OZ network scripts on 3.3.6 pre1??

[dansawyer]

All,

I am trying to debug a USB resume bug on 3.3.6 pre1. If the Zaurus is rebooted in the cradle it connects correctly over USB to a FC-1 box. If the Zaurus is suspended and resumed the USB connection is lost.

Below is the USB trace. It appears as though the Zaurus usb address is locked on resume. Is there a \'resume\' script? Or does suspend simply \'lock\' the Zaurus in suspended animation.

 hub.c: new USB device 00:07.2-1, assigned address 8Jan  4 08:11:35 600e kernel: usb-uhci.c: interrupt, status 2, frame# 1283
Jan  4 08:11:35 600e kernel: usb.c: USB device not accepting new address=8 (error=-110)
Jan  4 08:11:35 600e kernel: hub.c: new USB device 00:07.2-1, assigned address 9Jan  4 08:11:35 600e kernel: usb.c: USB device not accepting new address=9 (error=-110)

Dan

[VeeDubb]

I am having a very imilar problem.  The only way I can get a connection is if either the Z is on and in the cradle when I boot my desktop or if I reboot the z whilein the cradle.  other than that, my desktop can \"see\" the z in ifconfig, but I can\'t ping either direcion and qtopia desktop can\'t connect.

I\'m running OZ 3.3.6-pre1 on an sl-5500 connected via USB to a mandrake 9.2 Desktop.

I\'d be happy to share any other usefull info if anyone has a clue.

[Mickeyl]

install the hotplug from oz 3.2

[dansawyer]

That worked. I removed the hotplug package from 3.3.6 pre1 and installed the one from 3.2.

Thanks

[VeeDubb]

How does that work at all???

Aren\'t they compiled with different GCc versions?  I can\'t get the compat libs working and without them, how on earth does this work?  Or is it just a script and not an actal program?

[VeeDubb]

Well, one way or another, it works great.  If anyone else tries this release and has the same prblem, and you will have this problem most likely, then just uninstall the existing hotpluy and install the one from the old openzaurus 3.2 feed.  I tried everything I could to break sync and it works like a charm every time as long as the Z is on when you stick it in the cradle, and even putting it in the cradle while off wo\'nt break the usb connect, just remove and re-insert and you\'re golden.

Mickeyl, you\'re a God.  Thanks so much.  Now if Ican get the compat libs going I\'ll actualy be able to enjoy this rom!

[Mickeyl]

You\'re right. Hotplug is just a bunch of scripts.

[zenyatta]

I tried to diff the 3.2 <> 3.3.6-pre1 hotplug packages to see for my own curiousity what might be causing the trouble (I thought: well, it\'s just a bunch of scripts). Turns out the changes are quite major and I gave up eventually.

I did, however, notice a little mistake in the 3.3.6-pre1 version of /etc/init.d/hotplug. The usage instructions list \"force_reload\" as one of the possible arguments, yet the case statement expects \"force-reload\" instead. So if someone calls \"/etc/init.d/hotplug force_reload\" they will get an error. No such call is made in any of the files in the hotplug package but somewhere else, who knows?

I did not try changing the script to accept \"force_reload\" as I had more important things to do, and I did\'t know who to announce this error to so I wrote here. So there.

z.

[treke]

How does that work at all???

Aren\'t they compiled with different GCc versions?  I can\'t get the compat libs working and without them, how on earth does this work?  Or is it just a script and not an actal program?

In hotplug\'s case, it\'s just a bunch of shell scripts, but software between gcc versions can be mixed for the most part. There are specific cases when they can\'t involving C++ applications.