unmap_mapping_pages(vnode->netfs.inode.i_mapping, 0, 0, false);
}
-static void afs_server_init_callback(struct afs_server *server)
+static void afs_volume_init_callback(struct afs_volume *volume)
{
struct afs_vnode *vnode;
- struct afs_cell *cell = server->cell;
- down_read(&cell->fs_open_mmaps_lock);
+ down_read(&volume->open_mmaps_lock);
- list_for_each_entry(vnode, &cell->fs_open_mmaps, cb_mmap_link) {
- if (vnode->cb_server == server) {
- clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags);
+ list_for_each_entry(vnode, &volume->open_mmaps, cb_mmap_link) {
+ if (vnode->cb_v_check != atomic_read(&volume->cb_v_break)) {
+ atomic64_set(&vnode->cb_expires_at, AFS_NO_CB_PROMISE);
queue_work(system_unbound_wq, &vnode->cb_work);
}
}
- up_read(&cell->fs_open_mmaps_lock);
+ up_read(&volume->open_mmaps_lock);
}
/*
*/
void afs_init_callback_state(struct afs_server *server)
{
- struct afs_cell *cell = server->cell;
+ struct afs_server_entry *se;
- down_read(&cell->vs_lock);
+ down_read(&server->cell->vs_lock);
- do {
- server->cb_s_break++;
- atomic_inc(&server->cell->fs_s_break);
- if (!list_empty(&server->cell->fs_open_mmaps))
- afs_server_init_callback(server);
-
- } while ((server = rcu_dereference(server->uuid_next)));
+ list_for_each_entry(se, &server->volumes, slink) {
+ se->cb_expires_at = AFS_NO_CB_PROMISE;
+ se->volume->cb_expires_at = AFS_NO_CB_PROMISE;
+ trace_afs_cb_v_break(se->volume->vid, atomic_read(&se->volume->cb_v_break),
+ afs_cb_break_for_s_reinit);
+ if (!list_empty(&se->volume->open_mmaps))
+ afs_volume_init_callback(se->volume);
+ }
- up_read(&cell->vs_lock);
+ up_read(&server->cell->vs_lock);
}
/*
_enter("");
clear_bit(AFS_VNODE_NEW_CONTENT, &vnode->flags);
- if (test_and_clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags)) {
+ if (atomic64_xchg(&vnode->cb_expires_at, AFS_NO_CB_PROMISE) != AFS_NO_CB_PROMISE) {
vnode->cb_break++;
- vnode->cb_v_break = atomic_read(&vnode->volume->cb_v_break);
+ vnode->cb_v_check = atomic_read(&vnode->volume->cb_v_break);
afs_clear_permits(vnode);
if (vnode->lock_state == AFS_VNODE_LOCK_WAITING_FOR_CB)
return volume;
}
+/*
+ * Allow the fileserver to break callbacks at the volume-level. This is
+ * typically done when, for example, a R/W volume is snapshotted to a R/O
+ * volume (the only way to change an R/O volume). It may also, however, happen
+ * when a volserver takes control of a volume (offlining it, moving it, etc.).
+ *
+ * Every file in that volume will need to be reevaluated.
+ */
+static void afs_break_volume_callback(struct afs_server *server,
+ struct afs_volume *volume)
+ __releases(RCU)
+{
+ struct afs_server_list *slist = rcu_dereference(volume->servers);
+ unsigned int i, cb_v_break;
+
+ write_lock(&volume->cb_v_break_lock);
+
+ for (i = 0; i < slist->nr_servers; i++)
+ if (slist->servers[i].server == server)
+ slist->servers[i].cb_expires_at = AFS_NO_CB_PROMISE;
+ volume->cb_expires_at = AFS_NO_CB_PROMISE;
+
+ cb_v_break = atomic_inc_return_release(&volume->cb_v_break);
+ trace_afs_cb_v_break(volume->vid, cb_v_break, afs_cb_break_for_volume_callback);
+
+ write_unlock(&volume->cb_v_break_lock);
+ rcu_read_unlock();
+
+ if (!list_empty(&volume->open_mmaps))
+ afs_volume_init_callback(volume);
+}
+
/*
* allow the fileserver to explicitly break one callback
* - happens when
* - the backing file is changed
* - a lock is released
*/
-static void afs_break_one_callback(struct afs_volume *volume,
+static void afs_break_one_callback(struct afs_server *server,
+ struct afs_volume *volume,
struct afs_fid *fid)
{
struct super_block *sb;
struct afs_vnode *vnode;
struct inode *inode;
- unsigned int cb_v_break;
-
- if (fid->vnode == 0 && fid->unique == 0) {
- /* The callback break applies to an entire volume. */
- write_lock(&volume->cb_v_break_lock);
- cb_v_break = atomic_inc_return(&volume->cb_v_break);
- trace_afs_cb_break(fid, cb_v_break,
- afs_cb_break_for_volume_callback, false);
- write_unlock(&volume->cb_v_break_lock);
- return;
- }
/* See if we can find a matching inode - even an I_NEW inode needs to
* be marked as it can have its callback broken before we finish
rcu_read_lock();
volume = afs_lookup_volume_rcu(server->cell, vid);
- /* TODO: Find all matching volumes if we couldn't match the server and
- * break them anyway.
- */
- for (i = *_count; i > 0; cbb++, i--) {
- if (cbb->fid.vid == vid) {
- _debug("- Fid { vl=%08llx n=%llu u=%u }",
- cbb->fid.vid,
- cbb->fid.vnode,
- cbb->fid.unique);
- --*_count;
- if (volume)
- afs_break_one_callback(volume, &cbb->fid);
- } else {
- *residue++ = *cbb;
+ if (cbb->fid.vnode == 0 && cbb->fid.unique == 0) {
+ afs_break_volume_callback(server, volume);
+ *_count -= 1;
+ if (*_count)
+ memmove(cbb, cbb + 1, sizeof(*cbb) * *_count);
+ } else {
+ /* TODO: Find all matching volumes if we couldn't match the server and
+ * break them anyway.
+ */
+
+ for (i = *_count; i > 0; cbb++, i--) {
+ if (cbb->fid.vid == vid) {
+ _debug("- Fid { vl=%08llx n=%llu u=%u }",
+ cbb->fid.vid,
+ cbb->fid.vnode,
+ cbb->fid.unique);
+ --*_count;
+ if (volume)
+ afs_break_one_callback(server, volume, &cbb->fid);
+ } else {
+ *residue++ = *cbb;
+ }
}
+ rcu_read_unlock();
}
- rcu_read_unlock();
afs_put_volume(volume, afs_volume_trace_put_callback);
}
seqlock_init(&cell->volume_lock);
cell->fs_servers = RB_ROOT;
seqlock_init(&cell->fs_lock);
- INIT_LIST_HEAD(&cell->fs_open_mmaps);
- init_rwsem(&cell->fs_open_mmaps_lock);
rwlock_init(&cell->vl_servers_lock);
cell->flags = (1 << AFS_CELL_FL_CHECK_ALIAS);
dir = AFS_FS_I(d_inode(parent));
/* validate the parent directory */
- afs_validate(dir, key);
+ ret = afs_validate(dir, key);
+ if (ret == -ERESTARTSYS) {
+ dput(parent);
+ key_put(key);
+ return ret;
+ }
if (test_bit(AFS_VNODE_DELETED, &dir->flags)) {
_debug("%pd: parent dir deleted", dentry);
switch (afs_op_abort_code(op)) {
case VNOVNODE:
set_bit(AFS_VNODE_DELETED, &vnode->flags);
+ clear_nlink(&vnode->netfs.inode);
afs_break_callback(vnode, afs_cb_break_for_deleted);
}
}
clear_nlink(&vnode->netfs.inode);
set_bit(AFS_VNODE_DELETED, &vnode->flags);
- clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags);
+ atomic64_set(&vnode->cb_expires_at, AFS_NO_CB_PROMISE);
clear_bit(AFS_VNODE_DIR_VALID, &vnode->flags);
}
}
static void afs_add_open_mmap(struct afs_vnode *vnode)
{
if (atomic_inc_return(&vnode->cb_nr_mmap) == 1) {
- down_write(&vnode->volume->cell->fs_open_mmaps_lock);
+ down_write(&vnode->volume->open_mmaps_lock);
if (list_empty(&vnode->cb_mmap_link))
- list_add_tail(&vnode->cb_mmap_link,
- &vnode->volume->cell->fs_open_mmaps);
+ list_add_tail(&vnode->cb_mmap_link, &vnode->volume->open_mmaps);
- up_write(&vnode->volume->cell->fs_open_mmaps_lock);
+ up_write(&vnode->volume->open_mmaps_lock);
}
}
if (!atomic_dec_and_test(&vnode->cb_nr_mmap))
return;
- down_write(&vnode->volume->cell->fs_open_mmaps_lock);
+ down_write(&vnode->volume->open_mmaps_lock);
if (atomic_read(&vnode->cb_nr_mmap) == 0)
list_del_init(&vnode->cb_mmap_link);
- up_write(&vnode->volume->cell->fs_open_mmaps_lock);
+ up_write(&vnode->volume->open_mmaps_lock);
flush_work(&vnode->cb_work);
}
{
struct afs_vnode *vnode = AFS_FS_I(file_inode(vmf->vma->vm_file));
- if (afs_pagecache_valid(vnode))
+ if (afs_check_validity(vnode))
return filemap_map_pages(vmf, start_pgoff, end_pgoff);
return 0;
}
op->pre_volsync.creation = volume->creation_time;
op->pre_volsync.update = volume->update_time;
op->debug_id = atomic_inc_return(&afs_operation_debug_counter);
- op->nr_iterations = -1;
+ op->nr_iterations = -1;
afs_op_set_error(op, -EDESTADDRREQ);
_leave(" = [op=%08x]", op->debug_id);
op->call_responded = false;
op->call_error = 0;
op->call_abort_code = 0;
- op->cb_s_break = op->server->cb_s_break;
if (test_bit(AFS_SERVER_FL_IS_YFS, &op->server->flags) &&
op->ops->issue_yfs_rpc)
op->ops->issue_yfs_rpc(op);
write_seqlock(&vnode->cb_lock);
- vnode->cb_v_break = op->cb_v_break;
- vnode->cb_s_break = op->cb_s_break;
+ vnode->cb_v_check = op->cb_v_break;
vnode->status = *status;
t = status->mtime_client;
if (!vp->scb.have_cb) {
/* it's a symlink we just created (the fileserver
* didn't give us a callback) */
- vnode->cb_expires_at = ktime_get_real_seconds();
+ atomic64_set(&vnode->cb_expires_at, AFS_NO_CB_PROMISE);
} else {
- vnode->cb_expires_at = vp->scb.callback.expires_at;
vnode->cb_server = op->server;
- set_bit(AFS_VNODE_CB_PROMISED, &vnode->flags);
+ atomic64_set(&vnode->cb_expires_at, vp->scb.callback.expires_at);
}
write_sequnlock(&vnode->cb_lock);
vnode->status = *status;
if (vp->dv_before + vp->dv_delta != status->data_version) {
- if (test_bit(AFS_VNODE_CB_PROMISED, &vnode->flags))
+ if (vnode->cb_ro_snapshot == atomic_read(&vnode->volume->cb_ro_snapshot) &&
+ atomic64_read(&vnode->cb_expires_at) != AFS_NO_CB_PROMISE)
pr_warn("kAFS: vnode modified {%llx:%llu} %llx->%llx %s (op=%x)\n",
vnode->fid.vid, vnode->fid.vnode,
(unsigned long long)vp->dv_before + vp->dv_delta,
struct afs_vnode *vnode = vp->vnode;
if (!afs_cb_is_broken(vp->cb_break_before, vnode)) {
- vnode->cb_expires_at = cb->expires_at;
- vnode->cb_server = op->server;
- set_bit(AFS_VNODE_CB_PROMISED, &vnode->flags);
+ if (op->volume->type == AFSVL_RWVOL)
+ vnode->cb_server = op->server;
+ atomic64_set(&vnode->cb_expires_at, cb->expires_at);
}
}
BUG_ON(!(inode->i_state & I_NEW));
vnode = AFS_FS_I(inode);
- vnode->cb_v_break = atomic_read(&as->volume->cb_v_break),
+ vnode->cb_v_check = atomic_read(&as->volume->cb_v_break),
afs_set_netfs_context(vnode);
op = afs_alloc_operation(key, as->volume);
if (vnode->volume &&
!(query_flags & AT_STATX_DONT_SYNC) &&
- !test_bit(AFS_VNODE_CB_PROMISED, &vnode->flags)) {
+ atomic64_read(&vnode->cb_expires_at) == AFS_NO_CB_PROMISE) {
key = afs_request_key(vnode->volume->cell);
if (IS_ERR(key))
return PTR_ERR(key);
/* Active fileserver interaction state. */
struct rb_root fs_servers; /* afs_server (by server UUID) */
seqlock_t fs_lock; /* For fs_servers */
- struct rw_semaphore fs_open_mmaps_lock;
- struct list_head fs_open_mmaps; /* List of vnodes that are mmapped */
- atomic_t fs_s_break; /* Counter of CB.InitCallBackState messages */
/* VL server list. */
rwlock_t vl_servers_lock; /* Lock on vl_servers */
/* file service access */
rwlock_t fs_lock; /* access lock */
- /* callback promise management */
- unsigned cb_s_break; /* Break-everything counter. */
-
/* Probe state */
struct afs_endpoint_state __rcu *endpoint_state; /* Latest endpoint/probe state */
unsigned long probed_at; /* Time last probe was dispatched (jiffies) */
struct afs_server *server;
struct afs_volume *volume;
struct list_head slink; /* Link in server->volumes */
+ time64_t cb_expires_at; /* Time at which volume-level callback expires */
unsigned long flags;
#define AFS_SE_EXCLUDED 0 /* Set if server is to be excluded in rotation */
};
time64_t update_time; /* Volume update time (or TIME64_MIN) */
/* Callback management */
+ struct mutex cb_check_lock; /* Lock to control race to check after v_break */
+ time64_t cb_expires_at; /* Earliest volume callback expiry time */
atomic_t cb_ro_snapshot; /* RO volume update-from-snapshot counter */
atomic_t cb_v_break; /* Volume-break event counter. */
+ atomic_t cb_v_check; /* Volume-break has-been-checked counter. */
atomic_t cb_scrub; /* Scrub-all-data event counter. */
rwlock_t cb_v_break_lock;
+ struct rw_semaphore open_mmaps_lock;
+ struct list_head open_mmaps; /* List of vnodes that are mmapped */
afs_voltype_t type; /* type of volume */
char type_force; /* force volume type (suppress R/O -> R/W) */
spinlock_t wb_lock; /* lock for wb_keys */
spinlock_t lock; /* waitqueue/flags lock */
unsigned long flags;
