Re: umodem(4) with "no pointer to data interface"

[Ryo ONODERA <ryo_on%yk.rim.or.jp@localhost>]

From: Felix Deichmann <m4j0rd0m0%gmail.com@localhost>, Date: Tue, 21 Jan 2014 
20:11:19 +0100

> Am 21.01.2014 13:15, schrieb Brett Lymn:
>> Interface class is UICLASS_CDC, subclass
>> UISUBCLASS_ABSTRACT_CONTROL_MODEL
>> protocol is UIPROTO_CDC_AT.  Umodem should attach on this.  Maybe I
>> borked the match... try changing:
>>
>>          if (uaa->class != UICLASS_CDC ||
>>              uaa->subclass != UISUBCLASS_ABSTRACT_CONTROL_MODEL ||
>>              !(uaa->proto == UIPROTO_CDC_NOCLASS || uaa->proto ==
>>              !UIPROTO_CDC_AT))
>>
>> [...]
>>
>> see if that attaches... if it does then perhaps try:
>>
>>          if (uaa->class != UICLASS_CDC ||
>>              uaa->subclass != UISUBCLASS_ABSTRACT_CONTROL_MODEL ||
>>              (uaa->proto != UIPROTO_CDC_NOCLASS && uaa->proto !=
>>              UIPROTO_CDC_AT))
> 
> IMHO your *logic* is not to blame.
> Applying De Morgan's laws result in
> 
> !(uaa->proto == UIPROTO_CDC_NOCLASS || uaa->proto == UIPROTO_CDC_AT)
> 
>    and
> 
> (uaa->proto != UIPROTO_CDC_NOCLASS && uaa->proto != UIPROTO_CDC_AT)
> 
> being equal.
> 
> Must be something else then.
> Please also have a look at this site, issue seems identical:
> http://daemonforums.org/showthread.php?p=46644
> 
> Conexant has quirks?

Hi,

Thanks for your information.

According to this web site, OpenBSD can handle this kind of USB modem
properly.

http://www.openbsd.org/cgi-bin/cvsweb/src/sys/dev/usb/umodem.c?annotate=1.54
It seems that 246th line and later handles my problem.

OpenBSD's umodem_get_caps is different from NetBSD's one, and
it has "int *data_iface_no" argument. It is used for detecting
"no pointer to data interface".

--
Ryo ONODERA // ryo_on%yk.rim.or.jp@localhost
PGP fingerprint = 82A2 DC91 76E0 A10A 8ABB  FD1B F404 27FA C7D1 15F3