[linux-dvb] More on the cx24123
yeasah at schwide.com
Mon Mar 20 16:08:30 CET 2006
>>I've been messing around more with the KWorld DVB-S 100 (cx24123
>>frontend) card I have here trying to get something tuned, and finally
>>had some succcess this weekend.
>I have been trying to get this card to work, so far without success. I had posted a question to the
>forum on March 14 regarding the /dev/dvb entries not being set up.
>I would be especially interested to know what version of the kernel you are running, and whether you
>have applied any patches. Also, I would like to know how you have defined the card in .config.
>I have held off buying and installing a dish and lnb until I am sure the card is working. Is it
>necessary to have the cable from the lnb connected in order to set up /dev/dvb? I would think it
>would not be.
>Thanks and good luck with your card.
I'm running Fedora Core 4 with a 2.6.15 kernel, with a latest set of
v4l-dvb modules installed on top.
The device entries should (assuming you have a functioning udev) be
created automatically as soon as the module is loaded -- this is not
dependent upon having anything attached to the card, as you suspect.
There's a number of reasons why they might not be:
1) You haven't actually loaded the cx88-dvb module or prerequisites. Try
"modprobe cx88-dvb" if it isn't in the module list.
2) The kernel/drivers you have aren't recent enough to have support for
the card you have, or simply doesn't have the driver compiled.
3) Your system doesn't have a functioning udev, so the device entries
won't be automatically created. You can still make them yourself in the
fashion of the olde worlde.
4) Hardware issues.
Defintely make sure you're doing the modprobe, and then see if anything
dvb or cx related appears in dmesg.
Note that if you are trying to install modules from the latest v4l-dvb
tree on top of your existing kernel (which is what I do), there's some
trickery required to get all of the modules installed (mentioned before
on this list) -- basically the problem there is that some conditional
tests are done in v4l/Make.config to test which kernel version is in
use, but the main makefile includes that file before it does anything to
set the variables that Make.config is testing.
In particular, this block in Make.config (KERNELRELEASE, SUBLEVEL and
PATCHLEVEL are all unset in this context):
CONFIG_DVB_CORE := $(shell test $(SUBLEVEL) -ge 10 -a $(PATCHLEVEL) -ge
6 && e\
CONFIG_VIDEO_BUF_DVB := $(shell test $(SUBLEVEL) -ge 12 -a
$(PATCHLEVEL) -ge 6\
&& echo m)
CONFIG_VIDEO_DEV := $(shell test $(PATCHLEVEL) -ge 6 && echo m)
CONFIG_EM28XX := $(shell test $(PATCHLEVEL) -ge 6 && echo m)
and not having those variables sets causes undesirable effects, such as
certain modules not being installed. My quick solution to getting a
functioning build was to bypass the kernel tests with hardcoded values
for my kernel version, by adding a block like this right after the tests
(these values are correct for any kernel post-2.6.12):
CONFIG_DVB_CORE := m
CONFIG_VIDEO_BUF_DVB := m
CONFIG_VIDEO_DEV := m
CONFIG_EM28XX := m
I think somebody may have said recently that this issue was resolved
somehow, but I haven't noticed if it has, since I still have the bypass
Good luck, and don't give up! Though I kind of wish I had been smart
like you and not bought the dish/etc. at the same time as the card -- it
took a few weekends to figure out all my issues (not all of which were
driver-related of course...), and I was starting to wonder if I'd ever
get any singnal off of those things :-)
More information about the linux-dvb