summaryrefslogtreecommitdiff
path: root/fs/nfs/nfs4state.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/nfs/nfs4state.c')
-rw-r--r--fs/nfs/nfs4state.c450
1 files changed, 307 insertions, 143 deletions
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 02488b50534a..01179f7de322 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -49,6 +49,7 @@
#include <linux/workqueue.h>
#include <linux/bitops.h>
#include <linux/jiffies.h>
+#include <linux/sched/mm.h>
#include <linux/sunrpc/clnt.h>
@@ -60,11 +61,14 @@
#include "nfs4session.h"
#include "pnfs.h"
#include "netns.h"
+#include "nfs4trace.h"
#define NFSDBG_FACILITY NFSDBG_STATE
#define OPENOWNER_POOL_SIZE 8
+static void nfs4_state_start_reclaim_reboot(struct nfs_client *clp);
+
const nfs4_stateid zero_stateid = {
{ .data = { 0 } },
.type = NFS4_SPECIAL_STATEID_TYPE,
@@ -87,6 +91,25 @@ const nfs4_stateid current_stateid = {
static DEFINE_MUTEX(nfs_clid_init_mutex);
+static int nfs4_setup_state_renewal(struct nfs_client *clp)
+{
+ int status;
+ struct nfs_fsinfo fsinfo;
+
+ if (!test_bit(NFS_CS_CHECK_LEASE_TIME, &clp->cl_res_state)) {
+ nfs4_schedule_state_renewal(clp);
+ return 0;
+ }
+
+ status = nfs4_proc_get_lease_time(clp, &fsinfo);
+ if (status == 0) {
+ nfs4_set_lease_period(clp, fsinfo.lease_time * HZ);
+ nfs4_schedule_state_renewal(clp);
+ }
+
+ return status;
+}
+
int nfs4_init_clientid(struct nfs_client *clp, const struct cred *cred)
{
struct nfs4_setclientid_res clid = {
@@ -114,7 +137,7 @@ do_confirm:
if (status != 0)
goto out;
clear_bit(NFS4CLNT_LEASE_CONFIRM, &clp->cl_state);
- nfs4_schedule_state_renewal(clp);
+ nfs4_setup_state_renewal(clp);
out:
return status;
}
@@ -159,6 +182,10 @@ int nfs40_discover_server_trunking(struct nfs_client *clp,
/* Sustain the lease, even if it's empty. If the clientid4
* goes stale it's of no use for trunking discovery. */
nfs4_schedule_state_renewal(*result);
+
+ /* If the client state need to recover, do it. */
+ if (clp->cl_state)
+ nfs4_schedule_state_manager(clp);
}
out:
return status;
@@ -282,34 +309,13 @@ static int nfs4_begin_drain_session(struct nfs_client *clp)
#if defined(CONFIG_NFS_V4_1)
-static int nfs41_setup_state_renewal(struct nfs_client *clp)
-{
- int status;
- struct nfs_fsinfo fsinfo;
- unsigned long now;
-
- if (!test_bit(NFS_CS_CHECK_LEASE_TIME, &clp->cl_res_state)) {
- nfs4_schedule_state_renewal(clp);
- return 0;
- }
-
- now = jiffies;
- status = nfs4_proc_get_lease_time(clp, &fsinfo);
- if (status == 0) {
- nfs4_set_lease_period(clp, fsinfo.lease_time * HZ, now);
- nfs4_schedule_state_renewal(clp);
- }
-
- return status;
-}
-
static void nfs41_finish_session_reset(struct nfs_client *clp)
{
clear_bit(NFS4CLNT_LEASE_CONFIRM, &clp->cl_state);
clear_bit(NFS4CLNT_SESSION_RESET, &clp->cl_state);
/* create_session negotiated new slot table */
clear_bit(NFS4CLNT_BIND_CONN_TO_SESSION, &clp->cl_state);
- nfs41_setup_state_renewal(clp);
+ nfs4_setup_state_renewal(clp);
}
int nfs41_init_clientid(struct nfs_client *clp, const struct cred *cred)
@@ -326,6 +332,8 @@ do_confirm:
status = nfs4_proc_create_session(clp, cred);
if (status != 0)
goto out;
+ if (!(clp->cl_exchange_flags & EXCHGID4_FLAG_CONFIRMED_R))
+ nfs4_state_start_reclaim_reboot(clp);
nfs41_finish_session_reset(clp);
nfs_mark_client_ready(clp, NFS_CS_READY);
out:
@@ -493,12 +501,7 @@ nfs4_alloc_state_owner(struct nfs_server *server,
sp = kzalloc(sizeof(*sp), gfp_flags);
if (!sp)
return NULL;
- sp->so_seqid.owner_id = ida_simple_get(&server->openowner_id, 0, 0,
- gfp_flags);
- if (sp->so_seqid.owner_id < 0) {
- kfree(sp);
- return NULL;
- }
+ sp->so_seqid.owner_id = atomic64_inc_return(&server->owner_ctr);
sp->so_server = server;
sp->so_cred = get_cred(cred);
spin_lock_init(&sp->so_lock);
@@ -506,7 +509,6 @@ nfs4_alloc_state_owner(struct nfs_server *server,
nfs4_init_seqid_counter(&sp->so_seqid);
atomic_set(&sp->so_count, 1);
INIT_LIST_HEAD(&sp->so_lru);
- seqcount_init(&sp->so_reclaim_seqcount);
mutex_init(&sp->so_delegreturn_mutex);
return sp;
}
@@ -530,7 +532,6 @@ static void nfs4_free_state_owner(struct nfs4_state_owner *sp)
{
nfs4_destroy_seqid_counter(&sp->so_seqid);
put_cred(sp->so_cred);
- ida_simple_remove(&sp->so_server->openowner_id, sp->so_seqid.owner_id);
kfree(sp);
}
@@ -563,6 +564,7 @@ static void nfs4_gc_state_owners(struct nfs_server *server)
* nfs4_get_state_owner - Look up a state owner given a credential
* @server: nfs_server to search
* @cred: RPC credential to match
+ * @gfp_flags: allocation mode
*
* Returns a pointer to an instantiated nfs4_state_owner struct, or NULL.
