[PATCH 5.4 058/194] mtd: Fix gluebi NULL pointer dereference caused by ftl notifier

[Martin Kepplinger-Novakovic]

Am Montag, dem 22.01.2024 um 15:56 -0800 schrieb Greg Kroah-Hartman:
> 5.4-stable review patch.  If anyone has any objections, please let me
> know.
> 
> ------------------
> 
> From: ZhaoLong Wang <wangzhaolong1 at huawei.com>
> 
> [ Upstream commit a43bdc376deab5fff1ceb93dca55bcab8dbdc1d6 ]
> 
> If both ftl.ko and gluebi.ko are loaded, the notifier of ftl
> triggers NULL pointer dereference when trying to access
> ‘gluebi->desc’ in gluebi_read().
> 
> ubi_gluebi_init
>   ubi_register_volume_notifier
>     ubi_enumerate_volumes
>       ubi_notify_all
>         gluebi_notify    nb->notifier_call()
>           gluebi_create
>             mtd_device_register
>               mtd_device_parse_register
>                 add_mtd_device
>                   blktrans_notify_add   not->add()
>                     ftl_add_mtd         tr->add_mtd()
>                       scan_header
>                         mtd_read
>                           mtd_read_oob
>                             mtd_read_oob_std
>                               gluebi_read   mtd->read()
>                                 gluebi->desc - NULL
> 
> Detailed reproduction information available at the Link [1],
> 
> In the normal case, obtain gluebi->desc in the gluebi_get_device(),
> and access gluebi->desc in the gluebi_read(). However,
> gluebi_get_device() is not executed in advance in the
> ftl_add_mtd() process, which leads to NULL pointer dereference.
> 
> The solution for the gluebi module is to run jffs2 on the UBI
> volume without considering working with ftl or mtdblock [2].
> Therefore, this problem can be avoided by preventing gluebi from
> creating the mtdblock device after creating mtd partition of the
> type MTD_UBIVOLUME.
> 
> Fixes: 2ba3d76a1e29 ("UBI: make gluebi a separate module")
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=217992 [1]
> Link:
> https://lore.kernel.org/lkml/441107100.23734.1697904580252.JavaMail.zimbra@nod.at/
>  [2]
> Signed-off-by: ZhaoLong Wang <wangzhaolong1 at huawei.com>
> Reviewed-by: Zhihao Cheng <chengzhihao1 at huawei.com>
> Acked-by: Richard Weinberger <richard at nod.at>
> Signed-off-by: Miquel Raynal <miquel.raynal at bootlin.com>
> Link:
> https://lore.kernel.org/linux-mtd/20231220024619.2138625-1-wangzhaolong1@huawei.com
> Signed-off-by: Sasha Levin <sashal at kernel.org>
> ---
>  drivers/mtd/mtd_blkdevs.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/mtd/mtd_blkdevs.c b/drivers/mtd/mtd_blkdevs.c
> index 0c05f77f9b21..dd0d0bf5f57f 100644
> --- a/drivers/mtd/mtd_blkdevs.c
> +++ b/drivers/mtd/mtd_blkdevs.c
> @@ -533,7 +533,7 @@ static void blktrans_notify_add(struct mtd_info
> *mtd)
>  {
>         struct mtd_blktrans_ops *tr;
>  
> -       if (mtd->type == MTD_ABSENT)
> +       if (mtd->type == MTD_ABSENT || mtd->type == MTD_UBIVOLUME)
>                 return;
>  
>         list_for_each_entry(tr, &blktrans_majors, list)
> @@ -576,7 +576,7 @@ int register_mtd_blktrans(struct mtd_blktrans_ops
> *tr)
>         list_add(&tr->list, &blktrans_majors);
>  
>         mtd_for_each_device(mtd)
> -               if (mtd->type != MTD_ABSENT)
> +               if (mtd->type != MTD_ABSENT && mtd->type !=
> MTD_UBIVOLUME)
>                         tr->add_mtd(tr, mtd);
>  
>         mutex_unlock(&mtd_table_mutex);

Hi Greg, hi patch-developers,

wait a second. this already went into v5.4.268 but still: Doesn't this
break userspace?

According to
https://lore.kernel.org/lkml/441107100.23734.1697904580252.JavaMail.zimbra@nod.at/
where this solution seems to come from, the behaviour changes: "no
mtdblock (hence, also no FTLs) on top of gluebi."

I fell accross this because of an out-of-tree module that does
sys_mount() an mtdblock, so I won't complain about my code specifically
:) But doesn't it break mounting, say, jffs2 inside an ubi via
mtdblock? If so, is this really something that you want to see
backported to old kernels?

Or differently put: Has this patch been picked up for old stable
kernels by scripts or by a human?

I just want to make sure, and who knows, it might help others too, who
would just do a (possibly dangerous?) revert in their trees.

thanks!
                          martin