Re: older LFS partition boots, the new one does not - with same kernel

This is a multi-part message in MIME format...

------------=_1770114676-1253-3492
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


Inviato da iPhone

> Il giorno 2 feb 2026, alle ore 11:10, Thomas Seeling <[email protected]=
inuxfromscratch.org> ha scritto:
>=20
> =EF=BB=BFHallo dear list,
>=20
> I'd like to return to a problem I described a few weeks ago, but with a n=
ew twist, and ask for advice please.
>=20
> Over the weekend I compiled the latest LFS-dev book on an ancient machine=
, a Core2duo in 32 bit mode, i.e. the system I used to compile is LFS 12.0 =
32 bit.
> The full LFS compile went fine, I fixed /etc/fstab, added a root password=
, copied kernel modules to the correct position, and - again - had some str=
ange exceptions upon kernel boot about ELF signatures (see below).
>=20

Did you use the new kernel at all when building the system?

How did you copy the modules from the old system to the new one?


> What's absolutely strange to me: the exact same kernel works when booting=
 the older 12.0 without any errors, loading ethernet and USB drivers and is=
 nicely usable.
> So this means it's not a defect RAM module - else the 12.0 would show the=
 same symptoms. The system has i915 chipset and tg3 ethernet module.
>=20
> The only other explanation to me was a hardware defect on the disk, but I=
 did a binary compare and the kernel files are identical and there are no r=
ead errors from /mnt/lfs.
> The system *does* finally finish executing /sbin/init after reporting the=
 exceptions.
> I see the login prompt, but I don't have a usable keyboard, and there's a=
n error message that the ethernet interface does not exist.
> Another idea I had was the decompression code for the gz modules (opcodes=
 that are not present on the old CPU), but then the same symptoms would als=
o occur on the other LFS partition.
>=20
> root@buero ~ # find /lib/modules/6.12.68/ /boot/vmlinuz* -type f | while =
read f; do cmp -s {,/mnt/lfs}${f}; done
> <no output, i.e. no differences found>
>=20
> Jan 31 12:47:43 buero kernel: md: ... autorun DONE.
> Jan 31 12:47:44 buero kernel: EXT4-fs (sda1): orphan cleanup on readonly =
fs
> Jan 31 12:47:44 buero kernel: EXT4-fs (sda1): mounted filesystem 1a3b33ce=
-29c8-4523-b5da-ec2a3fac124b ro with ordered data mode. Quota mode: none.
> Jan 31 12:47:44 buero kernel: VFS: Mounted root (ext4 filesystem) readonl=
y on device 8:1.
> Jan 31 12:47:44 buero kernel: devtmpfs: mounted
> Jan 31 12:47:44 buero kernel: Freeing unused kernel image (initmem) memor=
y: 760K
> Jan 31 12:47:44 buero kernel: Write protecting kernel text and read-only =
data: 10736k
> Jan 31 12:47:44 buero kernel: Run /sbin/init as init process
> Jan 31 12:47:44 buero kernel:   with arguments:
> Jan 31 12:47:44 buero kernel:     /sbin/init
> Jan 31 12:47:44 buero kernel:     nosplash
> Jan 31 12:47:44 buero kernel:   with environment:
> Jan 31 12:47:44 buero kernel:     HOME=3D/
> Jan 31 12:47:44 buero kernel:     TERM=3Dlinux
> Jan 31 12:47:44 buero kernel: tsc: Refined TSC clocksource calibration: 3=
191.999 MHz
> Jan 31 12:47:44 buero kernel: clocksource: tsc: mask: 0xffffffffffffffff =
max_cycles: 0x2e02c398820, max_idle_ns: 440795273435 ns
> Jan 31 12:47:44 buero kernel: clocksource: Switched to clocksource tsc
> Jan 31 12:47:44 buero kernel: stty (581) used greatest stack depth: 5848 =
bytes left
> Jan 31 12:47:45 buero kernel: random: crng init done
> Jan 31 12:47:45 buero udevd[707]: Starting systemd-udevd version 259
> Jan 31 12:47:45 buero udevd[708]: Using default interface naming scheme '=
v259'.
> Jan 31 12:47:46 buero kernel: ------------[ cut here ]------------
> Jan 31 12:47:46 buero kernel: WARNING: CPU: 0 PID: 916 at mm/highmem.c:62=
2 kunmap_local_indexed+0xd8/0xf0
> Jan 31 12:47:46 buero kernel: Modules linked in:
> Jan 31 12:47:46 buero kernel: CPU: 0 UID: 0 PID: 916 Comm: (udev-worker) =
Not tainted 6.12.68 #1
> Jan 31 12:47:46 buero kernel: Hardware name: Hewlett-Packard HP Compaq dc=
7600 Small Form Factor/09F8h, BIOS 786D1 v01.61 07/01/2009
> Jan 31 12:47:46 buero kernel: EIP: kunmap_local_indexed+0xd8/0xf0
> Jan 31 12:47:46 buero kernel: Code: ec ff 64 a1 04 b2 cd c1 85 c0 74 26 e=
8 21 e9 ec ff 5b 5e 5f 5d c3 8d 74 26 00 0f 0b e9 62 ff ff ff 90 0f 0b 90 2=
