aboutsummaryrefslogtreecommitdiff
path: root/fs/ocfs2/cluster
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ocfs2/cluster')
-rw-r--r--fs/ocfs2/cluster/Makefile2
-rw-r--r--fs/ocfs2/cluster/heartbeat.c391
-rw-r--r--fs/ocfs2/cluster/masklog.c20
-rw-r--r--fs/ocfs2/cluster/masklog.h108
-rw-r--r--fs/ocfs2/cluster/netdebug.c102
-rw-r--r--fs/ocfs2/cluster/nodemanager.c4
-rw-r--r--fs/ocfs2/cluster/quorum.c8
-rw-r--r--fs/ocfs2/cluster/sys.c11
-rw-r--r--fs/ocfs2/cluster/tcp.c389
-rw-r--r--fs/ocfs2/cluster/tcp.h2
-rw-r--r--fs/ocfs2/cluster/tcp_internal.h2
-rw-r--r--fs/ocfs2/cluster/ver.c42
-rw-r--r--fs/ocfs2/cluster/ver.h31
13 files changed, 522 insertions, 590 deletions
diff --git a/fs/ocfs2/cluster/Makefile b/fs/ocfs2/cluster/Makefile
index bc8c5e7d860..1aefc0350ec 100644
--- a/fs/ocfs2/cluster/Makefile
+++ b/fs/ocfs2/cluster/Makefile
@@ -1,4 +1,4 @@
obj-$(CONFIG_OCFS2_FS) += ocfs2_nodemanager.o
ocfs2_nodemanager-objs := heartbeat.o masklog.o sys.o nodemanager.o \
- quorum.o tcp.o netdebug.o ver.o
+ quorum.o tcp.o netdebug.o
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index b108e863d8f..73039295d0d 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -35,6 +35,7 @@
#include <linux/time.h>
#include <linux/debugfs.h>
#include <linux/slab.h>
+#include <linux/bitmap.h>
#include "heartbeat.h"
#include "tcp.h"
@@ -176,7 +177,7 @@ static void o2hb_dead_threshold_set(unsigned int threshold)
}
}
-static int o2hb_global_hearbeat_mode_set(unsigned int hb_mode)
+static int o2hb_global_heartbeat_mode_set(unsigned int hb_mode)
{
int ret = -1;
@@ -216,6 +217,7 @@ struct o2hb_region {
struct list_head hr_all_item;
unsigned hr_unclean_stop:1,
+ hr_aborted_start:1,
hr_item_pinned:1,
hr_item_dropped:1;
@@ -254,6 +256,10 @@ struct o2hb_region {
* a more complete api that doesn't lead to this sort of fragility. */
atomic_t hr_steady_iterations;
+ /* terminate o2hb thread if it does not reach steady state
+ * (hr_steady_iterations == 0) within hr_unsteady_iterations */
+ atomic_t hr_unsteady_iterations;
+
char hr_dev_name[BDEVNAME_SIZE];
unsigned int hr_timeout_ms;
@@ -277,15 +283,6 @@ struct o2hb_bio_wait_ctxt {
int wc_error;
};
-static int o2hb_pop_count(void *map, int count)
-{
- int i = -1, pop = 0;
-
- while ((i = find_next_bit(map, count, i + 1)) < count)
- pop++;
- return pop;
-}
-
static void o2hb_write_timeout(struct work_struct *work)
{
int failed, quorum;
@@ -302,9 +299,9 @@ static void o2hb_write_timeout(struct work_struct *work)
spin_lock_irqsave(&o2hb_live_lock, flags);
if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
set_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
- failed = o2hb_pop_count(&o2hb_failed_region_bitmap,
+ failed = bitmap_weight(o2hb_failed_region_bitmap,
O2NM_MAX_REGIONS);
- quorum = o2hb_pop_count(&o2hb_quorum_region_bitmap,
+ quorum = bitmap_weight(o2hb_quorum_region_bitmap,
O2NM_MAX_REGIONS);
spin_unlock_irqrestore(&o2hb_live_lock, flags);
@@ -324,6 +321,10 @@ static void o2hb_write_timeout(struct work_struct *work)
static void o2hb_arm_write_timeout(struct o2hb_region *reg)
{
+ /* Arm writeout only after thread reaches steady state */
+ if (atomic_read(&reg->hr_steady_iterations) != 0)
+ return;
+
mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n",
O2HB_MAX_WRITE_TIMEOUT_MS);
@@ -367,11 +368,7 @@ static inline void o2hb_bio_wait_dec(struct o2hb_bio_wait_ctxt *wc,
static void o2hb_wait_on_io(struct o2hb_region *reg,
struct o2hb_bio_wait_ctxt *wc)
{
- struct address_space *mapping = reg->hr_bdev->bd_inode->i_mapping;
-
- blk_run_address_space(mapping);
o2hb_bio_wait_dec(wc, 1);
-
wait_for_completion(&wc->wc_io_complete);
}
@@ -416,7 +413,7 @@ static struct bio *o2hb_setup_one_bio(struct o2hb_region *reg,
}
/* Must put everything in 512 byte sectors for the bio... */
- bio->bi_sector = (reg->hr_start_block + cs) << (bits - 9);
+ bio->bi_iter.bi_sector = (reg->hr_start_block + cs) << (bits - 9);
bio->bi_bdev = reg->hr_bdev;
bio->bi_private = wc;
bio->bi_end_io = o2hb_bio_end_io;
@@ -495,7 +492,7 @@ static int o2hb_issue_node_write(struct o2hb_region *reg,
}
atomic_inc(&write_wc->wc_num_reqs);
- submit_bio(WRITE, bio);
+ submit_bio(WRITE_SYNC, bio);
status = 0;
bail:
@@ -541,27 +538,50 @@ static int o2hb_verify_crc(struct o2hb_region *reg,
return read == computed;
}
-/* We want to make sure that nobody is heartbeating on top of us --
- * this will help detect an invalid configuration. */
-static int o2hb_check_last_timestamp(struct o2hb_region *reg)
+/*
+ * Compare the slot data with what we wrote in the last iteration.
+ * If the match fails, print an appropriate error message. This is to
+ * detect errors like... another node hearting on the same slot,
+ * flaky device that is losing writes, etc.
+ * Returns 1 if check succeeds, 0 otherwise.
+ */
+static int o2hb_check_own_slot(struct o2hb_region *reg)
{
- int node_num, ret;
struct o2hb_disk_slot *slot;
struct o2hb_disk_heartbeat_block *hb_block;
+ char *errstr;
- node_num = o2nm_this_node();
-
- ret = 1;
- slot = &reg->hr_slots[node_num];
+ slot = &reg->hr_slots[o2nm_this_node()];
/* Don't check on our 1st timestamp */
- if (slot->ds_last_time) {
- hb_block = slot->ds_raw_block;
+ if (!slot->ds_last_time)
+ return 0;
- if (le64_to_cpu(hb_block->hb_seq) != slot->ds_last_time)
- ret = 0;
- }
+ hb_block = slot->ds_raw_block;
+ if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time &&
+ le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation &&
+ hb_block->hb_node == slot->ds_node_num)
+ return 1;
+
+#define ERRSTR1 "Another node is heartbeating on device"
+#define ERRSTR2 "Heartbeat generation mismatch on device"
+#define ERRSTR3 "Heartbeat sequence mismatch on device"
+
+ if (hb_block->hb_node != slot->ds_node_num)
+ errstr = ERRSTR1;
+ else if (le64_to_cpu(hb_block->hb_generation) !=
+ slot->ds_last_generation)
+ errstr = ERRSTR2;
+ else
+ errstr = ERRSTR3;
- return ret;
+ mlog(ML_ERROR, "%s (%s): expected(%u:0x%llx, 0x%llx), "
+ "ondisk(%u:0x%llx, 0x%llx)\n", errstr, reg->hr_dev_name,
+ slot->ds_node_num, (unsigned long long)slot->ds_last_generation,
+ (unsigned long long)slot->ds_last_time, hb_block->hb_node,
+ (unsigned long long)le64_to_cpu(hb_block->hb_generation),
+ (unsigned long long)le64_to_cpu(hb_block->hb_seq));
+
+ return 0;
}
static inline void o2hb_prepare_block(struct o2hb_region *reg,
@@ -600,11 +620,9 @@ static void o2hb_fire_callbacks(struct o2hb_callback *hbcall,
struct o2nm_node *node,
int idx)
{
- struct list_head *iter;
struct o2hb_callback_func *f;
- list_for_each(iter, &hbcall->list) {
- f = list_entry(iter, struct o2hb_callback_func, hc_item);
+ list_for_each_entry(f, &hbcall->list, hc_item) {
mlog(ML_HEARTBEAT, "calling funcs %p\n", f);
(f->hc_func)(node, idx, f->hc_data);
}
@@ -613,16 +631,9 @@ static void o2hb_fire_callbacks(struct o2hb_callback *hbcall,
/* Will run the list in order until we process the passed event */
static void o2hb_run_event_list(struct o2hb_node_event *queued_event)
{
- int empty;
struct o2hb_callback *hbcall;
struct o2hb_node_event *event;
- spin_lock(&o2hb_live_lock);
- empty = list_empty(&queued_event->hn_item);
- spin_unlock(&o2hb_live_lock);
- if (empty)
- return;
-
/* Holding callback sem assures we don't alter the callback
* lists when doing this, and serializes ourselves with other
* processes wanting callbacks. */
@@ -681,6 +692,7 @@ static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot)
struct o2hb_node_event event =
{ .hn_item = LIST_HEAD_INIT(event.hn_item), };
struct o2nm_node *node;
+ int queued = 0;
node = o2nm_get_node_by_num(slot->ds_node_num);
if (!node)
@@ -698,26 +710,35 @@ static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot)
o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, node,
slot->ds_node_num);
+ queued = 1;
}
}
spin_unlock(&o2hb_live_lock);
- o2hb_run_event_list(&event);
