root@testpi:~# lsusb | grep Conexant
Bus 001 Device 004: ID 0572:6831 Conexant Systems (Rockwell), Inc.
Though none of the LEDs on the unit itself are on - the driver also loads a firmware image, so will probably find the power LED doesn't do anything until that point
root@testpi:/home/pi# cd /usr/src
root@testpi:/usr/src# wget http://www.dvbsky.net/download/linux/media_build-bst-13-140619.tar.gz
root@testpi:/usr/src# tar xvzf media_build-bst-13-140619.tar.gz
root@testpi:/usr/src# cd media_build-bst-13
root@testpi:/usr/src/media_build-bst-13# wget http://www.dvbsky.net/download/linux/sit2-RPi.zip
root@testpi:/usr/src/media_build-bst-13# unzip sit2-RPi.zip
root@testpi:/usr/src/media_build-bst-13# tar xvzf sit2.o.RPi.tar.gz
root@testpi:/usr/src/media_build-bst-13# tar xvzf sit2.o.RPi.dvbc.tar.gz
root@testpi:/usr/src/media_build-bst-13# ./v4l/build_x86.sh # Seems to copy a file we need for make, without it make fails
root@testpi:/usr/src/media_build-bst-13# cp sit2.o.RPi v4l/
root@testpi:/usr/src/media_build-bst-13# cp sit2.o.RPi v4l/sit2.o
root@testpi:/usr/src/media_build-bst-13# cp sit2.o.RPi.dvbc v4l/
root@testpi:/usr/src/media_build-bst-13# make
The make takes quite a while to run, so I've left it to it for the time being.
WARNING: could not find /usr/src/media_build-bst-13/v4l/.sit2.o.cmd for /usr/src/media_build-bst-13/v4l/sit2.o
WARNING: "dma_buf_export" [/usr/src/media_build-bst-13/v4l/videobuf2-dma-contig.ko] undefined!
The first relates to our module, which isn't great. Will let the compile finish and see if it actually causes any issues though.
It would seem that the V4L modules have compiled incorrectly
root@testpi:/home/pi# insmod /lib/modules/3.12.28+/kernel/drivers/media/dvb-frontends/sit2fe.ko
Error: could not insert module /lib/modules/3.12.28+/kernel/drivers/media/dvb-frontends/sit2fe.ko: Invalid module format
Wondered if it was the use of the x86 specific script earlier, but turns out it doesn't do a huge amount.
Nov 6 16:11:58 testpi kernel: [ 247.383094] sit2fe: disagrees about version of symbol module_layout
Nov 6 16:17:06 testpi kernel: [ 555.058068] sit2fe: disagrees about version of symbol module_layout
Nov 6 16:28:48 testpi kernel: [ 1257.375108] sit2fe: disagrees about version of symbol module_layout
Suggests this compile is going to end in the same result, so have killed it off for the time being
make[3]: *** No rule to make target '/usr/src/media_build-bst/v4l/sit2.o', needed by '/usr/src/media_build-bst/v4l/msp3400.o'. Stop.
Makefile:1230: recipe for target '_module_/usr/src/media_build-bst/v4l' failed
make[2]: *** [_module_/usr/src/media_build-bst/v4l] Error 2
make[2]: Leaving directory '/usr/src/linux-headers-3.12.28+'
Was fixed previously by running the x86 build script but can't quite see how that could have fixed it as it copies a single module (which we wind up overwriting anyway),
Looking at it though, our make clean blew away the RPi module we copied in earlier
root@testpi:/usr/src/media_build-bst# cp sit2.o.RPi v4l/sit2.o
root@testpi:/usr/src/media_build-bst# make
WARNING: could not find /usr/src/media_build-bst/v4l/.sit2.o.cmd for /usr/src/media_build-bst/v4l/sit2.o
WARNING: "dma_buf_export" [/usr/src/media_build-bst/v4l/videobuf2-dma-contig.ko] undefined!
Will see at the end if it makes any difference (assuming we've fixed the initial issue).
Starting to wish I'd cross-compiled rather than getting the Pi to do it, but it's a good chunk of the way through now.
Activity
2014-11-06 12:09:42
So will need to build the V4L package
2014-11-06 12:31:32
2014-11-06 14:09:02
Get Kernel headers etc
Install Driver
The make takes quite a while to run, so I've left it to it for the time being.
2014-11-06 15:24:22
The first relates to our module, which isn't great. Will let the compile finish and see if it actually causes any issues though.
2014-11-06 16:16:59
Once the box comes back up, there should (hopefully) now be a /dev/dvb
Nuts.....
2014-11-06 16:25:38
Wondered if it was the use of the x86 specific script earlier, but turns out it doesn't do a huge amount.
2014-11-06 16:34:54
Now I need to find something else to do while that runs :)
2014-11-06 17:20:47
Syslog shows
Suggests this compile is going to end in the same result, so have killed it off for the time being
Got the GitHub link by looking for the relevant version in Module.symvers' history (https://github.com/raspberrypi/firmware/commits/master/extra/Module.symvers)
Have done a make clean and just started make again
2014-11-06 17:24:22
2014-11-06 18:20:46
Was fixed previously by running the x86 build script but can't quite see how that could have fixed it as it copies a single module (which we wind up overwriting anyway),
Looking at it though, our make clean blew away the RPi module we copied in earlier
2014-11-06 19:52:05
Will see at the end if it makes any difference (assuming we've fixed the initial issue).
Starting to wish I'd cross-compiled rather than getting the Pi to do it, but it's a good chunk of the way through now.
2014-11-06 19:52:28
2014-11-06 21:25:33
We're good to go :)
2014-11-06 21:25:40
2014-11-06 21:25:40
2014-11-06 21:25:43