From 892c4467e335e9050c95e0d8409c136c4dadaca2 Mon Sep 17 00:00:00 2001
From: David Teigland <teigland@redhat.com>
Date: Wed, 7 Jan 2009 16:48:52 -0600
Subject: dlm: fix seq_file usage in debugfs lock dump

The old code would leak iterators and leave reference counts on
rsbs because it was ignoring the "stop" seq callback.  The code
followed an example that used the seq operations differently.
This new code is based on actually understanding how the seq
operations work.  It also improves things by saving the hash bucket
in the position to avoid cycling through completed buckets in start.

Siged-off-by: Davd Teigland <teigland@redhat.com>
---
 fs/dlm/debug_fs.c | 696 ++++++++++++++++++++++++++----------------------------
 1 file changed, 337 insertions(+), 359 deletions(-)

(limited to 'fs')

diff --git a/fs/dlm/debug_fs.c b/fs/dlm/debug_fs.c
index 2f107d1a6a4..bc4af3ef65a 100644
--- a/fs/dlm/debug_fs.c
+++ b/fs/dlm/debug_fs.c
@@ -1,7 +1,7 @@
 /******************************************************************************
 *******************************************************************************
 **
-**  Copyright (C) 2005-2008 Red Hat, Inc.  All rights reserved.
+**  Copyright (C) 2005-2009 Red Hat, Inc.  All rights reserved.
 **
 **  This copyrighted material is made available to anyone wishing to use,
 **  modify, copy, or redistribute it subject to the terms and conditions
@@ -25,19 +25,6 @@ static struct mutex debug_buf_lock;
 
 static struct dentry *dlm_root;
 
-struct rsb_iter {
-	int entry;
-	int format;
-	int header;
-	struct dlm_ls *ls;
-	struct list_head *next;
-	struct dlm_rsb *rsb;
-};
-
-/*
- * dump all rsb's in the lockspace hash table
- */
-
 static char *print_lockmode(int mode)
 {
 	switch (mode) {
@@ -60,13 +47,13 @@ static char *print_lockmode(int mode)
 	}
 }
 
-static void print_format1_lock(struct seq_file *s, struct dlm_lkb *lkb,
-			       struct dlm_rsb *res)
+static int print_format1_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			      struct dlm_rsb *res)
 {
 	seq_printf(s, "%08x %s", lkb->lkb_id, print_lockmode(lkb->lkb_grmode));
 
-	if (lkb->lkb_status == DLM_LKSTS_CONVERT
-	    || lkb->lkb_status == DLM_LKSTS_WAITING)
+	if (lkb->lkb_status == DLM_LKSTS_CONVERT ||
+	    lkb->lkb_status == DLM_LKSTS_WAITING)
 		seq_printf(s, " (%s)", print_lockmode(lkb->lkb_rqmode));
 
 	if (lkb->lkb_nodeid) {
@@ -80,33 +67,42 @@ static void print_format1_lock(struct seq_file *s, struct dlm_lkb *lkb,
 	if (lkb->lkb_wait_type)
 		seq_printf(s, " wait_type: %d", lkb->lkb_wait_type);
 
-	seq_printf(s, "\n");
+	return seq_printf(s, "\n");
 }
 
 static int print_format1(struct dlm_rsb *res, struct seq_file *s)
 {
 	struct dlm_lkb *lkb;
 	int i, lvblen = res->res_ls->ls_lvblen, recover_list, root_list;
+	int rv;
 
 	lock_rsb(res);
 
-	seq_printf(s, "\nResource %p Name (len=%d) \"", res, res->res_length);
+	rv = seq_printf(s, "\nResource %p Name (len=%d) \"",
+			res, res->res_length);
+	if (rv)
+		goto out;
+
 	for (i = 0; i < res->res_length; i++) {
 		if (isprint(res->res_name[i]))
 			seq_printf(s, "%c", res->res_name[i]);
 		else
 			seq_printf(s, "%c", '.');
 	}
+
 	if (res->res_nodeid > 0)
-		seq_printf(s, "\"  \nLocal Copy, Master is node %d\n",
-			   res->res_nodeid);
+		rv = seq_printf(s, "\"  \nLocal Copy, Master is node %d\n",
+				res->res_nodeid);
 	else if (res->res_nodeid == 0)
-		seq_printf(s, "\"  \nMaster Copy\n");
+		rv = seq_printf(s, "\"  \nMaster Copy\n");
 	else if (res->res_nodeid == -1)
-		seq_printf(s, "\"  \nLooking up master (lkid %x)\n",
-			   res->res_first_lkid);
+		rv = seq_printf(s, "\"  \nLooking up master (lkid %x)\n",
+			   	res->res_first_lkid);
 	else
-		seq_printf(s, "\"  \nInvalid master %d\n", res->res_nodeid);
+		rv = seq_printf(s, "\"  \nInvalid master %d\n",
+				res->res_nodeid);
+	if (rv)
+		goto out;
 
 	/* Print the LVB: */
 	if (res->res_lvbptr) {
@@ -119,52 +115,66 @@ static int print_format1(struct dlm_rsb *res, struct seq_file *s)
 		}
 		if (rsb_flag(res, RSB_VALNOTVALID))
 			seq_printf(s, " (INVALID)");
-		seq_printf(s, "\n");
+		rv = seq_printf(s, "\n");
+		if (rv)
+			goto out;
 	}
 
 	root_list = !list_empty(&res->res_root_list);
 	recover_list = !list_empty(&res->res_recover_list);
 
 	if (root_list || recover_list) {
-		seq_printf(s, "Recovery: root %d recover %d flags %lx "
-			   "count %d\n", root_list, recover_list,
-			   res->res_flags, res->res_recover_locks_count);
+		rv = seq_printf(s, "Recovery: root %d recover %d flags %lx "
+				"count %d\n", root_list, recover_list,
+			   	res->res_flags, res->res_recover_locks_count);
+		if (rv)
+			goto out;
 	}
 
