[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CADYN=9JUDjj=KS=xTWXbVkkDPsCi2+GjGfWqAL8naOEHpu4nHQ@mail.gmail.com>
Date: Mon, 23 Oct 2023 15:48:14 +0200
From: Anders Roxell <anders.roxell@...aro.org>
To: Ian Kent <raven@...maw.net>
Cc: Arnd Bergmann <arnd@...db.de>,
Dan Carpenter <dan.carpenter@...aro.org>,
Christian Brauner <brauner@...nel.org>,
Naresh Kamboju <naresh.kamboju@...aro.org>,
"Bill O'Donnell" <bodonnel@...hat.com>,
Kernel Mailing List <linux-kernel@...r.kernel.org>,
autofs mailing list <autofs@...r.kernel.org>,
linux-fsdevel <linux-fsdevel@...r.kernel.org>,
lkft-triage@...ts.linaro.org,
Stephen Rothwell <sfr@...b.auug.org.au>,
Linux Kernel Functional Testing <lkft@...aro.org>
Subject: Re: [PATCH] autofs: fix add autofs_parse_fd()
On Mon, 23 Oct 2023 at 11:34, Ian Kent <raven@...maw.net> wrote:
>
> We are seeing systemd hang on its autofs direct mount at
> /proc/sys/fs/binfmt_misc.
>
> Historically this was due to a mismatch in the communication structure
> size between a 64 bit kernel and a 32 bit user space and was fixed by
> making the pipe communication record oriented.
>
> During autofs v5 development I decided to stay with the existing usage
> instead of changing to a packed structure for autofs <=> user space
> communications which turned out to be a mistake on my part.
>
> Problems arose and they were fixed by allowing for the 64 bit to 32
> bit size difference in the automount(8) code.
>
> Along the way systemd started to use autofs and eventually encountered
> this problem too. systemd refused to compensate for the length
> difference insisting it be fixed in the kernel. Fortunately Linus
> implemented the packetized pipe which resolved the problem in a
> straight forward and simple way.
>
> In the autofs mount api conversion series I inadvertatly dropped the
> packet pipe flag settings when adding the autofs_parse_fd() function.
> This patch fixes that omission.
>
> Fixes: 546694b8f658 ("autofs: add autofs_parse_fd()")
> Signed-off-by: Ian Kent <raven@...maw.net>
> Cc: Bill O'Donnell <bodonnel@...hat.com>
> Cc: Christian Brauner <brauner@...nel.org>
> Cc: Arnd Bergmann <arnd@...db.de>
> Cc: Dan Carpenter <dan.carpenter@...aro.org>
> Cc: Anders Roxell <anders.roxell@...aro.org>
> Cc: Naresh Kamboju <naresh.kamboju@...aro.org>
> Cc: Stephen Rothwell <sfr@...b.auug.org.au>
> Reported-by: Linux Kernel Functional Testing <lkft@...aro.org>
> Reported-by: Anders Roxell <anders.roxell@...aro.org>
Thank you Ian for finding the issue so quickly.
Tested-by: Anders Roxell <anders.roxell@...aro.org>
I built todays next-20231023 with this patch and the kernel booted up
fine in qemu.
Cheers,
Anders
> ---
> fs/autofs/autofs_i.h | 13 +++++++++----
> fs/autofs/inode.c | 2 ++
> 2 files changed, 11 insertions(+), 4 deletions(-)
>
> diff --git a/fs/autofs/autofs_i.h b/fs/autofs/autofs_i.h
> index 244f18cdf23c..8c1d587b3eef 100644
> --- a/fs/autofs/autofs_i.h
> +++ b/fs/autofs/autofs_i.h
> @@ -221,15 +221,20 @@ static inline int autofs_check_pipe(struct file *pipe)
> return 0;
> }
>
> -static inline int autofs_prepare_pipe(struct file *pipe)
> +static inline void autofs_set_packet_pipe_flags(struct file *pipe)
> {
> - int ret = autofs_check_pipe(pipe);
> - if (ret < 0)
> - return ret;
> /* We want a packet pipe */
> pipe->f_flags |= O_DIRECT;
> /* We don't expect -EAGAIN */
> pipe->f_flags &= ~O_NONBLOCK;
> +}
> +
> +static inline int autofs_prepare_pipe(struct file *pipe)
> +{
> + int ret = autofs_check_pipe(pipe);
> + if (ret < 0)
> + return ret;
> + autofs_set_packet_pipe_flags(pipe);
> return 0;
> }
>
> diff --git a/fs/autofs/inode.c b/fs/autofs/inode.c
> index 6d2e01c9057d..a3d62acc293a 100644
> --- a/fs/autofs/inode.c
> +++ b/fs/autofs/inode.c
> @@ -177,6 +177,8 @@ static int autofs_parse_fd(struct fs_context *fc, struct autofs_sb_info *sbi,
> return -EBADF;
> }
>
> + autofs_set_packet_pipe_flags(pipe);
> +
> if (sbi->pipe)
> fput(sbi->pipe);
>
> --
> 2.41.0
>
Powered by blists - more mailing lists