[HR10-250] Upgraded from 3.1.5f straight to 6.3c -- few minor issues

[iMagic]

I had a nicely hacked 3.1.5f HR10-250. I was very happy with it, especially after I got latest TWP and managed to get mfs_ftp to work.

I was about to dive into tuning my transfer speed, but this untimely DST thing was really getting me down, so I decided to upgrade to 6.3c

I used the steps given in Upgrading to 6.3a thread and everything went almost good. I copied over my killinitrd'd 3.1.5.f kernel and moved all my hacks. I copied my rc.sysinit.author and assumed that it would get launched from rc.sysinit.... Didn't even think to check. And of course I forgot to re-enable serial bash. :rolleyes:

Anyhow, I ended up having to pull the drive and to my surprise I discovered that rc.sysinit is quite different from what I had in 3.1.5f. I plugged my rc.sysinit.author after all the stages get executed, which works fine, but (!) for some reason my rc.sysinit.author is launched twice! I dunno why the heck it's doing it. Anybody came across this? Out of curiosity, where do you insert you rc.sysinit.author.

I was reading through a bunch of threads and I came across AW's idea of launching early bash and telnet from test.conf. I like that! I am not quite sure though when and how test.conf gets launched is it fired up by init process at some early stage?

Lastly, my Linksys USB200M (with 0x13b1/0x18) was not being recognized, so I had it running by insmod'ing usbnet and af_packet from my /hack/lib/modules directory. I added the above IDs to usb.map under usbnet. I am not fully familiar with usb.map, but it looks like usbnet depends on some mii module. I've commented my insmod's and rebooted... :o no network... :eek: And of course me i-d-i-o-t disabled serial bash again. Where is my freakin' star bit -- gotta pull the drive again.

Now I remember reading in the logs earlier that the stock usbnet was complaining about some mii ioctl missing or something. Does anybody know what it is? I can keep using my hack/lib usbnet, but it would be nice to be able to use the stock driver.

Any suggestions are appreciated.

DZ

[Jamie]

...

Any suggestions are appreciated.rc.sysinit.author was always launched for you, AFAIK. There is no nead to edit rc.sysinit to add that in. If you do, it will be run twice.

Your best bet with any 6.x,7.x,8.x release is to let the tivo software do all the network module loading for you. As long as your usb.map is up to date, it should load the right modules automatically. Now that patches are out there to enable the network configuration menus in the 6.x tivo UI, this is by far the easiest and best way to get networking going.

The stock tivo network modules have a dependency on the mii module. THe backport modules do not.

[iMagic]

Jamie,
Do you have any idears why it would run twice?

I can offload all the networking onto the tivoapp, which is my goal as long as I can get the stock usbnet to run. It looks like mii is MIA. I have to pull the drive out of my TiVo to see why it's not there. In worst case scenario I can overwrite stock usbnet with the backported one, which ran fine for me. What the hell is mii anyways?

Now, where do I launch my hacks from, if I don't have rc.sysinit.author? I am gonna run telnet and serial bash on an early stage as soon as I figure out how the test.conf works.

What about the rest? TWP, ftpd, tserver, vserver, mfs_ftp, etc? These would need to be attached somewhere to rc.sysinit. I was thinking of exporting a variable once .author runs and then check for its existence if it gets launched more then once, eg:

[ "$RC_SYSINIT_AUTHOR" == "1" ] && exit 0
export RC_SYSINIT_AUTHOR = "1"

Is there a simpler way to do it?

DZ

[Jamie]

Jamie,
Do you have any idears why it would run twice?The rc.sysinit scripts run it already. My understand is that you manually edited rc.sysinit to run it again. Did I misunderstand?
I can offload all the networking onto the tivoapp, which is my goal as long as I can get the stock usbnet to run. It looks like mii is MIA. I have to pull the drive out of my TiVo to see why it's not there. In worst case scenario I can overwrite stock usbnet with the backported one, which ran fine for me. What the hell is mii anyways?mii (http://en.wikipedia.org/wiki/Media_Independent_Interface) is a bus within ethernet devices. Linux has factored out the driver for it and it can be build as its own module, which other drivers have as a dependency.

Now, where do I launch my hacks from, if I don't have rc.sysinit.author? I am gonna run telnet and serial bash on an early stage as soon as I figure out how the test.conf works.

What about the rest? TWP, ftpd, tserver, vserver, mfs_ftp, etc? These would need to be attached somewhere to rc.sysinit. I was thinking of exporting a variable once .author runs and then check for its existence if it gets launched more then once, eg:

[ "$RC_SYSINIT_AUTHOR" == "1" ] && exit 0
export RC_SYSINIT_AUTHOR = "1"

Is there a simpler way to do it?

DZSeems that there is some confusion. If rc.sysinit.author exists, the standard tivo startup scripts will run it, and run it exactly once. If yours is running it twice, it is something peculiar to your hack setup.

[iMagic]

Jamie, I was under impression that if I don't manually add rc.sysinit.author to rc.sysinit, it doesn't get run at all. That's how I bricked my TiVo the first time after I upgraded. Well, I didn't brick it, but none of my hacks loaded, so I had to pull the drive.

It must get launched by the init process or by the rc.sysinit (although I didn't see where). rc.sysinit has undergone quite a change and it's way different than the one that I had on 3.1.5f.

The stock usbnet won't run -- complains about mii missing or something, so I am running backported one (I believe it's from usb2 package that you put together a while ago).

I am having some other issues at the moment. Let me see if I can sort them out or at least define what the problem is.

DZ

[Jamie]

Jamie, I was under impression that if I don't manually add rc.sysinit.author to rc.sysinit, it doesn't get run at all. That's how I bricked my TiVo the first time after I upgraded. Well, I didn't brick it, but none of my hacks loaded, so I had to pull the drive.I don't know where you got that impression, but it is wrong.


It must get launched by the init process or by the rc.sysinit (although I didn't see where). rc.sysinit has undergone quite a change and it's way different than the one that I had on 3.1.5f.It is in one of the Stage... directories. Grep for rc.sysinit.author and you should be able to find it.


The stock usbnet won't run -- complains about mii missing or something, so I am running backported one (I believe it's from usb2 package that you put together a while ago).I don't know what your issue is here. I don't think others have reported problems with the stock drivers on 6.3x. The backport driver should be fine though. I'd grab the latest version, if you have an older one.

[labbie48]

I don't know what your issue is here. I don't think others have reported problems with the stock drivers on 6.3x. The backport driver should be fine though. I'd grab the latest version, if you have an older one.

I have tried the stock and the backport drivers with 6.3c and there seems to be little if any increase in transfer speed using a FA120 and an ASOHOUSB.