Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 19 Mar 2026 21:56:50 -0700
From:      Warner Losh <imp@bsdimp.com>
To:        =?UTF-8?Q?Dag=2DErling_Sm=C3=B8rgrav?= <des@freebsd.org>
Cc:        Martin Matuska <mm@freebsd.org>, h v <henry.vogt@gmail.com>,  FreeBSD Current <freebsd-current@freebsd.org>
Subject:   Re: zfs module not loaded anymore on e.g MINIMAL
Message-ID:  <CANCZdfpF7GYLRPWvRSX1uKqKLZGPZB5Bg-Sd8t3xabVSHoCeOg@mail.gmail.com>
In-Reply-To: <86wlz73g23.fsf@ltc.des.dev>
References:  <5913b3ac-2ed7-42ab-9c26-0d204c38c578@gmail.com> <CANCZdfq8gxi2ywvfvCdC%2Bsoqu_9-XPFrvqNNeGqjsJbUfkLoXg@mail.gmail.com> <fdb20aa1-46e6-41ff-83a4-8549ed56c316@gmail.com> <CANCZdfpQnALFj%2B5Hd0xC-5yO3wzjOmZG9TYQL6tWbHr_%2By%2BRFQ@mail.gmail.com> <299d29ce-98b9-4daf-aed7-a8cc9cd3a45f@FreeBSD.org> <86a4w3517o.fsf@ltc.des.dev> <CANCZdfpB7HoeQe9_rxHRQocYGM8DcZn9un-9YdmjBS8tzr6vsQ@mail.gmail.com> <861phf4uu3.fsf@ltc.des.dev> <86wlz73g23.fsf@ltc.des.dev>

index | next in thread | previous in thread | raw e-mail

[-- Attachment #1 --]
On Thu, Mar 19, 2026 at 10:05 AM Dag-Erling Smørgrav <des@freebsd.org>
wrote:

> Dag-Erling Smørgrav <des@FreeBSD.org> writes:
> > Warner Losh <imp@bsdimp.com> writes:
> > > Any reason zstd can't be a module?
> > I had the same thought but it's going to take a bit of work.  We'll have
> > to split up sys/kern/subr_compress.c into separate files for each
> > backend (currently gzio and zstdio), add module Makefiles, and figure
> > out how to make dumps dynamically use whatever compression is available.
>
> Actually I think this last bit is already in place.  We just need to
> have each compression method register with subr_compress on load (and
> unregister on unload, which I don't think subr_compress currently
> supports).  We also need to keep a running count of compression /
> decompression contexts to prevent unloading a method while it's in use.
>

So I guess go ahead with the kernel config changes (though I don't like
having to
do that, there's no easy alternative). I wonder what the best way to
document this so
that it doesn't get list in the future...

Warner

[-- Attachment #2 --]
<div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">On Thu, Mar 19, 2026 at 10:05 AM Dag-Erling Smørgrav &lt;<a href="mailto:des@freebsd.org">des@freebsd.org</a>&gt; wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Dag-Erling Smørgrav &lt;des@FreeBSD.org&gt; writes:<br>
&gt; Warner Losh &lt;<a href="mailto:imp@bsdimp.com" target="_blank">imp@bsdimp.com</a>&gt; writes:<br>
&gt; &gt; Any reason zstd can&#39;t be a module?<br>
&gt; I had the same thought but it&#39;s going to take a bit of work.  We&#39;ll have<br>
&gt; to split up sys/kern/subr_compress.c into separate files for each<br>
&gt; backend (currently gzio and zstdio), add module Makefiles, and figure<br>
&gt; out how to make dumps dynamically use whatever compression is available.<br>
<br>
Actually I think this last bit is already in place.  We just need to<br>
have each compression method register with subr_compress on load (and<br>
unregister on unload, which I don&#39;t think subr_compress currently<br>
supports).  We also need to keep a running count of compression /<br>
decompression contexts to prevent unloading a method while it&#39;s in use.<br></blockquote><div><br></div><div>So I guess go ahead with the kernel config changes (though I don&#39;t like having to</div><div>do that, there&#39;s no easy alternative). I wonder what the best way to document this so</div><div>that it doesn&#39;t get list in the future...</div><div><br></div><div>Warner </div></div></div>
home | help

Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfpF7GYLRPWvRSX1uKqKLZGPZB5Bg-Sd8t3xabVSHoCeOg>