osmocom-bb.git
13 years agoRenamed gps_* functions to osmo_gps_* functions to avoid overlap with libgps functions.
Dario Lombardo [Thu, 10 Feb 2011 08:56:33 +0000 (09:56 +0100)]
Renamed gps_* functions to osmo_gps_* functions to avoid overlap with libgps functions.

13 years agotarget/fw: correct IO_CNTL_REG name and small cosmetic fixes
Steve Markgraf [Sun, 6 Feb 2011 16:33:59 +0000 (17:33 +0100)]
target/fw: correct IO_CNTL_REG name and small cosmetic fixes

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years ago[l1ctl] Adding missing msgb_free()
Andreas.Eversberg [Thu, 27 Jan 2011 16:33:04 +0000 (16:33 +0000)]
[l1ctl] Adding missing msgb_free()

When a corrupt frame cannot be delivered, it is dropped. Also it must be freed.

13 years agowireshark: update SMSCB dissector patch
Alex Badea [Wed, 26 Jan 2011 18:21:07 +0000 (20:21 +0200)]
wireshark: update SMSCB dissector patch

Add full support for reassembly, and dissection of Schedule messages.

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agocell_log: Added command line switches to change default gps device and baud rate.
Dario Lombardo [Wed, 26 Jan 2011 08:30:14 +0000 (09:30 +0100)]
cell_log: Added command line switches to change default gps device and baud rate.

13 years agofirmware: Add support for the Pirelli DP-L10 triband GSM/Wifi phone
Steve Markgraf [Mon, 24 Jan 2011 15:59:04 +0000 (16:59 +0100)]
firmware: Add support for the Pirelli DP-L10 triband GSM/Wifi phone

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agofw/app/layer1: unset the inverted attribute of the display after layer1_init()
Steve Markgraf [Tue, 18 Jan 2011 19:50:36 +0000 (20:50 +0100)]
fw/app/layer1: unset the inverted attribute of the display after layer1_init()

This attribute is toggled with the RTC interrupt, which is disabled
in layer1_init(). If an interrupt between rtc_init() and layer1_init()
occured, the display of the E88 phones remained inverted

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agofw/board/gta0x: move rffe configuration to board/gta0x
Steve Markgraf [Tue, 18 Jan 2011 16:18:11 +0000 (17:18 +0100)]
fw/board/gta0x: move rffe configuration to board/gta0x

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agofw/makefile: enable generation of Compal E86 images (Motorola C139/C140)
Steve Markgraf [Tue, 18 Jan 2011 15:51:41 +0000 (16:51 +0100)]
fw/makefile: enable generation of Compal E86 images (Motorola C139/C140)

This target now has its own init.c which correctly initializes the display
and the keypad backlight

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agofw/board/gta0x: remove unused define in RFFE configuration
Steve Markgraf [Mon, 24 Jan 2011 15:53:47 +0000 (16:53 +0100)]
fw/board/gta0x: remove unused define in RFFE configuration

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agofw/board/gta0x: implement TX support for gta0x devices
Steve Markgraf [Tue, 18 Jan 2011 15:27:56 +0000 (16:27 +0100)]
fw/board/gta0x: implement TX support for gta0x devices

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agotarget/fw: set the TRF6151 tsp IDs from the rffe initialization
Steve Markgraf [Mon, 24 Jan 2011 15:48:56 +0000 (16:48 +0100)]
target/fw: set the TRF6151 tsp IDs from the rffe initialization

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agoccch: Print paging of tyep2
Holger Hans Peter Freyther [Sun, 23 Jan 2011 13:29:30 +0000 (14:29 +0100)]
ccch: Print paging of tyep2

13 years agoccch: Decode the paging1 type paging requests
Holger Hans Peter Freyther [Sun, 23 Jan 2011 13:16:54 +0000 (14:16 +0100)]
ccch: Decode the paging1 type paging requests

Th size checks are a bit messy and I have not seen an optional
MI yet. So this code path is not tested at all.

