1 files changed, 221 insertions, 62 deletions

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 157fe59fbabe..23dbfae16156 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0
 #include <linux/ceph/ceph_debug.h>
 
 #include <linux/fs.h>
@@ -153,13 +154,19 @@ void ceph_adjust_min_caps(struct ceph_mds_client *mdsc, int delta)
 	spin_unlock(&mdsc->caps_list_lock);
 }
 
-void ceph_reserve_caps(struct ceph_mds_client *mdsc,
+/*
+ * Called under mdsc->mutex.
+ */
+int ceph_reserve_caps(struct ceph_mds_client *mdsc,
 		      struct ceph_cap_reservation *ctx, int need)
 {
-	int i;
+	int i, j;
 	struct ceph_cap *cap;
 	int have;
 	int alloc = 0;
+	int max_caps;
+	bool trimmed = false;
+	struct ceph_mds_session *s;
 	LIST_HEAD(newcaps);
 
 	dout("reserve caps ctx=%p need=%d\n", ctx, need);
@@ -177,17 +184,56 @@ void ceph_reserve_caps(struct ceph_mds_client *mdsc,
 					 mdsc->caps_avail_count);
 	spin_unlock(&mdsc->caps_list_lock);
 
-	for (i = have; i < need; i++) {
+	for (i = have; i < need; ) {
 		cap = kmem_cache_alloc(ceph_cap_cachep, GFP_NOFS);
-		if (!cap)
-			break;
-		list_add(&cap->caps_item, &newcaps);
-		alloc++;
-	}
-	/* we didn't manage to reserve as much as we needed */
-	if (have + alloc != need)
+		if (cap) {
+			list_add(&cap->caps_item, &newcaps);
+			alloc++;
+			i++;
+			continue;
+		}
+
+		if (!trimmed) {
+			for (j = 0; j < mdsc->max_sessions; j++) {
+				s = __ceph_lookup_mds_session(mdsc, j);
+				if (!s)
+					continue;
+				mutex_unlock(&mdsc->mutex);
+
+				mutex_lock(&s->s_mutex);
+				max_caps = s->s_nr_caps - (need - i);
+				ceph_trim_caps(mdsc, s, max_caps);
+				mutex_unlock(&s->s_mutex);
+
+				ceph_put_mds_session(s);
+				mutex_lock(&mdsc->mutex);
+			}
+			trimmed = true;
+
+			spin_lock(&mdsc->caps_list_lock);
+			if (mdsc->caps_avail_count) {
+				int more_have;
+				if (mdsc->caps_avail_count >= need - i)
+					more_have = need - i;
+				else
+					more_have = mdsc->caps_avail_count;
+
+				i += more_have;
+				have += more_have;
+				mdsc->caps_avail_count -= more_have;
+				mdsc->caps_reserve_count += more_have;
+
+			}
+			spin_unlock(&mdsc->caps_list_lock);
+
+			continue;
+		}
+
 		pr_warn("reserve caps ctx=%p ENOMEM need=%d got=%d\n",
 			ctx, need, have + alloc);
+		goto out_nomem;
+	}
+	BUG_ON(have + alloc != need);
 
 	spin_lock(&mdsc->caps_list_lock);
 	mdsc->caps_total_count += alloc;
@@ -203,17 +249,61 @@ void ceph_reserve_caps(struct ceph_mds_client *mdsc,
 	dout("reserve caps ctx=%p %d = %d used + %d resv + %d avail\n",
 	     ctx, mdsc->caps_total_count, mdsc->caps_use_count,
 	     mdsc->caps_reserve_count, mdsc->caps_avail_count);
+	return 0;
+
+out_nomem:
+
+	spin_lock(&mdsc->caps_list_lock);
+	mdsc->caps_avail_count += have;
+	mdsc->caps_reserve_count -= have;
+
+	while (!list_empty(&newcaps)) {
+		cap = list_first_entry(&newcaps,
+				struct ceph_cap, caps_item);
+		list_del(&cap->caps_item);
+
+		/* Keep some preallocated caps around (ceph_min_count), to
+		 * avoid lots of free/alloc churn. */
+		if (mdsc->caps_avail_count >=
+		    mdsc->caps_reserve_count + mdsc->caps_min_count) {
+			kmem_cache_free(ceph_cap_cachep, cap);
+		} else {
+			mdsc->caps_avail_count++;
+			mdsc->caps_total_count++;
+			list_add(&cap->caps_item, &mdsc->caps_list);
+		}
+	}
+
+	BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count +
+					 mdsc->caps_reserve_count +
+					 mdsc->caps_avail_count);
+	spin_unlock(&mdsc->caps_list_lock);
+	return -ENOMEM;
 }
 
