Re: [PATCH v6 6/9] nfsd: use the getattr operation to fetch i_version

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, 2022-10-06 at 08:14 +1100, NeilBrown wrote:
> On Wed, 05 Oct 2022, Jeff Layton wrote:
> > On Tue, 2022-10-04 at 10:39 +1100, NeilBrown wrote:
> > > On Fri, 30 Sep 2022, Jeff Layton wrote:
> > > > Now that we can call into vfs_getattr to get the i_version field, use
> > > > that facility to fetch it instead of doing it in nfsd4_change_attribute.
> > > > 
> > > > Neil also pointed out recently that IS_I_VERSION directory operations
> > > > are always logged, and so we only need to mitigate the rollback problem
> > > > on regular files. Also, we don't need to factor in the ctime when
> > > > reexporting NFS or Ceph.
> > > > 
> > > > Set the STATX_VERSION (and BTIME) bits in the request when we're dealing
> > > > with a v4 request. Then, instead of looking at IS_I_VERSION when
> > > > generating the change attr, look at the result mask and only use it if
> > > > STATX_VERSION is set. With this change, we can drop the fetch_iversion
> > > > export operation as well.
> > > > 
> > > > Move nfsd4_change_attribute into nfsfh.c, and change it to only factor
> > > > in the ctime if it's a regular file and the fs doesn't advertise
> > > > STATX_ATTR_VERSION_MONOTONIC.
> > > > 
> > > > Signed-off-by: Jeff Layton <jlayton@xxxxxxxxxx>
> > > > ---
> > > >  fs/nfs/export.c          |  7 -------
> > > >  fs/nfsd/nfs4xdr.c        |  4 +++-
> > > >  fs/nfsd/nfsfh.c          | 40 ++++++++++++++++++++++++++++++++++++++++
> > > >  fs/nfsd/nfsfh.h          | 29 +----------------------------
> > > >  fs/nfsd/vfs.h            |  7 ++++++-
> > > >  include/linux/exportfs.h |  1 -
> > > >  6 files changed, 50 insertions(+), 38 deletions(-)
> > > > 
> > > > diff --git a/fs/nfs/export.c b/fs/nfs/export.c
> > > > index 01596f2d0a1e..1a9d5aa51dfb 100644
> > > > --- a/fs/nfs/export.c
> > > > +++ b/fs/nfs/export.c
> > > > @@ -145,17 +145,10 @@ nfs_get_parent(struct dentry *dentry)
> > > >  	return parent;
> > > >  }
> > > >  
> > > > -static u64 nfs_fetch_iversion(struct inode *inode)
> > > > -{
> > > > -	nfs_revalidate_inode(inode, NFS_INO_INVALID_CHANGE);
> > > > -	return inode_peek_iversion_raw(inode);
> > > > -}
> > > > -
> > > >  const struct export_operations nfs_export_ops = {
> > > >  	.encode_fh = nfs_encode_fh,
> > > >  	.fh_to_dentry = nfs_fh_to_dentry,
> > > >  	.get_parent = nfs_get_parent,
> > > > -	.fetch_iversion = nfs_fetch_iversion,
> > > >  	.flags = EXPORT_OP_NOWCC|EXPORT_OP_NOSUBTREECHK|
> > > >  		EXPORT_OP_CLOSE_BEFORE_UNLINK|EXPORT_OP_REMOTE_FS|
> > > >  		EXPORT_OP_NOATOMIC_ATTR,
> > > > diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
> > > > index 1e9690a061ec..779c009314c6 100644
> > > > --- a/fs/nfsd/nfs4xdr.c
> > > > +++ b/fs/nfsd/nfs4xdr.c
> > > > @@ -2869,7 +2869,9 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp,
> > > >  			goto out;
> > > >  	}
> > > >  
> > > > -	err = vfs_getattr(&path, &stat, STATX_BASIC_STATS, AT_STATX_SYNC_AS_STAT);
> > > > +	err = vfs_getattr(&path, &stat,
> > > > +			  STATX_BASIC_STATS | STATX_BTIME | STATX_VERSION,
> > > > +			  AT_STATX_SYNC_AS_STAT);
> > > >  	if (err)
> > > >  		goto out_nfserr;
> > > >  	if (!(stat.result_mask & STATX_BTIME))
> > > > diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
> > > > index a5b71526cee0..9168bc657378 100644
> > > > --- a/fs/nfsd/nfsfh.c
> > > > +++ b/fs/nfsd/nfsfh.c
> > > > @@ -634,6 +634,10 @@ void fh_fill_pre_attrs(struct svc_fh *fhp)
> > > >  		stat.mtime = inode->i_mtime;
> > > >  		stat.ctime = inode->i_ctime;
> > > >  		stat.size  = inode->i_size;
> > > > +		if (v4 && IS_I_VERSION(inode)) {
> > > > +			stat.version = inode_query_iversion(inode);
> > > > +			stat.result_mask |= STATX_VERSION;
> > > > +		}
> > > 
> > > This is increasingly ugly.  I wonder if it is justified at all...
> > > 
> > 
> > I'm fine with dropping that. So if the getattrs fail, we should just not
> > offer up pre/post attrs?
> > 
> > > >  	}
> > > >  	if (v4)
> > > >  		fhp->fh_pre_change = nfsd4_change_attribute(&stat, inode);
> > > > @@ -665,6 +669,8 @@ void fh_fill_post_attrs(struct svc_fh *fhp)
> > > >  	if (err) {
> > > >  		fhp->fh_post_saved = false;
> > > >  		fhp->fh_post_attr.ctime = inode->i_ctime;
> > > > +		if (v4 && IS_I_VERSION(inode))
> > > > +			fhp->fh_post_attr.version = inode_query_iversion(inode);
> > > 
> > > ... ditto ...
> > > 
> > > >  	} else
> > > >  		fhp->fh_post_saved = true;
> > > >  	if (v4)
> > > > @@ -754,3 +760,37 @@ enum fsid_source fsid_source(const struct svc_fh *fhp)
> > > >  		return FSIDSOURCE_UUID;