 	/* Print the locks attached to this resource */
 	seq_printf(s, "Granted Queue\n");
-	list_for_each_entry(lkb, &res->res_grantqueue, lkb_statequeue)
-		print_format1_lock(s, lkb, res);
+	list_for_each_entry(lkb, &res->res_grantqueue, lkb_statequeue) {
+		rv = print_format1_lock(s, lkb, res);
+		if (rv)
+			goto out;
+	}
 
 	seq_printf(s, "Conversion Queue\n");
-	list_for_each_entry(lkb, &res->res_convertqueue, lkb_statequeue)
-		print_format1_lock(s, lkb, res);
+	list_for_each_entry(lkb, &res->res_convertqueue, lkb_statequeue) {
+		rv = print_format1_lock(s, lkb, res);
+		if (rv)
+			goto out;
+	}
 
 	seq_printf(s, "Waiting Queue\n");
-	list_for_each_entry(lkb, &res->res_waitqueue, lkb_statequeue)
-		print_format1_lock(s, lkb, res);
+	list_for_each_entry(lkb, &res->res_waitqueue, lkb_statequeue) {
+		rv = print_format1_lock(s, lkb, res);
+		if (rv)
+			goto out;
+	}
 
 	if (list_empty(&res->res_lookup))
 		goto out;
 
 	seq_printf(s, "Lookup Queue\n");
 	list_for_each_entry(lkb, &res->res_lookup, lkb_rsb_lookup) {
-		seq_printf(s, "%08x %s", lkb->lkb_id,
-			   print_lockmode(lkb->lkb_rqmode));
+		rv = seq_printf(s, "%08x %s", lkb->lkb_id,
+				print_lockmode(lkb->lkb_rqmode));
 		if (lkb->lkb_wait_type)
 			seq_printf(s, " wait_type: %d", lkb->lkb_wait_type);
-		seq_printf(s, "\n");
+		rv = seq_printf(s, "\n");
 	}
  out:
 	unlock_rsb(res);
-	return 0;
+	return rv;
 }
 
-static void print_format2_lock(struct seq_file *s, struct dlm_lkb *lkb,
-			       struct dlm_rsb *r)
+static int print_format2_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			      struct dlm_rsb *r)
 {
 	u64 xid = 0;
 	u64 us;
+	int rv;
 
 	if (lkb->lkb_flags & DLM_IFL_USER) {
 		if (lkb->lkb_ua)
@@ -177,69 +187,82 @@ static void print_format2_lock(struct seq_file *s, struct dlm_lkb *lkb,
 	/* id nodeid remid pid xid exflags flags sts grmode rqmode time_us
 	   r_nodeid r_len r_name */
 
-	seq_printf(s, "%x %d %x %u %llu %x %x %d %d %d %llu %u %d \"%s\"\n",
-		   lkb->lkb_id,
-		   lkb->lkb_nodeid,
-		   lkb->lkb_remid,
-		   lkb->lkb_ownpid,
-		   (unsigned long long)xid,
-		   lkb->lkb_exflags,
-		   lkb->lkb_flags,
-		   lkb->lkb_status,
-		   lkb->lkb_grmode,
-		   lkb->lkb_rqmode,
-		   (unsigned long long)us,
-		   r->res_nodeid,
-		   r->res_length,
-		   r->res_name);
+	rv = seq_printf(s, "%x %d %x %u %llu %x %x %d %d %d %llu %u %d \"%s\"\n",
+			lkb->lkb_id,
+			lkb->lkb_nodeid,
+			lkb->lkb_remid,
+			lkb->lkb_ownpid,
+			(unsigned long long)xid,
+			lkb->lkb_exflags,
+			lkb->lkb_flags,
+			lkb->lkb_status,
+			lkb->lkb_grmode,
+			lkb->lkb_rqmode,
+			(unsigned long long)us,
+			r->res_nodeid,
+			r->res_length,
+			r->res_name);
+	return rv;
 }
 
 static int print_format2(struct dlm_rsb *r, struct seq_file *s)
 {
 	struct dlm_lkb *lkb;
+	int rv = 0;
 
 	lock_rsb(r);
 
-	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue)
-		print_format2_lock(s, lkb, r);
-
-	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue)
-		print_format2_lock(s, lkb, r);
+	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue) {
+		rv = print_format2_lock(s, lkb, r);
+		if (rv)
+			goto out;
+	}
 
-	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue)
-		print_format2_lock(s, lkb, r);
+	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue) {
+		rv = print_format2_lock(s, lkb, r);
+		if (rv)
+			goto out;
+	}
 
+	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue) {
+		rv = print_format2_lock(s, lkb, r);
+		if (rv)
+			goto out;
+	}
+ out:
 	unlock_rsb(r);
-	return 0;
+	return rv;
 }
 
-static void print_format3_lock(struct seq_file *s, struct dlm_lkb *lkb,
-			       int rsb_lookup)
+static int print_format3_lock(struct seq_file *s, struct dlm_lkb *lkb,
+			      int rsb_lookup)
 {
 	u64 xid = 0;
+	int rv;
 
 	if (lkb->lkb_flags & DLM_IFL_USER) {
 		if (lkb->lkb_ua)
 			xid = lkb->lkb_ua->xid;
 	}
 
