summaryrefslogtreecommitdiff
path: root/fs/nfs/pnfs.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/nfs/pnfs.c')
-rw-r--r--fs/nfs/pnfs.c1003
1 files changed, 725 insertions, 278 deletions
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 53726da5c010..f157d43d1312 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -61,6 +61,7 @@ static void pnfs_free_returned_lsegs(struct pnfs_layout_hdr *lo,
u32 seq);
static bool pnfs_lseg_dec_and_remove_zero(struct pnfs_layout_segment *lseg,
struct list_head *tmp_list);
+static int pnfs_layout_return_on_reboot(struct pnfs_layout_hdr *lo);
/* Return the registered pnfs layout driver module matching given id */
static struct pnfs_layoutdriver_type *
@@ -92,6 +93,17 @@ find_pnfs_driver(u32 id)
return local;
}
+const struct pnfs_layoutdriver_type *pnfs_find_layoutdriver(u32 id)
+{
+ return find_pnfs_driver(id);
+}
+
+void pnfs_put_layoutdriver(const struct pnfs_layoutdriver_type *ld)
+{
+ if (ld)
+ module_put(ld->owner);
+}
+
void
unset_pnfs_layoutdriver(struct nfs_server *nfss)
{
@@ -268,11 +280,11 @@ pnfs_free_layout_hdr(struct pnfs_layout_hdr *lo)
struct nfs_server *server = NFS_SERVER(lo->plh_inode);
struct pnfs_layoutdriver_type *ld = server->pnfs_curr_ld;
- if (!list_empty(&lo->plh_layouts)) {
+ if (test_and_clear_bit(NFS_LAYOUT_HASHED, &lo->plh_flags)) {
struct nfs_client *clp = server->nfs_client;
spin_lock(&clp->cl_lock);
- list_del_init(&lo->plh_layouts);
+ list_del_rcu(&lo->plh_layouts);
spin_unlock(&clp->cl_lock);
}
put_cred(lo->plh_lc_cred);
@@ -304,11 +316,39 @@ pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo)
if (!list_empty(&lo->plh_segs))
WARN_ONCE(1, "NFS: BUG unfreed layout segments.\n");
pnfs_detach_layout_hdr(lo);
+ /* Notify pnfs_destroy_layout_final() that we're done */
+ if (inode_state_read(inode) & (I_FREEING | I_CLEAR))
+ wake_up_var_locked(lo, &inode->i_lock);
spin_unlock(&inode->i_lock);
pnfs_free_layout_hdr(lo);
}
}
+static struct inode *
+pnfs_grab_inode_layout_hdr(struct pnfs_layout_hdr *lo)
+{
+ struct inode *inode = igrab(lo->plh_inode);
+ if (inode)
+ return inode;
+ set_bit(NFS_LAYOUT_INODE_FREEING, &lo->plh_flags);
+ return NULL;
+}
+
+/*
+ * Compare 2 layout stateid sequence ids, to see which is newer,
+ * taking into account wraparound issues.
+ */
+static bool pnfs_seqid_is_newer(u32 s1, u32 s2)
+{
+ return (s32)(s1 - s2) > 0;
+}
+
+static void pnfs_barrier_update(struct pnfs_layout_hdr *lo, u32 newseq)
+{
+ if (pnfs_seqid_is_newer(newseq, lo->plh_barrier) || !lo->plh_barrier)
+ lo->plh_barrier = newseq;
+}
+
static void
pnfs_set_plh_return_info(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode,
u32 seq)
@@ -317,10 +357,15 @@ pnfs_set_plh_return_info(struct pnfs_layout_hdr *lo, enum pnfs_iomode iomode,
iomode = IOMODE_ANY;
lo->plh_return_iomode = iomode;
set_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags);
- if (seq != 0) {
- WARN_ON_ONCE(lo->plh_return_seq != 0 && lo->plh_return_seq != seq);
+ /*
+ * We must set lo->plh_return_seq to avoid livelocks with
+ * pnfs_layout_need_return()
+ */
+ if (seq == 0)
+ seq = be32_to_cpu(lo->plh_stateid.seqid);
+ if (!lo->plh_return_seq || pnfs_seqid_is_newer(seq, lo->plh_return_seq))
lo->plh_return_seq = seq;
- }
+ pnfs_barrier_update(lo, seq);
}
static void
@@ -359,9 +404,10 @@ pnfs_clear_lseg_state(struct pnfs_layout_segment *lseg,
}
/*
- * Update the seqid of a layout stateid
+ * Update the seqid of a layout stateid after receiving
+ * NFS4ERR_OLD_STATEID
*/
-bool nfs4_layoutreturn_refresh_stateid(nfs4_stateid *dst,
+bool nfs4_layout_refresh_old_stateid(nfs4_stateid *dst,
struct pnfs_layout_range *dst_range,
struct inode *inode)
{
@@ -377,7 +423,15 @@ bool nfs4_layoutreturn_refresh_stateid(nfs4_stateid *dst,
spin_lock(&inode->i_lock);
lo = NFS_I(inode)->layout;
- if (lo && nfs4_stateid_match_other(dst, &lo->plh_stateid)) {
+ if (lo && pnfs_layout_is_valid(lo) &&
+ nfs4_stateid_match_other(dst, &lo->plh_stateid)) {
+ /* Is our call using the most recent seqid? If so, bump it */
+ if (!nfs4_stateid_is_newer(&lo->plh_stateid, dst)) {
+ nfs4_stateid_seqid_inc(dst);
+ ret = true;
+ goto out;
+ }
+ /* Try to update the seqid to the most recent */
err = pnfs_mark_matching_lsegs_return(lo, &head, &range, 0);
if (err != -EBUSY) {
dst->seqid = lo->plh_stateid.seqid;
@@ -385,6 +439,7 @@ bool nfs4_layoutreturn_refresh_stateid(nfs4_stateid *dst,
ret = true;
}
}
+out:
spin_unlock(&inode->i_lock);
pnfs_free_lseg_list(&head);
return ret;
@@ -413,12 +468,25 @@ pnfs_mark_layout_stateid_invalid(struct pnfs_layout_hdr *lo,
pnfs_clear_lseg_state(lseg, lseg_list);
pnfs_clear_layoutreturn_info(lo);
pnfs_free_returned_lsegs(lo, lseg_list, &range, 0);
+ set_bit(NFS_LAYOUT_DRAIN, &lo->plh_flags);
if (test_bit(NFS_LAYOUT_RETURN, &lo->plh_flags) &&
!test_and_set_bit(NFS_LAYOUT_RETURN_LOCK, &lo->plh_flags))
pnfs_clear_layoutreturn_waitbit(lo);
return !list_empty(&lo->plh_segs);
}
+static int pnfs_mark_layout_stateid_return(struct pnfs_layout_hdr *lo,
+ struct list_head *lseg_list,
+ enum pnfs_iomode iomode, u32 seq)
+{
+ struct pnfs_layout_range range = {
+ .iomode = iomode,
+ .length = NFS4_MAX_UINT64,
+ };
+
+ return pnfs_mark_matching_lsegs_return(lo, lseg_list, &range, seq);
+}
+
static int
pnfs_iomode_to_fail_bit(u32 iomode)
{
@@ -454,7 +522,7 @@ pnfs_layout_io_set_failed(struct pnfs_layout_hdr *lo, u32 iomode)
spin_lock(&inode->i_lock);
pnfs_layout_set_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode));
- pnfs_mark_matching_lsegs_invalid(lo, &head, &range, 0);
+ pnfs_mark_matching_lsegs_return(lo, &head, &range, 0);
spin_unlock(&inode->i_lock);
pnfs_free_lseg_list(&head);
dprintk("%s Setting layout IOMODE_%s fail bit\n", __func__,
@@ -486,6 +554,7 @@ pnfs_init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg,
{
INIT_LIST_HEAD(&lseg->pls_list);
INIT_LIST_HEAD(&lseg->pls_lc_list);
+ INIT_LIST_HEAD(&lseg->pls_commits);
refcount_set(&lseg->pls_refcount, 1);
set_bit(NFS_LSEG_VALID, &lseg->pls_flags);
lseg->pls_layout = lo;
@@ -550,10 +619,6 @@ pnfs_put_lseg(struct pnfs_layout_segment *lseg)
inode = lo->plh_inode;
if (refcount_dec_and_lock(&lseg->pls_refcount, &inode->i_lock)) {
- if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags)) {
- spin_unlock(&inode->i_lock);
- return;
- }
pnfs_get_layout_hdr(lo);
pnfs_layout_remove_lseg(lo, lseg);
if (pnfs_cache_lseg_for_layoutreturn(lo, lseg))
@@ -613,15 +678,6 @@ static int mark_lseg_invalid(struct pnfs_layout_segment *lseg,
return rv;
}
-/*
- * Compare 2 layout stateid sequence ids, to see which is newer,
- * taking into account wraparound issues.
