1 files changed, 134 insertions, 79 deletions

diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c
index 1a2c5d390f7f..2c7b151a7c95 100644
--- a/fs/ceph/mdsmap.c
+++ b/fs/ceph/mdsmap.c
@@ -7,41 +7,54 @@
 #include <linux/slab.h>
 #include <linux/types.h>
 
-#include <linux/ceph/mdsmap.h>
 #include <linux/ceph/messenger.h>
 #include <linux/ceph/decode.h>
 
+#include "mdsmap.h"
+#include "mds_client.h"
 #include "super.h"
 
+#define CEPH_MDS_IS_READY(i, ignore_laggy) \
+	(m->m_info[i].state > 0 && ignore_laggy ? true : !m->m_info[i].laggy)
 
-/*
- * choose a random mds that is "up" (i.e. has a state > 0), or -1.
- */
-int ceph_mdsmap_get_random_mds(struct ceph_mdsmap *m)
+static int __mdsmap_get_random_mds(struct ceph_mdsmap *m, bool ignore_laggy)
 {
 	int n = 0;
-	int i;
-
-	/* special case for one mds */
-	if (1 == m->m_num_mds && m->m_info[0].state > 0)
-		return 0;
+	int i, j;
 
 	/* count */
-	for (i = 0; i < m->m_num_mds; i++)
-		if (m->m_info[i].state > 0)
+	for (i = 0; i < m->possible_max_rank; i++)
+		if (CEPH_MDS_IS_READY(i, ignore_laggy))
 			n++;
 	if (n == 0)
 		return -1;
 
 	/* pick */
-	n = prandom_u32() % n;
-	for (i = 0; n > 0; i++, n--)
-		while (m->m_info[i].state <= 0)
-			i++;
+	n = get_random_u32_below(n);
+	for (j = 0, i = 0; i < m->possible_max_rank; i++) {
+		if (CEPH_MDS_IS_READY(i, ignore_laggy))
+			j++;
+		if (j > n)
+			break;
+	}
 
 	return i;
 }
 
+/*
+ * choose a random mds that is "up" (i.e. has a state > 0), or -1.
+ */
+int ceph_mdsmap_get_random_mds(struct ceph_mdsmap *m)
+{
+	int mds;
+
+	mds = __mdsmap_get_random_mds(m, false);
+	if (mds == m->possible_max_rank || mds == -1)
+		mds = __mdsmap_get_random_mds(m, true);
+
+	return mds == m->possible_max_rank ? -1 : mds;
+}
+
 #define __decode_and_drop_type(p, end, type, bad)		\
 	do {							\
 		if (*p + sizeof(type) > end)			\
@@ -102,14 +115,17 @@ bad:
  * Ignore any fields we don't care about (there are quite a few of
  * them).
  */
-struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
+struct ceph_mdsmap *ceph_mdsmap_decode(struct ceph_mds_client *mdsc, void **p,
+				       void *end, bool msgr2)
 {
+	struct ceph_client *cl = mdsc->fsc->client;
 	struct ceph_mdsmap *m;
 	const void *start = *p;
 	int i, j, n;
-	int err = -EINVAL;
-	u8 mdsmap_v, mdsmap_cv;
+	int err;
+	u8 mdsmap_v;
 	u16 mdsmap_ev;
+	u32 target;
 
 	m = kzalloc(sizeof(*m), GFP_NOFS);
 	if (!m)
@@ -117,7 +133,7 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 
 	ceph_decode_need(p, end, 1 + 1, bad);
 	mdsmap_v = ceph_decode_8(p);
-	mdsmap_cv = ceph_decode_8(p);
+	*p += sizeof(u8);			/* mdsmap_cv */
 	if (mdsmap_v >= 4) {
 	       u32 mdsmap_len;
 	       ceph_decode_32_safe(p, end, mdsmap_len, bad);
@@ -135,19 +151,33 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 	m->m_session_autoclose = ceph_decode_32(p);
 	m->m_max_file_size = ceph_decode_64(p);
 	m->m_max_mds = ceph_decode_32(p);
-	m->m_num_mds = m->m_max_mds;
 
-	m->m_info = kcalloc(m->m_num_mds, sizeof(*m->m_info), GFP_NOFS);
+	/*
+	 * pick out the active nodes as the m_num_active_mds, the
+	 * m_num_active_mds maybe larger than m_max_mds when decreasing
+	 * the max_mds in cluster side, in other case it should less
+	 * than or equal to m_max_mds.
+	 */
+	m->m_num_active_mds = n = ceph_decode_32(p);
+
+	/*
+	 * the possible max rank, it maybe larger than the m_num_active_mds,
+	 * for example if the mds_max == 2 in the cluster, when the MDS(0)
+	 * was laggy and being replaced by a new MDS, we will temporarily
+	 * receive a new mds map with n_num_mds == 1 and the active MDS(1),
+	 * and the mds rank >= m_num_active_mds.
+	 */
+	m->possible_max_rank = max(m->m_num_active_mds, m->m_max_mds);
+
+	m->m_info = kcalloc(m->possible_max_rank, sizeof(*m->m_info), GFP_NOFS);
 	if (!m->m_info)
 		goto nomem;
 
