summaryrefslogtreecommitdiffstats
path: root/fs/ceph
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ceph')
-rw-r--r--fs/ceph/Kconfig2
-rw-r--r--fs/ceph/auth_x.c5
-rw-r--r--fs/ceph/caps.c129
-rw-r--r--fs/ceph/crush/mapper.c41
-rw-r--r--fs/ceph/debugfs.c2
-rw-r--r--fs/ceph/dir.c13
-rw-r--r--fs/ceph/file.c2
-rw-r--r--fs/ceph/inode.c27
-rw-r--r--fs/ceph/mds_client.c79
-rw-r--r--fs/ceph/mds_client.h7
-rw-r--r--fs/ceph/messenger.c75
-rw-r--r--fs/ceph/mon_client.c11
-rw-r--r--fs/ceph/osd_client.c9
-rw-r--r--fs/ceph/osdmap.c28
-rw-r--r--fs/ceph/super.c4
15 files changed, 275 insertions, 159 deletions
diff --git a/fs/ceph/Kconfig b/fs/ceph/Kconfig
index 04b8280582a..bc87b9c1d27 100644
--- a/fs/ceph/Kconfig
+++ b/fs/ceph/Kconfig
@@ -2,7 +2,7 @@ config CEPH_FS
tristate "Ceph distributed file system (EXPERIMENTAL)"
depends on INET && EXPERIMENTAL
select LIBCRC32C
- select CONFIG_CRYPTO_AES
+ select CRYPTO_AES
help
Choose Y or M here to include support for mounting the
experimental Ceph distributed file system. Ceph is an extremely
diff --git a/fs/ceph/auth_x.c b/fs/ceph/auth_x.c
index 83d4d2785ff..6d44053ecff 100644
--- a/fs/ceph/auth_x.c
+++ b/fs/ceph/auth_x.c
@@ -493,7 +493,7 @@ static int ceph_x_handle_reply(struct ceph_auth_client *ac, int result,
return -EAGAIN;
}
- op = le32_to_cpu(head->op);
+ op = le16_to_cpu(head->op);
result = le32_to_cpu(head->result);
dout("handle_reply op %d result %d\n", op, result);
switch (op) {
@@ -613,6 +613,9 @@ static void ceph_x_destroy(struct ceph_auth_client *ac)
remove_ticket_handler(ac, th);
}
+ if (xi->auth_authorizer.buf)
+ ceph_buffer_put(xi->auth_authorizer.buf);
+
kfree(ac->private);
ac->private = NULL;
}
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index ae3e3a30644..b81be9a5648 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -244,8 +244,14 @@ static struct ceph_cap *get_cap(struct ceph_cap_reservation *ctx)
struct ceph_cap *cap = NULL;
/* temporary, until we do something about cap import/export */
- if (!ctx)
- return kmem_cache_alloc(ceph_cap_cachep, GFP_NOFS);
+ if (!ctx) {
+ cap = kmem_cache_alloc(ceph_cap_cachep, GFP_NOFS);
+ if (cap) {
+ caps_use_count++;
+ caps_total_count++;
+ }
+ return cap;
+ }
spin_lock(&caps_list_lock);
dout("get_cap ctx=%p (%d) %d = %d used + %d resv + %d avail\n",
@@ -621,7 +627,7 @@ retry:
if (fmode >= 0)
__ceph_get_fmode(ci, fmode);
spin_unlock(&inode->i_lock);
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
return 0;
}
@@ -981,6 +987,46 @@ static int send_cap_msg(struct ceph_mds_session *session,
return 0;
}
+static void __queue_cap_release(struct ceph_mds_session *session,
+ u64 ino, u64 cap_id, u32 migrate_seq,
+ u32 issue_seq)
+{
+ struct ceph_msg *msg;
+ struct ceph_mds_cap_release *head;
+ struct ceph_mds_cap_item *item;
+
+ spin_lock(&session->s_cap_lock);
+ BUG_ON(!session->s_num_cap_releases);
+ msg = list_first_entry(&session->s_cap_releases,
+ struct ceph_msg, list_head);
+
+ dout(" adding %llx release to mds%d msg %p (%d left)\n",
+ ino, session->s_mds, msg, session->s_num_cap_releases);
+
+ BUG_ON(msg->front.iov_len + sizeof(*item) > PAGE_CACHE_SIZE);
+ head = msg->front.iov_base;
+ head->num = cpu_to_le32(le32_to_cpu(head->num) + 1);
+ item = msg->front.iov_base + msg->front.iov_len;
+ item->ino = cpu_to_le64(ino);
+ item->cap_id = cpu_to_le64(cap_id);
+ item->migrate_seq = cpu_to_le32(migrate_seq);
+ item->seq = cpu_to_le32(issue_seq);
+
+ session->s_num_cap_releases--;
+
+ msg->front.iov_len += sizeof(*item);
+ if (le32_to_cpu(head->num) == CEPH_CAPS_PER_RELEASE) {
+ dout(" release msg %p full\n", msg);
+ list_move_tail(&msg->list_head, &session->s_cap_releases_done);
+ } else {
+ dout(" release msg %p at %d/%d (%d)\n", msg,
+ (int)le32_to_cpu(head->num),
+ (int)CEPH_CAPS_PER_RELEASE,
+ (int)msg->front.iov_len);
+ }
+ spin_unlock(&session->s_cap_lock);
+}
+
/*
* Queue cap releases when an inode is dropped from our cache. Since
* inode is about to be destroyed, there is no need for i_lock.
@@ -994,41 +1040,9 @@ void ceph_queue_caps_release(struct inode *inode)
while (p) {
struct ceph_cap *cap = rb_entry(p, struct ceph_cap, ci_node);
struct ceph_mds_session *session = cap->session;
- struct ceph_msg *msg;
- struct ceph_mds_cap_release *head;
- struct ceph_mds_cap_item *item;
- spin_lock(&session->s_cap_lock);
- BUG_ON(!session->s_num_cap_releases);
- msg = list_first_entry(&session->s_cap_releases,
- struct ceph_msg, list_head);
-
- dout(" adding %p release to mds%d msg %p (%d left)\n",
- inode, session->s_mds, msg, session->s_num_cap_releases);
-
- BUG_ON(msg->front.iov_len + sizeof(*item) > PAGE_CACHE_SIZE);
- head = msg->front.iov_base;
- head->num = cpu_to_le32(le32_to_cpu(head->num) + 1);
- item = msg->front.iov_base + msg->front.iov_len;
- item->ino = cpu_to_le64(ceph_ino(inode));
- item->cap_id = cpu_to_le64(cap->cap_id);
- item->migrate_seq = cpu_to_le32(cap->mseq);
- item->seq = cpu_to_le32(cap->issue_seq);
-
- session->s_num_cap_releases--;
-
- msg->front.iov_len += sizeof(*item);
- if (le32_to_cpu(head->num) == CEPH_CAPS_PER_RELEASE) {
- dout(" release msg %p full\n", msg);
- list_move_tail(&msg->list_head,
- &session->s_cap_releases_done);
- } else {
- dout(" release msg %p at %d/%d (%d)\n", msg,
- (int)le32_to_cpu(head->num),
- (int)CEPH_CAPS_PER_RELEASE,
- (int)msg->front.iov_len);
- }
- spin_unlock(&session->s_cap_lock);
+ __queue_cap_release(session, ceph_ino(inode), cap->cap_id,
+ cap->mseq, cap->issue_seq);
p = rb_next(p);
__ceph_remove_cap(cap);
}
@@ -1167,7 +1181,7 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
}
if (wake)
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
return delayed;
}
@@ -2139,7 +2153,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had)
else if (flushsnaps)
ceph_flush_snaps(ci);
if (wake)
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
if (put)
iput(inode);
}
@@ -2215,7 +2229,7 @@ void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr,
iput(inode);
} else if (complete_capsnap) {
ceph_flush_snaps(ci);
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
}
if (drop_capsnap)
iput(inode);
@@ -2391,7 +2405,7 @@ static void handle_cap_grant(struct inode *inode, struct ceph_mds_caps *grant,
if (queue_invalidate)
ceph_queue_invalidate(inode);
if (wake)
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
if (check_caps == 1)
ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_AUTHONLY,
@@ -2446,7 +2460,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
struct ceph_inode_info,
i_flushing_item)->vfs_inode);
mdsc->num_cap_flushing--;
- wake_up(&mdsc->cap_flushing_wq);
+ wake_up_all(&mdsc->cap_flushing_wq);
dout(" inode %p now !flushing\n", inode);
if (ci->i_dirty_caps == 0) {
@@ -2458,7 +2472,7 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
}
}
spin_unlock(&mdsc->cap_dirty_lock);
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
out:
spin_unlock(&inode->i_lock);
@@ -2655,7 +2669,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
struct ceph_mds_caps *h;
int mds = session->s_mds;
int op;
- u32 seq;
+ u32 seq, mseq;
struct ceph_vino vino;
u64 cap_id;
u64 size, max_size;
@@ -2675,6 +2689,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
vino.snap = CEPH_NOSNAP;
cap_id = le64_to_cpu(h->cap_id);
seq = le32_to_cpu(h->seq);
+ mseq = le32_to_cpu(h->migrate_seq);
size = le64_to_cpu(h->size);
max_size = le64_to_cpu(h->max_size);
@@ -2689,6 +2704,18 @@ void ceph_handle_caps(struct ceph_mds_session *session,
vino.snap, inode);
if (!inode) {
dout(" i don't have ino %llx\n", vino.ino);
+
+ if (op == CEPH_CAP_OP_IMPORT)
+ __queue_cap_release(session, vino.ino, cap_id,
+ mseq, seq);
+
+ /*
+ * send any full release message to try to move things
+ * along for the mds (who clearly thinks we still have this
+ * cap).
+ */
+ ceph_add_cap_releases(mdsc, session, -1);
+ ceph_send_cap_releases(mdsc, session);
goto done;
}
@@ -2714,7 +2741,7 @@ void ceph_handle_caps(struct ceph_mds_session *session,
spin_lock(&inode->i_lock);
cap = __get_cap_for_mds(ceph_inode(inode), mds);
if (!cap) {
- dout("no cap on %p ino %llx.%llx from mds%d, releasing\n",
+ dout(" no cap on %p ino %llx.%llx from mds%d\n",
inode, ceph_ino(inode), ceph_snap(inode), mds);
spin_unlock(&inode->i_lock);
goto done;
@@ -2865,18 +2892,19 @@ int ceph_encode_inode_release(void **p, struct inode *inode,
struct ceph_inode_info *ci = ceph_inode(inode);
struct ceph_cap *cap;
struct ceph_mds_request_release *rel = *p;
+ int used, dirty;
int ret = 0;
- int used = 0;
spin_lock(&inode->i_lock);
used = __ceph_caps_used(ci);
+ dirty = __ceph_caps_dirty(ci);
- dout("encode_inode_release %p mds%d used %s drop %s unless %s\n", inode,
- mds, ceph_cap_string(used), ceph_cap_string(drop),
+ dout("encode_inode_release %p mds%d used|dirty %s drop %s unless %s\n",
+ inode, mds, ceph_cap_string(used|dirty), ceph_cap_string(drop),
ceph_cap_string(unless));
- /* only drop unused caps */
- drop &= ~used;
+ /* only drop unused, clean caps */
+ drop &= ~(used | dirty);
cap = __get_cap_for_mds(ci, mds);
if (cap && __cap_is_valid(cap)) {
@@ -2956,6 +2984,7 @@ int ceph_encode_dentry_release(void **p, struct dentry *dentry,
memcpy(*p, dentry->d_name.name, dentry->d_name.len);
*p += dentry->d_name.len;
rel->dname_seq = cpu_to_le32(di->lease_seq);
+ __ceph_mdsc_drop_dentry_lease(dentry);
}
spin_unlock(&dentry->d_lock);
return ret;
diff --git a/fs/ceph/crush/mapper.c b/fs/ceph/crush/mapper.c
index 9ba54efb654..a4eec133258 100644
--- a/fs/ceph/crush/mapper.c
+++ b/fs/ceph/crush/mapper.c
@@ -238,7 +238,7 @@ static int bucket_straw_choose(struct crush_bucket_straw *bucket,
static int crush_bucket_choose(struct crush_bucket *in, int x, int r)
{
- dprintk("choose %d x=%d r=%d\n", in->id, x, r);
+ dprintk(" crush_bucket_choose %d x=%d r=%d\n", in->id, x, r);
switch (in->alg) {
case CRUSH_BUCKET_UNIFORM:
return bucket_uniform_choose((struct crush_bucket_uniform *)in,
@@ -264,7 +264,7 @@ static int crush_bucket_choose(struct crush_bucket *in, int x, int r)
*/
static int is_out(struct crush_map *map, __u32 *weight, int item, int x)
{
- if (weight[item] >= 0x1000)
+ if (weight[item] >= 0x10000)
return 0;
if (weight[item] == 0)
return 1;
@@ -305,7 +305,9 @@ static int crush_choose(struct crush_map *map,
int itemtype;
int collide, reject;
const int orig_tries = 5; /* attempts before we fall back to search */
- dprintk("choose bucket %d x %d outpos %d\n", bucket->id, x, outpos);
+
+ dprintk("CHOOSE%s bucket %d x %d outpos %d numrep %d\n", recurse_to_leaf ? "_LEAF" : "",
+ bucket->id, x, outpos, numrep);
for (rep = outpos; rep < numrep; rep++) {
/* keep trying until we get a non-out, non-colliding item */
@@ -366,6 +368,7 @@ static int crush_choose(struct crush_map *map,
BUG_ON(item >= 0 ||
(-1-item) >= map->max_buckets);
in = map->buckets[-1-item];
+ retry_bucket = 1;
continue;
}
@@ -377,15 +380,25 @@ static int crush_choose(struct crush_map *map,
}
}
- if (recurse_to_leaf &&
- item < 0 &&
- crush_choose(map, map->buckets[-1-item],
- weight,
- x, outpos+1, 0,
- out2, outpos,
- firstn, 0, NULL) <= outpos) {
- reject = 1;
- } else {
+ reject = 0;
+ if (recurse_to_leaf) {
+ if (item < 0) {
+ if (crush_choose(map,
+ map->buckets[-1-item],
+ weight,
+ x, outpos+1, 0,
+ out2, outpos,
+ firstn, 0,
+ NULL) <= outpos)
+ /* didn't get leaf */
+ reject = 1;
+ } else {
+ /* we already have a leaf! */
+ out2[outpos] = item;
+ }
+ }
+
+ if (!reject) {
/* out? */
if (itemtype == 0)
reject = is_out(map, weight,
@@ -424,12 +437,12 @@ reject:
continue;
}
- dprintk("choose got %d\n", item);
+ dprintk("CHOOSE got %d\n", item);
out[outpos] = item;
outpos++;
}
- dprintk("choose returns %d\n", outpos);
+ dprintk("CHOOSE returns %d\n", outpos);
return outpos;
}
diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c
index 3be33fb066c..f2f5332ddbb 100644
--- a/fs/ceph/debugfs.c
+++ b/fs/ceph/debugfs.c
@@ -261,7 +261,7 @@ static int osdc_show(struct seq_file *s, void *pp)
static int caps_show(struct seq_file *s, void *p)
{
- struct ceph_client *client = p;
+ struct ceph_client *client = s->private;
int total, avail, used, reserved, min;
ceph_reservation_status(client, &total, &avail, &used, &reserved, &min);
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index f85719310db..f94ed3c7f6a 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -266,6 +266,7 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
spin_lock(&inode->i_lock);
if ((filp->f_pos == 2 || fi->dentry) &&
!ceph_test_opt(client, NOASYNCREADDIR) &&
+ ceph_snap(inode) != CEPH_SNAPDIR &&
(ci->i_ceph_flags & CEPH_I_COMPLETE) &&
__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) {
err = __dcache_readdir(filp, dirent, filldir);
@@ -1013,18 +1014,22 @@ out_touch:
/*
* When a dentry is released, clear the dir I_COMPLETE if it was part
- * of the current dir gen.
+ * of the current dir gen or if this is in the snapshot namespace.
*/
static void ceph_dentry_release(struct dentry *dentry)
{
struct ceph_dentry_info *di = ceph_dentry(dentry);
struct inode *parent_inode = dentry->d_parent->d_inode;
+ u64 snapid = ceph_snap(parent_inode);
- if (parent_inode) {
+ dout("dentry_release %p parent %p\n", dentry, parent_inode);
+
+ if (parent_inode && snapid != CEPH_SNAPDIR) {
struct ceph_inode_info *ci = ceph_inode(parent_inode);
spin_lock(&parent_inode->i_lock);
- if (ci->i_shared_gen == di->lease_shared_gen) {
+ if (ci->i_shared_gen == di->lease_shared_gen ||
+ snapid <= CEPH_MAXSNAP) {
dout(" clearing %p complete (d_release)\n",
parent_inode);
ci->i_ceph_flags &= ~CEPH_I_COMPLETE;
@@ -1241,7 +1246,9 @@ struct dentry_operations ceph_dentry_ops = {
struct dentry_operations ceph_snapdir_dentry_ops = {
.d_revalidate = ceph_snapdir_d_revalidate,
+ .d_release = ceph_dentry_release,
};
struct dentry_operations ceph_snap_dentry_ops = {
+ .d_release = ceph_dentry_release,
};
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 6251a1574b9..7c08698fad3 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -265,7 +265,7 @@ int ceph_release(struct inode *inode, struct file *file)
kmem_cache_free(ceph_file_cachep, cf);
/* wake up anyone waiting for caps on this inode */
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
return 0;
}
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 226f5a50d36..389f9dbd994 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -827,7 +827,7 @@ static void ceph_set_dentry_offset(struct dentry *dn)
spin_lock(&dcache_lock);
spin_lock(&dn->d_lock);
- list_move_tail(&dir->d_subdirs, &dn->d_u.d_child);
+ list_move(&dn->d_u.d_child, &dir->d_subdirs);
dout("set_dentry_offset %p %lld (%p %p)\n", dn, di->offset,
dn->d_u.d_child.prev, dn->d_u.d_child.next);
spin_unlock(&dn->d_lock);
@@ -854,8 +854,8 @@ static struct dentry *splice_dentry(struct dentry *dn, struct inode *in,
d_drop(dn);
realdn = d_materialise_unique(dn, in);
if (IS_ERR(realdn)) {
- pr_err("splice_dentry error %p inode %p ino %llx.%llx\n",
- dn, in, ceph_vinop(in));
+ pr_err("splice_dentry error %ld %p inode %p ino %llx.%llx\n",
+ PTR_ERR(realdn), dn, in, ceph_vinop(in));
if (prehash)
*prehash = false; /* don't rehash on error */
dn = realdn; /* note realdn contains the error */
@@ -1199,8 +1199,10 @@ retry_lookup:
goto out;
}
err = ceph_init_dentry(dn);
- if (err < 0)
+ if (err < 0) {
+ dput(dn);
goto out;
+ }
} else if (dn->d_inode &&
(ceph_ino(dn->d_inode) != vino.ino ||
ceph_snap(dn->d_inode) != vino.snap)) {
@@ -1234,18 +1236,23 @@ retry_lookup:
goto out;
}
dn = splice_dentry(dn, in, NULL);
+ if (IS_ERR(dn))
+ dn = NULL;
}
if (fill_inode(in, &rinfo->dir_in[i], NULL, session,
req->r_request_started, -1,
&req->r_caps_reservation) < 0) {
pr_err("fill_inode badness on %p\n", in);
- dput(dn);
- continue;
+ goto next_item;
}
- update_dentry_lease(dn, rinfo->dir_dlease[i],
- req->r_session, req->r_request_started);
- dput(dn);
+ if (dn)
+ update_dentry_lease(dn, rinfo->dir_dlease[i],
+ req->r_session,
+ req->r_request_started);
+next_item:
+ if (dn)
+ dput(dn);
}
req->r_did_prepopulate = true;
@@ -1494,7 +1501,7 @@ retry:
if (wrbuffer_refs == 0)
ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL);
if (wake)
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
}
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index b49f12822cb..dd440bd438a 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -868,7 +868,7 @@ static int wake_up_session_cb(struct inode *inode, struct ceph_cap *cap,
{
struct ceph_inode_info *ci = ceph_inode(inode);
- wake_up(&ci->i_cap_wq);
+ wake_up_all(&ci->i_cap_wq);
if (arg) {
spin_lock(&inode->i_lock);
ci->i_wanted_max_size = 0;
@@ -1066,9 +1066,9 @@ static int trim_caps(struct ceph_mds_client *mdsc,
*
* Called under s_mutex.
*/
-static int add_cap_releases(struct ceph_mds_client *mdsc,
- struct ceph_mds_session *session,
- int extra)
+int ceph_add_cap_releases(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session,
+ int extra)
{
struct ceph_msg *msg;
struct ceph_mds_cap_release *head;
@@ -1176,8 +1176,8 @@ static int check_cap_flush(struct ceph_mds_client *mdsc, u64 want_flush_seq)
/*
* called under s_mutex
*/
-static void send_cap_releases(struct ceph_mds_client *mdsc,
- struct ceph_mds_session *session)
+void ceph_send_cap_releases(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session)
{
struct ceph_msg *msg;
@@ -1514,6 +1514,9 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc,
ceph_encode_filepath(&p, end, ino1, path1);
ceph_encode_filepath(&p, end, ino2, path2);
+ /* make note of release offset, in case we need to replay */
+ req->r_request_release_offset = p - msg->front.iov_base;
+
/* cap releases */
releases = 0;
if (req->r_inode_drop)
@@ -1561,7 +1564,7 @@ static void complete_request(struct ceph_mds_client *mdsc,
if (req->r_callback)
req->r_callback(mdsc, req);
else
- complete(&req->r_completion);
+ complete_all(&req->r_completion);
}
/*
@@ -1580,6 +1583,32 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc,
dout("prepare_send_request %p tid %lld %s (attempt %d)\n", req,
req->r_tid, ceph_mds_op_name(req->r_op), req->r_attempts);
+ if (req->r_got_unsafe) {
+ /*
+ * Replay. Do not regenerate message (and rebuild
+ * paths, etc.); just use the original message.
+ * Rebuilding paths will break for renames because
+ * d_move mangles the src name.
+ */
+ msg = req->r_request;
+ rhead = msg->front.iov_base;
+
+ flags = le32_to_cpu(rhead->flags);
+ flags |= CEPH_MDS_FLAG_REPLAY;
+ rhead->flags = cpu_to_le32(flags);
+
+ if (req->r_target_inode)
+ rhead->ino = cpu_to_le64(ceph_ino(req->r_target_inode));
+
+ rhead->num_retry = req->r_attempts - 1;
+
+ /* remove cap/dentry releases from message */
+ rhead->num_releases = 0;
+ msg->hdr.front_len = cpu_to_le32(req->r_request_release_offset);
+ msg->front.iov_len = req->r_request_release_offset;
+ return 0;
+ }
+
if (req->r_request) {
ceph_msg_put(req->r_request);
req->r_request = NULL;
@@ -1601,13 +1630,9 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc,
rhead->flags = cpu_to_le32(flags);
rhead->num_fwd = req->r_num_fwd;
rhead->num_retry = req->r_attempts - 1;
+ rhead->ino = 0;
dout(" r_locked_dir = %p\n", req->r_locked_dir);
-
- if (req->r_target_inode && req->r_got_unsafe)
- rhead->ino = cpu_to_le64(ceph_ino(req->r_target_inode));
- else
- rhead->ino = 0;
return 0;
}
@@ -1907,7 +1932,7 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
if (head->safe) {
req->r_got_safe = true;
__unregister_request(mdsc, req);
- complete(&req->r_safe_completion);
+ complete_all(&req->r_safe_completion);
if (req->r_got_unsafe) {
/*
@@ -1922,7 +1947,7 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
/* last unsafe request during umount? */
if (mdsc->stopping && !__get_oldest_req(mdsc))
- complete(&mdsc->safe_umount_waiters);
+ complete_all(&mdsc->safe_umount_waiters);
mutex_unlock(&mdsc->mutex);
goto out;
}
@@ -1980,7 +2005,7 @@ out_err:
}
mutex_unlock(&mdsc->mutex);
- add_cap_releases(mdsc, req->r_session, -1);
+ ceph_add_cap_releases(mdsc, req->r_session, -1);
mutex_unlock(&session->s_mutex);
/* kick calling process */
@@ -2101,7 +2126,7 @@ static void handle_session(struct ceph_mds_session *session,
pr_info("mds%d reconnect denied\n", session->s_mds);
remove_session_caps(session);
wake = 1; /* for good measure */
- complete(&mdsc->session_close_waiters);
+ complete_all(&mdsc->session_close_waiters);
kick_requests(mdsc, mds);
break;
@@ -2433,6 +2458,7 @@ static void handle_lease(struct ceph_mds_client *mdsc,
struct ceph_dentry_info *di;
int mds = session->s_mds;
struct ceph_mds_lease *h = msg->front.iov_base;
+ u32 seq;
struct ceph_vino vino;
int mask;
struct qstr dname;
@@ -2446,6 +2472,7 @@ static void handle_lease(struct ceph_mds_client *mdsc,
vino.ino = le64_to_cpu(h->ino);
vino.snap = CEPH_NOSNAP;
mask = le16_to_cpu(h->mask);
+ seq = le32_to_cpu(h->seq);
dname.name = (void *)h + sizeof(*h) + sizeof(u32);
dname.len = msg->front.iov_len - sizeof(*h) - sizeof(u32);
if (dname.len != get_unaligned_le32(h+1))
@@ -2456,8 +2483,9 @@ static void handle_lease(struct ceph_mds_client *mdsc,
/* lookup inode */
inode = ceph_find_inode(sb, vino);
- dout("handle_lease '%s', mask %d, ino %llx %p\n",
- ceph_lease_op_name(h->action), mask, vino.ino, inode);
+ dout("handle_lease %s, mask %d, ino %llx %p %.*s\n",
+ ceph_lease_op_name(h->action), mask, vino.ino, inode,
+ dname.len, dname.name);
if (inode == NULL) {
dout("handle_lease no inode %llx\n", vino.ino);
goto release;
@@ -2482,7 +2510,8 @@ static void handle_lease(struct ceph_mds_client *mdsc,
switch (h->action) {
case CEPH_MDS_LEASE_REVOKE:
if (di && di->lease_session == session) {
- h->seq = cpu_to_le32(di->lease_seq);
+ if (ceph_seq_cmp(di->lease_seq, seq) > 0)
+ h->seq = cpu_to_le32(di->lease_seq);
__ceph_mdsc_drop_dentry_lease(dentry);
}
release = 1;
@@ -2496,7 +2525,7 @@ static void handle_lease(struct ceph_mds_client *mdsc,
unsigned long duration =
le32_to_cpu(h->duration_ms) * HZ / 1000;
- di->lease_seq = le32_to_cpu(h->seq);
+ di->lease_seq = seq;
dentry->d_time = di->lease_renew_from + duration;
di->lease_renew_after = di->lease_renew_from +
(duration >> 1);
@@ -2686,10 +2715,10 @@ static void delayed_work(struct work_struct *work)
send_renew_caps(mdsc, s);
else
ceph_con_keepalive(&s->s_con);