+ if (queued)
+ o2hb_run_event_list(&event);
o2nm_node_put(node);
}
-static void o2hb_set_quorum_device(struct o2hb_region *reg,
- struct o2hb_disk_slot *slot)
+static void o2hb_set_quorum_device(struct o2hb_region *reg)
{
- assert_spin_locked(&o2hb_live_lock);
-
if (!o2hb_global_heartbeat_active())
return;
- if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
+ /* Prevent race with o2hb_heartbeat_group_drop_item() */
+ if (kthread_should_stop())
+ return;
+
+ /* Tag region as quorum only after thread reaches steady state */
+ if (atomic_read(&reg->hr_steady_iterations) != 0)
return;
+ spin_lock(&o2hb_live_lock);
+
+ if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
+ goto unlock;
+
/*
* A region can be added to the quorum only when it sees all
* live nodes heartbeat on it. In other words, the region has been
@@ -725,13 +746,10 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg,
*/
if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap,
sizeof(o2hb_live_node_bitmap)))
- return;
-
- if (slot->ds_changed_samples < O2HB_LIVE_THRESHOLD)
- return;
+ goto unlock;
- printk(KERN_NOTICE "o2hb: Region %s is now a quorum device\n",
- config_item_name(&reg->hr_item));
+ printk(KERN_NOTICE "o2hb: Region %s (%s) is now a quorum device\n",
+ config_item_name(&reg->hr_item), reg->hr_dev_name);
set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
@@ -739,9 +757,11 @@ static void o2hb_set_quorum_device(struct o2hb_region *reg,
* If global heartbeat active, unpin all regions if the
* region count > CUT_OFF
*/
- if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
+ if (bitmap_weight(o2hb_quorum_region_bitmap,
O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF)
o2hb_region_unpin(NULL);
+unlock:
+ spin_unlock(&o2hb_live_lock);
}
static int o2hb_check_slot(struct o2hb_region *reg,
@@ -756,6 +776,7 @@ static int o2hb_check_slot(struct o2hb_region *reg,
unsigned int dead_ms = o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS;
unsigned int slot_dead_ms;
int tmp;
+ int queued = 0;
memcpy(hb_block, slot->ds_raw_block, reg->hr_block_bytes);
@@ -849,6 +870,7 @@ fire_callbacks:
slot->ds_node_num);
changed = 1;
+ queued = 1;
}
list_add_tail(&slot->ds_live_item,
@@ -900,6 +922,7 @@ fire_callbacks:
node, slot->ds_node_num);
changed = 1;
+ queued = 1;
}
/* We don't clear this because the node is still
@@ -913,39 +936,25 @@ fire_callbacks:
slot->ds_equal_samples = 0;
}
out:
- o2hb_set_quorum_device(reg, slot);
-
spin_unlock(&o2hb_live_lock);
- o2hb_run_event_list(&event);
+ if (queued)
+ o2hb_run_event_list(&event);
if (node)
o2nm_node_put(node);
return changed;
}
-/* This could be faster if we just implmented a find_last_bit, but I
- * don't think the circumstances warrant it. */
-static int o2hb_highest_node(unsigned long *nodes,
- int numbits)
+static int o2hb_highest_node(unsigned long *nodes, int numbits)
{
- int highest, node;
-
- highest = numbits;
- node = -1;
- while ((node = find_next_bit(nodes, numbits, node + 1)) != -1) {
- if (node >= numbits)
- break;
-
- highest = node;
- }
-
- return highest;
+ return find_last_bit(nodes, numbits);
}
static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
{
- int i, ret, highest_node, change = 0;
+ int i, ret, highest_node;
+ int membership_change = 0, own_slot_ok = 0;
unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)];
unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
struct o2hb_bio_wait_ctxt write_wc;
@@ -954,7 +963,7 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
sizeof(configured_nodes));
if (ret) {
mlog_errno(ret);
- return ret;
+ goto bail;
}
/*
@@ -970,8 +979,9 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES);
if (highest_node >= O2NM_MAX_NODES) {
- mlog(ML_NOTICE, "ocfs2_heartbeat: no configured nodes found!\n");
- return -EINVAL;
+ mlog(ML_NOTICE, "o2hb: No configured nodes found!\n");
+ ret = -EINVAL;
+ goto bail;
}
/* No sense in reading the slots of nodes that don't exist
@@ -981,31 +991,27 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
ret = o2hb_read_slots(reg, highest_node + 1);
if (ret < 0) {
mlog_errno(ret);
- return ret;
+ goto bail;
}
/* With an up to date view of the slots, we can check that no
* other node has been improperly configured to heartbeat in
* our slot. */
- if (!o2hb_check_last_timestamp(reg))
- mlog(ML_ERROR, "Device \"%s\": another node is heartbeating "
- "in our slot!\n", reg->hr_dev_name);
+ own_slot_ok = o2hb_check_own_slot(reg);
/* fill in the proper info for our next heartbeat */
o2hb_prepare_block(reg, reg->hr_generation);
- /* And fire off the write. Note that we don't wait on this I/O
- * until later. */
ret = o2hb_issue_node_write(reg, &write_wc);
if (ret < 0) {
mlog_errno(ret);
- return ret;
+ goto bail;
}
i = -1;
- while((i = find_next_bit(configured_nodes, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
-
- change |= o2hb_check_slot(reg, &reg->hr_slots[i]);
+ while((i = find_next_bit(configured_nodes,
+ O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
+ membership_change |= o2hb_check_slot(reg, &reg->hr_slots[i]);
}
/*
@@ -1020,18 +1026,39 @@ static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
* disk */
mlog(ML_ERROR, "Write error %d on device \"%s\"\n",
write_wc.wc_error, reg->hr_dev_name);
- return write_wc.wc_error;
+ ret = write_wc.wc_error;
+ goto bail;
}
- o2hb_arm_write_timeout(reg);
+ /* Skip disarming the timeout if own slot has stale/bad data */
+ if (own_slot_ok) {
+ o2hb_set_quorum_device(reg);
+ o2hb_arm_write_timeout(reg);
+ }
+bail:
/* let the person who launched us know when things are steady */
- if (!change && (atomic_read(&reg->hr_steady_iterations) != 0)) {
- if (atomic_dec_and_test(&reg->hr_steady_iterations))
+ if (atomic_read(&reg->hr_steady_iterations) != 0) {
+ if (!ret && own_slot_ok && !membership_change) {
+ if (atomic_dec_and_test(&reg->hr_steady_iterations))
+ wake_up(&o2hb_steady_queue);
+ }
+ }
+
+ if (atomic_read(&reg->hr_steady_iterations) != 0) {
+ if (atomic_dec_and_test(&reg->hr_unsteady_iterations)) {
+ printk(KERN_NOTICE "o2hb: Unable to stabilize "
+ "heartbeart on region %s (%s)\n",
+ config_item_name(&reg->hr_item),
+ reg->hr_dev_name);
+ atomic_set(&reg->hr_steady_iterations, 0);
+ reg->hr_aborted_start = 1;
wake_up(&o2hb_steady_queue);
+ ret = -EIO;
+ }
}
- return 0;
+ return ret;
}
/* Subtract b from a, storing the result in a. a *must* have a larger
@@ -1080,12 +1107,13 @@ static int o2hb_thread(void *data)
mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread running\n");
- set_user_nice(current, -20);
+ set_user_nice(current, MIN_NICE);
/* Pin node */
o2nm_depend_this_node();
- while (!kthread_should_stop() && !reg->hr_unclean_stop) {
+ while (!kthread_should_stop() &&
+ !reg->hr_unclean_stop && !reg->hr_aborted_start) {
/* We track the time spent inside
* o2hb_do_disk_heartbeat so that we avoid more than
* hr_timeout_ms between disk writes. On busy systems
@@ -1093,10 +1121,7 @@ static int o2hb_thread(void *data)
* likely to time itself out. */
do_gettimeofday(&before_hb);
- i = 0;
- do {
- ret = o2hb_do_disk_heartbeat(reg);
- } while (ret && ++i < 2);
+ ret = o2hb_do_disk_heartbeat(reg);
do_gettimeofday(&after_hb);
elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb);
@@ -1107,7 +1132,8 @@ static int o2hb_thread(void *data)
after_hb.tv_sec, (unsigned long) after_hb.tv_usec,
elapsed_msec);
- if (elapsed_msec < reg->hr_timeout_ms) {
+ if (!kthread_should_stop() &&
+ elapsed_msec < reg->hr_timeout_ms) {
/* the kthread api has blocked signals for us so no
* need to record the return value. */
msleep_interruptible(reg->hr_timeout_ms - elapsed_msec);
@@ -1124,20 +1150,20 @@ static int o2hb_thread(void *data)
* to timeout on this region when we could just as easily
* write a clear generation - thus indicating to them that
* this node has left this region.