13 years agoccch: Print the decoded immediate assignment again, ignore two messages
Holger Hans Peter Freyther [Sun, 23 Jan 2011 11:20:31 +0000 (12:20 +0100)]
ccch: Print the decoded immediate assignment again, ignore two messages

Decode the immediate assignment and print the messages as a
notice, ignore the Notification for voice group services and
use a magic number for SI 2quater that I could not find in my
version of the spec.

13 years agoccch: Stop using fprintf for the SI
Holger Hans Peter Freyther [Sun, 23 Jan 2011 11:06:35 +0000 (12:06 +0100)]
ccch: Stop using fprintf for the SI

Stop using fprintf to print the System Information number of the
data we get. For the check use LOGP with LOGL_ERROR. There is
little use in this information.

13 years agologging: Move DRSL, DLAPDM, DL1C to LOGL_NOTICE as default
Holger Hans Peter Freyther [Sun, 23 Jan 2011 11:05:39 +0000 (12:05 +0100)]
logging: Move DRSL, DLAPDM, DL1C to LOGL_NOTICE as default

All these layers are mostly stable, increase the default log
level to LOGL_NOTICE.

13 years agomisc: Ignore two misc application binaries
Holger Hans Peter Freyther [Sun, 16 Jan 2011 10:30:31 +0000 (11:30 +0100)]
misc: Ignore two misc application binaries

13 years agoccch: Stop following assignment requests and sending LUs
Holger Hans Peter Freyther [Sun, 16 Jan 2011 10:28:23 +0000 (11:28 +0100)]
ccch: Stop following assignment requests and sending LUs

Make the ccch_scan only look at the immediate assignment
but do not follow it. The old behavior was dangerous as
it could interfere with the real receiver of the channel.

13 years agoccch: Fix the cast of the frequency list
Holger Hans Peter Freyther [Sun, 16 Jan 2011 10:22:58 +0000 (11:22 +0100)]
ccch: Fix the cast of the frequency list

13 years agoccch: Remove the app_phone.c, move it to the app_ccch_scan file
Holger Hans Peter Freyther [Sun, 16 Jan 2011 10:20:09 +0000 (11:20 +0100)]
ccch: Remove the app_phone.c, move it to the app_ccch_scan file

13 years agoccch: Add a copyright header to that file.
Holger Hans Peter Freyther [Sun, 16 Jan 2011 10:16:01 +0000 (11:16 +0100)]
ccch: Add a copyright header to that file.

13 years agolayer23: Rename layer23 to ccch_scan as discussed on the ml
Holger Hans Peter Freyther [Sun, 16 Jan 2011 10:14:37 +0000 (11:14 +0100)]
layer23: Rename layer23 to ccch_scan as discussed on the ml

13 years agoupdate .gitignore
Steve Markgraf [Tue, 18 Jan 2011 12:24:28 +0000 (13:24 +0100)]
update .gitignore

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agoMakefile: be more robust against toolchains without syscalls
Michael Grzeschik [Tue, 18 Jan 2011 10:58:26 +0000 (11:58 +0100)]
Makefile: be more robust against toolchains without syscalls

Several toolchains are missing syscalls provided by the libc used. For example,
if the newlib was build with the configure flag "--disable-newlib-supplied-syscalls".
To prevent the configure check for things like "_exit" in osmocom
the CFLAGS+="-nostartfiles -nodefaultlibs" helps a lot.

