|
List Info
Thread: NetBSD driver update: Help?
|
|
| NetBSD driver update: Help? |

|
2006-09-15 23:16:41 |
Hello Peter. Yes, I care. I'm running ASterisk zaptel on
NetBSD, and
I'm still running NetBSD-2.0, with the pre-port version of
the zaptel and
ztdummy drivers. I did some work on the stuff that went
into the pkgsrc
tree, corresponding with Jeff while he was doing the work,
and I'd be happy
to help out in any way I can, trying to answer specific
questions, loading
test drivers on a test box.
Questions for you:
1. Are you trying to make something compatible with 4.x and
higher, or are
you still including 3.x compatibility?
2. What lead you to believe it was best to re-port the
entire driver,
rather than just patching the changes in?
My understanding is that FreeBSD has evolved quite a bit in
the
interveening two years, and so if you could just distill
what functional
additions were made, it might be easier to just add those to
the existing
working infrastructure.
-Brian
On Sep 15, 1:57pm, Peter Seebach wrote:
} Subject: Re: [Asterisk-bsd] NetBSD driver update: Help?
} In message <200609122040.k8CKekM6005817 guild.plethora.net>, Peter Seebach writ
} es:
} >If I load wcfxs, I generally get panics or hangs, and
I don't think I
} >understand the code entirely well enough to get a good
start on debugging
} >this. I would appreciate second opinions. It looks
as though the interrupt
} >routine is what's causing the most trouble, but I'm
not sure. (On a secondary
} >note; is it normal for a 3215-based FXO module on the
card to fail to probe
} >if there's no phone line on it yet? It gave some
warning about failing to
} >come ready.)
}
} Hi! I am still stumped.
}
} Does anyone have ANY input? Suggestions? Even just a
basic understanding
} of what the driver does and how the card works, so I can
ask questions about
} what a given chunk of code is supposed to do?
}
} I am reliably getting various weirdness, such as
"kernel hangs completely
} when driver is loaded" or "panic in unrelated
parts of kernel", all of which
} suggest to me that there's some key step in porting from
FreeBSD that I don't
} know about, but I don't really have anything to go on
here.
}
} Any interest? Is the zaptel-BSD code still maintained?
Does anyone care
} whether it runs on *BSD other than FreeBSD/x86?
}
} -s
} _______________________________________________
} Asterisk-BSD mailing list
} Asterisk-BSD lists.digium.com
} http://lists.digium.com/mailman/listinfo/asterisk-bsd
>-- End of excerpt from Peter Seebach
_______________________________________________
Asterisk-BSD mailing list
Asterisk-BSD lists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-bsd
|
|
| NetBSD driver update: Help? |

|
2006-09-16 01:21:05 |
In message <200609152316.k8FNGfue012494 lothlorien.nfbcal.org>, Brian Buhrow wr
ites:
>1. Are you trying to make something compatible with 4.x
and higher, or are
>you still including 3.x compatibility?
3.0.1.
>2. What lead you to believe it was best to re-port the
entire driver,
>rather than just patching the changes in?
Major refactoring in the FreeBSD driver.
>My understanding is that FreeBSD has evolved quite a bit
in the
>interveening two years, and so if you could just distill
what functional
>additions were made, it might be easier to just add
those to the existing
>working infrastructure.
This way, I now have a single unified source tree that, with
a bit of kicking,
should compile on both systems, because I #ifdef'd all my
changes one way or
another.
I think. I may have missed some.
Anyway, I finally tracked down the last thing; the DMA
allocation was
different between the drivers, so the NetBSD driver was
dutifully trying to
dereference the null pointers that would have been
initialized by the FreeBSD
code.
I now have a driver which loads just fine with a TDM400P
(two FXS, two FXO):
Zapata Telephony Interface Registered on major 196
Echo Canceller: MARK2
FXS device: vendor=e159 product=1 subvendor=b119
wcfxs0 at pci0 dev 19 function 0: Wildcard TDM400P REV I
Registered Span 1 ('WCTDM/0') with 4 channels
Span ('WCTDM/0') is new master
wcfxs0: interrupting at irq 11
Freshmaker version: 73
Freshmaker passed register test
Module 0: Installed -- AUTO FXS
Module 1: Installed -- AUTO FXS
VoiceDAA did not bring up ISO link properly!
Module 2: Not installed
VoiceDAA did not bring up ISO link properly!
Module 3: Not installed
This leads to the question: Do I need a phone line for the
FXO modules to
come up correctly? I was hoping to have this as close as
possible to fully
debugged before I connected it to the phone lines.
-s
_______________________________________________
Asterisk-BSD mailing list
Asterisk-BSD lists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-bsd
|
|
| NetBSD driver update: Help? |