- *
- * XXX: Should we skip this on unclean_stop? */
- o2hb_prepare_block(reg, 0);
- ret = o2hb_issue_node_write(reg, &write_wc);
- if (ret == 0) {
- o2hb_wait_on_io(reg, &write_wc);
- } else {
- mlog_errno(ret);
+ */
+ if (!reg->hr_unclean_stop && !reg->hr_aborted_start) {
+ o2hb_prepare_block(reg, 0);
+ ret = o2hb_issue_node_write(reg, &write_wc);
+ if (ret == 0)
+ o2hb_wait_on_io(reg, &write_wc);
+ else
+ mlog_errno(ret);
}
/* Unpin node */
o2nm_undepend_this_node();
- mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread exiting\n");
+ mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n");
return 0;
}
@@ -1148,6 +1174,7 @@ static int o2hb_debug_open(struct inode *inode, struct file *file)
struct o2hb_debug_buf *db = inode->i_private;
struct o2hb_region *reg;
unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)];
+ unsigned long lts;
char *buf = NULL;
int i = -1;
int out = 0;
@@ -1184,9 +1211,11 @@ static int o2hb_debug_open(struct inode *inode, struct file *file)
case O2HB_DB_TYPE_REGION_ELAPSED_TIME:
reg = (struct o2hb_region *)db->db_data;
- out += snprintf(buf + out, PAGE_SIZE - out, "%u\n",
- jiffies_to_msecs(jiffies -
- reg->hr_last_timeout_start));
+ lts = reg->hr_last_timeout_start;
+ /* If 0, it has never been set before */
+ if (lts)
+ lts = jiffies_to_msecs(jiffies - lts);
+ out += snprintf(buf + out, PAGE_SIZE - out, "%lu\n", lts);
goto done;
case O2HB_DB_TYPE_REGION_PINNED:
@@ -1416,8 +1445,9 @@ static void o2hb_region_release(struct config_item *item)
struct page *page;
struct o2hb_region *reg = to_o2hb_region(item);
- if (reg->hr_tmp_block)
- kfree(reg->hr_tmp_block);
+ mlog(ML_HEARTBEAT, "hb region release (%s)\n", reg->hr_dev_name);
+
+ kfree(reg->hr_tmp_block);
if (reg->hr_slot_data) {
for (i = 0; i < reg->hr_num_pages; i++) {
@@ -1431,8 +1461,7 @@ static void o2hb_region_release(struct config_item *item)
if (reg->hr_bdev)
blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
- if (reg->hr_slots)
- kfree(reg->hr_slots);
+ kfree(reg->hr_slots);
kfree(reg->hr_db_regnum);
kfree(reg->hr_db_livenodes);
@@ -1658,8 +1687,6 @@ static int o2hb_populate_slot_data(struct o2hb_region *reg)
struct o2hb_disk_slot *slot;
struct o2hb_disk_heartbeat_block *hb_block;
- mlog_entry_void();
-
ret = o2hb_read_slots(reg, reg->hr_blocks);
if (ret) {
mlog_errno(ret);
@@ -1681,7 +1708,6 @@ static int o2hb_populate_slot_data(struct o2hb_region *reg)
}
out:
- mlog_exit(ret);
return ret;
}
@@ -1694,9 +1720,10 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
long fd;
int sectsize;
char *p = (char *)page;
- struct file *filp = NULL;
- struct inode *inode = NULL;
+ struct fd f;
+ struct inode *inode;
ssize_t ret = -EINVAL;
+ int live_threshold;
if (reg->hr_bdev)
goto out;
@@ -1713,26 +1740,26 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
if (fd < 0 || fd >= INT_MAX)
goto out;
- filp = fget(fd);
- if (filp == NULL)
+ f = fdget(fd);
+ if (f.file == NULL)
goto out;
if (reg->hr_blocks == 0 || reg->hr_start_block == 0 ||
reg->hr_block_bytes == 0)
- goto out;
+ goto out2;
- inode = igrab(filp->f_mapping->host);
+ inode = igrab(f.file->f_mapping->host);
if (inode == NULL)
- goto out;
+ goto out2;
if (!S_ISBLK(inode->i_mode))
- goto out;
+ goto out3;
- reg->hr_bdev = I_BDEV(filp->f_mapping->host);
+ reg->hr_bdev = I_BDEV(f.file->f_mapping->host);
ret = blkdev_get(reg->hr_bdev, FMODE_WRITE | FMODE_READ, NULL);
if (ret) {
reg->hr_bdev = NULL;
- goto out;
+ goto out3;
}
inode = NULL;
@@ -1744,7 +1771,7 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
"blocksize %u incorrect for device, expected %d",
reg->hr_block_bytes, sectsize);
ret = -EINVAL;
- goto out;
+ goto out3;
}
o2hb_init_region_params(reg);
@@ -1758,13 +1785,13 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
ret = o2hb_map_slot_data(reg);
if (ret) {
mlog_errno(ret);
- goto out;
+ goto out3;
}
ret = o2hb_populate_slot_data(reg);
if (ret) {
mlog_errno(ret);
- goto out;
+ goto out3;
}
INIT_DELAYED_WORK(&reg->hr_write_timeout_work, o2hb_write_timeout);
@@ -1773,15 +1800,28 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
* A node is considered live after it has beat LIVE_THRESHOLD
* times. We're not steady until we've given them a chance
* _after_ our first read.
+ * The default threshold is bare minimum so as to limit the delay
+ * during mounts. For global heartbeat, the threshold doubled for the
+ * first region.
*/
- atomic_set(&reg->hr_steady_iterations, O2HB_LIVE_THRESHOLD + 1);
+ live_threshold = O2HB_LIVE_THRESHOLD;
+ if (o2hb_global_heartbeat_active()) {
+ spin_lock(&o2hb_live_lock);
+ if (bitmap_weight(o2hb_region_bitmap, O2NM_MAX_REGIONS) == 1)
+ live_threshold <<= 1;
+ spin_unlock(&o2hb_live_lock);
+ }
+ ++live_threshold;
+ atomic_set(&reg->hr_steady_iterations, live_threshold);
+ /* unsteady_iterations is double the steady_iterations */
+ atomic_set(&reg->hr_unsteady_iterations, (live_threshold << 1));
hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s",
reg->hr_item.ci_name);
if (IS_ERR(hb_task)) {
ret = PTR_ERR(hb_task);
mlog_errno(ret);
- goto out;
+ goto out3;
}
spin_lock(&o2hb_live_lock);
@@ -1791,15 +1831,13 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
ret = wait_event_interruptible(o2hb_steady_queue,
atomic_read(&reg->hr_steady_iterations) == 0);
if (ret) {
- /* We got interrupted (hello ptrace!). Clean up */
- spin_lock(&o2hb_live_lock);
- hb_task = reg->hr_task;
- reg->hr_task = NULL;
- spin_unlock(&o2hb_live_lock);
+ atomic_set(&reg->hr_steady_iterations, 0);
+ reg->hr_aborted_start = 1;
+ }
- if (hb_task)
- kthread_stop(hb_task);
- goto out;
+ if (reg->hr_aborted_start) {
+ ret = -EIO;
+ goto out3;
}
/* Ok, we were woken. Make sure it wasn't by drop_item() */
@@ -1815,14 +1853,14 @@ static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
ret = -EIO;
if (hb_task && o2hb_global_heartbeat_active())
- printk(KERN_NOTICE "o2hb: Heartbeat started on region %s\n",
- config_item_name(&reg->hr_item));
+ printk(KERN_NOTICE "o2hb: Heartbeat started on region %s (%s)\n",
+ config_item_name(&reg->hr_item), reg->hr_dev_name);
+out3:
+ iput(inode);
+out2:
+ fdput(f);
out:
- if (filp)
- fput(filp);
- if (inode)
- iput(inode);
if (ret < 0) {
if (reg->hr_bdev) {
blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
@@ -2074,13 +2112,6 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group,
/* stop the thread when the user removes the region dir */
spin_lock(&o2hb_live_lock);
- if (o2hb_global_heartbeat_active()) {
- clear_bit(reg->hr_region_num, o2hb_region_bitmap);
- clear_bit(reg->hr_region_num, o2hb_live_region_bitmap);
- if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
- quorum_region = 1;
- clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
- }
hb_task = reg->hr_task;
reg->hr_task = NULL;
reg->hr_item_dropped = 1;
@@ -2089,19 +2120,30 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group,
if (hb_task)
kthread_stop(hb_task);
+ if (o2hb_global_heartbeat_active()) {
+ spin_lock(&o2hb_live_lock);
+ clear_bit(reg->hr_region_num, o2hb_region_bitmap);
+ clear_bit(reg->hr_region_num, o2hb_live_region_bitmap);
+ if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
+ quorum_region = 1;
+ clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
+ spin_unlock(&o2hb_live_lock);
+ printk(KERN_NOTICE "o2hb: Heartbeat %s on region %s (%s)\n",
+ ((atomic_read(&reg->hr_steady_iterations) == 0) ?