Signed-off-by: Michael Grzeschik <mgr@xviews.de>
Acked-by: Wolfram Sang <wolfram@the-dreams.de>
Acked-by: Peter Stuge <peter@stuge.se>
13 years agotarget/loader: disable bootrom mapping
Steve Markgraf [Wed, 12 Jan 2011 18:09:16 +0000 (19:09 +0100)]
target/loader: disable bootrom mapping

This is for being able to read the whole flash on devices that use the bootrom,
and also fixes flash detection on the C139/C140/J100i

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agofw/calypso/dsp: Add comment about meaning of bit 11 of ndb->d_tch_mode
Sylvain Munaut [Wed, 12 Jan 2011 14:39:25 +0000 (15:39 +0100)]
fw/calypso/dsp: Add comment about meaning of bit 11 of ndb->d_tch_mode

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agotarget/simtest: remove delay functions
Steve Markgraf [Tue, 11 Jan 2011 19:03:35 +0000 (20:03 +0100)]
target/simtest: remove delay functions

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agofixed typos in comments
Thomas Waldmann [Tue, 11 Jan 2011 18:38:59 +0000 (19:38 +0100)]
fixed typos in comments

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agodocumentation: fix some typos
Steve Markgraf [Sat, 8 Jan 2011 19:58:25 +0000 (20:58 +0100)]
documentation: fix some typos

Thanks to Thomas Waldmann who provided this patch

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agoFix flash chip detection by using a volatile pointer
Andreas Oberritter [Fri, 7 Jan 2011 22:50:04 +0000 (23:50 +0100)]
Fix flash chip detection by using a volatile pointer

* This might be a workaround for a compiler bug
  (gcc 4.5.2, binutils 2.21)

Signed-off-by: Andreas Oberritter <obi@saftware.de>
Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agoMerge commit '4a4f96d1f67441e2902a0806f1d7c4b29a072ca3'
Steve Markgraf [Wed, 5 Jan 2011 22:48:06 +0000 (23:48 +0100)]
Merge commit '4a4f96d1f67441e2902a0806f1d7c4b29a072ca3'

13 years agovty: Fix typo testing return value of bind
Sylvain Munaut [Mon, 3 Jan 2011 21:19:40 +0000 (22:19 +0100)]
vty: Fix typo testing return value of bind

Thanks to playya__ (Dr. Fred) on IRC for pointing this out

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agodisplay/st7558: remove delays, since the bug in the i2c driver is fixed
Steve Markgraf [Sun, 2 Jan 2011 15:54:54 +0000 (16:54 +0100)]
display/st7558: remove delays, since the bug in the i2c driver is fixed

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agoTesting for the idle i2c bus was done backwards, we have
Christian Vogel [Mon, 4 Oct 2010 18:10:42 +0000 (20:10 +0200)]
Testing for the idle i2c bus was done backwards, we have
to wait for the bit to become 0, not 1.

13 years agodisplay/td014: add short delay after initializing uwire
Steve Markgraf [Sun, 2 Jan 2011 00:36:04 +0000 (01:36 +0100)]
display/td014: add short delay after initializing uwire

This should fix the initialization of the C139/C140 display

Signed-off-by: Steve Markgraf <steve@steve-m.de>
13 years agosmscb: Add definition of a warning message
Holger Hans Peter Freyther [Wed, 29 Dec 2010 20:14:20 +0000 (21:14 +0100)]
smscb: Add definition of a warning message

13 years agosmscb: Ignore the very simple test application.
Holger Hans Peter Freyther [Wed, 29 Dec 2010 20:13:44 +0000 (21:13 +0100)]
smscb: Ignore the very simple test application.

13 years agosmscb: Mention 23.041 and fix the size of the update number
Holger Hans Peter Freyther [Wed, 29 Dec 2010 11:00:01 +0000 (12:00 +0100)]
smscb: Mention 23.041 and fix the size of the update number

13 years agosmscb: Add some simple decoding routines for SMSCB.
Holger Hans Peter Freyther [Wed, 29 Dec 2010 10:57:48 +0000 (11:57 +0100)]
smscb: Add some simple decoding routines for SMSCB.

