Zaptel oddities

I'm playing with FreeBSD 6.0 and Asterisk 1.2.7.1, with a
checkout of zaptel-bsd
from the 1.2.7.1 timeframe. (If someone could tell me how to
check which SVN
revision it is, I'll be happy to add that).

For hardware, I have one Digium analog card with one FXS and
one FXO daughter
card onboard, and a "generic" FXO clone.

I added the following into /boot/loader.conf, so I could
load "automatically":
zaptel_load = "YES"
wcfxs_load = "YES"
wcfxo_load = "YES"

I'm not sure this is the "proper" way to do
this (should it be in hints? loaded
manually by kldload later?), but I see in the probes:

Jun  9 04:19:49  kernel: Zapata Telephony Interface
Registered on major 196
Jun  9 04:19:49  kernel: Echo Canceller: MARK2
Jun  9 04:19:49  kernel: FXS device: vendor=e159 device=1
subvendor=b100
Jun  9 04:19:49  kernel: wcfxs0: <Wildcard TDM400P REV
E/F> port 0xd000-0xd0ff m
em 0xdfffd000-0xdfffdfff irq 12 at device 16.0 on pci0
Jun  9 04:19:49  kernel: FXS Attach for wcfxs0: deviceID :
0xe159
Jun  9 04:19:49  kernel: wcfxs0: [FAST]
Jun  9 04:19:49  kernel: Freshmaker version: 71
Jun  9 04:19:49  kernel: Freshmaker passed register test
Jun  9 04:19:49  kernel: Module 0: Installed -- AUTO FXS
Jun  9 04:19:49  kernel: ProSLIC sanity check failed
Jun  9 04:19:49  kernel: Module 1: Not installed
Jun  9 04:19:49  kernel: ProSLIC sanity check failed
Jun  9 04:19:49  kernel: Module 2: Not installed
Jun  9 04:19:49  kernel: Module 3: Installed -- AUTO FXO
Jun  9 04:19:49  kernel: Found a Wildcard TDM: Wildcard
TDM400P REV E/F (4 modul
es)
Jun  9 04:19:49  kernel: ZapTel device: vendor=e159 device=1
subvendor=8086
Jun  9 04:19:49  kernel: wcfxo0: <Generic Clone> port
0xcc00-0xccff mem 0xdfffc0
00-0xdfffcfff irq 5 at device 20.0 on pci0
Jun  9 04:19:49  kernel: ZapTel Attach for wcfxo0: deviceID
: 0xe159
Jun  9 04:19:49  kernel: Can't use fast interrupts, falling
back to normal
Jun  9 04:19:49  kernel: wcfxo0: [GIANT-LOCKED]
Jun  9 04:19:49  kernel: Failed to initailize DAA, giving
up...
Jun  9 04:19:49  kernel: Freed a Wildcard
Jun  9 04:19:49  kernel: ZapTel detach!
Jun  9 04:19:49  kernel: We failed: 5
Jun  9 04:19:49  kernel: device_attach: wcfxo0 attach
returned 5

When I run ztcfg on boot, of course the one card that
initilized properly is
up, and the other card gives an error.

If I "kldunload" and then "kldload"
the fxo device, all is happy and ztcfg
runs without errors. Is this normal? Did I munge something
up?

Also, when I'm running, I get these occasionally:
Jun  7 20:03:56 WARNING[480] chan_zap.c: zt hook failed:
Device busy
Jun  7 21:20:25 WARNING[480] chan_zap.c: Unable to reset
default ring on 'Zap/1-
1'
Jun  7 22:38:00 NOTICE[480] app_dial.c: Unable to create
channel of type 'Zap' (
cause 17 - User busy)

I'm not sure why this is. I did have the card running under
linux, and I never
saw these kinds of issues, so my first inkling is that it
might be something
related to the driver; but it could be my setup, so I
thought I'd ask.

Thanks,
J.
_______________________________________________
Asterisk-BSD mailing list
Asterisk-BSDlists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-bsd

A while back, I wrote:

> I'm playing with FreeBSD 6.0 and Asterisk 1.2.7.1,
with a checkout of 
> zaptel-bsd from the 1.2.7.1 timeframe. (If someone
could tell me how 
> to check
> which SVN revision it is, I'll be happy to add that).
>
> For hardware, I have one Digium analog card with one
FXS and one FXO daughter
> card onboard, and a "generic" FXO clone.
>
> When I run ztcfg on boot, of course the one card that
initilized properly is
> up, and the other card gives an error.
>
> If I "kldunload" and then
"kldload" the fxo device, all is happy and ztcfg
> runs without errors. Is this normal? Did I munge
something up?
>
> Also, when I'm running, I get these occasionally:
> Jun  7 20:03:56 WARNING[480] chan_zap.c: zt hook
failed: Device busy
> Jun  7 21:20:25 WARNING[480] chan_zap.c: Unable to
reset default ring 
> on 'Zap/1-1'
> Jun  7 22:38:00 NOTICE[480] app_dial.c: Unable to
create channel of 
> type 'Zap' (cause 17 - User busy)

I'm still getting this, but I did manage to find a sort of
unrelated issue --
I found out my onboard ethernet is sharing interrupts with
the clone card,
by grepping through dmesg. I tried some bios tricks, but
that seems to just
move the interrupt they share around, not actually get them
to not share.

The reason I'm trying to track this down is that I'm
having some strange
issues with CallerID and popping/echos, and I thought it
might be the driver
that is causing it. Now, some of this might actually be
related to missed
interrupts, but I'm not sure if there is anything I can do,
FreeBSD wise,
to get FreeBSD not to allocate two devices the same IRQ. (I
tried turning
some of the unneeded stuff off in BIOS, but that doesn't
seem to do 
anything --
is there something BIOS or motherboard related that would
tie a PCI slot
and something onboard (LAN) to the same IRQ?)

If anyone has a suggestion, I'd be happy to try.

Thanks,
J

_______________________________________________
Asterisk-BSD mailing list
Asterisk-BSDlists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-bsd

On Jun 13, 2006, at 11:40 PM, jerryvoiptower.com wrote:

> is there something BIOS or motherboard related that
would tie a PCI  
> slot
> and something onboard (LAN) to the same IRQ?)


Some BIOS let you assign specific PCI slots (including those
non- 
physical ones like onboard LAN) to specific interrupts if I
remember  
the screen correctly.  I have never had to mess around with
it but  
have seen the screen on several of my servers.

Chad
_______________________________________________
Asterisk-BSD mailing list
Asterisk-BSDlists.digium.com
http://lists.digium.com/mailman/listinfo/asterisk-bsd