*/
@@ -619,24 +621,39 @@ void nfs4_put_state_owner(struct nfs4_state_owner *sp)
/**
* nfs4_purge_state_owners - Release all cached state owners
* @server: nfs_server with cached state owners to release
+ * @head: resulting list of state owners
*
* Called at umount time. Remaining state owners will be on
* the LRU with ref count of zero.
+ * Note that the state owners are not freed, but are added
+ * to the list @head, which can later be used as an argument
+ * to nfs4_free_state_owners.
*/
-void nfs4_purge_state_owners(struct nfs_server *server)
+void nfs4_purge_state_owners(struct nfs_server *server, struct list_head *head)
{
struct nfs_client *clp = server->nfs_client;
struct nfs4_state_owner *sp, *tmp;
- LIST_HEAD(doomed);
spin_lock(&clp->cl_lock);
list_for_each_entry_safe(sp, tmp, &server->state_owners_lru, so_lru) {
- list_move(&sp->so_lru, &doomed);
+ list_move(&sp->so_lru, head);
nfs4_remove_state_owner_locked(sp);
}
spin_unlock(&clp->cl_lock);
+}
- list_for_each_entry_safe(sp, tmp, &doomed, so_lru) {
+/**
+ * nfs4_free_state_owners - Release all cached state owners
+ * @head: resulting list of state owners
+ *
+ * Frees a list of state owners that was generated by
+ * nfs4_purge_state_owners
+ */
+void nfs4_free_state_owners(struct list_head *head)
+{
+ struct nfs4_state_owner *sp, *tmp;
+
+ list_for_each_entry_safe(sp, tmp, head, so_lru) {
list_del(&sp->so_lru);
nfs4_free_state_owner(sp);
}
@@ -647,7 +664,7 @@ nfs4_alloc_open_state(void)
{
struct nfs4_state *state;
- state = kzalloc(sizeof(*state), GFP_NOFS);
+ state = kzalloc(sizeof(*state), GFP_KERNEL_ACCOUNT);
if (!state)
return NULL;
refcount_set(&state->count, 1);
@@ -715,9 +732,9 @@ nfs4_get_open_state(struct inode *inode, struct nfs4_state_owner *owner)
state = new;
state->owner = owner;
atomic_inc(&owner->so_count);
- list_add_rcu(&state->inode_states, &nfsi->open_states);
ihold(inode);
state->inode = inode;
+ list_add_rcu(&state->inode_states, &nfsi->open_states);
spin_unlock(&inode->i_lock);
/* Note: The reclaim code dictates that we add stateless
* and read-only stateids to the end of the list */
@@ -745,6 +762,7 @@ void nfs4_put_open_state(struct nfs4_state *state)
list_del(&state->open_states);
spin_unlock(&inode->i_lock);
spin_unlock(&owner->so_lock);
+ nfs4_inode_return_delegation_on_close(inode);
iput(inode);
nfs4_free_open_state(state);
nfs4_put_state_owner(owner);
@@ -800,7 +818,7 @@ static void __nfs4_close(struct nfs4_state *state,
void nfs4_close_state(struct nfs4_state *state, fmode_t fmode)
{
- __nfs4_close(state, fmode, GFP_NOFS, 0);
+ __nfs4_close(state, fmode, GFP_KERNEL, 0);
}
void nfs4_close_sync(struct nfs4_state *state, fmode_t fmode)
@@ -823,15 +841,15 @@ void nfs4_close_sync(struct nfs4_state *state, fmode_t fmode)
*/
static struct nfs4_lock_state *
__nfs4_find_lock_state(struct nfs4_state *state,
- fl_owner_t fl_owner, fl_owner_t fl_owner2)
+ fl_owner_t owner, fl_owner_t owner2)
{
struct nfs4_lock_state *pos, *ret = NULL;
list_for_each_entry(pos, &state->lock_states, ls_locks) {
- if (pos->ls_owner == fl_owner) {
+ if (pos->ls_owner == owner) {
ret = pos;
break;
}
- if (pos->ls_owner == fl_owner2)
+ if (pos->ls_owner == owner2)
ret = pos;
}
if (ret)
@@ -844,31 +862,25 @@ __nfs4_find_lock_state(struct nfs4_state *state,
* exists, return an uninitialized one.
