Kernel 2.6.17 Files
Kernel 2.6.17 Files
Hello everyone,
is there a chance to see any 2.6.17 patched for the XBOX anymore ? I've looked on the SourceForge site for many weeks now, but nothing happens.
Is it planned to get the 2.6.17 XBOX Patched against the Kernel up in anytime ?
Wishes
Florian Strankowski
http://xbox.coresec.de
is there a chance to see any 2.6.17 patched for the XBOX anymore ? I've looked on the SourceForge site for many weeks now, but nothing happens.
Is it planned to get the 2.6.17 XBOX Patched against the Kernel up in anytime ?
Wishes
Florian Strankowski
http://xbox.coresec.de
I would reccomend searching the forum as this has been asked before. The FatX driver is bugged in 2.6 (maybe even 2.4) so it is not recommended to run 2.6 on non native xboxes.
files might be found (site was down when i posted this)
http://cvs.xbox-linux.org/cgi-bin/viewc ... box-linux/
im sure you can run natively just fine with 2.6, but rootfs loopback is basically out of the question
files might be found (site was down when i posted this)
http://cvs.xbox-linux.org/cgi-bin/viewc ... box-linux/
im sure you can run natively just fine with 2.6, but rootfs loopback is basically out of the question
awesome, huge thanks.
by the way, i am running native xbox linux on my box, so i really need it.
as many ppl. said "you dont need the 2.6 kernel at all", i can say that i have many users on my xbox, because its for development usage in my university and i dont want to run a vulnerable kernel (see latest kernel vulnerables at securityfocus or milw0rm).
many thanks and a good evening
florian
by the way, i am running native xbox linux on my box, so i really need it.
as many ppl. said "you dont need the 2.6 kernel at all", i can say that i have many users on my xbox, because its for development usage in my university and i dont want to run a vulnerable kernel (see latest kernel vulnerables at securityfocus or milw0rm).
many thanks and a good evening
florian
Hey Shallax, i just tested out the new Kernelpatch and it does not work.
Afterwards restarting the box makes it hang and the first screen where it says "starting linux". Even nothing is logged in dmesg or messages log.
Any suggestions ?
Florian
my linuxboot.cfg
Code: Select all
extract kernel.tar.bz
[...]
patch -p1 < linux-2.6.17.7-xbox.patch
cp kernel.config .config
make oldconfig
make clean
make
make modules
make modules_install
make install
[editing linuxboot.cfg etc pp...]
Afterwards restarting the box makes it hang and the first screen where it says "starting linux". Even nothing is logged in dmesg or messages log.
Any suggestions ?
Florian
my linuxboot.cfg
Code: Select all
devnode:/boot# cat linuxboot.cfg
title v2_6_16
kernel /boot/vmlinuz
append root=/dev/hda2 devfs=mount kbd-reset xbox=hdd
xboxfb y
title v2_6_17
kernel /boot/vmlinuz.new
append root=/dev/hda2 devfs=mount kbd-reset xbox=hdd
xboxfb y
default v2_6_16
Hi,
I have been playing with the 2.6 series kernel on the xbox and I have encountered the same problem...
I have spent 2 days now crawilng around in the kernel compressed/head.S and related sources trying to track down the boot issue (lots of printk and strput messages)... but I dont even get the "decompile" message...
I have dug through the cromwell source trying to see if its a problem there, but I can not spot it if it exists.
If someone could attach a debugger to the kernel(cromwell?) during a 2.6.17 boot to see where it is halting it would be much appreciated (I dont have a serial output on my xbox).
I have been playing with the 2.6 series kernel on the xbox and I have encountered the same problem...
I have spent 2 days now crawilng around in the kernel compressed/head.S and related sources trying to track down the boot issue (lots of printk and strput messages)... but I dont even get the "decompile" message...
I have dug through the cromwell source trying to see if its a problem there, but I can not spot it if it exists.
If someone could attach a debugger to the kernel(cromwell?) during a 2.6.17 boot to see where it is halting it would be much appreciated (I dont have a serial output on my xbox).
I cant even get 2.6.16 to boot on my xbox... but 2.4 series work fine... is there somthing obvious I am missing?
If it helps any I can paste my .config here...
I have built countless kernel's in the past and I have never had this much trouble before. I have a development background, so fixing the problem if I can find it shouldn't be too hard.
Any assistance you can offer would be much appreciated
If it helps any I can paste my .config here...
I have built countless kernel's in the past and I have never had this much trouble before. I have a development background, so fixing the problem if I can find it shouldn't be too hard.
