[PATCH] knfsd: add some new fsid types

Add support for using a filesystem UUID to identify and export point in the
filehandle.

For NFSv2, this UUID is xor-ed down to 4 or 8 bytes so that it doesn't take up
too much room.  For NFSv3+, we use the full 16 bytes, and possibly also a
64bit inode number for exports beneath the root of a filesystem.

When generating an fsid to return in 'stat' information, use the UUID (hashed
down to size) if it is available and a small 'fsid' was not specifically
provided.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

5 files changed, 162 insertions, 92 deletions

diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c
index 49c310b84923..bb3c314e2479 100644
--- a/fs/nfsd/export.c
+++ b/fs/nfsd/export.c
@@ -190,18 +190,17 @@ static int expkey_show(struct seq_file *m,
                        struct cache_head *h)
 {
         struct svc_expkey *ek ;
+        int i;
 
         if (h ==NULL) {
                 seq_puts(m, "#domain fsidtype fsid [path]\n");
                 return 0;
         }
         ek = container_of(h, struct svc_expkey, h);
-        seq_printf(m, "%s %d 0x%08x", ek->ek_client->name,
-                   ek->ek_fsidtype, ek->ek_fsid[0]);
-        if (ek->ek_fsidtype != 1)
-                seq_printf(m, "%08x", ek->ek_fsid[1]);
-        if (ek->ek_fsidtype == 2)
-                seq_printf(m, "%08x", ek->ek_fsid[2]);
+        seq_printf(m, "%s %d 0x", ek->ek_client->name,
+                   ek->ek_fsidtype);
+        for (i=0; i < key_len(ek->ek_fsidtype)/4; i++)
+                seq_printf(m, "%08x", ek->ek_fsid[i]);
         if (test_bit(CACHE_VALID, &h->flags) && 
             !test_bit(CACHE_NEGATIVE, &h->flags)) {
                 seq_printf(m, " ");
@@ -232,9 +231,8 @@ static inline void expkey_init(struct cache_head *cnew,
         kref_get(&item->ek_client->ref);
         new->ek_client = item->ek_client;
         new->ek_fsidtype = item->ek_fsidtype;
-        new->ek_fsid[0] = item->ek_fsid[0];
-        new->ek_fsid[1] = item->ek_fsid[1];
-        new->ek_fsid[2] = item->ek_fsid[2];
+
+        memcpy(new->ek_fsid, item->ek_fsid, sizeof(new->ek_fsid));
 }
 
 static inline void expkey_update(struct cache_head *cnew,
@@ -363,7 +361,7 @@ static struct svc_export *svc_export_update(struct svc_export *new,
                                             struct svc_export *old);
 static struct svc_export *svc_export_lookup(struct svc_export *);
 
-static int check_export(struct inode *inode, int flags)
+static int check_export(struct inode *inode, int flags, unsigned char *uuid)
 {
 
         /* We currently export only dirs and regular files.
@@ -376,12 +374,13 @@ static int check_export(struct inode *inode, int flags)
         /* There are two requirements on a filesystem to be exportable.
          * 1:  We must be able to identify the filesystem from a number.
          *       either a device number (so FS_REQUIRES_DEV needed)
-         *       or an FSID number (so NFSEXP_FSID needed).
+         *       or an FSID number (so NFSEXP_FSID or ->uuid is needed).
          * 2:  We must be able to find an inode from a filehandle.
          *       This means that s_export_op must be set.
          */
         if (!(inode->i_sb->s_type->fs_flags & FS_REQUIRES_DEV) &&
-            !(flags & NFSEXP_FSID)) {
+            !(flags & NFSEXP_FSID) &&
+            uuid == NULL) {
                 dprintk("exp_export: export of non-dev fs without fsid\n");
                 return -EINVAL;
         }
@@ -406,10 +405,6 @@ fsloc_parse(char **mesg, char *buf, struct nfsd4_fs_locations *fsloc)
         int len;
         int migrated, i, err;
 