e 8d 74 26 00 <0f> 0b eb 9a 8d 74 26 00 e8 6b 89 65 00 eb d3 90 2e 8d b4 26=
 00 00
> Jan 31 12:47:46 buero kernel: EAX: 00000001 EBX: fffba000 ECX: 00046000 E=
DX: fffb9000
> Jan 31 12:47:46 buero kernel: ESI: c237e000 EDI: 00000001 EBP: c28b9e60 E=
SP: c28b9e54
> Jan 31 12:47:46 buero kernel: DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 006=
8 EFLAGS: 00010206
> Jan 31 12:47:46 buero kernel: CR0: 80050033 CR2: 010204cc CR3: 02f54000 C=
R4: 000006d0
> Jan 31 12:47:46 buero kernel: Call Trace:
> Jan 31 12:47:46 buero kernel:  module_decompress+0x12e/0x230
> Jan 31 12:47:46 buero kernel:  init_module_from_file+0x92/0xb0
> Jan 31 12:47:46 buero kernel:  __ia32_sys_finit_module+0x115/0x2c0
> Jan 31 12:47:46 buero kernel:  ia32_sys_call+0x1814/0x1e30
> Jan 31 12:47:46 buero kernel:  __do_fast_syscall_32+0x50/0xd0
> Jan 31 12:47:46 buero kernel:  do_fast_syscall_32+0x29/0x60
> Jan 31 12:47:46 buero kernel:  do_SYSENTER_32+0x15/0x20
> Jan 31 12:47:46 buero kernel:  entry_SYSENTER_32+0x98/0xf9
> Jan 31 12:47:46 buero kernel: EIP: 0xb7f15579
> Jan 31 12:47:46 buero kernel: Code: b8 01 10 06 03 74 b4 01 10 07 03 74 b=
0 01 10 08 03 74 d8 01 00 00 00 00 00 00 00 00 00 00 00 00 00 51 52 55 89 e=
5 0f 34 cd 80 <5d> 5a 59 c3 90 90 90 90 8d 76 00 58 b8 77 00 00 00 cd 80 90=
 8d 76
> Jan 31 12:47:46 buero kernel: EAX: ffffffda EBX: 00000019 ECX: b754d44f E=
DX: 00000004
> Jan 31 12:47:46 buero kernel: ESI: 0101b530 EDI: 00fc4240 EBP: 00000000 E=
SP: bfea394c
> Jan 31 12:47:46 buero kernel: DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007=
b EFLAGS: 00000296
> Jan 31 12:47:46 buero kernel: ---[ end trace 0000000000000000 ]---
> Jan 31 12:47:46 buero kernel: Invalid ELF section hdr(type 569694567)
> Jan 31 12:47:46 buero kernel: ------------[ cut here ]------------
> Jan 31 12:47:46 buero kernel: DEBUG_LOCKS_WARN_ON(current->kmap_ctrl.idx)
> Jan 31 12:47:46 buero kernel: WARNING: CPU: 0 PID: 916 at include/linux/h=
ighmem-internal.h:17 syscall_exit_to_user_mode+0xfa/0x120
> Jan 31 12:47:46 buero kernel: Modules linked in:
> Jan 31 12:47:46 buero kernel: CPU: 0 UID: 0 PID: 916 Comm: (udev-worker) =
Tainted: G        W          6.12.68 #1
> Jan 31 12:47:46 buero kernel: Tainted: [W]=3DWARN
> Jan 31 12:47:46 buero kernel: Hardware name: Hewlett-Packard HP Compaq dc=
7600 Small Form Factor/09F8h, BIOS 786D1 v01.61 07/01/2009
> Jan 31 12:47:46 buero kernel: EIP: syscall_exit_to_user_mode+0xfa/0x120
> Jan 31 12:47:46 buero kernel: Code: e8 cb 91 b9 ff 85 c0 0f 84 6f ff ff f=
f 8b 0d 60 59 c0 c1 85 c9 0f 85 61 ff ff ff 68 89 23 9b c1 68 e1 b0 9a c1 e=
8 b6 de 84 ff <0f> 0b 58 5a e9 49 ff ff ff 2e 8d 74 26 00 e8 e3 27 8d ff e9=
 0b ff
> Jan 31 12:47:46 buero kernel: EAX: 0000002b EBX: 00000001 ECX: 00000002 E=
DX: 00000000
> Jan 31 12:47:46 buero kernel: ESI: 00000000 EDI: 00fc4240 EBP: c28b9f84 E=
SP: c28b9f70
> Jan 31 12:47:46 buero kernel: DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 006=
8 EFLAGS: 00010086
> Jan 31 12:47:46 buero kernel: CR0: 80050033 CR2: 010204cc CR3: 02f54000 C=
R4: 000006d0
> Jan 31 12:47:46 buero kernel: Call Trace:
> Jan 31 12:47:46 buero kernel:  __do_fast_syscall_32+0x5a/0xd0
> Jan 31 12:47:46 buero kernel:  do_fast_syscall_32+0x29/0x60
> Jan 31 12:47:46 buero kernel:  do_SYSENTER_32+0x15/0x20
> Jan 31 12:47:46 buero kernel:  entry_SYSENTER_32+0x98/0xf9
> Jan 31 12:47:46 buero kernel: EIP: 0xb7f15579
> Jan 31 12:47:46 buero kernel: Code: b8 01 10 06 03 74 b4 01 10 07 03 74 b=
0 01 10 08 03 74 d8 01 00 00 00 00 00 00 00 00 00 00 00 00 00 51 52 55 89 e=
5 0f 34 cd 80 <5d> 5a 59 c3 90 90 90 90 8d 76 00 58 b8 77 00 00 00 cd 80 90=
 8d 76
> Jan 31 12:47:46 buero kernel: EAX: fffffff8 EBX: 00000019 ECX: b754d44f E=
DX: 00000004
> Jan 31 12:47:46 buero kernel: ESI: 0101b530 EDI: 00fc4240 EBP: 00000000 E=
SP: bfea394c
> Jan 31 12:47:46 buero kernel: DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007=
b EFLAGS: 00000296
> Jan 31 12:47:46 buero kernel: ---[ end trace 0000000000000000 ]---
> Jan 31 12:47:46 buero kernel: Invalid ELF section hdr(type 3901880924)
> Jan 31 12:47:46 buero kernel: Invalid ELF section hdr(type 4123171279)
> Jan 31 12:47:46 buero kernel: Invalid ELF section hdr(type 1219650857)
> Jan 31 12:47:46 buero kernel: Invalid ELF header magic: !=3D \x7fELF