- */
-static bool pnfs_seqid_is_newer(u32 s1, u32 s2)
-{
- return (s32)(s1 - s2) > 0;
-}
-
static bool
pnfs_should_free_range(const struct pnfs_layout_range *lseg_range,
const struct pnfs_layout_range *recall_range)
@@ -665,6 +721,7 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
u32 seq)
{
struct pnfs_layout_segment *lseg, *next;
+ struct nfs_server *server = NFS_SERVER(lo->plh_inode);
int remaining = 0;
dprintk("%s:Begin lo %p\n", __func__, lo);
@@ -677,13 +734,23 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
"offset %llu length %llu\n", __func__,
lseg, lseg->pls_range.iomode, lseg->pls_seq,
lseg->pls_range.offset, lseg->pls_range.length);
- if (!mark_lseg_invalid(lseg, tmp_list))
- remaining++;
+ if (mark_lseg_invalid(lseg, tmp_list))
+ continue;
+ remaining++;
+ pnfs_lseg_cancel_io(server, lseg);
}
dprintk("%s:Return %i\n", __func__, remaining);
return remaining;
}
+static void pnfs_reset_return_info(struct pnfs_layout_hdr *lo)
+{
+ struct pnfs_layout_segment *lseg;
+
+ list_for_each_entry(lseg, &lo->plh_return_segs, pls_list)
+ pnfs_set_plh_return_info(lo, lseg->pls_range.iomode, 0);
+}
+
static void
pnfs_free_returned_lsegs(struct pnfs_layout_hdr *lo,
struct list_head *free_me,
@@ -713,8 +780,7 @@ pnfs_free_lseg_list(struct list_head *free_me)
}
}
-void
-pnfs_destroy_layout(struct nfs_inode *nfsi)
+static struct pnfs_layout_hdr *__pnfs_destroy_layout(struct nfs_inode *nfsi)
{
struct pnfs_layout_hdr *lo;
LIST_HEAD(tmp_list);
@@ -732,9 +798,28 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
pnfs_put_layout_hdr(lo);
} else
spin_unlock(&nfsi->vfs_inode.i_lock);
+ return lo;
+}
+
+void pnfs_destroy_layout(struct nfs_inode *nfsi)
+{
+ __pnfs_destroy_layout(nfsi);
}
EXPORT_SYMBOL_GPL(pnfs_destroy_layout);
+void pnfs_destroy_layout_final(struct nfs_inode *nfsi)
+{
+ struct pnfs_layout_hdr *lo = __pnfs_destroy_layout(nfsi);
+ struct inode *inode = &nfsi->vfs_inode;
+
+ if (lo) {
+ spin_lock(&inode->i_lock);
+ wait_var_event_spinlock(lo, nfsi->layout != lo,
+ &inode->i_lock);
+ spin_unlock(&inode->i_lock);
+ }
+}
+
static bool
pnfs_layout_add_bulk_destroy_list(struct inode *inode,
struct list_head *layout_list)
@@ -758,22 +843,33 @@ static int
pnfs_layout_bulk_destroy_byserver_locked(struct nfs_client *clp,
struct nfs_server *server,
struct list_head *layout_list)
+ __must_hold(&clp->cl_lock)
+ __must_hold(RCU)
{
struct pnfs_layout_hdr *lo, *next;
struct inode *inode;
list_for_each_entry_safe(lo, next, &server->layouts, plh_layouts) {
- if (test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags))
- continue;
- inode = igrab(lo->plh_inode);
- if (inode == NULL)
+ if (test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags) ||
+ test_bit(NFS_LAYOUT_INODE_FREEING, &lo->plh_flags) ||
+ !list_empty(&lo->plh_bulk_destroy))
continue;
- list_del_init(&lo->plh_layouts);
- if (pnfs_layout_add_bulk_destroy_list(inode, layout_list))
- continue;
- rcu_read_unlock();
- spin_unlock(&clp->cl_lock);
- iput(inode);
+ /* If the sb is being destroyed, just bail */
+ if (!nfs_sb_active(server->super))
+ break;
+ inode = pnfs_grab_inode_layout_hdr(lo);
+ if (inode != NULL) {
+ if (pnfs_layout_add_bulk_destroy_list(inode,
+ layout_list))
+ continue;
+ rcu_read_unlock();
+ spin_unlock(&clp->cl_lock);
+ iput(inode);
+ } else {
+ rcu_read_unlock();
+ spin_unlock(&clp->cl_lock);
+ }
+ nfs_sb_deactive(server->super);
spin_lock(&clp->cl_lock);
rcu_read_lock();
return -EAGAIN;
@@ -783,7 +879,7 @@ pnfs_layout_bulk_destroy_byserver_locked(struct nfs_client *clp,
static int
pnfs_layout_free_bulk_destroy_list(struct list_head *layout_list,
- bool is_bulk_recall)
+ enum pnfs_layout_destroy_mode mode)
{
struct pnfs_layout_hdr *lo;
struct inode *inode;
@@ -801,8 +897,11 @@ pnfs_layout_free_bulk_destroy_list(struct list_head *layout_list,
spin_lock(&inode->i_lock);
list_del_init(&lo->plh_bulk_destroy);
- if (pnfs_mark_layout_stateid_invalid(lo, &lseg_list)) {
- if (is_bulk_recall)
+ if (mode == PNFS_LAYOUT_FILE_BULK_RETURN) {
+ pnfs_mark_layout_stateid_return(lo, &lseg_list,
+ IOMODE_ANY, 0);
+ } else if (pnfs_mark_layout_stateid_invalid(lo, &lseg_list)) {
+ if (mode == PNFS_LAYOUT_BULK_RETURN)
set_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags);
ret = -EAGAIN;
}
@@ -811,15 +910,13 @@ pnfs_layout_free_bulk_destroy_list(struct list_head *layout_list,
/* Free all lsegs that are attached to commit buckets */
nfs_commit_inode(inode, 0);
pnfs_put_layout_hdr(lo);
- iput(inode);
+ nfs_iput_and_deactive(inode);
}
return ret;
}
-int
-pnfs_destroy_layouts_byfsid(struct nfs_client *clp,
- struct nfs_fsid *fsid,
- bool is_recall)
+int pnfs_layout_destroy_byfsid(struct nfs_client *clp, struct nfs_fsid *fsid,
+ enum pnfs_layout_destroy_mode mode)
{
struct nfs_server *server;
LIST_HEAD(layout_list);
@@ -838,37 +935,44 @@ restart:
rcu_read_unlock();
spin_unlock(&clp->cl_lock);
- if (list_empty(&layout_list))
- return 0;
- return pnfs_layout_free_bulk_destroy_list(&layout_list, is_recall);
+ return pnfs_layout_free_bulk_destroy_list(&layout_list, mode);
}
-int
-pnfs_destroy_layouts_byclid(struct nfs_client *clp,
- bool is_recall)
+static void pnfs_layout_build_destroy_list_byclient(struct nfs_client *clp,
+ struct list_head *list)
{
struct nfs_server *server;
- LIST_HEAD(layout_list);
spin_lock(&clp->cl_lock);
rcu_read_lock();
restart:
list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) {
- if (pnfs_layout_bulk_destroy_byserver_locked(clp,
- server,
- &layout_list) != 0)
+ if (pnfs_layout_bulk_destroy_byserver_locked(clp, server,
+ list) != 0)
goto restart;
}
rcu_read_unlock();
spin_unlock(&clp->cl_lock);
+}
- if (list_empty(&layout_list))
- return 0;
- return pnfs_layout_free_bulk_destroy_list(&layout_list, is_recall);
+static int pnfs_layout_do_destroy_byclid(struct nfs_client *clp,
+ struct list_head *list,
+ enum pnfs_layout_destroy_mode mode)
+{
+ pnfs_layout_build_destroy_list_byclient(clp, list);
+ return pnfs_layout_free_bulk_destroy_list(list, mode);
+}
+
+int pnfs_layout_destroy_byclid(struct nfs_client *clp,
+ enum pnfs_layout_destroy_mode mode)
+{
+ LIST_HEAD(layout_list);
+
+ return pnfs_layout_do_destroy_byclid(clp, &layout_list, mode);
}
/*
- * Called by the state manger to remove all layouts established under an
+ * Called by the state manager to remove all layouts established under an
* expired lease.