-	seq_printf(s, "lkb %x %d %x %u %llu %x %x %d %d %d %d %d %d %u %llu %llu\n",
-		   lkb->lkb_id,
-		   lkb->lkb_nodeid,
-		   lkb->lkb_remid,
-		   lkb->lkb_ownpid,
-		   (unsigned long long)xid,
-		   lkb->lkb_exflags,
-		   lkb->lkb_flags,
-		   lkb->lkb_status,
-		   lkb->lkb_grmode,
-		   lkb->lkb_rqmode,
-		   lkb->lkb_highbast,
-		   rsb_lookup,
-		   lkb->lkb_wait_type,
-		   lkb->lkb_lvbseq,
-		   (unsigned long long)ktime_to_ns(lkb->lkb_timestamp),
-		   (unsigned long long)ktime_to_ns(lkb->lkb_time_bast));
+	rv = seq_printf(s, "lkb %x %d %x %u %llu %x %x %d %d %d %d %d %d %u %llu %llu\n",
+			lkb->lkb_id,
+			lkb->lkb_nodeid,
+			lkb->lkb_remid,
+			lkb->lkb_ownpid,
+			(unsigned long long)xid,
+			lkb->lkb_exflags,
+			lkb->lkb_flags,
+			lkb->lkb_status,
+			lkb->lkb_grmode,
+			lkb->lkb_rqmode,
+			lkb->lkb_highbast,
+			rsb_lookup,
+			lkb->lkb_wait_type,
+			lkb->lkb_lvbseq,
+			(unsigned long long)ktime_to_ns(lkb->lkb_timestamp),
+			(unsigned long long)ktime_to_ns(lkb->lkb_time_bast));
+	return rv;
 }
 
 static int print_format3(struct dlm_rsb *r, struct seq_file *s)
@@ -247,18 +270,21 @@ static int print_format3(struct dlm_rsb *r, struct seq_file *s)
 	struct dlm_lkb *lkb;
 	int i, lvblen = r->res_ls->ls_lvblen;
 	int print_name = 1;
+	int rv;
 
 	lock_rsb(r);
 
-	seq_printf(s, "rsb %p %d %x %lx %d %d %u %d ",
-		   r,
-		   r->res_nodeid,
-		   r->res_first_lkid,
-		   r->res_flags,
-		   !list_empty(&r->res_root_list),
-		   !list_empty(&r->res_recover_list),
-		   r->res_recover_locks_count,
-		   r->res_length);
+	rv = seq_printf(s, "rsb %p %d %x %lx %d %d %u %d ",
+			r,
+			r->res_nodeid,
+			r->res_first_lkid,
+			r->res_flags,
+			!list_empty(&r->res_root_list),
+			!list_empty(&r->res_recover_list),
+			r->res_recover_locks_count,
+			r->res_length);
+	if (rv)
+		goto out;
 
 	for (i = 0; i < r->res_length; i++) {
 		if (!isascii(r->res_name[i]) || !isprint(r->res_name[i]))
@@ -273,7 +299,9 @@ static int print_format3(struct dlm_rsb *r, struct seq_file *s)
 		else
 			seq_printf(s, " %02x", (unsigned char)r->res_name[i]);
 	}
-	seq_printf(s, "\n");
+	rv = seq_printf(s, "\n");
+	if (rv)
+		goto out;
 
 	if (!r->res_lvbptr)
 		goto do_locks;
@@ -282,344 +310,294 @@ static int print_format3(struct dlm_rsb *r, struct seq_file *s)
 
 	for (i = 0; i < lvblen; i++)
 		seq_printf(s, " %02x", (unsigned char)r->res_lvbptr[i]);
-	seq_printf(s, "\n");
+	rv = seq_printf(s, "\n");
+	if (rv)
+		goto out;
 
  do_locks:
-	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue)
-		print_format3_lock(s, lkb, 0);
-
-	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue)
-		print_format3_lock(s, lkb, 0);
-
-	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue)
-		print_format3_lock(s, lkb, 0);
-
-	list_for_each_entry(lkb, &r->res_lookup, lkb_rsb_lookup)
-		print_format3_lock(s, lkb, 1);
-
-	unlock_rsb(r);
-	return 0;
-}
-
-static int rsb_iter_next(struct rsb_iter *ri)
-{
-	struct dlm_ls *ls = ri->ls;
-	int i;
-
-	if (!ri->next) {
- top:
-		/* Find the next non-empty hash bucket */
-		for (i = ri->entry; i < ls->ls_rsbtbl_size; i++) {
-			read_lock(&ls->ls_rsbtbl[i].lock);
-			if (!list_empty(&ls->ls_rsbtbl[i].list)) {
-				ri->next = ls->ls_rsbtbl[i].list.next;
-				ri->rsb = list_entry(ri->next, struct dlm_rsb,
-							res_hashchain);
-				dlm_hold_rsb(ri->rsb);
-				read_unlock(&ls->ls_rsbtbl[i].lock);
-				break;
-			}
-			read_unlock(&ls->ls_rsbtbl[i].lock);
-		}
-		ri->entry = i;
-
-		if (ri->entry >= ls->ls_rsbtbl_size)
-			return 1;
-	} else {
-		struct dlm_rsb *old = ri->rsb;
-		i = ri->entry;
-		read_lock(&ls->ls_rsbtbl[i].lock);
-		ri->next = ri->next->next;
-		if (ri->next->next == ls->ls_rsbtbl[i].list.next) {
-			/* End of list - move to next bucket */
-			ri->next = NULL;
-			ri->entry++;
-			read_unlock(&ls->ls_rsbtbl[i].lock);
-			dlm_put_rsb(old);
-			goto top;
-		}
-		ri->rsb = list_entry(ri->next, struct dlm_rsb, res_hashchain);
-		dlm_hold_rsb(ri->rsb);
-		read_unlock(&ls->ls_rsbtbl[i].lock);
-		dlm_put_rsb(old);
+	list_for_each_entry(lkb, &r->res_grantqueue, lkb_statequeue) {
+		rv = print_format3_lock(s, lkb, 0);
+		if (rv)
+			goto out;
 	}
 
