Chrisian Richter
ae7ba0a377
preparing first release 1_0_4
2006-11-13 14:55:43 +00:00
Chrisian Richter
a340296f95
added makefile targets for very basic release management, based on cvs tags
2006-11-13 14:53:48 +00:00
Chrisian Richter
1b39ae1623
removed some useless logs
2006-11-02 09:40:53 +00:00
srichter
b2f839bd51
Applied patch from David Woodhouse:
...
- fix old mistake breaking HFC-PCI on bigendian (blame Simon)
- get rid of config.h. Breaks builds against old kernels, so minor regression
2006-10-20 14:10:25 +00:00
Chrisian Richter
7baab3c770
added dsp_poll option to handle different dsp_polls from hfc polls, they might be multiples of 80 if the Timersetting is set 100Hz in the Kernel, default is none the dsp will autosense what is necessary
2006-10-20 12:51:56 +00:00
Chrisian Richter
53e5e1d7c7
added notfatal option for misdn-init, in case lsusb is not installed, we still go on.
2006-10-16 11:40:20 +00:00
Chrisian Richter
1ed9289f48
Modified change from gkelleter a bit. If you change things in the TE State machine, check the whole code! Disconnects might also be sent out in other States like outgoing Setup.. i've did your change only in state 7 == In Setup + Out Alerting, please check if that works for you.
2006-10-16 07:43:57 +00:00
Chrisian Richter
60be2d76f8
removed a lot of verbosity. Better commented the pcm_slave and ignore_pcm_frameclock options
2006-10-09 12:51:33 +00:00
Daniel Potts
36ade5e5f5
James Harper identified that locks were being taken twice along the same code path, causing a hang (on SMP). These have been removed.
2006-10-02 12:51:33 +00:00
Chrisian Richter
7de26421e1
we work now with 32bit access on the fifos. Also for the b410P we don't adjust the txgain anymore since -4 is quite too much
2006-09-27 08:28:50 +00:00
Günther Kelleter
fca2dd211f
kernels < 2.6.9 don't implement kcalloc
2006-09-15 08:38:20 +00:00
Günther Kelleter
643710ddbf
endian fixes for BIGENDIAN systems. capi_profile and messages have to be little endian by definition
2006-09-14 15:51:46 +00:00
Günther Kelleter
2b7e0b8e84
in l3dss1_disconnect() the cause value was not taken from the DISCONNECT message, when it supplies one. Therefore an incoming DISCONNECT after incoming SETUP and outgoing ALERTING was not answered with RELEASE, since the cause was not evaluated properly. State 12 was entered instead of state 19
2006-09-14 15:47:22 +00:00
Günther Kelleter
e42aada6cf
set message type to MT_RELEASE_COMPLETE in message for CC_RELEASE_COMPLETE_REQ. MT_DISCONNECT was sent instead of MT_RELEASE_COMPLETE as required
2006-09-14 15:43:09 +00:00
Günther Kelleter
d87bf83950
fix potential skb overflow when user-user-info is sent in an ALERTING message
2006-09-14 15:41:05 +00:00
Günther Kelleter
f85a2af479
make compilable on non-i386 architectures
2006-09-14 15:39:52 +00:00
Günther Kelleter
0f20f5bdc6
fix debug output
2006-09-14 15:36:47 +00:00
Günther Kelleter
15f0d65a7f
fix typo in hex constant
2006-09-14 15:34:12 +00:00
Günther Kelleter
70914a8e6f
make kzalloc macro for kernels < 2.6.14 clear allocated memory
2006-09-14 15:30:42 +00:00
Chrisian Richter
216a4b1edd
litle fix to make the leds work when the watchdog is enabled
2006-09-13 15:53:01 +00:00
Chrisian Richter
5787705119
added a 2 clocking options for the hfc e1 chips
2006-09-13 14:01:04 +00:00
Martin Bachem
89626fbade
configure end of pulse control for ST mode (TE & NT)
2006-09-11 11:44:02 +00:00
Chrisian Richter
a9c1ddd31c
added some minipci hfcmulti based cards
2006-09-08 10:39:54 +00:00
Chrisian Richter
5d275ef7d7
don't print the NEXT_SKB EXISTS error.. for testing
2006-09-07 13:02:34 +00:00
Chrisian Richter
8ab0c38df3
added the new B410P HFC-4S based card. This card has Echocancelation in hardware. Modified mISDN_dsp to check for hw_echocan facilities. added possibility to control the gain from lower layers. Thanks to Matthew Fredrickson for most of the patches.