-#define AFS_VNODE_CB_PROMISED 0 /* Set if vnode has a callback promise */
#define AFS_VNODE_UNSET 1 /* set if vnode attributes not yet set */
#define AFS_VNODE_DIR_VALID 2 /* Set if dir contents are valid */
#define AFS_VNODE_ZAP_DATA 3 /* set if vnode's data should be invalidated */
struct list_head cb_mmap_link; /* Link in cell->fs_open_mmaps */
void *cb_server; /* Server with callback/filelock */
atomic_t cb_nr_mmap; /* Number of mmaps */
- unsigned int cb_fs_s_break; /* Mass server break counter (cell->fs_s_break) */
- unsigned int cb_s_break; /* Mass break counter on ->server */
- unsigned int cb_v_break; /* Mass break counter on ->volume */
+ unsigned int cb_ro_snapshot; /* RO volume release counter on ->volume */
+ unsigned int cb_scrub; /* Scrub counter on ->volume */
unsigned int cb_break; /* Break counter on vnode */
+ unsigned int cb_v_check; /* Break check counter on ->volume */
seqlock_t cb_lock; /* Lock for ->cb_server, ->status, ->cb_*break */
- time64_t cb_expires_at; /* time at which callback expires */
+ atomic64_t cb_expires_at; /* time at which callback expires */
+#define AFS_NO_CB_PROMISE TIME64_MIN
};
static inline struct fscache_cookie *afs_vnode_cache(struct afs_vnode *vnode)
struct afs_fid fid; /* Fid to access */
struct afs_status_cb scb; /* Returned status and callback promise */
afs_dataversion_t dv_before; /* Data version before the call */
- unsigned int cb_break_before; /* cb_break + cb_s_break before the call */
+ unsigned int cb_break_before; /* cb_break before the call */
u8 dv_delta; /* Expected change in data version */
bool put_vnode:1; /* T if we have a ref on the vnode */
bool need_io_lock:1; /* T if we need the I/O lock on this */
unsigned int debug_id;
unsigned int cb_v_break; /* Volume break counter before op */
- unsigned int cb_s_break; /* Server break counter before op */
union {
struct {
static inline unsigned int afs_calc_vnode_cb_break(struct afs_vnode *vnode)
{
- return vnode->cb_break + vnode->cb_v_break;
+ return vnode->cb_break + vnode->cb_ro_snapshot + vnode->cb_scrub;
}
static inline bool afs_cb_is_broken(unsigned int cb_break,
const struct afs_vnode *vnode)
{
- return cb_break != (vnode->cb_break + atomic_read(&vnode->volume->cb_v_break));
+ return cb_break != (vnode->cb_break +
+ atomic_read(&vnode->volume->cb_ro_snapshot) +
+ atomic_read(&vnode->volume->cb_scrub));
}
/*
/*
* validation.c
*/
+bool afs_check_validity(const struct afs_vnode *vnode);
int afs_update_volume_state(struct afs_operation *op);
-bool afs_check_validity(struct afs_vnode *vnode);
-bool afs_pagecache_valid(struct afs_vnode *vnode);
int afs_validate(struct afs_vnode *vnode, struct key *key);
/*
refcount_read(&server->ref),
atomic_read(&server->active),
server->cell->name);
- seq_printf(m, " - info: fl=%lx rtt=%u brk=%x\n",
- server->flags, server->rtt, server->cb_s_break);
+ seq_printf(m, " - info: fl=%lx rtt=%u\n",
+ server->flags, server->rtt);
seq_printf(m, " - probe: last=%d\n",
(int)(jiffies - server->probed_at) / HZ);
failed = estate->failed_set;
write_seqlock(&vnode->cb_lock);
ASSERTCMP(cb_server, ==, vnode->cb_server);
vnode->cb_server = NULL;
- if (test_and_clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags))
+ if (atomic64_xchg(&vnode->cb_expires_at, AFS_NO_CB_PROMISE) != AFS_NO_CB_PROMISE)
vnode->cb_break++;
write_sequnlock(&vnode->cb_lock);
}
switch (op->call_error) {
case 0:
op->cumul_error.responded = true;
+
+ /* We succeeded, but we may need to redo the op from another
+ * server if we're looking at a set of RO volumes where some of
+ * the servers have not yet been brought up to date lest we
+ * regress the data. We only switch to the new version once
+ * >=50% of the servers are updated.
+ */
+ error = afs_update_volume_state(op);
+ if (error != 0) {
+ if (error == 1) {
+ afs_sleep_and_retry(op);
+ goto restart_from_beginning;
+ }
+ afs_op_set_error(op, error);
+ goto failed;
+ }
fallthrough;
default:
/* Success or local failure. Stop. */
op->server = server;
if (vnode->cb_server != server) {
vnode->cb_server = server;
- vnode->cb_s_break = server->cb_s_break;
- vnode->cb_fs_s_break = atomic_read(&server->cell->fs_s_break);
- vnode->cb_v_break = atomic_read(&vnode->volume->cb_v_break);
- clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags);
+ vnode->cb_v_check = atomic_read(&vnode->volume->cb_v_break);
+ atomic64_set(&vnode->cb_expires_at, AFS_NO_CB_PROMISE);
}
read_lock(&server->fs_lock);
slist->servers[j].server = server;
slist->servers[j].volume = volume;
slist->servers[j].flags = se_flags;
+ slist->servers[j].cb_expires_at = AFS_NO_CB_PROMISE;
slist->nr_servers++;
}
int diff;
if (pn && po && pn->server == po->server) {
+ pn->cb_expires_at = po->cb_expires_at;
list_replace(&po->slink, &pn->slink);
n++;
o++;
#include <linux/sched.h>
#include "internal.h"
+/*
+ * Data validation is managed through a number of mechanisms from the server:
+ *
+ * (1) On first contact with a server (such as if it has just been rebooted),
+ * the server sends us a CB.InitCallBackState* request.
+ *
+ * (2) On a RW volume, in response to certain vnode (inode)-accessing RPC
+ * calls, the server maintains a time-limited per-vnode promise that it
+ * will send us a CB.CallBack request if a third party alters the vnodes
+ * accessed.
+ *
+ * Note that a vnode-level callbacks may also be sent for other reasons,
+ * such as filelock release.
+ *
+ * (3) On a RO (or Backup) volume, in response to certain vnode-accessing RPC
+ * calls, each server maintains a time-limited per-volume promise that it
+ * will send us a CB.CallBack request if the RO volume is updated to a
+ * snapshot of the RW volume ("vos release"). This is an atomic event
+ * that cuts over all instances of the RO volume across multiple servers
+ * simultaneously.
+ *
+ * Note that a volume-level callbacks may also be sent for other reasons,
+ * such as the volumeserver taking over control of the volume from the
+ * fileserver.
+ *
+ * Note also that each server maintains an independent time limit on an
+ * independent callback.
+ *
+ * (4) Certain RPC calls include a volume information record "VolSync" in
+ * their reply. This contains a creation date for the volume that should
+ * remain unchanged for a RW volume (but will be changed if the volume is
+ * restored from backup) or will be bumped to the time of snapshotting
+ * when a RO volume is released.
+ *
+ * In order to track this events, the following are provided:
+ *
+ * ->cb_v_break. A counter of events that might mean that the contents of
+ * a volume have been altered since we last checked a vnode.
+ *
+ * ->cb_v_check. A counter of the number of events that we've sent a
+ * query to the server for. Everything's up to date if this equals
+ * cb_v_break.
+ *
+ * ->cb_scrub. A counter of the number of regression events for which we
+ * have to completely wipe the cache.
+ *
+ * ->cb_ro_snapshot. A counter of the number of times that we've
+ * recognised that a RO volume has been updated.