- add_cap_releases(mdsc, s, -1);
+ ceph_add_cap_releases(mdsc, s, -1);
if (s->s_state == CEPH_MDS_SESSION_OPEN ||
s->s_state == CEPH_MDS_SESSION_HUNG)
- send_cap_releases(mdsc, s);
+ ceph_send_cap_releases(mdsc, s);
mutex_unlock(&s->s_mutex);
ceph_put_mds_session(s);
@@ -2779,6 +2808,12 @@ void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc)
drop_leases(mdsc);
ceph_flush_dirty_caps(mdsc);
wait_requests(mdsc);
+
+ /*
+ * wait for reply handlers to drop their request refs and
+ * their inode/dcache refs
+ */
+ ceph_msgr_flush();
}
/*
diff --git a/fs/ceph/mds_client.h b/fs/ceph/mds_client.h
index d9936c4f121..952410c60d0 100644
--- a/fs/ceph/mds_client.h
+++ b/fs/ceph/mds_client.h
@@ -188,6 +188,7 @@ struct ceph_mds_request {
int r_old_inode_drop, r_old_inode_unless;
struct ceph_msg *r_request; /* original request */
+ int r_request_release_offset;
struct ceph_msg *r_reply;
struct ceph_mds_reply_info_parsed r_reply_info;
int r_err;
@@ -322,6 +323,12 @@ static inline void ceph_mdsc_put_request(struct ceph_mds_request *req)
kref_put(&req->r_kref, ceph_mdsc_release_request);
}
+extern int ceph_add_cap_releases(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session,
+ int extra);
+extern void ceph_send_cap_releases(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session);
+
extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc);
extern char *ceph_mdsc_build_path(struct dentry *dentry, int *plen, u64 *base,
diff --git a/fs/ceph/messenger.c b/fs/ceph/messenger.c
index 64b8b1f7863..15167b2daa5 100644
--- a/fs/ceph/messenger.c
+++ b/fs/ceph/messenger.c
@@ -43,7 +43,8 @@ static void ceph_fault(struct ceph_connection *con);
* nicely render a sockaddr as a string.
*/
#define MAX_ADDR_STR 20
-static char addr_str[MAX_ADDR_STR][40];
+#define MAX_ADDR_STR_LEN 60
+static char addr_str[MAX_ADDR_STR][MAX_ADDR_STR_LEN];
static DEFINE_SPINLOCK(addr_str_lock);
static int last_addr_str;
@@ -52,7 +53,6 @@ const char *pr_addr(const struct sockaddr_storage *ss)
int i;
char *s;
struct sockaddr_in *in4 = (void *)ss;
- unsigned char *quad = (void *)&in4->sin_addr.s_addr;
struct sockaddr_in6 *in6 = (void *)ss;
spin_lock(&addr_str_lock);
@@ -64,25 +64,13 @@ const char *pr_addr(const struct sockaddr_storage *ss)
switch (ss->ss_family) {
case AF_INET:
- sprintf(s, "%u.%u.%u.%u:%u",
- (unsigned int)quad[0],
- (unsigned int)quad[1],
- (unsigned int)quad[2],
- (unsigned int)quad[3],
- (unsigned int)ntohs(in4->sin_port));
+ snprintf(s, MAX_ADDR_STR_LEN, "%pI4:%u", &in4->sin_addr,
+ (unsigned int)ntohs(in4->sin_port));
break;
case AF_INET6:
- sprintf(s, "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x:%u",
- in6->sin6_addr.s6_addr16[0],
- in6->sin6_addr.s6_addr16[1],
- in6->sin6_addr.s6_addr16[2],
- in6->sin6_addr.s6_addr16[3],
- in6->sin6_addr.s6_addr16[4],
- in6->sin6_addr.s6_addr16[5],
- in6->sin6_addr.s6_addr16[6],
- in6->sin6_addr.s6_addr16[7],
- (unsigned int)ntohs(in6->sin6_port));
+ snprintf(s, MAX_ADDR_STR_LEN, "[%pI6c]:%u", &in6->sin6_addr,
+ (unsigned int)ntohs(in6->sin6_port));
break;
default:
@@ -215,12 +203,13 @@ static void set_sock_callbacks(struct socket *sock,
*/
static struct socket *ceph_tcp_connect(struct ceph_connection *con)
{
- struct sockaddr *paddr = (struct sockaddr *)&con->peer_addr.in_addr;
+ struct sockaddr_storage *paddr = &con->peer_addr.in_addr;
struct socket *sock;
int ret;
BUG_ON(con->sock);
- ret = sock_create_kern(AF_INET, SOCK_STREAM, IPPROTO_TCP, &sock);
+ ret = sock_create_kern(con->peer_addr.in_addr.ss_family, SOCK_STREAM,
+ IPPROTO_TCP, &sock);
if (ret)
return ERR_PTR(ret);
con->sock = sock;
@@ -234,7 +223,8 @@ static struct socket *ceph_tcp_connect(struct ceph_connection *con)
dout("connect %s\n", pr_addr(&con->peer_addr.in_addr));
- ret = sock->ops->connect(sock, paddr, sizeof(*paddr), O_NONBLOCK);
+ ret = sock->ops->connect(sock, (struct sockaddr *)paddr, sizeof(*paddr),
+ O_NONBLOCK);
if (ret == -EINPROGRESS) {
dout("connect %s EINPROGRESS sk_state = %u\n",
pr_addr(&con->peer_addr.in_addr),