+ "stopped" : "start aborted"), config_item_name(item),
+ reg->hr_dev_name);
+ }
+
/*
* If we're racing a dev_write(), we need to wake them. They will
* check reg->hr_task
*/
if (atomic_read(&reg->hr_steady_iterations) != 0) {
+ reg->hr_aborted_start = 1;
atomic_set(&reg->hr_steady_iterations, 0);
wake_up(&o2hb_steady_queue);
}
- if (o2hb_global_heartbeat_active())
- printk(KERN_NOTICE "o2hb: Heartbeat stopped on region %s\n",
- config_item_name(&reg->hr_item));
-
config_item_put(item);
if (!o2hb_global_heartbeat_active() || !quorum_region)
@@ -2116,7 +2158,7 @@ static void o2hb_heartbeat_group_drop_item(struct config_group *group,
if (!o2hb_dependent_users)
goto unlock;
- if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
+ if (bitmap_weight(o2hb_quorum_region_bitmap,
O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
o2hb_region_pin(NULL);
@@ -2205,7 +2247,7 @@ ssize_t o2hb_heartbeat_group_mode_store(struct o2hb_heartbeat_group *group,
if (strnicmp(page, o2hb_heartbeat_mode_desc[i], len))
continue;
- ret = o2hb_global_hearbeat_mode_set(i);
+ ret = o2hb_global_heartbeat_mode_set(i);
if (!ret)
printk(KERN_NOTICE "o2hb: Heartbeat mode set to %s\n",
o2hb_heartbeat_mode_desc[i]);
@@ -2238,7 +2280,7 @@ static struct configfs_attribute *o2hb_heartbeat_group_attrs[] = {
NULL,
};
-static struct configfs_item_operations o2hb_hearbeat_group_item_ops = {
+static struct configfs_item_operations o2hb_heartbeat_group_item_ops = {
.show_attribute = o2hb_heartbeat_group_show,
.store_attribute = o2hb_heartbeat_group_store,
};
@@ -2250,7 +2292,7 @@ static struct configfs_group_operations o2hb_heartbeat_group_group_ops = {
static struct config_item_type o2hb_heartbeat_group_type = {
.ct_group_ops = &o2hb_heartbeat_group_group_ops,
- .ct_item_ops = &o2hb_hearbeat_group_item_ops,
+ .ct_item_ops = &o2hb_heartbeat_group_item_ops,
.ct_attrs = o2hb_heartbeat_group_attrs,
.ct_owner = THIS_MODULE,
};
@@ -2282,7 +2324,7 @@ void o2hb_free_hb_set(struct config_group *group)
kfree(hs);
}
-/* hb callback registration and issueing */
+/* hb callback registration and issuing */
static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type)
{
@@ -2323,6 +2365,9 @@ static int o2hb_region_pin(const char *region_uuid)
assert_spin_locked(&o2hb_live_lock);
list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
+ if (reg->hr_item_dropped)
+ continue;
+
uuid = config_item_name(&reg->hr_item);
/* local heartbeat */
@@ -2373,6 +2418,9 @@ static void o2hb_region_unpin(const char *region_uuid)
assert_spin_locked(&o2hb_live_lock);
list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
+ if (reg->hr_item_dropped)
+ continue;
+
uuid = config_item_name(&reg->hr_item);
if (region_uuid) {
if (strcmp(region_uuid, uuid))
@@ -2410,7 +2458,7 @@ static int o2hb_region_inc_user(const char *region_uuid)
if (o2hb_dependent_users > 1)
goto unlock;
- if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
+ if (bitmap_weight(o2hb_quorum_region_bitmap,
O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
ret = o2hb_region_pin(NULL);
@@ -2444,8 +2492,7 @@ unlock:
int o2hb_register_callback(const char *region_uuid,
struct o2hb_callback_func *hc)
{
- struct o2hb_callback_func *tmp;
- struct list_head *iter;
+ struct o2hb_callback_func *f;
struct o2hb_callback *hbcall;
int ret;
@@ -2468,10 +2515,9 @@ int o2hb_register_callback(const char *region_uuid,
down_write(&o2hb_callback_sem);
- list_for_each(iter, &hbcall->list) {
- tmp = list_entry(iter, struct o2hb_callback_func, hc_item);
- if (hc->hc_priority < tmp->hc_priority) {
- list_add_tail(&hc->hc_item, iter);
+ list_for_each_entry(f, &hbcall->list, hc_item) {
+ if (hc->hc_priority < f->hc_priority) {
+ list_add_tail(&hc->hc_item, &f->hc_item);
break;
}
}
@@ -2588,6 +2634,9 @@ int o2hb_get_all_regions(char *region_uuids, u8 max_regions)
p = region_uuids;
list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
+ if (reg->hr_item_dropped)
+ continue;
+
mlog(0, "Region: %s\n", config_item_name(&reg->hr_item));
if (numregs < max_regions) {
memcpy(p, config_item_name(&reg->hr_item),
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 6c61771469a..07ac24fd925 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -30,7 +30,7 @@
struct mlog_bits mlog_and_bits = MLOG_BITS_RHS(MLOG_INITIAL_AND_MASK);
EXPORT_SYMBOL_GPL(mlog_and_bits);
-struct mlog_bits mlog_not_bits = MLOG_BITS_RHS(MLOG_INITIAL_NOT_MASK);
+struct mlog_bits mlog_not_bits = MLOG_BITS_RHS(0);
EXPORT_SYMBOL_GPL(mlog_not_bits);
static ssize_t mlog_mask_show(u64 mask, char *buf)
@@ -80,8 +80,6 @@ struct mlog_attribute {
}
static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
- define_mask(ENTRY),
- define_mask(EXIT),
define_mask(TCP),
define_mask(MSG),
define_mask(SOCKET),
@@ -93,27 +91,11 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_THREAD),
define_mask(DLM_MASTER),
define_mask(DLM_RECOVERY),
- define_mask(AIO),
- define_mask(JOURNAL),
- define_mask(DISK_ALLOC),
- define_mask(SUPER),
- define_mask(FILE_IO),
- define_mask(EXTENT_MAP),
define_mask(DLM_GLUE),
- define_mask(BH_IO),
- define_mask(UPTODATE),
- define_mask(NAMEI),
- define_mask(INODE),
define_mask(VOTE),
- define_mask(DCACHE),
define_mask(CONN),
define_mask(QUORUM),
- define_mask(EXPORT),
- define_mask(XATTR),
- define_mask(QUOTA),
- define_mask(REFCOUNT),
define_mask(BASTS),
- define_mask(RESERVATIONS),
define_mask(CLUSTER),
define_mask(ERROR),
define_mask(NOTICE),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 34d6544357d..2260fb9e650 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -82,41 +82,23 @@
/* bits that are frequently given and infrequently matched in the low word */
/* NOTE: If you add a flag, you need to also update masklog.c! */
-#define ML_ENTRY 0x0000000000000001ULL /* func call entry */
-#define ML_EXIT 0x0000000000000002ULL /* func call exit */
-#define ML_TCP 0x0000000000000004ULL /* net cluster/tcp.c */
-#define ML_MSG 0x0000000000000008ULL /* net network messages */
-#define ML_SOCKET 0x0000000000000010ULL /* net socket lifetime */
-#define ML_HEARTBEAT 0x0000000000000020ULL /* hb all heartbeat tracking */
-#define ML_HB_BIO 0x0000000000000040ULL /* hb io tracing */
-#define ML_DLMFS 0x0000000000000080ULL /* dlm user dlmfs */
-#define ML_DLM 0x0000000000000100ULL /* dlm general debugging */
-#define ML_DLM_DOMAIN 0x0000000000000200ULL /* dlm domain debugging */
-#define ML_DLM_THREAD 0x0000000000000400ULL /* dlm domain thread */
-#define ML_DLM_MASTER 0x0000000000000800ULL /* dlm master functions */
-#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
-#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
-#define ML_JOURNAL 0x0000000000004000ULL /* ocfs2 journalling functions */
-#define ML_DISK_ALLOC 0x0000000000008000ULL /* ocfs2 disk allocation */
-#define ML_SUPER 0x0000000000010000ULL /* ocfs2 mount / umount */
