From: Harald Hoyer <harald-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
To: Joe Lawrence <Joe.Lawrence-7+ureL1bLXNBDgjK7y7TUQ@public.gmane.org>
Cc: initramfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: Re: Out of tree drivers missing from initramfs
Date: Wed, 23 Jan 2013 16:19:47 +0100 [thread overview]
Message-ID: <50FFFF93.2080209@redhat.com> (raw)
In-Reply-To: <alpine.DEB.2.02.1301221612280.31163-ceYW5R1vr2hcrvxNGtJwk767FWEIOpWeVpNB7YpNyf8@public.gmane.org>
Am 22.01.2013 23:04, schrieb Joe Lawrence:
> Hello Harold / list,
>
> I'm trying to build an initramfs that includes an out-of-tree kernel
> module for the mpt2sas driver without much success (by default).
>
> The modified driver in question resides at:
>
> /lib/modules/3.8.0-rc4.sra+/updates/lsb-ft/extra/mpt2sas/mpt2sas.ko
>
> and recent versions of dracut exclude this driver from the generated
> initramfs. This is on a F18 install, 3.8.0-rc4 kernel and a newly
> recompiled out-of-tree mpt2sas ko file.
>
> Using git bisect, I found dracut *would* include the updated .ko until...
>
> d8aeb3a72afe03e426f4ea9f07c857c9a499144f is the first bad commit
> commit d8aeb3a72afe03e426f4ea9f07c857c9a499144f
> Author: Harald Hoyer <harald-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> Date: Wed Jun 27 09:35:39 2012 +0200
>
> dracut.sh, dracut-functions.sh: do lazy kernel module dep resolving
>
> First, we just install the kernel module and keep track, what we
> installed. At the very end, call modprobe and modinfo for firmwares to
> resolve all kernel module dependencies. This speeds up image creation,
> because we can call modprobe and modinfo with many modules.
>
> I notice that find_kernel_modules_by_path() filters out any string that
> does not start with 'kernel' as well as the given path, ie 'drivers/scsi'
> (by design I would assume).
>
> If the in-tree mpt2sas driver is loaded, then
> find_kernel_modules_by_path() would echo the driver since:
>
> $1 = drivers/scsi
> $a = kernel/drivers/scsi/mpt2sas/mpt2sas.ko
>
> but when the out-of-tree driver is used, then
>
> $1 = drivers/scsi
> $a = updates/lsb-ft/extra/mpt2sas/mpt2sas.ko
>
> fails this check.
>
> 1 - Should Dracut be checking the /lib/modules/ver/{updates,extras}
> directories as well as the in-tree kernel directory for drivers?
yes
>
> 2 - Do modules underneath /lib/modules/ver/{updates,extras} need to
> include their respective directory structure (ie, 'drivers/scsi').
yes
>
> I can avoid this by forcing dracut to include out-of-tree drivers using
> the add_drivers directive, but I defer to the dracut coders here on how I
> can best include them by default.
>
> Regards,
>
> -- Joe
Thanks for reporting this, will fix this bug!
next prev parent reply other threads:[~2013-01-23 15:19 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-01-22 22:04 Out of tree drivers missing from initramfs Joe Lawrence
[not found] ` <alpine.DEB.2.02.1301221612280.31163-ceYW5R1vr2hcrvxNGtJwk767FWEIOpWeVpNB7YpNyf8@public.gmane.org>
2013-01-23 14:58 ` Joe Lawrence
2013-01-23 15:19 ` Harald Hoyer [this message]
[not found] ` <50FFFF93.2080209-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-01-24 22:26 ` Joe Lawrence
[not found] ` <alpine.DEB.2.02.1301241718240.31163-ceYW5R1vr2hcrvxNGtJwk767FWEIOpWeVpNB7YpNyf8@public.gmane.org>
2013-01-25 12:08 ` Harald Hoyer
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=50FFFF93.2080209@redhat.com \
--to=harald-h+wxahxf7alqt0dzr+alfa@public.gmane.org \
--cc=Joe.Lawrence-7+ureL1bLXNBDgjK7y7TUQ@public.gmane.org \
--cc=initramfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox