[PATCH] fuse: show io_uring mount option in /proc/mounts

[PATCH] fuse: show io_uring mount option in /proc/mounts

[leo.lilong]

From: Long Li <leo.lilong@huawei.com>

When mounting a FUSE filesystem with io_uring option and using io_uring
for communication, this option was not shown in /proc/mounts or mount
command output. This made it difficult for users to verify whether
io_uring was being used for communication in their FUSE mounts.

Add io_uring to the list of mount options displayed in fuse_show_options()
when the fuse-over-io_uring feature is enabled and being used.

Signed-off-by: Long Li <leo.lilong@huawei.com>
---
 fs/fuse/inode.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index ecb869e895ab..a6a8cd84fdde 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -913,6 +913,8 @@ static int fuse_show_options(struct seq_file *m, struct dentry *root)
 			seq_puts(m, ",default_permissions");
 		if (fc->allow_other)
 			seq_puts(m, ",allow_other");
+		if (fc->io_uring)
+			seq_puts(m, ",io_uring");
 		if (fc->max_read != ~0)
 			seq_printf(m, ",max_read=%u", fc->max_read);
 		if (sb->s_bdev && sb->s_blocksize != FUSE_DEFAULT_BLKSIZE)
-- 
2.39.2

Re: [PATCH] fuse: show io_uring mount option in /proc/mounts

[Bernd Schubert]

On 7/9/25 04:02, leo.lilong@huaweicloud.com wrote:
> From: Long Li <leo.lilong@huawei.com>
> 
> When mounting a FUSE filesystem with io_uring option and using io_uring
> for communication, this option was not shown in /proc/mounts or mount
> command output. This made it difficult for users to verify whether
> io_uring was being used for communication in their FUSE mounts.
> 
> Add io_uring to the list of mount options displayed in fuse_show_options()
> when the fuse-over-io_uring feature is enabled and being used.
> 
> Signed-off-by: Long Li <leo.lilong@huawei.com>
> ---
>  fs/fuse/inode.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
> index ecb869e895ab..a6a8cd84fdde 100644
> --- a/fs/fuse/inode.c
> +++ b/fs/fuse/inode.c
> @@ -913,6 +913,8 @@ static int fuse_show_options(struct seq_file *m, struct dentry *root)
>  			seq_puts(m, ",default_permissions");
>  		if (fc->allow_other)
>  			seq_puts(m, ",allow_other");
> +		if (fc->io_uring)
> +			seq_puts(m, ",io_uring");
>  		if (fc->max_read != ~0)
>  			seq_printf(m, ",max_read=%u", fc->max_read);
>  		if (sb->s_bdev && sb->s_blocksize != FUSE_DEFAULT_BLKSIZE)

I agree with you that is impossible to see, but issue is that io_uring
is not a mount option. Maybe we should add a sysfs file?

Thanks,
Bernd

Re: [PATCH] fuse: show io_uring mount option in /proc/mounts

[Joanne Koong]

On Wed, Jul 9, 2025 at 4:24 AM Bernd Schubert <bernd@bsbernd.com> wrote:
>
> On 7/9/25 04:02, leo.lilong@huaweicloud.com wrote:
> > From: Long Li <leo.lilong@huawei.com>
> >
> > When mounting a FUSE filesystem with io_uring option and using io_uring
> > for communication, this option was not shown in /proc/mounts or mount
> > command output. This made it difficult for users to verify whether
> > io_uring was being used for communication in their FUSE mounts.
> >
> > Add io_uring to the list of mount options displayed in fuse_show_options()
> > when the fuse-over-io_uring feature is enabled and being used.
> >
> > Signed-off-by: Long Li <leo.lilong@huawei.com>
> > ---
> >  fs/fuse/inode.c | 2 ++
> >  1 file changed, 2 insertions(+)
> >
> > diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
> > index ecb869e895ab..a6a8cd84fdde 100644
> > --- a/fs/fuse/inode.c
> > +++ b/fs/fuse/inode.c
> > @@ -913,6 +913,8 @@ static int fuse_show_options(struct seq_file *m, struct dentry *root)
> >                       seq_puts(m, ",default_permissions");
> >               if (fc->allow_other)
> >                       seq_puts(m, ",allow_other");
> > +             if (fc->io_uring)
> > +                     seq_puts(m, ",io_uring");
> >               if (fc->max_read != ~0)
> >                       seq_printf(m, ",max_read=%u", fc->max_read);
> >               if (sb->s_bdev && sb->s_blocksize != FUSE_DEFAULT_BLKSIZE)
>
> I agree with you that is impossible to see, but issue is that io_uring
> is not a mount option. Maybe we should add a sysfs file?
>

Libfuse knows so what about just relaying that information from libfuse?


