Currently when re-exporting a NFS share the NFS cross mount feature does
not work [0].
This patch series outlines an approach to address the problem.
Crossing mounts does not work for two reasons:
1. As soon the NFS client (on the re-exporting server) sees a different
filesystem id, it installs an automount. That way the other filesystem
will be mounted automatically when someone enters the directory.
But the cross mount logic of KNFS does not know about automount.
This patch series addresses the problem and teach both KNFSD
and the exportfs logic of NFS to deal with automount.
2. When KNFSD detects crossing of a mount point, it asks rpc.mountd to install
a new export for the target mount point. Beside of authentication rpc.mountd
also has to find a filesystem id for the new export. Is the to be exported
filesystem a NFS share, rpc.mountd cannot derive a filesystem id from it and
refuses to export. In the logs you'll see errors such as:
mountd: Cannot export /srv/nfs/vol0, possibly unsupported filesystem or fsid= required
To deal with that I've changed rpc.mountd to use generate and store fsids [1].
Since the kernel side of my changes did change for a long time I decided to
try upstreaming it first.
A 3rd iteration of my rpc.mountd will happen soon.
[0] https://marc.info/?l=linux-nfs&m=161653016627277&w=2
[1] https://lore.kernel.org/linux-nfs/[email protected]/
Changes since v1:
https://lore.kernel.org/linux-nfs/[email protected]/
- Use LOOKUP_AUTOMOUNT only when NFSEXP_CROSSMOUNT is set (Jeff Layton)
Richard Weinberger (3):
NFSD: Teach nfsd_mountpoint() auto mounts
fs: namei: Allow follow_down() to uncover auto mounts
NFS: nfs_encode_fh: Remove S_AUTOMOUNT check
fs/namei.c | 6 +++---
fs/nfs/export.c | 2 +-
fs/nfsd/vfs.c | 8 ++++++--
include/linux/namei.h | 2 +-
4 files changed, 11 insertions(+), 7 deletions(-)
--
2.26.2
This function is only used by NFSD to cross mount points.
If a mount point is of type auto mount, follow_down() will
not uncover it. Add LOOKUP_AUTOMOUNT to the lookup flags
to have ->d_automount() called when NFSD walks down the
mount tree.
Signed-off-by: Richard Weinberger <[email protected]>
---
fs/namei.c | 6 +++---
fs/nfsd/vfs.c | 6 +++++-
include/linux/namei.h | 2 +-
3 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/fs/namei.c b/fs/namei.c
index 578c2110df02..a6bb6863bf0c 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -1458,11 +1458,11 @@ EXPORT_SYMBOL(follow_down_one);
* point, the filesystem owning that dentry may be queried as to whether the
* caller is permitted to proceed or not.
*/
-int follow_down(struct path *path)
+int follow_down(struct path *path, unsigned int flags)
{
struct vfsmount *mnt = path->mnt;
bool jumped;
- int ret = traverse_mounts(path, &jumped, NULL, 0);
+ int ret = traverse_mounts(path, &jumped, NULL, flags);
if (path->mnt != mnt)
mntput(mnt);
@@ -2864,7 +2864,7 @@ int path_pts(struct path *path)
path->dentry = child;
dput(parent);
- follow_down(path);
+ follow_down(path, 0);
return 0;
}
#endif
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 157f0df0e93a..ced04fc2b947 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -63,9 +63,13 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
struct dentry *dentry = *dpp;
struct path path = {.mnt = mntget(exp->ex_path.mnt),
.dentry = dget(dentry)};
+ unsigned int follow_flags = 0;
int err = 0;
- err = follow_down(&path);
+ if (exp->ex_flags & NFSEXP_CROSSMOUNT)
+ follow_flags = LOOKUP_AUTOMOUNT;
+
+ err = follow_down(&path, follow_flags);
if (err < 0)
goto out;
if (path.mnt == exp->ex_path.mnt && path.dentry == dentry &&
diff --git a/include/linux/namei.h b/include/linux/namei.h
index 00fee52df842..6f96db73a70a 100644
--- a/include/linux/namei.h
+++ b/include/linux/namei.h
@@ -77,7 +77,7 @@ struct dentry *lookup_one_positive_unlocked(struct user_namespace *mnt_userns,
struct dentry *base, int len);
extern int follow_down_one(struct path *);
-extern int follow_down(struct path *);
+extern int follow_down(struct path *, unsigned int flags);
extern int follow_up(struct path *);
extern struct dentry *lock_rename(struct dentry *, struct dentry *);
--
2.26.2
> On Dec 7, 2022, at 3:43 AM, Richard Weinberger <[email protected]> wrote:
>
> Currently when re-exporting a NFS share the NFS cross mount feature does
> not work [0].
> This patch series outlines an approach to address the problem.
>
> Crossing mounts does not work for two reasons:
>
> 1. As soon the NFS client (on the re-exporting server) sees a different
> filesystem id, it installs an automount. That way the other filesystem
> will be mounted automatically when someone enters the directory.
> But the cross mount logic of KNFS does not know about automount.
> This patch series addresses the problem and teach both KNFSD
> and the exportfs logic of NFS to deal with automount.
>
> 2. When KNFSD detects crossing of a mount point, it asks rpc.mountd to install
> a new export for the target mount point. Beside of authentication rpc.mountd
> also has to find a filesystem id for the new export. Is the to be exported
> filesystem a NFS share, rpc.mountd cannot derive a filesystem id from it and
> refuses to export. In the logs you'll see errors such as:
>
> mountd: Cannot export /srv/nfs/vol0, possibly unsupported filesystem or fsid= required
>
> To deal with that I've changed rpc.mountd to use generate and store fsids [1].
> Since the kernel side of my changes did change for a long time I decided to
> try upstreaming it first.
> A 3rd iteration of my rpc.mountd will happen soon.
>
> [0] https://marc.info/?l=linux-nfs&m=161653016627277&w=2
> [1] https://lore.kernel.org/linux-nfs/[email protected]/
>
> Changes since v1:
> https://lore.kernel.org/linux-nfs/[email protected]/
>
> - Use LOOKUP_AUTOMOUNT only when NFSEXP_CROSSMOUNT is set (Jeff Layton)
>
> Richard Weinberger (3):
> NFSD: Teach nfsd_mountpoint() auto mounts
> fs: namei: Allow follow_down() to uncover auto mounts
> NFS: nfs_encode_fh: Remove S_AUTOMOUNT check
>
> fs/namei.c | 6 +++---
> fs/nfs/export.c | 2 +-
> fs/nfsd/vfs.c | 8 ++++++--
> include/linux/namei.h | 2 +-
> 4 files changed, 11 insertions(+), 7 deletions(-)
>
> --
> 2.26.2
>
This series is a bit late for inclusion in v6.2. The next opportunity
will be v6.3 in a couple of months. I prefer to have a "final" version
of patches by -rc5.
I'm waiting for review comments on v2 of this series.
--
Chuck Lever
----- Ursprüngliche Mail -----
> Von: "chuck lever" <[email protected]>
>> Richard Weinberger (3):
>> NFSD: Teach nfsd_mountpoint() auto mounts
>> fs: namei: Allow follow_down() to uncover auto mounts
>> NFS: nfs_encode_fh: Remove S_AUTOMOUNT check
>>
>> fs/namei.c | 6 +++---
>> fs/nfs/export.c | 2 +-
>> fs/nfsd/vfs.c | 8 ++++++--
>> include/linux/namei.h | 2 +-
>> 4 files changed, 11 insertions(+), 7 deletions(-)
>>
>> --
>> 2.26.2
>>
>
> This series is a bit late for inclusion in v6.2. The next opportunity
> will be v6.3 in a couple of months. I prefer to have a "final" version
> of patches by -rc5.
>
> I'm waiting for review comments on v2 of this series.
Ok! Do you want me to resend the series in any case by v6.2-rc5 or only
if new comments arise?
Thanks,
//richard
> On Dec 10, 2022, at 4:52 PM, Richard Weinberger <[email protected]> wrote:
>
> ----- Ursprüngliche Mail -----
>> Von: "chuck lever" <[email protected]>
>>> Richard Weinberger (3):
>>> NFSD: Teach nfsd_mountpoint() auto mounts
>>> fs: namei: Allow follow_down() to uncover auto mounts
>>> NFS: nfs_encode_fh: Remove S_AUTOMOUNT check
>>>
>>> fs/namei.c | 6 +++---
>>> fs/nfs/export.c | 2 +-
>>> fs/nfsd/vfs.c | 8 ++++++--
>>> include/linux/namei.h | 2 +-
>>> 4 files changed, 11 insertions(+), 7 deletions(-)
>>>
>>> --
>>> 2.26.2
>>>
>>
>> This series is a bit late for inclusion in v6.2. The next opportunity
>> will be v6.3 in a couple of months. I prefer to have a "final" version
>> of patches by -rc5.
>>
>> I'm waiting for review comments on v2 of this series.