13 years ago[layer23] Enable half rate support in support.c
Andreas.Eversberg [Tue, 28 Dec 2010 12:22:39 +0000 (12:22 +0000)]
[layer23] Enable half rate support in support.c

13 years ago[layer23] Timer 3126 must be stopped even if starting time is given
Andreas.Eversberg [Tue, 28 Dec 2010 12:21:18 +0000 (12:21 +0000)]
[layer23] Timer 3126 must be stopped even if starting time is given

13 years agolayer2: Make it possible to disable or increase the number of RA bursts
Holger Hans Peter Freyther [Mon, 27 Dec 2010 19:03:27 +0000 (20:03 +0100)]
layer2: Make it possible to disable or increase the number of RA bursts

13 years agolayer23: Enable combining the long options
Holger Hans Peter Freyther [Mon, 27 Dec 2010 13:51:13 +0000 (14:51 +0100)]
layer23: Enable combining the long options

Combine the long options from the base and the application. Provide
the long option for the cell log application.

13 years agolayer23: Make the logfile configurable with cell_log
Holger Hans Peter Freyther [Mon, 27 Dec 2010 07:58:57 +0000 (08:58 +0100)]
layer23: Make the logfile configurable with cell_log

Be able to add extra (short) options from the 'applet' to
the main application. Use this to print the help mentioning
app specific options, pass the getopt string and handle the
command line parsing for it.

Change cell_log to keep the logname in the app_cell_log.c
and then access it from the cell_log.c implementation.

13 years agolayer23: Not every application supports every option add enum
Holger Hans Peter Freyther [Sun, 26 Dec 2010 20:29:51 +0000 (21:29 +0100)]
layer23: Not every application supports every option add enum

Allow each application to specify the options it is supporting.

13 years agolayer23: Create an l23_app_info for various information and cb's
Holger Hans Peter Freyther [Sun, 26 Dec 2010 18:15:48 +0000 (19:15 +0100)]
layer23: Create an l23_app_info for various information and cb's

Make it possible that each l23 app can inject the copyright string,
also prepare to have callbacks for the config handling and other
places. This will be useful to add app specific config options.

13 years ago[cell_log] Use GPS signal, even if it becomes valid during scanning
Andreas.Eversberg [Mon, 27 Dec 2010 17:11:28 +0000 (17:11 +0000)]
[cell_log] Use GPS signal, even if it becomes valid during scanning

13 years agofw/layer1: Resync TCH primitive on mode changes
Sylvain Munaut [Mon, 27 Dec 2010 15:07:30 +0000 (16:07 +0100)]
fw/layer1: Resync TCH primitive on mode changes

This is needed for audio to work properly in VEA mode

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agoUse the app_info->name instead of the hostname
Harald Welte [Fri, 24 Dec 2010 14:10:14 +0000 (15:10 +0100)]
Use the app_info->name instead of the hostname

This makes more sense in case you run BCS, SGSN and other components
on the same host.  Having multiple telnet sessions with the same
prompt can otherwise be confusing.

13 years agostdint.h: Next attempt at making this work with various compilers
Holger Hans Peter Freyther [Mon, 20 Dec 2010 20:01:22 +0000 (21:01 +0100)]
stdint.h: Next attempt at making this work with various compilers

Newer GCC with GNU LIBC do not like our minimalistic version of
stdint.h and will have conflicts. Older GCC with older C Libs do
not have a stdint.h yet and the #include_next trick is failing. To
make matters worse NEWLIB does not export its version via the
pre-processor.

We will have to guess once more about the compiler. This code now
assumes that if we have a GCC < 4 that it does not have a stdint.h
and we will not try to include the next stdint.h file.

13 years agorate_ctr: No need to include the inttypes.h
Holger Hans Peter Freyther [Mon, 20 Dec 2010 12:38:22 +0000 (13:38 +0100)]
rate_ctr: No need to include the inttypes.h

There should not be any u_int*_t types in this file, no need
to include this file. It is breaking compilation with the last
x86 build of GNU ARM for GCC 3.4.

