Trying to get Fedora running on an old Mybook World, but can not seem to get it to run at all..
From what I understand this is the same that Peter Lemenkov successfully
got Fedora 11 running on some time ago, but I only get "illegal instruction" when trying to run anything..
I am using the original kernel shipped by WDC.
Regards Henrik
Henrik Nordström píše v So 24. 07. 2010 v 03:28 +0200:
Trying to get Fedora running on an old Mybook World, but can not seem to get it to run at all..
From what I understand this is the same that Peter Lemenkov successfully
got Fedora 11 running on some time ago, but I only get "illegal instruction" when trying to run anything..
it most likely means that the CPU is lower architecture than the Fedora required minimum (armv5te), like armv4, what is the exact spec of the CPU?
Dan
lör 2010-07-24 klockan 09:37 +0200 skrev Dan Horák:
it most likely means that the CPU is lower architecture than the Fedora required minimum (armv5te), like armv4, what is the exact spec of the CPU?
It should be supported from what I gather.
/proc/cpuinfo:
Processor : ARM926EJ-Sid(wb) rev 5 (v5l) BogoMIPS : 99.73 Features : swp half thumb fastmult edsp java CPU implementer : 0x41 CPU architecture: 5TEJ CPU variant : 0x0 CPU part : 0x926 CPU revision : 5 Cache type : write-back Cache clean : cp15 c7 ops Cache lockdown : format C Cache format : Harvard I size : 32768 I assoc : 4 I line length : 32 I sets : 256 D size : 32768 D assoc : 4 D line length : 32 D sets : 256
Hardware : Oxsemi NAS Revision : 0000
Regards Henrik
On 07/24/10 10:12, Somebody in the thread at some point said:
lör 2010-07-24 klockan 09:37 +0200 skrev Dan Horák:
it most likely means that the CPU is lower architecture than the Fedora required minimum (armv5te), like armv4, what is the exact spec of the CPU?
It should be supported from what I gather.
/proc/cpuinfo:
Processor : ARM926EJ-Sid(wb) rev 5 (v5l) BogoMIPS : 99.73 Features : swp half thumb fastmult edsp java CPU implementer : 0x41 CPU architecture: 5TEJ
Is it an issue with EABI not in the kernel?
That would cause exactly illegal instruction faults IIUI.
Check for
CONFIG_AEABI=y
in your kernel config.
-Andy
lör 2010-07-24 klockan 13:27 +0200 skrev Henrik Nordström:
lör 2010-07-24 klockan 10:49 +0100 skrev Andy Green:
Is it an issue with EABI not in the kernel?
That would cause exactly illegal instruction faults IIUI.
Thanks. Looks like I need to build my own kernel.. # CONFIG_AEABI is not set
Getting a little further.. instead of illegal instruction I now have
FATAL: kernel too old
Kernel is 2.6.17.14 (patched with board support)
Using the F11 rootfs. (F12 gives same result, obviously)
Is 2.6.17.14 really too old, or am I missing more kernel options?
Regards Henrik
On 07/25/10 11:49, Somebody in the thread at some point said:
lör 2010-07-24 klockan 13:27 +0200 skrev Henrik Nordström:
lör 2010-07-24 klockan 10:49 +0100 skrev Andy Green:
Is it an issue with EABI not in the kernel?
That would cause exactly illegal instruction faults IIUI.
Thanks. Looks like I need to build my own kernel.. # CONFIG_AEABI is not set
Getting a little further.. instead of illegal instruction I now have
FATAL: kernel too old
Who is actually saying that? Presumably it's too old to have some specific feature something needs, eg, udev?
Kernel is 2.6.17.14 (patched with board support)
Using the F11 rootfs. (F12 gives same result, obviously)
Is 2.6.17.14 really too old, or am I missing more kernel options?
Too old for "something" I guess it is true.
Did you try coming up with init=/bin/bash and see if what creates that message is bypassed?
-Andy
Henrik Nordström píše v Ne 25. 07. 2010 v 12:49 +0200:
lör 2010-07-24 klockan 13:27 +0200 skrev Henrik Nordström:
lör 2010-07-24 klockan 10:49 +0100 skrev Andy Green:
Is it an issue with EABI not in the kernel?
That would cause exactly illegal instruction faults IIUI.
Thanks. Looks like I need to build my own kernel.. # CONFIG_AEABI is not set
Getting a little further.. instead of illegal instruction I now have
FATAL: kernel too old
it's glibc message ...
Kernel is 2.6.17.14 (patched with board support)
Using the F11 rootfs. (F12 gives same result, obviously)
Is 2.6.17.14 really too old, or am I missing more kernel options?
Fedora glibc required kernel at least 2.6.18 until this Wednesday, now it requires 2.6.32, the minimum kernel version defines how much compatibility stuff is included into glibc
Dan
sön 2010-07-25 klockan 12:57 +0200 skrev Dan Horák:
Fedora glibc required kernel at least 2.6.18 until this Wednesday, now it requires 2.6.32, the minimum kernel version defines how much compatibility stuff is included into glibc
Ok. So I'll either have to update kernel to at least 2.6.18, or rebuild glibc with lower kernel requirement..
But first I will grab the F10 root so I get something running.. that should work fine.. (requires 2.6.9)
Regards Henrik
sön 2010-07-25 klockan 15:30 +0200 skrev Henrik Nordström:
Ok. So I'll either have to update kernel to at least 2.6.18, or rebuild glibc with lower kernel requirement..
Working on updating the kernel. Have 2.6.34.1 booting, but have not yet ported the sATA driver.
F10 seems to run quite fine with the "original" kernel + EABI support, but for this project to be worthwhile I need to go beyond that.
Regards Henrik
2010/7/29 Henrik Nordström henrik@henriknordstrom.net:
sön 2010-07-25 klockan 15:30 +0200 skrev Henrik Nordström:
Ok. So I'll either have to update kernel to at least 2.6.18, or rebuild glibc with lower kernel requirement..
Working on updating the kernel. Have 2.6.34.1 booting, but have not yet ported the sATA driver.
Great work! Could you, please, set up public git repository for this task?
-
Andy Green -
Dan Horák -
Henrik Nordström -
Peter Lemenkov