Re: kernel errors

["Thomas Schmitt" <scdbackup@gmx.net>]

Hi,

i wrote:
> > Maybe the script which runs blkid or alike has vanished during the recent
> > reconstruction of the initrd which fixed the problems ?

Richmond wrote:
> I didn't need to reconstruct initrd to cause the problems. As far as I
> remember all I did was destroy the swap space, having commented-it-out
> of fstab. So I imagin that the scripts are the same.

I rather meant the run which you meantioned in your mail of Thu,
02 Feb 2023 14:05:37 +0000:

> > > sudo update-initramfs -u
> > > After I did this, the errors went away.

It is possible that there were files in /scripts/local-block before this
run.


> find /tmp/initrd21/ -print|xargs grep -i blkid|less
> /tmp/initrd21/scripts/functions: DEV="$(blkid -l -t "$DEV" -o device)" || return 1

This is the kind of blkid run which i expect to cause the mislead read
attempts to the pseudo-end of /dev/sr0. If the given attribute is not
found, then it cycles over all block devices, not knowing about the
purpose of resuming and the usefulness of the device types.

It is in function resolve_device() which obviously shall find a device
matching either LABEL, UUID, PARTLABEL, or PARTUUID.

A plausible caller of resolve_device() would be in hooks/resume:

	if resume_dev_node="$(resolve_device "$RESUME")" && \
	   blkid -p -n swap "$resume_dev_node" >/dev/null 2>&1; then

But i fail to make the connection to the loop with the messages about
/scripts/local-block.


Have a nice day :)

Thomas