*
*/
-static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, fl_owner_t fl_owner)
+static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, fl_owner_t owner)
{
struct nfs4_lock_state *lsp;
struct nfs_server *server = state->owner->so_server;
- lsp = kzalloc(sizeof(*lsp), GFP_NOFS);
+ lsp = kzalloc(sizeof(*lsp), GFP_KERNEL_ACCOUNT);
if (lsp == NULL)
return NULL;
nfs4_init_seqid_counter(&lsp->ls_seqid);
refcount_set(&lsp->ls_count, 1);
lsp->ls_state = state;
- lsp->ls_owner = fl_owner;
- lsp->ls_seqid.owner_id = ida_simple_get(&server->lockowner_id, 0, 0, GFP_NOFS);
- if (lsp->ls_seqid.owner_id < 0)
- goto out_free;
+ lsp->ls_owner = owner;
+ lsp->ls_seqid.owner_id = atomic64_inc_return(&server->owner_ctr);
INIT_LIST_HEAD(&lsp->ls_locks);
return lsp;
-out_free:
- kfree(lsp);
- return NULL;
}
void nfs4_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp)
{
- ida_simple_remove(&server->lockowner_id, lsp->ls_seqid.owner_id);
nfs4_destroy_seqid_counter(&lsp->ls_seqid);
kfree(lsp);
}
@@ -956,7 +968,7 @@ int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl)
if (fl->fl_ops != NULL)
return 0;
- lsp = nfs4_get_lock_state(state, fl->fl_owner);
+ lsp = nfs4_get_lock_state(state, fl->c.flc_owner);
if (lsp == NULL)
return -ENOMEM;
fl->fl_u.nfs4_fl.owner = lsp;
@@ -969,7 +981,7 @@ static int nfs4_copy_lock_stateid(nfs4_stateid *dst,
const struct nfs_lock_context *l_ctx)
{
struct nfs4_lock_state *lsp;
- fl_owner_t fl_owner, fl_flock_owner;
+ fl_owner_t owner, fl_flock_owner;
int ret = -ENOENT;
if (l_ctx == NULL)
@@ -978,11 +990,11 @@ static int nfs4_copy_lock_stateid(nfs4_stateid *dst,
if (test_bit(LK_STATE_IN_USE, &state->flags) == 0)
goto out;
- fl_owner = l_ctx->lockowner;
+ owner = l_ctx->lockowner;
fl_flock_owner = l_ctx->open_context->flock_owner;
spin_lock(&state->state_lock);
- lsp = __nfs4_find_lock_state(state, fl_owner, fl_flock_owner);
+ lsp = __nfs4_find_lock_state(state, owner, fl_flock_owner);
if (lsp && test_bit(NFS_LOCK_LOST, &lsp->ls_flags))
ret = -EIO;
else if (lsp != NULL && test_bit(NFS_LOCK_INITIALIZED, &lsp->ls_flags) != 0) {
@@ -995,22 +1007,6 @@ out:
return ret;
}
-bool nfs4_refresh_open_stateid(nfs4_stateid *dst, struct nfs4_state *state)
-{
- bool ret;
- int seq;
-
- do {
- ret = false;
- seq = read_seqbegin(&state->seqlock);
- if (nfs4_state_match_open_stateid_other(state, dst)) {
- dst->seqid = state->open_stateid.seqid;
- ret = true;
- }
- } while (read_seqretry(&state->seqlock, seq));
- return ret;
-}
-
bool nfs4_copy_open_stateid(nfs4_stateid *dst, struct nfs4_state *state)
{
bool ret;
@@ -1059,8 +1055,7 @@ int nfs4_select_rw_stateid(struct nfs4_state *state,
* choose to use.
*/
goto out;
- nfs4_copy_open_stateid(dst, state);
- ret = 0;
+ ret = nfs4_copy_open_stateid(dst, state) ? 0 : -EAGAIN;
out:
if (nfs_server_capable(state->inode, NFS_CAP_STATEID_NFSV41))
dst->seqid = 0;
@@ -1088,14 +1083,12 @@ void nfs_release_seqid(struct nfs_seqid *seqid)
return;
sequence = seqid->sequence;
spin_lock(&sequence->lock);
- list_del_init(&seqid->list);
- if (!list_empty(&sequence->list)) {
- struct nfs_seqid *next;
-
- next = list_first_entry(&sequence->list,
- struct nfs_seqid, list);
+ if (list_is_first(&seqid->list, &sequence->list) &&
+ !list_is_singular(&sequence->list)) {
+ struct nfs_seqid *next = list_next_entry(seqid, list);
rpc_wake_up_queued_task(&sequence->wait, next->task);
}
+ list_del_init(&seqid->list);
spin_unlock(&sequence->lock);
}
@@ -1122,6 +1115,7 @@ static void nfs_increment_seqid(int status, struct nfs_seqid *seqid)
" sequence-id error on an"
" unconfirmed sequence %p!\n",
seqid->sequence);
+ return;
case -NFS4ERR_STALE_CLIENTID:
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_BAD_STATEID:
@@ -1131,7 +1125,7 @@ static void nfs_increment_seqid(int status, struct nfs_seqid *seqid)
case -NFS4ERR_MOVED:
/* Non-seqid mutating errors */
return;
- };
+ }
/*
* Note: no locking needed as we are guaranteed to be first
* on the sequence list
@@ -1190,10 +1184,7 @@ static int nfs4_run_state_manager(void *);
static void nfs4_clear_state_manager_bit(struct nfs_client *clp)