Any assistance you can offer would be much appreciated
I used that as a reference, even tried one verbatim... still no go.
After that I enabled the kernel hacking debug stuff hoping to get atleast a bit of output before the crash... but its not even decompressing the image!
The kernel config should not be causing a problem at this stage... the kernel should still decompress.
There is a documented case of the "misc.c" crashing on the xbox when compiled with "-O2", the patch makes it compile with "-O0", i have double, even triple checked that it is getting compiled properly.
Could it be a problem with the version of gcc i am using? (4.1)
I have tried to compile with 3.4 and 3.3, but it doesnt seem to make any difference.
I also moved from cromwell to gentoox (nice job on keeping it updated, might be an idea to post to the xbox-linux wiki that it exists, it took me a good 6-8 hours to find it).
Also, I am building the kernel on my main PC, not on the xbox, I dont see how this could cause a problem unless gcc is trying to optomise for an amd processor, but since its set for "Microsoft XBOX" in the kernel config, i wouldn't think this was the problem (i could be mistaken).
I have also used diff on the "arch/i386/boot/compressed/head.S" and "arch/i386/boot/compressed/misc.c" files between 2.6.8, 2.6.16 and 2.6.17. The differences are only minor, and when I revert to a previous version it does not help, so I think the 2.6.17 kernel not working properly for the user in this thread is how he compiled it.
Has there been any 3rd party confirmation that the 2.6.17 kernel doesnt work other then from "fs"?
After that I enabled the kernel hacking debug stuff hoping to get atleast a bit of output before the crash... but its not even decompressing the image!
The kernel config should not be causing a problem at this stage... the kernel should still decompress.
There is a documented case of the "misc.c" crashing on the xbox when compiled with "-O2", the patch makes it compile with "-O0", i have double, even triple checked that it is getting compiled properly.
Could it be a problem with the version of gcc i am using? (4.1)
I have tried to compile with 3.4 and 3.3, but it doesnt seem to make any difference.
I also moved from cromwell to gentoox (nice job on keeping it updated, might be an idea to post to the xbox-linux wiki that it exists, it took me a good 6-8 hours to find it).
Also, I am building the kernel on my main PC, not on the xbox, I dont see how this could cause a problem unless gcc is trying to optomise for an amd processor, but since its set for "Microsoft XBOX" in the kernel config, i wouldn't think this was the problem (i could be mistaken).
I have also used diff on the "arch/i386/boot/compressed/head.S" and "arch/i386/boot/compressed/misc.c" files between 2.6.8, 2.6.16 and 2.6.17. The differences are only minor, and when I revert to a previous version it does not help, so I think the 2.6.17 kernel not working properly for the user in this thread is how he compiled it.
Has there been any 3rd party confirmation that the 2.6.17 kernel doesnt work other then from "fs"?
By the way, this hasnt been time wasted
I have learnt how cromwell/gentoox bootstraps the system, exploits the overflow bug, loads the kernel and jumps to the start_32() method.
Also, then how the kernel proceeds to decompress itself, then jump to the 2nd decompressed start_32() method.
All very interesting... and I think I have learnt enough to fix this 2.6.17 bug should I be able to even get a 2.6.16 kernel booting.
Also, about the FatX support... I am happy to work on it and try and get it working in 2.6... but no promises. I am running it in native mode (loopback is so sloooow), so its not a high priority for me.
Oh.. and in GentooX... what happend to etherboot?? I use it on one of my xbox's since it has no HDD (got it off ebay for $30AUD with a faulty IDE controller), I cant upgrade it to gentoox .
I have learnt how cromwell/gentoox bootstraps the system, exploits the overflow bug, loads the kernel and jumps to the start_32() method.
Also, then how the kernel proceeds to decompress itself, then jump to the 2nd decompressed start_32() method.
All very interesting... and I think I have learnt enough to fix this 2.6.17 bug should I be able to even get a 2.6.16 kernel booting.
Also, about the FatX support... I am happy to work on it and try and get it working in 2.6... but no promises. I am running it in native mode (loopback is so sloooow), so its not a high priority for me.
Oh.. and in GentooX... what happend to etherboot?? I use it on one of my xbox's since it has no HDD (got it off ebay for $30AUD with a faulty IDE controller), I cant upgrade it to gentoox .
[Edit]
Rereading my post, if the ide is bad my solution would not work
Rereading my post, if the ide is bad my solution would not work
Last edited by orochi on Thu Sep 14, 2006 9:13 am, edited 1 time in total.