GP2X The F200 Telnet Problem

[paeryn]

I don't think that works on mine - have you tested that on a F200??

I'll have another look tonight.

Yup, it was the F200 that gave me that output. Try insmod usbcore first, that should be all you need for insmod rt2570 to work - it did for me.

Simonb:
There is no /proc/config.gz and I can't see anything similar.

The F200 only runs /etc/rc.d/rc.sysinit which mounts /tmp, makes some dirs on it, starts devfsd, exports PATH and LD_LIBRARY and runs gp2xmenu. All runlevels other than 0 are commented out in /etc/inittab.

The full lot of insmod's needed to enable all the usb stuff are (in order, as loaded when enabling USB host from the system menu):CODE

insmod usbcore
insmod mmsp2_usbh
insmod usb-ohci
insmod usb-ohci-mmsp2
insmod scsi_mod
insmod sd_mod
insmod usb-storage
insmod input
insmod hid
insmod keybdev
insmod mousedev
insmod evdev

 

[deluded]

Cheers!!

Will give it another try over the weekend.

 

[simonb]

The F200 only runs /etc/rc.d/rc.sysinit which mounts /tmp, makes some dirs on it, starts devfsd, exports PATH and LD_LIBRARY and runs gp2xmenu. All runlevels other than 0 are commented out in /etc/inittab.

whoa! is there an init process running? spawning an init to just execute a shell script is about as expensive a way to do it as possible, they could just set init=/myscript.sh in the kernel boot params...

I've ordered a wireless card now, so I'll try and get a shell on this thing asap. Out of interest, is there any benefit/interest in me trying to hack around with the boot processes to see what/if can be speeded up?

 

[Orkie]

whoa! is there an init process running? spawning an init to just execute a shell script is about as expensive a way to do it as possible, they could just set init=/myscript.sh in the kernel boot params...

I've ordered a wireless card now, so I'll try and get a shell on this thing asap. Out of interest, is there any benefit/interest in me trying to hack around with the boot processes to see what/if can be speeded up?

This isn't the problem. The init launch time is negligible (I've looked into it before). The main slowness comes from gp2xmenu, which handles pretty much everything, including setting up USB and mounting the user NAND on the F100, where the correct places for these things is in boot scripts so that they can be disabled, adding no time to bootup if needed (gp2xmenu is so huge since it has a load of rubbish in there, it adds 1-2 seconds to the boot time). The GPH firmware is just a complete mess - Open2x actually has 3-4MB of RAM extra free due to a lack of all this bloat. It uses init scripts and manages a noticeably shorter boot time.

If I thought it was worth the effort to get the GPH firmware booting a tiny bit faster, I would've released my modifications a long time ago but the fact is, gp2xmenu is the real problem. However, if you do that you may as well work on GMenu2X and Open2x instead.

 

[simonb]

If I thought it was worth the effort to get the GPH firmware booting a tiny bit faster, I would've released my modifications a long time ago but the fact is, gp2xmenu is the real problem. However, if you do that you may as well work on GMenu2X and Open2x instead.

Probably. In reality I'll get distracted now, and start trying to get NetBSD running on it for the hell of it, at which point I'll brick the device and friends will tell me yet again to leave well alone that I do not fully understand ;-)

Will checkout the open2x / gmenu2x stuff. Chrs for the pointer :-D

 

[deluded]

Hmm - it detects my wireless dongle, but doesn't do anything when I start it, so suspect my dongle isn't 2570 compatible.

 

[simonb]

Hmm - it detects my wireless dongle, but doesn't do anything when I start it, so suspect my dongle isn't 2570 compatible.

Well, my wifi dongle turned up. Managed to insmod the module fine, however its not working[1]. I'm not sure what linux version this module was built against, but it looks very much like we need to recompile against the same kernel version. Which, well, is nigh on impossible until GPH release the new src code...

Mines running 2.4.25 ; anyone tried to recompile against a vanilla version ?

[1] hangs at ifconfig xxx up for some reason; also the device doesn't appear to have a mac address

 

[deluded]

Did the light on the dongle (assuming it has one) turn on?

Mine didn't but I'm using a 2571 chipset and don't know if that's supported with the 2570 driver.

 

[simonb]

Did the light on the dongle (assuming it has one) turn on?

Mine didn't but I'm using a 2571 chipset and don't know if that's supported with the 2570 driver.