-        len = qword_get(mesg, buf, PAGE_SIZE);
-        if (len != 5 || memcmp(buf, "fsloc", 5))
-                return 0;
-
         /* listsize */
         err = get_int(mesg, &fsloc->locations_count);
         if (err)
@@ -520,6 +515,8 @@ static int svc_export_parse(struct cache_detail *cd, char *mesg, int mlen)
         exp.ex_fslocs.locations_count = 0;
         exp.ex_fslocs.migrated = 0;
 
+        exp.ex_uuid = NULL;
+
         /* flags */
         err = get_int(&mesg, &an_int);
         if (err == -ENOENT)
@@ -543,12 +540,33 @@ static int svc_export_parse(struct cache_detail *cd, char *mesg, int mlen)
                 if (err) goto out;
                 exp.ex_fsid = an_int;
 
-                err = check_export(nd.dentry->d_inode, exp.ex_flags);
-                if (err) goto out;
+                while ((len = qword_get(&mesg, buf, PAGE_SIZE)) > 0) {
+                        if (strcmp(buf, "fsloc") == 0)
+                                err = fsloc_parse(&mesg, buf, &exp.ex_fslocs);
+                        else if (strcmp(buf, "uuid") == 0) {
+                                /* expect a 16 byte uuid encoded as \xXXXX... */
+                                len = qword_get(&mesg, buf, PAGE_SIZE);
+                                if (len != 16)
+                                        err  = -EINVAL;
+                                else {
+                                        exp.ex_uuid =
+                                                kmemdup(buf, 16, GFP_KERNEL);
+                                        if (exp.ex_uuid == NULL)
+                                                err = -ENOMEM;
+                                }
+                        } else
+                                /* quietly ignore unknown words and anything
+                                 * following. Newer user-space can try to set
+                                 * new values, then see what the result was.
+                                 */
+                                break;
+                        if (err)
+                                goto out;
+                }
 
-                err = fsloc_parse(&mesg, buf, &exp.ex_fslocs);
-                if (err)
-                        goto out;
+                err = check_export(nd.dentry->d_inode, exp.ex_flags,
+                                   exp.ex_uuid);
+                if (err) goto out;
         }
 
         expp = svc_export_lookup(&exp);
@@ -562,6 +580,8 @@ static int svc_export_parse(struct cache_detail *cd, char *mesg, int mlen)
         else
                 exp_put(expp);
  out:
+        nfsd4_fslocs_free(&exp.ex_fslocs);
+        kfree(exp.ex_uuid);
         kfree(exp.ex_path);
         if (nd.dentry)
                 path_release(&nd);
@@ -591,9 +611,19 @@ static int svc_export_show(struct seq_file *m,
         seq_escape(m, exp->ex_client->name, " \t\n\\");
         seq_putc(m, '(');
         if (test_bit(CACHE_VALID, &h->flags) && 
-            !test_bit(CACHE_NEGATIVE, &h->flags))
+            !test_bit(CACHE_NEGATIVE, &h->flags)) {
                 exp_flags(m, exp->ex_flags, exp->ex_fsid,
                           exp->ex_anon_uid, exp->ex_anon_gid, &exp->ex_fslocs);
+                if (exp->ex_uuid) {
+                        int i;
+                        seq_puts(m, ",uuid=");
+                        for (i=0; i<16; i++) {
+                                if ((i&3) == 0 && i)
+                                        seq_putc(m, ':');
+                                seq_printf(m, "%02x", exp->ex_uuid[i]);
+                        }
+                }
+        }
         seq_puts(m, ")\n");
         return 0;
 }
@@ -630,6 +660,8 @@ static void export_update(struct cache_head *cnew, struct cache_head *citem)
         new->ex_anon_uid = item->ex_anon_uid;
         new->ex_anon_gid = item->ex_anon_gid;
         new->ex_fsid = item->ex_fsid;
+        new->ex_uuid = item->ex_uuid;
+        item->ex_uuid = NULL;
         new->ex_path = item->ex_path;
         item->ex_path = NULL;
         new->ex_fslocs.locations = item->ex_fslocs.locations;
@@ -752,11 +784,11 @@ exp_get_key(svc_client *clp, dev_t dev, ino_t ino)
         u32 fsidv[3];
         