*/
void
@@ -877,40 +981,112 @@ pnfs_destroy_all_layouts(struct nfs_client *clp)
nfs4_deviceid_mark_client_invalid(clp);
nfs4_deviceid_purge_client(clp);
- pnfs_destroy_layouts_byclid(clp, false);
+ pnfs_layout_destroy_byclid(clp, PNFS_LAYOUT_INVALIDATE);
+}
+
+static void pnfs_layout_build_recover_list_byclient(struct nfs_client *clp,
+ struct list_head *list)
+{
+ struct nfs_server *server;
+
+ spin_lock(&clp->cl_lock);
+ rcu_read_lock();
+restart:
+ list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) {
+ if (!(server->caps & NFS_CAP_REBOOT_LAYOUTRETURN))
+ continue;
+ if (pnfs_layout_bulk_destroy_byserver_locked(clp, server,
+ list) != 0)
+ goto restart;
+ }
+ rcu_read_unlock();
+ spin_unlock(&clp->cl_lock);
+}
+
+static int pnfs_layout_bulk_list_reboot(struct list_head *list)
+{
+ struct pnfs_layout_hdr *lo;
+ struct nfs_server *server;
+ int ret;
+
+ list_for_each_entry(lo, list, plh_bulk_destroy) {
+ server = NFS_SERVER(lo->plh_inode);
+ ret = pnfs_layout_return_on_reboot(lo);
+ switch (ret) {
+ case 0:
+ continue;
+ case -NFS4ERR_BAD_STATEID:
+ server->caps &= ~NFS_CAP_REBOOT_LAYOUTRETURN;
+ break;
+ case -NFS4ERR_NO_GRACE:
+ break;
+ default:
+ goto err;
+ }
+ break;
+ }
+ return 0;
+err:
+ return ret;
+}
+
+int pnfs_layout_handle_reboot(struct nfs_client *clp)
+{
+ LIST_HEAD(list);
+ int ret = 0, ret2;
+
+ pnfs_layout_build_recover_list_byclient(clp, &list);
+ if (!list_empty(&list))
+ ret = pnfs_layout_bulk_list_reboot(&list);
+ ret2 = pnfs_layout_do_destroy_byclid(clp, &list,
+ PNFS_LAYOUT_INVALIDATE);
+ if (!ret)
+ ret = ret2;
+ return (ret == 0) ? 0 : -EAGAIN;
+}
+
+static void
+pnfs_set_layout_cred(struct pnfs_layout_hdr *lo, const struct cred *cred)
+{
+ const struct cred *old;
+
+ if (cred && cred_fscmp(lo->plh_lc_cred, cred) != 0) {
+ old = xchg(&lo->plh_lc_cred, get_cred(cred));
+ put_cred(old);
+ }
}
/* update lo->plh_stateid with new if is more recent */
void
pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo, const nfs4_stateid *new,
- bool update_barrier)
+ const struct cred *cred, bool update_barrier)
{
- u32 oldseq, newseq, new_barrier = 0;
-
- oldseq = be32_to_cpu(lo->plh_stateid.seqid);
- newseq = be32_to_cpu(new->seqid);
+ u32 oldseq = be32_to_cpu(lo->plh_stateid.seqid);
+ u32 newseq = be32_to_cpu(new->seqid);
if (!pnfs_layout_is_valid(lo)) {
+ pnfs_set_layout_cred(lo, cred);
nfs4_stateid_copy(&lo->plh_stateid, new);
lo->plh_barrier = newseq;
pnfs_clear_layoutreturn_info(lo);
clear_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags);
return;
}
- if (pnfs_seqid_is_newer(newseq, oldseq)) {
+
+ if (pnfs_seqid_is_newer(newseq, oldseq))
nfs4_stateid_copy(&lo->plh_stateid, new);
- /*
- * Because of wraparound, we want to keep the barrier
- * "close" to the current seqids.
- */
- new_barrier = newseq - atomic_read(&lo->plh_outstanding);
- }
- if (update_barrier)
- new_barrier = be32_to_cpu(new->seqid);
- else if (new_barrier == 0)
+
+ if (update_barrier) {
+ pnfs_barrier_update(lo, newseq);
return;
- if (pnfs_seqid_is_newer(new_barrier, lo->plh_barrier))
- lo->plh_barrier = new_barrier;
+ }
+ /*
+ * Because of wraparound, we want to keep the barrier
+ * "close" to the current seqids. We really only want to
+ * get here from a layoutget call.
+ */
+ if (atomic_read(&lo->plh_outstanding) == 1)
+ pnfs_barrier_update(lo, be32_to_cpu(lo->plh_stateid.seqid));
}
static bool
@@ -919,7 +1095,7 @@ pnfs_layout_stateid_blocked(const struct pnfs_layout_hdr *lo,
{
u32 seqid = be32_to_cpu(stateid->seqid);
- return !pnfs_seqid_is_newer(seqid, lo->plh_barrier);
+ return lo->plh_barrier && pnfs_seqid_is_newer(lo->plh_barrier, seqid);
}
/* lget is set to 1 if called from inside send_layoutget call chain */
@@ -1038,7 +1214,7 @@ pnfs_alloc_init_layoutget_args(struct inode *ino,
lgp->args.ctx = get_nfs_open_context(ctx);
nfs4_stateid_copy(&lgp->args.stateid, stateid);
lgp->gfp_flags = gfp_flags;
- lgp->cred = get_cred(ctx->cred);
+ lgp->cred = ctx->cred;
return lgp;
}
@@ -1047,9 +1223,7 @@ void pnfs_layoutget_free(struct nfs4_layoutget *lgp)
size_t max_pages = lgp->args.layout.pglen / PAGE_SIZE;
nfs4_free_pages(lgp->args.layout.pages, max_pages);
- if (lgp->args.inode)
- pnfs_put_layout_hdr(NFS_I(lgp->args.inode)->layout);
- put_cred(lgp->cred);
+ pnfs_put_layout_hdr(lgp->lo);
put_nfs_open_context(lgp->args.ctx);
kfree(lgp);
}
@@ -1069,6 +1243,33 @@ static void pnfs_clear_layoutcommit(struct inode *inode,
}
}
+static void
+pnfs_layoutreturn_retry_later_locked(struct pnfs_layout_hdr *lo,
+ const nfs4_stateid *arg_stateid,
+ const struct pnfs_layout_range *range,
+ struct list_head *freeme)
+{
+ if (pnfs_layout_is_valid(lo) &&
+ nfs4_stateid_match_other(&lo->plh_stateid, arg_stateid))
+ pnfs_reset_return_info(lo);
+ else
+ pnfs_mark_layout_stateid_invalid(lo, freeme);
+ pnfs_clear_layoutreturn_waitbit(lo);
+}
+
+void pnfs_layoutreturn_retry_later(struct pnfs_layout_hdr *lo,
+ const nfs4_stateid *arg_stateid,
+ const struct pnfs_layout_range *range)
+{
+ struct inode *inode = lo->plh_inode;
+ LIST_HEAD(freeme);
+
+ spin_lock(&inode->i_lock);
+ pnfs_layoutreturn_retry_later_locked(lo, arg_stateid, range, &freeme);
+ spin_unlock(&inode->i_lock);
+ pnfs_free_lseg_list(&freeme);
+}
+
void pnfs_layoutreturn_free_lsegs(struct pnfs_layout_hdr *lo,
const nfs4_stateid *arg_stateid,
const struct pnfs_layout_range *range,
@@ -1078,15 +1279,15 @@ void pnfs_layoutreturn_free_lsegs(struct pnfs_layout_hdr *lo,
LIST_HEAD(freeme);
spin_lock(&inode->i_lock);
- if (!pnfs_layout_is_valid(lo) || !arg_stateid ||
- !nfs4_stateid_match_other(&lo->plh_stateid, arg_stateid))
+ if (!nfs4_stateid_match_other(&lo->plh_stateid, arg_stateid))
goto out_unlock;
- if (stateid) {
+ if (stateid && pnfs_layout_is_valid(lo)) {
u32 seq = be32_to_cpu(arg_stateid->seqid);
pnfs_mark_matching_lsegs_invalid(lo, &freeme, range, seq);
pnfs_free_returned_lsegs(lo, &freeme, range, seq);
- pnfs_set_layout_stateid(lo, stateid, true);
+ pnfs_set_layout_stateid(lo, stateid, NULL, true);
+ pnfs_reset_return_info(lo);
} else
pnfs_mark_layout_stateid_invalid(lo, &freeme);
out_unlock:
@@ -1099,30 +1300,27 @@ out_unlock:
static bool
pnfs_prepare_layoutreturn(struct pnfs_layout_hdr *lo,
nfs4_stateid *stateid,
+ const struct cred **cred,
enum pnfs_iomode *iomode)
{
/* Serialise LAYOUTGET/LAYOUTRETURN */
- if (atomic_read(&lo->plh_outstanding) != 0)
+ if (atomic_read(&lo->plh_outstanding) != 0 && lo->plh_return_seq == 0)
return false;
if (test_and_set_bit(NFS_LAYOUT_RETURN_LOCK, &lo->plh_flags))
return false;
set_bit(NFS_LAYOUT_RETURN, &lo->plh_flags);
pnfs_get_layout_hdr(lo);
+ nfs4_stateid_copy(stateid, &lo->plh_stateid);
+ *cred = get_cred(lo->plh_lc_cred);
if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags)) {
- if (stateid != NULL) {
- nfs4_stateid_copy(stateid, &lo->plh_stateid);
- if (lo->plh_return_seq != 0)
- stateid->seqid = cpu_to_be32(lo->plh_return_seq);
- }
+ if (lo->plh_return_seq != 0)
+ stateid->seqid = cpu_to_be32(lo->plh_return_seq);
if (iomode != NULL)
*iomode = lo->plh_return_iomode;
pnfs_clear_layoutreturn_info(lo);
- return true;
- }