-#define ML_FILE_IO 0x0000000000020000ULL /* ocfs2 file I/O */
-#define ML_EXTENT_MAP 0x0000000000040000ULL /* ocfs2 extent map caching */
-#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
-#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
-#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
-#define ML_NAMEI 0x0000000000400000ULL /* ocfs2 directory / namespace */
-#define ML_INODE 0x0000000000800000ULL /* ocfs2 inode manipulation */
-#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
-#define ML_DCACHE 0x0000000002000000ULL /* ocfs2 dcache operations */
-#define ML_CONN 0x0000000004000000ULL /* net connection management */
-#define ML_QUORUM 0x0000000008000000ULL /* net connection quorum */
-#define ML_EXPORT 0x0000000010000000ULL /* ocfs2 export operations */
-#define ML_XATTR 0x0000000020000000ULL /* ocfs2 extended attributes */
-#define ML_QUOTA 0x0000000040000000ULL /* ocfs2 quota operations */
-#define ML_REFCOUNT 0x0000000080000000ULL /* refcount tree operations */
-#define ML_BASTS 0x0000000100000000ULL /* dlmglue asts and basts */
-#define ML_RESERVATIONS 0x0000000200000000ULL /* ocfs2 alloc reservations */
-#define ML_CLUSTER 0x0000000400000000ULL /* cluster stack */
+#define ML_TCP 0x0000000000000001ULL /* net cluster/tcp.c */
+#define ML_MSG 0x0000000000000002ULL /* net network messages */
+#define ML_SOCKET 0x0000000000000004ULL /* net socket lifetime */
+#define ML_HEARTBEAT 0x0000000000000008ULL /* hb all heartbeat tracking */
+#define ML_HB_BIO 0x0000000000000010ULL /* hb io tracing */
+#define ML_DLMFS 0x0000000000000020ULL /* dlm user dlmfs */
+#define ML_DLM 0x0000000000000040ULL /* dlm general debugging */
+#define ML_DLM_DOMAIN 0x0000000000000080ULL /* dlm domain debugging */
+#define ML_DLM_THREAD 0x0000000000000100ULL /* dlm domain thread */
+#define ML_DLM_MASTER 0x0000000000000200ULL /* dlm master functions */
+#define ML_DLM_RECOVERY 0x0000000000000400ULL /* dlm master functions */
+#define ML_DLM_GLUE 0x0000000000000800ULL /* ocfs2 dlm glue layer */
+#define ML_VOTE 0x0000000000001000ULL /* ocfs2 node messaging */
+#define ML_CONN 0x0000000000002000ULL /* net connection management */
+#define ML_QUORUM 0x0000000000004000ULL /* net connection quorum */
+#define ML_BASTS 0x0000000000008000ULL /* dlmglue asts and basts */
+#define ML_CLUSTER 0x0000000000010000ULL /* cluster stack */
/* bits that are infrequently given and frequently matched in the high word */
#define ML_ERROR 0x1000000000000000ULL /* sent to KERN_ERR */
@@ -124,7 +106,6 @@
#define ML_KTHREAD 0x4000000000000000ULL /* kernel thread activity */
#define MLOG_INITIAL_AND_MASK (ML_ERROR|ML_NOTICE)
-#define MLOG_INITIAL_NOT_MASK (ML_ENTRY|ML_EXIT)
#ifndef MLOG_MASK_PREFIX
#define MLOG_MASK_PREFIX 0
#endif
@@ -218,62 +199,11 @@ extern struct mlog_bits mlog_and_bits, mlog_not_bits;
#define mlog_errno(st) do { \
int _st = (st); \
if (_st != -ERESTARTSYS && _st != -EINTR && \
- _st != AOP_TRUNCATED_PAGE && _st != -ENOSPC) \
+ _st != AOP_TRUNCATED_PAGE && _st != -ENOSPC && \
+ _st != -EDQUOT) \
mlog(ML_ERROR, "status = %lld\n", (long long)_st); \
} while (0)
-#if defined(CONFIG_OCFS2_DEBUG_MASKLOG)
-#define mlog_entry(fmt, args...) do { \
- mlog(ML_ENTRY, "ENTRY:" fmt , ##args); \
-} while (0)
-
-#define mlog_entry_void() do { \
- mlog(ML_ENTRY, "ENTRY:\n"); \
-} while (0)
-
-/*
- * We disable this for sparse.
- */
-#if !defined(__CHECKER__)
-#define mlog_exit(st) do { \
- if (__builtin_types_compatible_p(typeof(st), unsigned long)) \
- mlog(ML_EXIT, "EXIT: %lu\n", (unsigned long) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), signed long)) \
- mlog(ML_EXIT, "EXIT: %ld\n", (signed long) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), unsigned int) \
- || __builtin_types_compatible_p(typeof(st), unsigned short) \
- || __builtin_types_compatible_p(typeof(st), unsigned char)) \
- mlog(ML_EXIT, "EXIT: %u\n", (unsigned int) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), signed int) \
- || __builtin_types_compatible_p(typeof(st), signed short) \
- || __builtin_types_compatible_p(typeof(st), signed char)) \
- mlog(ML_EXIT, "EXIT: %d\n", (signed int) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), long long)) \
- mlog(ML_EXIT, "EXIT: %lld\n", (long long) (st)); \
- else \
- mlog(ML_EXIT, "EXIT: %llu\n", (unsigned long long) (st)); \
-} while (0)
-#else
-#define mlog_exit(st) do { \
- mlog(ML_EXIT, "EXIT: %lld\n", (long long) (st)); \
-} while (0)
-#endif
-
-#define mlog_exit_ptr(ptr) do { \
- mlog(ML_EXIT, "EXIT: %p\n", ptr); \
-} while (0)
-
-#define mlog_exit_void() do { \
- mlog(ML_EXIT, "EXIT\n"); \
-} while (0)
-#else
-#define mlog_entry(...) do { } while (0)
-#define mlog_entry_void(...) do { } while (0)
-#define mlog_exit(...) do { } while (0)
-#define mlog_exit_ptr(...) do { } while (0)
-#define mlog_exit_void(...) do { } while (0)
-#endif /* defined(CONFIG_OCFS2_DEBUG_MASKLOG) */
-
#define mlog_bug_on_msg(cond, fmt, args...) do { \
if (cond) { \
mlog(ML_ERROR, "bug expression: " #cond "\n"); \
diff --git a/fs/ocfs2/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c
index 3a5835904b3..73ba81928bc 100644
--- a/fs/ocfs2/cluster/netdebug.c
+++ b/fs/ocfs2/cluster/netdebug.c
@@ -47,6 +47,7 @@
#define SC_DEBUG_NAME "sock_containers"
#define NST_DEBUG_NAME "send_tracking"
#define STATS_DEBUG_NAME "stats"
+#define NODES_DEBUG_NAME "connected_nodes"
#define SHOW_SOCK_CONTAINERS 0
#define SHOW_SOCK_STATS 1
@@ -55,6 +56,7 @@ static struct dentry *o2net_dentry;
static struct dentry *sc_dentry;
static struct dentry *nst_dentry;
static struct dentry *stats_dentry;
+static struct dentry *nodes_dentry;
static DEFINE_SPINLOCK(o2net_debug_lock);
@@ -491,53 +493,87 @@ static const struct file_operations sc_seq_fops = {
.release = sc_fop_release,
};
-int o2net_debugfs_init(void)
+static int o2net_fill_bitmap(char *buf, int len)
{
- o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL);
- if (!o2net_dentry) {
- mlog_errno(-ENOMEM);
- goto bail;
- }
+ unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)];
+ int i = -1, out = 0;
- nst_dentry = debugfs_create_file(NST_DEBUG_NAME, S_IFREG|S_IRUSR,
- o2net_dentry, NULL,
- &nst_seq_fops);
- if (!nst_dentry) {
- mlog_errno(-ENOMEM);
- goto bail;
- }
+ o2net_fill_node_map(map, sizeof(map));
- sc_dentry = debugfs_create_file(SC_DEBUG_NAME, S_IFREG|S_IRUSR,
- o2net_dentry, NULL,
- &sc_seq_fops);
- if (!sc_dentry) {
- mlog_errno(-ENOMEM);
- goto bail;
- }
+ while ((i = find_next_bit(map, O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES)
+ out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i);
+ out += snprintf(buf + out, PAGE_SIZE - out, "\n");
- stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, S_IFREG|S_IRUSR,
- o2net_dentry, NULL,
- &stats_seq_fops);
- if (!stats_dentry) {
- mlog_errno(-ENOMEM);
- goto bail;
- }
+ return out;
+}
+