|
2006-09-16 02:58:12 |
Followup: It works well enough for "ztcfg" to
look plausible.
http://ww
w.seebs.net/tmp/zaptel-bsd-mod.tgz
The wcfxs directory contains my patched driver, the
wcfxs.orig directory
contains the FreeBSD driver unpatched. Same for ztdummy and
zaptel.
Not tested very much, no warranty, and since I don't know
anything about
asterisk, I can't test it either.
-s
_______________________________________________
Asterisk-BSD mailing list
Asterisk-BSD lists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-bsd
|
|
| NetBSD driver update: Help? |

|
2006-09-18 02:56:05 |
peter,
i have a Wildcard TDM400P REV E/F (2 FXS, 2 FXO) that loads
find in
NetBSD 3.0 using the current comms/zaptel-netbsd package
with one tiny
hand made patch.
> This leads to the question: Do I need a phone line
for the FXO
> modules to come up correctly?
No, you do not. my dmesg shows the following with no
phones/lines connected
FXS device: vendor=e159 product=1 subvendor=b100
wcfxs0 at pci2 dev 9 function 0: Wildcard TDM400P REV E/F
zaptel at wcfxs0 not configured
wcfxs0: interrupting at irq 10
Registered Span 1 ('WCTDM/0') with 4 channels
Span ('WCTDM/0') is new master
Freshmaker version: 71
Freshmaker passed register test
Module 0: Installed -- AUTO FXS
Module 1: Installed -- AUTO FXS
Module 2: Installed -- AUTO FXO
Module 3: Installed -- AUTO FXO
generate_desc, cards = 4
generate_desc, cards = 4
i can have a go at installing your zaptel-bsd-mod.tgz and
tell you how
things go.
michael.
Peter Seebach wrote:
> In message <200609152316.k8FNGfue012494 lothlorien.nfbcal.org>, Brian Buhrow wr
> ites:
>> 1. Are you trying to make something compatible
with 4.x and higher, or are
>> you still including 3.x compatibility?
>
> 3.0.1.
>
>> 2. What lead you to believe it was best to re-port
the entire driver,
>> rather than just patching the changes in?
>
> Major refactoring in the FreeBSD driver.
>
>> My understanding is that FreeBSD has evolved quite
a bit in the
>> interveening two years, and so if you could just
distill what functional
>> additions were made, it might be easier to just add
those to the existing
>> working infrastructure.
>
> This way, I now have a single unified source tree that,
with a bit of kicking,
> should compile on both systems, because I #ifdef'd all
my changes one way or
> another.
>
> I think. I may have missed some.
>
> Anyway, I finally tracked down the last thing; the DMA
allocation was
> different between the drivers, so the NetBSD driver was
dutifully trying to
> dereference the null pointers that would have been
initialized by the FreeBSD
> code.
>
> I now have a driver which loads just fine with a
TDM400P (two FXS, two FXO):
>
> Zapata Telephony Interface Registered on major 196
> Echo Canceller: MARK2
> FXS device: vendor=e159 product=1 subvendor=b119
> wcfxs0 at pci0 dev 19 function 0: Wildcard TDM400P REV
I
> Registered Span 1 ('WCTDM/0') with 4 channels
> Span ('WCTDM/0') is new master
> wcfxs0: interrupting at irq 11
> Freshmaker version: 73
> Freshmaker passed register test
> Module 0: Installed -- AUTO FXS
> Module 1: Installed -- AUTO FXS
> VoiceDAA did not bring up ISO link properly!
> Module 2: Not installed
> VoiceDAA did not bring up ISO link properly!
> Module 3: Not installed
>
> This leads to the question: Do I need a phone line for
the FXO modules to
> come up correctly? I was hoping to have this as close
as possible to fully
> debugged before I connected it to the phone lines.
>
> -s
> _______________________________________________
> Asterisk-BSD mailing list
> Asterisk-BSD lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-bsd
>
>
>
--
Omniscient Australia Pty Ltd
(03) 9614 0112
http://www.omniscient.co
m.au
_______________________________________________
Asterisk-BSD mailing list
Asterisk-BSD lists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-bsd
|
|
[1-4]
|
|