         if (old_valid_dev(dev)) {
-                mk_fsid_v0(fsidv, dev, ino);
-                return exp_find_key(clp, 0, fsidv, NULL);
+                mk_fsid(FSID_DEV, fsidv, dev, ino, 0, NULL);
+                return exp_find_key(clp, FSID_DEV, fsidv, NULL);
         }
-        mk_fsid_v3(fsidv, dev, ino);
-        return exp_find_key(clp, 3, fsidv, NULL);
+        mk_fsid(FSID_ENCODE_DEV, fsidv, dev, ino, 0, NULL);
+        return exp_find_key(clp, FSID_ENCODE_DEV, fsidv, NULL);
 }
 
 /*
@@ -767,9 +799,9 @@ exp_get_fsid_key(svc_client *clp, int fsid)
 {
         u32 fsidv[2];
 
-        mk_fsid_v1(fsidv, fsid);
+        mk_fsid(FSID_NUM, fsidv, 0, 0, fsid, NULL);
 
-        return exp_find_key(clp, 1, fsidv, NULL);
+        return exp_find_key(clp, FSID_NUM, fsidv, NULL);
 }
 
 svc_export *
@@ -883,8 +915,8 @@ static int exp_fsid_hash(svc_client *clp, struct svc_export *exp)
         if ((exp->ex_flags & NFSEXP_FSID) == 0)
                 return 0;
 
-        mk_fsid_v1(fsid, exp->ex_fsid);
-        return exp_set_key(clp, 1, fsid, exp);
+        mk_fsid(FSID_NUM, fsid, 0, 0, exp->ex_fsid, NULL);
+        return exp_set_key(clp, FSID_NUM, fsid, exp);
 }
 
 static int exp_hash(struct auth_domain *clp, struct svc_export *exp)
@@ -894,11 +926,11 @@ static int exp_hash(struct auth_domain *clp, struct svc_export *exp)
         dev_t dev = inode->i_sb->s_dev;
 
         if (old_valid_dev(dev)) {
-                mk_fsid_v0(fsid, dev, inode->i_ino);
-                return exp_set_key(clp, 0, fsid, exp);
+                mk_fsid(FSID_DEV, fsid, dev, inode->i_ino, 0, NULL);
+                return exp_set_key(clp, FSID_DEV, fsid, exp);
         }
-        mk_fsid_v3(fsid, dev, inode->i_ino);
-        return exp_set_key(clp, 3, fsid, exp);
+        mk_fsid(FSID_ENCODE_DEV, fsid, dev, inode->i_ino, 0, NULL);
+        return exp_set_key(clp, FSID_ENCODE_DEV, fsid, exp);
 }
 
 static void exp_unhash(struct svc_export *exp)
@@ -977,7 +1009,7 @@ exp_export(struct nfsctl_export *nxp)
                 goto finish;
         }
 
-        err = check_export(nd.dentry->d_inode, nxp->ex_flags);
+        err = check_export(nd.dentry->d_inode, nxp->ex_flags, NULL);
         if (err) goto finish;
 
         err = -ENOMEM;
@@ -1170,9 +1202,9 @@ exp_pseudoroot(struct auth_domain *clp, struct svc_fh *fhp,
         __be32 rv;
         u32 fsidv[2];
 
-        mk_fsid_v1(fsidv, 0);
+        mk_fsid(FSID_NUM, fsidv, 0, 0, 0, NULL);
 
-        exp = exp_find(clp, 1, fsidv, creq);
+        exp = exp_find(clp, FSID_NUM, fsidv, creq);
         if (IS_ERR(exp))
                 return nfserrno(PTR_ERR(exp));
         if (exp == NULL)
diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c
index e695660921ec..6f677988c71d 100644
--- a/fs/nfsd/nfs3xdr.c
+++ b/fs/nfsd/nfs3xdr.c
@@ -149,6 +149,27 @@ decode_sattr3(__be32 *p, struct iattr *iap)
         return p;
 }
 
