aboutsummaryrefslogtreecommitdiff
path: root/fs/ceph/mds_client.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ceph/mds_client.c')
-rw-r--r--fs/ceph/mds_client.c1512
1 files changed, 1071 insertions, 441 deletions
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 5c7920be642..92a2548278f 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -1,15 +1,22 @@
-#include "ceph_debug.h"
+#include <linux/ceph/ceph_debug.h>
+#include <linux/fs.h>
#include <linux/wait.h>
+#include <linux/slab.h>
+#include <linux/gfp.h>
#include <linux/sched.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
-#include "mds_client.h"
-#include "mon_client.h"
#include "super.h"
-#include "messenger.h"
-#include "decode.h"
-#include "auth.h"
-#include "pagelist.h"
+#include "mds_client.h"
+
+#include <linux/ceph/ceph_features.h>
+#include <linux/ceph/messenger.h>
+#include <linux/ceph/decode.h>
+#include <linux/ceph/pagelist.h>
+#include <linux/ceph/auth.h>
+#include <linux/ceph/debugfs.h>
/*
* A cluster of MDS (metadata server) daemons is responsible for
@@ -36,10 +43,16 @@
* are no longer valid.
*/
+struct ceph_reconnect_state {
+ int nr_caps;
+ struct ceph_pagelist *pagelist;
+ bool flock;
+};
+
static void __wake_requests(struct ceph_mds_client *mdsc,
struct list_head *head);
-const static struct ceph_connection_operations mds_con_ops;
+static const struct ceph_connection_operations mds_con_ops;
/*
@@ -50,7 +63,8 @@ const static struct ceph_connection_operations mds_con_ops;
* parse individual inode info
*/
static int parse_reply_info_in(void **p, void *end,
- struct ceph_mds_reply_info_in *info)
+ struct ceph_mds_reply_info_in *info,
+ u64 features)
{
int err = -EIO;
@@ -64,6 +78,12 @@ static int parse_reply_info_in(void **p, void *end,
info->symlink = *p;
*p += info->symlink_len;
+ if (features & CEPH_FEATURE_DIRLAYOUTHASH)
+ ceph_decode_copy_safe(p, end, &info->dir_layout,
+ sizeof(info->dir_layout), bad);
+ else
+ memset(&info->dir_layout, 0, sizeof(info->dir_layout));
+
ceph_decode_32_safe(p, end, info->xattr_len, bad);
ceph_decode_need(p, end, info->xattr_len, bad);
info->xattr_data = *p;
@@ -78,12 +98,13 @@ bad:
* target inode.
*/
static int parse_reply_info_trace(void **p, void *end,
- struct ceph_mds_reply_info_parsed *info)
+ struct ceph_mds_reply_info_parsed *info,
+ u64 features)
{
int err;
if (info->head->is_dentry) {
- err = parse_reply_info_in(p, end, &info->diri);
+ err = parse_reply_info_in(p, end, &info->diri, features);
if (err < 0)
goto out_bad;
@@ -104,7 +125,7 @@ static int parse_reply_info_trace(void **p, void *end,
}
if (info->head->is_target) {
- err = parse_reply_info_in(p, end, &info->targeti);
+ err = parse_reply_info_in(p, end, &info->targeti, features);
if (err < 0)
goto out_bad;
}
@@ -124,7 +145,8 @@ out_bad:
* parse readdir results
*/
static int parse_reply_info_dir(void **p, void *end,
- struct ceph_mds_reply_info_parsed *info)
+ struct ceph_mds_reply_info_parsed *info,
+ u64 features)
{
u32 num, i = 0;
int err;
@@ -144,21 +166,18 @@ static int parse_reply_info_dir(void **p, void *end,
if (num == 0)
goto done;
- /* alloc large array */
- info->dir_nr = num;
- info->dir_in = kcalloc(num, sizeof(*info->dir_in) +
- sizeof(*info->dir_dname) +
- sizeof(*info->dir_dname_len) +
- sizeof(*info->dir_dlease),
- GFP_NOFS);
- if (info->dir_in == NULL) {
- err = -ENOMEM;
- goto out_bad;
- }
+ BUG_ON(!info->dir_in);
info->dir_dname = (void *)(info->dir_in + num);
info->dir_dname_len = (void *)(info->dir_dname + num);
info->dir_dlease = (void *)(info->dir_dname_len + num);
+ if ((unsigned long)(info->dir_dlease + num) >
+ (unsigned long)info->dir_in + info->dir_buf_size) {
+ pr_err("dir contents are larger than expected\n");
+ WARN_ON(1);
+ goto bad;
+ }
+ info->dir_nr = num;
while (num) {
/* dentry */
ceph_decode_need(p, end, sizeof(u32)*2, bad);
@@ -172,7 +191,7 @@ static int parse_reply_info_dir(void **p, void *end,
*p += sizeof(struct ceph_mds_reply_lease);
/* inode */
- err = parse_reply_info_in(p, end, &info->dir_in[i]);
+ err = parse_reply_info_in(p, end, &info->dir_in[i], features);
if (err < 0)
goto out_bad;
i++;
@@ -192,10 +211,74 @@ out_bad:
}
/*
+ * parse fcntl F_GETLK results
+ */
+static int parse_reply_info_filelock(void **p, void *end,
+ struct ceph_mds_reply_info_parsed *info,
+ u64 features)
+{
+ if (*p + sizeof(*info->filelock_reply) > end)
+ goto bad;
+
+ info->filelock_reply = *p;
+ *p += sizeof(*info->filelock_reply);
+
+ if (unlikely(*p != end))
+ goto bad;
+ return 0;
+
+bad:
+ return -EIO;
+}
+
+/*
+ * parse create results
+ */
+static int parse_reply_info_create(void **p, void *end,
+ struct ceph_mds_reply_info_parsed *info,
+ u64 features)
+{
+ if (features & CEPH_FEATURE_REPLY_CREATE_INODE) {
+ if (*p == end) {
+ info->has_create_ino = false;
+ } else {
+ info->has_create_ino = true;
+ info->ino = ceph_decode_64(p);
+ }
+ }
+
+ if (unlikely(*p != end))
+ goto bad;
+ return 0;
+
+bad:
+ return -EIO;
+}
+
+/*
+ * parse extra results
+ */
+static int parse_reply_info_extra(void **p, void *end,
+ struct ceph_mds_reply_info_parsed *info,
+ u64 features)
+{
+ if (info->head->op == CEPH_MDS_OP_GETFILELOCK)
+ return parse_reply_info_filelock(p, end, info, features);
+ else if (info->head->op == CEPH_MDS_OP_READDIR ||
+ info->head->op == CEPH_MDS_OP_LSSNAP)
+ return parse_reply_info_dir(p, end, info, features);
+ else if (info->head->op == CEPH_MDS_OP_CREATE)
+ return parse_reply_info_create(p, end, info, features);
+ else
+ return -EIO;
+}
+
+/*
* parse entire mds reply
*/
static int parse_reply_info(struct ceph_msg *msg,
- struct ceph_mds_reply_info_parsed *info)
+ struct ceph_mds_reply_info_parsed *info,
+ u64 features)
{
void *p, *end;
u32 len;
@@ -208,15 +291,17 @@ static int parse_reply_info(struct ceph_msg *msg,
/* trace */
ceph_decode_32_safe(&p, end, len, bad);
if (len > 0) {
- err = parse_reply_info_trace(&p, p+len, info);
+ ceph_decode_need(&p, end, len, bad);
+ err = parse_reply_info_trace(&p, p+len, info, features);
if (err < 0)
goto out_bad;
}
- /* dir content */
+ /* extra */
ceph_decode_32_safe(&p, end, len, bad);
if (len > 0) {
- err = parse_reply_info_dir(&p, p+len, info);
+ ceph_decode_need(&p, end, len, bad);
+ err = parse_reply_info_extra(&p, p+len, info, features);
if (err < 0)
goto out_bad;
}
@@ -240,7 +325,9 @@ out_bad:
static void destroy_reply_info(struct ceph_mds_reply_info_parsed *info)
{
- kfree(info->dir_in);
+ if (!info->dir_in)
+ return;
+ free_pages((unsigned long)info->dir_in, get_order(info->dir_buf_size));
}
@@ -278,9 +365,10 @@ void ceph_put_mds_session(struct ceph_mds_session *s)
dout("mdsc put_session %p %d -> %d\n", s,
atomic_read(&s->s_ref), atomic_read(&s->s_ref)-1);
if (atomic_dec_and_test(&s->s_ref)) {
- if (s->s_authorizer)
- s->s_mdsc->client->monc.auth->ops->destroy_authorizer(
- s->s_mdsc->client->monc.auth, s->s_authorizer);
+ if (s->s_auth.authorizer)
+ ceph_auth_destroy_authorizer(
+ s->s_mdsc->fsc->client->monc.auth,
+ s->s_auth.authorizer);
kfree(s);
}
}
@@ -327,6 +415,9 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc,
{
struct ceph_mds_session *s;
+ if (mds >= mdsc->mdsmap->m_max_mds)
+ return ERR_PTR(-EINVAL);
+
s = kzalloc(sizeof(*s), GFP_NOFS);
if (!s)
return ERR_PTR(-ENOMEM);
@@ -337,15 +428,13 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc,
s->s_seq = 0;
mutex_init(&s->s_mutex);
- ceph_con_init(mdsc->client->msgr, &s->s_con);
- s->s_con.private = s;
- s->s_con.ops = &mds_con_ops;
- s->s_con.peer_name.type = CEPH_ENTITY_TYPE_MDS;
- s->s_con.peer_name.num = cpu_to_le64(mds);
+ ceph_con_init(&s->s_con, s, &mds_con_ops, &mdsc->fsc->client->msgr);
- spin_lock_init(&s->s_cap_lock);
+ spin_lock_init(&s->s_gen_ttl_lock);
s->s_cap_gen = 0;
- s->s_cap_ttl = 0;
+ s->s_cap_ttl = jiffies - 1;
+
+ spin_lock_init(&s->s_cap_lock);
s->s_renew_requested = 0;
s->s_renew_seq = 0;
INIT_LIST_HEAD(&s->s_caps);
@@ -355,6 +444,7 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc,
INIT_LIST_HEAD(&s->s_waiting);
INIT_LIST_HEAD(&s->s_unsafe);
s->s_num_cap_releases = 0;
+ s->s_cap_reconnect = 0;
s->s_cap_iterator = NULL;
INIT_LIST_HEAD(&s->s_cap_releases);
INIT_LIST_HEAD(&s->s_cap_releases_done);
@@ -381,7 +471,8 @@ static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc,
mdsc->sessions[mds] = s;
atomic_inc(&s->s_ref); /* one ref to sessions[], one to caller */
- ceph_con_open(&s->s_con, ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
+ ceph_con_open(&s->s_con, CEPH_ENTITY_TYPE_MDS, mds,
+ ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
return s;
@@ -421,34 +512,38 @@ void ceph_mdsc_release_request(struct kref *kref)
struct ceph_mds_request *req = container_of(kref,
struct ceph_mds_request,
r_kref);
+ destroy_reply_info(&req->r_reply_info);
if (req->r_request)
ceph_msg_put(req->r_request);
- if (req->r_reply) {
+ if (req->r_reply)
ceph_msg_put(req->r_reply);
- destroy_reply_info(&req->r_reply_info);
- }
if (req->r_inode) {
- ceph_put_cap_refs(ceph_inode(req->r_inode),
- CEPH_CAP_PIN);
+ ceph_put_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN);
iput(req->r_inode);
}
if (req->r_locked_dir)
- ceph_put_cap_refs(ceph_inode(req->r_locked_dir),
- CEPH_CAP_PIN);
+ ceph_put_cap_refs(ceph_inode(req->r_locked_dir), CEPH_CAP_PIN);
if (req->r_target_inode)
iput(req->r_target_inode);
if (req->r_dentry)
dput(req->r_dentry);
- if (req->r_old_dentry) {
- ceph_put_cap_refs(
- ceph_inode(req->r_old_dentry->d_parent->d_inode),
- CEPH_CAP_PIN);
+ if (req->r_old_dentry)
dput(req->r_old_dentry);
+ if (req->r_old_dentry_dir) {
+ /*
+ * track (and drop pins for) r_old_dentry_dir
+ * separately, since r_old_dentry's d_parent may have
+ * changed between the dir mutex being dropped and
+ * this request being freed.