>
> Ok! Do you want me to resend the series in any case by v6.2-rc5 or only
> if new comments arise?
If v2 garners no new comments, then send a reminder with a URL to this
thread on lore.kernel.org. I will pull the series from there.
--
Chuck Lever
On Wed, 2022-12-07 at 09:43 +0100, Richard Weinberger wrote:
> Currently when re-exporting a NFS share the NFS cross mount feature does
> not work [0].
> This patch series outlines an approach to address the problem.
>
> Crossing mounts does not work for two reasons:
>
> 1. As soon the NFS client (on the re-exporting server) sees a different
> filesystem id, it installs an automount. That way the other filesystem
> will be mounted automatically when someone enters the directory.
> But the cross mount logic of KNFS does not know about automount.
> This patch series addresses the problem and teach both KNFSD
> and the exportfs logic of NFS to deal with automount.
>
> 2. When KNFSD detects crossing of a mount point, it asks rpc.mountd to install
> a new export for the target mount point. Beside of authentication rpc.mountd
> also has to find a filesystem id for the new export. Is the to be exported
> filesystem a NFS share, rpc.mountd cannot derive a filesystem id from it and
> refuses to export. In the logs you'll see errors such as:
>
> mountd: Cannot export /srv/nfs/vol0, possibly unsupported filesystem or fsid= required
>
> To deal with that I've changed rpc.mountd to use generate and store fsids [1].
> Since the kernel side of my changes did change for a long time I decided to
> try upstreaming it first.
> A 3rd iteration of my rpc.mountd will happen soon.
>
> [0] https://marc.info/?l=linux-nfs&m=161653016627277&w=2
> [1] https://lore.kernel.org/linux-nfs/[email protected]/
>
> Changes since v1:
> https://lore.kernel.org/linux-nfs/[email protected]/
>
> - Use LOOKUP_AUTOMOUNT only when NFSEXP_CROSSMOUNT is set (Jeff Layton)
>
> Richard Weinberger (3):
> NFSD: Teach nfsd_mountpoint() auto mounts
> fs: namei: Allow follow_down() to uncover auto mounts
> NFS: nfs_encode_fh: Remove S_AUTOMOUNT check
>
> fs/namei.c | 6 +++---
> fs/nfs/export.c | 2 +-
> fs/nfsd/vfs.c | 8 ++++++--
> include/linux/namei.h | 2 +-
> 4 files changed, 11 insertions(+), 7 deletions(-)
>
This set looks reasonable to me.
Reviewed-by: Jeff Layton <[email protected]>
On 13/12/22 01:06, Jeff Layton wrote:
> On Wed, 2022-12-07 at 09:43 +0100, Richard Weinberger wrote:
>> Currently when re-exporting a NFS share the NFS cross mount feature does
>> not work [0].
>> This patch series outlines an approach to address the problem.
>>
>> Crossing mounts does not work for two reasons:
>>
>> 1. As soon the NFS client (on the re-exporting server) sees a different
>> filesystem id, it installs an automount. That way the other filesystem
>> will be mounted automatically when someone enters the directory.
>> But the cross mount logic of KNFS does not know about automount.
>> This patch series addresses the problem and teach both KNFSD
>> and the exportfs logic of NFS to deal with automount.
>>
>> 2. When KNFSD detects crossing of a mount point, it asks rpc.mountd to install
>> a new export for the target mount point. Beside of authentication rpc.mountd
>> also has to find a filesystem id for the new export. Is the to be exported
>> filesystem a NFS share, rpc.mountd cannot derive a filesystem id from it and
>> refuses to export. In the logs you'll see errors such as:
>>
>> mountd: Cannot export /srv/nfs/vol0, possibly unsupported filesystem or fsid= required
>>
>> To deal with that I've changed rpc.mountd to use generate and store fsids [1].
>> Since the kernel side of my changes did change for a long time I decided to
>> try upstreaming it first.
>> A 3rd iteration of my rpc.mountd will happen soon.
>>
>> [0] https://marc.info/?l=linux-nfs&m=161653016627277&w=2
>> [1] https://lore.kernel.org/linux-nfs/[email protected]/
>>
>> Changes since v1:
>> https://lore.kernel.org/linux-nfs/[email protected]/
>>
>> - Use LOOKUP_AUTOMOUNT only when NFSEXP_CROSSMOUNT is set (Jeff Layton)
>>
>> Richard Weinberger (3):
>> NFSD: Teach nfsd_mountpoint() auto mounts
>> fs: namei: Allow follow_down() to uncover auto mounts
>> NFS: nfs_encode_fh: Remove S_AUTOMOUNT check
>>
>> fs/namei.c | 6 +++---
>> fs/nfs/export.c | 2 +-
>> fs/nfsd/vfs.c | 8 ++++++--
>> include/linux/namei.h | 2 +-
>> 4 files changed, 11 insertions(+), 7 deletions(-)
>>
> This set looks reasonable to me.
