MIME-Version: 1.0
References: <a13cd3b5-cc41-bf2f-c8ac-e031ad0d5dd7@leemhuis.info>
 <CAPhsuW7ZWthh0PZt71hQh1_51C0yMSpOqWYJKc_+VzzTmW_r5A@mail.gmail.com>
 <167805126796.8008.3635368722810568057@noble.neil.brown.name>
 <CAPhsuW64R2ze1AYZhEmQcGf0cKBjjX=4EZZowD+=Cr=VPg1QYg@mail.gmail.com> <167814210165.8008.4497120634222267375@noble.neil.brown.name>
In-Reply-To: <167814210165.8008.4497120634222267375@noble.neil.brown.name>
From:   Song Liu <song@kernel.org>
Date:   Mon, 13 Mar 2023 13:35:20 -0700
Message-ID: <CAPhsuW4oAsp=8dBcf_mmpp-bG_+wTHAZzvtRSn3x1QU0N3irXA@mail.gmail.com>
Subject: Re: [regression] Bug 217074 - upgrading to kernel 6.1.12 from 5.15.x
 can no longer assemble software raid0
To:     NeilBrown <neilb@suse.de>
Cc:     Linux regressions mailing list <regressions@lists.linux.dev>,
        Jes.Sorensen@gmail.com, linux-raid <linux-raid@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Nikolay Kichukov <hijacker@oldum.net>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Mar 6, 2023 at 2:35=E2=80=AFPM NeilBrown <neilb@suse.de> wrote:
>
> On Tue, 07 Mar 2023, Song Liu wrote:
> > On Sun, Mar 5, 2023 at 1:21=E2=80=AFPM NeilBrown <neilb@suse.de> wrote:
> > >
> > > On Sat, 04 Mar 2023, Song Liu wrote:
> > > > + Jes.
> > > >
> > > > It appeared to me that we can assemble the array if we have any of =
the
> > > > following:
> > > > 1. Enable CONFIG_BLOCK_LEGACY_AUTOLOAD;
> > > > 2. Have a valid /etc/mdadm.conf;
> > > > 3. Update mdadm to handle this case. (I tried some ugly hacks, whic=
h worked but
> > > >     weren't clean).
> > > >
> > > > Since we eventually would like to get rid of CONFIG_BLOCK_LEGACY_AU=
TOLOAD, I
> > > > think we need mdadm to handle this properly. But the logistics migh=
t
> > > > be complicated, as
> > > > mdadm are shipped separately.
> > > >
> > > > Jes, what do you think about this? AFAICT, we need to update the lo=
gic in
> > > > mdopen.c:create_mddev().
> > >
> > > mdadm already handles this, but only if
> > >    CREATE names=3Dyes
> > > is present in /etc/mdadm.conf
> > >
> > > Maybe we should flip the default for the next mdadm release, and patc=
h
> > > the kernel (with a stable backport) to select BLOCK_LEGACY_AUTOLOAD i=
f
> > > BLK_DEV_MD=3Dm
> > > Then revert that - say - 6 months after the new mdadm is released.
> >
> > I like this idea. I guess we also need to select BLOCK_LEGACY_AUTOLOAD
> > if BLK_DEV_MD=3Dy?
>
> Yes of course - sorry.
>
> Something like the following.
>
> Thanks,
> NeilBrown

Thanks Neil!

I applied this to md-fixes.

Song

>
>
> From: NeilBrown <neilb@suse.de>
> Subject: md: select BLOCK_LEGACY_AUTOLOAD
>
> When BLOCK_LEGACY_AUTOLOAD is not enable, mdadm is not able to
> activate new arrays unless "CREATE names=3Dyes" appears in
> mdadm.conf
>
> As this is a regression we need to always enable BLOCK_LEGACY_AUTOLOAD
> for when MD is selected - at least until mdadm is updated and the
> updates widely available.
>
> Fixes: fbdee71bb5d8 ("block: deprecate autoloading based on dev_t")
> Signed-off-by: NeilBrown <neilb@suse.de>
>
> diff --git a/drivers/md/Kconfig b/drivers/md/Kconfig
> index 998a5cfdbc4e..662d219c39bf 100644
> --- a/drivers/md/Kconfig
> +++ b/drivers/md/Kconfig
> @@ -16,6 +16,10 @@ if MD
>
>  config BLK_DEV_MD
>         tristate "RAID support"
>         select BLOCK_HOLDER_DEPRECATED if SYSFS
> +       # BLOCK_LEGACY_AUTOLOAD requirement should be removed
> +       # after relevant mdadm enhancements - to make "names=3Dyes"
> +       # the default - are widely available.
> +       select BLOCK_LEGACY_AUTOLOAD
>         help
>           This driver lets you combine several hard disk partitions into =
one
>           logical block device. This can be used to simply append one
>
>