Re: [PATCH] exportfs: handle CONFIG_EXPORTFS=m also

From: Amir Goldstein
Date: Thu Oct 26 2023 - 15:46:22 EST


On Thu, Oct 26, 2023 at 10:28 PM Randy Dunlap <rdunlap@xxxxxxxxxxxxx> wrote:
>
> When CONFIG_EXPORTFS=m, there are multiple build errors due to
> the header <linux/exportfs.h> not handling the =m setting correctly.
> Change the header file to check for CONFIG_EXPORTFS enabled at all
> instead of just set =y.
>
> Fixes: dfaf653dc415 ("exportfs: make ->encode_fh() a mandatory method for NFS export")
> Signed-off-by: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
> Cc: Chuck Lever <chuck.lever@xxxxxxxxxx>
> Cc: Jeff Layton <jlayton@xxxxxxxxxx>
> Cc: linux-nfs@xxxxxxxxxxxxxxx
> Cc: Amir Goldstein <amir73il@xxxxxxxxx>
> Cc: Christian Brauner <brauner@xxxxxxxxxx>
> Cc: Alexander Viro <viro@xxxxxxxxxxxxxxxxxx>
> Cc: linux-fsdevel@xxxxxxxxxxxxxxx
>
> ---
> include/linux/exportfs.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff -- a/include/linux/exportfs.h b/include/linux/exportfs.h
> --- a/include/linux/exportfs.h
> +++ b/include/linux/exportfs.h
> @@ -314,7 +314,7 @@ extern struct dentry *exportfs_decode_fh
> /*
> * Generic helpers for filesystems.
> */
> -#ifdef CONFIG_EXPORTFS
> +#if IS_ENABLED(CONFIG_EXPORTFS)
> int generic_encode_ino32_fh(struct inode *inode, __u32 *fh, int *max_len,
> struct inode *parent);
> #else

Thanks for the fix, but Arnd reported that this fix could cause a link
issue in some configuration - I did not verify.

I would much rather turn EXPORTFS into a bool config
and avoid the unneeded build test matrix.

See comparison to the amount of code of EXPORTFS
to BUFFER_HEAD and FS_IOMAP code which are bool:

~/src/linux$ wc -l fs/exportfs/*.c
636 fs/exportfs/expfs.c

~/src/linux$ wc -l fs/buffer.c fs/mpage.c
3164 fs/buffer.c
685 fs/mpage.c
3849 total

~/src/linux$ wc -l fs/iomap/*.c
2002 fs/iomap/buffered-io.c
754 fs/iomap/direct-io.c
124 fs/iomap/fiemap.c
97 fs/iomap/iter.c
104 fs/iomap/seek.c
195 fs/iomap/swapfile.c
13 fs/iomap/trace.c
3289 total

Thanks,
Amir.