Thanks,
Joanne

> Thanks,
> Bernd
>

Re: [PATCH] fuse: show io_uring mount option in /proc/mounts

[Christian Brauner]

On Wed, Jul 09, 2025 at 10:55:31AM -0700, Joanne Koong wrote:
> On Wed, Jul 9, 2025 at 4:24 AM Bernd Schubert <bernd@bsbernd.com> wrote:
> >
> > On 7/9/25 04:02, leo.lilong@huaweicloud.com wrote:
> > > From: Long Li <leo.lilong@huawei.com>
> > >
> > > When mounting a FUSE filesystem with io_uring option and using io_uring
> > > for communication, this option was not shown in /proc/mounts or mount
> > > command output. This made it difficult for users to verify whether
> > > io_uring was being used for communication in their FUSE mounts.
> > >
> > > Add io_uring to the list of mount options displayed in fuse_show_options()
> > > when the fuse-over-io_uring feature is enabled and being used.
> > >
> > > Signed-off-by: Long Li <leo.lilong@huawei.com>
> > > ---
> > >  fs/fuse/inode.c | 2 ++
> > >  1 file changed, 2 insertions(+)
> > >
> > > diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
> > > index ecb869e895ab..a6a8cd84fdde 100644
> > > --- a/fs/fuse/inode.c
> > > +++ b/fs/fuse/inode.c
> > > @@ -913,6 +913,8 @@ static int fuse_show_options(struct seq_file *m, struct dentry *root)
> > >                       seq_puts(m, ",default_permissions");
> > >               if (fc->allow_other)
> > >                       seq_puts(m, ",allow_other");
> > > +             if (fc->io_uring)
> > > +                     seq_puts(m, ",io_uring");
> > >               if (fc->max_read != ~0)
> > >                       seq_printf(m, ",max_read=%u", fc->max_read);
> > >               if (sb->s_bdev && sb->s_blocksize != FUSE_DEFAULT_BLKSIZE)
> >
> > I agree with you that is impossible to see, but issue is that io_uring
> > is not a mount option. Maybe we should add a sysfs file?
> >
> 
> Libfuse knows so what about just relaying that information from libfuse?

Yeah, if it's not a mount option we shouldn't actually display it in
mountinfo...

Re: [PATCH] fuse: show io_uring mount option in /proc/mounts

[Long Li]

On Wed, Jul 09, 2025 at 01:00:05PM +0200, Bernd Schubert wrote:
> 
> 
> On 7/9/25 04:02, leo.lilong@huaweicloud.com wrote:
> > From: Long Li <leo.lilong@huawei.com>
> > 
> > When mounting a FUSE filesystem with io_uring option and using io_uring
> > for communication, this option was not shown in /proc/mounts or mount
> > command output. This made it difficult for users to verify whether
> > io_uring was being used for communication in their FUSE mounts.
> > 
> > Add io_uring to the list of mount options displayed in fuse_show_options()
> > when the fuse-over-io_uring feature is enabled and being used.
> > 
> > Signed-off-by: Long Li <leo.lilong@huawei.com>
> > ---
> >  fs/fuse/inode.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
> > index ecb869e895ab..a6a8cd84fdde 100644
> > --- a/fs/fuse/inode.c
> > +++ b/fs/fuse/inode.c
> > @@ -913,6 +913,8 @@ static int fuse_show_options(struct seq_file *m, struct dentry *root)
> >  			seq_puts(m, ",default_permissions");
> >  		if (fc->allow_other)
> >  			seq_puts(m, ",allow_other");
> > +		if (fc->io_uring)
> > +			seq_puts(m, ",io_uring");
> >  		if (fc->max_read != ~0)
> >  			seq_printf(m, ",max_read=%u", fc->max_read);
> >  		if (sb->s_bdev && sb->s_blocksize != FUSE_DEFAULT_BLKSIZE)
> 
> I agree with you that is impossible to see, but issue is that io_uring
> is not a mount option. Maybe we should add a sysfs file?
> 

Yes, this is not a mount option. Thank you for your suggestion.

Thanks,
Long Li