@@ -657,7 +647,7 @@ static void prepare_write_connect(struct ceph_messenger *msgr,
dout("prepare_write_connect %p cseq=%d gseq=%d proto=%d\n", con,
con->connect_seq, global_seq, proto);
- con->out_connect.features = CEPH_FEATURE_SUPPORTED_CLIENT;
+ con->out_connect.features = cpu_to_le64(CEPH_FEATURE_SUPPORTED_CLIENT);
con->out_connect.host_type = cpu_to_le32(CEPH_ENTITY_TYPE_CLIENT);
con->out_connect.connect_seq = cpu_to_le32(con->connect_seq);
con->out_connect.global_seq = cpu_to_le32(global_seq);
@@ -1009,19 +999,32 @@ int ceph_parse_ips(const char *c, const char *end,
struct sockaddr_in *in4 = (void *)ss;
struct sockaddr_in6 *in6 = (void *)ss;
int port;
+ char delim = ',';
+
+ if (*p == '[') {
+ delim = ']';
+ p++;
+ }
memset(ss, 0, sizeof(*ss));
if (in4_pton(p, end - p, (u8 *)&in4->sin_addr.s_addr,
- ',', &ipend)) {
+ delim, &ipend))
ss->ss_family = AF_INET;
- } else if (in6_pton(p, end - p, (u8 *)&in6->sin6_addr.s6_addr,
- ',', &ipend)) {
+ else if (in6_pton(p, end - p, (u8 *)&in6->sin6_addr.s6_addr,
+ delim, &ipend))
ss->ss_family = AF_INET6;
- } else {
+ else
goto bad;
- }
p = ipend;
+ if (delim == ']') {
+ if (*p != ']') {
+ dout("missing matching ']'\n");
+ goto bad;
+ }
+ p++;
+ }
+
/* port? */
if (p < end && *p == ':') {
port = 0;
@@ -1055,7 +1058,7 @@ int ceph_parse_ips(const char *c, const char *end,
return 0;
bad:
- pr_err("parse_ips bad ip '%s'\n", c);
+ pr_err("parse_ips bad ip '%.*s'\n", (int)(end - c), c);
return -EINVAL;
}
@@ -1396,10 +1399,12 @@ static int read_partial_message(struct ceph_connection *con)
if (!con->in_msg) {
dout("got hdr type %d front %d data %d\n", con->in_hdr.type,
con->in_hdr.front_len, con->in_hdr.data_len);
+ skip = 0;
con->in_msg = ceph_alloc_msg(con, &con->in_hdr, &skip);
if (skip) {
/* skip this message */
dout("alloc_msg said skip message\n");
+ BUG_ON(con->in_msg);
con->in_base_pos = -front_len - middle_len - data_len -
sizeof(m->footer);
con->in_tag = CEPH_MSGR_TAG_READY;
@@ -2013,20 +2018,20 @@ void ceph_con_revoke(struct ceph_connection *con, struct ceph_msg *msg)
{
mutex_lock(&con->mutex);
if (!list_empty(&msg->list_head)) {
- dout("con_revoke %p msg %p\n", con, msg);
+ dout("con_revoke %p msg %p - was on queue\n", con, msg);
list_del_init(&msg->list_head);
ceph_msg_put(msg);
msg->hdr.seq = 0;
- if (con->out_msg == msg) {
- ceph_msg_put(con->out_msg);
- con->out_msg = NULL;
- }
+ }
+ if (con->out_msg == msg) {
+ dout("con_revoke %p msg %p - was sending\n", con, msg);
+ con->out_msg = NULL;
if (con->out_kvec_is_msg) {
con->out_skip = con->out_kvec_bytes;
con->out_kvec_is_msg = false;
}
- } else {
- dout("con_revoke %p msg %p - not queued (sent?)\n", con, msg);
+ ceph_msg_put(msg);
+ msg->hdr.seq = 0;
}
mutex_unlock(&con->mutex);
}
diff --git a/fs/ceph/mon_client.c b/fs/ceph/mon_client.c
index 21c62e9b7d1..54fe01c5070 100644
--- a/fs/ceph/mon_client.c
+++ b/fs/ceph/mon_client.c
@@ -345,7 +345,7 @@ static void ceph_monc_handle_map(struct ceph_mon_client *monc,
out:
mutex_unlock(&monc->mutex);
- wake_up(&client->auth_wq);
+ wake_up_all(&client->auth_wq);
}
/*
@@ -400,6 +400,8 @@ static void release_generic_request(struct kref *kref)
ceph_msg_put(req->reply);
if (req->request)
ceph_msg_put(req->request);
+
+ kfree(req);
}
static void put_generic_request(struct ceph_mon_generic_request *req)
@@ -460,7 +462,7 @@ static void handle_statfs_reply(struct ceph_mon_client *monc,
}
mutex_unlock(&monc->mutex);
if (req) {
- complete(&req->completion);
+ complete_all(&req->completion);
put_generic_request(req);
}
return;
@@ -716,14 +718,15 @@ static void handle_auth_reply(struct ceph_mon_client *monc,
monc->m_auth->front_max);
if (ret < 0) {
monc->client->auth_err = ret;
- wake_up(&monc->client->auth_wq);
+ wake_up_all(&monc->client->auth_wq);
} else if (ret > 0) {
__send_prepared_auth_request(monc, ret);
} else if (!was_auth && monc->auth->ops->is_authenticated(monc->auth)) {
dout("authenticated, starting session\n");
monc->client->msgr->inst.name.type = CEPH_ENTITY_TYPE_CLIENT;
- monc->client->msgr->inst.name.num = monc->auth->global_id;
+ monc->client->msgr->inst.name.num =
+ cpu_to_le64(monc->auth->global_id);
__send_subscribe(monc);
__resend_generic_request(monc);
diff --git a/fs/ceph/osd_client.c b/fs/ceph/osd_client.c
index d25b4add85b..e3852234789 100644
--- a/fs/ceph/osd_client.c
+++ b/fs/ceph/osd_client.c
@@ -862,12 +862,12 @@ static void handle_reply(struct ceph_osd_client *osdc, struct ceph_msg *msg,
if (req->r_callback)
req->r_callback(req, msg);
else
- complete(&req->r_completion);
+ complete_all(&req->r_completion);
if (flags & CEPH_OSD_FLAG_ONDISK) {
if (req->r_safe_callback)
req->r_safe_callback(req, msg);
- complete(&req->r_safe_completion); /* fsync waiter */
+ complete_all(&req->r_safe_completion); /* fsync waiter */
}
done:
@@ -1083,7 +1083,7 @@ done:
if (newmap)
kick_requests(osdc, NULL);
up_read(&osdc->map_sem);
- wake_up(&osdc->client->auth_wq);
+ wake_up_all(&osdc->client->auth_wq);
return;
bad:
@@ -1344,7 +1344,7 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
int type = le16_to_cpu(msg->hdr.type);
if (!osd)
- return;
+ goto out;
osdc = osd->o_osdc;
switch (type) {
@@ -1359,6 +1359,7 @@ static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
pr_err("received unknown message type %d %s\n", type,
ceph_msg_type_name(type));
}
+out:
ceph_msg_put(msg);
}
diff --git a/fs/ceph/osdmap.c b/fs/ceph/osdmap.c
index ddc656fb5c0..416d46adbf8 100644
--- a/fs/ceph/osdmap.c
+++ b/fs/ceph/osdmap.c
@@ -568,6 +568,7 @@ struct ceph_osdmap *osdmap_decode(void **p, void *end)
if (ev > CEPH_PG_POOL_VERSION) {
pr_warning("got unknown v %d > %d of ceph_pg_pool\n",
ev, CEPH_PG_POOL_VERSION);
+ kfree(pi);
goto bad;
}
__decode_pool(p, pi);
@@ -707,6 +708,7 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
newcrush = crush_decode(*p, min(*p+len, end));
if (IS_ERR(newcrush))
return ERR_CAST(newcrush);
+ *p += len;
}
/* new flags? */
@@ -829,12 +831,13 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
/* remove any? */
while (rbp && pgid_cmp(rb_entry(rbp, struct ceph_pg_mapping,
node)->pgid, pgid) <= 0) {
- struct rb_node *cur = rbp;
+ struct ceph_pg_mapping *cur =
+ rb_entry(rbp, struct ceph_pg_mapping, node);
+
rbp = rb_next(rbp);
- dout(" removed pg_temp %llx\n",
- *(u64 *)&rb_entry(cur, struct ceph_pg_mapping,
- node)->pgid);
- rb_erase(cur, &map->pg_temp);
+ dout(" removed pg_temp %llx\n", *(u64 *)&cur->pgid);
+ rb_erase(&cur->node, &map->pg_temp);
+ kfree(cur);
}
if (pglen) {
@@ -850,19 +853,22 @@ struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
for (j = 0; j < pglen; j++)
pg->osds[j] = ceph_decode_32(p);
err = __insert_pg_mapping(pg, &map->pg_temp);
- if (err)
+ if (err) {
+ kfree(pg);
goto bad;
+ }
dout(" added pg_temp %llx len %d\n", *(u64 *)&pgid,
pglen);
}
}
while (rbp) {
- struct rb_node *cur = rbp;
+ struct ceph_pg_mapping *cur =
+ rb_entry(rbp, struct ceph_pg_mapping, node);
+
rbp = rb_next(rbp);
- dout(" removed pg_temp %llx\n",
- *(u64 *)&rb_entry(cur, struct ceph_pg_mapping,
- node)->pgid);
- rb_erase(cur, &map->pg_temp);
+ dout(" removed pg_temp %llx\n", *(u64 *)&cur->pgid);
+ rb_erase(&cur->node, &map->pg_temp);
+ kfree(cur);
}
/* ignore the rest */
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index 4e0bee240b9..fa87f51e38e 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -89,7 +89,7 @@ static int ceph_statfs(struct dentry *dentry, struct kstatfs *buf)
buf->f_files = le64_to_cpu(st.num_objects);
buf->f_ffree = -1;
- buf->f_namelen = PATH_MAX;
+ buf->f_namelen = NAME_MAX;
buf->f_frsize = PAGE_CACHE_SIZE;
/* leave fsid little-endian, regardless of host endianness */
@@ -926,7 +926,7 @@ static int ceph_compare_super(struct super_block *sb, void *data)
/*
* construct our own bdi so we can control readahead, etc.
*/
-static atomic_long_t bdi_seq = ATOMIC_INIT(0);
+static atomic_long_t bdi_seq = ATOMIC_LONG_INIT(0);
static int ceph_register_bdi(struct super_block *sb, struct ceph_client *client)
{
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2024-12-03 15:21:27 +0000