- if (stateid != NULL)
- nfs4_stateid_copy(stateid, &lo->plh_stateid);
- if (iomode != NULL)
+ } else if (iomode != NULL)
*iomode = IOMODE_ANY;
+ pnfs_barrier_update(lo, be32_to_cpu(stateid->seqid));
return true;
}
@@ -1144,20 +1342,26 @@ pnfs_init_layoutreturn_args(struct nfs4_layoutreturn_args *args,
}
static int
-pnfs_send_layoutreturn(struct pnfs_layout_hdr *lo, const nfs4_stateid *stateid,
- enum pnfs_iomode iomode, bool sync)
+pnfs_send_layoutreturn(struct pnfs_layout_hdr *lo,
+ const nfs4_stateid *stateid,
+ const struct cred **pcred,
+ enum pnfs_iomode iomode,
+ unsigned int flags)
{
struct inode *ino = lo->plh_inode;
struct pnfs_layoutdriver_type *ld = NFS_SERVER(ino)->pnfs_curr_ld;
struct nfs4_layoutreturn *lrp;
+ const struct cred *cred = *pcred;
int status = 0;
- lrp = kzalloc(sizeof(*lrp), GFP_NOFS);
+ *pcred = NULL;
+ lrp = kzalloc(sizeof(*lrp), nfs_io_gfp_mask());
if (unlikely(lrp == NULL)) {
status = -ENOMEM;
spin_lock(&ino->i_lock);
pnfs_clear_layoutreturn_waitbit(lo);
spin_unlock(&ino->i_lock);
+ put_cred(cred);
pnfs_put_layout_hdr(lo);
goto out;
}
@@ -1165,11 +1369,11 @@ pnfs_send_layoutreturn(struct pnfs_layout_hdr *lo, const nfs4_stateid *stateid,
pnfs_init_layoutreturn_args(&lrp->args, lo, stateid, iomode);
lrp->args.ld_private = &lrp->ld_private;
lrp->clp = NFS_SERVER(ino)->nfs_client;
- lrp->cred = lo->plh_lc_cred;
+ lrp->cred = cred;
if (ld->prepare_layoutreturn)
ld->prepare_layoutreturn(&lrp->args);
- status = nfs4_proc_layoutreturn(lrp, sync);
+ status = nfs4_proc_layoutreturn(lrp, flags);
out:
dprintk("<-- %s status: %d\n", __func__, status);
return status;
@@ -1179,27 +1383,11 @@ out:
static bool
pnfs_layout_need_return(struct pnfs_layout_hdr *lo)
{
- struct pnfs_layout_segment *s;
- enum pnfs_iomode iomode;
- u32 seq;
-
if (!test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags))
return false;
-
- seq = lo->plh_return_seq;
- iomode = lo->plh_return_iomode;
-
- /* Defer layoutreturn until all recalled lsegs are done */
- list_for_each_entry(s, &lo->plh_segs, pls_list) {
- if (seq && pnfs_seqid_is_newer(s->pls_seq, seq))
- continue;
- if (iomode != IOMODE_ANY && s->pls_range.iomode != iomode)
- continue;
- if (test_bit(NFS_LSEG_LAYOUTRETURN, &s->pls_flags))
- return false;
- }
-
- return true;
+ return pnfs_mark_layout_stateid_return(lo, &lo->plh_return_segs,
+ lo->plh_return_iomode,
+ lo->plh_return_seq) != EBUSY;
}
static void pnfs_layoutreturn_before_put_layout_hdr(struct pnfs_layout_hdr *lo)
@@ -1210,15 +1398,17 @@ static void pnfs_layoutreturn_before_put_layout_hdr(struct pnfs_layout_hdr *lo)
return;
spin_lock(&inode->i_lock);
if (pnfs_layout_need_return(lo)) {
+ const struct cred *cred;
nfs4_stateid stateid;
enum pnfs_iomode iomode;
bool send;
- send = pnfs_prepare_layoutreturn(lo, &stateid, &iomode);
+ send = pnfs_prepare_layoutreturn(lo, &stateid, &cred, &iomode);
spin_unlock(&inode->i_lock);
if (send) {
/* Send an async layoutreturn so we dont deadlock */
- pnfs_send_layoutreturn(lo, &stateid, iomode, false);
+ pnfs_send_layoutreturn(lo, &stateid, &cred, iomode,
+ PNFS_FL_LAYOUTRETURN_ASYNC);
}
} else
spin_unlock(&inode->i_lock);
@@ -1237,7 +1427,13 @@ _pnfs_return_layout(struct inode *ino)
{
struct pnfs_layout_hdr *lo = NULL;
struct nfs_inode *nfsi = NFS_I(ino);
+ struct pnfs_layout_range range = {
+ .iomode = IOMODE_ANY,
+ .offset = 0,
+ .length = NFS4_MAX_UINT64,
+ };
LIST_HEAD(tmp_list);
+ const struct cred *cred;
nfs4_stateid stateid;
int status = 0;
bool send, valid_layout;
@@ -1263,29 +1459,26 @@ _pnfs_return_layout(struct inode *ino)
}
valid_layout = pnfs_layout_is_valid(lo);
pnfs_clear_layoutcommit(ino, &tmp_list);
- pnfs_mark_matching_lsegs_invalid(lo, &tmp_list, NULL, 0);
+ pnfs_mark_matching_lsegs_return(lo, &tmp_list, &range, 0);
- if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) {
- struct pnfs_layout_range range = {
- .iomode = IOMODE_ANY,
- .offset = 0,
- .length = NFS4_MAX_UINT64,
- };
+ if (NFS_SERVER(ino)->pnfs_curr_ld->return_range)
NFS_SERVER(ino)->pnfs_curr_ld->return_range(lo, &range);
- }
/* Don't send a LAYOUTRETURN if list was initially empty */
if (!test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags) ||
!valid_layout) {
spin_unlock(&ino->i_lock);
dprintk("NFS: %s no layout segments to return\n", __func__);
- goto out_put_layout_hdr;
+ goto out_wait_layoutreturn;
}
- send = pnfs_prepare_layoutreturn(lo, &stateid, NULL);
+ send = pnfs_prepare_layoutreturn(lo, &stateid, &cred, NULL);
spin_unlock(&ino->i_lock);
if (send)
- status = pnfs_send_layoutreturn(lo, &stateid, IOMODE_ANY, true);
+ status = pnfs_send_layoutreturn(lo, &stateid, &cred, IOMODE_ANY,
+ 0);
+out_wait_layoutreturn:
+ wait_on_bit(&lo->plh_flags, NFS_LAYOUT_RETURN, TASK_UNINTERRUPTIBLE);
out_put_layout_hdr:
pnfs_free_lseg_list(&tmp_list);
pnfs_put_layout_hdr(lo);
@@ -1321,6 +1514,24 @@ pnfs_commit_and_return_layout(struct inode *inode)
return ret;
}
+static int pnfs_layout_return_on_reboot(struct pnfs_layout_hdr *lo)
+{
+ struct inode *inode = lo->plh_inode;
+ const struct cred *cred;
+
+ spin_lock(&inode->i_lock);
+ if (!pnfs_layout_is_valid(lo)) {
+ spin_unlock(&inode->i_lock);
+ return 0;
+ }
+ cred = get_cred(lo->plh_lc_cred);
+ pnfs_get_layout_hdr(lo);
+ spin_unlock(&inode->i_lock);
+
+ return pnfs_send_layoutreturn(lo, &zero_stateid, &cred, IOMODE_ANY,
+ PNFS_FL_LAYOUTRETURN_PRIVILEGED);
+}
+
bool pnfs_roc(struct inode *ino,
struct nfs4_layoutreturn_args *args,
struct nfs4_layoutreturn_res *res,
@@ -1331,6 +1542,7 @@ bool pnfs_roc(struct inode *ino,
struct nfs4_state *state;
struct pnfs_layout_hdr *lo;
struct pnfs_layout_segment *lseg, *next;
+ const struct cred *lc_cred;
nfs4_stateid stateid;
enum pnfs_iomode iomode = 0;
bool layoutreturn = false, roc = false;
@@ -1400,15 +1612,16 @@ retry:
* 2. we don't send layoutreturn
*/
/* lo ref dropped in pnfs_roc_release() */
- layoutreturn = pnfs_prepare_layoutreturn(lo, &stateid, &iomode);
+ layoutreturn = pnfs_prepare_layoutreturn(lo, &stateid, &lc_cred, &iomode);
/* If the creds don't match, we can't compound the layoutreturn */
- if (!layoutreturn || cred != lo->plh_lc_cred)
+ if (!layoutreturn || cred_fscmp(cred, lc_cred) != 0)
goto out_noroc;
roc = layoutreturn;
pnfs_init_layoutreturn_args(args, lo, &stateid, iomode);
res->lrs_present = 0;
layoutreturn = false;
+ put_cred(lc_cred);
out_noroc:
spin_unlock(&ino->i_lock);
@@ -1422,31 +1635,99 @@ out_noroc:
return true;
}
if (layoutreturn)
- pnfs_send_layoutreturn(lo, &stateid, iomode, true);
+ pnfs_send_layoutreturn(lo, &stateid, &lc_cred, iomode, 0);
pnfs_put_layout_hdr(lo);
return false;
}
+int pnfs_roc_done(struct rpc_task *task, struct nfs4_layoutreturn_args **argpp,
+ struct nfs4_layoutreturn_res **respp, int *ret)
+{
+ struct nfs4_layoutreturn_args *arg = *argpp;
+ int retval = -EAGAIN;
+
+ if (!arg)
+ return 0;
+ /* Handle Layoutreturn errors */
+ switch (*ret) {
+ case 0:
+ retval = 0;
+ break;
+ case -NFS4ERR_NOMATCHING_LAYOUT:
+ /* Was there an RPC level error? If not, retry */
+ if (task->tk_rpc_status == 0)
+ break;
+ /*
+ * Is there a fatal network level error?
+ * If so release the layout, but flag the error.
+ */
+ if ((task->tk_rpc_status == -ENETDOWN ||
+ task->tk_rpc_status == -ENETUNREACH) &&
+ task->tk_flags & RPC_TASK_NETUNREACH_FATAL) {
+ *ret = 0;
+ (*respp)->lrs_present = 0;
+ retval = -EIO;
+ break;
+ }
+ /* If the call was not sent, let caller handle it */
+ if (!RPC_WAS_SENT(task))
+ return 0;
+ /*
+ * Otherwise, assume the call succeeded and
+ * that we need to release the layout
+ */
+ *ret = 0;
+ (*respp)->lrs_present = 0;
+ retval = 0;
+ break;
+ case -NFS4ERR_DELAY:
+ /* Let the caller handle the retry */
+ *ret = -NFS4ERR_NOMATCHING_LAYOUT;
+ return 0;
+ case -NFS4ERR_OLD_STATEID:
+ if (!nfs4_layout_refresh_old_stateid(&arg->stateid,
+ &arg->range, arg->inode))
+ break;
+ *ret = -NFS4ERR_NOMATCHING_LAYOUT;
+ return -EAGAIN;
+ }
+ *argpp = NULL;
+ *respp = NULL;
+ return retval;
+}
+
void pnfs_roc_release(struct nfs4_layoutreturn_args *args,
- struct nfs4_layoutreturn_res *res,
- int ret)
+ struct nfs4_layoutreturn_res *res, int ret)
{
struct pnfs_layout_hdr *lo = args->layout;
- const nfs4_stateid *arg_stateid = NULL;
+ struct inode *inode = args->inode;
const nfs4_stateid *res_stateid = NULL;
struct nfs4_xdr_opaque_data *ld_private = args->ld_private;
+ LIST_HEAD(freeme);
- if (ret == 0) {
- arg_stateid = &args->stateid;
+ switch (ret) {
+ case -NFS4ERR_BADSESSION:
+ case -NFS4ERR_DEADSESSION:
+ case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
+ case -NFS4ERR_NOMATCHING_LAYOUT:
+ spin_lock(&inode->i_lock);
+ pnfs_layoutreturn_retry_later_locked(lo, &args->stateid,
+ &args->range, &freeme);
+ spin_unlock(&inode->i_lock);
+ pnfs_free_lseg_list(&freeme);
+ break;
+ case 0:
if (res->lrs_present)
res_stateid = &res->stateid;
+ fallthrough;
+ default:
+ pnfs_layoutreturn_free_lsegs(lo, &args->stateid, &args->range,
+ res_stateid);
}
- pnfs_layoutreturn_free_lsegs(lo, arg_stateid, &args->range,
- res_stateid);
+ trace_nfs4_layoutreturn_on_close(args->inode, &args->stateid, ret);
if (ld_private && ld_private->ops && ld_private->ops->free)
ld_private->ops->free(ld_private);
pnfs_put_layout_hdr(lo);
- trace_nfs4_layoutreturn_on_close(args->inode, 0);
}
bool pnfs_wait_on_layoutreturn(struct inode *ino, struct rpc_task *task)
@@ -1759,7 +2040,7 @@ static int pnfs_prepare_to_retry_layoutget(struct pnfs_layout_hdr *lo)
pnfs_layoutcommit_inode(lo->plh_inode, false);
return wait_on_bit_action(&lo->plh_flags, NFS_LAYOUT_RETURN,
nfs_wait_bit_killable,
- TASK_KILLABLE);
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
}
static void nfs_layoutget_begin(struct pnfs_layout_hdr *lo)
@@ -1769,8 +2050,16 @@ static void nfs_layoutget_begin(struct pnfs_layout_hdr *lo)
static void nfs_layoutget_end(struct pnfs_layout_hdr *lo)
{
- if (atomic_dec_and_test(&lo->plh_outstanding))
- wake_up_var(&lo->plh_outstanding);
+ if (atomic_dec_and_test(&lo->plh_outstanding) &&
+ test_and_clear_bit(NFS_LAYOUT_DRAIN, &lo->plh_flags)) {
+ smp_mb__after_atomic();
+ wake_up_bit(&lo->plh_flags, NFS_LAYOUT_DRAIN);
+ }
+}
+
+static bool pnfs_is_first_layoutget(struct pnfs_layout_hdr *lo)
+{
+ return test_bit(NFS_LAYOUT_FIRST_LAYOUTGET, &lo->plh_flags);
}
static void pnfs_clear_first_layoutget(struct pnfs_layout_hdr *lo)
@@ -1785,15 +2074,14 @@ static void pnfs_clear_first_layoutget(struct pnfs_layout_hdr *lo)
static void _add_to_server_list(struct pnfs_layout_hdr *lo,
struct nfs_server *server)
{
- if (list_empty(&lo->plh_layouts)) {
+ if (!test_and_set_bit(NFS_LAYOUT_HASHED, &lo->plh_flags)) {
struct nfs_client *clp = server->nfs_client;
/* The lo must be on the clp list if there is any
* chance of a CB_LAYOUTRECALL(FILE) coming in.
*/
spin_lock(&clp->cl_lock);
- if (list_empty(&lo->plh_layouts))
- list_add_tail(&lo->plh_layouts, &server->layouts);
+ list_add_tail_rcu(&lo->plh_layouts, &server->layouts);
spin_unlock(&clp->cl_lock);
}
}
@@ -1823,7 +2111,9 @@ pnfs_update_layout(struct inode *ino,
struct pnfs_layout_segment *lseg = NULL;
struct nfs4_layoutget *lgp;
nfs4_stateid stateid;
- long timeout = 0;
+ struct nfs4_exception exception = {
+ .inode = ino,
+ };
unsigned long giveup = jiffies + (clp->cl_lease_time << 1);
bool first;
@@ -1840,6 +2130,14 @@ pnfs_update_layout(struct inode *ino,
}
lookup_again:
+ if (!nfs4_valid_open_stateid(ctx->state)) {
+ trace_pnfs_update_layout(ino, pos, count,
+ iomode, lo, lseg,
+ PNFS_UPDATE_LAYOUT_INVALID_OPEN);
+ lseg = ERR_PTR(-EIO);
+ goto out;
+ }
+
lseg = ERR_PTR(nfs4_client_recover_expired_lease(clp));
if (IS_ERR(lseg))
goto out;
@@ -1848,6 +2146,7 @@ lookup_again:
lo = pnfs_find_alloc_layout(ino, ctx, gfp_flags);
if (lo == NULL) {
spin_unlock(&ino->i_lock);
+ lseg = ERR_PTR(-ENOMEM);
trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg,
PNFS_UPDATE_LAYOUT_NOMEM);
goto out;
@@ -1872,27 +2171,42 @@ lookup_again:
* If the layout segment list is empty, but there are outstanding
* layoutget calls, then they might be subject to a layoutrecall.
*/
- if (list_empty(&lo->plh_segs) &&
+ if (test_bit(NFS_LAYOUT_DRAIN, &lo->plh_flags) &&
atomic_read(&lo->plh_outstanding) != 0) {
spin_unlock(&ino->i_lock);
- lseg = ERR_PTR(wait_var_event_killable(&lo->plh_outstanding,
- atomic_read(&lo->plh_outstanding)));
- if (IS_ERR(lseg) || !list_empty(&lo->plh_segs))
+ lseg = ERR_PTR(wait_on_bit(&lo->plh_flags, NFS_LAYOUT_DRAIN,
+ TASK_KILLABLE));
+ if (IS_ERR(lseg))
goto out_put_layout_hdr;
pnfs_put_layout_hdr(lo);
goto lookup_again;
}
- lseg = pnfs_find_lseg(lo, &arg, strict_iomode);
- if (lseg) {
+ /*
+ * Because we free lsegs when sending LAYOUTRETURN, we need to wait
+ * for LAYOUTRETURN.
+ */
+ if (test_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) {
+ spin_unlock(&ino->i_lock);
+ dprintk("%s wait for layoutreturn\n", __func__);
+ lseg = ERR_PTR(pnfs_prepare_to_retry_layoutget(lo));
+ if (!IS_ERR(lseg)) {
+ pnfs_put_layout_hdr(lo);
+ dprintk("%s retrying\n", __func__);
+ trace_pnfs_update_layout(ino, pos, count, iomode, lo,
+ lseg,
+ PNFS_UPDATE_LAYOUT_RETRY);
+ goto lookup_again;
+ }
trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg,
- PNFS_UPDATE_LAYOUT_FOUND_CACHED);
- goto out_unlock;
+ PNFS_UPDATE_LAYOUT_RETURN);
+ goto out_put_layout_hdr;
}
- if (!nfs4_valid_open_stateid(ctx->state)) {
+ lseg = pnfs_find_lseg(lo, &arg, strict_iomode);
+ if (lseg) {
trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg,
- PNFS_UPDATE_LAYOUT_INVALID_OPEN);
+ PNFS_UPDATE_LAYOUT_FOUND_CACHED);
goto out_unlock;
}
@@ -1902,6 +2216,7 @@ lookup_again:
* stateid.
*/
if (test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags)) {
+ int status;
/*
* The first layoutget for the file. Need to serialize per
@@ -1920,39 +2235,24 @@ lookup_again:
goto lookup_again;
}
+ spin_unlock(&ino->i_lock);
first = true;
- if (nfs4_select_rw_stateid(ctx->state,
+ status = nfs4_select_rw_stateid(ctx->state,
iomode == IOMODE_RW ? FMODE_WRITE : FMODE_READ,
- NULL, &stateid, NULL) != 0) {
+ NULL, &stateid, NULL);
+ if (status != 0) {
+ lseg = ERR_PTR(status);
trace_pnfs_update_layout(ino, pos, count,
iomode, lo, lseg,
PNFS_UPDATE_LAYOUT_INVALID_OPEN);
- goto out_unlock;
- }
- } else {
- nfs4_stateid_copy(&stateid, &lo->plh_stateid);
- }
-
- /*
- * Because we free lsegs before sending LAYOUTRETURN, we need to wait
- * for LAYOUTRETURN even if first is true.
- */
- if (test_bit(NFS_LAYOUT_RETURN, &lo->plh_flags)) {
- spin_unlock(&ino->i_lock);
- dprintk("%s wait for layoutreturn\n", __func__);
- lseg = ERR_PTR(pnfs_prepare_to_retry_layoutget(lo));
- if (!IS_ERR(lseg)) {
- if (first)
- pnfs_clear_first_layoutget(lo);
+ nfs4_schedule_stateid_recovery(server, ctx->state);
+ pnfs_clear_first_layoutget(lo);
pnfs_put_layout_hdr(lo);
- dprintk("%s retrying\n", __func__);
- trace_pnfs_update_layout(ino, pos, count, iomode, lo,
- lseg, PNFS_UPDATE_LAYOUT_RETRY);
goto lookup_again;
}
- trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg,
- PNFS_UPDATE_LAYOUT_RETURN);
- goto out_put_layout_hdr;
+ spin_lock(&ino->i_lock);
+ } else {
+ nfs4_stateid_copy(&stateid, &lo->plh_stateid);
}
if (pnfs_layoutgets_blocked(lo)) {
@@ -1975,13 +2275,17 @@ lookup_again:
lgp = pnfs_alloc_init_layoutget_args(ino, ctx, &stateid, &arg, gfp_flags);
if (!lgp) {
+ lseg = ERR_PTR(-ENOMEM);
trace_pnfs_update_layout(ino, pos, count, iomode, lo, NULL,
PNFS_UPDATE_LAYOUT_NOMEM);
nfs_layoutget_end(lo);
goto out_put_layout_hdr;
}
- lseg = nfs4_proc_layoutget(lgp, &timeout);
+ lgp->lo = lo;
+ pnfs_get_layout_hdr(lo);
+
+ lseg = nfs4_proc_layoutget(lgp, &exception);
trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg,
PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET);
nfs_layoutget_end(lo);
@@ -1994,6 +2298,12 @@ lookup_again:
case -ERECALLCONFLICT:
case -EAGAIN:
break;
+ case -ENODATA:
+ /* The server returned NFS4ERR_LAYOUTUNAVAILABLE */
+ pnfs_layout_set_fail_bit(
+ lo, pnfs_iomode_to_fail_bit(iomode));
+ lseg = NULL;
+ goto out_put_layout_hdr;
default:
if (!nfs_error_is_fatal(PTR_ERR(lseg))) {
pnfs_layout_clear_fail_bit(lo, pnfs_iomode_to_fail_bit(iomode));
@@ -2002,6 +2312,8 @@ lookup_again:
goto out_put_layout_hdr;
}
if (lseg) {
+ if (!exception.retry)
+ goto out_put_layout_hdr;
if (first)
pnfs_clear_first_layoutget(lo);
trace_pnfs_update_layout(ino, pos, count,
@@ -2016,6 +2328,8 @@ lookup_again:
out_put_layout_hdr:
if (first)
pnfs_clear_first_layoutget(lo);
+ trace_pnfs_update_layout(ino, pos, count, iomode, lo, lseg,
+ PNFS_UPDATE_LAYOUT_EXIT);
pnfs_put_layout_hdr(lo);
out:
dprintk("%s: inode %s/%llu pNFS layout segment %s for "
@@ -2059,7 +2373,7 @@ _pnfs_grab_empty_layout(struct inode *ino, struct nfs_open_context *ctx)
struct pnfs_layout_hdr *lo;
spin_lock(&ino->i_lock);
- lo = pnfs_find_alloc_layout(ino, ctx, GFP_KERNEL);
+ lo = pnfs_find_alloc_layout(ino, ctx, nfs_io_gfp_mask());
if (!lo)
goto out_unlock;
if (!test_bit(NFS_LAYOUT_INVALID_STID, &lo->plh_flags))
@@ -2081,8 +2395,6 @@ out_unlock:
return NULL;
}
-extern const nfs4_stateid current_stateid;
-
static void _lgopen_prepare_attached(struct nfs4_opendata *data,
struct nfs_open_context *ctx)
{
@@ -2104,13 +2416,15 @@ static void _lgopen_prepare_attached(struct nfs4_opendata *data,
lo = _pnfs_grab_empty_layout(ino, ctx);
if (!lo)
return;
- lgp = pnfs_alloc_init_layoutget_args(ino, ctx, &current_stateid,
- &rng, GFP_KERNEL);
+ lgp = pnfs_alloc_init_layoutget_args(ino, ctx, &current_stateid, &rng,
+ nfs_io_gfp_mask());
if (!lgp) {
pnfs_clear_first_layoutget(lo);
+ nfs_layoutget_end(lo);
pnfs_put_layout_hdr(lo);
return;
}
+ lgp->lo = lo;
data->lgp = lgp;
data->o_arg.lg_args = &lgp->args;
data->o_res.lg_res = &lgp->res;
@@ -2119,6 +2433,7 @@ static void _lgopen_prepare_attached(struct nfs4_opendata *data,
static void _lgopen_prepare_floating(struct nfs4_opendata *data,
struct nfs_open_context *ctx)
{
+ struct inode *ino = data->dentry->d_inode;
struct pnfs_layout_range rng = {
.iomode = (data->o_arg.fmode & FMODE_WRITE) ?