-	return 0;
-}
-
-static void rsb_iter_free(struct rsb_iter *ri)
-{
-	kfree(ri);
-}
-
-static struct rsb_iter *rsb_iter_init(struct dlm_ls *ls)
-{
-	struct rsb_iter *ri;
-
-	ri = kzalloc(sizeof *ri, GFP_KERNEL);
-	if (!ri)
-		return NULL;
-
-	ri->ls = ls;
-	ri->entry = 0;
-	ri->next = NULL;
-	ri->format = 1;
-
-	if (rsb_iter_next(ri)) {
-		rsb_iter_free(ri);
-		return NULL;
+	list_for_each_entry(lkb, &r->res_convertqueue, lkb_statequeue) {
+		rv = print_format3_lock(s, lkb, 0);
+		if (rv)
+			goto out;
 	}
 
-	return ri;
-}
-
-static void *rsb_seq_start(struct seq_file *file, loff_t *pos)
-{
-	struct rsb_iter *ri;
-	loff_t n = *pos;
-
-	ri = rsb_iter_init(file->private);
-	if (!ri)
-		return NULL;
-
-	while (n--) {
-		if (rsb_iter_next(ri)) {
-			rsb_iter_free(ri);
-			return NULL;
-		}
+	list_for_each_entry(lkb, &r->res_waitqueue, lkb_statequeue) {
+		rv = print_format3_lock(s, lkb, 0);
+		if (rv)
+			goto out;
 	}
 
-	return ri;
-}
-
-static void *rsb_seq_next(struct seq_file *file, void *iter_ptr, loff_t *pos)
-{
-	struct rsb_iter *ri = iter_ptr;
-
-	(*pos)++;
-
-	if (rsb_iter_next(ri)) {
-		rsb_iter_free(ri);
-		return NULL;
+	list_for_each_entry(lkb, &r->res_lookup, lkb_rsb_lookup) {
+		rv = print_format3_lock(s, lkb, 1);
+		if (rv)
+			goto out;
 	}
-
-	return ri;
+ out:
+	unlock_rsb(r);
+	return rv;
 }
 
-static void rsb_seq_stop(struct seq_file *file, void *iter_ptr)
-{
-	/* nothing for now */
-}
+struct rsbtbl_iter {
+	struct dlm_rsb *rsb;
+	unsigned bucket;
+	int format;
+	int header;
+};
 
-static int rsb_seq_show(struct seq_file *file, void *iter_ptr)
+/* seq_printf returns -1 if the buffer is full, and 0 otherwise.
+   If the buffer is full, seq_printf can be called again, but it
+   does nothing and just returns -1.  So, the these printing routines
+   periodically check the return value to avoid wasting too much time
+   trying to print to a full buffer. */
+
+static int table_seq_show(struct seq_file *seq, void *iter_ptr)
 {
-	struct rsb_iter *ri = iter_ptr;
+	struct rsbtbl_iter *ri = iter_ptr;
+	int rv = 0;
 
 	switch (ri->format) {
 	case 1:
-		print_format1(ri->rsb, file);
+		rv = print_format1(ri->rsb, seq);
 		break;
 	case 2:
 		if (ri->header) {
-			seq_printf(file, "id nodeid remid pid xid exflags "
-					 "flags sts grmode rqmode time_ms "
-					 "r_nodeid r_len r_name\n");
+			seq_printf(seq, "id nodeid remid pid xid exflags "
+					"flags sts grmode rqmode time_ms "
+					"r_nodeid r_len r_name\n");
 			ri->header = 0;
 		}
-		print_format2(ri->rsb, file);
+		rv = print_format2(ri->rsb, seq);
 		break;
 	case 3:
 		if (ri->header) {
-			seq_printf(file, "version rsb 1.1 lvb 1.1 lkb 1.1\n");
+			seq_printf(seq, "version rsb 1.1 lvb 1.1 lkb 1.1\n");
 			ri->header = 0;
 		}
-		print_format3(ri->rsb, file);
+		rv = print_format3(ri->rsb, seq);
 		break;
 	}
 
-	return 0;
+	return rv;
 }
 
-static struct seq_operations rsb_seq_ops = {
-	.start = rsb_seq_start,
-	.next  = rsb_seq_next,
-	.stop  = rsb_seq_stop,
-	.show  = rsb_seq_show,
-};
+static struct seq_operations format1_seq_ops;
+static struct seq_operations format2_seq_ops;
+static struct seq_operations format3_seq_ops;
 
-static int rsb_open(struct inode *inode, struct file *file)
+static void *table_seq_start(struct seq_file *seq, loff_t *pos)
 {
-	struct seq_file *seq;
-	int ret;
-
-	ret = seq_open(file, &rsb_seq_ops);
-	if (ret)
-		return ret;
-
-	seq = file->private_data;
-	seq->private = inode->i_private;
-
-	return 0;
-}
-
-static const struct file_operations rsb_fops = {
-	.owner   = THIS_MODULE,
-	.open    = rsb_open,
-	.read    = seq_read,
-	.llseek  = seq_lseek,
-	.release = seq_release
-};
+	struct dlm_ls *ls = seq->private;
+	struct rsbtbl_iter *ri;
+	struct dlm_rsb *r;
+	loff_t n = *pos;
+	unsigned bucket, entry;
 
-/*
- * Dump state in compact per-lock listing
- */
+	bucket = n >> 32;
+	entry = n & ((1LL << 32) - 1);
 
-static struct rsb_iter *locks_iter_init(struct dlm_ls *ls, loff_t *pos)
-{
-	struct rsb_iter *ri;
+	if (bucket >= ls->ls_rsbtbl_size)
+		return NULL;
 