{
- smp_mb__before_atomic();
- clear_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state);
- smp_mb__after_atomic();
- wake_up_bit(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING);
+ clear_and_wake_up_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state);
rpc_wake_up(&clp->cl_rpcwaitq);
}
@@ -1204,10 +1195,26 @@ void nfs4_schedule_state_manager(struct nfs_client *clp)
{
struct task_struct *task;
char buf[INET6_ADDRSTRLEN + sizeof("-manager") + 1];
+ struct rpc_clnt *clnt = clp->cl_rpcclient;
+ bool swapon = false;
+
+ if (clp->cl_cons_state < 0)
+ return;
set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
+
+ if (atomic_read(&clnt->cl_swapper)) {
+ swapon = !test_and_set_bit(NFS4CLNT_MANAGER_AVAILABLE,
+ &clp->cl_state);
+ if (!swapon) {
+ wake_up_var(&clp->cl_state);
+ return;
+ }
+ }
+
if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
return;
+
__module_get(THIS_MODULE);
refcount_inc(&clp->cl_count);
@@ -1222,6 +1229,10 @@ void nfs4_schedule_state_manager(struct nfs_client *clp)
if (IS_ERR(task)) {
printk(KERN_ERR "%s: kthread_run: %ld\n",
__func__, PTR_ERR(task));
+ if (!nfs_client_init_is_complete(clp))
+ nfs_mark_client_ready(clp, PTR_ERR(task));
+ if (swapon)
+ clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state);
nfs4_clear_state_manager_bit(clp);
nfs_put_client(clp);
module_put(THIS_MODULE);
@@ -1303,7 +1314,8 @@ int nfs4_wait_clnt_recover(struct nfs_client *clp)
refcount_inc(&clp->cl_count);
res = wait_on_bit_action(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING,
- nfs_wait_bit_killable, TASK_KILLABLE);
+ nfs_wait_bit_killable,
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
if (res)
goto out;
if (clp->cl_cons_state < 0)
@@ -1391,7 +1403,7 @@ int nfs4_schedule_stateid_recovery(const struct nfs_server *server, struct nfs4_
dprintk("%s: scheduling stateid recovery for server %s\n", __func__,
clp->cl_hostname);
nfs4_schedule_state_manager(clp);
- return 0;
+ return clp->cl_cons_state < 0 ? clp->cl_cons_state : 0;
}
EXPORT_SYMBOL_GPL(nfs4_schedule_stateid_recovery);
@@ -1404,7 +1416,7 @@ nfs_state_find_lock_state_by_stateid(struct nfs4_state *state,
list_for_each_entry(pos, &state->lock_states, ls_locks) {
if (!test_bit(NFS_LOCK_INITIALIZED, &pos->ls_flags))
continue;
- if (nfs4_stateid_match_other(&pos->ls_stateid, stateid))
+ if (nfs4_stateid_match_or_older(&pos->ls_stateid, stateid))
return pos;
}
return NULL;
@@ -1438,12 +1450,13 @@ void nfs_inode_find_state_and_recover(struct inode *inode,
state = ctx->state;
if (state == NULL)
continue;
- if (nfs4_stateid_match_other(&state->stateid, stateid) &&
+ if (nfs4_stateid_match_or_older(&state->stateid, stateid) &&
nfs4_state_mark_reclaim_nograce(clp, state)) {
found = true;
continue;
}
- if (nfs4_stateid_match_other(&state->open_stateid, stateid) &&
+ if (test_bit(NFS_OPEN_STATE, &state->flags) &&
+ nfs4_stateid_match_or_older(&state->open_stateid, stateid) &&
nfs4_state_mark_reclaim_nograce(clp, state)) {
found = true;
continue;
@@ -1459,7 +1472,7 @@ void nfs_inode_find_state_and_recover(struct inode *inode,
nfs4_schedule_state_manager(clp);
}
-static void nfs4_state_mark_open_context_bad(struct nfs4_state *state)
+static void nfs4_state_mark_open_context_bad(struct nfs4_state *state, int err)
{
struct inode *inode = state->inode;
struct nfs_inode *nfsi = NFS_I(inode);
@@ -1470,6 +1483,8 @@ static void nfs4_state_mark_open_context_bad(struct nfs4_state *state)
if (ctx->state != state)
continue;
set_bit(NFS_CONTEXT_BAD, &ctx->flags);
+ pr_warn("NFSv4: state recovery failed for open file %pd2, "
+ "error = %d\n", ctx->dentry, err);
}
rcu_read_unlock();
}
@@ -1477,7 +1492,7 @@ static void nfs4_state_mark_open_context_bad(struct nfs4_state *state)
static void nfs4_state_mark_recovery_failed(struct nfs4_state *state, int error)
{
set_bit(NFS_STATE_RECOVERY_FAILED, &state->flags);
- nfs4_state_mark_open_context_bad(state);
+ nfs4_state_mark_open_context_bad(state, error);
}
@@ -1488,7 +1503,7 @@ static int nfs4_reclaim_locks(struct nfs4_state *state, const struct nfs4_state_