+ */
+ ceph_put_cap_refs(ceph_inode(req->r_old_dentry_dir),
+ CEPH_CAP_PIN);
+ iput(req->r_old_dentry_dir);
}
kfree(req->r_path1);
kfree(req->r_path2);
put_request_session(req);
- ceph_unreserve_caps(&req->r_caps_reservation);
+ ceph_unreserve_caps(req->r_mdsc, &req->r_caps_reservation);
kfree(req);
}
@@ -511,14 +606,19 @@ static void __register_request(struct ceph_mds_client *mdsc,
{
req->r_tid = ++mdsc->last_tid;
if (req->r_num_caps)
- ceph_reserve_caps(&req->r_caps_reservation, req->r_num_caps);
+ ceph_reserve_caps(mdsc, &req->r_caps_reservation,
+ req->r_num_caps);
dout("__register_request %p tid %lld\n", req, req->r_tid);
ceph_mdsc_get_request(req);
__insert_request(mdsc, req);
+ req->r_uid = current_fsuid();
+ req->r_gid = current_fsgid();
+
if (dir) {
struct ceph_inode_info *ci = ceph_inode(dir);
+ ihold(dir);
spin_lock(&ci->i_unsafe_lock);
req->r_unsafe_dir = dir;
list_add_tail(&req->r_unsafe_dir_item, &ci->i_unsafe_dirops);
@@ -539,8 +639,13 @@ static void __unregister_request(struct ceph_mds_client *mdsc,
spin_lock(&ci->i_unsafe_lock);
list_del_init(&req->r_unsafe_dir_item);
spin_unlock(&ci->i_unsafe_lock);
+
+ iput(req->r_unsafe_dir);
+ req->r_unsafe_dir = NULL;
}
+ complete_all(&req->r_safe_completion);
+
ceph_mdsc_put_request(req);
}
@@ -552,6 +657,19 @@ static void __unregister_request(struct ceph_mds_client *mdsc,
*
* Called under mdsc->mutex.
*/
+static struct dentry *get_nonsnap_parent(struct dentry *dentry)
+{
+ /*
+ * we don't need to worry about protecting the d_parent access
+ * here because we never renaming inside the snapped namespace
+ * except to resplice to another snapdir, and either the old or new
+ * result is a valid result.
+ */
+ while (!IS_ROOT(dentry) && ceph_snap(dentry->d_inode) != CEPH_NOSNAP)
+ dentry = dentry->d_parent;
+ return dentry;
+}
+
static int __choose_mds(struct ceph_mds_client *mdsc,
struct ceph_mds_request *req)
{
@@ -582,14 +700,31 @@ static int __choose_mds(struct ceph_mds_client *mdsc,
if (req->r_inode) {
inode = req->r_inode;
} else if (req->r_dentry) {
- if (req->r_dentry->d_inode) {
+ /* ignore race with rename; old or new d_parent is okay */
+ struct dentry *parent = req->r_dentry->d_parent;
+ struct inode *dir = parent->d_inode;
+
+ if (dir->i_sb != mdsc->fsc->sb) {
+ /* not this fs! */
inode = req->r_dentry->d_inode;
+ } else if (ceph_snap(dir) != CEPH_NOSNAP) {
+ /* direct snapped/virtual snapdir requests
+ * based on parent dir inode */
+ struct dentry *dn = get_nonsnap_parent(parent);
+ inode = dn->d_inode;
+ dout("__choose_mds using nonsnap parent %p\n", inode);
} else {
- inode = req->r_dentry->d_parent->d_inode;
- hash = req->r_dentry->d_name.hash;
- is_hash = true;
+ /* dentry target */
+ inode = req->r_dentry->d_inode;
+ if (!inode || mode == USE_AUTH_MDS) {
+ /* dir + name */
+ inode = dir;
+ hash = ceph_dentry_hash(dir, req->r_dentry);
+ is_hash = true;
+ }
}
}
+
dout("__choose_mds %p is_hash=%d (%d) mode %d\n", inode, (int)is_hash,
(int)hash, mode);
if (!inode)
@@ -612,9 +747,11 @@ static int __choose_mds(struct ceph_mds_client *mdsc,
dout("choose_mds %p %llx.%llx "
"frag %u mds%d (%d/%d)\n",
inode, ceph_vinop(inode),
- frag.frag, frag.mds,
+ frag.frag, mds,
(int)r, frag.ndist);
- return mds;
+ if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
+ CEPH_MDS_STATE_ACTIVE)
+ return mds;
}
/* since this file/dir wasn't known to be
@@ -627,26 +764,28 @@ static int __choose_mds(struct ceph_mds_client *mdsc,
dout("choose_mds %p %llx.%llx "
"frag %u mds%d (auth)\n",
inode, ceph_vinop(inode), frag.frag, mds);
- return mds;
+ if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
+ CEPH_MDS_STATE_ACTIVE)
+ return mds;
}
}
}
- spin_lock(&inode->i_lock);
+ spin_lock(&ci->i_ceph_lock);
cap = NULL;
if (mode == USE_AUTH_MDS)
cap = ci->i_auth_cap;
if (!cap && !RB_EMPTY_ROOT(&ci->i_caps))
cap = rb_entry(rb_first(&ci->i_caps), struct ceph_cap, ci_node);
if (!cap) {
- spin_unlock(&inode->i_lock);
+ spin_unlock(&ci->i_ceph_lock);
goto random;
}
mds = cap->session->s_mds;
dout("choose_mds %p %llx.%llx mds%d (%scap %p)\n",
inode, ceph_vinop(inode), mds,
cap == ci->i_auth_cap ? "auth " : "", cap);
- spin_unlock(&inode->i_lock);
+ spin_unlock(&ci->i_ceph_lock);
return mds;
random:
@@ -664,10 +803,11 @@ static struct ceph_msg *create_session_msg(u32 op, u64 seq)
struct ceph_msg *msg;
struct ceph_mds_session_head *h;
- msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), 0, 0, NULL);
- if (IS_ERR(msg)) {
+ msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), GFP_NOFS,
+ false);
+ if (!msg) {
pr_err("create_session_msg ENOMEM creating msg\n");
- return ERR_PTR(PTR_ERR(msg));
+ return NULL;
}
h = msg->front.iov_base;
h->op = cpu_to_le32(op);
@@ -686,7 +826,6 @@ static int __open_session(struct ceph_mds_client *mdsc,
struct ceph_msg *msg;
int mstate;
int mds = session->s_mds;
- int err = 0;
/* wait for mds to go active? */
mstate = ceph_mdsmap_get_state(mdsc->mdsmap, mds);
@@ -697,17 +836,79 @@ static int __open_session(struct ceph_mds_client *mdsc,
/* send connect message */
msg = create_session_msg(CEPH_SESSION_REQUEST_OPEN, session->s_seq);
- if (IS_ERR(msg)) {
- err = PTR_ERR(msg);
- goto out;
- }
+ if (!msg)
+ return -ENOMEM;
ceph_con_send(&session->s_con, msg);
-
-out:
return 0;
}
/*
+ * open sessions for any export targets for the given mds
+ *
+ * called under mdsc->mutex
+ */
+static struct ceph_mds_session *
+__open_export_target_session(struct ceph_mds_client *mdsc, int target)
+{
+ struct ceph_mds_session *session;
+
+ session = __ceph_lookup_mds_session(mdsc, target);
+ if (!session) {
+ session = register_session(mdsc, target);
+ if (IS_ERR(session))
+ return session;
+ }
+ if (session->s_state == CEPH_MDS_SESSION_NEW ||
+ session->s_state == CEPH_MDS_SESSION_CLOSING)
+ __open_session(mdsc, session);
+
+ return session;
+}
+
+struct ceph_mds_session *
+ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target)
+{
+ struct ceph_mds_session *session;
+
+ dout("open_export_target_session to mds%d\n", target);
+
+ mutex_lock(&mdsc->mutex);
+ session = __open_export_target_session(mdsc, target);
+ mutex_unlock(&mdsc->mutex);
+
+ return session;
+}
+
+static void __open_export_target_sessions(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session)
+{
+ struct ceph_mds_info *mi;
+ struct ceph_mds_session *ts;
+ int i, mds = session->s_mds;
+
+ if (mds >= mdsc->mdsmap->m_max_mds)
+ return;
+
+ mi = &mdsc->mdsmap->m_info[mds];
+ dout("open_export_target_sessions for mds%d (%d targets)\n",
+ session->s_mds, mi->num_export_targets);
+
+ for (i = 0; i < mi->num_export_targets; i++) {
+ ts = __open_export_target_session(mdsc, mi->export_targets[i]);
+ if (!IS_ERR(ts))
+ ceph_put_mds_session(ts);
+ }
+}
+
+void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session)
+{
+ mutex_lock(&mdsc->mutex);
+ __open_export_target_sessions(mdsc, session);
+ mutex_unlock(&mdsc->mutex);
+}
+
+/*
* session caps
*/
@@ -735,9 +936,10 @@ static void cleanup_cap_releases(struct ceph_mds_session *session)
}
/*
- * Helper to safely iterate over all caps associated with a session.
+ * Helper to safely iterate over all caps associated with a session, with
+ * special care taken to handle a racing __ceph_remove_cap().
*
- * caller must hold session s_mutex
+ * Caller must hold session s_mutex.
*/
static int iterate_session_caps(struct ceph_mds_session *session,
int (*cb)(struct inode *, struct ceph_cap *,
@@ -767,7 +969,7 @@ static int iterate_session_caps(struct ceph_mds_session *session,
last_inode = NULL;
}
if (old_cap) {
- ceph_put_cap(old_cap);
+ ceph_put_cap(session->s_mdsc, old_cap);
old_cap = NULL;
}
@@ -796,18 +998,55 @@ out:
if (last_inode)
iput(last_inode);
if (old_cap)
- ceph_put_cap(old_cap);
+ ceph_put_cap(session->s_mdsc, old_cap);
return ret;
}
static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
- void *arg)
+ void *arg)
{
struct ceph_inode_info *ci = ceph_inode(inode);
+ int drop = 0;
+
dout("removing cap %p, ci is %p, inode is %p\n",
cap, ci, &ci->vfs_inode);
- ceph_remove_cap(cap);
+ spin_lock(&ci->i_ceph_lock);
+ __ceph_remove_cap(cap, false);
+ if (!__ceph_is_any_real_caps(ci)) {
+ struct ceph_mds_client *mdsc =
+ ceph_sb_to_client(inode->i_sb)->mdsc;
+
+ spin_lock(&mdsc->cap_dirty_lock);
+ if (!list_empty(&ci->i_dirty_item)) {
+ pr_info(" dropping dirty %s state for %p %lld\n",
+ ceph_cap_string(ci->i_dirty_caps),
+ inode, ceph_ino(inode));
+ ci->i_dirty_caps = 0;
+ list_del_init(&ci->i_dirty_item);
+ drop = 1;
+ }
+ if (!list_empty(&ci->i_flushing_item)) {
+ pr_info(" dropping dirty+flushing %s state for %p %lld\n",
+ ceph_cap_string(ci->i_flushing_caps),
+ inode, ceph_ino(inode));
+ ci->i_flushing_caps = 0;
+ list_del_init(&ci->i_flushing_item);
+ mdsc->num_cap_flushing--;
+ drop = 1;
+ }
+ if (drop && ci->i_wrbuffer_ref) {
+ pr_info(" dropping dirty data for %p %lld\n",
+ inode, ceph_ino(inode));
+ ci->i_wrbuffer_ref = 0;
+ ci->i_wrbuffer_ref_head = 0;
+ drop++;
+ }
+ spin_unlock(&mdsc->cap_dirty_lock);
+ }
+ spin_unlock(&ci->i_ceph_lock);
+ while (drop--)
+ iput(inode);
return 0;
}
@@ -818,7 +1057,39 @@ static void remove_session_caps(struct ceph_mds_session *session)
{
dout("remove_session_caps on %p\n", session);
iterate_session_caps(session, remove_session_caps_cb, NULL);
+
+ spin_lock(&session->s_cap_lock);
+ if (session->s_nr_caps > 0) {
+ struct super_block *sb = session->s_mdsc->fsc->sb;
+ struct inode *inode;
+ struct ceph_cap *cap, *prev = NULL;
+ struct ceph_vino vino;
+ /*
+ * iterate_session_caps() skips inodes that are being
+ * deleted, we need to wait until deletions are complete.
+ * __wait_on_freeing_inode() is designed for the job,
+ * but it is not exported, so use lookup inode function
+ * to access it.
+ */
+ while (!list_empty(&session->s_caps)) {
+ cap = list_entry(session->s_caps.next,
+ struct ceph_cap, session_caps);
+ if (cap == prev)
+ break;
+ prev = cap;
+ vino = cap->ci->i_vino;
+ spin_unlock(&session->s_cap_lock);
+
+ inode = ceph_find_inode(sb, vino);
+ iput(inode);
+
+ spin_lock(&session->s_cap_lock);
+ }
+ }
+ spin_unlock(&session->s_cap_lock);
+
BUG_ON(session->s_nr_caps > 0);
+ BUG_ON(!list_empty(&session->s_cap_flushing));
cleanup_cap_releases(session);
}
@@ -833,12 +1104,12 @@ 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);
+ spin_lock(&ci->i_ceph_lock);
ci->i_wanted_max_size = 0;
ci->i_requested_max_size = 0;
- spin_unlock(&inode->i_lock);
+ spin_unlock(&ci->i_ceph_lock);
}
return 0;
}
@@ -881,12 +1152,27 @@ static int send_renew_caps(struct ceph_mds_client *mdsc,
ceph_mds_state_name(state));
msg = create_session_msg(CEPH_SESSION_REQUEST_RENEWCAPS,
++session->s_renew_seq);
- if (IS_ERR(msg))
- return PTR_ERR(msg);
+ if (!msg)
+ return -ENOMEM;
ceph_con_send(&session->s_con, msg);
return 0;
}
+static int send_flushmsg_ack(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session, u64 seq)
+{
+ struct ceph_msg *msg;
+
+ dout("send_flushmsg_ack to mds%d (%s)s seq %lld\n",
+ session->s_mds, session_state_name(session->s_state), seq);
+ msg = create_session_msg(CEPH_SESSION_FLUSHMSG_ACK, seq);
+ if (!msg)
+ return -ENOMEM;
+ ceph_con_send(&session->s_con, msg);
+ return 0;
+}
+
+
/*
* Note new cap ttl, and any transition from stale -> not stale (fresh?).
*
@@ -899,8 +1185,7 @@ static void renewed_caps(struct ceph_mds_client *mdsc,
int wake = 0;
spin_lock(&session->s_cap_lock);
- was_stale = is_renew && (session->s_cap_ttl == 0 ||
- time_after_eq(jiffies, session->s_cap_ttl));
+ was_stale = is_renew && time_after_eq(jiffies, session->s_cap_ttl);
session->s_cap_ttl = session->s_renew_requested +
mdsc->mdsmap->m_session_timeout*HZ;
@@ -929,17 +1214,15 @@ static int request_close_session(struct ceph_mds_client *mdsc,
struct ceph_mds_session *session)
{
struct ceph_msg *msg;
- int err = 0;
dout("request_close_session mds%d state %s seq %lld\n",
session->s_mds, session_state_name(session->s_state),
session->s_seq);
msg = create_session_msg(CEPH_SESSION_REQUEST_CLOSE, session->s_seq);
- if (IS_ERR(msg))
- err = PTR_ERR(msg);
- else
- ceph_con_send(&session->s_con, msg);
- return err;
+ if (!msg)
+ return -ENOMEM;
+ ceph_con_send(&session->s_con, msg);
+ return 0;
}
/*
@@ -968,31 +1251,36 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg)
{
struct ceph_mds_session *session = arg;
struct ceph_inode_info *ci = ceph_inode(inode);
- int used, oissued, mine;
+ int used, wanted, oissued, mine;
if (session->s_trim_caps <= 0)
return -1;
- spin_lock(&inode->i_lock);
+ spin_lock(&ci->i_ceph_lock);
mine = cap->issued | cap->implemented;
used = __ceph_caps_used(ci);
+ wanted = __ceph_caps_file_wanted(ci);
oissued = __ceph_caps_issued_other(ci, cap);
- dout("trim_caps_cb %p cap %p mine %s oissued %s used %s\n",
+ dout("trim_caps_cb %p cap %p mine %s oissued %s used %s wanted %s\n",
inode, cap, ceph_cap_string(mine), ceph_cap_string(oissued),
- ceph_cap_string(used));
- if (ci->i_dirty_caps)
- goto out; /* dirty caps */
- if ((used & ~oissued) & mine)
+ ceph_cap_string(used), ceph_cap_string(wanted));
+ if (cap == ci->i_auth_cap) {
+ if (ci->i_dirty_caps | ci->i_flushing_caps)
+ goto out;
+ if ((used | wanted) & CEPH_CAP_ANY_WR)
+ goto out;
+ }
+ if ((used | wanted) & ~oissued & mine)
goto out; /* we need these caps */
session->s_trim_caps--;
if (oissued) {
/* we aren't the only cap.. just remove us */
- __ceph_remove_cap(cap);
+ __ceph_remove_cap(cap, true);
} else {
/* try to drop referring dentries */
- spin_unlock(&inode->i_lock);
+ spin_unlock(&ci->i_ceph_lock);
d_prune_aliases(inode);
dout("trim_caps_cb %p cap %p pruned, count now %d\n",
inode, cap, atomic_read(&inode->i_count));
@@ -1000,7 +1288,7 @@ static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg)
}
out:
- spin_unlock(&inode->i_lock);
+ spin_unlock(&ci->i_ceph_lock);
return 0;
}
@@ -1023,6 +1311,9 @@ static int trim_caps(struct ceph_mds_client *mdsc,
trim_caps - session->s_trim_caps);
session->s_trim_caps = 0;
}
+
+ ceph_add_cap_releases(mdsc, session);
+ ceph_send_cap_releases(mdsc, session);
return 0;
}
@@ -1033,16 +1324,17 @@ 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)
{
- struct ceph_msg *msg;
+ struct ceph_msg *msg, *partial = NULL;
struct ceph_mds_cap_release *head;
int err = -ENOMEM;
+ int extra = mdsc->fsc->mount_options->cap_release_safety;
+ int num;
- if (extra < 0)
- extra = mdsc->client->mount_args->cap_release_safety;
+ dout("add_cap_releases %p mds%d extra %d\n", session, session->s_mds,
+ extra);
spin_lock(&session->s_cap_lock);
@@ -1051,13 +1343,18 @@ static int add_cap_releases(struct ceph_mds_client *mdsc,
struct ceph_msg,
list_head);
head = msg->front.iov_base;
- extra += CEPH_CAPS_PER_RELEASE - le32_to_cpu(head->num);
+ num = le32_to_cpu(head->num);
+ if (num) {
+ dout(" partial %p with (%d/%d)\n", msg, num,
+ (int)CEPH_CAPS_PER_RELEASE);
+ extra += CEPH_CAPS_PER_RELEASE - num;
+ partial = msg;
+ }
}
-
while (session->s_num_cap_releases < session->s_nr_caps + extra) {
spin_unlock(&session->s_cap_lock);
msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPRELEASE, PAGE_CACHE_SIZE,
- 0, 0, NULL);
+ GFP_NOFS, false);
if (!msg)
goto out_unlocked;
dout("add_cap_releases %p msg %p now %d\n", session, msg,
@@ -1070,19 +1367,14 @@ static int add_cap_releases(struct ceph_mds_client *mdsc,
session->s_num_cap_releases += CEPH_CAPS_PER_RELEASE;
}
- if (!list_empty(&session->s_cap_releases)) {
- msg = list_first_entry(&session->s_cap_releases,
- struct ceph_msg,
- list_head);
- head = msg->front.iov_base;
- if (head->num) {
- dout(" queueing non-full %p (%d)\n", msg,
- le32_to_cpu(head->num));
- list_move_tail(&msg->list_head,
- &session->s_cap_releases_done);
- session->s_num_cap_releases -=
- CEPH_CAPS_PER_RELEASE - le32_to_cpu(head->num);
- }
+ if (partial) {
+ head = partial->front.iov_base;
+ num = le32_to_cpu(head->num);
+ dout(" queueing partial %p with %d/%d\n", partial, num,
+ (int)CEPH_CAPS_PER_RELEASE);
+ list_move_tail(&partial->list_head,
+ &session->s_cap_releases_done);
+ session->s_num_cap_releases -= CEPH_CAPS_PER_RELEASE - num;
}
err = 0;
spin_unlock(&session->s_cap_lock);
@@ -1117,7 +1409,7 @@ static int check_cap_flush(struct ceph_mds_client *mdsc, u64 want_flush_seq)
i_flushing_item);
struct inode *inode = &ci->vfs_inode;
- spin_lock(&inode->i_lock);
+ spin_lock(&ci->i_ceph_lock);
if (ci->i_cap_flush_seq <= want_flush_seq) {
dout("check_cap_flush still flushing %p "
"seq %lld <= %lld to mds%d\n", inode,
@@ -1125,7 +1417,7 @@ static int check_cap_flush(struct ceph_mds_client *mdsc, u64 want_flush_seq)
session->s_mds);
ret = 0;
}
- spin_unlock(&inode->i_lock);
+ spin_unlock(&ci->i_ceph_lock);
}
mutex_unlock(&session->s_mutex);
ceph_put_mds_session(session);
@@ -1143,16 +1435,14 @@ 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;
dout("send_cap_releases mds%d\n", session->s_mds);
- while (1) {
- spin_lock(&session->s_cap_lock);
- if (list_empty(&session->s_cap_releases_done))
- break;
+ spin_lock(&session->s_cap_lock);
+ while (!list_empty(&session->s_cap_releases_done)) {
msg = list_first_entry(&session->s_cap_releases_done,
struct ceph_msg, list_head);
list_del_init(&msg->list_head);
@@ -1160,14 +1450,91 @@ static void send_cap_releases(struct ceph_mds_client *mdsc,
msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
dout("send_cap_releases mds%d %p\n", session->s_mds, msg);
ceph_con_send(&session->s_con, msg);
+ spin_lock(&session->s_cap_lock);
}
spin_unlock(&session->s_cap_lock);
}
+static void discard_cap_releases(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session)
+{
+ struct ceph_msg *msg;
+ struct ceph_mds_cap_release *head;
+ unsigned num;
+
+ dout("discard_cap_releases mds%d\n", session->s_mds);
+
+ if (!list_empty(&session->s_cap_releases)) {
+ /* zero out the in-progress message */
+ msg = list_first_entry(&session->s_cap_releases,
+ struct ceph_msg, list_head);
+ head = msg->front.iov_base;
+ num = le32_to_cpu(head->num);
+ dout("discard_cap_releases mds%d %p %u\n",
+ session->s_mds, msg, num);
+ head->num = cpu_to_le32(0);
+ msg->front.iov_len = sizeof(*head);
+ session->s_num_cap_releases += num;
+ }
+
+ /* requeue completed messages */
+ while (!list_empty(&session->s_cap_releases_done)) {
+ msg = list_first_entry(&session->s_cap_releases_done,
+ struct ceph_msg, list_head);
+ list_del_init(&msg->list_head);
+
+ head = msg->front.iov_base;
+ num = le32_to_cpu(head->num);
+ dout("discard_cap_releases mds%d %p %u\n", session->s_mds, msg,
+ num);
+ session->s_num_cap_releases += num;
+ head->num = cpu_to_le32(0);
+ msg->front.iov_len = sizeof(*head);
+ list_add(&msg->list_head, &session->s_cap_releases);
+ }
+}
+
/*
* requests
*/
+int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req,
+ struct inode *dir)
+{
+ struct ceph_inode_info *ci = ceph_inode(dir);
+ struct ceph_mds_reply_info_parsed *rinfo = &req->r_reply_info;
+ struct ceph_mount_options *opt = req->r_mdsc->fsc->mount_options;
+ size_t size = sizeof(*rinfo->dir_in) + sizeof(*rinfo->dir_dname_len) +
+ sizeof(*rinfo->dir_dname) + sizeof(*rinfo->dir_dlease);
+ int order, num_entries;
+
+ spin_lock(&ci->i_ceph_lock);
+ num_entries = ci->i_files + ci->i_subdirs;
+ spin_unlock(&ci->i_ceph_lock);
+ num_entries = max(num_entries, 1);
+ num_entries = min(num_entries, opt->max_readdir);
+
+ order = get_order(size * num_entries);
+ while (order >= 0) {
+ rinfo->dir_in = (void*)__get_free_pages(GFP_NOFS | __GFP_NOWARN,
+ order);
+ if (rinfo->dir_in)
+ break;
+ order--;
+ }
+ if (!rinfo->dir_in)
+ return -ENOMEM;
+
+ num_entries = (PAGE_SIZE << order) / size;
+ num_entries = min(num_entries, opt->max_readdir);
+
+ rinfo->dir_buf_size = PAGE_SIZE << order;
+ req->r_num_caps = num_entries + 1;
+ req->r_args.readdir.max_entries = cpu_to_le32(num_entries);
+ req->r_args.readdir.max_bytes = cpu_to_le32(opt->max_readdir_bytes);
+ return 0;
+}
+
/*
* Create an mds request.
*/
@@ -1179,6 +1546,8 @@ ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode)
if (!req)
return ERR_PTR(-ENOMEM);
+ mutex_init(&req->r_fill_mutex);
+ req->r_mdsc = mdsc;
req->r_started = jiffies;
req->r_resend_mds = -1;
INIT_LIST_HEAD(&req->r_unsafe_dir_item);
@@ -1189,6 +1558,8 @@ ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode)
init_completion(&req->r_safe_completion);
INIT_LIST_HEAD(&req->r_unsafe_item);
+ req->r_stamp = CURRENT_TIME;
+
req->r_op = op;
req->r_direct_mode = mode;
return req;
@@ -1232,12 +1603,15 @@ char *ceph_mdsc_build_path(struct dentry *dentry, int *plen, u64 *base,
struct dentry *temp;
char *path;
int len, pos;
+ unsigned seq;
if (dentry == NULL)
return ERR_PTR(-EINVAL);
retry:
len = 0;
+ seq = read_seqbegin(&rename_lock);
+ rcu_read_lock();
for (temp = dentry; !IS_ROOT(temp);) {
struct inode *inode = temp->d_inode;
if (inode && ceph_snap(inode) == CEPH_SNAPDIR)
@@ -1248,11 +1622,8 @@ retry:
else
len += 1 + temp->d_name.len;
temp = temp->d_parent;
- if (temp == NULL) {
- pr_err("build_path_dentry corrupt dentry %p\n", dentry);
- return ERR_PTR(-EINVAL);
- }
}
+ rcu_read_unlock();
if (len)
len--; /* no leading '/' */
@@ -1261,35 +1632,36 @@ retry:
return ERR_PTR(-ENOMEM);
pos = len;
path[pos] = 0; /* trailing null */
+ rcu_read_lock();
for (temp = dentry; !IS_ROOT(temp) && pos != 0; ) {
- struct inode *inode = temp->d_inode;
+ struct inode *inode;
+ spin_lock(&temp->d_lock);
+ inode = temp->d_inode;
if (inode && ceph_snap(inode) == CEPH_SNAPDIR) {
- dout("build_path_dentry path+%d: %p SNAPDIR\n",
+ dout("build_path path+%d: %p SNAPDIR\n",
pos, temp);
} else if (stop_on_nosnap && inode &&
ceph_snap(inode) == CEPH_NOSNAP) {
+ spin_unlock(&temp->d_lock);
break;
} else {
pos -= temp->d_name.len;
- if (pos < 0)
+ if (pos < 0) {
+ spin_unlock(&temp->d_lock);
break;
+ }
strncpy(path + pos, temp->d_name.name,
temp->d_name.len);
- dout("build_path_dentry path+%d: %p '%.*s'\n",
- pos, temp, temp->d_name.len, path + pos);
}
+ spin_unlock(&temp->d_lock);
if (pos)
path[--pos] = '/';
temp = temp->d_parent;
- if (temp == NULL) {
- pr_err("build_path_dentry corrupt dentry\n");
- kfree(path);
- return ERR_PTR(-EINVAL);
- }
}
- if (pos != 0) {
- pr_err("build_path_dentry did not end path lookup where "
+ rcu_read_unlock();
+ if (pos != 0 || read_seqretry(&rename_lock, seq)) {
+ pr_err("build_path did not end path lookup where "
"expected, namelen is %d, pos is %d\n", len, pos);
/* presumably this is only possible if racing with a
rename of one of the parent directories (we can not
@@ -1301,8 +1673,8 @@ retry:
*base = ceph_ino(temp->d_inode);
*plen = len;
- dout("build_path_dentry on %p %d built %llx '%.*s'\n",
- dentry, atomic_read(&dentry->d_count), *base, len, path);
+ dout("build_path on %p %d built %llx '%.*s'\n",
+ dentry, d_count(dentry), *base, len, path);
return path;
}
@@ -1367,10 +1739,10 @@ static int set_request_path_attr(struct inode *rinode, struct dentry *rdentry,
r = build_dentry_path(rdentry, ppath, pathlen, ino, freepath);
dout(" dentry %p %llx/%.*s\n", rdentry, *ino, *pathlen,
*ppath);
- } else if (rpath) {
+ } else if (rpath || rino) {
*ino = rino;
*ppath = rpath;
- *pathlen = strlen(rpath);
+ *pathlen = rpath ? strlen(rpath) : 0;
dout(" path %.*s\n", *pathlen, rpath);
}
@@ -1413,7 +1785,8 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc,
}
len = sizeof(*head) +
- pathlen1 + pathlen2 + 2*(1 + sizeof(u32) + sizeof(u64));
+ pathlen1 + pathlen2 + 2*(1 + sizeof(u32) + sizeof(u64)) +
+ sizeof(struct timespec);
/* calculate (max) length for cap releases */
len += sizeof(struct ceph_mds_request_release) *
@@ -1424,10 +1797,13 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc,
if (req->r_old_dentry_drop)
len += req->r_old_dentry->d_name.len;
- msg = ceph_msg_new(CEPH_MSG_CLIENT_REQUEST, len, 0, 0, NULL);
- if (IS_ERR(msg))
+ msg = ceph_msg_new(CEPH_MSG_CLIENT_REQUEST, len, GFP_NOFS, false);
+ if (!msg) {
+ msg = ERR_PTR(-ENOMEM);
goto out_free2;
+ }
+ msg->hdr.version = 2;
msg->hdr.tid = cpu_to_le64(req->r_tid);
head = msg->front.iov_base;
@@ -1436,13 +1812,16 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc,
head->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch);
head->op = cpu_to_le32(req->r_op);
- head->caller_uid = cpu_to_le32(current_fsuid());
- head->caller_gid = cpu_to_le32(current_fsgid());
+ head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns, req->r_uid));
+ head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns, req->r_gid));
head->args = req->r_args;
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)
@@ -1461,12 +1840,19 @@ static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc,
mds, req->r_old_inode_drop, req->r_old_inode_unless, 0);
head->num_releases = cpu_to_le16(releases);
+ /* time stamp */
+ ceph_encode_copy(&p, &req->r_stamp, sizeof(req->r_stamp));
+
BUG_ON(p > end);
msg->front.iov_len = p - msg->front.iov_base;
msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
- msg->pages = req->r_pages;
- msg->nr_pages = req->r_num_pages;
+ if (req->r_data_len) {
+ /* outbound data set only by ceph_sync_setxattr() */
+ BUG_ON(!req->r_pages);
+ ceph_msg_data_add_pages(msg, req->r_pages, req->r_data_len, 0);
+ }
+
msg->hdr.data_len = cpu_to_le32(req->r_data_len);
msg->hdr.data_off = cpu_to_le16(0);
@@ -1490,7 +1876,7 @@ 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);
}
/*
@@ -1504,20 +1890,54 @@ static int __prepare_send_request(struct ceph_mds_client *mdsc,
struct ceph_msg *msg;
int flags = 0;
- req->r_mds = mds;
req->r_attempts++;
+ if (req->r_inode) {
+ struct ceph_cap *cap =
+ ceph_get_cap_for_mds(ceph_inode(req->r_inode), mds);
+
+ if (cap)
+ req->r_sent_on_mseq = cap->mseq;
+ else
+ req->r_sent_on_mseq = -1;
+ }
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;
}
msg = create_request_message(mdsc, req, mds);
if (IS_ERR(msg)) {
- req->r_reply = ERR_PTR(PTR_ERR(msg));
+ req->r_err = PTR_ERR(msg);
complete_request(mdsc, req);
- return -PTR_ERR(msg);
+ return PTR_ERR(msg);
}
req->r_request = msg;
@@ -1530,13 +1950,9 @@ 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;
}
@@ -1550,8 +1966,11 @@ static int __do_request(struct ceph_mds_client *mdsc,
int mds = -1;
int err = -EAGAIN;
- if (req->r_reply)
+ if (req->r_err || req->r_got_result) {
+ if (req->r_aborted)
+ __unregister_request(mdsc, req);
goto out;
+ }
if (req->r_timeout &&
time_after_eq(jiffies, req->r_started + req->r_timeout)) {
@@ -1560,6 +1979,8 @@ static int __do_request(struct ceph_mds_client *mdsc,
goto finish;
}
+ put_request_session(req);
+
mds = __choose_mds(mdsc, req);
if (mds < 0 ||
ceph_mdsmap_get_state(mdsc->mdsmap, mds) < CEPH_MDS_STATE_ACTIVE) {
@@ -1577,6 +1998,8 @@ static int __do_request(struct ceph_mds_client *mdsc,
goto finish;
}
}
+ req->r_session = get_session(session);
+
dout("do_request mds%d session %p state %s\n", mds, session,
session_state_name(session->s_state));
if (session->s_state != CEPH_MDS_SESSION_OPEN &&
@@ -1589,7 +2012,6 @@ static int __do_request(struct ceph_mds_client *mdsc,
}
/* send request */
- req->r_session = get_session(session);
req->r_resend_mds = -1; /* forget any previous mds hint */
if (req->r_request_started == 0) /* note request start time */
@@ -1607,7 +2029,7 @@ out:
return err;
finish:
- req->r_reply = ERR_PTR(err);
+ req->r_err = err;
complete_request(mdsc, req);
goto out;
}
@@ -1618,20 +2040,25 @@ finish:
static void __wake_requests(struct ceph_mds_client *mdsc,
struct list_head *head)
{
- struct ceph_mds_request *req, *nreq;
+ struct ceph_mds_request *req;
+ LIST_HEAD(tmp_list);
+
+ list_splice_init(head, &tmp_list);
- list_for_each_entry_safe(req, nreq, head, r_wait) {
+ while (!list_empty(&tmp_list)) {
+ req = list_entry(tmp_list.next,
+ struct ceph_mds_request, r_wait);
list_del_init(&req->r_wait);
+ dout(" wake request %p tid %llu\n", req, req->r_tid);
__do_request(mdsc, req);
}
}
/*
* Wake up threads with requests pending for @mds, so that they can
- * resubmit their requests to a possibly different mds. If @all is set,
- * wake up if their requests has been forwarded to @mds, too.
+ * resubmit their requests to a possibly different mds.
*/
-static void kick_requests(struct ceph_mds_client *mdsc, int mds, int all)
+static void kick_requests(struct ceph_mds_client *mdsc, int mds)
{
struct ceph_mds_request *req;
struct rb_node *p;
@@ -1644,7 +2071,6 @@ static void kick_requests(struct ceph_mds_client *mdsc, int mds, int all)
if (req->r_session &&
req->r_session->s_mds == mds) {
dout(" kicking tid %llu\n", req->r_tid);
- put_request_session(req);
__do_request(mdsc, req);
}
}
@@ -1677,74 +2103,83 @@ int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
ceph_get_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN);
if (req->r_locked_dir)
ceph_get_cap_refs(ceph_inode(req->r_locked_dir), CEPH_CAP_PIN);
- if (req->r_old_dentry)
- ceph_get_cap_refs(
- ceph_inode(req->r_old_dentry->d_parent->d_inode),
- CEPH_CAP_PIN);
+ if (req->r_old_dentry_dir)
+ ceph_get_cap_refs(ceph_inode(req->r_old_dentry_dir),
+ CEPH_CAP_PIN);
/* issue */
mutex_lock(&mdsc->mutex);
__register_request(mdsc, req, dir);
__do_request(mdsc, req);
- /* wait */
- if (!req->r_reply) {
- mutex_unlock(&mdsc->mutex);
- if (req->r_timeout) {
- err = (long)wait_for_completion_interruptible_timeout(
- &req->r_completion, req->r_timeout);
- if (err == 0)
- req->r_reply = ERR_PTR(-EIO);
- else if (err < 0)
- req->r_reply = ERR_PTR(err);
- } else {
- err = wait_for_completion_interruptible(
- &req->r_completion);
- if (err)
- req->r_reply = ERR_PTR(err);
- }
- mutex_lock(&mdsc->mutex);
+ if (req->r_err) {
+ err = req->r_err;
+ __unregister_request(mdsc, req);
+ dout("do_request early error %d\n", err);
+ goto out;
}
- if (IS_ERR(req->r_reply)) {
- err = PTR_ERR(req->r_reply);
- req->r_reply = NULL;
+ /* wait */
+ mutex_unlock(&mdsc->mutex);
+ dout("do_request waiting\n");
+ if (req->r_timeout) {
+ err = (long)wait_for_completion_killable_timeout(
+ &req->r_completion, req->r_timeout);
+ if (err == 0)
+ err = -EIO;
+ } else {
+ err = wait_for_completion_killable(&req->r_completion);
+ }
+ dout("do_request waited, got %d\n", err);
+ mutex_lock(&mdsc->mutex);
- if (err == -ERESTARTSYS) {
- /* aborted */
- req->r_aborted = true;
+ /* only abort if we didn't race with a real reply */
+ if (req->r_got_result) {
+ err = le32_to_cpu(req->r_reply_info.head->result);
+ } else if (err < 0) {
+ dout("aborted request %lld with %d\n", req->r_tid, err);
- if (req->r_locked_dir &&
- (req->r_op & CEPH_MDS_OP_WRITE)) {
- struct ceph_inode_info *ci =
- ceph_inode(req->r_locked_dir);
+ /*
+ * ensure we aren't running concurrently with
+ * ceph_fill_trace or ceph_readdir_prepopulate, which
+ * rely on locks (dir mutex) held by our caller.
+ */
+ mutex_lock(&req->r_fill_mutex);
+ req->r_err = err;
+ req->r_aborted = true;
+ mutex_unlock(&req->r_fill_mutex);
- dout("aborted, clearing I_COMPLETE on %p\n",
- req->r_locked_dir);
- spin_lock(&req->r_locked_dir->i_lock);
- ci->i_ceph_flags &= ~CEPH_I_COMPLETE;
- ci->i_release_count++;
- spin_unlock(&req->r_locked_dir->i_lock);
- }
- } else {
- /* clean up this request */
- __unregister_request(mdsc, req);
- if (!list_empty(&req->r_unsafe_item))
- list_del_init(&req->r_unsafe_item);
- complete(&req->r_safe_completion);
- }
- } else if (req->r_err) {
- err = req->r_err;
+ if (req->r_locked_dir &&
+ (req->r_op & CEPH_MDS_OP_WRITE))
+ ceph_invalidate_dir_request(req);
} else {
- err = le32_to_cpu(req->r_reply_info.head->result);
+ err = req->r_err;
}
- mutex_unlock(&mdsc->mutex);
+out:
+ mutex_unlock(&mdsc->mutex);
dout("do_request %p done, result %d\n", req, err);
return err;
}
/*
+ * Invalidate dir's completeness, dentry lease state on an aborted MDS
+ * namespace request.
+ */
+void ceph_invalidate_dir_request(struct ceph_mds_request *req)
+{
+ struct inode *inode = req->r_locked_dir;
+
+ dout("invalidate_dir_request %p (complete, lease(s))\n", inode);
+
+ ceph_dir_clear_complete(inode);
+ if (req->r_dentry)
+ ceph_invalidate_dentry_lease(req->r_dentry);
+ if (req->r_old_dentry)
+ ceph_invalidate_dentry_lease(req->r_old_dentry);
+}
+
+/*
* Handle mds reply.
*
* We take the session mutex and parse and process the reply immediately.
@@ -1790,34 +2225,51 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
/* dup? */
if ((req->r_got_unsafe && !head->safe) ||
(req->r_got_safe && head->safe)) {
- pr_warning("got a dup %s reply on %llu from mds%d\n",
+ pr_warn("got a dup %s reply on %llu from mds%d\n",
head->safe ? "safe" : "unsafe", tid, mds);
mutex_unlock(&mdsc->mutex);
goto out;
}
+ if (req->r_got_safe && !head->safe) {
+ pr_warn("got unsafe after safe on %llu from mds%d\n",
+ tid, mds);
+ mutex_unlock(&mdsc->mutex);
+ goto out;
+ }
result = le32_to_cpu(head->result);
/*
- * Tolerate 2 consecutive ESTALEs from the same mds.
- * FIXME: we should be looking at the cap migrate_seq.
+ * Handle an ESTALE
+ * if we're not talking to the authority, send to them
+ * if the authority has changed while we weren't looking,
+ * send to new authority
+ * Otherwise we just have to return an ESTALE
*/
if (result == -ESTALE) {
- req->r_direct_mode = USE_AUTH_MDS;
- req->r_num_stale++;
- if (req->r_num_stale <= 2) {
+ dout("got ESTALE on request %llu", req->r_tid);
+ if (req->r_direct_mode != USE_AUTH_MDS) {
+ dout("not using auth, setting for that now");
+ req->r_direct_mode = USE_AUTH_MDS;
__do_request(mdsc, req);
mutex_unlock(&mdsc->mutex);
goto out;
+ } else {
+ int mds = __choose_mds(mdsc, req);
+ if (mds >= 0 && mds != req->r_session->s_mds) {
+ dout("but auth changed, so resending");
+ __do_request(mdsc, req);
+ mutex_unlock(&mdsc->mutex);
+ goto out;
+ }
}
- } else {
- req->r_num_stale = 0;
+ dout("have to return ESTALE on request %llu", req->r_tid);
}
+
if (head->safe) {
req->r_got_safe = true;
__unregister_request(mdsc, req);
- complete(&req->r_safe_completion);
if (req->r_got_unsafe) {
/*
@@ -1832,27 +2284,23 @@ 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;
}
- }
-
- BUG_ON(req->r_reply);
-
- if (!head->safe) {
+ } else {
req->r_got_unsafe = true;
list_add_tail(&req->r_unsafe_item, &req->r_session->s_unsafe);
}
dout("handle_reply tid %lld result %d\n", tid, result);
rinfo = &req->r_reply_info;
- err = parse_reply_info(msg, rinfo);
+ err = parse_reply_info(msg, rinfo, session->s_con.peer_features);
mutex_unlock(&mdsc->mutex);
mutex_lock(&session->s_mutex);
if (err < 0) {
- pr_err("mdsc_handle_reply got corrupt reply mds%d\n", mds);
+ pr_err("mdsc_handle_reply got corrupt reply mds%d(tid:%lld)\n", mds, tid);
ceph_msg_dump(msg);
goto out_err;
}
@@ -1869,23 +2317,33 @@ static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
}
/* insert trace into our cache */
- err = ceph_fill_trace(mdsc->client->sb, req, req->r_session);
+ mutex_lock(&req->r_fill_mutex);
+ err = ceph_fill_trace(mdsc->fsc->sb, req, req->r_session);
if (err == 0) {
- if (result == 0 && rinfo->dir_nr)
+ if (result == 0 && (req->r_op == CEPH_MDS_OP_READDIR ||
+ req->r_op == CEPH_MDS_OP_LSSNAP))
ceph_readdir_prepopulate(req, req->r_session);
- ceph_unreserve_caps(&req->r_caps_reservation);
+ ceph_unreserve_caps(mdsc, &req->r_caps_reservation);
}
+ mutex_unlock(&req->r_fill_mutex);
up_read(&mdsc->snap_rwsem);
out_err:
- if (err) {
- req->r_err = err;
+ mutex_lock(&mdsc->mutex);
+ if (!req->r_aborted) {
+ if (err) {
+ req->r_err = err;
+ } else {
+ req->r_reply = msg;
+ ceph_msg_get(msg);
+ req->r_got_result = true;
+ }
} else {
- req->r_reply = msg;
- ceph_msg_get(msg);
+ dout("reply arrived after request %lld was aborted\n", tid);
}
+ mutex_unlock(&mdsc->mutex);
- add_cap_releases(mdsc, req->r_session, -1);
+ ceph_add_cap_releases(mdsc, req->r_session);
mutex_unlock(&session->s_mutex);
/* kick calling process */
@@ -1919,16 +2377,21 @@ static void handle_forward(struct ceph_mds_client *mdsc,
mutex_lock(&mdsc->mutex);
req = __lookup_request(mdsc, tid);
if (!req) {
- dout("forward %llu to mds%d - req dne\n", tid, next_mds);
+ dout("forward tid %llu to mds%d - req dne\n", tid, next_mds);
goto out; /* dup reply? */
}
- if (fwd_seq <= req->r_num_fwd) {
- dout("forward %llu to mds%d - old seq %d <= %d\n",
+ if (req->r_aborted) {
+ dout("forward tid %llu aborted, unregistering\n", tid);
+ __unregister_request(mdsc, req);
+ } else if (fwd_seq <= req->r_num_fwd) {
+ dout("forward tid %llu to mds%d - old seq %d <= %d\n",
tid, next_mds, req->r_num_fwd, fwd_seq);
} else {
/* resend. forward race not possible; mds would drop */
- dout("forward %llu to mds%d (we resend)\n", tid, next_mds);
+ dout("forward tid %llu to mds%d (we resend)\n", tid, next_mds);
+ BUG_ON(req->r_err);
+ BUG_ON(req->r_got_result);
req->r_num_fwd = fwd_seq;
req->r_resend_mds = next_mds;
put_request_session(req);
@@ -1982,6 +2445,8 @@ static void handle_session(struct ceph_mds_session *session,
switch (op) {
case CEPH_SESSION_OPEN:
+ if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
+ pr_info("mds%d reconnect success\n", session->s_mds);
session->s_state = CEPH_MDS_SESSION_OPEN;
renewed_caps(mdsc, session, 0);
wake = 1;
@@ -1995,19 +2460,21 @@ static void handle_session(struct ceph_mds_session *session,
break;
case CEPH_SESSION_CLOSE:
+ if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
+ pr_info("mds%d reconnect denied\n", session->s_mds);
remove_session_caps(session);
wake = 1; /* for good measure */
- complete(&mdsc->session_close_waiters);
- kick_requests(mdsc, mds, 0); /* cur only */
+ wake_up_all(&mdsc->session_close_wq);
+ kick_requests(mdsc, mds);
break;
case CEPH_SESSION_STALE:
pr_info("mds%d caps went stale, renewing\n",
session->s_mds);
- spin_lock(&session->s_cap_lock);
+ spin_lock(&session->s_gen_ttl_lock);
session->s_cap_gen++;
- session->s_cap_ttl = 0;
- spin_unlock(&session->s_cap_lock);
+ session->s_cap_ttl = jiffies - 1;
+ spin_unlock(&session->s_gen_ttl_lock);
send_renew_caps(mdsc, session);
break;
@@ -2015,6 +2482,10 @@ static void handle_session(struct ceph_mds_session *session,
trim_caps(mdsc, session, le32_to_cpu(h->max_caps));
break;
+ case CEPH_SESSION_FLUSHMSG:
+ send_flushmsg_ack(mdsc, session, seq);
+ break;
+
default:
pr_err("mdsc_handle_session bad op %d mds%d\n", op, mds);
WARN_ON(1);
@@ -2064,9 +2535,14 @@ static void replay_unsafe_requests(struct ceph_mds_client *mdsc,
static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap,
void *arg)
{
- struct ceph_mds_cap_reconnect rec;
+ union {
+ struct ceph_mds_cap_reconnect v2;
+ struct ceph_mds_cap_reconnect_v1 v1;
+ } rec;
+ size_t reclen;
struct ceph_inode_info *ci;
- struct ceph_pagelist *pagelist = arg;
+ struct ceph_reconnect_state *recon_state = arg;
+ struct ceph_pagelist *pagelist = recon_state->pagelist;
char *path;
int pathlen, err;
u64 pathbase;
@@ -2086,7 +2562,7 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap,
path = ceph_mdsc_build_path(dentry, &pathlen, &pathbase, 0);
if (IS_ERR(path)) {
err = PTR_ERR(path);
- BUG_ON(err);
+ goto out_dput;
}
} else {
path = NULL;
@@ -2094,25 +2570,80 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap,
}
err = ceph_pagelist_encode_string(pagelist, path, pathlen);
if (err)
- goto out;
+ goto out_free;
- spin_lock(&inode->i_lock);
+ spin_lock(&ci->i_ceph_lock);
cap->seq = 0; /* reset cap seq */
cap->issue_seq = 0; /* and issue_seq */
- rec.cap_id = cpu_to_le64(cap->cap_id);
- rec.pathbase = cpu_to_le64(pathbase);
- rec.wanted = cpu_to_le32(__ceph_caps_wanted(ci));
- rec.issued = cpu_to_le32(cap->issued);
- rec.size = cpu_to_le64(inode->i_size);
- ceph_encode_timespec(&rec.mtime, &inode->i_mtime);
- ceph_encode_timespec(&rec.atime, &inode->i_atime);
- rec.snaprealm = cpu_to_le64(ci->i_snap_realm->ino);
- spin_unlock(&inode->i_lock);
-
- err = ceph_pagelist_append(pagelist, &rec, sizeof(rec));
+ cap->mseq = 0; /* and migrate_seq */
+ cap->cap_gen = cap->session->s_cap_gen;
+
+ if (recon_state->flock) {
+ rec.v2.cap_id = cpu_to_le64(cap->cap_id);
+ rec.v2.wanted = cpu_to_le32(__ceph_caps_wanted(ci));
+ rec.v2.issued = cpu_to_le32(cap->issued);
+ rec.v2.snaprealm = cpu_to_le64(ci->i_snap_realm->ino);
+ rec.v2.pathbase = cpu_to_le64(pathbase);
+ rec.v2.flock_len = 0;
+ reclen = sizeof(rec.v2);
+ } else {
+ rec.v1.cap_id = cpu_to_le64(cap->cap_id);
+ rec.v1.wanted = cpu_to_le32(__ceph_caps_wanted(ci));
+ rec.v1.issued = cpu_to_le32(cap->issued);
+ rec.v1.size = cpu_to_le64(inode->i_size);
+ ceph_encode_timespec(&rec.v1.mtime, &inode->i_mtime);
+ ceph_encode_timespec(&rec.v1.atime, &inode->i_atime);
+ rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino);
+ rec.v1.pathbase = cpu_to_le64(pathbase);
+ reclen = sizeof(rec.v1);
+ }
+ spin_unlock(&ci->i_ceph_lock);
+
+ if (recon_state->flock) {
+ int num_fcntl_locks, num_flock_locks;
+ struct ceph_filelock *flocks;
+
+encode_again:
+ spin_lock(&inode->i_lock);
+ ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks);
+ spin_unlock(&inode->i_lock);
+ flocks = kmalloc((num_fcntl_locks+num_flock_locks) *
+ sizeof(struct ceph_filelock), GFP_NOFS);
+ if (!flocks) {
+ err = -ENOMEM;
+ goto out_free;
+ }
+ spin_lock(&inode->i_lock);
+ err = ceph_encode_locks_to_buffer(inode, flocks,
+ num_fcntl_locks,
+ num_flock_locks);
+ spin_unlock(&inode->i_lock);
+ if (err) {
+ kfree(flocks);
+ if (err == -ENOSPC)
+ goto encode_again;
+ goto out_free;
+ }
+ /*
+ * number of encoded locks is stable, so copy to pagelist
+ */
+ rec.v2.flock_len = cpu_to_le32(2*sizeof(u32) +
+ (num_fcntl_locks+num_flock_locks) *
+ sizeof(struct ceph_filelock));
+ err = ceph_pagelist_append(pagelist, &rec, reclen);
+ if (!err)
+ err = ceph_locks_to_pagelist(flocks, pagelist,
+ num_fcntl_locks,
+ num_flock_locks);
+ kfree(flocks);
+ } else {
+ err = ceph_pagelist_append(pagelist, &rec, reclen);
+ }
-out:
+ recon_state->nr_caps++;
+out_free:
kfree(path);
+out_dput:
dput(dentry);
return err;
}
@@ -2130,61 +2661,76 @@ out:
*
* called with mdsc->mutex held.
*/
-static void send_mds_reconnect(struct ceph_mds_client *mdsc, int mds)
+static void send_mds_reconnect(struct ceph_mds_client *mdsc,
+ struct ceph_mds_session *session)
{
- struct ceph_mds_session *session = NULL;
struct ceph_msg *reply;
struct rb_node *p;
- int err;
+ int mds = session->s_mds;
+ int err = -ENOMEM;
+ int s_nr_caps;
struct ceph_pagelist *pagelist;
+ struct ceph_reconnect_state recon_state;
- pr_info("reconnect to recovering mds%d\n", mds);
+ pr_info("mds%d reconnect start\n", mds);
pagelist = kmalloc(sizeof(*pagelist), GFP_NOFS);
if (!pagelist)
goto fail_nopagelist;
ceph_pagelist_init(pagelist);
- reply = ceph_msg_new(CEPH_MSG_CLIENT_RECONNECT, 0, 0, 0, NULL);
- if (IS_ERR(reply)) {
- err = PTR_ERR(reply);
+ reply = ceph_msg_new(CEPH_MSG_CLIENT_RECONNECT, 0, GFP_NOFS, false);
+ if (!reply)
goto fail_nomsg;
- }
-
- /* find session */
- session = __ceph_lookup_mds_session(mdsc, mds);
- mutex_unlock(&mdsc->mutex); /* drop lock for duration */
- if (session) {
- mutex_lock(&session->s_mutex);
+ mutex_lock(&session->s_mutex);
+ session->s_state = CEPH_MDS_SESSION_RECONNECTING;
+ session->s_seq = 0;
- session->s_state = CEPH_MDS_SESSION_RECONNECTING;
- session->s_seq = 0;
+ ceph_con_close(&session->s_con);
+ ceph_con_open(&session->s_con,
+ CEPH_ENTITY_TYPE_MDS, mds,
+ ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
- ceph_con_open(&session->s_con,
- ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
-
- /* replay unsafe requests */
- replay_unsafe_requests(mdsc, session);
- } else {
- dout("no session for mds%d, will send short reconnect\n",
- mds);
- }
+ /* replay unsafe requests */
+ replay_unsafe_requests(mdsc, session);
down_read(&mdsc->snap_rwsem);
- if (!session)
- goto send;
dout("session %p state %s\n", session,
session_state_name(session->s_state));
+ spin_lock(&session->s_gen_ttl_lock);
+ session->s_cap_gen++;
+ spin_unlock(&session->s_gen_ttl_lock);
+
+ spin_lock(&session->s_cap_lock);
+ /*
+ * notify __ceph_remove_cap() that we are composing cap reconnect.
+ * If a cap get released before being added to the cap reconnect,
+ * __ceph_remove_cap() should skip queuing cap release.
