static void
get_layout_hdr_locked(struct pnfs_layout_hdr *lo)
{
- assert_spin_locked(&lo->inode->i_lock);
- lo->refcount++;
+ assert_spin_locked(&lo->plh_inode->i_lock);
+ lo->plh_refcount++;
}
static void
put_layout_hdr_locked(struct pnfs_layout_hdr *lo)
{
- assert_spin_locked(&lo->inode->i_lock);
- BUG_ON(lo->refcount == 0);
+ assert_spin_locked(&lo->plh_inode->i_lock);
+ BUG_ON(lo->plh_refcount == 0);
- lo->refcount--;
- if (!lo->refcount) {
+ lo->plh_refcount--;
+ if (!lo->plh_refcount) {
dprintk("%s: freeing layout cache %p\n", __func__, lo);
- BUG_ON(!list_empty(&lo->layouts));
- NFS_I(lo->inode)->layout = NULL;
+ BUG_ON(!list_empty(&lo->plh_layouts));
+ NFS_I(lo->plh_inode)->layout = NULL;
kfree(lo);
}
}
static void
init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg)
{
- INIT_LIST_HEAD(&lseg->fi_list);
- kref_init(&lseg->kref);
- lseg->layout = lo;
+ INIT_LIST_HEAD(&lseg->pls_list);
+ atomic_set(&lseg->pls_refcount, 1);
+ smp_mb();
+ set_bit(NFS_LSEG_VALID, &lseg->pls_flags);
+ lseg->pls_layout = lo;
}
-/* Called without i_lock held, as the free_lseg call may sleep */
-static void
-destroy_lseg(struct kref *kref)
+static void free_lseg(struct pnfs_layout_segment *lseg)
{
- struct pnfs_layout_segment *lseg =
- container_of(kref, struct pnfs_layout_segment, kref);
- struct inode *ino = lseg->layout->inode;
+ struct inode *ino = lseg->pls_layout->plh_inode;
- dprintk("--> %s\n", __func__);
NFS_SERVER(ino)->pnfs_curr_ld->free_lseg(lseg);
- /* Matched by get_layout_hdr_locked in pnfs_insert_layout */
+ /* Matched by get_layout_hdr in pnfs_insert_layout */
put_layout_hdr(ino);
}
-static void
-put_lseg(struct pnfs_layout_segment *lseg)
+/* The use of tmp_list is necessary because pnfs_curr_ld->free_lseg
+ * could sleep, so must be called outside of the lock.
+ * Returns 1 if object was removed, otherwise return 0.
+ */
+static int
+put_lseg_locked(struct pnfs_layout_segment *lseg,
+ struct list_head *tmp_list)
+{
+ dprintk("%s: lseg %p ref %d valid %d\n", __func__, lseg,
+ atomic_read(&lseg->pls_refcount),
+ test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
+ if (atomic_dec_and_test(&lseg->pls_refcount)) {
+ struct inode *ino = lseg->pls_layout->plh_inode;
+
+ BUG_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
+ list_del(&lseg->pls_list);
+ if (list_empty(&lseg->pls_layout->plh_segs)) {
+ struct nfs_client *clp;
+
+ clp = NFS_SERVER(ino)->nfs_client;
+ spin_lock(&clp->cl_lock);
+ /* List does not take a reference, so no need for put here */
+ list_del_init(&lseg->pls_layout->plh_layouts);
+ spin_unlock(&clp->cl_lock);
+ }
+ list_add(&lseg->pls_list, tmp_list);
+ return 1;
+ }
+ return 0;
+}
+
+static bool
+should_free_lseg(u32 lseg_iomode, u32 recall_iomode)
{
- if (!lseg)
- return;
+ return (recall_iomode == IOMODE_ANY ||
+ lseg_iomode == recall_iomode);
+}
- dprintk("%s: lseg %p ref %d\n", __func__, lseg,
- atomic_read(&lseg->kref.refcount));
- kref_put(&lseg->kref, destroy_lseg);
+/* Returns 1 if lseg is removed from list, 0 otherwise */
+static int mark_lseg_invalid(struct pnfs_layout_segment *lseg,
+ struct list_head *tmp_list)
+{
+ int rv = 0;
+
+ if (test_and_clear_bit(NFS_LSEG_VALID, &lseg->pls_flags)) {
+ /* Remove the reference keeping the lseg in the
+ * list. It will now be removed when all
+ * outstanding io is finished.
+ */
+ rv = put_lseg_locked(lseg, tmp_list);
+ }
+ return rv;
}
-static void
-pnfs_clear_lseg_list(struct pnfs_layout_hdr *lo, struct list_head *tmp_list)
+/* Returns count of number of matching invalid lsegs remaining in list
+ * after call.