struct file_lock *fl;
struct nfs4_lock_state *lsp;
int status = 0;
- struct file_lock_context *flctx = inode->i_flctx;
+ struct file_lock_context *flctx = locks_inode_context(inode);
struct list_head *list;
if (flctx == NULL)
@@ -1500,14 +1515,15 @@ static int nfs4_reclaim_locks(struct nfs4_state *state, const struct nfs4_state_
down_write(&nfsi->rwsem);
spin_lock(&flctx->flc_lock);
restart:
- list_for_each_entry(fl, list, fl_list) {
- if (nfs_file_open_context(fl->fl_file)->state != state)
+ for_each_file_lock(fl, list) {
+ if (nfs_file_open_context(fl->c.flc_file)->state != state)
continue;
spin_unlock(&flctx->flc_lock);
status = ops->recover_lock(state, fl);
switch (status) {
case 0:
break;
+ case -ETIMEDOUT:
case -ESTALE:
case -NFS4ERR_ADMIN_REVOKED:
case -NFS4ERR_STALE_STATEID:
@@ -1523,7 +1539,7 @@ restart:
default:
pr_err("NFS: %s: unhandled error %d\n",
__func__, status);
- /* Fall through */
+ fallthrough;
case -ENOMEM:
case -NFS4ERR_DENIED:
case -NFS4ERR_RECLAIM_BAD:
@@ -1550,16 +1566,32 @@ static void nfs42_complete_copies(struct nfs4_state_owner *sp, struct nfs4_state
{
struct nfs4_copy_state *copy;
- if (!test_bit(NFS_CLNT_DST_SSC_COPY_STATE, &state->flags))
+ if (!test_bit(NFS_CLNT_DST_SSC_COPY_STATE, &state->flags) &&
+ !test_bit(NFS_CLNT_SRC_SSC_COPY_STATE, &state->flags))
return;
spin_lock(&sp->so_server->nfs_client->cl_lock);
list_for_each_entry(copy, &sp->so_server->ss_copies, copies) {
- if (!nfs4_stateid_match_other(&state->stateid, &copy->parent_state->stateid))
- continue;
+ if ((test_bit(NFS_CLNT_DST_SSC_COPY_STATE, &state->flags) &&
+ !nfs4_stateid_match_other(&state->stateid,
+ &copy->parent_dst_state->stateid)))
+ continue;
copy->flags = 1;
- complete(&copy->completion);
- break;
+ if (test_and_clear_bit(NFS_CLNT_DST_SSC_COPY_STATE,
+ &state->flags)) {
+ clear_bit(NFS_CLNT_SRC_SSC_COPY_STATE, &state->flags);
+ complete(&copy->completion);
+ }
+ }
+ list_for_each_entry(copy, &sp->so_server->ss_src_copies, src_copies) {
+ if ((test_bit(NFS_CLNT_SRC_SSC_COPY_STATE, &state->flags) &&
+ !nfs4_stateid_match_other(&state->stateid,
+ &copy->parent_src_state->stateid)))
+ continue;
+ copy->flags = 1;
+ if (test_and_clear_bit(NFS_CLNT_DST_SSC_COPY_STATE,
+ &state->flags))
+ complete(&copy->completion);
}
spin_unlock(&sp->so_server->nfs_client->cl_lock);
}
@@ -1571,7 +1603,8 @@ static inline void nfs42_complete_copies(struct nfs4_state_owner *sp,
#endif /* CONFIG_NFS_V4_2 */
static int __nfs4_reclaim_open_state(struct nfs4_state_owner *sp, struct nfs4_state *state,
- const struct nfs4_state_recovery_ops *ops)
+ const struct nfs4_state_recovery_ops *ops,
+ int *lost_locks)
{
struct nfs4_lock_state *lock;
int status;
@@ -1587,8 +1620,10 @@ static int __nfs4_reclaim_open_state(struct nfs4_state_owner *sp, struct nfs4_st
if (!test_bit(NFS_DELEGATED_STATE, &state->flags)) {
spin_lock(&state->state_lock);
list_for_each_entry(lock, &state->lock_states, ls_locks) {
- if (!test_bit(NFS_LOCK_INITIALIZED, &lock->ls_flags))
- pr_warn_ratelimited("NFS: %s: Lock reclaim failed!\n", __func__);
+ trace_nfs4_state_lock_reclaim(state, lock);
+ if (!test_bit(NFS_LOCK_INITIALIZED, &lock->ls_flags) &&
+ !test_bit(NFS_LOCK_UNLOCKING, &lock->ls_flags))
+ *lost_locks += 1;
}
spin_unlock(&state->state_lock);
}
@@ -1598,10 +1633,16 @@ static int __nfs4_reclaim_open_state(struct nfs4_state_owner *sp, struct nfs4_st
return status;
}
-static int nfs4_reclaim_open_state(struct nfs4_state_owner *sp, const struct nfs4_state_recovery_ops *ops)
+static int nfs4_reclaim_open_state(struct nfs4_state_owner *sp,
+ const struct nfs4_state_recovery_ops *ops,
+ int *lost_locks)
{
struct nfs4_state *state;
+ unsigned int loop = 0;
int status = 0;
+#ifdef CONFIG_NFS_V4_2
+ bool found_ssc_copy_state = false;
+#endif /* CONFIG_NFS_V4_2 */
/* Note: we rely on the sp->so_states list being ordered
* so that we always reclaim open(O_RDWR) and/or open(O_WRITE)
@@ -1612,7 +1653,6 @@ static int nfs4_reclaim_open_state(struct nfs4_state_owner *sp, const struct nfs
* server that doesn't support a grace period.