+ */
+ session->s_cap_reconnect = 1;
+ /* drop old cap expires; we're about to reestablish that state */
+ discard_cap_releases(mdsc, session);
+ spin_unlock(&session->s_cap_lock);
+
/* traverse this session's caps */
- err = ceph_pagelist_encode_32(pagelist, session->s_nr_caps);
+ s_nr_caps = session->s_nr_caps;
+ err = ceph_pagelist_encode_32(pagelist, s_nr_caps);
if (err)
goto fail;
- err = iterate_session_caps(session, encode_caps_cb, pagelist);
+
+ recon_state.nr_caps = 0;
+ recon_state.pagelist = pagelist;
+ recon_state.flock = session->s_con.peer_features & CEPH_FEATURE_FLOCK;
+ err = iterate_session_caps(session, encode_caps_cb, &recon_state);
if (err < 0)
- goto out;
+ goto fail;
+
+ spin_lock(&session->s_cap_lock);
+ session->s_cap_reconnect = 0;
+ spin_unlock(&session->s_cap_lock);
/*
* snaprealms. we provide mds with the ino, seq (version), and
@@ -2206,34 +2752,41 @@ static void send_mds_reconnect(struct ceph_mds_client *mdsc, int mds)
goto fail;
}
-send:
- reply->pagelist = pagelist;
+ if (recon_state.flock)
+ reply->hdr.version = cpu_to_le16(2);
+
+ /* raced with cap release? */
+ if (s_nr_caps != recon_state.nr_caps) {
+ struct page *page = list_first_entry(&pagelist->head,
+ struct page, lru);
+ __le32 *addr = kmap_atomic(page);
+ *addr = cpu_to_le32(recon_state.nr_caps);
+ kunmap_atomic(addr);
+ }
+
reply->hdr.data_len = cpu_to_le32(pagelist->length);
- reply->nr_pages = calc_pages_for(0, pagelist->length);
+ ceph_msg_data_add_pagelist(reply, pagelist);
ceph_con_send(&session->s_con, reply);
- if (session) {
- session->s_state = CEPH_MDS_SESSION_OPEN;
- __wake_requests(mdsc, &session->s_waiting);
- }
+ mutex_unlock(&session->s_mutex);
-out:
- up_read(&mdsc->snap_rwsem);
- if (session) {
- mutex_unlock(&session->s_mutex);
- ceph_put_mds_session(session);
- }
mutex_lock(&mdsc->mutex);
+ __wake_requests(mdsc, &session->s_waiting);
+ mutex_unlock(&mdsc->mutex);
+
+ up_read(&mdsc->snap_rwsem);
return;
fail:
ceph_msg_put(reply);
+ up_read(&mdsc->snap_rwsem);
+ mutex_unlock(&session->s_mutex);
fail_nomsg:
ceph_pagelist_release(pagelist);
kfree(pagelist);
fail_nopagelist:
- pr_err("ENOMEM preparing reconnect for mds%d\n", mds);
- goto out;
+ pr_err("error %d preparing reconnect for mds%d\n", err, mds);
+ return;
}
@@ -2261,12 +2814,15 @@ static void check_new_map(struct ceph_mds_client *mdsc,
oldstate = ceph_mdsmap_get_state(oldmap, i);
newstate = ceph_mdsmap_get_state(newmap, i);
- dout("check_new_map mds%d state %s -> %s (session %s)\n",
+ dout("check_new_map mds%d state %s%s -> %s%s (session %s)\n",
i, ceph_mds_state_name(oldstate),
+ ceph_mdsmap_is_laggy(oldmap, i) ? " (laggy)" : "",
ceph_mds_state_name(newstate),
+ ceph_mdsmap_is_laggy(newmap, i) ? " (laggy)" : "",
session_state_name(s->s_state));
- if (memcmp(ceph_mdsmap_get_addr(oldmap, i),
+ if (i >= newmap->m_max_mds ||
+ memcmp(ceph_mdsmap_get_addr(oldmap, i),
ceph_mdsmap_get_addr(newmap, i),
sizeof(struct ceph_entity_addr))) {
if (s->s_state == CEPH_MDS_SESSION_OPENING) {
@@ -2285,7 +2841,7 @@ static void check_new_map(struct ceph_mds_client *mdsc,
}
/* kick any requests waiting on the recovering mds */
- kick_requests(mdsc, i, 1);
+ kick_requests(mdsc, i);
} else if (oldstate == newstate) {
continue; /* nothing new with this mds */
}
@@ -2294,26 +2850,40 @@ static void check_new_map(struct ceph_mds_client *mdsc,
* send reconnect?
*/
if (s->s_state == CEPH_MDS_SESSION_RESTARTING &&
- newstate >= CEPH_MDS_STATE_RECONNECT)
- send_mds_reconnect(mdsc, i);
+ newstate >= CEPH_MDS_STATE_RECONNECT) {
+ mutex_unlock(&mdsc->mutex);
+ send_mds_reconnect(mdsc, s);
+ mutex_lock(&mdsc->mutex);
+ }
/*
- * kick requests on any mds that has gone active.
- *
- * kick requests on cur or forwarder: we may have sent
- * the request to mds1, mds1 told us it forwarded it
- * to mds2, but then we learn mds1 failed and can't be
- * sure it successfully forwarded our request before
- * it died.
+ * kick request on any mds that has gone active.
*/
if (oldstate < CEPH_MDS_STATE_ACTIVE &&
newstate >= CEPH_MDS_STATE_ACTIVE) {
- pr_info("mds%d reconnect completed\n", s->s_mds);
- kick_requests(mdsc, i, 1);
+ if (oldstate != CEPH_MDS_STATE_CREATING &&
+ oldstate != CEPH_MDS_STATE_STARTING)
+ pr_info("mds%d recovery completed\n", s->s_mds);
+ kick_requests(mdsc, i);
ceph_kick_flushing_caps(mdsc, s);
wake_up_session_caps(s, 1);
}
}
+
+ for (i = 0; i < newmap->m_max_mds && i < mdsc->max_sessions; i++) {
+ s = mdsc->sessions[i];
+ if (!s)
+ continue;
+ if (!ceph_mdsmap_is_laggy(newmap, i))
+ continue;
+ if (s->s_state == CEPH_MDS_SESSION_OPEN ||
+ s->s_state == CEPH_MDS_SESSION_HUNG ||
+ s->s_state == CEPH_MDS_SESSION_CLOSING) {
+ dout(" connecting to export targets of laggy mds%d\n",
+ i);
+ __open_export_target_sessions(mdsc, s);
+ }
+ }
}
@@ -2337,15 +2907,14 @@ static void handle_lease(struct ceph_mds_client *mdsc,
struct ceph_mds_session *session,
struct ceph_msg *msg)
{
- struct super_block *sb = mdsc->client->sb;
+ struct super_block *sb = mdsc->fsc->sb;
struct inode *inode;
- struct ceph_inode_info *ci;
struct dentry *parent, *dentry;
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;
int release = 0;
@@ -2356,7 +2925,7 @@ static void handle_lease(struct ceph_mds_client *mdsc,
goto bad;
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))
@@ -2367,13 +2936,13 @@ 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, ino %llx %p %.*s\n",
+ ceph_lease_op_name(h->action), vino.ino, inode,
+ dname.len, dname.name);
if (inode == NULL) {
dout("handle_lease no inode %llx\n", vino.ino);
goto release;
}
- ci = ceph_inode(inode);
/* dentry */
parent = d_find_alias(inode);
@@ -2392,22 +2961,23 @@ static void handle_lease(struct ceph_mds_client *mdsc,
di = ceph_dentry(dentry);
switch (h->action) {
case CEPH_MDS_LEASE_REVOKE:
- if (di && di->lease_session == session) {
- h->seq = cpu_to_le32(di->lease_seq);
+ if (di->lease_session == session) {
+ 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;
break;
case CEPH_MDS_LEASE_RENEW:
- if (di && di->lease_session == session &&
+ if (di->lease_session == session &&
di->lease_gen == session->s_cap_gen &&
di->lease_renew_from &&
di->lease_renew_after == 0) {
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);
@@ -2452,12 +3022,11 @@ void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
dnamelen = dentry->d_name.len;
len += dnamelen;
- msg = ceph_msg_new(CEPH_MSG_CLIENT_LEASE, len, 0, 0, NULL);
- if (IS_ERR(msg))
+ msg = ceph_msg_new(CEPH_MSG_CLIENT_LEASE, len, GFP_NOFS, false);
+ if (!msg)
return;
lease = msg->front.iov_base;
lease->action = action;
- lease->mask = cpu_to_le16(CEPH_LOCK_DN);
lease->ino = cpu_to_le64(ceph_vino(inode).ino);
lease->first = lease->last = cpu_to_le64(ceph_vino(inode).snap);
lease->seq = cpu_to_le32(seq);
@@ -2479,7 +3048,7 @@ void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
* Pass @inode always, @dentry is optional.
*/
void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode,
- struct dentry *dentry, int mask)
+ struct dentry *dentry)
{
struct ceph_dentry_info *di;
struct ceph_mds_session *session;
@@ -2487,7 +3056,6 @@ void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode,
BUG_ON(inode == NULL);
BUG_ON(dentry == NULL);
- BUG_ON(mask != CEPH_LOCK_DN);
/* is dentry lease valid? */
spin_lock(&dentry->d_lock);
@@ -2497,8 +3065,8 @@ void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode,
di->lease_gen != di->lease_session->s_cap_gen ||
!time_before(jiffies, dentry->d_time)) {
dout("lease_release inode %p dentry %p -- "
- "no lease on %d\n",
- inode, dentry, mask);
+ "no lease\n",
+ inode, dentry);
spin_unlock(&dentry->d_lock);
return;
}
@@ -2509,8 +3077,8 @@ void ceph_mdsc_lease_release(struct ceph_mds_client *mdsc, struct inode *inode,
__ceph_mdsc_drop_dentry_lease(dentry);
spin_unlock(&dentry->d_lock);
- dout("lease_release inode %p dentry %p mask %d to mds%d\n",
- inode, dentry, mask, session->s_mds);
+ dout("lease_release inode %p dentry %p to mds%d\n",
+ inode, dentry, session->s_mds);
ceph_mdsc_lease_send_msg(session, inode, dentry,
CEPH_MDS_LEASE_RELEASE, seq);
ceph_put_mds_session(session);
@@ -2597,8 +3165,10 @@ 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);
- send_cap_releases(mdsc, s);
+ ceph_add_cap_releases(mdsc, s);
+ if (s->s_state == CEPH_MDS_SESSION_OPEN ||
+ s->s_state == CEPH_MDS_SESSION_HUNG)
+ ceph_send_cap_releases(mdsc, s);
mutex_unlock(&s->s_mutex);
ceph_put_mds_session(s);
@@ -2609,14 +3179,25 @@ static void delayed_work(struct work_struct *work)
schedule_delayed(mdsc);
}
+int ceph_mdsc_init(struct ceph_fs_client *fsc)
-int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client)
{
- mdsc->client = client;
+ struct ceph_mds_client *mdsc;
+
+ mdsc = kzalloc(sizeof(struct ceph_mds_client), GFP_NOFS);
+ if (!mdsc)
+ return -ENOMEM;
+ mdsc->fsc = fsc;
+ fsc->mdsc = mdsc;
mutex_init(&mdsc->mutex);
mdsc->mdsmap = kzalloc(sizeof(*mdsc->mdsmap), GFP_NOFS);
+ if (mdsc->mdsmap == NULL) {
+ kfree(mdsc);
+ return -ENOMEM;
+ }
+
init_completion(&mdsc->safe_umount_waiters);
- init_completion(&mdsc->session_close_waiters);
+ init_waitqueue_head(&mdsc->session_close_wq);
INIT_LIST_HEAD(&mdsc->waiting_for_map);
mdsc->sessions = NULL;
mdsc->max_sessions = 0;
@@ -2635,11 +3216,16 @@ int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client)
spin_lock_init(&mdsc->snap_flush_lock);
mdsc->cap_flush_seq = 0;
INIT_LIST_HEAD(&mdsc->cap_dirty);
+ INIT_LIST_HEAD(&mdsc->cap_dirty_migrating);
mdsc->num_cap_flushing = 0;
spin_lock_init(&mdsc->cap_dirty_lock);
init_waitqueue_head(&mdsc->cap_flushing_wq);
spin_lock_init(&mdsc->dentry_lru_lock);
INIT_LIST_HEAD(&mdsc->dentry_lru);
+
+ ceph_caps_init(mdsc);
+ ceph_adjust_min_caps(mdsc, fsc->min_caps);
+
return 0;
}
@@ -2650,7 +3236,7 @@ int ceph_mdsc_init(struct ceph_mds_client *mdsc, struct ceph_client *client)
static void wait_requests(struct ceph_mds_client *mdsc)
{
struct ceph_mds_request *req;
- struct ceph_client *client = mdsc->client;
+ struct ceph_fs_client *fsc = mdsc->fsc;
mutex_lock(&mdsc->mutex);
if (__get_oldest_req(mdsc)) {
@@ -2658,7 +3244,7 @@ static void wait_requests(struct ceph_mds_client *mdsc)
dout("wait_requests waiting for requests\n");
wait_for_completion_timeout(&mdsc->safe_umount_waiters,
- client->mount_args->mount_timeout * HZ);
+ fsc->client->options->mount_timeout * HZ);
/* tear down remaining requests */
mutex_lock(&mdsc->mutex);
@@ -2684,6 +3270,12 @@ 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();
}
/*
@@ -2735,6 +3327,9 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc)
{
u64 want_tid, want_flush;
+ if (mdsc->fsc->mount_state == CEPH_MOUNT_SHUTDOWN)
+ return;
+
dout("sync\n");
mutex_lock(&mdsc->mutex);
want_tid = mdsc->last_tid;
@@ -2748,6 +3343,23 @@ void ceph_mdsc_sync(struct ceph_mds_client *mdsc)
wait_event(mdsc->cap_flushing_wq, check_cap_flush(mdsc, want_flush));
}
+/*
+ * true if all sessions are closed, or we force unmount
+ */
+static bool done_closing_sessions(struct ceph_mds_client *mdsc)
+{
+ int i, n = 0;
+
+ if (mdsc->fsc->mount_state == CEPH_MOUNT_SHUTDOWN)
+ return true;
+
+ mutex_lock(&mdsc->mutex);
+ for (i = 0; i < mdsc->max_sessions; i++)
+ if (mdsc->sessions[i])
+ n++;
+ mutex_unlock(&mdsc->mutex);
+ return n == 0;
+}
/*
* called after sb is ro.