> > > >  	return FSIDSOURCE_DEV;
> > > >  }
> > > > +
> > > > +/*
> > > > + * We could use i_version alone as the change attribute.  However, i_version
> > > > + * can go backwards on a regular file after an unclean shutdown.  On its own
> > > > + * that doesn't necessarily cause a problem, but if i_version goes backwards
> > > > + * and then is incremented again it could reuse a value that was previously
> > > > + * used before boot, and a client who queried the two values might incorrectly
> > > > + * assume nothing changed.
> > > > + *
> > > > + * By using both ctime and the i_version counter we guarantee that as long as
> > > > + * time doesn't go backwards we never reuse an old value. If the filesystem
> > > > + * advertises STATX_ATTR_VERSION_MONOTONIC, then this mitigation is not needed.
> > > > + *
> > > > + * We only need to do this for regular files as well. For directories, we
> > > > + * assume that the new change attr is always logged to stable storage in some
> > > > + * fashion before the results can be seen.
> > > > + */
> > > > +u64 nfsd4_change_attribute(struct kstat *stat, struct inode *inode)
> > > > +{
> > > > +	u64 chattr;
> > > > +
> > > > +	if (stat->result_mask & STATX_VERSION) {
> > > > +		chattr = stat->version;
> > > > +
> > > > +		if (S_ISREG(inode->i_mode) &&
> > > > +		    !(stat->attributes & STATX_ATTR_VERSION_MONOTONIC)) {
> > > 
> > > I would really rather that the fs got to make this decision.
> > > If it can guarantee that the i_version is monotonic even over a crash
> > > (which is probably can for directory, and might need changes to do for
> > > files) then it sets STATX_ATTR_VERSION_MONOTONIC and nfsd trusts it
> > > completely.
> > > If it cannot, then it doesn't set the flag.
> > > i.e. the S_ISREG() test should be in the filesystem, not in nfsd.
> > > 
> > 
> > This sounds reasonable, but for one thing.
> > 
> > From RFC 7862:
> > 
> >    While Section 5.4 of [RFC5661] discusses
> >    per-file system attributes, it is expected that the value of
> >    change_attr_type will not depend on the value of "homogeneous" and
> >    will only change in the event of a migration.
> > 
> > The change_attr_type4 must be the same for all filehandles under a
> > particular filesystem.
> > 
> > If we do what you suggest though, then it's easily possible for the fs
> > to set STATX_ATTR_VERSION_MONOTONIC on directories but not files. If we
> > later want to allow nfsd to advertise a change_attr_type4, we won't be
> > able to rely on the STATX_ATTR_VERSION_MONOTONIC to tell us how to fill
> > that out.
> > 
> > Maybe that's ok. I suppose we could add a new field to the export
> > options that filesystems can set to advertise what sort of change attr
> > they offer?
> > 
> 
> There are 3 cases:
> 1/ a file/dir which advertises MONOTONIC is easy to handle.
> 2/ an IS_I_VERSION file/dir that does not advertise MONOTONIC will only fail
>    to be MONOTONIC across unclean restart (correct?).  nfsd can
>    compensate using an xattr on the root to count crashes, or just adding ctime.
> 3/ a non-IS_I_VERSION fs that does not advertise MONOTONIC cannot
>    be compensated for by nfsd.
> 
> If we ever want nfsd to advertise MONOTONIC, then we must be able to
> reject non-IS_I_VERSION filesystems that don't advertise MONOTONIC on
> all files.
> 
> Maybe we need a global nfsd option which defaults to "monotoric" and
> causes those files to be rejected, but can be set to "non-monotonic" and
> then allows all files to be exported.
> 
> It would be nice to make it easy to run multiple nfsd instances each on a
> different IP address.  Each can then have different options.  This could
> also be used to reexport an NFS mount using unmodified filehandles.
> 
> Currently you need a network namespace to create a new nfsd.  I wonder
> if that is a little too much of a barrier.  But maybe we could automate
> the creation of working network namespaces for nfsd....
> 

My current thinking is to just allow the filesystem to set
STATX_ATTR_VERSION_MONOTONIC flag on a per-inode basis, and create a new
change_attr_type() export operation and leave it up to the filesystem to
fill that out appropriately.

I think that'll give us maximum flexibility, and would also allow NFS to
pass the change attr type from the server directly through to the client
when reexporting.
-- 
Jeff Layton <jlayton@xxxxxxxxxx>




[Index of Archives]     [XFS Filesystem Development (older mail)]     [Linux Filesystem Development]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux RAID]     [Linux SCSI]


  Powered by Linux