IOMODE_RW: IOMODE_READ,
@@ -2127,8 +2442,8 @@ static void _lgopen_prepare_floating(struct nfs4_opendata *data,
};
struct nfs4_layoutget *lgp;
- lgp = pnfs_alloc_init_layoutget_args(NULL, ctx, &current_stateid,
- &rng, GFP_KERNEL);
+ lgp = pnfs_alloc_init_layoutget_args(ino, ctx, &current_stateid, &rng,
+ nfs_io_gfp_mask());
if (!lgp)
return;
data->lgp = lgp;
@@ -2147,6 +2462,8 @@ void pnfs_lgopen_prepare(struct nfs4_opendata *data,
/* Could check on max_ops, but currently hardcoded high enough */
if (!nfs_server_capable(data->dir->d_inode, NFS_CAP_LGOPEN))
return;
+ if (data->lgp)
+ return;
if (data->state)
_lgopen_prepare_attached(data, ctx);
else
@@ -2186,13 +2503,13 @@ void pnfs_parse_lgopen(struct inode *ino, struct nfs4_layoutget *lgp,
}
return;
}
- if (!lgp->args.inode) {
+ if (!lgp->lo) {
lo = _pnfs_grab_empty_layout(ino, ctx);
if (!lo)
return;
- lgp->args.inode = ino;
+ lgp->lo = lo;
} else
- lo = NFS_I(lgp->args.inode)->layout;
+ lo = lgp->lo;
lseg = pnfs_layout_process(lgp);
if (!IS_ERR(lseg)) {
@@ -2205,11 +2522,9 @@ void pnfs_parse_lgopen(struct inode *ino, struct nfs4_layoutget *lgp,
void nfs4_lgopen_release(struct nfs4_layoutget *lgp)
{
if (lgp != NULL) {
- struct inode *inode = lgp->args.inode;
- if (inode) {
- struct pnfs_layout_hdr *lo = NFS_I(inode)->layout;
- pnfs_clear_first_layoutget(lo);
- nfs_layoutget_end(lo);
+ if (lgp->lo) {
+ pnfs_clear_first_layoutget(lgp->lo);
+ nfs_layoutget_end(lgp->lo);
}
pnfs_layoutget_free(lgp);
}
@@ -2218,7 +2533,7 @@ void nfs4_lgopen_release(struct nfs4_layoutget *lgp)
struct pnfs_layout_segment *
pnfs_layout_process(struct nfs4_layoutget *lgp)
{
- struct pnfs_layout_hdr *lo = NFS_I(lgp->args.inode)->layout;
+ struct pnfs_layout_hdr *lo = lgp->lo;
struct nfs4_layoutget_res *res = &lgp->res;
struct pnfs_layout_segment *lseg;
struct inode *ino = lo->plh_inode;
@@ -2246,23 +2561,33 @@ pnfs_layout_process(struct nfs4_layoutget *lgp)
goto out_forget;
}
- if (!pnfs_layout_is_valid(lo)) {
- /* We have a completely new layout */
- pnfs_set_layout_stateid(lo, &res->stateid, true);
- } else if (nfs4_stateid_match_other(&lo->plh_stateid, &res->stateid)) {
+ if (test_bit(NFS_LAYOUT_DRAIN, &lo->plh_flags) &&
+ !pnfs_is_first_layoutget(lo))
+ goto out_forget;
+
+ if (nfs4_stateid_match_other(&lo->plh_stateid, &res->stateid)) {
/* existing state ID, make sure the sequence number matches. */
if (pnfs_layout_stateid_blocked(lo, &res->stateid)) {
+ if (!pnfs_layout_is_valid(lo))
+ lo->plh_barrier = 0;
dprintk("%s forget reply due to sequence\n", __func__);
goto out_forget;
}
- pnfs_set_layout_stateid(lo, &res->stateid, false);
- } else {
+ pnfs_set_layout_stateid(lo, &res->stateid, lgp->cred, false);
+ } else if (pnfs_layout_is_valid(lo)) {
/*
* We got an entirely new state ID. Mark all segments for the
* inode invalid, and retry the layoutget
*/
- pnfs_mark_layout_stateid_invalid(lo, &free_me);
+ struct pnfs_layout_range range = {
+ .iomode = IOMODE_ANY,
+ .length = NFS4_MAX_UINT64,
+ };
+ pnfs_mark_matching_lsegs_return(lo, &free_me, &range, 0);
goto out_forget;
+ } else {
+ /* We have a completely new layout */
+ pnfs_set_layout_stateid(lo, &res->stateid, lgp->cred, true);
}
pnfs_get_lseg(lseg);
@@ -2283,16 +2608,6 @@ out_forget:
return ERR_PTR(-EAGAIN);
}
-static int
-mark_lseg_invalid_or_return(struct pnfs_layout_segment *lseg,
- struct list_head *tmp_list)
-{
- if (!mark_lseg_invalid(lseg, tmp_list))
- return 0;
- pnfs_cache_lseg_for_layoutreturn(lseg->pls_layout, lseg);
- return 1;
-}
-
/**
* pnfs_mark_matching_lsegs_return - Free or return matching layout segments
* @lo: pointer to layout header
@@ -2316,12 +2631,16 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo,
u32 seq)
{
struct pnfs_layout_segment *lseg, *next;
+ struct nfs_server *server = NFS_SERVER(lo->plh_inode);
int remaining = 0;
dprintk("%s:Begin lo %p\n", __func__, lo);
assert_spin_locked(&lo->plh_inode->i_lock);
+ if (test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags))
+ tmp_list = &lo->plh_return_segs;
+
list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list)
if (pnfs_match_lseg_recall(lseg, return_range, seq)) {
dprintk("%s: marking lseg %p iomode %d "
@@ -2329,10 +2648,13 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo,
lseg, lseg->pls_range.iomode,
lseg->pls_range.offset,
lseg->pls_range.length);
- if (mark_lseg_invalid_or_return(lseg, tmp_list))
+ if (test_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags))
+ tmp_list = &lo->plh_return_segs;
+ if (mark_lseg_invalid(lseg, tmp_list))
continue;
remaining++;
set_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags);
+ pnfs_lseg_cancel_io(server, lseg);
}
if (remaining) {
@@ -2348,87 +2670,208 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo,
return -ENOENT;
}
-void pnfs_error_mark_layout_for_return(struct inode *inode,
- struct pnfs_layout_segment *lseg)
+static void
+pnfs_mark_layout_for_return(struct inode *inode,
+ const struct pnfs_layout_range *range)
{
- struct pnfs_layout_hdr *lo = NFS_I(inode)->layout;
- struct pnfs_layout_range range = {
- .iomode = lseg->pls_range.iomode,
- .offset = 0,
- .length = NFS4_MAX_UINT64,
- };
+ struct pnfs_layout_hdr *lo;
bool return_now = false;
spin_lock(&inode->i_lock);
+ lo = NFS_I(inode)->layout;
if (!pnfs_layout_is_valid(lo)) {
spin_unlock(&inode->i_lock);
return;
}
- pnfs_set_plh_return_info(lo, range.iomode, 0);
+ pnfs_set_plh_return_info(lo, range->iomode, 0);
/*
* mark all matching lsegs so that we are sure to have no live
* segments at hand when sending layoutreturn. See pnfs_put_lseg()
* for how it works.
*/
- if (pnfs_mark_matching_lsegs_return(lo, &lo->plh_return_segs, &range, 0) != -EBUSY) {
+ if (pnfs_mark_matching_lsegs_return(lo, &lo->plh_return_segs, range, 0) != -EBUSY) {
+ const struct cred *cred;
nfs4_stateid stateid;
enum pnfs_iomode iomode;
- return_now = pnfs_prepare_layoutreturn(lo, &stateid, &iomode);
+ return_now = pnfs_prepare_layoutreturn(lo, &stateid, &cred, &iomode);
spin_unlock(&inode->i_lock);
if (return_now)
- pnfs_send_layoutreturn(lo, &stateid, iomode, false);
+ pnfs_send_layoutreturn(lo, &stateid, &cred, iomode,
+ PNFS_FL_LAYOUTRETURN_ASYNC);
} else {
spin_unlock(&inode->i_lock);
nfs_commit_inode(inode, 0);
}
}
+
+void pnfs_error_mark_layout_for_return(struct inode *inode,
+ struct pnfs_layout_segment *lseg)
+{
+ struct pnfs_layout_range range = {
+ .iomode = lseg->pls_range.iomode,
+ .offset = 0,
+ .length = NFS4_MAX_UINT64,
+ };
+
+ pnfs_mark_layout_for_return(inode, &range);
+}
EXPORT_SYMBOL_GPL(pnfs_error_mark_layout_for_return);
+static bool
+pnfs_layout_can_be_returned(struct pnfs_layout_hdr *lo)
+{
+ return pnfs_layout_is_valid(lo) &&
+ !test_bit(NFS_LAYOUT_INODE_FREEING, &lo->plh_flags) &&
+ !test_bit(NFS_LAYOUT_RETURN, &lo->plh_flags);
+}
+
+static struct pnfs_layout_segment *
+pnfs_find_first_lseg(struct pnfs_layout_hdr *lo,
+ const struct pnfs_layout_range *range,
+ enum pnfs_iomode iomode)
+{
+ struct pnfs_layout_segment *lseg;
+
+ list_for_each_entry(lseg, &lo->plh_segs, pls_list) {
+ if (!test_bit(NFS_LSEG_VALID, &lseg->pls_flags))
+ continue;
+ if (test_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags))
+ continue;
+ if (lseg->pls_range.iomode != iomode && iomode != IOMODE_ANY)
+ continue;
+ if (pnfs_lseg_range_intersecting(&lseg->pls_range, range))
+ return lseg;
+ }
+ return NULL;
+}
+
+/* Find open file states whose mode matches that of the range */
+static bool
+pnfs_should_return_unused_layout(struct pnfs_layout_hdr *lo,
+ const struct pnfs_layout_range *range)
+{
+ struct list_head *head;
+ struct nfs_open_context *ctx;
+ fmode_t mode = 0;
+
+ if (!pnfs_layout_can_be_returned(lo) ||
+ !pnfs_find_first_lseg(lo, range, range->iomode))
+ return false;
+
+ head = &NFS_I(lo->plh_inode)->open_files;
+ list_for_each_entry_rcu(ctx, head, list) {
+ if (ctx->state)
+ mode |= ctx->state->state & (FMODE_READ|FMODE_WRITE);
+ }
+
+ switch (range->iomode) {
+ default:
+ break;
+ case IOMODE_READ:
+ mode &= ~FMODE_WRITE;
+ break;
+ case IOMODE_RW:
+ if (pnfs_find_first_lseg(lo, range, IOMODE_READ))
+ mode &= ~FMODE_READ;
+ }
+ return mode == 0;
+}
+
+static int pnfs_layout_return_unused_byserver(struct nfs_server *server,
+ void *data)
+{
+ const struct pnfs_layout_range *range = data;
+ const struct cred *cred;
+ struct pnfs_layout_hdr *lo;
+ struct inode *inode;
+ nfs4_stateid stateid;
+ enum pnfs_iomode iomode;
+
+restart:
+ rcu_read_lock();
+ list_for_each_entry_rcu(lo, &server->layouts, plh_layouts) {
+ inode = lo->plh_inode;
+ if (!inode || !pnfs_layout_can_be_returned(lo) ||
+ test_bit(NFS_LAYOUT_RETURN_REQUESTED, &lo->plh_flags))
+ continue;
+ spin_lock(&inode->i_lock);
+ if (!lo->plh_inode ||
+ !pnfs_should_return_unused_layout(lo, range)) {
+ spin_unlock(&inode->i_lock);
+ continue;
+ }
+ pnfs_get_layout_hdr(lo);
+ pnfs_set_plh_return_info(lo, range->iomode, 0);
+ if (pnfs_mark_matching_lsegs_return(lo, &lo->plh_return_segs,
+ range, 0) != 0 ||
+ !pnfs_prepare_layoutreturn(lo, &stateid, &cred, &iomode)) {
+ spin_unlock(&inode->i_lock);
+ rcu_read_unlock();
+ pnfs_put_layout_hdr(lo);
+ cond_resched();
+ goto restart;
+ }
+ spin_unlock(&inode->i_lock);
+ rcu_read_unlock();
+ pnfs_send_layoutreturn(lo, &stateid, &cred, iomode,
+ PNFS_FL_LAYOUTRETURN_ASYNC);
+ pnfs_put_layout_hdr(lo);
+ cond_resched();
+ goto restart;
+ }
+ rcu_read_unlock();
+ return 0;
+}
+
void
-pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio)
+pnfs_layout_return_unused_byclid(struct nfs_client *clp,
+ enum pnfs_iomode iomode)
+{
+ struct pnfs_layout_range range = {
+ .iomode = iomode,
+ .offset = 0,
+ .length = NFS4_MAX_UINT64,
+ };
+
+ nfs_client_for_each_server(clp, pnfs_layout_return_unused_byserver,
+ &range);
+}
+
+/* Check if we have we have a valid layout but if there isn't an intersection
+ * between the request and the pgio->pg_lseg, put this pgio->pg_lseg away.
