Powering up, Almost there...reboot loop.

[spaceranger]

After having had some success hacking my original 40gb drive for practice this morning, I decided to hack my 250gb upgraded drive.
This drive had been InstantCaked, a 250 gb drive that had been upgraded to 7.1 (from 5.3) before prom mod. It has my entire
James Bond movie collection (19 of 'em)on it. I think.

I used bootpage and found 4 to be boot root.
I dd'ed the kernel(5.4.1) into hda3 and hacks into hda4 /hack. Also dd'ed the kernel(5.4.1) into hda6 for good measure.

BIG mistake. It went into a reboot loop.
Hacks were still in place. I had modified ADH's rc.sysinit file
(posted in the Series2.5 dev thread)as my rc.sysinit.author.
I deleted the last line and rebooted in tivo. Same loop. Deleted more lines. SOS. Deleted rc.sysinit.author entirely. SOS. I guess it wasn't the .author file.

My guess is the kernel is not compatible with the already upgraded 7.1 software. If it had been hacked before the sw
upgrade, it would have assimilated ok, I think.

An uneducated guess... at best.

Anyone got a better one? Solutions? Diagnosis?
Anyone else hacked a Series2.5 after the 7.1 upgrade?

Flat spot developing on forehead from banging head against screen. After this, I am now an "Experienced user". HA!
You could call it that.

[PlainBill]

I will only point out the obvious - without the console output (possible only with a serial cable), troubleshooting this is very difficult.

PlainBill

[spaceranger]

I will only point out the obvious - without the console output (possible only with a serial cable), troubleshooting this is very difficult.

PlainBill

Yes, I built a cable from an old serial mouse. Tip to 2, ring to
3, and ground to 5, as prescribed here and elsewhere. Checked it
with ohmmeter, no shorts, connections ok. The only output I ever got was when I put my old (apparently toasted) prom in the socket and tried to boot. Showed garbage. I have gotten no output from serial port since.

I see that it would all be guesswork without serial output.

I'll put the bash command back into rc.sysinit.author if you think it'll help.
I thought it wasn't called until after the almost
there screen.
I'm kind of scared to edit the rc.sysinit file...
Anyway, I did have it in the .author file, but
never got any output during boot---reboot process.


Does this .author file look ok?

export PATH=/bin:/tvbin:/sbin:/var/hack:/hack
export PATH=/MFS_DEVICE=/dev/hda10
export TIVO_ROOT=
insmod /lib/modules/bcm7315tty.o
stty 115200 sane < /dev/ttyS1
bash < /dev/ttyS1 >& /dev/ttyS1
tnlited 23 /bin/bash -login &
/hack/tivoftpd.mips

[hayreass]

I would imagine that the procedure is the same on a S2.5 as on an S2.
(I have a very wild imagination though, so do some research 1st)

On an S2 you need to add

dsscon=true console=2,115200

to the bootpage parameters, and run your terminal program at the same speed to see the console output

[PlainBill]

Yes, I built a cable from an old serial mouse. Tip to 2, ring to
3, and ground to 5, as prescribed here and elsewhere. Checked it
with ohmmeter, no shorts, connections ok. The only output I ever got was when I put my old (apparently toasted) prom in the socket and tried to boot. Showed garbage. I have gotten no output from serial port since.

I see that it would all be guesswork without serial output.

I'll put the bash command back into rc.sysinit.author if you think it'll help.
I thought it wasn't called until after the almost
there screen.
I'm kind of scared to edit the rc.sysinit file...
Anyway, I did have it in the .author file, but
never got any output during boot---reboot process.


Does this .author file look ok?

export PATH=/bin:/tvbin:/sbin:/var/hack:/hack
export PATH=/MFS_DEVICE=/dev/hda10
export TIVO_ROOT=
insmod /lib/modules/bcm7315tty.o
stty 115200 sane < /dev/ttyS1
bash < /dev/ttyS1 >& /dev/ttyS1
tnlited 23 /bin/bash -login &
/hack/tivoftpd.mips

No, it doesn't BUT I'm not familiar with Series 2.5 hardware; particularly which comm port is used for the serial port. If I had to guess (a really bad idea), I'd use S2.

PlainBill

[spaceranger]

No, it doesn't BUT I'm not familiar with Series 2.5 hardware; particularly which comm port is used for the serial port. If I had to guess (a really bad idea), I'd use S2.