13 years agoupdate list of networks with information from ITU
Harald Welte [Sat, 18 Dec 2010 23:25:46 +0000 (00:25 +0100)]
update list of networks with information from ITU

http://www.itu.int/dms_pub/itu-t/opb/sp/T-SP-E.212B-2010-PDF-E.pdf

13 years ago[mtk] Add initial / unfinished MT6139 driver
Harald Welte [Sat, 11 Dec 2010 19:04:31 +0000 (20:04 +0100)]
[mtk] Add initial / unfinished MT6139 driver

13 years agoadd perl scripts to compute MTK RF Transceiver PLL settings
Harald Welte [Sat, 11 Dec 2010 04:23:58 +0000 (05:23 +0100)]
add perl scripts to compute MTK RF Transceiver PLL settings

13 years agoREADME update about transmit as well as development
Harald Welte [Sat, 11 Dec 2010 03:38:59 +0000 (11:38 +0800)]
README update about transmit as well as development

13 years agolayer23: use correct sizeof(sockaddr) for connecting to L1CTL socket
David Wilson [Thu, 9 Dec 2010 03:33:02 +0000 (03:33 +0000)]
layer23: use correct sizeof(sockaddr) for connecting to L1CTL socket

connect's addrlen should be sizeof(local), not the contained path's length.

With the previous code, on OS X connect() will fail with ENOENT.

This permits layer23 to work on OS X using the pl2303 driver,
/dev/tty.usbserial , MacPorts arm-elf-gcc and RANLIB=arm-elf-ranlib

Signed-off-by: Harald Welte <laforge@gnumonks.org>
13 years agoMerge commit 'ab1246e0b5ffa0ac8d7b8e8fbe7c51bc22c8e751'
Sylvain Munaut [Sun, 28 Nov 2010 14:31:20 +0000 (15:31 +0100)]
Merge commit 'ab1246e0b5ffa0ac8d7b8e8fbe7c51bc22c8e751'

13 years agoprotocol: introduce gsm_03_41.h
Alex Badea [Sat, 27 Nov 2010 21:35:08 +0000 (23:35 +0200)]
protocol: introduce gsm_03_41.h

This currently contains definitions for the BTS->MS SMSCB message.

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agogsm_08_58: add struct and constants for RSL_IE_CB_CMD_TYPE
Alex Badea [Sat, 27 Nov 2010 21:34:46 +0000 (23:34 +0200)]
gsm_08_58: add struct and constants for RSL_IE_CB_CMD_TYPE

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agogsm_04_12: fix 04.13 typos
Alex Badea [Sat, 27 Nov 2010 18:00:39 +0000 (20:00 +0200)]
gsm_04_12: fix 04.13 typos

There are two occurrences of "413" in the 04.12 header file.
These are probably typos; correct them to "412".

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agowireshark: fixes for the SMSCB dissector
Alex Badea [Sun, 28 Nov 2010 13:45:17 +0000 (15:45 +0200)]
wireshark: fixes for the SMSCB dissector

Fix dissection of message ID, and cosmetic adjustments
to line up the bitfields.

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agofw/layer1: Fix reset of dedicated channel state
Sylvain Munaut [Sun, 28 Nov 2010 13:40:08 +0000 (14:40 +0100)]
fw/layer1: Fix reset of dedicated channel state

Thanks to Alex Badea <vamposdecampos@gmail.com> for pointing
this out.

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agobuild: Use absolute path in the CFLAGS for libosmocore target build
Sylvain Munaut [Wed, 24 Nov 2010 09:05:49 +0000 (10:05 +0100)]
build: Use absolute path in the CFLAGS for libosmocore target build

There is now subdirectories to build !

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agowireshark: update SMSCB dissector patch
Alex Badea [Sat, 27 Nov 2010 16:31:49 +0000 (18:31 +0200)]
wireshark: update SMSCB dissector patch

Now factored out to its own dissector module, and supporting
message block reassembly.