-	ri = kzalloc(sizeof *ri, GFP_KERNEL);
+	ri = kzalloc(sizeof(struct rsbtbl_iter), GFP_KERNEL);
 	if (!ri)
 		return NULL;
-
-	ri->ls = ls;
-	ri->entry = 0;
-	ri->next = NULL;
-	ri->format = 2;
-
-	if (*pos == 0)
+	if (n == 0)
 		ri->header = 1;
-
-	if (rsb_iter_next(ri)) {
-		rsb_iter_free(ri);
-		return NULL;
+	if (seq->op == &format1_seq_ops)
+		ri->format = 1;
+	if (seq->op == &format2_seq_ops)
+		ri->format = 2;
+	if (seq->op == &format3_seq_ops)
+		ri->format = 3;
+
+	read_lock(&ls->ls_rsbtbl[bucket].lock);
+	if (!list_empty(&ls->ls_rsbtbl[bucket].list)) {
+		list_for_each_entry(r, &ls->ls_rsbtbl[bucket].list,
+				    res_hashchain) {
+			if (!entry--) {
+				dlm_hold_rsb(r);
+				ri->rsb = r;
+				ri->bucket = bucket;
+				read_unlock(&ls->ls_rsbtbl[bucket].lock);
+				return ri;
+			}
+		}
 	}
+	read_unlock(&ls->ls_rsbtbl[bucket].lock);
 
-	return ri;
-}
+	/*
+	 * move to the first rsb in the next non-empty bucket
+	 */
 
-static void *locks_seq_start(struct seq_file *file, loff_t *pos)
-{
-	struct rsb_iter *ri;
-	loff_t n = *pos;
+	/* zero the entry */
+	n &= ~((1LL << 32) - 1);
 
-	ri = locks_iter_init(file->private, pos);
-	if (!ri)
-		return NULL;
+	while (1) {
+		bucket++;
+		n += 1LL << 32;
 
-	while (n--) {
-		if (rsb_iter_next(ri)) {
-			rsb_iter_free(ri);
+		if (bucket >= ls->ls_rsbtbl_size) {
+			kfree(ri);
 			return NULL;
 		}
-	}
 
-	return ri;
+		read_lock(&ls->ls_rsbtbl[bucket].lock);
+		if (!list_empty(&ls->ls_rsbtbl[bucket].list)) {
+			r = list_first_entry(&ls->ls_rsbtbl[bucket].list,
+					     struct dlm_rsb, res_hashchain);
+			dlm_hold_rsb(r);
+			ri->rsb = r;
+			ri->bucket = bucket;
+			read_unlock(&ls->ls_rsbtbl[bucket].lock);
+			*pos = n;
+			return ri;
+		}
+		read_unlock(&ls->ls_rsbtbl[bucket].lock);
+	}
 }
 
-static struct seq_operations locks_seq_ops = {
-	.start = locks_seq_start,
-	.next  = rsb_seq_next,
-	.stop  = rsb_seq_stop,
-	.show  = rsb_seq_show,
-};
-
-static int locks_open(struct inode *inode, struct file *file)
+static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)
 {
-	struct seq_file *seq;
-	int ret;
-
-	ret = seq_open(file, &locks_seq_ops);
-	if (ret)
-		return ret;
-
-	seq = file->private_data;
-	seq->private = inode->i_private;
-
-	return 0;
-}
-
-static const struct file_operations locks_fops = {
-	.owner   = THIS_MODULE,
-	.open    = locks_open,
-	.read    = seq_read,
-	.llseek  = seq_lseek,
-	.release = seq_release
-};
-
-/*
- * Dump all rsb/lvb/lkb state in compact listing, more complete than _locks
- * This can replace both formats 1 and 2 eventually.
- */
+	struct dlm_ls *ls = seq->private;
+	struct rsbtbl_iter *ri = iter_ptr;
+	struct list_head *next;
+	struct dlm_rsb *r, *rp;
+	loff_t n = *pos;
+	unsigned bucket;
+
+	bucket = n >> 32;
+
+	/*
+	 * move to the next rsb in the same bucket
+	 */
+
+	read_lock(&ls->ls_rsbtbl[bucket].lock);
+	rp = ri->rsb;
+	next = rp->res_hashchain.next;
+
+	if (next != &ls->ls_rsbtbl[bucket].list) {
+		r = list_entry(next, struct dlm_rsb, res_hashchain);
+		dlm_hold_rsb(r);
+		ri->rsb = r;
+		read_unlock(&ls->ls_rsbtbl[bucket].lock);
+		dlm_put_rsb(rp);
+		++*pos;
+		return ri;
+	}
+	read_unlock(&ls->ls_rsbtbl[bucket].lock);
+	dlm_put_rsb(rp);
 
-static struct rsb_iter *all_iter_init(struct dlm_ls *ls, loff_t *pos)
-{
-	struct rsb_iter *ri;
+	/*
+	 * move to the first rsb in the next non-empty bucket
+	 */
 
-	ri = kzalloc(sizeof *ri, GFP_KERNEL);
-	if (!ri)
-		return NULL;
+	/* zero the entry */
+	n &= ~((1LL << 32) - 1);
 
-	ri->ls = ls;
-	ri->entry = 0;
-	ri->next = NULL;
-	ri->format = 3;
+	while (1) {
+		bucket++;
+		n += 1LL << 32;
 
-	if (*pos == 0)
-		ri->header = 1;
+		if (bucket >= ls->ls_rsbtbl_size) {
+			kfree(ri);
+			return NULL;
+		}
 