+ */
+void
+pnfs_generic_pg_check_layout(struct nfs_pageio_descriptor *pgio,
+ struct nfs_page *req)
{
if (pgio->pg_lseg == NULL ||
- test_bit(NFS_LSEG_VALID, &pgio->pg_lseg->pls_flags))
+ (test_bit(NFS_LSEG_VALID, &pgio->pg_lseg->pls_flags) &&
+ pnfs_lseg_request_intersecting(pgio->pg_lseg, req)))
return;
pnfs_put_lseg(pgio->pg_lseg);
pgio->pg_lseg = NULL;
}
EXPORT_SYMBOL_GPL(pnfs_generic_pg_check_layout);
-/*
- * Check for any intersection between the request and the pgio->pg_lseg,
- * and if none, put this pgio->pg_lseg away.
- */
-static void
-pnfs_generic_pg_check_range(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
-{
- if (pgio->pg_lseg && !pnfs_lseg_request_intersecting(pgio->pg_lseg, req)) {
- pnfs_put_lseg(pgio->pg_lseg);
- pgio->pg_lseg = NULL;
- }
-}
-
void
pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
{
- u64 rd_size = req->wb_bytes;
+ u64 rd_size;
- pnfs_generic_pg_check_layout(pgio);
- pnfs_generic_pg_check_range(pgio, req);
+ pnfs_generic_pg_check_layout(pgio, req);
if (pgio->pg_lseg == NULL) {
if (pgio->pg_dreq == NULL)
rd_size = i_size_read(pgio->pg_inode) - req_offset(req);
else
- rd_size = nfs_dreq_bytes_left(pgio->pg_dreq);
-
- pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
- req->wb_context,
- req_offset(req),
- rd_size,
- IOMODE_READ,
- false,
- GFP_KERNEL);
+ rd_size = nfs_dreq_bytes_left(pgio->pg_dreq,
+ req_offset(req));
+
+ pgio->pg_lseg =
+ pnfs_update_layout(pgio->pg_inode, nfs_req_openctx(req),
+ req_offset(req), rd_size,
+ IOMODE_READ, false,
+ nfs_io_gfp_mask());
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
@@ -2446,16 +2889,12 @@ void
pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,
struct nfs_page *req, u64 wb_size)
{
- pnfs_generic_pg_check_layout(pgio);
- pnfs_generic_pg_check_range(pgio, req);
+ pnfs_generic_pg_check_layout(pgio, req);
if (pgio->pg_lseg == NULL) {
- pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
- req->wb_context,
- req_offset(req),
- wb_size,
- IOMODE_RW,
- false,
- GFP_NOFS);
+ pgio->pg_lseg =
+ pnfs_update_layout(pgio->pg_inode, nfs_req_openctx(req),
+ req_offset(req), wb_size, IOMODE_RW,
+ false, nfs_io_gfp_mask());
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
@@ -2531,7 +2970,6 @@ int pnfs_write_done_resend_to_mds(struct nfs_pgio_header *hdr)
/* Resend all requests through the MDS */
nfs_pageio_init_write(&pgio, hdr->inode, FLUSH_STABLE, true,
hdr->completion_ops);
- set_bit(NFS_CONTEXT_RESEND_WRITES, &hdr->args.context->flags);
return nfs_pageio_resend(&pgio, hdr);
}
EXPORT_SYMBOL_GPL(pnfs_write_done_resend_to_mds);
@@ -2612,6 +3050,7 @@ pnfs_do_write(struct nfs_pageio_descriptor *desc,
switch (trypnfs) {
case PNFS_NOT_ATTEMPTED:
pnfs_write_through_mds(desc, hdr);
+ break;
case PNFS_ATTEMPTED:
break;
case PNFS_TRY_AGAIN:
@@ -2727,7 +3166,8 @@ pnfs_try_to_read_data(struct nfs_pgio_header *hdr,
}
/* Resend all requests through pnfs. */
-void pnfs_read_resend_pnfs(struct nfs_pgio_header *hdr)
+void pnfs_read_resend_pnfs(struct nfs_pgio_header *hdr,
+ unsigned int mirror_idx)
{
struct nfs_pageio_descriptor pgio;
@@ -2738,6 +3178,7 @@ void pnfs_read_resend_pnfs(struct nfs_pgio_header *hdr)
nfs_pageio_init_read(&pgio, hdr->inode, false,
hdr->completion_ops);
+ pgio.pg_mirror_idx = mirror_idx;
hdr->task.tk_status = nfs_pageio_resend(&pgio, hdr);
}
}
@@ -2754,6 +3195,7 @@ pnfs_do_read(struct nfs_pageio_descriptor *desc, struct nfs_pgio_header *hdr)
switch (trypnfs) {
case PNFS_NOT_ATTEMPTED:
pnfs_read_through_mds(desc, hdr);
+ break;
case PNFS_ATTEMPTED:
break;
case PNFS_TRY_AGAIN:
@@ -2890,6 +3332,7 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
struct nfs_inode *nfsi = NFS_I(inode);
loff_t end_pos;
int status;
+ bool mark_as_dirty = false;
if (!pnfs_layoutcommit_outstanding(inode))
return 0;
@@ -2903,14 +3346,14 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
status = wait_on_bit_lock_action(&nfsi->flags,
NFS_INO_LAYOUTCOMMITTING,
nfs_wait_bit_killable,
- TASK_KILLABLE);
+ TASK_KILLABLE|TASK_FREEZABLE_UNSAFE);
if (status)
goto out;
}
status = -ENOMEM;
/* Note kzalloc ensures data->res.seq_res.sr_slot == NULL */
- data = kzalloc(sizeof(*data), GFP_NOFS);
+ data = kzalloc(sizeof(*data), nfs_io_gfp_mask());
if (!data)
goto clear_layoutcommitting;
@@ -2925,10 +3368,10 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
end_pos = nfsi->layout->plh_lwb;
nfs4_stateid_copy(&data->args.stateid, &nfsi->layout->plh_stateid);
+ data->cred = get_cred(nfsi->layout->plh_lc_cred);
spin_unlock(&inode->i_lock);
data->args.inode = inode;
- data->cred = get_cred(nfsi->layout->plh_lc_cred);
nfs_fattr_init(&data->fattr);
data->args.bitmask = NFS_SERVER(inode)->cache_consistency_bitmask;
data->res.fattr = &data->fattr;
@@ -2941,19 +3384,23 @@ pnfs_layoutcommit_inode(struct inode *inode, bool sync)
if (ld->prepare_layoutcommit) {
status = ld->prepare_layoutcommit(&data->args);
if (status) {
- put_cred(data->cred);
+ if (status != -ENOSPC)
+ put_cred(data->cred);
spin_lock(&inode->i_lock);
set_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags);
if (end_pos > nfsi->layout->plh_lwb)
nfsi->layout->plh_lwb = end_pos;
- goto out_unlock;
+ if (status != -ENOSPC)
+ goto out_unlock;
+ spin_unlock(&inode->i_lock);
+ mark_as_dirty = true;
}
}
status = nfs4_proc_layoutcommit(data, sync);
out:
- if (status)
+ if (status || mark_as_dirty)
mark_inode_dirty_sync(inode);
dprintk("<-- %s status %d\n", __func__, status);
return status;
@@ -2977,7 +3424,7 @@ struct nfs4_threshold *pnfs_mdsthreshold_alloc(void)
{
struct nfs4_threshold *thp;
- thp = kzalloc(sizeof(*thp), GFP_NOFS);
+ thp = kzalloc(sizeof(*thp), nfs_io_gfp_mask());
if (!thp) {
dprintk("%s mdsthreshold allocation failed\n", __func__);
return NULL;
generated by cgit (git 2.34.1) at 2025-12-24 21:12:13 +0000