 	/* pick out active nodes from mds_info (state > 0) */
-	n = ceph_decode_32(p);
 	for (i = 0; i < n; i++) {
 		u64 global_id;
 		u32 namelen;
 		s32 mds, inc, state;
-		u64 state_seq;
 		u8 info_v;
 		void *info_end = NULL;
 		struct ceph_entity_addr addr;
@@ -155,15 +185,15 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 		void *pexport_targets = NULL;
 		struct ceph_timespec laggy_since;
 		struct ceph_mds_info *info;
+		bool laggy;
 
 		ceph_decode_need(p, end, sizeof(u64) + 1, bad);
 		global_id = ceph_decode_64(p);
 		info_v= ceph_decode_8(p);
 		if (info_v >= 4) {
 			u32 info_len;
-			u8 info_cv;
 			ceph_decode_need(p, end, 1 + sizeof(u32), bad);
-			info_cv = ceph_decode_8(p);
+			*p += sizeof(u8);	/* info_cv */
 			info_len = ceph_decode_32(p);
 			info_end = *p + info_len;
 			if (info_end > end)
@@ -175,17 +205,20 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 		namelen = ceph_decode_32(p);  /* skip mds name */
 		*p += namelen;
 
-		ceph_decode_need(p, end,
-				 4*sizeof(u32) + sizeof(u64) +
-				 sizeof(addr) + sizeof(struct ceph_timespec),
-				 bad);
-		mds = ceph_decode_32(p);
-		inc = ceph_decode_32(p);
-		state = ceph_decode_32(p);
-		state_seq = ceph_decode_64(p);
-		ceph_decode_copy(p, &addr, sizeof(addr));
-		ceph_decode_addr(&addr);
-		ceph_decode_copy(p, &laggy_since, sizeof(laggy_since));
+		ceph_decode_32_safe(p, end, mds, bad);
+		ceph_decode_32_safe(p, end, inc, bad);
+		ceph_decode_32_safe(p, end, state, bad);
+		*p += sizeof(u64);		/* state_seq */
+		if (info_v >= 8)
+			err = ceph_decode_entity_addrvec(p, end, msgr2, &addr);
+		else
+			err = ceph_decode_entity_addr(p, end, &addr);
+		if (err)
+			goto corrupt;
+
+		ceph_decode_copy_safe(p, end, &laggy_since, sizeof(laggy_since),
+				      bad);
+		laggy = laggy_since.tv_sec != 0 || laggy_since.tv_nsec != 0;
 		*p += sizeof(u32);
 		ceph_decode_32_safe(p, end, namelen, bad);
 		*p += namelen;
@@ -203,52 +236,40 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 			*p = info_end;
 		}
 
-		dout("mdsmap_decode %d/%d %lld mds%d.%d %s %s\n",
-		     i+1, n, global_id, mds, inc,
-		     ceph_pr_addr(&addr.in_addr),
-		     ceph_mds_state_name(state));
+		doutc(cl, "%d/%d %lld mds%d.%d %s %s%s\n", i+1, n, global_id,
+		      mds, inc, ceph_pr_addr(&addr),
+		      ceph_mds_state_name(state), laggy ? "(laggy)" : "");
 
-		if (mds < 0 || state <= 0)
+		if (mds < 0 || mds >= m->possible_max_rank) {
+			pr_warn_client(cl, "got incorrect mds(%d)\n", mds);
 			continue;
+		}
 
-		if (mds >= m->m_num_mds) {
-			int new_num = max(mds + 1, m->m_num_mds * 2);
-			void *new_m_info = krealloc(m->m_info,
-						new_num * sizeof(*m->m_info),
-						GFP_NOFS | __GFP_ZERO);
-			if (!new_m_info)
-				goto nomem;
-			m->m_info = new_m_info;
-			m->m_num_mds = new_num;
+		if (state <= 0) {
+			doutc(cl, "got incorrect state(%s)\n",
+			      ceph_mds_state_name(state));
+			continue;
 		}
 
 		info = &m->m_info[mds];
 		info->global_id = global_id;
 		info->state = state;
 		info->addr = addr;
-		info->laggy = (laggy_since.tv_sec != 0 ||
-			       laggy_since.tv_nsec != 0);
+		info->laggy = laggy;
 		info->num_export_targets = num_export_targets;
 		if (num_export_targets) {
 			info->export_targets = kcalloc(num_export_targets,
 						       sizeof(u32), GFP_NOFS);
 			if (!info->export_targets)
 				goto nomem;
-			for (j = 0; j < num_export_targets; j++)
-				info->export_targets[j] =
-				       ceph_decode_32(&pexport_targets);
+			for (j = 0; j < num_export_targets; j++) {
+				target = ceph_decode_32(&pexport_targets);
+				info->export_targets[j] = target;
+			}
 		} else {
 			info->export_targets = NULL;
 		}
 	}
-	if (m->m_num_mds > m->m_max_mds) {
-		/* find max up mds */
-		for (i = m->m_num_mds; i >= m->m_max_mds; i--) {
-			if (i == 0 || m->m_info[i-1].state > 0)
-				break;
-		}
-		m->m_num_mds = i;
-	}
 