>
> Reviewed-by: Jeff Layton <[email protected]>
Right, looks ok to me too, at least from the POV of that follow_down()
change.
Reviewed-by: Ian Kent <[email protected]>
Ian
> On Dec 7, 2022, at 3:43 AM, Richard Weinberger <[email protected]> wrote:
>
> This function is only used by NFSD to cross mount points.
> If a mount point is of type auto mount, follow_down() will
> not uncover it. Add LOOKUP_AUTOMOUNT to the lookup flags
> to have ->d_automount() called when NFSD walks down the
> mount tree.
>
> Signed-off-by: Richard Weinberger <[email protected]>
> ---
> fs/namei.c | 6 +++---
> fs/nfsd/vfs.c | 6 +++++-
> include/linux/namei.h | 2 +-
> 3 files changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/fs/namei.c b/fs/namei.c
> index 578c2110df02..a6bb6863bf0c 100644
> --- a/fs/namei.c
> +++ b/fs/namei.c
> @@ -1458,11 +1458,11 @@ EXPORT_SYMBOL(follow_down_one);
> * point, the filesystem owning that dentry may be queried as to whether the
> * caller is permitted to proceed or not.
> */
> -int follow_down(struct path *path)
> +int follow_down(struct path *path, unsigned int flags)
> {
> struct vfsmount *mnt = path->mnt;
> bool jumped;
> - int ret = traverse_mounts(path, &jumped, NULL, 0);
> + int ret = traverse_mounts(path, &jumped, NULL, flags);
>
> if (path->mnt != mnt)
> mntput(mnt);
> @@ -2864,7 +2864,7 @@ int path_pts(struct path *path)
>
> path->dentry = child;
> dput(parent);
> - follow_down(path);
> + follow_down(path, 0);
> return 0;
> }
> #endif
> diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
> index 157f0df0e93a..ced04fc2b947 100644
> --- a/fs/nfsd/vfs.c
> +++ b/fs/nfsd/vfs.c
> @@ -63,9 +63,13 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
> struct dentry *dentry = *dpp;
> struct path path = {.mnt = mntget(exp->ex_path.mnt),
> .dentry = dget(dentry)};
> + unsigned int follow_flags = 0;
> int err = 0;
>
> - err = follow_down(&path);
> + if (exp->ex_flags & NFSEXP_CROSSMOUNT)
> + follow_flags = LOOKUP_AUTOMOUNT;
> +
> + err = follow_down(&path, follow_flags);
> if (err < 0)
> goto out;
> if (path.mnt == exp->ex_path.mnt && path.dentry == dentry &&
> diff --git a/include/linux/namei.h b/include/linux/namei.h
> index 00fee52df842..6f96db73a70a 100644
> --- a/include/linux/namei.h
> +++ b/include/linux/namei.h
> @@ -77,7 +77,7 @@ struct dentry *lookup_one_positive_unlocked(struct user_namespace *mnt_userns,
> struct dentry *base, int len);
>
> extern int follow_down_one(struct path *);
> -extern int follow_down(struct path *);
> +extern int follow_down(struct path *, unsigned int flags);
> extern int follow_up(struct path *);
>
> extern struct dentry *lock_rename(struct dentry *, struct dentry *);
> --
> 2.26.2
>
I plan to take this through the nfsd tree, thus this one needs
an Ack from the VFS maintainer(s).
--
Chuck Lever
> On Dec 7, 2022, at 3:43 AM, Richard Weinberger <[email protected]> wrote:
>
> This function is only used by NFSD to cross mount points.
> If a mount point is of type auto mount, follow_down() will
> not uncover it. Add LOOKUP_AUTOMOUNT to the lookup flags
> to have ->d_automount() called when NFSD walks down the
> mount tree.
>
> Signed-off-by: Richard Weinberger <[email protected]>
Hello Al, you are top of the maintainers listed for fs/namei.c.
I'd like to take this series for v6.3 via the nfsd tree. Can
I get your Acked-by: for this one?