The light turns on, the device exists, but it just doesn't work. Its only when the kernel / userland try and configure the device that the thing hangs. Annoyingly, I'm wondering if theres a stack trace happening I just can't see...

I'm getting somewhat irritated now. I really really need the official gph linux source to be able to build this module :-((

 

[simonb]

OK, done some more digging. This driver apparently locks up sometimes (no idea why) on other linux versions and distributions, not just on arm but on x86-64 too. I've downloaded the CVS version of the driver, and will try and get it to compile... I'm at work at the moment so can't test it, but assuming it does compile if someones in a position to test it they'd save me quite some time tonight...

 

[simonb]

OK, *think* i've managed to create a newer version of the driver. I've uploaded it to:

http://kaizo.org/misc/rt2570.o

if anyone can test it out and report back, that'd be cool.
The driver is based on the driver from:
http://rt2x00.serialmonkey.com/wiki/index.php/Downloads

It's the:
rt2570 (USB) Last beta release: v1.1.0-b2

Having /some/ trouble with the CVS version, but will persevere...

Also, if you've got a "normal" USB -> ethernet adaptor, this kernel module may well prove fruitful for you:
http://kaizo.org/misc/pegasus.o

Please let me know results. P.S. this may well eat your cat, crash your gp2x, or do all manner of evil wrong things. Not my fault, sorry... ;-)

 

[deluded]

any news on this - not had chance to try it myself yet but don't think the 2571 will work anyway.

 

[simonb]

any news on this - not had chance to try it myself yet but don't think the 2571 will work anyway.

Hey, sorry meant to get back. My USB ethernet adaptor works fine now, but the rt2750 doesn't work still - seems to hang the box for some reason. I'm waiting on a serial cable now, so I may be in a better position to debug it...

I'll try and compile you a 2571 module today.

 

[Crog]

Hi, I'm fairly new to the Gp2x development and know this is a bit off the current subject.

But I was wondering if anybody had though about creating a batch file/program which could disable the USB device while also make the Gp2x notice that fact somehow. You would also need to re-enable it and in a batch file you could enable - copy_files - disable. This would save the hassle of disconnecting cables and could be ran as a post-build event when compiling.

I know this doesn't replace the telnet issue but it would be a nice option with an 'out of the box' machine

 

[deluded]

You probably can do that - I've got my build script on linux to attempt to copy via the USB if the gp2x is plugged in. The gp2x doesn't always seem to detect that the USB has been disconnected though.

However, serial (or ethernet) is needed to get debug strings out of the program which is what you used to be able to do (apparently!) via the standard USB.

 

[naomaru]

Hi,

I'm a GP2X newbie. I have just got F200 and started coding on it.

By reading this thread, I gather it's not possible to telnet to F200 via USB.

Is it possible to telnet via a serial cable? Unfortunately, my laptop doesn't have serial connection. Is it possible to use a USB to serial converter?

I also see there are several options in terms of serial cables. Which one is the easiest way? I could do hardware hacking like soldering, but would rather avoid it if possible.

Thank you so much for your help,

 

[Orkie]

I would just buy a cradle. That has serial all wired up for you, and the advantage of serial is that you don't need to set up USB networking and telnet - you just plug it in and you get data from the moment you turn on the GP2X. I always use serial, it's much more convenient .

 

[naomaru]

Thanks, Orkie. I wasn't aware of the cradle.

But my laptop doesn't have a serial connector. If I buy a USB-to-serial adapter (e.g. http://sewelldirect.com/usbtoserial.asp), will it let me get the debug output through USB?

Thanks,

I would just buy a cradle. That has serial all wired up for you, and the advantage of serial is that you don't need to set up USB networking and telnet - you just plug it in and you get data from the moment you turn on the GP2X. I always use serial, it's much more convenient .

 

[DJWillis]

Thanks, Orkie. I wasn't aware of the cradle.

But my laptop doesn't have a serial connector. If I buy a USB-to-serial adapter (e.g. http://sewelldirect.com/usbtoserial.asp), will it let me get the debug output through USB?



Simple answer. Yes .

The most common USB<>Serial leads have a Prolific chip and that works well if you want to be 100% sure but any USB>Serial IC should work great.

 

[deluded]

That's what I've done, although I did it DIY rather than using the cradle.