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agofw/layer1: Fix missing include in prim_pm.c
Sylvain Munaut [Wed, 24 Nov 2010 08:43:11 +0000 (09:43 +0100)]
fw/layer1: Fix missing include in prim_pm.c

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agowireshark: add CBS/SMSCB dissector patch [WIP]
Alex Badea [Tue, 23 Nov 2010 19:03:03 +0000 (21:03 +0200)]
wireshark: add CBS/SMSCB dissector patch [WIP]

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agohost layer23: add a small cbch_sniff application
Alex Badea [Sun, 21 Nov 2010 20:46:11 +0000 (22:46 +0200)]
host layer23: add a small cbch_sniff application

Tune to the ARFCN specified on the commandline (-a).  Then, if
a CBCH Channel Description IE is found in System Information Type 4,
switch to dedicated mode on that particular channel to receive
the CBCH.

13 years agolayer23 sysinfo: fix parsing of CBCH Mobile Allocation
Alex Badea [Sun, 21 Nov 2010 20:45:23 +0000 (22:45 +0200)]
layer23 sysinfo: fix parsing of CBCH Mobile Allocation

When calling gsm48_decode_mobile_alloc(), read the length
byte from payload data[], rather than from si->data[].

13 years agolayer23 sysinfo: store chan_nr when decoding CBCH Channel Description
Alex Badea [Sun, 21 Nov 2010 20:45:22 +0000 (22:45 +0200)]
layer23 sysinfo: store chan_nr when decoding CBCH Channel Description

13 years agoMakefile: introduce 'make nofirmware' target to build host-only tools
Harald Welte [Sun, 21 Nov 2010 17:24:10 +0000 (18:24 +0100)]
Makefile: introduce 'make nofirmware' target to build host-only tools

13 years agorita_pll: Compute values for T-GSM 810
Harald Welte [Sun, 21 Nov 2010 17:10:46 +0000 (18:10 +0100)]
rita_pll: Compute values for T-GSM 810

13 years ago[layer23] VTY uses CAPITAL letters for all optional parameter names
Andreas.Eversberg [Sat, 20 Nov 2010 08:58:44 +0000 (08:58 +0000)]
[layer23] VTY uses CAPITAL letters for all optional parameter names

This makes it easier to understand the VTY help. Every parameter name
is shown in capital letters, every keyword is shown in lower case.

13 years ago[layer23] VTY supports "show this" to show current MS node only
Andreas.Eversberg [Sat, 20 Nov 2010 08:34:31 +0000 (08:34 +0000)]
[layer23] VTY supports "show this" to show current MS node only

13 years ago[layer23] Adding LAC and TMSI (both optional) to test card (rplmn)
Andreas.Eversberg [Sat, 20 Nov 2010 08:32:32 +0000 (08:32 +0000)]
[layer23] Adding LAC and TMSI (both optional) to test card (rplmn)

13 years ago[layer23] Adding LAI to debug output of LOCATION UPDATING REQEST
Andreas.Eversberg [Sat, 20 Nov 2010 08:30:45 +0000 (08:30 +0000)]
[layer23] Adding LAI to debug output of LOCATION UPDATING REQEST

13 years ago[layer23] Bugfix: Exit, if init function fails
Andreas.Eversberg [Sat, 20 Nov 2010 08:19:38 +0000 (08:19 +0000)]
[layer23] Bugfix: Exit, if init function fails

13 years agoutils: Fix typo. It is ascii.
Holger Hans Peter Freyther [Fri, 19 Nov 2010 18:20:46 +0000 (19:20 +0100)]
utils: Fix typo. It is ascii.