+ */
+static int
+mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
+ struct list_head *tmp_list,
+ u32 iomode)
{
struct pnfs_layout_segment *lseg, *next;
- struct nfs_client *clp;
+ int invalid = 0, removed = 0;
dprintk("%s:Begin lo %p\n", __func__, lo);
- assert_spin_locked(&lo->inode->i_lock);
- list_for_each_entry_safe(lseg, next, &lo->segs, fi_list) {
- dprintk("%s: freeing lseg %p\n", __func__, lseg);
- list_move(&lseg->fi_list, tmp_list);
- }
- clp = NFS_SERVER(lo->inode)->nfs_client;
- spin_lock(&clp->cl_lock);
- /* List does not take a reference, so no need for put here */
- list_del_init(&lo->layouts);
- spin_unlock(&clp->cl_lock);
- write_seqlock(&lo->seqlock);
- clear_bit(NFS_LAYOUT_STATEID_SET, &lo->state);
- write_sequnlock(&lo->seqlock);
-
- dprintk("%s:Return\n", __func__);
+ list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list)
+ if (should_free_lseg(lseg->pls_range.iomode, iomode)) {
+ dprintk("%s: freeing lseg %p iomode %d "
+ "offset %llu length %llu\n", __func__,
+ lseg, lseg->pls_range.iomode, lseg->pls_range.offset,
+ lseg->pls_range.length);
+ invalid++;
+ removed += mark_lseg_invalid(lseg, tmp_list);
+ }
+ dprintk("%s:Return %i\n", __func__, invalid - removed);
+ return invalid - removed;
}
static void
-pnfs_free_lseg_list(struct list_head *tmp_list)
+pnfs_free_lseg_list(struct list_head *free_me)
{
- struct pnfs_layout_segment *lseg;
+ struct pnfs_layout_segment *lseg, *tmp;
- while (!list_empty(tmp_list)) {
- lseg = list_entry(tmp_list->next, struct pnfs_layout_segment,
- fi_list);
- dprintk("%s calling put_lseg on %p\n", __func__, lseg);
- list_del(&lseg->fi_list);
- put_lseg(lseg);
+ list_for_each_entry_safe(lseg, tmp, free_me, pls_list) {
+ list_del(&lseg->pls_list);
+ free_lseg(lseg);
}
}
spin_lock(&nfsi->vfs_inode.i_lock);
lo = nfsi->layout;
if (lo) {
- pnfs_clear_lseg_list(lo, &tmp_list);
+ set_bit(NFS_LAYOUT_DESTROYED, &nfsi->layout->plh_flags);
+ mark_matching_lsegs_invalid(lo, &tmp_list, IOMODE_ANY);
/* Matched by refcount set to 1 in alloc_init_layout_hdr */
put_layout_hdr_locked(lo);
}
while (!list_empty(&tmp_list)) {
lo = list_entry(tmp_list.next, struct pnfs_layout_hdr,
- layouts);
+ plh_layouts);
dprintk("%s freeing layout for inode %lu\n", __func__,
- lo->inode->i_ino);
- pnfs_destroy_layout(NFS_I(lo->inode));
+ lo->plh_inode->i_ino);
+ pnfs_destroy_layout(NFS_I(lo->plh_inode));
}
}
-/* update lo->stateid with new if is more recent
- *
- * lo->stateid could be the open stateid, in which case we just use what given.
- */
+/* update lo->plh_stateid with new if is more recent */
static void
pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo,
const nfs4_stateid *new)
{
- nfs4_stateid *old = &lo->stateid;
- bool overwrite = false;
-
- write_seqlock(&lo->seqlock);
- if (!test_bit(NFS_LAYOUT_STATEID_SET, &lo->state) ||
- memcmp(old->stateid.other, new->stateid.other, sizeof(new->stateid.other)))
- overwrite = true;
- else {
- u32 oldseq, newseq;
+ u32 oldseq, newseq;
- oldseq = be32_to_cpu(old->stateid.seqid);
- newseq = be32_to_cpu(new->stateid.seqid);
- if ((int)(newseq - oldseq) > 0)
- overwrite = true;
- }
- if (overwrite)
- memcpy(&old->stateid, &new->stateid, sizeof(new->stateid));
- write_sequnlock(&lo->seqlock);
+ oldseq = be32_to_cpu(lo->plh_stateid.stateid.seqid);
+ newseq = be32_to_cpu(new->stateid.seqid);
+ if ((int)(newseq - oldseq) > 0)
+ memcpy(&lo->plh_stateid, &new->stateid, sizeof(new->stateid));
}
-static void
-pnfs_layout_from_open_stateid(struct pnfs_layout_hdr *lo,