 int ceph_unreserve_caps(struct ceph_mds_client *mdsc,
 			struct ceph_cap_reservation *ctx)
 {
+	int i;
+	struct ceph_cap *cap;
+
 	dout("unreserve caps ctx=%p count=%d\n", ctx, ctx->count);
 	if (ctx->count) {
 		spin_lock(&mdsc->caps_list_lock);
 		BUG_ON(mdsc->caps_reserve_count < ctx->count);
 		mdsc->caps_reserve_count -= ctx->count;
-		mdsc->caps_avail_count += ctx->count;
+		if (mdsc->caps_avail_count >=
+		    mdsc->caps_reserve_count + mdsc->caps_min_count) {
+			mdsc->caps_total_count -= ctx->count;
+			for (i = 0; i < ctx->count; i++) {
+				cap = list_first_entry(&mdsc->caps_list,
+					struct ceph_cap, caps_item);
+				list_del(&cap->caps_item);
+				kmem_cache_free(ceph_cap_cachep, cap);
+			}
+		} else {
+			mdsc->caps_avail_count += ctx->count;
+		}
 		ctx->count = 0;
 		dout("unreserve caps %d = %d used + %d resv + %d avail\n",
 		     mdsc->caps_total_count, mdsc->caps_use_count,
@@ -239,7 +329,23 @@ struct ceph_cap *ceph_get_cap(struct ceph_mds_client *mdsc,
 			mdsc->caps_use_count++;
 			mdsc->caps_total_count++;
 			spin_unlock(&mdsc->caps_list_lock);
+		} else {
+			spin_lock(&mdsc->caps_list_lock);
+			if (mdsc->caps_avail_count) {
+				BUG_ON(list_empty(&mdsc->caps_list));
+
+				mdsc->caps_avail_count--;
+				mdsc->caps_use_count++;
+				cap = list_first_entry(&mdsc->caps_list,
+						struct ceph_cap, caps_item);
+				list_del(&cap->caps_item);
+
+				BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count +
+				       mdsc->caps_reserve_count + mdsc->caps_avail_count);
+			}
+			spin_unlock(&mdsc->caps_list_lock);
 		}
+
 		return cap;
 	}
 
@@ -295,6 +401,8 @@ void ceph_reservation_status(struct ceph_fs_client *fsc,
 {
 	struct ceph_mds_client *mdsc = fsc->mdsc;
 
+	spin_lock(&mdsc->caps_list_lock);
+
 	if (total)
 		*total = mdsc->caps_total_count;
 	if (avail)
@@ -305,6 +413,8 @@ void ceph_reservation_status(struct ceph_fs_client *fsc,
 		*reserved = mdsc->caps_reserve_count;
 	if (min)
 		*min = mdsc->caps_min_count;
+
+	spin_unlock(&mdsc->caps_list_lock);
 }
 
 /*
@@ -497,7 +607,7 @@ static void __check_cap_issue(struct ceph_inode_info *ci, struct ceph_cap *cap,
 	 */
 	if ((issued & CEPH_CAP_FILE_SHARED) != (had & CEPH_CAP_FILE_SHARED)) {
 		if (issued & CEPH_CAP_FILE_SHARED)
-			ci->i_shared_gen++;
+			atomic_inc(&ci->i_shared_gen);
 		if (S_ISDIR(ci->vfs_inode.i_mode)) {
 			dout(" marking %p NOT complete\n", &ci->vfs_inode);
 			__ceph_dir_clear_complete(ci);
@@ -576,18 +686,32 @@ void ceph_add_cap(struct inode *inode,
 		}
 	}
 