+ *
+ * ->cb_break. A counter of events that might mean that the contents of a
+ * vnode have been altered.
+ *
+ * ->cb_expires_at. The time at which the callback promise expires or
+ * AFS_NO_CB_PROMISE if we have no promise.
+ *
+ * The way we manage things is:
+ *
+ * (1) When a volume-level CB.CallBack occurs, we increment ->cb_v_break on
+ * the volume and reset ->cb_expires_at (ie. set AFS_NO_CB_PROMISE) on the
+ * volume and volume's server record.
+ *
+ * (2) When a CB.InitCallBackState occurs, we treat this as a volume-level
+ * callback break on all the volumes that have been using that volume
+ * (ie. increment ->cb_v_break and reset ->cb_expires_at).
+ *
+ * (3) When a vnode-level CB.CallBack occurs, we increment ->cb_break on the
+ * vnode and reset its ->cb_expires_at. If the vnode is mmapped, we also
+ * dispatch a work item to unmap all PTEs to the vnode's pagecache to
+ * force reentry to the filesystem for revalidation.
+ *
+ * (4) When entering the filesystem, we call afs_validate() to check the
+ * validity of a vnode. This first checks to see if ->cb_v_check and
+ * ->cb_v_break match, and if they don't, we lock volume->cb_check_lock
+ * exclusively and perform an FS.FetchStatus on the vnode.
+ *
+ * After checking the volume, we check the vnode. If there's a mismatch
+ * between the volume counters and the vnode's mirrors of those counters,
+ * we lock vnode->validate_lock and issue an FS.FetchStatus on the vnode.
+ *
+ * (5) When the reply from FS.FetchStatus arrives, the VolSync record is
+ * parsed:
+ *
+ * (A) If the Creation timestamp has changed on a RW volume or regressed
+ * on a RO volume, we try to increment ->cb_scrub; if it advances on a
+ * RO volume, we assume "vos release" happened and try to increment
+ * ->cb_ro_snapshot.
+ *
+ * (B) If the Update timestamp has regressed, we try to increment
+ * ->cb_scrub.
+ *
+ * Note that in both of these cases, we only do the increment if we can
+ * cmpxchg the value of the timestamp from the value we noted before the
+ * op. This tries to prevent parallel ops from fighting one another.
+ *
+ * volume->cb_v_check is then set to ->cb_v_break.
+ *
+ * (6) The AFSCallBack record included in the FS.FetchStatus reply is also
+ * parsed and used to set the promise in ->cb_expires_at for the vnode,
+ * the volume and the volume's server record.
+ *
+ * (7) If ->cb_scrub is seen to have advanced, we invalidate the pagecache for
+ * the vnode.
+ */
+
+/*
+ * Check the validity of a vnode/inode and its parent volume.
+ */
+bool afs_check_validity(const struct afs_vnode *vnode)
+{
+ const struct afs_volume *volume = vnode->volume;
+ time64_t deadline = ktime_get_real_seconds() + 10;
+
+ if (atomic_read(&volume->cb_v_check) != atomic_read(&volume->cb_v_break) ||
+ atomic64_read(&vnode->cb_expires_at) <= deadline ||
+ volume->cb_expires_at <= deadline ||
+ vnode->cb_ro_snapshot != atomic_read(&volume->cb_ro_snapshot) ||
+ vnode->cb_scrub != atomic_read(&volume->cb_scrub) ||
+ test_bit(AFS_VNODE_ZAP_DATA, &vnode->flags)) {
+ _debug("inval");
+ return false;
+ }
+
+ return true;
+}
+
/*
* See if the server we've just talked to is currently excluded.
*/
}
/*
- * Update the state of a volume. Returns 1 to redo the operation from the start.
+ * Update the state of a volume, including recording the expiration time of the
+ * callback promise. Returns 1 to redo the operation from the start.
*/
int afs_update_volume_state(struct afs_operation *op)
{
+ struct afs_server_list *slist = op->server_list;
+ struct afs_server_entry *se = &slist->servers[op->server_index];
+ struct afs_callback *cb = &op->file[0].scb.callback;
struct afs_volume *volume = op->volume;
+ unsigned int cb_v_break = atomic_read(&volume->cb_v_break);
+ unsigned int cb_v_check = atomic_read(&volume->cb_v_check);
int ret;
_enter("%llx", op->volume->vid);
}
}
+ if (op->cb_v_break == cb_v_break &&
+ (op->file[0].scb.have_cb || op->file[1].scb.have_cb)) {
+ time64_t expires_at = cb->expires_at;
+
+ if (!op->file[0].scb.have_cb)
+ expires_at = op->file[1].scb.callback.expires_at;
+
+ se->cb_expires_at = expires_at;
+ volume->cb_expires_at = expires_at;
+ }
+ if (cb_v_check < op->cb_v_break)
+ atomic_cmpxchg(&volume->cb_v_check, cb_v_check, op->cb_v_break);
return 0;
}
invalidate_inode_pages2(vnode->netfs.inode.i_mapping);
}
-/*
- * Check to see if we have a server currently serving this volume and that it
- * hasn't been reinitialised or dropped from the list.