It seems to me like the modules have some problems to load. I had a similar=
 issue when I tried to load a kernel module compiled previously with anothe=
r kernel (same version) that I built later.

My suggestion would be to compile a newer 32bit kernel as instructect by th=
e book

Michele

> Jan 31 12:47:46 buero last message buffered 1 times
> Jan 31 12:47:46 buero kernel: Invalid ELF section hdr(type 3192168103)
> Jan 31 12:47:46 buero kernel: Invalid ELF section hdr(type 1079286796)
> Jan 31 12:47:46 buero kernel: Invalid ELF section hdr(type 3502702874)
>=20
> I have no clue where to look further or how to do any more diagnosis.
> Why does it work on one partition, but not the fresh compile on another o=
ne?
>=20
> Stay healthy,
> Thomas Seeling
> --
> "Do you wanna be a legend or a passing footprint on the sands of time?"
>=20
> Bleiben Sie gesund,
> Thomas Seeling
> --
> "Do you wanna be a legend or a passing footprint on the sands of time?"
> --
> http://lists.linuxfromscratch.org/sympa/info/lfs-support
> Unsubscribe: See the above information page

------------=_1770114676-1253-3492
Content-Type: text/plain; charset="UTF-8"
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
MIME-Version: 1.0

-- 
http://lists.linuxfromscratch.org/sympa/info/lfs-support
Unsubscribe: See the above information page
------------=_1770114676-1253-3492--