> ---
> fs/namei.c | 6 +++---
> fs/nfsd/vfs.c | 6 +++++-
> include/linux/namei.h | 2 +-
> 3 files changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/fs/namei.c b/fs/namei.c
> index 578c2110df02..a6bb6863bf0c 100644
> --- a/fs/namei.c
> +++ b/fs/namei.c
> @@ -1458,11 +1458,11 @@ EXPORT_SYMBOL(follow_down_one);
> * point, the filesystem owning that dentry may be queried as to whether the
> * caller is permitted to proceed or not.
> */
> -int follow_down(struct path *path)
> +int follow_down(struct path *path, unsigned int flags)
> {
> struct vfsmount *mnt = path->mnt;
> bool jumped;
> - int ret = traverse_mounts(path, &jumped, NULL, 0);
> + int ret = traverse_mounts(path, &jumped, NULL, flags);
>
> if (path->mnt != mnt)
> mntput(mnt);
> @@ -2864,7 +2864,7 @@ int path_pts(struct path *path)
>
> path->dentry = child;
> dput(parent);
> - follow_down(path);
> + follow_down(path, 0);
> return 0;
> }
> #endif
> diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
> index 157f0df0e93a..ced04fc2b947 100644
> --- a/fs/nfsd/vfs.c
> +++ b/fs/nfsd/vfs.c
> @@ -63,9 +63,13 @@ nfsd_cross_mnt(struct svc_rqst *rqstp, struct dentry **dpp,
> struct dentry *dentry = *dpp;
> struct path path = {.mnt = mntget(exp->ex_path.mnt),
> .dentry = dget(dentry)};
> + unsigned int follow_flags = 0;
> int err = 0;
>
> - err = follow_down(&path);
> + if (exp->ex_flags & NFSEXP_CROSSMOUNT)
> + follow_flags = LOOKUP_AUTOMOUNT;
> +
> + err = follow_down(&path, follow_flags);
> if (err < 0)
> goto out;
> if (path.mnt == exp->ex_path.mnt && path.dentry == dentry &&
> diff --git a/include/linux/namei.h b/include/linux/namei.h
> index 00fee52df842..6f96db73a70a 100644
> --- a/include/linux/namei.h
> +++ b/include/linux/namei.h
> @@ -77,7 +77,7 @@ struct dentry *lookup_one_positive_unlocked(struct user_namespace *mnt_userns,
> struct dentry *base, int len);
>
> extern int follow_down_one(struct path *);
> -extern int follow_down(struct path *);
> +extern int follow_down(struct path *, unsigned int flags);
> extern int follow_up(struct path *);
>
> extern struct dentry *lock_rename(struct dentry *, struct dentry *);
> --
> 2.26.2
>
--
Chuck Lever
----- Ursprüngliche Mail -----
> Von: "chuck lever" <[email protected]>
>> On Dec 7, 2022, at 3:43 AM, Richard Weinberger <[email protected]> wrote:
>>
>> This function is only used by NFSD to cross mount points.
>> If a mount point is of type auto mount, follow_down() will
>> not uncover it. Add LOOKUP_AUTOMOUNT to the lookup flags
>> to have ->d_automount() called when NFSD walks down the
>> mount tree.
>>
>> Signed-off-by: Richard Weinberger <[email protected]>
>
> Hello Al, you are top of the maintainers listed for fs/namei.c.
> I'd like to take this series for v6.3 via the nfsd tree. Can
> I get your Acked-by: for this one?
ping?
Thanks,
//richard
On Sun, Jan 29, 2023 at 04:42:39PM +0100, Richard Weinberger wrote:
> ----- Urspr?ngliche Mail -----
> > Von: "chuck lever" <[email protected]>
> >> On Dec 7, 2022, at 3:43 AM, Richard Weinberger <[email protected]> wrote:
> >>
> >> This function is only used by NFSD to cross mount points.
> >> If a mount point is of type auto mount, follow_down() will
> >> not uncover it. Add LOOKUP_AUTOMOUNT to the lookup flags
> >> to have ->d_automount() called when NFSD walks down the
> >> mount tree.
> >>
> >> Signed-off-by: Richard Weinberger <[email protected]>
> >
> > Hello Al, you are top of the maintainers listed for fs/namei.c.
> > I'd like to take this series for v6.3 via the nfsd tree. Can
> > I get your Acked-by: for this one?
>
> ping?
modulo clumsy wording ("mount point is of type auto mount")
Acked-by: Al Viro <[email protected]>
Commit message sounds as if it refered to autofs, rather than NFS referrals
et.al. and AFAICS those are the cases it's really about...