and network
> interfaces.

I apologize for just "blurting" this out at you, my bad. I get a
little anxious sometimes :{

> If you create a network interface from USB you are currently running
> into the problem that CURVNETs are not properly setup.  The indexlim,
> should just increase itself up to 64k if needed - see if_grow() in
> if.c.

...and of course, I sound like a dork again for making absurd
assertions about code I know nothing about LOL :)

It totally slipped my mind that all of the bits are not yet in the
tree to allow adding interfaces after the kernel is already running --
patience, patience...

-Brandon