- struct nfs4_state *state)
+/* lget is set to 1 if called from inside send_layoutget call chain */
+static bool
+pnfs_layoutgets_blocked(struct pnfs_layout_hdr *lo, int lget)
{
- int seq;
-
- dprintk("--> %s\n", __func__);
- write_seqlock(&lo->seqlock);
- do {
- seq = read_seqbegin(&state->seqlock);
- memcpy(lo->stateid.data, state->stateid.data,
- sizeof(state->stateid.data));
- } while (read_seqretry(&state->seqlock, seq));
- set_bit(NFS_LAYOUT_STATEID_SET, &lo->state);
- write_sequnlock(&lo->seqlock);
- dprintk("<-- %s\n", __func__);
+ return (list_empty(&lo->plh_segs) &&
+ (atomic_read(&lo->plh_outstanding) > lget));
}
-void
-pnfs_get_layout_stateid(nfs4_stateid *dst, struct pnfs_layout_hdr *lo,
- struct nfs4_state *open_state)
+int
+pnfs_choose_layoutget_stateid(nfs4_stateid *dst, struct pnfs_layout_hdr *lo,
+ struct nfs4_state *open_state)
{
- int seq;
+ int status = 0;
dprintk("--> %s\n", __func__);
- do {
- seq = read_seqbegin(&lo->seqlock);
- if (!test_bit(NFS_LAYOUT_STATEID_SET, &lo->state)) {
- /* This will trigger retry of the read */
- pnfs_layout_from_open_stateid(lo, open_state);
- } else
- memcpy(dst->data, lo->stateid.data,
- sizeof(lo->stateid.data));
- } while (read_seqretry(&lo->seqlock, seq));
+ spin_lock(&lo->plh_inode->i_lock);
+ if (pnfs_layoutgets_blocked(lo, 1)) {
+ status = -EAGAIN;
+ } else if (list_empty(&lo->plh_segs)) {
+ int seq;
+
+ do {
+ seq = read_seqbegin(&open_state->seqlock);
+ memcpy(dst->data, open_state->stateid.data,
+ sizeof(open_state->stateid.data));
+ } while (read_seqretry(&open_state->seqlock, seq));
+ } else
+ memcpy(dst->data, lo->plh_stateid.data, sizeof(lo->plh_stateid.data));
+ spin_unlock(&lo->plh_inode->i_lock);
dprintk("<-- %s\n", __func__);
+ return status;
}
/*
struct nfs_open_context *ctx,
u32 iomode)
{
- struct inode *ino = lo->inode;
+ struct inode *ino = lo->plh_inode;
struct nfs_server *server = NFS_SERVER(ino);
struct nfs4_layoutget *lgp;
struct pnfs_layout_segment *lseg = NULL;
BUG_ON(ctx == NULL);
lgp = kzalloc(sizeof(*lgp), GFP_KERNEL);
- if (lgp == NULL) {
- put_layout_hdr(lo->inode);
+ if (lgp == NULL)
return NULL;
- }
lgp->args.minlength = NFS4_MAX_UINT64;
lgp->args.maxcount = PNFS_LAYOUT_MAXSIZE;
lgp->args.range.iomode = iomode;
nfs4_proc_layoutget(lgp);
if (!lseg) {
/* remember that LAYOUTGET failed and suspend trying */
- set_bit(lo_fail_bit(iomode), &lo->state);
+ set_bit(lo_fail_bit(iomode), &lo->plh_flags);
}
return lseg;
}
dprintk("%s:Begin\n", __func__);
- assert_spin_locked(&lo->inode->i_lock);
- if (list_empty(&lo->segs)) {
- struct nfs_client *clp = NFS_SERVER(lo->inode)->nfs_client;
+ assert_spin_locked(&lo->plh_inode->i_lock);
+ if (list_empty(&lo->plh_segs)) {
+ struct nfs_client *clp = NFS_SERVER(lo->plh_inode)->nfs_client;
spin_lock(&clp->cl_lock);
- BUG_ON(!list_empty(&lo->layouts));
- list_add_tail(&lo->layouts, &clp->cl_layouts);
+ BUG_ON(!list_empty(&lo->plh_layouts));
+ list_add_tail(&lo->plh_layouts, &clp->cl_layouts);
spin_unlock(&clp->cl_lock);
}
- list_for_each_entry(lp, &lo->segs, fi_list) {
- if (cmp_layout(lp->range.iomode, lseg->range.iomode) > 0)
+ list_for_each_entry(lp, &lo->plh_segs, pls_list) {
+ if (cmp_layout(lp->pls_range.iomode, lseg->pls_range.iomode) > 0)
continue;
- list_add_tail(&lseg->fi_list, &lp->fi_list);
+ list_add_tail(&lseg->pls_list, &lp->pls_list);
dprintk("%s: inserted lseg %p "
"iomode %d offset %llu length %llu before "
"lp %p iomode %d offset %llu length %llu\n",
