killhdinitrd 0.9.x Support Thread

[mattdb]

NO, you can extract the kernels from the ISO file. Use a program like WINISO.

It might be worth $5 to make sure you have the right kernels, your choice.

I have the correct kernels stored on my drive. Just wanting to be able to double check from the TiVo without ftp'ng those files to the tivo, especially when you are low on space.

Besides those kernels are available for free elsewhere. Push comes to shove you can pull the drive, kill the kernel and then put it back, but that defeats the purpose of an upgrade with out pulling the drives.

[newbie]

I have the correct kernels stored on my drive. Just wanting to be able to double check from the TiVo without ftp'ng those files to the tivo, especially when you are low on space.

Besides those kernels are available for free elsewhere. Push comes to shove you can pull the drive, kill the kernel and then put it back, but that defeats the purpose of an upgrade with out pulling the drives.

Not everyone can do that. I don't think you can run killhdintrid with the kernel that comes with 6.2X. I'm not even sure you can run killhdintrid with any of the kernels currently being distributed with tivo software, although the older kernels that can be killed work fine. Users have to find either original kernels or kernels that have already been killed.

You previously said you couldn't use the kernel on the $5 disk without pulling the drive. Now you're saying you have the kernels. Different point. edited to say I've never done business with PTVupgrade, just think that's the fastest source of known good killed kernels.

$5 gives you the correct kernel immediately. Most of the sites that had the kernels for free are down. I'm sure you can still find them somewhere, some people think it's worth $5 to get kernels that are know to be correct immediately.

[lenoxb]

$5 gives you the correct kernel immediately. Most of the sites that had the kernels for free are down. I'm sure you can still find them somewhere, some people think it's worth $5 to get kernels that are know to be correct immediately.

$0 and ten minutes' work gives you the virgin kernel if you follow the directions in this post.

[mattdb]

$0 and ten minutes' work gives you the virgin kernel if you follow the directions in this post.

Thx I knew it was on here somewhere.

[ScanMan]

The best way to tell if your kernel is killed may be to examine the px header with a tool like this (requires perl). It will show an initrd length of 0 on a killed kernel.

Jamie, I tried 'splitppceval.pl' on a couple of kernels, first on a stock 8.1a kernel then on a 8.1a that had a nulled initrd. I got the "Bad magic in header" error. Then I tried an old 3.1.1c killhdinitrd kernel and I got the out.info I was looking for. In examing the perl script and the kernel header in a hex editor I see the first bytes are totally different that the script expects (hence the error).

Anyway, does this only work on killhdinitrd kernels or do newer kernels have a changed header that is incompatible with this tool? Thanks in advance.

[Jamie]

Jamie, I tried 'splitppceval.pl' on a couple of kernels, first on a stock 8.1a kernel then on a 8.1a that had a nulled initrd. I got the "Bad magic in header" error. Then I tried an old 3.1.1c killhdinitrd kernel and I got the out.info I was looking for. In examing the perl script and the kernel header in a hex editor I see the first bytes are totally different that the script expects (hence the error).

Anyway, does this only work on killhdinitrd kernels or do newer kernels have a changed header that is incompatible with this tool? Thanks in advance.

Works just fine for me on the 8.1a Series2 stock kernel.

# md5sum vmlinux.px
22abaff5ec98444f4a37fed777357c21  vmlinux.px
# strings vmlinux.px | grep "Linux version"
Linux version 2.4.20 (build@buildmaster59) (gcc version 3.3.4) #1 Thu Oct 26 01:35:03 PDT 2006
# splitppceval.pl  vmlinux.px  foo
Header info:
  Magic:                  0052504f
  Kernel load address:    80001fe0 (offset 0x00000020)
  Total blocks - 1:       00000d50
  Debug:                  00000000
  Kernel entry address:   800043a0 (offset 0x000023e0)
  Kernel length:          0015cc10 (addr 0x8015ebd0)
  Kernel+initrd length:   001a9fc0 (addr 0x801abf80)
  Reserved:               00000000
Initrd length:            0004d3b0
Signature length:         00000111

[ScanMan]

Thanks for the md5sum Jamie; I must have borked the kernel when I extracted it from GZkernel. Once I extracted it the right way, I got the same results from splitppceval.pl. Thanks again for showing me the error of my ways and pointing out a great utility.