+static int nodes_fop_open(struct inode *inode, struct file *file)
+{
+ char *buf;
+
+ buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
+ if (!buf)
+ return -ENOMEM;
+
+ i_size_write(inode, o2net_fill_bitmap(buf, PAGE_SIZE));
+
+ file->private_data = buf;
return 0;
-bail:
- debugfs_remove(stats_dentry);
- debugfs_remove(sc_dentry);
- debugfs_remove(nst_dentry);
- debugfs_remove(o2net_dentry);
- return -ENOMEM;
}
+static int o2net_debug_release(struct inode *inode, struct file *file)
+{
+ kfree(file->private_data);
+ return 0;
+}
+
+static ssize_t o2net_debug_read(struct file *file, char __user *buf,
+ size_t nbytes, loff_t *ppos)
+{
+ return simple_read_from_buffer(buf, nbytes, ppos, file->private_data,
+ i_size_read(file->f_mapping->host));
+}
+
+static const struct file_operations nodes_fops = {
+ .open = nodes_fop_open,
+ .release = o2net_debug_release,
+ .read = o2net_debug_read,
+ .llseek = generic_file_llseek,
+};
+
void o2net_debugfs_exit(void)
{
+ debugfs_remove(nodes_dentry);
debugfs_remove(stats_dentry);
debugfs_remove(sc_dentry);
debugfs_remove(nst_dentry);
debugfs_remove(o2net_dentry);
}
+int o2net_debugfs_init(void)
+{
+ umode_t mode = S_IFREG|S_IRUSR;
+
+ o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL);
+ if (o2net_dentry)
+ nst_dentry = debugfs_create_file(NST_DEBUG_NAME, mode,
+ o2net_dentry, NULL, &nst_seq_fops);
+ if (nst_dentry)
+ sc_dentry = debugfs_create_file(SC_DEBUG_NAME, mode,
+ o2net_dentry, NULL, &sc_seq_fops);
+ if (sc_dentry)
+ stats_dentry = debugfs_create_file(STATS_DEBUG_NAME, mode,
+ o2net_dentry, NULL, &stats_seq_fops);
+ if (stats_dentry)
+ nodes_dentry = debugfs_create_file(NODES_DEBUG_NAME, mode,
+ o2net_dentry, NULL, &nodes_fops);
+ if (nodes_dentry)
+ return 0;
+
+ o2net_debugfs_exit();
+ mlog_errno(-ENOMEM);
+ return -ENOMEM;
+}
+
#endif /* CONFIG_DEBUG_FS */
diff --git a/fs/ocfs2/cluster/nodemanager.c b/fs/ocfs2/cluster/nodemanager.c
index bb240647ca5..441c84e169e 100644
--- a/fs/ocfs2/cluster/nodemanager.c
+++ b/fs/ocfs2/cluster/nodemanager.c
@@ -29,7 +29,6 @@
#include "heartbeat.h"
#include "masklog.h"
#include "sys.h"
-#include "ver.h"
/* for now we operate under the assertion that there can be only one
* cluster active at a time. Changing this will require trickling
@@ -945,8 +944,6 @@ static int __init init_o2nm(void)
{
int ret = -1;
- cluster_print_version();
-
ret = o2hb_init();
if (ret)
goto out;
@@ -984,6 +981,7 @@ out:
MODULE_AUTHOR("Oracle");
MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("OCFS2 cluster management");
module_init(init_o2nm)
module_exit(exit_o2nm)
diff --git a/fs/ocfs2/cluster/quorum.c b/fs/ocfs2/cluster/quorum.c
index a87366750f2..1ec141e758d 100644
--- a/fs/ocfs2/cluster/quorum.c
+++ b/fs/ocfs2/cluster/quorum.c
@@ -89,7 +89,7 @@ static void o2quo_fence_self(void)
};
}
-/* Indicate that a timeout occured on a hearbeat region write. The
+/* Indicate that a timeout occurred on a hearbeat region write. The
* other nodes in the cluster may consider us dead at that time so we
* want to "fence" ourselves so that we don't scribble on the disk
* after they think they've recovered us. This can't solve all
@@ -261,10 +261,10 @@ void o2quo_hb_still_up(u8 node)
spin_unlock(&qs->qs_lock);
}
-/* This is analagous to hb_up. as a node's connection comes up we delay the
+/* This is analogous to hb_up. as a node's connection comes up we delay the
* quorum decision until we see it heartbeating. the hold will be droped in
* hb_up or hb_down. it might be perpetuated by con_err until hb_down. if
- * it's already heartbeating we we might be dropping a hold that conn_up got.
+ * it's already heartbeating we might be dropping a hold that conn_up got.
* */
void o2quo_conn_up(u8 node)
{
@@ -327,5 +327,5 @@ void o2quo_exit(void)
{
struct o2quo_state *qs = &o2quo_state;
- flush_work_sync(&qs->qs_work);
+ flush_work(&qs->qs_work);
}
diff --git a/fs/ocfs2/cluster/sys.c b/fs/ocfs2/cluster/sys.c
index bc702dab5d1..b7f57271d49 100644
--- a/fs/ocfs2/cluster/sys.c
+++ b/fs/ocfs2/cluster/sys.c
@@ -41,7 +41,7 @@ static ssize_t version_show(struct kobject *kobj, struct kobj_attribute *attr,
return snprintf(buf, PAGE_SIZE, "%u\n", O2NM_API_VERSION);
}
static struct kobj_attribute attr_version =
- __ATTR(interface_revision, S_IFREG | S_IRUGO, version_show, NULL);
+ __ATTR(interface_revision, S_IRUGO, version_show, NULL);
static struct attribute *o2cb_attrs[] = {
&attr_version.attr,
@@ -57,7 +57,6 @@ static struct kset *o2cb_kset;
void o2cb_sys_shutdown(void)
{
mlog_sys_shutdown();
- sysfs_remove_link(NULL, "o2cb");
kset_unregister(o2cb_kset);
}
@@ -69,14 +68,6 @@ int o2cb_sys_init(void)
if (!o2cb_kset)
return -ENOMEM;
- /*
- * Create this symlink for backwards compatibility with old
- * versions of ocfs2-tools which look for things in /sys/o2cb.
- */
- ret = sysfs_create_link(NULL, &o2cb_kset->kobj, "o2cb");
- if (ret)
- goto error;
-
ret = sysfs_create_group(&o2cb_kset->kobj, &o2cb_attr_group);
if (ret)
goto error;
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index 3b11cb1e38f..681691bc233 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -59,6 +59,7 @@
#include <linux/idr.h>
#include <linux/kref.h>
#include <linux/net.h>
+#include <linux/export.h>
#include <net/tcp.h>
#include <asm/uaccess.h>
@@ -107,7 +108,7 @@ static struct rb_root o2net_handler_tree = RB_ROOT;
static struct o2net_node o2net_nodes[O2NM_MAX_NODES];
/* XXX someday we'll need better accounting */
-static struct socket *o2net_listen_sock = NULL;
+static struct socket *o2net_listen_sock;
/*
* listen work is only queued by the listening socket callbacks on the
@@ -136,7 +137,7 @@ static int o2net_sys_err_translations[O2NET_ERR_MAX] =
static void o2net_sc_connect_completed(struct work_struct *work);
static void o2net_rx_until_empty(struct work_struct *work);
static void o2net_shutdown_sc(struct work_struct *work);
-static void o2net_listen_data_ready(struct sock *sk, int bytes);
+static void o2net_listen_data_ready(struct sock *sk);
static void o2net_sc_send_keep_req(struct work_struct *work);
static void o2net_idle_timer(unsigned long data);
static void o2net_sc_postpone_idle(struct o2net_sock_container *sc);
@@ -210,10 +211,6 @@ static inline void o2net_set_func_stop_time(struct o2net_sock_container *sc)
sc->sc_tv_func_stop = ktime_get();
}
-static ktime_t o2net_get_func_run_time(struct o2net_sock_container *sc)
-{
- return ktime_sub(sc->sc_tv_func_stop, sc->sc_tv_func_start);
-}
#else /* CONFIG_DEBUG_FS */
# define o2net_init_nst(a, b, c, d, e)
# define o2net_set_nst_sock_time(a)
@@ -227,10 +224,14 @@ static ktime_t o2net_get_func_run_time(struct o2net_sock_container *sc)
# define o2net_set_advance_stop_time(a)
# define o2net_set_func_start_time(a)
# define o2net_set_func_stop_time(a)
-# define o2net_get_func_run_time(a) (ktime_t)0
#endif /* CONFIG_DEBUG_FS */
#ifdef CONFIG_OCFS2_FS_STATS