PlainBill

PlainBill

Because it works for him,
I used ADH's rc.sysinit from Series2.5 dev thread, adding
tnlited 23/bin/bash -login &.
rc.sysinit.author contains /hack/tivoftpd

Still no bash or any kind of serial output.
Tried pressing enter at boot as well.
Still stuck in reboot loop.

[alldeadhomiez]

Series2.5 units use ttyS1 for the serial port.

(How can you tell? /dev/ttyDSS is a symlink to /dev/ttyS1 on Series2.5, /dev/ttyS2 on Series2.0, /dev/ttyS3 on Series1, etc.)

[hayreass]

Series2.5 units use ttyS1 for the serial port.

(How can you tell? /dev/ttyDSS is a symlink to /dev/ttyS1 on Series2.5, /dev/ttyS2 on Series2.0, /dev/ttyS3 on Series1, etc.)

So then would it be fair to say that setting

dsscon=true console=1,115200

in the bootpage parameter would turn on the serial console?

[spaceranger]

So then would it be fair to say that setting

dsscon=true console=1,115200

in the bootpage parameter would turn on the serial console?

Yes, I think that would be fair. IT WORKED! (used -C to write) Progress!
linux:~ # bootpage -p /dev/hda
root=/dev/hda4 dsscon=true console=1,115200

Thanks, all of you who replied! I've stopped beating my head on the screen.

I was looking at rc.sysinit.orig (renamed from rc.sysinit, and
called by ADH's series2.5 rc.sysinit file I'm using now) and it
has no call to rc.sysinit.author, so tivoftpd won't be called.

Is this something new in Series2.5?

Of course, it's still in reboot loop. HELP!

Attached see bootlog:

[spaceranger]

A note to anyone considering hacking a TCD540040:
Prom mod is a necessity. The 5.4.1 kernel works.
But an upgrade will wipe that kernel.
Use what ADH posted (rc.sysinit and kernel)Set console in bootpage as above and ip fix below.

ADH has generously provided what you need.
ADH is undoubtedly a genius and one should pay very close attention to everything he says.
Read ALL related posts every day.
Do it his way and it WILL work.

[alldeadhomiez]

Are you still having trouble booting 7.1 on a TCD540040?

Judging from the bootlog 2 posts up, you probably have not applied the iptables fix to that system. 7.1 will assert if the netfilter stuff fails.

[spaceranger]

Are you still having trouble booting 7.1 on a TCD540040?

7.1 will assert if the netfilter stuff fails.

Yes, I was until I until I applied that fix. Now it works great!
I just need to configure my wireless now. And learn to use bash
and read some more and learn some more! Serendipitously, I was
reading that thread that the ip fix was in when you posted.

Thank you very much for all the files and help!

Your last comment went right over my head, but I'm learning.

Gone to edit post above .

[alldeadhomiez]

Your last comment went right over my head, but I'm learning.

assert = big ugly crash (like what you saw)
netfilter = iptables / firewall stuff
fails = iptables returns nonzero because it can't find the netfilter stuff in the kernel

Basically, they're checking return values from iptables (sometimes) in 7.1, so if something goes wrong when they're setting up the firewall, startup halts. Which is actually a good coding practice, but it's bad for us.

[NutKase]

The 5.4.1 kernel works.
But an upgrade will wipe that kernel.

Adding...

upgradesoftware=false

to your bootpage parameters should prevent the upgrade from installing and wiping anything just as it does in every other software version. Alternatively you can export it just before the check in /etc/rc.sysinit unless they've changed it in this version (can't check to confirm right now as I have the drive pulled.)

Keep in mind that I don't have a 2.5 but I do have 7.1a.


NutKase

[Elstevo]

After working fine for a week (first time hacking a TiVo, used the 4.x+RID guide and some image off eMule that im told isnt good because it was pre kill-whatever'd) my HDVR2 is now rebooting by itself too, never loads guide data. Does it with or without everything plugged into it, let it sit unplugged a day, still does it. Plopped in original drive, works fine.

No cable, just going to do it right this time, using Tivo-S2-4.0.1b-02-2-xxx.mfs freshly downloaded from eMule.

It worked fine for that week, but then after it was done recording the superbowl, it was responding slowly, and the superbowl was still showing up as recording (no red light on) in the Now Playing List - and that the duration was 4:30. Attempts to delete it failed, so i reset, and that was all she wrote.

All of this brought to you by the letter T, the number 4, and a bad prehacked image right?

I think im actually going to miss having transparent menus, as the audio and everything continued, it was more like a feature than a bug. Go go gadget learning experience though.