13 years agogsmtap.h: Add GSMTAP_TYPE_SIM
Harald Welte [Thu, 18 Nov 2010 22:54:20 +0000 (23:54 +0100)]
gsmtap.h: Add GSMTAP_TYPE_SIM

13 years agotoplevel Makefile: accept arguments for host ./configure calls
Alex Badea [Wed, 17 Nov 2010 21:35:18 +0000 (23:35 +0200)]
toplevel Makefile: accept arguments for host ./configure calls

Append $(HOST_CONFARGS) to ./configure scripts for 'host' applications.

This allows e.g. cross-compiling on an x86 build system for an OpenMoko
gta0x host, using an invocation such as:

$ make HOST_CONFARGS="--host=arm-angstrom-linux-gnueabi"

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agotarget uart: remove REG_OFFS() macro side-effect
Alex Badea [Wed, 17 Nov 2010 21:35:17 +0000 (23:35 +0200)]
target uart: remove REG_OFFS() macro side-effect

Don't assign to the variable given as argument.  This prevents
clobbering the local 'reg' variables in uart_reg_{read,write}(),
which would in turn prevent the latch bits from being restored
correctly.

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agotarget uart: fix preservation of LCR
Alex Badea [Wed, 17 Nov 2010 21:35:16 +0000 (23:35 +0200)]
target uart: fix preservation of LCR

Store old_lcr only when switching to LCR == 0xBF.  We don't want
to clobber old_lcr when switching back, otherwise we can't restore
the previous LCR value.

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agoosmocon: make beacon interval configurable via cmdline
Alex Badea [Wed, 17 Nov 2010 21:35:15 +0000 (23:35 +0200)]
osmocon: make beacon interval configurable via cmdline

Beacons with the default 50 mS interval are too far apart to
be picked up by the OpenMoko gta0x Calypso chip.  Make them
configurable via a -i commandline argument.

As recommended in the OpenMoko wiki[1], an interval of 13 mS works.

[1] http://wiki.openmoko.org/wiki/GSM/Flashing (-od fluid argument)

Signed-off-by: Alex Badea <vamposdecampos@gmail.com>
13 years agofw/l1/mframe_sched: Fix the TCH_A entries to include the MF_F_SACCH flag
Sylvain Munaut [Sun, 14 Nov 2010 09:13:04 +0000 (10:13 +0100)]
fw/l1/mframe_sched: Fix the TCH_A entries to include the MF_F_SACCH flag

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agofw/layer1: Use dbm2rxlev instead of just '+110' so that overflows are handled
Sylvain Munaut [Sun, 14 Nov 2010 15:08:08 +0000 (16:08 +0100)]
fw/layer1: Use dbm2rxlev instead of just '+110' so that overflows are handled

Without this, we would sometime get +143 dBm reported ... which is obviously
false.

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years ago[layer23/mobile] Better overview of "show ms" VTY command
Andreas.Eversberg [Sun, 14 Nov 2010 15:15:38 +0000 (15:15 +0000)]
[layer23/mobile] Better overview of "show ms" VTY command

Combined "sh ms" and "sh states" to a more detailed "sh ms". The
look-and-feel is more router like now.

13 years ago[layer23] Cleanup of mobile application
Andreas.Eversberg [Sun, 14 Nov 2010 11:52:57 +0000 (11:52 +0000)]
[layer23] Cleanup of mobile application

All functions for handling mobile instances and mobile relevant parts are
moved to mobile/app_mobile.c, the mobile/main.c and mobile/mncc.c become a
simple out-of-the-box mobile application. (making calls)

The mobile/main.c can be replaced easily by a different application now.
this application may have it's own call control implementation (layer 4).
Full configurations via VTY is still possible and required in this case.