+static ktime_t o2net_get_func_run_time(struct o2net_sock_container *sc)
+{
+ return ktime_sub(sc->sc_tv_func_stop, sc->sc_tv_func_start);
+}
+
static void o2net_update_send_stats(struct o2net_send_tracking *nst,
struct o2net_sock_container *sc)
{
@@ -261,17 +262,17 @@ static void o2net_update_recv_stats(struct o2net_sock_container *sc)
#endif /* CONFIG_OCFS2_FS_STATS */
-static inline int o2net_reconnect_delay(void)
+static inline unsigned int o2net_reconnect_delay(void)
{
return o2nm_single_cluster->cl_reconnect_delay_ms;
}
-static inline int o2net_keepalive_delay(void)
+static inline unsigned int o2net_keepalive_delay(void)
{
return o2nm_single_cluster->cl_keepalive_delay_ms;
}
-static inline int o2net_idle_timeout(void)
+static inline unsigned int o2net_idle_timeout(void)
{
return o2nm_single_cluster->cl_idle_timeout_ms;
}
@@ -303,28 +304,22 @@ static u8 o2net_num_from_nn(struct o2net_node *nn)
static int o2net_prep_nsw(struct o2net_node *nn, struct o2net_status_wait *nsw)
{
- int ret = 0;
-
- do {
- if (!idr_pre_get(&nn->nn_status_idr, GFP_ATOMIC)) {
- ret = -EAGAIN;
- break;
- }
- spin_lock(&nn->nn_lock);
- ret = idr_get_new(&nn->nn_status_idr, nsw, &nsw->ns_id);
- if (ret == 0)
- list_add_tail(&nsw->ns_node_item,
- &nn->nn_status_list);
- spin_unlock(&nn->nn_lock);
- } while (ret == -EAGAIN);
+ int ret;
- if (ret == 0) {
- init_waitqueue_head(&nsw->ns_wq);
- nsw->ns_sys_status = O2NET_ERR_NONE;
- nsw->ns_status = 0;
+ spin_lock(&nn->nn_lock);
+ ret = idr_alloc(&nn->nn_status_idr, nsw, 0, 0, GFP_ATOMIC);
+ if (ret >= 0) {
+ nsw->ns_id = ret;
+ list_add_tail(&nsw->ns_node_item, &nn->nn_status_list);
}
+ spin_unlock(&nn->nn_lock);
+ if (ret < 0)
+ return ret;
- return ret;
+ init_waitqueue_head(&nsw->ns_wq);
+ nsw->ns_sys_status = O2NET_ERR_NONE;
+ nsw->ns_status = 0;
+ return 0;
}
static void o2net_complete_nsw_locked(struct o2net_node *nn,
@@ -411,6 +406,9 @@ static void sc_kref_release(struct kref *kref)
sc->sc_node = NULL;
o2net_debug_del_sc(sc);
+
+ if (sc->sc_page)
+ __free_page(sc->sc_page);
kfree(sc);
}
@@ -545,8 +543,9 @@ static void o2net_set_nn_state(struct o2net_node *nn,
}
if (was_valid && !valid) {
- printk(KERN_NOTICE "o2net: no longer connected to "
- SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc));
+ if (old_sc)
+ printk(KERN_NOTICE "o2net: No longer connected to "
+ SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc));
o2net_complete_nodes_nsw(nn);
}
@@ -555,7 +554,7 @@ static void o2net_set_nn_state(struct o2net_node *nn,
cancel_delayed_work(&nn->nn_connect_expired);
printk(KERN_NOTICE "o2net: %s " SC_NODEF_FMT "\n",
o2nm_this_node() > sc->sc_node->nd_num ?
- "connected to" : "accepted connection from",
+ "Connected to" : "Accepted connection from",
SC_NODEF_ARGS(sc));
}
@@ -565,7 +564,7 @@ static void o2net_set_nn_state(struct o2net_node *nn,
* the work queue actually being up. */
if (!valid && o2net_wq) {
unsigned long delay;
- /* delay if we're withing a RECONNECT_DELAY of the
+ /* delay if we're within a RECONNECT_DELAY of the
* last attempt */
delay = (nn->nn_last_connect_attempt +
msecs_to_jiffies(o2net_reconnect_delay()))
@@ -598,9 +597,9 @@ static void o2net_set_nn_state(struct o2net_node *nn,
}
/* see o2net_register_callbacks() */
-static void o2net_data_ready(struct sock *sk, int bytes)
+static void o2net_data_ready(struct sock *sk)
{
- void (*ready)(struct sock *sk, int bytes);
+ void (*ready)(struct sock *sk);
read_lock(&sk->sk_callback_lock);
if (sk->sk_user_data) {
@@ -614,7 +613,7 @@ static void o2net_data_ready(struct sock *sk, int bytes)
}
read_unlock(&sk->sk_callback_lock);
- ready(sk, bytes);
+ ready(sk);
}
/* see o2net_register_callbacks() */
@@ -635,19 +634,19 @@ static void o2net_state_change(struct sock *sk)
state_change = sc->sc_state_change;
switch(sk->sk_state) {
- /* ignore connecting sockets as they make progress */
- case TCP_SYN_SENT:
- case TCP_SYN_RECV:
- break;
- case TCP_ESTABLISHED:
- o2net_sc_queue_work(sc, &sc->sc_connect_work);
- break;
- default:
- printk(KERN_INFO "o2net: connection to " SC_NODEF_FMT
- " shutdown, state %d\n",
- SC_NODEF_ARGS(sc), sk->sk_state);
- o2net_sc_queue_work(sc, &sc->sc_shutdown_work);
- break;
+ /* ignore connecting sockets as they make progress */
+ case TCP_SYN_SENT:
+ case TCP_SYN_RECV:
+ break;
+ case TCP_ESTABLISHED:
+ o2net_sc_queue_work(sc, &sc->sc_connect_work);
+ break;
+ default:
+ printk(KERN_INFO "o2net: Connection to " SC_NODEF_FMT
+ " shutdown, state %d\n",
+ SC_NODEF_ARGS(sc), sk->sk_state);
+ o2net_sc_queue_work(sc, &sc->sc_shutdown_work);
+ break;
}
out:
read_unlock(&sk->sk_callback_lock);
@@ -767,32 +766,32 @@ static struct o2net_msg_handler *
o2net_handler_tree_lookup(u32 msg_type, u32 key, struct rb_node ***ret_p,
struct rb_node **ret_parent)
{
- struct rb_node **p = &o2net_handler_tree.rb_node;
- struct rb_node *parent = NULL;
+ struct rb_node **p = &o2net_handler_tree.rb_node;
+ struct rb_node *parent = NULL;
struct o2net_msg_handler *nmh, *ret = NULL;
int cmp;
- while (*p) {
- parent = *p;
- nmh = rb_entry(parent, struct o2net_msg_handler, nh_node);
+ while (*p) {
+ parent = *p;
+ nmh = rb_entry(parent, struct o2net_msg_handler, nh_node);
cmp = o2net_handler_cmp(nmh, msg_type, key);
- if (cmp < 0)
- p = &(*p)->rb_left;
- else if (cmp > 0)
- p = &(*p)->rb_right;
- else {
+ if (cmp < 0)
+ p = &(*p)->rb_left;
+ else if (cmp > 0)
+ p = &(*p)->rb_right;
+ else {
ret = nmh;
- break;
+ break;
}
- }
+ }
- if (ret_p != NULL)
- *ret_p = p;
- if (ret_parent != NULL)
- *ret_parent = parent;
+ if (ret_p != NULL)
+ *ret_p = p;
+ if (ret_parent != NULL)
+ *ret_parent = parent;
- return ret;
+ return ret;
}
static void o2net_handler_kref_release(struct kref *kref)
@@ -869,7 +868,7 @@ int o2net_register_handler(u32 msg_type, u32 key, u32 max_len,
/* we've had some trouble with handlers seemingly vanishing. */
mlog_bug_on_msg(o2net_handler_tree_lookup(msg_type, key, &p,
&parent) == NULL,
- "couldn't find handler we *just* registerd "
+ "couldn't find handler we *just* registered "
"for type %u key %08x\n", msg_type, key);
}
write_unlock(&o2net_handler_lock);
@@ -917,57 +916,30 @@ static struct o2net_msg_handler *o2net_handler_get(u32 msg_type, u32 key)
static int o2net_recv_tcp_msg(struct socket *sock, void *data, size_t len)
{
- int ret;
- mm_segment_t oldfs;
- struct kvec vec = {
- .iov_len = len,
- .iov_base = data,
- };
- struct msghdr msg = {
- .msg_iovlen = 1,
- .msg_iov = (struct iovec *)&vec,
- .msg_flags = MSG_DONTWAIT,
- };
-
- oldfs = get_fs();
- set_fs(get_ds());
- ret = sock_recvmsg(sock, &msg, len, msg.msg_flags);
- set_fs(oldfs);
-
- return ret;
+ struct kvec vec = { .iov_len = len, .iov_base = data, };
+ struct msghdr msg = { .msg_flags = MSG_DONTWAIT, };
+ return kernel_recvmsg(sock, &msg, &vec, 1, len, msg.msg_flags);
}
static int o2net_send_tcp_msg(struct socket *sock, struct kvec *vec,
