On Thu, May 28, 2020 at 4:20 PM Alfonso Sanchez-Beato <
<email address hidden>> wrote:
> Some comments on the debdiff:
>
> Use
> blkid -s TYPE "$writable_part" -o value
> instead of
> blkid -s TYPE "$writable_part" | awk '{print $2}' | grep -o '".*"' | tr -d
> '"'
>
I've revised it accordingly.
> No need to redirect to >>$LOGFILE in resize_filesystem() which is
> already redirected when called. Also, you should not redirect to
> /dev/kmsg - I guess that was for debugging purposes.
>
Revised this as well.
> Is it really necessary to mount the partition to find out the filesystem
> usage?
> "btrfs filesystem show" apparently can work on devices according to [1].
> But it also looks like btrfs commands are mostly used with mounted systems,
> so it is fine if you want to keep things like that.
>
> Yeap, thanks for improving this. I also found out the original command
will have a potential bug as I should not cut the specific length for
space. I've changed this for one line only.
Finally, you will need to use copy_exec for btrfs and blkid to the
> fixtrtc hook to get them included in the initramfs (although maybe blkid
> comes from busybox in this case, not sure).
>
> From my observation, btrfs-progs has a hook
(/usr/share/initramfs-tools/hooks/btrfs) to include btrfs related binaries.
I think it should be fine here.
blkid is indeed included by udev's hook as udev is also one of dependencies
for initramfs-tools-core.
Hi Alfonso,
On Thu, May 28, 2020 at 4:20 PM Alfonso Sanchez-Beato <
<email address hidden>> wrote:
> Some comments on the debdiff:
>
> Use
> blkid -s TYPE "$writable_part" -o value
> instead of
> blkid -s TYPE "$writable_part" | awk '{print $2}' | grep -o '".*"' | tr -d
> '"'
>
I've revised it accordingly.
> No need to redirect to >>$LOGFILE in resize_filesystem() which is
> already redirected when called. Also, you should not redirect to
> /dev/kmsg - I guess that was for debugging purposes.
>
Revised this as well.
> Is it really necessary to mount the partition to find out the filesystem
> usage?
> "btrfs filesystem show" apparently can work on devices according to [1].
> But it also looks like btrfs commands are mostly used with mounted systems,
> so it is fine if you want to keep things like that.
>
> Yeap, thanks for improving this. I also found out the original command
will have a potential bug as I should not cut the specific length for
space. I've changed this for one line only.
Finally, you will need to use copy_exec for btrfs and blkid to the initramfs- tools/hooks/ btrfs) to include btrfs related binaries. tools-core.
> fixtrtc hook to get them included in the initramfs (although maybe blkid
> comes from busybox in this case, not sure).
>
> From my observation, btrfs-progs has a hook
(/usr/share/
I think it should be fine here.
blkid is indeed included by udev's hook as udev is also one of dependencies
for initramfs-
Thanks,
Woodrow
Otherwise, looks good to me. /btrfs. wiki.kernel. org/index. php/Manpage/ btrfs-filesyste m /bugs.launchpad .net/bugs/ 1881074 tools-devices package in Ubuntu: /bugs.launchpad .net/ubuntu/ +source/ initramfs- tools-devices/ +bug/1881074/ +subscriptions
>
> [1] https:/
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https:/
>
> Title:
> Support btrfs when writable is converted to btrfs
>
> Status in initramfs-
> New
>
> Bug description:
> It's a special case that we may need to convert rootfs from ext4 to
> btrfs, and we have to resize btrfs inside initramfs as well.
>
> To manage notifications about this bug go to:
>
> https:/
>
--
Woodrow Shen
Software Engineer, Canonical ltd.
Devices Enablement | OEM Enablement, Taipei