- __func__, lseg, lseg->range.iomode,
- lseg->range.offset, lseg->range.length,
- lp, lp->range.iomode, lp->range.offset,
- lp->range.length);
+ __func__, lseg, lseg->pls_range.iomode,
+ lseg->pls_range.offset, lseg->pls_range.length,
+ lp, lp->pls_range.iomode, lp->pls_range.offset,
+ lp->pls_range.length);
found = 1;
break;
}
if (!found) {
- list_add_tail(&lseg->fi_list, &lo->segs);
+ list_add_tail(&lseg->pls_list, &lo->plh_segs);
dprintk("%s: inserted lseg %p "
"iomode %d offset %llu length %llu at tail\n",
- __func__, lseg, lseg->range.iomode,
- lseg->range.offset, lseg->range.length);
+ __func__, lseg, lseg->pls_range.iomode,
+ lseg->pls_range.offset, lseg->pls_range.length);
}
get_layout_hdr_locked(lo);
lo = kzalloc(sizeof(struct pnfs_layout_hdr), GFP_KERNEL);
if (!lo)
return NULL;
- lo->refcount = 1;
- INIT_LIST_HEAD(&lo->layouts);
- INIT_LIST_HEAD(&lo->segs);
- seqlock_init(&lo->seqlock);
- lo->inode = ino;
+ lo->plh_refcount = 1;
+ INIT_LIST_HEAD(&lo->plh_layouts);
+ INIT_LIST_HEAD(&lo->plh_segs);
+ lo->plh_inode = ino;
return lo;
}
dprintk("%s Begin ino=%p layout=%p\n", __func__, ino, nfsi->layout);
assert_spin_locked(&ino->i_lock);
- if (nfsi->layout)
- return nfsi->layout;
-
+ if (nfsi->layout) {
+ if (test_bit(NFS_LAYOUT_DESTROYED, &nfsi->layout->plh_flags))
+ return NULL;
+ else
+ return nfsi->layout;
+ }
spin_unlock(&ino->i_lock);
new = alloc_init_layout_hdr(ino);
spin_lock(&ino->i_lock);
static int
is_matching_lseg(struct pnfs_layout_segment *lseg, u32 iomode)
{
- return (iomode != IOMODE_RW || lseg->range.iomode == IOMODE_RW);
+ return (iomode != IOMODE_RW || lseg->pls_range.iomode == IOMODE_RW);
}
/*
dprintk("%s:Begin\n", __func__);
- assert_spin_locked(&lo->inode->i_lock);
- list_for_each_entry(lseg, &lo->segs, fi_list) {
- if (is_matching_lseg(lseg, iomode)) {
+ assert_spin_locked(&lo->plh_inode->i_lock);
+ list_for_each_entry(lseg, &lo->plh_segs, pls_list) {
+ if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags) &&
+ is_matching_lseg(lseg, iomode)) {
ret = lseg;
break;
}
- if (cmp_layout(iomode, lseg->range.iomode) > 0)
+ if (cmp_layout(iomode, lseg->pls_range.iomode) > 0)
break;
}
dprintk("%s:Return lseg %p ref %d\n",
- __func__, ret, ret ? atomic_read(&ret->kref.refcount) : 0);
+ __func__, ret, ret ? atomic_read(&ret->pls_refcount) : 0);
return ret;
}
}
/* if LAYOUTGET already failed once we don't try again */
- if (test_bit(lo_fail_bit(iomode), &nfsi->layout->state))
+ if (test_bit(lo_fail_bit(iomode), &nfsi->layout->plh_flags))
+ goto out_unlock;
+
+ if (pnfs_layoutgets_blocked(lo, 0))
goto out_unlock;
+ atomic_inc(&lo->plh_outstanding);
- get_layout_hdr_locked(lo); /* Matched in nfs4_layoutget_release */
+ get_layout_hdr_locked(lo);
spin_unlock(&ino->i_lock);
lseg = send_layoutget(lo, ctx, iomode);
+ atomic_dec(&lo->plh_outstanding);
+ put_layout_hdr(ino);
out:
dprintk("%s end, state 0x%lx lseg %p\n", __func__,
- nfsi->layout->state, lseg);
+ nfsi->layout->plh_flags, lseg);
return lseg;
out_unlock:
spin_unlock(&ino->i_lock);
struct pnfs_layout_hdr *lo = NFS_I(lgp->args.inode)->layout;
struct nfs4_layoutget_res *res = &lgp->res;
struct pnfs_layout_segment *lseg;
- struct inode *ino = lo->inode;
+ struct inode *ino = lo->plh_inode;
int status = 0;
/* Inject layout blob into I/O device driver */
spin_lock(&ino->i_lock);
init_lseg(lo, lseg);
- lseg->range = res->range;
+ lseg->pls_range = res->range;
*lgp->lsegpp = lseg;
pnfs_insert_layout(lo, lseg);
projects / mirror_ubuntu-bionic-kernel.git / blobdiff