-	if (rsb_iter_next(ri)) {
-		rsb_iter_free(ri);
-		return NULL;
+		read_lock(&ls->ls_rsbtbl[bucket].lock);
+		if (!list_empty(&ls->ls_rsbtbl[bucket].list)) {
+			r = list_first_entry(&ls->ls_rsbtbl[bucket].list,
+					     struct dlm_rsb, res_hashchain);
+			dlm_hold_rsb(r);
+			ri->rsb = r;
+			ri->bucket = bucket;
+			read_unlock(&ls->ls_rsbtbl[bucket].lock);
+			*pos = n;
+			return ri;
+		}
+		read_unlock(&ls->ls_rsbtbl[bucket].lock);
 	}
-
-	return ri;
 }
 
-static void *all_seq_start(struct seq_file *file, loff_t *pos)
+static void table_seq_stop(struct seq_file *seq, void *iter_ptr)
 {
-	struct rsb_iter *ri;
-	loff_t n = *pos;
-
-	ri = all_iter_init(file->private, pos);
-	if (!ri)
-		return NULL;
+	struct rsbtbl_iter *ri = iter_ptr;
 
-	while (n--) {
-		if (rsb_iter_next(ri)) {
-			rsb_iter_free(ri);
-			return NULL;
-		}
+	if (ri) {
+		dlm_put_rsb(ri->rsb);
+		kfree(ri);
 	}
-
-	return ri;
 }
 
-static struct seq_operations all_seq_ops = {
-	.start = all_seq_start,
-	.next  = rsb_seq_next,
-	.stop  = rsb_seq_stop,
-	.show  = rsb_seq_show,
+static struct seq_operations format1_seq_ops = {
+	.start = table_seq_start,
+	.next  = table_seq_next,
+	.stop  = table_seq_stop,
+	.show  = table_seq_show,
 };
 
-static int all_open(struct inode *inode, struct file *file)
+static struct seq_operations format2_seq_ops = {
+	.start = table_seq_start,
+	.next  = table_seq_next,
+	.stop  = table_seq_stop,
+	.show  = table_seq_show,
+};
+
+static struct seq_operations format3_seq_ops = {
+	.start = table_seq_start,
+	.next  = table_seq_next,
+	.stop  = table_seq_stop,
+	.show  = table_seq_show,
+};
+
+static const struct file_operations format1_fops;
+static const struct file_operations format2_fops;
+static const struct file_operations format3_fops;
+
+static int table_open(struct inode *inode, struct file *file)
 {
 	struct seq_file *seq;
-	int ret;
+	int ret = -1;
+
+	if (file->f_op == &format1_fops)
+		ret = seq_open(file, &format1_seq_ops);
+	else if (file->f_op == &format2_fops)
+		ret = seq_open(file, &format2_seq_ops);
+	else if (file->f_op == &format3_fops)
+		ret = seq_open(file, &format3_seq_ops);
 
-	ret = seq_open(file, &all_seq_ops);
 	if (ret)
 		return ret;
 
 	seq = file->private_data;
-	seq->private = inode->i_private;
-
+	seq->private = inode->i_private; /* the dlm_ls */
 	return 0;
 }
 
-static const struct file_operations all_fops = {
+static const struct file_operations format1_fops = {
+	.owner   = THIS_MODULE,
+	.open    = table_open,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+static const struct file_operations format2_fops = {
+	.owner   = THIS_MODULE,
+	.open    = table_open,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.release = seq_release
+};
+
+static const struct file_operations format3_fops = {
 	.owner   = THIS_MODULE,
-	.open    = all_open,
+	.open    = table_open,
 	.read    = seq_read,
 	.llseek  = seq_lseek,
 	.release = seq_release
@@ -689,7 +667,7 @@ int dlm_create_debug_file(struct dlm_ls *ls)
 						      S_IFREG | S_IRUGO,
 						      dlm_root,
 						      ls,
-						      &rsb_fops);
+						      &format1_fops);
 	if (!ls->ls_debug_rsb_dentry)
 		goto fail;
 
@@ -702,7 +680,7 @@ int dlm_create_debug_file(struct dlm_ls *ls)
 							S_IFREG | S_IRUGO,
 							dlm_root,
 							ls,
-							&locks_fops);
+							&format2_fops);
 	if (!ls->ls_debug_locks_dentry)
 		goto fail;
 
@@ -715,7 +693,7 @@ int dlm_create_debug_file(struct dlm_ls *ls)
 						      S_IFREG | S_IRUGO,
 						      dlm_root,
 						      ls,
-						      &all_fops);
+						      &format3_fops);
 	if (!ls->ls_debug_all_dentry)
 		goto fail;
 
-- 
cgit v1.2.3


From c7be761a8163d2f1ac0b606c21e4316b7abc5af7 Mon Sep 17 00:00:00 2001
From: David Teigland <teigland@redhat.com>
Date: Wed, 7 Jan 2009 16:50:41 -0600
Subject: dlm: change rsbtbl rwlock to spinlock

The rwlock is almost always used in write mode, so there's no reason
to not use a spinlock instead.

Signed-off-by: David Teigland <teigland@redhat.com>
---
 fs/dlm/debug_fs.c     | 24 ++++++++++++------------
 fs/dlm/dlm_internal.h |  2 +-
 fs/dlm/lock.c         | 26 +++++++++++++-------------
 fs/dlm/lockspace.c    |  2 +-
 fs/dlm/recover.c      | 10 +++++-----
 5 files changed, 32 insertions(+), 32 deletions(-)

(limited to 'fs')

diff --git a/fs/dlm/debug_fs.c b/fs/dlm/debug_fs.c
index bc4af3ef65a..1d1d2744223 100644
--- a/fs/dlm/debug_fs.c
+++ b/fs/dlm/debug_fs.c
@@ -416,7 +416,7 @@ static void *table_seq_start(struct seq_file *seq, loff_t *pos)
 	if (seq->op == &format3_seq_ops)
 		ri->format = 3;
 
