Reply
Thread Tools
Posts: 162 | Thanked: 351 times | Joined on Apr 2006 @ Cotswolds, UK
#41
Originally Posted by armtwister View Post
I'm not a USB expert. Graham - do you know how I can talk to it (on the tablet or the PC) to find out how it identifies itself? Then perhaps you could tell me if I can modify something in your stuff to see it?
The best information is available if you can connect it to a PC running linux -- in that case you can use lsusb -v to get a lot of information about it. It probably looks like several devices, the interesting one is, of course, the display device.

If you don't have a linux PC available, the best option is to get syslog working on the tablet (it may be as simple as installing sysklogd and then saying /etc/init.d/sysklogd start, or you may need to edit /etc/syslog.conf first, I can't remember. But certainly either remove sysklogd or use /etc/syslog.conf to turn off the logging once you have finished as it can fill up the internal disk and/or slow the tablet down). Once syslog is working, plug the device in and see what messages get reported in the log.

Unfortunately, even if you find out what it is it may not help. The kernel module is responsible for recognising the device. If it is actually the same chip but with a different ID then it would be possible to rebuild the kernel module to support it. But if it is a different chip then, presumably, a different kernel module would be needed. But see if you can find out what it claims to be and report back here!

Graham
 
armtwister's Avatar
Posts: 16 | Thanked: 11 times | Joined on Mar 2009 @ Cambridge, England
#42
from GeneralAntilles: On the swift aand convenient note, having to use a powered hub is a particularly irritating inconvenience A AA-powered hub would probably be an option, but what about some sort of inline power injector?

I've been experimenting with hub and power options. A possible battery-powered hub is the CyberPower CPH420MP. However, it has mixed reviews; may be out of production; and has AAA batteries, so would last only a couple of hours or so supplying a USB2VGA2 at ~500 mA. The benefit is that it combines power with hub, of course.

An alternative is to couple a tiny hub, like the 4-port Targus ACH63EU (also sold as Synapsis by Maplin) with a decent battery. The USB Battery Pack (GBP 16.50) is compact and will last about 6 hours at 500 mA (max output 700 mA). It comes with stick-on Velcro, so you can mount the hub on the battery (it helps to have some little stick-on feet as well). There's a beefier battery from Bixnet.com (~53 hours @ 500 mA; max o/p 1,500 mA), but it costs USD 79.95.

Stacked together with Velcro, feet and customised wiring connectors, the USBVGA2 + battery + hub looks quite neat:



For use with a projector (which has AC, of course), there is the opportunity to use a 5V power-supply, but I don't want the clumsiness of a separate AC plug. There is a nice in-line 5V supply from Lindy (GBP 29.50 - probably there are cheaper alternatives), which sits nicely on Velcro at the back of my projector (Optima EX530) and I use a Y-cable to supply it and the projector:



Here is the complete set-up, using a micro-A(M) to USB A(F) adapter and a nice long USB A to mini-B cable. The jpegs are on a micro SD card on the dongle plugged into the hub - very handy. I sometimes plug a USB keyboard into the hub as well. The picture on the wall is being projected!

__________________
"I think." Charles Darwin (1809-1882)
 

The Following 5 Users Say Thank You to armtwister For This Useful Post:
armtwister's Avatar
Posts: 16 | Thanked: 11 times | Joined on Mar 2009 @ Cambridge, England
#43
I've installed sysklogd on the tablet, and it seems to be working - when started, adding text to /var/log/syslog .

(BTW, am I right in assuming that /etc/init.d/sysklogd stop is adequate to stop it?)

The log doesn't show anything very informative (to me, anyway ), and I wonder if I'm looking in the right file?

From Graham Cobb: you may need to edit /etc/syslog.conf first
I've uncommented kern and daemon in syslog.conf, but the respective log files don't get written to. Anything else I should be doing?
__________________
"I think." Charles Darwin (1809-1882)
 
qwerty12's Avatar
Posts: 4,274 | Thanked: 5,358 times | Joined on Sep 2007 @ Looking at y'all and sighing
#44
Originally Posted by armtwister View Post
(BTW, am I right in assuming that /etc/init.d/sysklogd stop is adequate to stop it?)
Yes, and "/etc/init.d/klogd stop"
 

The Following User Says Thank You to qwerty12 For This Useful Post:
Posts: 3,841 | Thanked: 1,079 times | Joined on Nov 2006
#45
Originally Posted by armtwister View Post
I've uncommented kern and daemon in syslog.conf, but the respective log files don't get written to. Anything else I should be doing?
You must restart sysklogd (or, alternatively, send it a HUP signal)
__________________
N800/OS2007|N900/Maemo5
-- Metalayer-crawler delenda est.
-- Current state: Fed up with everything MeeGo.
 

