Linux Btrfs filesystem development
 help / color / mirror / Atom feed
From: Josef Bacik <josef@toxicpanda.com>
To: Roman Mamedov <rm@romanrm.net>
Cc: linux-btrfs@vger.kernel.org, kernel-team@fb.com
Subject: Re: [PATCH] btrfs: do not allow -o compress-force to override per-inode settings
Date: Mon, 30 Nov 2020 10:10:42 -0500	[thread overview]
Message-ID: <df787c30-8a5e-0256-a4c9-baa3e3556a39@toxicpanda.com> (raw)
In-Reply-To: <20201130200116.79a710fe@natsu>

On 11/30/20 10:01 AM, Roman Mamedov wrote:
> On Mon, 30 Nov 2020 09:50:13 -0500
> Josef Bacik <josef@toxicpanda.com> wrote:
> 
>> The thing you're missing is that when we do chattr -c we're setting NOCOMPRESS
>> on the file.
> 
> Wow, and does this need a previously set +c to work? Or just -c on an already
> -c file will change the Btrfs flag under the hood? Seems to be very weird in
> any case, as from the user perspective there's no way to view the current
> status of that flag, with the only way to change it being via a side-effect of
> another operation.
> 
>>    If chattr -c is supposed to just be the removal of +c, then btrfs is doing the
>> wrong thing by setting NOCOMPRESS.
> 
> I would agree with that.
> 
>> I guess the question is what do we want?  Do we want to only allow the user to
>> indicate we want compression, or do we want to allow them to also indicate that
>> they don't want compression?  If we don't want to enable them to disable
>> compression for a file, then this patch needs to be thrown away, but then we
>> also need to fix up all the places we set NOCOMPRESS when we clear these flags.
> 
> The patch also seems to prioritize "no compress if compression ratio is bad"
> over compress-force, whereas the whole point of compress-force feels to be to
> compress no matter what, especially no matter what are the possibly imperfect
> compression ratio estimates.
> 

Right, but if we have compress-force we don't set NOCOMPRESS if the compression 
is bad, so theoretically we shouldn't ever really have that problem?  But I 
agree, this is a weird source of ambiguity.  I'm thinking the best solution is 
to stop setting NOCOMPRESS except in the bad compression case, and then figure 
out a different mechanism to force no compression and deal with that separately. 
  Thanks,

Josef

  reply	other threads:[~2020-11-30 15:11 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-11-30 13:46 [PATCH] btrfs: do not allow -o compress-force to override per-inode settings Josef Bacik
2020-11-30 14:08 ` Roman Mamedov
2020-11-30 14:27   ` Amy Parker
2020-11-30 14:50   ` Josef Bacik
2020-11-30 15:01     ` Roman Mamedov
2020-11-30 15:10       ` Josef Bacik [this message]
2020-11-30 15:17         ` Roman Mamedov
2020-11-30 15:04     ` Hugo Mills
2020-11-30 15:12       ` Josef Bacik
2020-11-30 17:28       ` sys

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=df787c30-8a5e-0256-a4c9-baa3e3556a39@toxicpanda.com \
    --to=josef@toxicpanda.com \
    --cc=kernel-team@fb.com \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=rm@romanrm.net \
    /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