[gundan]

I have an updated 9.1 Series 2 SA Tivo(s). Do the killhdinitrd kernels work with the new sw?

[ScanMan]

I have an updated 9.1 Series 2 SA Tivo(s). Do the killhdinitrd kernels work with the new sw?

Yes, absolutely; the '3.1.5' and the '7.2.2-oth.k1' both work fine with the 9.1 software.

[richr69]

Why can't I take the stock 6.2(a) software kernel (2.4.20) and apply the killhdinitrd on it as oppose to taking the 3.1.5 software kernel (2.4.20). I thought I understood at one point, but know I guess I'm confused again.

[ScanMan]

Why can't I take the stock 6.2(a) software kernel (2.4.20) and apply the killhdinitrd on it as oppose to taking the 3.1.5 software kernel (2.4.20). I thought I understood at one point, but know I guess I'm confused again.

That won't work. Only a few specific kernels were vulnerable to the killhdinitrd exploit which defeats the chain of trust. Although ADH's 'chain of trust' post was in the context of why Series2.5 needs a PROM mod, the same concept applies to Series2; it's just that some S2 kernels were found vulnerable to a software hack (i.e., killhdinitrd). See here for the specific killhdinitrd kernels vulnerable.

[gundan]

Scanman,

Thanks for your response. One other question. I am in the process of updating the stock kernel with the hacked one. I already have the 9.1 version of the Tivo SW, further more I am sure tivo will release updates(sooner/later). I am assuming these updates also include kernels and will over write my hacked kernel.

What are the best practises to prevent the updates, or people just let the update happen and then go in and reapply the modified kernel/hacks again?

After going through all the apps/utilities, I am just interested in TivoWebPlus(for its renaming capabilities). I am happy with pyTivo/MRV and would like to continue using that. Can I just ran that one app(no disabling encryption etc) after I have hacked the tivo?

TIA,
G

[ScanMan]

To prevent updates from automatically happening, most people add the 'upgradesoftware=false' parameter to their bootpage. Then, when Tivo sends an upgrade, you will get a 'pending restart' status of last call which is your clue. Then you can perform a manual upgrade without losing hacks.

Yes, you can just use TivoWebPlus without disabling encryption or adding other hacks.

[gundan]

Scanman,

My tivo is a TCD230(40 gig unit). I took the stock 40 gig(9.1 rev s/w) expanded it to a 80 gig unit and I have been fooling around with the kernel patch etc on that unit. That being said, most of the discussions here keep alluding to the root being at /dev/hda4.

But in my upgraded/resized 80 gig unit the root is /dev/hda7, any idea why this is? Further I was following instructions from some where in the web and it kept alluding to putting all the tivotools/tivowebplus package in the /var/hacks directory(located in /dev/hda9). But every time I put it there, it just seems to get wiped away or disappears. Again any thoughts? I finally put the stuff in root(/hacks in /dev/hda7) and it seems to be there.

Further to prevent an upgrade, should my bootpage command be instead? When I boot up in my pc, my tivo hd is /dev/hde(hence the /dev/hde7 in the command below.

bootpage -P "root=/dev/hde7 dsscon=true console=2,115200 upgradesoftware=false" /dev/hde

I have a question on the usb backport drivers, so I will post it in that thread.

Thanks,
G

[ScanMan]

Further to prevent an upgrade, should my bootpage command be instead? When I boot up in my pc, my tivo hd is /dev/hde(hence the /dev/hde7 in the command below.

bootpage -P "root=/dev/hde7 dsscon=true console=2,115200 upgradesoftware=false" /dev/hde

To your first question, tivo uses two sets of partition pairs 3/4 and 6/7 for the boot/root; one set is the active and the other is the alternate. When tivo sends a software upgrade it installs it into the alternate and flips the bootpage to boot from the "other" set. So since your active root is partition 7 you should be installing your killhdinitrd kernel in partition 6.

When your drive is in your Tivo it is always primary master; so your bootpage command should be:

bootpage -P "root=/dev/hda7 dsscon=true console=2,115200 upgradesoftware=false" /dev/hde