The Following User Says Thank You to TA-t3 For This Useful Post:
armtwister's Avatar
Posts: 16 | Thanked: 11 times | Joined on Mar 2009 @ Cambridge, England
#46
OK, I have some info from kern.log . It finds:
  • USB MultiMedia Audio Device (sound gadget?)
  • USB-MAC Controller (Ethernet?)
  • SCISSORS Keyboard (plugged into one of the USB ports).
I cannot see anything looking like another device . What sort of runes should I expect?
__________________
"I think." Charles Darwin (1809-1882)
 
armtwister's Avatar
Posts: 16 | Thanked: 11 times | Joined on Mar 2009 @ Cambridge, England
#47
I had the bright (?) idea of connecting the USB2VGA2 dongle to one of the ports on the USBVGADOCK to see what appears in the logs.

The message device v0711 p0900 is not supported appears, and then sisusb wakes up and finds it.

Looking back, the message device v0711 p1000 appears earlier in the log on a sensible-looking port number.

Putting 2 and 2 together , I wonder if v0711 p0900 is a unique identifier for the USB2VGA2 and sisusb is configured to respond to that? If so, perhaps sisusb could be configured also to respond to v0711 p1000, which might be a unique identifier for the newer product?

I hope this might offer a way forward: if it's complete garbage, then I'm sorry .
__________________
"I think." Charles Darwin (1809-1882)

Last edited by armtwister; 2009-04-17 at 15:48. Reason: corrected 0911 to 0711
 
armtwister's Avatar
Posts: 16 | Thanked: 11 times | Joined on Mar 2009 @ Cambridge, England
#48
I've done some more digging around, and I have some more information. I'm sure it will be obvious to some, but it's all new to me .

The v0000 p0000 ID refers to vendor and product, with v0711 being Magic Control Technology (MCT) and p0900 the USB to VGA adapter used by StarTech for the USB2VGA2. So the ID of v0711 p1000 for the USBVGADOCK indicates another MCT product.



Looking at Thomas Winischhofer's source for the USB2VGA driver, there is a structure in his file sisusb.c (around line 3489) which I believe defines the the IDs it responds to. These are:
  • 0711 0900
  • 182d 021c
  • 182d 0269
The first we know, of course: the other two are SiteCom products which presumably use the same chip.

I imagine that if we add 0711 1000 to this structure, then the driver will respond to the USBVGADOCK. We don't know if it uses the same chip (SiS315), but its spec is the same as the USB2VGA2, so I'd say it's a good bet .

Graham - would you be able to make this alteration to the driver so that we can try it?
__________________
"I think." Charles Darwin (1809-1882)
 
armtwister's Avatar
Posts: 16 | Thanked: 11 times | Joined on Mar 2009 @ Cambridge, England
#49
Success!

I've taken the bull by the horns and hacked my way to a solution .


I looked at the binary of the driver (sisusbvga.ko) and found the table that matched the assumed device ID structure in the source. It looks like the table has the following entries:
  • 0711 0550
  • 0711 0900
  • 0711 0901
  • 0711 0902
  • 182d 021c
  • 182d 0269
I changed the appropriate 0x09 to 0x10 and bingo - it works !

So I conclude that the USBVGADOCK has an ID of v0711 p1000 and behaves like a USB2VGA2.

Of course, poking the driver like this stops the USB2VGA2 from working, but it should now be certain that adding the extra ID to the driver will work. I wonder if the 0901 and 0902 entries are placeholders that you have added, Graham, and that one of these could become 1000?
__________________
"I think." Charles Darwin (1809-1882)

Last edited by armtwister; 2009-04-27 at 22:22. Reason: Changed ambiguous 0xhhhh to 0xhh
 

The Following 2 Users Say Thank You to armtwister For This Useful Post:
Posts: 2 | Thanked: 0 times | Joined on Apr 2009
#50
Hi, someone could explain for a non-expert linux user how can I make this work w/ my N800?

Thanks.
 
Reply

Thread Tools

 
Forum Jump


All times are GMT. The time now is 14:30.