-	if (!ci->i_snap_realm) {
+	if (!ci->i_snap_realm ||
+	    ((flags & CEPH_CAP_FLAG_AUTH) &&
+	     realmino != (u64)-1 && ci->i_snap_realm->ino != realmino)) {
 		/*
 		 * add this inode to the appropriate snap realm
 		 */
 		struct ceph_snap_realm *realm = ceph_lookup_snap_realm(mdsc,
 							       realmino);
 		if (realm) {
+			struct ceph_snap_realm *oldrealm = ci->i_snap_realm;
+			if (oldrealm) {
+				spin_lock(&oldrealm->inodes_with_caps_lock);
+				list_del_init(&ci->i_snap_realm_item);
+				spin_unlock(&oldrealm->inodes_with_caps_lock);
+			}
+
 			spin_lock(&realm->inodes_with_caps_lock);
-			ci->i_snap_realm = realm;
 			list_add(&ci->i_snap_realm_item,
 				 &realm->inodes_with_caps);
+			ci->i_snap_realm = realm;
+			if (realm->ino == ci->i_vino.ino)
+				realm->inode = inode;
 			spin_unlock(&realm->inodes_with_caps_lock);
+
+			if (oldrealm)
+				ceph_put_snap_realm(mdsc, oldrealm);
 		} else {
 			pr_err("ceph_add_cap: couldn't find snap realm %llx\n",
 			       realmino);
@@ -889,6 +1013,11 @@ int __ceph_caps_mds_wanted(struct ceph_inode_info *ci, bool check)
 /*
  * called under i_ceph_lock
  */
+static int __ceph_is_single_caps(struct ceph_inode_info *ci)
+{
+	return rb_first(&ci->i_caps) == rb_last(&ci->i_caps);
+}
+
 static int __ceph_is_any_caps(struct ceph_inode_info *ci)
 {
 	return !RB_EMPTY_ROOT(&ci->i_caps);
@@ -1159,7 +1288,7 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
 	struct ceph_inode_info *ci = cap->ci;
 	struct inode *inode = &ci->vfs_inode;
 	struct cap_msg_args arg;
-	int held, revoking, dropping;
+	int held, revoking;
 	int wake = 0;
 	int delayed = 0;
 	int ret;
@@ -1167,7 +1296,6 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
 	held = cap->issued | cap->implemented;
 	revoking = cap->implemented & ~cap->issued;
 	retain &= ~revoking;
-	dropping = cap->issued & ~retain;
 
 	dout("__send_cap %p cap %p session %p %s -> %s (revoking %s)\n",
 	     inode, cap, cap->session,
@@ -1703,21 +1831,24 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
 	int mds = -1;   /* keep track of how far we've gone through i_caps list
 			   to avoid an infinite loop on retry */
 	struct rb_node *p;
-	int delayed = 0, sent = 0, num;
-	bool is_delayed = flags & CHECK_CAPS_NODELAY;
+	int delayed = 0, sent = 0;
+	bool no_delay = flags & CHECK_CAPS_NODELAY;
 	bool queue_invalidate = false;
-	bool force_requeue = false;
 	bool tried_invalidate = false;
 
 	/* if we are unmounting, flush any unused caps immediately. */
 	if (mdsc->stopping)
-		is_delayed = 1;
+		no_delay = true;
 
 	spin_lock(&ci->i_ceph_lock);
 
 	if (ci->i_ceph_flags & CEPH_I_FLUSH)
 		flags |= CHECK_CAPS_FLUSH;
 
+	if (!(flags & CHECK_CAPS_AUTHONLY) ||
+	    (ci->i_auth_cap && __ceph_is_single_caps(ci)))
+		__cap_delay_cancel(mdsc, ci);
+
 	goto retry_locked;
 retry:
 	spin_lock(&ci->i_ceph_lock);
@@ -1772,7 +1903,7 @@ retry_locked:
 	 * have cached pages, but don't want them, then try to invalidate.
 	 * If we fail, it's because pages are locked.... try again later.
 	 */
-	if ((!is_delayed || mdsc->stopping) &&
+	if ((!no_delay || mdsc->stopping) &&
 	    !S_ISDIR(inode->i_mode) &&		/* ignore readdir cache */
 	    !(ci->i_wb_ref || ci->i_wrbuffer_ref) &&   /* no dirty pages... */
 	    inode->i_data.nrpages &&		/* have cached pages */
@@ -1781,27 +1912,16 @@ retry_locked:
 	    !tried_invalidate) {
 		dout("check_caps trying to invalidate on %p\n", inode);
 		if (try_nonblocking_invalidate(inode) < 0) {
-			if (revoking & (CEPH_CAP_FILE_CACHE|
-					CEPH_CAP_FILE_LAZYIO)) {
-				dout("check_caps queuing invalidate\n");
-				queue_invalidate = true;
-				ci->i_rdcache_revoking = ci->i_rdcache_gen;
-			} else {
-				dout("check_caps failed to invalidate pages\n");
-				/* we failed to invalidate pages.  check these
-				   caps again later. */
-				force_requeue = true;
-				__cap_set_timeouts(mdsc, ci);
-			}
+			dout("check_caps queuing invalidate\n");
+			queue_invalidate = true;
+			ci->i_rdcache_revoking = ci->i_rdcache_gen;
 		}
 		tried_invalidate = true;
 		goto retry_locked;
 	}
 
-	num = 0;
 	for (p = rb_first(&ci->i_caps); p; p = rb_next(p)) {
 		cap = rb_entry(p, struct ceph_cap, ci_node);
-		num++;
 
 		/* avoid looping forever */
 		if (mds >= cap->mds ||
@@ -1864,7 +1984,7 @@ retry_locked:
 		    cap->mds_wanted == want)
 			continue;     /* nope, all good */
 
-		if (is_delayed)
+		if (no_delay)
 			goto ack;
 
 		/* delay? */
@@ -1955,15 +2075,8 @@ ack:
 		goto retry; /* retake i_ceph_lock and restart our cap scan. */
 	}
 
-	/*
-	 * Reschedule delayed caps release if we delayed anything,
-	 * otherwise cancel.
-	 */
-	if (delayed && is_delayed)
-		force_requeue = true;   /* __send_cap delayed release; requeue */
-	if (!delayed && !is_delayed)
-		__cap_delay_cancel(mdsc, ci);
-	else if (!is_delayed || force_requeue)
+	/* Reschedule delayed caps release if we delayed anything */
+	if (delayed)
 		__cap_delay_requeue(mdsc, ci);
 
 	spin_unlock(&ci->i_ceph_lock);
@@ -1991,6 +2104,7 @@ static int try_flush_caps(struct inode *inode, u64 *ptid)
 retry:
 	spin_lock(&ci->i_ceph_lock);
 	if (ci->i_ceph_flags & CEPH_I_NOFLUSH) {
+		spin_unlock(&ci->i_ceph_lock);
 		dout("try_flush_caps skipping %p I_NOFLUSH set\n", inode);
 		goto out;
 	}
@@ -2008,8 +2122,10 @@ retry:
 			mutex_lock(&session->s_mutex);
 			goto retry;
 		}
-		if (cap->session->s_state < CEPH_MDS_SESSION_OPEN)
+		if (cap->session->s_state < CEPH_MDS_SESSION_OPEN) {
+			spin_unlock(&ci->i_ceph_lock);
 			goto out;
+		}
 
 		flushing = __mark_caps_flushing(inode, session, true,
 						&flush_tid, &oldest_flush_tid);
@@ -2157,7 +2273,7 @@ int ceph_write_inode(struct inode *inode, struct writeback_control *wbc)
 	u64 flush_tid;
 	int err = 0;
 	int dirty;
-	int wait = wbc->sync_mode == WB_SYNC_ALL;
+	int wait = (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync);
 
 	dout("write_inode %p wait=%d\n", inode, wait);
 	if (wait) {
@@ -3185,8 +3301,8 @@ static void handle_cap_flush_ack(struct inode *inode, u64 flush_tid,
 	int dirty = le32_to_cpu(m->dirty);
 	int cleaned = 0;
 	bool drop = false;
-	bool wake_ci = 0;
-	bool wake_mdsc = 0;
+	bool wake_ci = false;
+	bool wake_mdsc = false;
 
 	list_for_each_entry_safe(cf, tmp_cf, &ci->i_cap_flush_list, i_list) {
 		if (cf->tid == flush_tid)
@@ -3423,7 +3539,14 @@ retry:
 	 */
 
 	issued = cap->issued;
-	WARN_ON(issued != cap->implemented);
+	if (issued != cap->implemented)
+		pr_err_ratelimited("handle_cap_export: issued != implemented: "
+				"ino (%llx.%llx) mds%d seq %d mseq %d "
+				"issued %s implemented %s\n",
+				ceph_vinop(inode), mds, cap->seq, cap->mseq,
+				ceph_cap_string(issued),
+				ceph_cap_string(cap->implemented));
+
 
 	tcap = __get_cap_for_mds(ci, target);
 	if (tcap) {
@@ -3569,12 +3692,13 @@ retry:
 		if ((ph->flags & CEPH_CAP_FLAG_AUTH) &&
 		    (ocap->seq != le32_to_cpu(ph->seq) ||
 		     ocap->mseq != le32_to_cpu(ph->mseq))) {
-			pr_err("handle_cap_import: mismatched seq/mseq: "
-			       "ino (%llx.%llx) mds%d seq %d mseq %d "
-			       "importer mds%d has peer seq %d mseq %d\n",
-			       ceph_vinop(inode), peer, ocap->seq,
-			       ocap->mseq, mds, le32_to_cpu(ph->seq),
-			       le32_to_cpu(ph->mseq));
+			pr_err_ratelimited("handle_cap_import: "
+					"mismatched seq/mseq: ino (%llx.%llx) "
+					"mds%d seq %d mseq %d importer mds%d "
+					"has peer seq %d mseq %d\n",
+					ceph_vinop(inode), peer, ocap->seq,
+					ocap->mseq, mds, le32_to_cpu(ph->seq),
+					le32_to_cpu(ph->mseq));
 		}
 		__ceph_remove_cap(ocap, (ph->flags & CEPH_CAP_FLAG_RELEASE));
 	}
@@ -3907,6 +4031,32 @@ void ceph_put_fmode(struct ceph_inode_info *ci, int fmode)
 }
 
 /*
+ * For a soon-to-be unlinked file, drop the AUTH_RDCACHE caps. If it
+ * looks like the link count will hit 0, drop any other caps (other
+ * than PIN) we don't specifically want (due to the file still being
+ * open).
+ */
+int ceph_drop_caps_for_unlink(struct inode *inode)
+{
+	struct ceph_inode_info *ci = ceph_inode(inode);
+	int drop = CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL;
+
+	spin_lock(&ci->i_ceph_lock);
+	if (inode->i_nlink == 1) {
+		drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN);
+
+		ci->i_ceph_flags |= CEPH_I_NODELAY;
+		if (__ceph_caps_dirty(ci)) {
+			struct ceph_mds_client *mdsc =
+				ceph_inode_to_client(inode)->mdsc;
+			__cap_delay_requeue_front(mdsc, ci);
+		}
+	}
+	spin_unlock(&ci->i_ceph_lock);
+	return drop;
+}
+
+/*
  * Helpers for embedding cap and dentry lease releases into mds
  * requests.
  *
@@ -3936,11 +4086,20 @@ int ceph_encode_inode_release(void **p, struct inode *inode,
 
 	cap = __get_cap_for_mds(ci, mds);
 	if (cap && __cap_is_valid(cap)) {
-		if (force ||
-		    ((cap->issued & drop) &&
-		     (cap->issued & unless) == 0)) {
-			if ((cap->issued & drop) &&
-			    (cap->issued & unless) == 0) {
+		unless &= cap->issued;
+		if (unless) {
+			if (unless & CEPH_CAP_AUTH_EXCL)
+				drop &= ~CEPH_CAP_AUTH_SHARED;
+			if (unless & CEPH_CAP_LINK_EXCL)
+				drop &= ~CEPH_CAP_LINK_SHARED;
+			if (unless & CEPH_CAP_XATTR_EXCL)
+				drop &= ~CEPH_CAP_XATTR_SHARED;
+			if (unless & CEPH_CAP_FILE_EXCL)
+				drop &= ~CEPH_CAP_FILE_SHARED;
+		}
+
+		if (force || (cap->issued & drop)) {
+			if (cap->issued & drop) {
 				int wanted = __ceph_caps_wanted(ci);
 				if ((ci->i_ceph_flags & CEPH_I_NODELAY) == 0)
 					wanted |= cap->mds_wanted;
@@ -3972,7 +4131,7 @@ int ceph_encode_inode_release(void **p, struct inode *inode,
 			*p += sizeof(*rel);
 			ret = 1;
 		} else {
-			dout("encode_inode_release %p cap %p %s\n",
+			dout("encode_inode_release %p cap %p %s (noop)\n",
 			     inode, cap, ceph_cap_string(cap->issued));
 		}
 	}