[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

linux 3.10.1 with initrd (was Re: linux 3.8.12-1 (atari flavour) does not boot)

Dixi quod…

>I guess if it’s like that, I can have a go as well.

Did one now, and it boots. It uses initrd now, which
means one has to copy-out the initrd every time it is
regenerated, but also adds flexibility (root filesystem
not ext4fs, root on nfs, nbd, etc).

I disabled quite a lot of stuff (like USB, WLAN) that
we could, in theory, have (with the NetUSBee or PCMCIA)
for now. Also, NTFS and Ceph and several network protocols.

My opinion here is: most are modules, but they also add
dependencies to the core kernel image, which we’d like
to keep smaller. We can always add back things later
that people miss in production.

I also threw out SELinux/AppArmor/TOMOYO/whatever support.

I’d like the m68k people to read the whole eMail and test
on all their boxen, at least Amiga/Atari/Macintosh
though the image also has BVME and MVME support
(no Sun3 or Q40 but that can easily be added ofc).

Also, I’d like the m68k people *and* the Debian Linux
kernel people to review the configuration, to see whether
there are things either can absolutely not live with:

>It’s probably a bit tricky getting it into the Debian
>source package, which apparently generates the configs
>from smaller fragments… I’ll read the kernel handbook
>in the hopes to find instructions there (for my own
>one-off testbuilds I just put full configs in, but

I did put the full configs in, for now. I’ve got no
idea how to reliably reduce them, and since I got rid
of all the kernel flavours it’s just a few KiB.

-rw-r--r-- 1 root root 3970373 Jul 20 17:10 /boot/initrd.img-3.10-0+m68k.1-m68k

Someone seems to have put “set -x” into the initrd,
but it boots fine:

root@ara3:~ # uname -a
Linux ara3.mirbsd.org 3.10-0+m68k.1-m68k #1 Debian 3.10.1-1+m68k.1 (2013-07-19) m68k GNU/Linux
root@ara3:~ # lsmod
Module                  Size  Used by
evdev                   6613  0
ext3                   96402  1
mbcache                 3606  1 ext3
jbd                    29270  1 ext3

(This is ARAnyM, so the NatFeat disc/net/… drivers are
compiled in; the ide-core and SCSI stuff is put in modules.
By the way, this is MODULES=most; is there anything we’ll
need for MODULES=dep to work, considering we have modules
like ide-core (instead of libata) and HW specific drivers
that other platforms don’t? To reduce the initrd size.)

So good, so far. An ARAnyM log of the build (kernel
messages) is attached. There’s one thing sticking out:

[    0.160000] WARNING: at /root/linux-3.10.1/init/main.c:698 do_one_initcall+0x12e/0x13a()
[    0.160000] initcall param_sysfs_init+0x0/0x1a4 returned with disabled interrupts

This seems to happen for several initcalls. Geert et al,
can you have a look at them, they’re scary ☺

(gnutls can also be used, but if you are compiling lynx for your own use,
there is no reason to consider using that package)
	-- Thomas E. Dickey on the Lynx mailing list, about OpenSSL

Attachment: log.gz
Description: Binary data

Reply to: