On Wednesday, 5 September 2018 17:41:01 CEST Daniel P. Berrangé wrote:
On Wed, Sep 05, 2018 at 04:37:05PM +0100, Richard W.M. Jones wrote:
>
> Here's a fun one:
>
> + guestfish -N test-virt-sparsify-in-place-fstrim-unsupported.img=fs:vfat exit
> + virt-sparsify --in-place test-virt-sparsify-in-place-fstrim-unsupported.img
> + tee test-virt-sparsify-in-place-fstrim-unsupported.log
> [ 2.4] Trimming /dev/sda1
> [ 7.5] Sparsify in-place operation completed with no errors
> + grep 'warning:.*fstrim'
test-virt-sparsify-in-place-fstrim-unsupported.log
> FAIL test-virt-sparsify-in-place-fstrim-unsupported.sh (exit status: 1)
>
> We expect (for the purposes of the regression test) that vfat
> filesystems cannot be trimmed. It turns out that fstrim for vfat has
> now been implemented in Linux (commit f663b5b38fff) :-) Thanks
> Wentao Wang (this is actually great for virt-v2v).
>
> So we need to find another filesystem which doesn't support fstrim.
> Or maybe just delete this regression test.
>
> Thoughts?
What about filesystems such as minix or cramfs (their mkfs helpers are
part of util-linux), or nilfs2/jfs/hfsplus?
Presumably something ancient like ext2 will not support it, and is
unlikely
to be given it given that its ancient with no active development.
Nowadays there is no more separate ext2 module in linux, and the ext4
module handles it as well.
--
Pino Toscano