-	read_lock(&ls->ls_rsbtbl[bucket].lock);
+	spin_lock(&ls->ls_rsbtbl[bucket].lock);
 	if (!list_empty(&ls->ls_rsbtbl[bucket].list)) {
 		list_for_each_entry(r, &ls->ls_rsbtbl[bucket].list,
 				    res_hashchain) {
@@ -424,12 +424,12 @@ static void *table_seq_start(struct seq_file *seq, loff_t *pos)
 				dlm_hold_rsb(r);
 				ri->rsb = r;
 				ri->bucket = bucket;
-				read_unlock(&ls->ls_rsbtbl[bucket].lock);
+				spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 				return ri;
 			}
 		}
 	}
-	read_unlock(&ls->ls_rsbtbl[bucket].lock);
+	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 
 	/*
 	 * move to the first rsb in the next non-empty bucket
@@ -447,18 +447,18 @@ static void *table_seq_start(struct seq_file *seq, loff_t *pos)
 			return NULL;
 		}
 
-		read_lock(&ls->ls_rsbtbl[bucket].lock);
+		spin_lock(&ls->ls_rsbtbl[bucket].lock);
 		if (!list_empty(&ls->ls_rsbtbl[bucket].list)) {
 			r = list_first_entry(&ls->ls_rsbtbl[bucket].list,
 					     struct dlm_rsb, res_hashchain);
 			dlm_hold_rsb(r);
 			ri->rsb = r;
 			ri->bucket = bucket;
-			read_unlock(&ls->ls_rsbtbl[bucket].lock);
+			spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 			*pos = n;
 			return ri;
 		}
-		read_unlock(&ls->ls_rsbtbl[bucket].lock);
+		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 	}
 }
 
@@ -477,7 +477,7 @@ static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)
 	 * move to the next rsb in the same bucket
 	 */
 
-	read_lock(&ls->ls_rsbtbl[bucket].lock);
+	spin_lock(&ls->ls_rsbtbl[bucket].lock);
 	rp = ri->rsb;
 	next = rp->res_hashchain.next;
 
@@ -485,12 +485,12 @@ static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)
 		r = list_entry(next, struct dlm_rsb, res_hashchain);
 		dlm_hold_rsb(r);
 		ri->rsb = r;
-		read_unlock(&ls->ls_rsbtbl[bucket].lock);
+		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 		dlm_put_rsb(rp);
 		++*pos;
 		return ri;
 	}
-	read_unlock(&ls->ls_rsbtbl[bucket].lock);
+	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 	dlm_put_rsb(rp);
 
 	/*
@@ -509,18 +509,18 @@ static void *table_seq_next(struct seq_file *seq, void *iter_ptr, loff_t *pos)
 			return NULL;
 		}
 
-		read_lock(&ls->ls_rsbtbl[bucket].lock);
+		spin_lock(&ls->ls_rsbtbl[bucket].lock);
 		if (!list_empty(&ls->ls_rsbtbl[bucket].list)) {
 			r = list_first_entry(&ls->ls_rsbtbl[bucket].list,
 					     struct dlm_rsb, res_hashchain);
 			dlm_hold_rsb(r);
 			ri->rsb = r;
 			ri->bucket = bucket;
-			read_unlock(&ls->ls_rsbtbl[bucket].lock);
+			spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 			*pos = n;
 			return ri;
 		}
-		read_unlock(&ls->ls_rsbtbl[bucket].lock);
+		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 	}
 }
 
diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h
index ef2f1e35396..076e86f38bc 100644
--- a/fs/dlm/dlm_internal.h
+++ b/fs/dlm/dlm_internal.h
@@ -105,7 +105,7 @@ struct dlm_dirtable {
 struct dlm_rsbtable {
 	struct list_head	list;
 	struct list_head	toss;
-	rwlock_t		lock;
+	spinlock_t		lock;
 };
 
 struct dlm_lkbtable {
diff --git a/fs/dlm/lock.c b/fs/dlm/lock.c
index 6cfe65bbf4a..01e7d39c5fb 100644
--- a/fs/dlm/lock.c
+++ b/fs/dlm/lock.c
@@ -412,9 +412,9 @@ static int search_rsb(struct dlm_ls *ls, char *name, int len, int b,
 		      unsigned int flags, struct dlm_rsb **r_ret)
 {
 	int error;
-	write_lock(&ls->ls_rsbtbl[b].lock);
+	spin_lock(&ls->ls_rsbtbl[b].lock);
 	error = _search_rsb(ls, name, len, b, flags, r_ret);
-	write_unlock(&ls->ls_rsbtbl[b].lock);
+	spin_unlock(&ls->ls_rsbtbl[b].lock);
 	return error;
 }
 
@@ -478,16 +478,16 @@ static int find_rsb(struct dlm_ls *ls, char *name, int namelen,
 		r->res_nodeid = nodeid;
 	}
 
-	write_lock(&ls->ls_rsbtbl[bucket].lock);
+	spin_lock(&ls->ls_rsbtbl[bucket].lock);
 	error = _search_rsb(ls, name, namelen, bucket, 0, &tmp);
 	if (!error) {
-		write_unlock(&ls->ls_rsbtbl[bucket].lock);
+		spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 		dlm_free_rsb(r);
 		r = tmp;
 		goto out;
 	}
 	list_add(&r->res_hashchain, &ls->ls_rsbtbl[bucket].list);
-	write_unlock(&ls->ls_rsbtbl[bucket].lock);
+	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 	error = 0;
  out:
 	*r_ret = r;
@@ -530,9 +530,9 @@ static void put_rsb(struct dlm_rsb *r)
 	struct dlm_ls *ls = r->res_ls;
 	uint32_t bucket = r->res_bucket;
 