@@ -2756,45 +3368,32 @@ void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc)
{
struct ceph_mds_session *session;
int i;
- int n;
- struct ceph_client *client = mdsc->client;
- unsigned long started, timeout = client->mount_args->mount_timeout * HZ;
+ struct ceph_fs_client *fsc = mdsc->fsc;
+ unsigned long timeout = fsc->client->options->mount_timeout * HZ;
dout("close_sessions\n");
- mutex_lock(&mdsc->mutex);
-
/* close sessions */
- started = jiffies;
- while (time_before(jiffies, started + timeout)) {
- dout("closing sessions\n");
- n = 0;
- for (i = 0; i < mdsc->max_sessions; i++) {
- session = __ceph_lookup_mds_session(mdsc, i);
- if (!session)
- continue;
- mutex_unlock(&mdsc->mutex);
- mutex_lock(&session->s_mutex);
- __close_session(mdsc, session);
- mutex_unlock(&session->s_mutex);
- ceph_put_mds_session(session);
- mutex_lock(&mdsc->mutex);
- n++;
- }
- if (n == 0)
- break;
-
- if (client->mount_state == CEPH_MOUNT_SHUTDOWN)
- break;
-
- dout("waiting for sessions to close\n");
+ mutex_lock(&mdsc->mutex);
+ for (i = 0; i < mdsc->max_sessions; i++) {
+ session = __ceph_lookup_mds_session(mdsc, i);
+ if (!session)
+ continue;
mutex_unlock(&mdsc->mutex);
- wait_for_completion_timeout(&mdsc->session_close_waiters,
- timeout);
+ mutex_lock(&session->s_mutex);
+ __close_session(mdsc, session);
+ mutex_unlock(&session->s_mutex);
+ ceph_put_mds_session(session);
mutex_lock(&mdsc->mutex);
}
+ mutex_unlock(&mdsc->mutex);
+
+ dout("waiting for sessions to close\n");
+ wait_event_timeout(mdsc->session_close_wq, done_closing_sessions(mdsc),
+ timeout);
/* tear down remaining sessions */
+ mutex_lock(&mdsc->mutex);
for (i = 0; i < mdsc->max_sessions; i++) {
if (mdsc->sessions[i]) {
session = get_session(mdsc->sessions[i]);
@@ -2807,9 +3406,7 @@ void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc)
mutex_lock(&mdsc->mutex);
}
}
-
WARN_ON(!list_empty(&mdsc->cap_delay_list));
-
mutex_unlock(&mdsc->mutex);
ceph_cleanup_empty_realms(mdsc);
@@ -2819,13 +3416,29 @@ void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc)
dout("stopped\n");
}
-void ceph_mdsc_stop(struct ceph_mds_client *mdsc)
+static void ceph_mdsc_stop(struct ceph_mds_client *mdsc)
{
dout("stop\n");
cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */
if (mdsc->mdsmap)
ceph_mdsmap_destroy(mdsc->mdsmap);
kfree(mdsc->sessions);
+ ceph_caps_finalize(mdsc);
+}
+
+void ceph_mdsc_destroy(struct ceph_fs_client *fsc)
+{
+ struct ceph_mds_client *mdsc = fsc->mdsc;
+
+ dout("mdsc_destroy %p\n", mdsc);
+ ceph_mdsc_stop(mdsc);
+
+ /* flush out any connection work with references to us */
+ ceph_msgr_flush();
+
+ fsc->mdsc = NULL;
+ kfree(mdsc);
+ dout("mdsc_destroy %p done\n", mdsc);
}
@@ -2844,14 +3457,14 @@ void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
ceph_decode_need(&p, end, sizeof(fsid)+2*sizeof(u32), bad);
ceph_decode_copy(&p, &fsid, sizeof(fsid));
- if (ceph_check_fsid(mdsc->client, &fsid) < 0)
+ if (ceph_check_fsid(mdsc->fsc->client, &fsid) < 0)
return;
epoch = ceph_decode_32(&p);
maplen = ceph_decode_32(&p);
dout("handle_map epoch %u len %d\n", epoch, (int)maplen);
/* do we need it? */
- ceph_monc_got_mdsmap(&mdsc->client->monc, epoch);
+ ceph_monc_got_mdsmap(&mdsc->fsc->client->monc, epoch);
mutex_lock(&mdsc->mutex);
if (mdsc->mdsmap && epoch <= mdsc->mdsmap->m_epoch) {
dout("handle_map epoch %u <= our %u\n",
@@ -2875,7 +3488,7 @@ void ceph_mdsc_handle_map(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
} else {
mdsc->mdsmap = newmap; /* first mds map */
}
- mdsc->client->sb->s_maxbytes = mdsc->mdsmap->m_max_file_size;
+ mdsc->fsc->sb->s_maxbytes = mdsc->mdsmap->m_max_file_size;
__wake_requests(mdsc, &mdsc->waiting_for_map);
@@ -2906,8 +3519,8 @@ static void con_put(struct ceph_connection *con)
{
struct ceph_mds_session *s = con->private;
+ dout("mdsc con_put %p (%d)\n", s, atomic_read(&s->s_ref) - 1);
ceph_put_mds_session(s);
- dout("mdsc con_put %p (%d)\n", s, atomic_read(&s->s_ref));
}
/*
@@ -2917,9 +3530,10 @@ static void con_put(struct ceph_connection *con)
static void peer_reset(struct ceph_connection *con)
{
struct ceph_mds_session *s = con->private;
+ struct ceph_mds_client *mdsc = s->s_mdsc;
- pr_err("mds%d gave us the boot. IMPLEMENT RECONNECT.\n",
- s->s_mds);
+ pr_warn("mds%d closed our session\n", s->s_mds);
+ send_mds_reconnect(mdsc, s);
}
static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
@@ -2969,39 +3583,37 @@ out:
/*
* authentication
*/
-static int get_authorizer(struct ceph_connection *con,
- void **buf, int *len, int *proto,
- void **reply_buf, int *reply_len, int force_new)
+
+/*
+ * Note: returned pointer is the address of a structure that's
+ * managed separately. Caller must *not* attempt to free it.
+ */
+static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
+ int *proto, int force_new)
{
struct ceph_mds_session *s = con->private;
struct ceph_mds_client *mdsc = s->s_mdsc;
- struct ceph_auth_client *ac = mdsc->client->monc.auth;
- int ret = 0;
-
- if (force_new && s->s_authorizer) {
- ac->ops->destroy_authorizer(ac, s->s_authorizer);
- s->s_authorizer = NULL;
- }
- if (s->s_authorizer == NULL) {
- if (ac->ops->create_authorizer) {
- ret = ac->ops->create_authorizer(
- ac, CEPH_ENTITY_TYPE_MDS,
- &s->s_authorizer,
- &s->s_authorizer_buf,
- &s->s_authorizer_buf_len,
- &s->s_authorizer_reply_buf,
- &s->s_authorizer_reply_buf_len);
- if (ret)
- return ret;
- }
+ struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
+ struct ceph_auth_handshake *auth = &s->s_auth;
+
+ if (force_new && auth->authorizer) {
+ ceph_auth_destroy_authorizer(ac, auth->authorizer);
+ auth->authorizer = NULL;
+ }
+ if (!auth->authorizer) {
+ int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_MDS,
+ auth);
+ if (ret)
+ return ERR_PTR(ret);
+ } else {
+ int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_MDS,
+ auth);
+ if (ret)
+ return ERR_PTR(ret);
}
-
*proto = ac->protocol;
- *buf = s->s_authorizer_buf;
- *len = s->s_authorizer_buf_len;
- *reply_buf = s->s_authorizer_reply_buf;
- *reply_len = s->s_authorizer_reply_buf_len;
- return 0;
+
+ return auth;
}
@@ -3009,24 +3621,44 @@ static int verify_authorizer_reply(struct ceph_connection *con, int len)
{
struct ceph_mds_session *s = con->private;
struct ceph_mds_client *mdsc = s->s_mdsc;
- struct ceph_auth_client *ac = mdsc->client->monc.auth;
+ struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
- return ac->ops->verify_authorizer_reply(ac, s->s_authorizer, len);
+ return ceph_auth_verify_authorizer_reply(ac, s->s_auth.authorizer, len);
}
static int invalidate_authorizer(struct ceph_connection *con)
{
struct ceph_mds_session *s = con->private;
struct ceph_mds_client *mdsc = s->s_mdsc;
- struct ceph_auth_client *ac = mdsc->client->monc.auth;
+ struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
- if (ac->ops->invalidate_authorizer)
- ac->ops->invalidate_authorizer(ac, CEPH_ENTITY_TYPE_MDS);
+ ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_MDS);
- return ceph_monc_validate_auth(&mdsc->client->monc);
+ return ceph_monc_validate_auth(&mdsc->fsc->client->monc);
}
-const static struct ceph_connection_operations mds_con_ops = {
+static struct ceph_msg *mds_alloc_msg(struct ceph_connection *con,
+ struct ceph_msg_header *hdr, int *skip)
+{
+ struct ceph_msg *msg;
+ int type = (int) le16_to_cpu(hdr->type);
+ int front_len = (int) le32_to_cpu(hdr->front_len);
+
+ if (con->in_msg)
+ return con->in_msg;
+
+ *skip = 0;
+ msg = ceph_msg_new(type, front_len, GFP_NOFS, false);
+ if (!msg) {
+ pr_err("unable to allocate msg type %d len %d\n",
+ type, front_len);
+ return NULL;
+ }
+
+ return msg;
+}
+
+static const struct ceph_connection_operations mds_con_ops = {
.get = con_get,
.put = con_put,
.dispatch = dispatch,
@@ -3034,9 +3666,7 @@ const static struct ceph_connection_operations mds_con_ops = {
.verify_authorizer_reply = verify_authorizer_reply,
.invalidate_authorizer = invalidate_authorizer,
.peer_reset = peer_reset,
+ .alloc_msg = mds_alloc_msg,
};
-
-
-
/* eof */
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-12-05 23:05:52 +0000