13 years ago[layer23] Prevent layer23 apps from abort, if socket to layer1 fails
Andreas.Eversberg [Sun, 14 Nov 2010 11:50:14 +0000 (11:50 +0000)]
[layer23] Prevent layer23 apps from abort, if socket to layer1 fails

13 years agoMerge commit '7f6615a37df0dadbe86fdfc980e0a8a15013a80e'
Sylvain Munaut [Sat, 13 Nov 2010 21:48:15 +0000 (22:48 +0100)]
Merge commit '7f6615a37df0dadbe86fdfc980e0a8a15013a80e'

13 years agoMove the generate backtrace call from MSGB_ABORT to osmo_panic handler
Sylvain Munaut [Sat, 13 Nov 2010 21:47:47 +0000 (22:47 +0100)]
Move the generate backtrace call from MSGB_ABORT to osmo_panic handler

This has two benefits:
 - All people calling osmo_panic() will have the backtrace
 - It makes the thing build in 'target' mode in osmocom-bb

And one downside:
 - The osmo_panic handler is now in the backtrace
(I can live with that :)

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years ago[layer23] Added support for multiple MS instances
Andreas.Eversberg [Sat, 13 Nov 2010 18:45:09 +0000 (18:45 +0000)]
[layer23] Added support for multiple MS instances

To create another instance: 'ms <name> create'

To remove an instance: 'no ms <name>'

If no instance exists, 'ms 1' is created automatically on startup.

Each instance can be enabled / disabled by using 'shutdown' or
'no shutdown'. Multiple instances may share the same layer2 socket (same
phone hardware), but in this case only one instance can be enabled at the
same time. This makes it much easier to select different settings without
modifying them.

A 'shutdown' initiates the IMSI detach procedure before shutdown is
completed. A 'shutdown force' will immidiately shutdown.

There is no need to restart the software anymore, if fundamental settings
are changed. In this case, a 'shutdown' followed by a 'no shutdown' will
do the job.

If you already have an old osmocom.cfg, you need to "no shutdown" it.
Everything else behaves as before.

13 years ago[layer23] msgb_free() must in l1ctl.c must called after reading its header
Andreas.Eversberg [Sat, 13 Nov 2010 18:23:57 +0000 (18:23 +0000)]
[layer23] msgb_free() must in l1ctl.c must called after reading its header

13 years agopanic: Fix type warning - osmo_panic_handler_t is already a pointer ...
Sylvain Munaut [Sat, 13 Nov 2010 17:00:25 +0000 (18:00 +0100)]
panic: Fix type warning - osmo_panic_handler_t is already a pointer ...

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agogsm_utils: Define 4 upper bits as "flags" and mask them out in utility functions
Sylvain Munaut [Sat, 13 Nov 2010 16:51:37 +0000 (17:51 +0100)]
gsm_utils: Define 4 upper bits as "flags" and mask them out in utility functions

This way those function don't care about the flags they don't know about

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agoadd definitions for sciphone G2
Harald Welte [Tue, 9 Nov 2010 18:17:48 +0000 (19:17 +0100)]
add definitions for sciphone G2

13 years agoadd some header files about MTK GSM related peripherals
Harald Welte [Tue, 9 Nov 2010 17:19:06 +0000 (18:19 +0100)]
add some header files about MTK GSM related peripherals

13 years agomsgb: fix printing of msgb pointer during MSGB_ABORT()
Harald Welte [Tue, 9 Nov 2010 12:42:26 +0000 (13:42 +0100)]
msgb: fix printing of msgb pointer during MSGB_ABORT()

13 years agomsgb: don't compare unsigned int with int (fix lcr build)
Harald Welte [Tue, 9 Nov 2010 12:41:48 +0000 (13:41 +0100)]
msgb: don't compare unsigned int with int (fix lcr build)

13 years agotarget_dsp/calypso: Makefile rework
Sylvain Munaut [Thu, 30 Sep 2010 21:12:06 +0000 (23:12 +0200)]
target_dsp/calypso: Makefile rework

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
13 years agotarget_dsp/calypso: Small updates bin2cfile.py script
Sylvain Munaut [Wed, 29 Sep 2010 16:45:58 +0000 (18:45 +0200)]
target_dsp/calypso: Small updates bin2cfile.py script

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>