-	write_lock(&ls->ls_rsbtbl[bucket].lock);
+	spin_lock(&ls->ls_rsbtbl[bucket].lock);
 	kref_put(&r->res_ref, toss_rsb);
-	write_unlock(&ls->ls_rsbtbl[bucket].lock);
+	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 }
 
 void dlm_put_rsb(struct dlm_rsb *r)
@@ -967,7 +967,7 @@ static int shrink_bucket(struct dlm_ls *ls, int b)
 
 	for (;;) {
 		found = 0;
-		write_lock(&ls->ls_rsbtbl[b].lock);
+		spin_lock(&ls->ls_rsbtbl[b].lock);
 		list_for_each_entry_reverse(r, &ls->ls_rsbtbl[b].toss,
 					    res_hashchain) {
 			if (!time_after_eq(jiffies, r->res_toss_time +
@@ -978,20 +978,20 @@ static int shrink_bucket(struct dlm_ls *ls, int b)
 		}
 
 		if (!found) {
-			write_unlock(&ls->ls_rsbtbl[b].lock);
+			spin_unlock(&ls->ls_rsbtbl[b].lock);
 			break;
 		}
 
 		if (kref_put(&r->res_ref, kill_rsb)) {
 			list_del(&r->res_hashchain);
-			write_unlock(&ls->ls_rsbtbl[b].lock);
+			spin_unlock(&ls->ls_rsbtbl[b].lock);
 
 			if (is_master(r))
 				dir_remove(r);
 			dlm_free_rsb(r);
 			count++;
 		} else {
-			write_unlock(&ls->ls_rsbtbl[b].lock);
+			spin_unlock(&ls->ls_rsbtbl[b].lock);
 			log_error(ls, "tossed rsb in use %s", r->res_name);
 		}
 	}
@@ -4224,7 +4224,7 @@ static struct dlm_rsb *find_purged_rsb(struct dlm_ls *ls, int bucket)
 {
 	struct dlm_rsb *r, *r_ret = NULL;
 
-	read_lock(&ls->ls_rsbtbl[bucket].lock);
+	spin_lock(&ls->ls_rsbtbl[bucket].lock);
 	list_for_each_entry(r, &ls->ls_rsbtbl[bucket].list, res_hashchain) {
 		if (!rsb_flag(r, RSB_LOCKS_PURGED))
 			continue;
@@ -4233,7 +4233,7 @@ static struct dlm_rsb *find_purged_rsb(struct dlm_ls *ls, int bucket)
 		r_ret = r;
 		break;
 	}
-	read_unlock(&ls->ls_rsbtbl[bucket].lock);
+	spin_unlock(&ls->ls_rsbtbl[bucket].lock);
 	return r_ret;
 }
 
diff --git a/fs/dlm/lockspace.c b/fs/dlm/lockspace.c
index 8d86b7960f0..aa32e5f0249 100644
--- a/fs/dlm/lockspace.c
+++ b/fs/dlm/lockspace.c
@@ -464,7 +464,7 @@ static int new_lockspace(char *name, int namelen, void **lockspace,
 	for (i = 0; i < size; i++) {
 		INIT_LIST_HEAD(&ls->ls_rsbtbl[i].list);
 		INIT_LIST_HEAD(&ls->ls_rsbtbl[i].toss);
-		rwlock_init(&ls->ls_rsbtbl[i].lock);
+		spin_lock_init(&ls->ls_rsbtbl[i].lock);
 	}
 
 	size = dlm_config.ci_lkbtbl_size;
diff --git a/fs/dlm/recover.c b/fs/dlm/recover.c
index 80aba5bdd4a..eda43f36261 100644
--- a/fs/dlm/recover.c
+++ b/fs/dlm/recover.c
@@ -726,7 +726,7 @@ int dlm_create_root_list(struct dlm_ls *ls)
 	}
 
 	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		read_lock(&ls->ls_rsbtbl[i].lock);
+		spin_lock(&ls->ls_rsbtbl[i].lock);
 		list_for_each_entry(r, &ls->ls_rsbtbl[i].list, res_hashchain) {
 			list_add(&r->res_root_list, &ls->ls_root_list);
 			dlm_hold_rsb(r);
@@ -737,7 +737,7 @@ int dlm_create_root_list(struct dlm_ls *ls)
 		   but no other recovery steps should do anything with them. */
 
 		if (dlm_no_directory(ls)) {
-			read_unlock(&ls->ls_rsbtbl[i].lock);
+			spin_unlock(&ls->ls_rsbtbl[i].lock);
 			continue;
 		}
 
@@ -745,7 +745,7 @@ int dlm_create_root_list(struct dlm_ls *ls)
 			list_add(&r->res_root_list, &ls->ls_root_list);
 			dlm_hold_rsb(r);
 		}
-		read_unlock(&ls->ls_rsbtbl[i].lock);
+		spin_unlock(&ls->ls_rsbtbl[i].lock);
 	}
  out:
 	up_write(&ls->ls_root_sem);
@@ -775,7 +775,7 @@ void dlm_clear_toss_list(struct dlm_ls *ls)
 	int i;
 
 	for (i = 0; i < ls->ls_rsbtbl_size; i++) {
-		write_lock(&ls->ls_rsbtbl[i].lock);
+		spin_lock(&ls->ls_rsbtbl[i].lock);
 		list_for_each_entry_safe(r, safe, &ls->ls_rsbtbl[i].toss,
 					 res_hashchain) {
 			if (dlm_no_directory(ls) || !is_master(r)) {
@@ -783,7 +783,7 @@ void dlm_clear_toss_list(struct dlm_ls *ls)
 				dlm_free_rsb(r);
 			}
 		}
-		write_unlock(&ls->ls_rsbtbl[i].lock);
+		spin_unlock(&ls->ls_rsbtbl[i].lock);
 	}
 }
 
-- 
cgit v1.2.3