2006-09-06 17:24:22 +00:00
Nadi Sarrar
18df072e60
fixed typo
2006-08-25 11:50:41 +00:00
Martin Bachem
c9d11dda5f
rx skb mem range check
2006-08-15 09:38:07 +00:00
Chrisian Richter
01e13c7883
kzalloc is very new, only since 2.6.14
2006-08-10 15:00:03 +00:00
Karsten Keil
3606a753cd
fix protocol parameter handling (bug 0000067)
2006-08-07 23:35:59 +00:00
Daniel Potts
7fa0d32c3e
Pointer next_skb test fix.
...
Take lock on initialization in an attempt to fix SMP race.
2006-08-07 11:53:02 +00:00
Karsten Keil
b2629bc4b7
fix for overflow in copy_pid
2006-08-04 17:08:31 +00:00
Karsten Keil
71aa0d3dc1
some addons from root@netsentry.superset.co.za
...
1)added a new card to hfc_multi seems to be junghans ...
2)added hotplug support for hfcsusb by exporting it to the modules list to
be picked up by hotplug works realy nicely
3)dont compile in devfs bits for new kernel (2.6.18 does not have any devfs)
2006-08-01 11:25:10 +00:00
Chrisian Richter
cb7fbf931b
added new Subdevice ID for new 2E1 beroNet card with watchdog for relais for the tranparent switched mode.
2006-07-31 12:55:31 +00:00
Chrisian Richter
9d468d79be
moved misdn-init from /etc/init.d to /usr/sbin, added checks for bc / modprobe / ... in misdn-init
2006-07-28 12:12:55 +00:00
Chrisian Richter
6fa421dc88
T310 can be between 30-120 seconds. we use 120 seconds now.
2006-07-20 08:14:18 +00:00
Chrisian Richter
a217cc8071
reworked the echocancel_chunk function, it works now exactly as under zaptel and cancels the echo much better now.
2006-07-18 14:44:51 +00:00
Chrisian Richter
7c9d3427c2
removed unused variable
2006-07-14 15:53:53 +00:00
Chrisian Richter
50bfd90684
readded the class_create_file stuff, that didn't harm anything but created compiler warnings
2006-07-14 15:48:17 +00:00
Chrisian Richter
ab4597fde4
splitted the release_port function in a release_port and release_ports_hw functions, the ports_hw function disables all the hw related stuff and the release_port function removes lists and deregisters channels and stuff like that. Both functions are now only called in the unload module routine. On higher Kernelversions the release_port function is not called because it creates kernel oopses. ifdefed some sysfs_create_files.
2006-07-14 15:30:22 +00:00
Chrisian Richter
1c93a4f475
for kernels >2.6.10 we need to release the port in the RELEASE message, this is not everytime true, but it doesn't harm anything to do it.. only the leaving mSDNstackd processes. This really needs rework
2006-07-13 09:11:32 +00:00
Chrisian Richter
fed7fcf52d
remove printks
2006-07-13 07:22:27 +00:00
Chrisian Richter
86beb33815
default threshold of 100 is more accurate, also we unload the card modules after we've unloaded the layers
2006-07-12 15:57:24 +00:00
Chrisian Richter
25f37a1d92
made the dtmftreshold value changeable by modul parameter. This might help reducing wrong identified dtmftones.
2006-07-04 13:38:45 +00:00
Chrisian Richter
4d82d4de70
added the instance id for the l2mgr Log, so we can distinguish the Port which might be faulty
2006-07-03 12:44:37 +00:00
Chrisian Richter
f15296da42
added poll option for the hfcsusb driver, also we pass the poll option for this driver and for the dsp modul now from the misdn-init script. This fixes the no-sound issue with the hfcsub cards. Many Thanks for James Harper who gave us this patch. In the future we might add some sort of autoconfiguration of the packetsizes.
2006-07-03 11:48:02 +00:00
Chrisian Richter
3ead35f330
we can have much more the 4ports in the hfcmulti
2006-07-03 10:04:04 +00:00
Karsten Keil
5036dcc9bc
fix debug declaration
2006-06-29 10:40:23 +00:00
Karsten Keil
63a0047915
add some compatibility stuff
2006-06-29 09:11:29 +00:00
Karsten Keil
06b2b6d540
cleanup compatibility stuff
2006-06-29 09:11:08 +00:00
Karsten Keil
5d3fa852e7
fix warning about MODULE_PARM_T
2006-06-29 08:52:39 +00:00