+static __be32 *encode_fsid(__be32 *p, struct svc_fh *fhp)
+{
+        u64 f;
+        switch(fsid_source(fhp)) {
+        default:
+        case FSIDSOURCE_DEV:
+                p = xdr_encode_hyper(p, (u64)huge_encode_dev
+                                     (fhp->fh_dentry->d_inode->i_sb->s_dev));
+                break;
+        case FSIDSOURCE_FSID:
+                p = xdr_encode_hyper(p, (u64) fhp->fh_export->ex_fsid);
+                break;
+        case FSIDSOURCE_UUID:
+                f = ((u64*)fhp->fh_export->ex_uuid)[0];
+                f ^= ((u64*)fhp->fh_export->ex_uuid)[1];
+                p = xdr_encode_hyper(p, f);
+                break;
+        }
+        return p;
+}
+
 static __be32 *
 encode_fattr3(struct svc_rqst *rqstp, __be32 *p, struct svc_fh *fhp,
               struct kstat *stat)
@@ -169,10 +190,7 @@ encode_fattr3(struct svc_rqst *rqstp, __be32 *p, struct svc_fh *fhp,
         p = xdr_encode_hyper(p, ((u64)stat->blocks) << 9);
         *p++ = htonl((u32) MAJOR(stat->rdev));
         *p++ = htonl((u32) MINOR(stat->rdev));
-        if (is_fsid(fhp, rqstp->rq_reffh))
-                p = xdr_encode_hyper(p, (u64) fhp->fh_export->ex_fsid);
-        else
-                p = xdr_encode_hyper(p, (u64) huge_encode_dev(stat->dev));
+        p = encode_fsid(p, fhp);
         p = xdr_encode_hyper(p, (u64) stat->ino);
         p = encode_time3(p, &stat->atime);
         lease_get_mtime(dentry->d_inode, &time); 
@@ -203,10 +221,7 @@ encode_saved_post_attr(struct svc_rqst *rqstp, __be32 *p, struct svc_fh *fhp)
         p = xdr_encode_hyper(p, ((u64)fhp->fh_post_blocks) << 9);
         *p++ = fhp->fh_post_rdev[0];
         *p++ = fhp->fh_post_rdev[1];
-        if (is_fsid(fhp, rqstp->rq_reffh))
-                p = xdr_encode_hyper(p, (u64) fhp->fh_export->ex_fsid);
-        else
-                p = xdr_encode_hyper(p, (u64)huge_encode_dev(inode->i_sb->s_dev));
+        p = encode_fsid(p, fhp);
         p = xdr_encode_hyper(p, (u64) inode->i_ino);
         p = encode_time3(p, &fhp->fh_post_atime);
         p = encode_time3(p, &fhp->fh_post_mtime);
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index 18aa9440df14..0efba557fb55 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -1563,14 +1563,20 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp,
                 if (exp->ex_fslocs.migrated) {
                         WRITE64(NFS4_REFERRAL_FSID_MAJOR);
                         WRITE64(NFS4_REFERRAL_FSID_MINOR);
-                } else if (is_fsid(fhp, rqstp->rq_reffh)) {
+                } else switch(fsid_source(fhp)) {
+                case FSIDSOURCE_FSID:
                         WRITE64((u64)exp->ex_fsid);
                         WRITE64((u64)0);
-                } else {
+                        break;
+                case FSIDSOURCE_DEV:
                         WRITE32(0);
                         WRITE32(MAJOR(stat.dev));
                         WRITE32(0);
                         WRITE32(MINOR(stat.dev));
+                        break;
+                case FSIDSOURCE_UUID:
+                        WRITEMEM(exp->ex_uuid, 16);
+                        break;
                 }
         }
         if (bmval0 & FATTR4_WORD0_UNIQUE_HANDLES) {
diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
index 12c5e7421374..286bc4d356f4 100644
--- a/fs/nfsd/nfsfh.c
+++ b/fs/nfsd/nfsfh.c
@@ -119,9 +119,6 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access)
 
         dprintk("nfsd: fh_verify(%s)\n", SVCFH_fmt(fhp));
 
-        /* keep this filehandle for possible reference  when encoding attributes */
-        rqstp->rq_reffh = fh;
-
         if (!fhp->fh_dentry) {
                 __u32 *datap=NULL;
                 __u32 tfh[3];           /* filehandle fragment for oldstyle filehandles */
@@ -146,10 +143,10 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access)
                         }
                         len = key_len(fh->fh_fsid_type) / 4;
                         if (len == 0) goto out;
-                        if  (fh->fh_fsid_type == 2) {
+                        if  (fh->fh_fsid_type == FSID_MAJOR_MINOR) {
                                 /* deprecated, convert to type 3 */
-                                len = 3;
-                                fh->fh_fsid_type = 3;
+                                len = key_len(FSID_ENCODE_DEV)/4;
+                                fh->fh_fsid_type = FSID_ENCODE_DEV;
                                 fh->fh_fsid[0] = new_encode_dev(MKDEV(ntohl(fh->fh_fsid[0]), ntohl(fh->fh_fsid[1])));
                                 fh->fh_fsid[1] = fh->fh_fsid[2];
                         }
@@ -164,8 +161,9 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access)
                         /* assume old filehandle format */
                         xdev = old_decode_dev(fh->ofh_xdev);
                         xino = u32_to_ino_t(fh->ofh_xino);
-                        mk_fsid_v0(tfh, xdev, xino);
-                        exp = exp_find(rqstp->rq_client, 0, tfh, &rqstp->rq_chandle);
+                        mk_fsid(FSID_DEV, tfh, xdev, xino, 0, NULL);
+                        exp = exp_find(rqstp->rq_client, FSID_DEV, tfh,
+                                       &rqstp->rq_chandle);
                 }
 
                 if (IS_ERR(exp) && (PTR_ERR(exp) == -EAGAIN
@@ -334,6 +332,7 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
         struct dentry *parent = dentry->d_parent;
         __u32 *datap;
         dev_t ex_dev = exp->ex_dentry->d_inode->i_sb->s_dev;
+        int root_export = (exp->ex_dentry == exp->ex_dentry->d_sb->s_root);
 
         dprintk("nfsd: fh_compose(exp %02x:%02x/%ld %s/%s, ino=%ld)\n",
                 MAJOR(ex_dev), MINOR(ex_dev),
@@ -348,19 +347,31 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
         if (ref_fh && ref_fh->fh_export == exp) {
                 version = ref_fh->fh_handle.fh_version;
                 if (version == 0xca)
-                        fsid_type = 0;
+                        fsid_type = FSID_DEV;
                 else
                         fsid_type = ref_fh->fh_handle.fh_fsid_type;
                 /* We know this version/type works for this export
                  * so there is no need for further checks.
                  */
+        } else if (exp->ex_uuid) {
+                if (fhp->fh_maxsize >= 64) {
+                        if (root_export)
+                                fsid_type = FSID_UUID16;
+                        else
+                                fsid_type = FSID_UUID16_INUM;
+                } else {
+                        if (root_export)
+                                fsid_type = FSID_UUID8;
+                        else
+                                fsid_type = FSID_UUID4_INUM;
+                }
         } else if (exp->ex_flags & NFSEXP_FSID)
-                fsid_type = 1;
+                fsid_type = FSID_NUM;
         else if (!old_valid_dev(ex_dev))
                 /* for newer device numbers, we must use a newer fsid format */
-                fsid_type = 3;
+                fsid_type = FSID_ENCODE_DEV;
         else
-                fsid_type = 0;
+                fsid_type = FSID_DEV;
 
         if (ref_fh == fhp)
                 fh_put(ref_fh);
@@ -396,36 +407,10 @@ fh_compose(struct svc_fh *fhp, struct svc_export *exp, struct dentry *dentry,
                 fhp->fh_handle.fh_auth_type = 0;
                 datap = fhp->fh_handle.fh_auth+0;
                 fhp->fh_handle.fh_fsid_type = fsid_type;
-                switch (fsid_type) {
-                case 0:
-                        /*
-                         * fsid_type 0:
-                         * 2byte major, 2byte minor, 4byte inode
-                         */
-                        mk_fsid_v0(datap, ex_dev,
-                                   exp->ex_dentry->d_inode->i_ino);
-                        break;
-                case 1:
-                        /* fsid_type 1 == 4 bytes filesystem id */
-                        mk_fsid_v1(datap, exp->ex_fsid);
-                        break;
-                case 2:
-                        /*
-                         * fsid_type 2:
-                         * 4byte major, 4byte minor, 4byte inode
-                         */
-                        mk_fsid_v2(datap, ex_dev,
-                                   exp->ex_dentry->d_inode->i_ino);
-                        break;
-                case 3:
-                        /*
-                         * fsid_type 3:
-                         * 4byte devicenumber, 4byte inode
-                         */
-                        mk_fsid_v3(datap, ex_dev,
-                                   exp->ex_dentry->d_inode->i_ino);
-                        break;
-                }
+                mk_fsid(fsid_type, datap, ex_dev,
+                        exp->ex_dentry->d_inode->i_ino,
+                        exp->ex_fsid, exp->ex_uuid);
+
                 len = key_len(fsid_type);
                 datap += len/4;
                 fhp->fh_handle.fh_size = 4 + len;
@@ -530,3 +515,22 @@ char * SVCFH_fmt(struct svc_fh *fhp)
                 fh->fh_base.fh_pad[5]);
         return buf;
 }
+
+enum fsid_source fsid_source(struct svc_fh *fhp)
+{
+        if (fhp->fh_handle.fh_version != 1)
+                return FSIDSOURCE_DEV;
+        switch(fhp->fh_handle.fh_fsid_type) {
+        case FSID_DEV:
+        case FSID_ENCODE_DEV:
+        case FSID_MAJOR_MINOR:
+                return FSIDSOURCE_DEV;
+        case FSID_NUM:
+                return FSIDSOURCE_FSID;
+        default:
+                if (fhp->fh_export->ex_flags & NFSEXP_FSID)
+                        return FSIDSOURCE_FSID;
+                else
+                        return FSIDSOURCE_UUID;
+        }
+}
diff --git a/fs/nfsd/nfsxdr.c b/fs/nfsd/nfsxdr.c
index 6555c50d9006..0c24b9e24fe8 100644
--- a/fs/nfsd/nfsxdr.c
+++ b/fs/nfsd/nfsxdr.c
@@ -153,6 +153,7 @@ encode_fattr(struct svc_rqst *rqstp, __be32 *p, struct svc_fh *fhp,
         struct dentry   *dentry = fhp->fh_dentry;
         int type;
         struct timespec time;
+        u32 f;
 
         type = (stat->mode & S_IFMT);
 
@@ -173,10 +174,22 @@ encode_fattr(struct svc_rqst *rqstp, __be32 *p, struct svc_fh *fhp,
         else
                 *p++ = htonl(0xffffffff);
         *p++ = htonl((u32) stat->blocks);
-        if (is_fsid(fhp, rqstp->rq_reffh))
-                *p++ = htonl((u32) fhp->fh_export->ex_fsid);
-        else
+        switch (fsid_source(fhp)) {
+        default:
+        case FSIDSOURCE_DEV:
                 *p++ = htonl(new_encode_dev(stat->dev));
+                break;
+        case FSIDSOURCE_FSID:
+                *p++ = htonl((u32) fhp->fh_export->ex_fsid);
+                break;
+        case FSIDSOURCE_UUID:
+                f = ((u32*)fhp->fh_export->ex_uuid)[0];
+                f ^= ((u32*)fhp->fh_export->ex_uuid)[1];
+                f ^= ((u32*)fhp->fh_export->ex_uuid)[2];
+                f ^= ((u32*)fhp->fh_export->ex_uuid)[3];
+                *p++ = htonl(f);
+                break;
+        }
         *p++ = htonl((u32) stat->ino);
         *p++ = htonl((u32) stat->atime.tv_sec);
         *p++ = htonl(stat->atime.tv_nsec ? stat->atime.tv_nsec / 1000 : 0);