Insmod errors at boot time before usplash

Bug #21540 reported by Charles Majola
54
Affects Status Importance Assigned to Milestone
usplash (Ubuntu)
Fix Released
Low
Jeff Bailey

Bug Description

/init: 63: cannot open /sys/bus/scsi/devices/*/type: No such file
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/console/bitblit.ko': -1 File Exists
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/console/font.ko': -1 File Exists
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/console/tileblit.ko': -1 File Exists
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/console/fbcon.ko': -1 File Exists
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/cfbcopyarea.ko': -1 File Exists
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/cfbfillrect.ko': -1 File Exists
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/cfbimgblit.ko': -1 File Exists
insmod: error inserting
'/lib/modules/2.6.12-8-386/kernel/drivers/video/softcursor.ko': -1 File Exists

Revision history for this message
Charles Majola (chmj) wrote :

This happens with the preview release

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Jeff, do we attempt to load module twice?

Charles, these errors are harmless.

Revision history for this message
Matt Zimmerman (mdz) wrote :

*** Bug 21823 has been marked as a duplicate of this bug. ***

Revision history for this message
Alvaro Muñoz (alvaro-munoz) wrote :

I get the exact same output right after grub loads the kernel. Then it goes to
Usplash.
It doesnt affect the system (as far as I've experienced), but they're still
errors...

Revision history for this message
Alvaro Muñoz (alvaro-munoz) wrote :

FIXED in 2.6.12-9-386 kernel.

Revision history for this message
Alvaro Muñoz (alvaro-munoz) wrote :

after todays upgrade, the messages appear again, but now pointing at the -9 kernel.
Also, at restart/shutdown, something along the lines of "K01usplash: bootsplash
command not found" appears...

Revision history for this message
Marc Wiriadisastra (strikeforce) wrote :

Yeah exactly the same for me as well. Only just noticed in on reboot but thats
about it. Hasn't affected me in any way but the errors are there.

Revision history for this message
Matt Zimmerman (mdz) wrote :

*** Bug 22723 has been marked as a duplicate of this bug. ***

Revision history for this message
Matt Zimmerman (mdz) wrote :

Why can we not simply use modprobe for this?

Revision history for this message
Matthew Garrett (mjg59) wrote :

A good question. I asked Jeff, but I can't remember the explanation. Jeff?

Revision history for this message
Jeff Bailey (jbailey) wrote :

(In reply to comment #10)
> A good question. I asked Jeff, but I can't remember the explanation. Jeff?

It runs before depmod is run.

We could use modprobe --dry-run to generate the insmod commands at built time
and make sure we get them right, though.

Revision history for this message
Matt Zimmerman (mdz) wrote :

(In reply to comment #11)
> (In reply to comment #10)
> > A good question. I asked Jeff, but I can't remember the explanation. Jeff?
>
> It runs before depmod is run.
>
> We could use modprobe --dry-run to generate the insmod commands at built time
> and make sure we get them right, though.

depmod should be run when the initramfs is generated. Since no modules are
added to the initramfs at boot time, there should not even be a need to run
depmod from initramfs.

But if this is necessary for some reason of which I'm unaware, why don't we just
run it earlier?

Revision history for this message
Yvan AMILIN (yvanamilin-deactivatedaccount) wrote :

Hint: the file /usr/share/initramfs-tools/scripts/init-top/usplash contains the
following lines:

if [ $SPLASH = "true" -a $VESA = "false" ]; then
 insmod /lib/modules/`uname -r`/kernel/drivers/video/console/bitblit.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/console/font.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/console/tileblit.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/console/fbcon.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/cfbcopyarea.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/cfbfillrect.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/cfbimgblt.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/softcursor.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/vgastate.ko
 insmod /lib/modules/`uname -r`/kernel/drivers/video/vga16fb.ko
 mknod /dev/fb0 c 29 0
 mknod /dev/tty0 c 4 0
 /sbin/usplash -c &
fi

Revision history for this message
Jeff Bailey (jbailey) wrote :

(In reply to comment #12)
>
> depmod should be run when the initramfs is generated. Since no modules are
> added to the initramfs at boot time, there should not even be a need to run
> depmod from initramfs.
>
> But if this is necessary for some reason of which I'm unaware, why don't we just
> run it earlier?

initramfs allows other modules to be concatenated on to the end, either after
its built or at boot time by the boot loader. Long term, the idea is that the
bootloader should be able to add the modules that it wants and not load others
(So that an initramfs that doesn't need network drivers doesn't have them and
it's assembled at boot time out of generic pieces). Because the modules in the
initramfs then isn't known at boot time, the depmod has to be done at runtime.

(Sorry about the lag, I apparnetly didn't click "commit")

Tks,
Jeff Bailey

Revision history for this message
Matt Zimmerman (mdz) wrote :

Since we don't actually do this yet, and since usplash works by hooking into the
initramfs build, it is safe for depmod to run before usplash's runtime hook

Revision history for this message
Matt Zimmerman (mdz) wrote :

*** Bug 22823 has been marked as a duplicate of this bug. ***

Revision history for this message
Matt Zimmerman (mdz) wrote :

*** Bug 22831 has been marked as a duplicate of this bug. ***

Revision history for this message
Matt Zimmerman (mdz) wrote :

At any rate, in case it wasn't clear, this needs to be addressed for 5.10. The
user community is confused and worried by these errors.

Revision history for this message
Jeff Bailey (jbailey) wrote :

I've seen two confirmations that this is fixed by usplash 0.1-12. If you're
seeing this, please regenerate, your initramfs with dpkg-reconfigure
linux-image-$(uname -r)

Tks,
Jeff Bailey

Revision history for this message
Sam Argiro (argirosa) wrote :

(In reply to comment #19)
> I've seen two confirmations that this is fixed by usplash 0.1-12. If you're
> seeing this, please regenerate, your initramfs with dpkg-reconfigure
> linux-image-$(uname -r)
>
> Tks,
> Jeff Bailey

Hi Mr Bailey,

I had this error and it disappeared w/ one of the Usplash updates and now it's
back...I currently have Usplash 0.1-16. Should we wait for a fix via synaptic
or apply the reconfigure mentioned here? I'm not quite sure how the command
above works, would you mind replying w/ and example? Thank you.

Revision history for this message
Jeff Bailey (jbailey) wrote :

sudo dpkg-reconfigure linux-image-$(uname -r)

is the command you literally type in to regenerate the initramfs. Then reboot
and the errors should be gone.

Tks,
Jeff Bailey

Revision history for this message
Matt Zimmerman (mdz) wrote :

*** Bug 22927 has been marked as a duplicate of this bug. ***

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.