*/
spin_lock(&sp->so_lock);
- raw_write_seqcount_begin(&sp->so_reclaim_seqcount);
restart:
list_for_each_entry(state, &sp->so_states, open_states) {
if (!test_and_clear_bit(ops->state_flag_bit, &state->flags))
@@ -1621,16 +1661,25 @@ restart:
continue;
if (state->state == 0)
continue;
+#ifdef CONFIG_NFS_V4_2
+ if (test_bit(NFS_SRV_SSC_COPY_STATE, &state->flags)) {
+ nfs4_state_mark_recovery_failed(state, -EIO);
+ found_ssc_copy_state = true;
+ continue;
+ }
+#endif /* CONFIG_NFS_V4_2 */
refcount_inc(&state->count);
spin_unlock(&sp->so_lock);
- status = __nfs4_reclaim_open_state(sp, state, ops);
+ status = __nfs4_reclaim_open_state(sp, state, ops, lost_locks);
switch (status) {
default:
- if (status >= 0)
+ if (status >= 0) {
+ loop = 0;
break;
+ }
printk(KERN_ERR "NFS: %s: unhandled error %d\n", __func__, status);
- /* Fall through */
+ fallthrough;
case -ENOENT:
case -ENOMEM:
case -EACCES:
@@ -1642,7 +1691,11 @@ restart:
break;
case -EAGAIN:
ssleep(1);
- /* Fall through */
+ if (loop++ < 10) {
+ set_bit(ops->state_flag_bit, &state->flags);
+ break;
+ }
+ fallthrough;
case -NFS4ERR_ADMIN_REVOKED:
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_OLD_STATEID:
@@ -1654,24 +1707,28 @@ restart:
case -NFS4ERR_EXPIRED:
case -NFS4ERR_NO_GRACE:
nfs4_state_mark_reclaim_nograce(sp->so_server->nfs_client, state);
+ fallthrough;
case -NFS4ERR_STALE_CLIENTID:
case -NFS4ERR_BADSESSION:
case -NFS4ERR_BADSLOT:
case -NFS4ERR_BAD_HIGH_SLOT:
case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
+ case -ETIMEDOUT:
goto out_err;
}
nfs4_put_open_state(state);
spin_lock(&sp->so_lock);
goto restart;
}
- raw_write_seqcount_end(&sp->so_reclaim_seqcount);
spin_unlock(&sp->so_lock);
+#ifdef CONFIG_NFS_V4_2
+ if (found_ssc_copy_state)
+ return -EIO;
+#endif /* CONFIG_NFS_V4_2 */
return 0;
out_err:
nfs4_put_open_state(state);
spin_lock(&sp->so_lock);
- raw_write_seqcount_end(&sp->so_reclaim_seqcount);
spin_unlock(&sp->so_lock);
return status;
}
@@ -1729,6 +1786,7 @@ static void nfs4_state_mark_reclaim_helper(struct nfs_client *clp,
static void nfs4_state_start_reclaim_reboot(struct nfs_client *clp)
{
+ set_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state);
/* Mark all delegations for reclaim */
nfs_delegation_mark_reclaim(clp);
nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_reboot);
@@ -1792,6 +1850,7 @@ static void nfs4_state_end_reclaim_reboot(struct nfs_client *clp)
if (!nfs4_state_clear_reclaim_reboot(clp))
return;
+ pnfs_destroy_all_layouts(clp);
ops = clp->cl_mvops->reboot_recovery_ops;
cred = nfs4_get_clid_cred(clp);
err = nfs4_reclaim_complete(clp, ops, cred);
@@ -1852,12 +1911,17 @@ static int nfs4_do_reclaim(struct nfs_client *clp, const struct nfs4_state_recov
struct nfs4_state_owner *sp;
struct nfs_server *server;
struct rb_node *pos;
- int status = 0;
+ LIST_HEAD(freeme);
+ int lost_locks = 0;
+ int status;
+ status = nfs4_begin_drain_session(clp);
+ if (status < 0)
+ return status;
restart:
rcu_read_lock();
list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) {
- nfs4_purge_state_owners(server);
+ nfs4_purge_state_owners(server, &freeme);
spin_lock(&clp->cl_lock);
for (pos = rb_first(&server->state_owners);
pos != NULL;
@@ -1872,11 +1936,15 @@ restart:
spin_unlock(&clp->cl_lock);
rcu_read_unlock();
- status = nfs4_reclaim_open_state(sp, ops);
+ status = nfs4_reclaim_open_state(sp, ops, &lost_locks);
if (status < 0) {
+ if (lost_locks)
+ pr_warn("NFS: %s: lost %d locks\n",
+ clp->cl_hostname, lost_locks);
set_bit(ops->owner_flag_bit, &sp->so_flags);
nfs4_put_state_owner(sp);
status = nfs4_recovery_handle_error(clp, status);
+ nfs4_free_state_owners(&freeme);
return (status != 0) ? status : -EAGAIN;
}
@@ -1886,6 +1954,11 @@ restart:
spin_unlock(&clp->cl_lock);
}
rcu_read_unlock();
+ nfs4_free_state_owners(&freeme);
+ nfs_local_probe_async(clp);
+ if (lost_locks)
+ pr_warn("NFS: %s: lost %d locks\n",
+ clp->cl_hostname, lost_locks);
return 0;
}
@@ -1939,9 +2012,14 @@ static int nfs4_handle_reclaim_lease_error(struct nfs_client *clp, int status)
nfs_mark_client_ready(clp, -EPERM);
clear_bit(NFS4CLNT_LEASE_CONFIRM, &clp->cl_state);
return -EPERM;
+ case -ETIMEDOUT:
+ if (clp->cl_cons_state == NFS_CS_SESSION_INITING) {
+ nfs_mark_client_ready(clp, -EIO);
+ return -EIO;
+ }
+ fallthrough;
case -EACCES:
case -NFS4ERR_DELAY:
- case -ETIMEDOUT:
case -EAGAIN:
ssleep(1);
break;
@@ -1952,6 +2030,10 @@ static int nfs4_handle_reclaim_lease_error(struct nfs_client *clp, int status)
dprintk("%s: exit with error %d for server %s\n",
