On Sun, 18 Jan 2009, Uri Shkolnik wrote: > > You know, it might help if I actually looked at the files > > I'm > > hacking on, instead of blindly assuming they work like > > `MAKEDEV' > > and create the node in the current directory :-)
Well, I guess that has blown my chances of ever getting a job again ;-) Now the world will see this post and say that something about my claim to having written an operating system in my spare time just doesn't add up... But now that the cat is out of the bag, and you've made the Linux library available to all, I figure it's time to post my hack to the public. Attached as a file is a replacement for the MAKEDEV-like script which can be found in contrib/ in the download from Siano. It tackles the following issues: * eliminates the DOS-type line-endings, which did not agree with my flavour of /bin/sh * attempts to figure out where the Siano library should be looking for its devices -- this is useful for the case where the desired major number 251 is already in use, and one has to specify a module load parameter (I think I've described this as best I could in an earlier message to linux-dvb) * allows the user to specify the major and/or minor numbers given as module parameters * unlinks potentially existing nodes in case of the need to recreate them with different major or minors This is a total hack, and can be rewritten to probably half the number of lines by anyone with a Clue, say, using a for loop and simplifying my logic (note to future employers: look into my eyes, my eyes, not around the eyes, you did not see this, you are suddenly interested in the shiny toys on your desk) So, feel free to improve on this, or to use it in place of that supplied from the Siano ftp site. > > > It was not hardware debugging needed, so it seems. On > > Or... was it? Not only with major 249 on the newer build, > > but now, again with 249 on my notebook, I also see success. > > > > Could it be that the USB device into which I plugged the > > TerraTec Piranha caused the problems? Maybe, because into > > a different USB hub, I have success on the notebook... > > Now, the interesting thing is that a USB2 DVB-S connected > > through this same hub delivered streams that were corrupt > > every few minutes, while the same device connected to a > > different hub has been delivering flawless streams. Now > > I need to check whether the USB1 Piranha can deliver the > > clean streams, or if again, cheap hardware will cause me > > grief... An update to this, in case it would be of interest... After several attempts to tune potential if weak signals, I once again reached the point where attempts to initialise the device would timeout -- exactly what I had seen originally, but this time after half a dozen successful attempts to use it and at least get to the tuning stage. So now I've moved the device to a different USB hub port, and once again it's working fine. How long will this last? I still haven't tried it in the hub where my other DVB device works flawlessly, but it may come to that... > Cool, I'm just CC the ML > I get questions (sometimes the exact same questions) from various of people. > Lets use the ML to sync all... Well, I hope that by making a fool of myself, I can post this info so that others can benefit from it, and you won't be bothered by these same questions. In summary: I've had success with DAB signals with the following: The default_mode= module loading parameter works set to =2. The firmware I've used is that which I downloaded some time back as part of the DVB-T in-kernel siano support, and looks like... -r-xr-xr-x 1 beer besoffen 40096 May 17 2007 /usr/lib/hotplug/firmware/tdmb_stellar_usb.inp There are firmware files of different size on the Siano ftp server, such as -rw-rw-r-- 1 beer staff 38428 2008-12-31 16:26 tdmb_stellar_usb_12mhz_downld.inp -rw-rw-r-- 1 beer staff 38720 2008-12-31 16:26 tdmb_stellar_usb_12mhz_eeprom_a2.brn I have not yet done anything with these to see if these might be a better choice. If, like me, your `cat /proc/devices` is full, and major 251 is assigned to something else, when you go to load the siano smsmdtv.ko kernel module, you will need to specify an alternate major number (minor can also be specified, should you need) smschar_major=249 (tweak as you need), for example. You can give the major number as 0... smschar_major=0 In this case, the first available major number will be automatically assigned. The script which I attach attempts to find this, assuming `procfs' is available and mounted, and does its best to create the proper devices. Should that fail, you can optionally specify the major (and, should you wish, the minor) as options to the script, which should explain itself in the comments if you read it. Here's hoping this is useful... (followup as appropriate; /dev/null is where this should have gone) barry bouwsma
create_char_dev.sh
Description: Hacked version of contrib/create_char_dev.sh
