linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Lennart Poettering <lennart@poettering.net>
To: Josef Bacik <josef@toxicpanda.com>,
	David Sterba <dsterba@suse.com>,
	linux-btrfs@vger.kernel.org
Subject: btrfs: please undo removal of "norecovery" (userspace breakage)
Date: Tue, 21 May 2024 10:20:31 +0200	[thread overview]
Message-ID: <ZkxZT0J-z0GYvfy8@gardel-login> (raw)

Hi!

So we learnt the hard way that btrfs dropped support for the the
"norecovery" mount option recently (kernel 6.8).

This basically broke any released version of systemd, because this
option is what we use to ensure that btrfs images mounted off loopback
block devices do not modify the backing file (because unfortunately,
the loopback block devices allow writes go through even if marked
read-only).

I understand that the option was marked "obsolete" for a while, but
this was not visible to us.

Also why even make this obsolete at all? I mean, the functionality is
still there, so is there any value in randomly renaming this option?
In particular as the other big file systems (i.e. ext3, ext4, xfs) all
have a flag of the same name doing the same thing. So why depart from
this at all as the only outlier? Where's the value in that?

(This issue is tracked in systemd here:
https://github.com/systemd/systemd/pull/32892 – we commited a
work-around for future versions of systemd, but that doesn't fix
things if older systemd versions are used on newer kernels.)

Hence, please reconsider this, this seems to be breakage for the sake
of breakage with no real benefit at all?

I mean, I respect your right to deprecate and obsolete stuff, but in
this case this seems just entirely random and ignoring how much
userspace this will break.

It's fine to hide stuff in docs and so on you think is redundant, but
if you take the kernel developers mantra of "we never break userspace"
seriously, you should still keep the option around to be parsed,
because this *did* break userspace, quite massively. And there's so
little effort necessary to keep compat here, just treat "norecovery"
as an alias "rescue=nologreplay".

Thank you,

Lennart

             reply	other threads:[~2024-05-21  8:26 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-05-21  8:20 Lennart Poettering [this message]
2024-05-21  8:55 ` btrfs: please undo removal of "norecovery" (userspace breakage) Qu Wenruo

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=ZkxZT0J-z0GYvfy8@gardel-login \
    --to=lennart@poettering.net \
    --cc=dsterba@suse.com \
    --cc=josef@toxicpanda.com \
    --cc=linux-btrfs@vger.kernel.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;
as well as URLs for NNTP newsgroup(s).