__func__, -EPROTONOSUPPORT, clp->cl_hostname);
return -EPROTONOSUPPORT;
+ case -ENOSPC:
+ if (clp->cl_cons_state == NFS_CS_SESSION_INITING)
+ nfs_mark_client_ready(clp, -EIO);
+ return -EIO;
case -NFS4ERR_NOT_SAME: /* FixMe: implement recovery
* in nfs4_exchange_id */
default:
@@ -1982,7 +2064,6 @@ static int nfs4_establish_lease(struct nfs_client *clp)
put_cred(cred);
if (status != 0)
return status;
- pnfs_destroy_all_layouts(clp);
return 0;
}
@@ -2030,6 +2111,7 @@ static int nfs4_try_migration(struct nfs_server *server, const struct cred *cred
{
struct nfs_client *clp = server->nfs_client;
struct nfs4_fs_locations *locations = NULL;
+ struct nfs_fattr *fattr;
struct inode *inode;
struct page *page;
int status, result;
@@ -2039,16 +2121,19 @@ static int nfs4_try_migration(struct nfs_server *server, const struct cred *cred
(unsigned long long)server->fsid.minor,
clp->cl_hostname);
- result = 0;
page = alloc_page(GFP_KERNEL);
locations = kmalloc(sizeof(struct nfs4_fs_locations), GFP_KERNEL);
- if (page == NULL || locations == NULL) {
+ fattr = nfs_alloc_fattr();
+ if (page == NULL || locations == NULL || fattr == NULL) {
dprintk("<-- %s: no memory\n", __func__);
+ result = 0;
goto out;
}
+ locations->fattr = fattr;
inode = d_inode(server->super->s_root);
- result = nfs4_proc_get_locations(inode, locations, page, cred);
+ result = nfs4_proc_get_locations(server, NFS_FH(inode), locations,
+ page, cred);
if (result) {
dprintk("<-- %s: failed to retrieve fs_locations: %d\n",
__func__, result);
@@ -2056,15 +2141,20 @@ static int nfs4_try_migration(struct nfs_server *server, const struct cred *cred
}
result = -NFS4ERR_NXIO;
- if (!(locations->fattr.valid & NFS_ATTR_FATTR_V4_LOCATIONS)) {
+ if (!locations->nlocations)
+ goto out;
+
+ if (!(locations->fattr->valid & NFS_ATTR_FATTR_V4_LOCATIONS)) {
dprintk("<-- %s: No fs_locations data, migration skipped\n",
__func__);
goto out;
}
status = nfs4_begin_drain_session(clp);
- if (status != 0)
- return status;
+ if (status != 0) {
+ result = status;
+ goto out;
+ }
status = nfs4_replace_transport(server, locations);
if (status != 0) {
@@ -2079,6 +2169,8 @@ static int nfs4_try_migration(struct nfs_server *server, const struct cred *cred
out:
if (page != NULL)
__free_page(page);
+ if (locations != NULL)
+ kfree(locations->fattr);
kfree(locations);
if (result) {
pr_err("NFS: migration recovery failed (server %s)\n",
@@ -2223,11 +2315,11 @@ again:
case -ETIMEDOUT:
if (clnt->cl_softrtry)
break;
- /* Fall through */
+ fallthrough;
case -NFS4ERR_DELAY:
case -EAGAIN:
ssleep(1);
- /* Fall through */
+ fallthrough;
case -NFS4ERR_STALE_CLIENTID:
dprintk("NFS: %s after status %d, retrying\n",
__func__, status);
@@ -2239,7 +2331,7 @@ again:
}
if (clnt->cl_auth->au_flavor == RPC_AUTH_UNIX)
break;
- /* Fall through */
+ fallthrough;
case -NFS4ERR_CLID_INUSE:
case -NFS4ERR_WRONGSEC:
/* No point in retrying if we already used RPC_AUTH_UNIX */
@@ -2345,8 +2437,7 @@ static void nfs41_handle_recallable_state_revoked(struct nfs_client *clp)
{
/* FIXME: For now, we destroy all layouts. */
pnfs_destroy_all_layouts(clp);
- /* FIXME: For now, we test all delegations+open state+locks. */
- nfs41_handle_some_state_revoked(clp);
+ nfs_test_expired_all_delegations(clp);
dprintk("%s: Recallable state revoked on server %s!\n", __func__,
clp->cl_hostname);
}
@@ -2475,6 +2566,21 @@ static int nfs4_bind_conn_to_session(struct nfs_client *clp)
}
return 0;
}
+
+static void nfs4_layoutreturn_any_run(struct nfs_client *clp)
+{
+ int iomode = 0;
+
+ if (test_and_clear_bit(NFS4CLNT_RECALL_ANY_LAYOUT_READ, &clp->cl_state))
+ iomode += IOMODE_READ;
+ if (test_and_clear_bit(NFS4CLNT_RECALL_ANY_LAYOUT_RW, &clp->cl_state))
+ iomode += IOMODE_RW;
+ /* Note: IOMODE_READ + IOMODE_RW == IOMODE_ANY */
+ if (iomode) {
+ pnfs_layout_return_unused_byclid(clp, iomode);
+ set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
+ }
+}
#else /* CONFIG_NFS_V4_1 */
static int nfs4_reset_session(struct nfs_client *clp) { return 0; }
@@ -2482,15 +2588,28 @@ static int nfs4_bind_conn_to_session(struct nfs_client *clp)
{
return 0;
}
+
+static void nfs4_layoutreturn_any_run(struct nfs_client *clp)
+{
+}
#endif /* CONFIG_NFS_V4_1 */
static void nfs4_state_manager(struct nfs_client *clp)
{
+ unsigned int memflags;
int status = 0;
const char *section = "", *section_sep = "";
+ /*
+ * State recovery can deadlock if the direct reclaim code tries
+ * start NFS writeback. So ensure memory allocations are all
+ * GFP_NOFS.
