MIME-Version: 1.0
References: <20230221135506.296074-1-yukuai1@huaweicloud.com>
In-Reply-To: <20230221135506.296074-1-yukuai1@huaweicloud.com>
From:   Xiao Ni <xni@redhat.com>
Date:   Wed, 22 Feb 2023 08:39:03 +0800
Message-ID: <CALTww2-eUz7__wmu6Cbxs0yV3OGD23WggRUbn6ry_z+W9MApKA@mail.gmail.com>
Subject: Re: [PATCH] md: fix null-ptr-deference in md_free_disk()
To:     Yu Kuai <yukuai1@huaweicloud.com>
Cc:     song@kernel.org, linux-raid@vger.kernel.org,
        linux-kernel@vger.kernel.org, yukuai3@huawei.com,
        yi.zhang@huawei.com, yangerkun@huawei.com,
        Heinz Mauelshagen <heinzm@redhat.com>,
        Nigel Croxon <ncroxon@redhat.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi Yu Kuai

Thanks for reporting this problem.

For creating raid device, md raid calls do_md_run->md_run, dm raid
calls md_run. We alloc those memory in md_run.
For stopping raid device, md raid calls do_md_stop->__md_stop, dm raid
calls md_stop->__md_stop.

So how about this patch:
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 927a43db5dfb..f5480778e2f7 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c

On Tue, Feb 21, 2023 at 9:33 PM Yu Kuai <yukuai1@huaweicloud.com> wrote:
>
> From: Yu Kuai <yukuai3@huawei.com>
>
> If md_run() failed after 'acitive_io' is initialized, then
> percpu_ref_exit() is called in error path, however, later md_free_disk()
> will call percpu_ref_exit() again, which lead to following
> null-ptr-deference:
>
> BUG: kernel NULL pointer dereference, address: 0000000000000038
> Oops: 0000 [#1] PREEMPT SMP
> CPU: 41 PID: 585 Comm: kworker/41:1 Not tainted 6.2.0-rc8-next-20230220 #1452
> Workqueue: md_misc mddev_delayed_delete
> RIP: 0010:free_percpu+0x110/0x630
> Call Trace:
>  <TASK>
>  __percpu_ref_exit+0x44/0x70
>  percpu_ref_exit+0x16/0x90
>  md_free_disk+0x2f/0x80
>  disk_release+0x101/0x180
>  device_release+0x84/0x110
>  kobject_put+0x12a/0x380
>  kobject_put+0x160/0x380
>  mddev_delayed_delete+0x19/0x30
>  process_one_work+0x269/0x680
>  worker_thread+0x266/0x640
>  kthread+0x151/0x1b0
>  ret_from_fork+0x1f/0x30
>
> Since freeing mddev will exit 'active_io' unconditionally, fix the
> problem by removing exiting 'active_io' from error path, this way
> it will be delayed to free mddev.
>
> Fixes: 72adae23a72c ("md: Change active_io to percpu")
> Signed-off-by: Yu Kuai <yukuai3@huawei.com>
> ---
>  drivers/md/md.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index 927a43db5dfb..77124679b3fd 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -5851,7 +5851,7 @@ int md_run(struct mddev *mddev)
>         if (!bioset_initialized(&mddev->bio_set)) {
>                 err = bioset_init(&mddev->bio_set, BIO_POOL_SIZE, 0, BIOSET_NEED_BVECS);
>                 if (err)
> -                       goto exit_active_io;
> +                       return err;
>         }
>         if (!bioset_initialized(&mddev->sync_set)) {
>                 err = bioset_init(&mddev->sync_set, BIO_POOL_SIZE, 0, BIOSET_NEED_BVECS);
> @@ -6039,8 +6039,6 @@ int md_run(struct mddev *mddev)
>         bioset_exit(&mddev->sync_set);
>  exit_bio_set:
>         bioset_exit(&mddev->bio_set);
> -exit_active_io:
> -       percpu_ref_exit(&mddev->active_io);
>         return err;
>  }
>  EXPORT_SYMBOL_GPL(md_run);
> --
> 2.31.1
>