- */
-static bool afs_check_server_good(struct afs_vnode *vnode)
-{
- struct afs_server_list *slist;
- struct afs_server *server;
- bool good;
- int i;
-
- if (vnode->cb_fs_s_break == atomic_read(&vnode->volume->cell->fs_s_break))
- return true;
-
- rcu_read_lock();
-
- slist = rcu_dereference(vnode->volume->servers);
- for (i = 0; i < slist->nr_servers; i++) {
- server = slist->servers[i].server;
- if (server == vnode->cb_server) {
- good = (vnode->cb_s_break == server->cb_s_break);
- rcu_read_unlock();
- return good;
- }
- }
-
- rcu_read_unlock();
- return false;
-}
-
-/*
- * Check the validity of a vnode/inode.
- */
-bool afs_check_validity(struct afs_vnode *vnode)
-{
- enum afs_cb_break_reason need_clear = afs_cb_break_no_break;
- time64_t now = ktime_get_real_seconds();
- unsigned int cb_break;
- int seq;
-
- do {
- seq = read_seqbegin(&vnode->cb_lock);
- cb_break = vnode->cb_break;
-
- if (test_bit(AFS_VNODE_CB_PROMISED, &vnode->flags)) {
- if (vnode->cb_v_break != atomic_read(&vnode->volume->cb_v_break))
- need_clear = afs_cb_break_for_v_break;
- else if (!afs_check_server_good(vnode))
- need_clear = afs_cb_break_for_s_reinit;
- else if (test_bit(AFS_VNODE_ZAP_DATA, &vnode->flags))
- need_clear = afs_cb_break_for_zap;
- else if (vnode->cb_expires_at - 10 <= now)
- need_clear = afs_cb_break_for_lapsed;
- } else if (test_bit(AFS_VNODE_DELETED, &vnode->flags)) {
- ;
- } else {
- need_clear = afs_cb_break_no_promise;
- }
-
- } while (read_seqretry(&vnode->cb_lock, seq));
-
- if (need_clear == afs_cb_break_no_break)
- return true;
-
- write_seqlock(&vnode->cb_lock);
- if (need_clear == afs_cb_break_no_promise)
- vnode->cb_v_break = atomic_read(&vnode->volume->cb_v_break);
- else if (cb_break == vnode->cb_break)
- __afs_break_callback(vnode, need_clear);
- else
- trace_afs_cb_miss(&vnode->fid, need_clear);
- write_sequnlock(&vnode->cb_lock);
- return false;
-}
-
-/*
- * Returns true if the pagecache is still valid. Does not sleep.
- */
-bool afs_pagecache_valid(struct afs_vnode *vnode)
-{
- if (unlikely(test_bit(AFS_VNODE_DELETED, &vnode->flags))) {
- if (vnode->netfs.inode.i_nlink)
- clear_nlink(&vnode->netfs.inode);
- return true;
- }
-
- if (test_bit(AFS_VNODE_CB_PROMISED, &vnode->flags) &&
- afs_check_validity(vnode))
- return true;
-
- return false;
-}
-
/*
* validate a vnode/inode
* - there are several things we need to check
*/
int afs_validate(struct afs_vnode *vnode, struct key *key)
{
+ struct afs_volume *volume = vnode->volume;
+ unsigned int cb_ro_snapshot, cb_scrub;
+ time64_t deadline = ktime_get_real_seconds() + 10;
+ bool zap = false, locked_vol = false;
int ret;
_enter("{v={%llx:%llu} fl=%lx},%x",
vnode->fid.vid, vnode->fid.vnode, vnode->flags,
key_serial(key));
- if (afs_pagecache_valid(vnode))
- goto valid;
+ if (afs_check_validity(vnode))
+ return 0;
- down_write(&vnode->validate_lock);
+ ret = down_write_killable(&vnode->validate_lock);
+ if (ret < 0)
+ goto error;
+
+ /* Validate a volume after the v_break has changed or the volume
+ * callback expired. We only want to do this once per volume per
+ * v_break change. The actual work will be done when parsing the
+ * status fetch reply.
+ */
+ if (volume->cb_expires_at <= deadline ||
+ atomic_read(&volume->cb_v_check) != atomic_read(&volume->cb_v_break)) {
+ ret = mutex_lock_interruptible(&volume->cb_check_lock);
+ if (ret < 0)
+ goto error_unlock;
+ locked_vol = true;
+ }
- /* if the promise has expired, we need to check the server again to get
- * a new promise - note that if the (parent) directory's metadata was
- * changed then the security may be different and we may no longer have
- * access */
- if (!test_bit(AFS_VNODE_CB_PROMISED, &vnode->flags)) {
- _debug("not promised");
+ cb_ro_snapshot = atomic_read(&volume->cb_ro_snapshot);
+ cb_scrub = atomic_read(&volume->cb_scrub);
+ if (vnode->cb_ro_snapshot != cb_ro_snapshot ||
+ vnode->cb_scrub != cb_scrub)
+ unmap_mapping_pages(vnode->netfs.inode.i_mapping, 0, 0, false);
+
+ if (vnode->cb_ro_snapshot != cb_ro_snapshot ||
+ vnode->cb_scrub != cb_scrub ||
+ volume->cb_expires_at <= deadline ||
+ atomic_read(&volume->cb_v_check) != atomic_read(&volume->cb_v_break) ||
+ atomic64_read(&vnode->cb_expires_at) <= deadline
+ ) {
ret = afs_fetch_status(vnode, key, false, NULL);
if (ret < 0) {
if (ret == -ENOENT) {
}
goto error_unlock;
}
+
_debug("new promise [fl=%lx]", vnode->flags);
}
+ /* We can drop the volume lock now as. */
+ if (locked_vol) {
+ mutex_unlock(&volume->cb_check_lock);
+ locked_vol = false;
+ }
+
+ cb_ro_snapshot = atomic_read(&volume->cb_ro_snapshot);
+ cb_scrub = atomic_read(&volume->cb_scrub);
+ _debug("vnode inval %x==%x %x==%x",
+ vnode->cb_ro_snapshot, cb_ro_snapshot,
+ vnode->cb_scrub, cb_scrub);
+ if (vnode->cb_scrub != cb_scrub)
+ zap = true;
+ vnode->cb_ro_snapshot = cb_ro_snapshot;
+ vnode->cb_scrub = cb_scrub;
+
if (test_bit(AFS_VNODE_DELETED, &vnode->flags)) {
_debug("file already deleted");
ret = -ESTALE;
/* if the vnode's data version number changed then its contents are
* different */
- if (test_and_clear_bit(AFS_VNODE_ZAP_DATA, &vnode->flags))
+ zap |= test_and_clear_bit(AFS_VNODE_ZAP_DATA, &vnode->flags);
+ if (zap)
afs_zap_data(vnode);
up_write(&vnode->validate_lock);
-valid:
_leave(" = 0");
return 0;
error_unlock:
+ if (locked_vol)
+ mutex_unlock(&volume->cb_check_lock);
up_write(&vnode->validate_lock);
+error:
_leave(" = %d", ret);
return ret;
}
INIT_WORK(&volume->destructor, afs_destroy_volume);
rwlock_init(&volume->servers_lock);
mutex_init(&volume->volsync_lock);
+ mutex_init(&volume->cb_check_lock);
rwlock_init(&volume->cb_v_break_lock);
+ INIT_LIST_HEAD(&volume->open_mmaps);
+ init_rwsem(&volume->open_mmaps_lock);
memcpy(volume->name, vldb->name, vldb->name_len + 1);
for (i = 0; i < AFS_MAXTYPES; i++)
#define afs_cb_break_reasons \
EM(afs_cb_break_no_break, "no-break") \
- EM(afs_cb_break_no_promise, "no-promise") \
EM(afs_cb_break_for_callback, "break-cb") \
EM(afs_cb_break_for_creation_regress, "creation-regress") \
EM(afs_cb_break_for_deleted, "break-del") \
- EM(afs_cb_break_for_lapsed, "break-lapsed") \
EM(afs_cb_break_for_s_reinit, "s-reinit") \
EM(afs_cb_break_for_unlink, "break-unlink") \
EM(afs_cb_break_for_update_regress, "update-regress") \
- EM(afs_cb_break_for_v_break, "break-v") \
EM(afs_cb_break_for_volume_callback, "break-v-cb") \
EM(afs_cb_break_for_vos_release, "break-vos-release") \
- EM(afs_cb_break_for_zap, "break-zap") \
E_(afs_cb_break_volume_excluded, "vol-excluded")
/*
projects / linux-2.6-block.git / commitdiff