+ */
+ memflags = memalloc_nofs_save();
+
/* Ensure exclusive access to NFSv4 state */
do {
+ trace_nfs4_state_mgr(clp);
clear_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
if (test_bit(NFS4CLNT_PURGE_STATE, &clp->cl_state)) {
section = "purge state";
@@ -2556,22 +2675,28 @@ static void nfs4_state_manager(struct nfs_client *clp)
section = "reclaim reboot";
status = nfs4_do_reclaim(clp,
clp->cl_mvops->reboot_recovery_ops);
+ if (status == 0)
+ status = pnfs_layout_handle_reboot(clp);
if (status == -EAGAIN)
continue;
if (status < 0)
goto out_error;
nfs4_state_end_reclaim_reboot(clp);
+ continue;
}
/* Detect expired delegations... */
if (test_and_clear_bit(NFS4CLNT_DELEGATION_EXPIRED, &clp->cl_state)) {
section = "detect expired delegations";
+ status = nfs4_begin_drain_session(clp);
+ if (status < 0)
+ goto out_error;
nfs_reap_expired_delegations(clp);
continue;
}
/* Now recover expired state... */
- if (test_and_clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state)) {
+ if (test_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state)) {
section = "reclaim nograce";
status = nfs4_do_reclaim(clp,
clp->cl_mvops->nograce_recovery_ops);
@@ -2579,35 +2704,55 @@ static void nfs4_state_manager(struct nfs_client *clp)
continue;
if (status < 0)
goto out_error;
+ clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
}
+ memalloc_nofs_restore(memflags);
nfs4_end_drain_session(clp);
nfs4_clear_state_manager_bit(clp);
- if (!test_and_set_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state)) {
+ if (test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state) &&
+ !test_and_set_bit(NFS4CLNT_MANAGER_RUNNING,
+ &clp->cl_state)) {
+ memflags = memalloc_nofs_save();
+ continue;
+ }
+
+ if (!test_and_set_bit(NFS4CLNT_RECALL_RUNNING, &clp->cl_state)) {
if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) {
nfs_client_return_marked_delegations(clp);
set_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state);
}
- clear_bit(NFS4CLNT_DELEGRETURN_RUNNING, &clp->cl_state);
+ nfs4_layoutreturn_any_run(clp);
+ clear_bit(NFS4CLNT_RECALL_RUNNING, &clp->cl_state);
}
- /* Did we race with an attempt to give us more work? */
- if (!test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state))
- return;
- if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
- return;
+ return;
+
} while (refcount_read(&clp->cl_count) > 1 && !signalled());
goto out_drain;
out_error:
if (strlen(section))
section_sep = ": ";
+ trace_nfs4_state_mgr_failed(clp, section, status);
pr_warn_ratelimited("NFS: state manager%s%s failed on NFSv4 server %s"
" with error %d\n", section_sep, section,
clp->cl_hostname, -status);
- ssleep(1);
+ switch (status) {
+ case -ENETDOWN:
+ case -ENETUNREACH:
+ nfs_mark_client_ready(clp, -EIO);
+ break;
+ case -EINVAL:
+ nfs_mark_client_ready(clp, status);
+ break;
+ default:
+ ssleep(1);
+ break;
+ }
out_drain:
+ memalloc_nofs_restore(memflags);
nfs4_end_drain_session(clp);
nfs4_clear_state_manager_bit(clp);
}
@@ -2615,16 +2760,35 @@ out_drain:
static int nfs4_run_state_manager(void *ptr)
{
struct nfs_client *clp = ptr;
+ struct rpc_clnt *cl = clp->cl_rpcclient;
+
+ while (cl != cl->cl_parent)
+ cl = cl->cl_parent;
allow_signal(SIGKILL);
+again:
nfs4_state_manager(clp);
+
+ if (test_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state) &&
+ !test_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state)) {
+ wait_var_event_interruptible(&clp->cl_state,
+ test_bit(NFS4CLNT_RUN_MANAGER,
+ &clp->cl_state));
+ if (!atomic_read(&cl->cl_swapper))
+ clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state);
+ if (refcount_read(&clp->cl_count) > 1 && !signalled() &&
+ !test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state))
+ goto again;
+ /* Either no longer a swapper, or were signalled */
+ clear_bit(NFS4CLNT_MANAGER_AVAILABLE, &clp->cl_state);
+ }
+
+ if (refcount_read(&clp->cl_count) > 1 && !signalled() &&
+ test_bit(NFS4CLNT_RUN_MANAGER, &clp->cl_state) &&
+ !test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state))
+ goto again;
+
nfs_put_client(clp);
- module_put_and_exit(0);
+ module_put_and_kthread_exit(0);
return 0;
}
-
-/*
- * Local variables:
- * c-basic-offset: 8
- * End:
- */
generated by cgit (git 2.34.1) at 2025-12-24 21:37:34 +0000