 	/* pg_pools */
 	ceph_decode_32_safe(p, end, n, bad);
@@ -290,14 +311,14 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 
 		for (i = 0; i < n; i++) {
 			s32 mds = ceph_decode_32(p);
-			if (mds >= 0 && mds < m->m_num_mds) {
+			if (mds >= 0 && mds < m->possible_max_rank) {
 				if (m->m_info[mds].laggy)
 					num_laggy++;
 			}
 		}
 		m->m_num_laggy = num_laggy;
 
-		if (n > m->m_num_mds) {
+		if (n > m->possible_max_rank) {
 			void *new_m_info = krealloc(m->m_info,
 						    n * sizeof(*m->m_info),
 						    GFP_NOFS | __GFP_ZERO);
@@ -305,7 +326,7 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 				goto nomem;
 			m->m_info = new_m_info;
 		}
-		m->m_num_mds = n;
+		m->possible_max_rank = n;
 	}
 
 	/* inc */
@@ -332,12 +353,22 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 		__decode_and_drop_type(p, end, u8, bad_ext);
 	}
 	if (mdsmap_ev >= 8) {
-		u32 name_len;
+		u32 fsname_len;
 		/* enabled */
 		ceph_decode_8_safe(p, end, m->m_enabled, bad_ext);
-		ceph_decode_32_safe(p, end, name_len, bad_ext);
-		ceph_decode_need(p, end, name_len, bad_ext);
-		*p += name_len;
+		/* fs_name */
+		ceph_decode_32_safe(p, end, fsname_len, bad_ext);
+
+		/* validate fsname against mds_namespace */
+		if (!namespace_equals(mdsc->fsc->mount_options, *p,
+				      fsname_len)) {
+			pr_warn_client(cl, "fsname %*pE doesn't match mds_namespace %s\n",
+				       (int)fsname_len, (char *)*p,
+				       mdsc->fsc->mount_options->mds_namespace);
+			goto bad;
+		}
+		/* skip fsname after validation */
+		ceph_decode_skip_n(p, end, fsname_len, bad);
 	}
 	/* damaged */
 	if (mdsmap_ev >= 9) {
@@ -350,30 +381,54 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end)
 	} else {
 		m->m_damaged = false;
 	}
+	if (mdsmap_ev >= 17) {
+		/* balancer */
+		ceph_decode_skip_string(p, end, bad_ext);
+		/* standby_count_wanted */
+		ceph_decode_skip_32(p, end, bad_ext);
+		/* old_max_mds */
+		ceph_decode_skip_32(p, end, bad_ext);
+		/* min_compat_client */
+		ceph_decode_skip_8(p, end, bad_ext);
+		/* required_client_features */
+		ceph_decode_skip_set(p, end, 64, bad_ext);
+		/* bal_rank_mask */
+		ceph_decode_skip_string(p, end, bad_ext);
+	}
+	if (mdsmap_ev >= 18) {
+		ceph_decode_64_safe(p, end, m->m_max_xattr_size, bad_ext);
+	}
 bad_ext:
+	doutc(cl, "m_enabled: %d, m_damaged: %d, m_num_laggy: %d\n",
+	      !!m->m_enabled, !!m->m_damaged, m->m_num_laggy);
 	*p = end;
-	dout("mdsmap_decode success epoch %u\n", m->m_epoch);
+	doutc(cl, "success epoch %u\n", m->m_epoch);
 	return m;
 nomem:
 	err = -ENOMEM;
 	goto out_err;
-bad:
-	pr_err("corrupt mdsmap\n");
+corrupt:
+	pr_err_client(cl, "corrupt mdsmap\n");
 	print_hex_dump(KERN_DEBUG, "mdsmap: ",
 		       DUMP_PREFIX_OFFSET, 16, 1,
 		       start, end - start, true);
 out_err:
 	ceph_mdsmap_destroy(m);
 	return ERR_PTR(err);
+bad:
+	err = -EINVAL;
+	goto corrupt;
 }
 
 void ceph_mdsmap_destroy(struct ceph_mdsmap *m)
 {
 	int i;
 
-	for (i = 0; i < m->m_num_mds; i++)
-		kfree(m->m_info[i].export_targets);
-	kfree(m->m_info);
+	if (m->m_info) {
+		for (i = 0; i < m->possible_max_rank; i++)
+			kfree(m->m_info[i].export_targets);
+		kfree(m->m_info);
+	}
 	kfree(m->m_data_pg_pools);
 	kfree(m);
 }
@@ -385,9 +440,9 @@ bool ceph_mdsmap_is_cluster_available(struct ceph_mdsmap *m)
 		return false;
 	if (m->m_damaged)
 		return false;
-	if (m->m_num_laggy > 0)
+	if (m->m_num_laggy == m->m_num_active_mds)
 		return false;
-	for (i = 0; i < m->m_num_mds; i++) {
+	for (i = 0; i < m->possible_max_rank; i++) {
 		if (m->m_info[i].state == CEPH_MDS_STATE_ACTIVE)
 			nr_active++;
 	}