size_t veclen, size_t total)
{
int ret;
- mm_segment_t oldfs;
- struct msghdr msg = {
- .msg_iov = (struct iovec *)vec,
- .msg_iovlen = veclen,
- };
+ struct msghdr msg;
if (sock == NULL) {
ret = -EINVAL;
goto out;
}
- oldfs = get_fs();
- set_fs(get_ds());
- ret = sock_sendmsg(sock, &msg, total);
- set_fs(oldfs);
- if (ret != total) {
- mlog(ML_ERROR, "sendmsg returned %d instead of %zu\n", ret,
- total);
- if (ret >= 0)
- ret = -EPIPE; /* should be smarter, I bet */
- goto out;
- }
-
- ret = 0;
+ ret = kernel_sendmsg(sock, &msg, vec, veclen, total);
+ if (likely(ret == total))
+ return 0;
+ mlog(ML_ERROR, "sendmsg returned %d instead of %zu\n", ret, total);
+ if (ret >= 0)
+ ret = -EPIPE; /* should be smarter, I bet */
out:
- if (ret < 0)
- mlog(0, "returning error: %d\n", ret);
+ mlog(0, "returning error: %d\n", ret);
return ret;
}
@@ -1034,6 +1006,25 @@ static int o2net_tx_can_proceed(struct o2net_node *nn,
return ret;
}
+/* Get a map of all nodes to which this node is currently connected to */
+void o2net_fill_node_map(unsigned long *map, unsigned bytes)
+{
+ struct o2net_sock_container *sc;
+ int node, ret;
+
+ BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long)));
+
+ memset(map, 0, bytes);
+ for (node = 0; node < O2NM_MAX_NODES; ++node) {
+ o2net_tx_can_proceed(o2net_nn_from_num(node), &sc, &ret);
+ if (!ret) {
+ set_bit(node, map);
+ sc_put(sc);
+ }
+ }
+}
+EXPORT_SYMBOL_GPL(o2net_fill_node_map);
+
int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec,
size_t caller_veclen, u8 target_node, int *status)
{
@@ -1145,10 +1136,8 @@ out:
o2net_debug_del_nst(&nst); /* must be before dropping sc and node */
if (sc)
sc_put(sc);
- if (vec)
- kfree(vec);
- if (msg)
- kfree(msg);
+ kfree(vec);
+ kfree(msg);
o2net_complete_nsw(nn, &nsw, 0, 0, 0);
return ret;
}
@@ -1284,11 +1273,11 @@ static int o2net_check_handshake(struct o2net_sock_container *sc)
struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num);
if (hand->protocol_version != cpu_to_be64(O2NET_PROTOCOL_VERSION)) {
- mlog(ML_NOTICE, SC_NODEF_FMT " advertised net protocol "
- "version %llu but %llu is required, disconnecting\n",
- SC_NODEF_ARGS(sc),
- (unsigned long long)be64_to_cpu(hand->protocol_version),
- O2NET_PROTOCOL_VERSION);
+ printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " Advertised net "
+ "protocol version %llu but %llu is required. "
+ "Disconnecting.\n", SC_NODEF_ARGS(sc),
+ (unsigned long long)be64_to_cpu(hand->protocol_version),
+ O2NET_PROTOCOL_VERSION);
/* don't bother reconnecting if its the wrong version. */
o2net_ensure_shutdown(nn, sc, -ENOTCONN);
@@ -1302,33 +1291,33 @@ static int o2net_check_handshake(struct o2net_sock_container *sc)
*/
if (be32_to_cpu(hand->o2net_idle_timeout_ms) !=
o2net_idle_timeout()) {
- mlog(ML_NOTICE, SC_NODEF_FMT " uses a network idle timeout of "
- "%u ms, but we use %u ms locally. disconnecting\n",
- SC_NODEF_ARGS(sc),
- be32_to_cpu(hand->o2net_idle_timeout_ms),
- o2net_idle_timeout());
+ printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a network "
+ "idle timeout of %u ms, but we use %u ms locally. "
+ "Disconnecting.\n", SC_NODEF_ARGS(sc),
+ be32_to_cpu(hand->o2net_idle_timeout_ms),
+ o2net_idle_timeout());
o2net_ensure_shutdown(nn, sc, -ENOTCONN);
return -1;
}
if (be32_to_cpu(hand->o2net_keepalive_delay_ms) !=
o2net_keepalive_delay()) {
- mlog(ML_NOTICE, SC_NODEF_FMT " uses a keepalive delay of "
- "%u ms, but we use %u ms locally. disconnecting\n",
- SC_NODEF_ARGS(sc),
- be32_to_cpu(hand->o2net_keepalive_delay_ms),
- o2net_keepalive_delay());
+ printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a keepalive "
+ "delay of %u ms, but we use %u ms locally. "
+ "Disconnecting.\n", SC_NODEF_ARGS(sc),
+ be32_to_cpu(hand->o2net_keepalive_delay_ms),
+ o2net_keepalive_delay());
o2net_ensure_shutdown(nn, sc, -ENOTCONN);
return -1;
}
if (be32_to_cpu(hand->o2hb_heartbeat_timeout_ms) !=
O2HB_MAX_WRITE_TIMEOUT_MS) {
- mlog(ML_NOTICE, SC_NODEF_FMT " uses a heartbeat timeout of "
- "%u ms, but we use %u ms locally. disconnecting\n",
- SC_NODEF_ARGS(sc),
- be32_to_cpu(hand->o2hb_heartbeat_timeout_ms),
- O2HB_MAX_WRITE_TIMEOUT_MS);
+ printk(KERN_NOTICE "o2net: " SC_NODEF_FMT " uses a heartbeat "
+ "timeout of %u ms, but we use %u ms locally. "
+ "Disconnecting.\n", SC_NODEF_ARGS(sc),
+ be32_to_cpu(hand->o2hb_heartbeat_timeout_ms),
+ O2HB_MAX_WRITE_TIMEOUT_MS);
o2net_ensure_shutdown(nn, sc, -ENOTCONN);
return -1;
}
@@ -1539,28 +1528,16 @@ static void o2net_idle_timer(unsigned long data)
{
struct o2net_sock_container *sc = (struct o2net_sock_container *)data;
struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num);
-
#ifdef CONFIG_DEBUG_FS
- ktime_t now = ktime_get();
+ unsigned long msecs = ktime_to_ms(ktime_get()) -
+ ktime_to_ms(sc->sc_tv_timer);
+#else
+ unsigned long msecs = o2net_idle_timeout();
#endif
- printk(KERN_NOTICE "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u "
- "seconds, shutting it down.\n", SC_NODEF_ARGS(sc),
- o2net_idle_timeout() / 1000,
- o2net_idle_timeout() % 1000);
-
-#ifdef CONFIG_DEBUG_FS
- mlog(ML_NOTICE, "Here are some times that might help debug the "
- "situation: (Timer: %lld, Now %lld, DataReady %lld, Advance %lld-%lld, "
- "Key 0x%08x, Func %u, FuncTime %lld-%lld)\n",
- (long long)ktime_to_us(sc->sc_tv_timer), (long long)ktime_to_us(now),
- (long long)ktime_to_us(sc->sc_tv_data_ready),
- (long long)ktime_to_us(sc->sc_tv_advance_start),
- (long long)ktime_to_us(sc->sc_tv_advance_stop),
- sc->sc_msg_key, sc->sc_msg_type,
- (long long)ktime_to_us(sc->sc_tv_func_start),
- (long long)ktime_to_us(sc->sc_tv_func_stop));
-#endif
+ printk(KERN_NOTICE "o2net: Connection to " SC_NODEF_FMT " has been "
+ "idle for %lu.%lu secs, shutting it down.\n", SC_NODEF_ARGS(sc),
+ msecs / 1000, msecs % 1000);
/*
* Initialize the nn_timeout so that the next connection attempt
@@ -1692,13 +1669,12 @@ static void o2net_start_connect(struct work_struct *work)
ret = 0;
out:
- if (ret) {
- mlog(ML_NOTICE, "connect attempt to " SC_NODEF_FMT " failed "
- "with errno %d\n", SC_NODEF_ARGS(sc), ret);
+ if (ret && sc) {
+ printk(KERN_NOTICE "o2net: Connect attempt to " SC_NODEF_FMT
+ " failed with errno %d\n", SC_NODEF_ARGS(sc), ret);
/* 0 err so that another will be queued and attempted
* from set_nn_state */
- if (sc)
- o2net_ensure_shutdown(nn, sc, 0);
+ o2net_ensure_shutdown(nn, sc, 0);
}
if (sc)
sc_put(sc);
@@ -1717,8 +1693,8 @@ static void o2net_connect_expired(struct work_struct *work)
spin_lock(&nn->nn_lock);
if (!nn->nn_sc_valid) {
- mlog(ML_ERROR, "no connection established with node %u after "
- "%u.%u seconds, giving up and returning errors.\n",
+ printk(KERN_NOTICE "o2net: No connection established with "
+ "node %u after %u.%u seconds, giving up.\n",
o2net_num_from_nn(nn),
o2net_idle_timeout() / 1000,
o2net_idle_timeout() % 1000);
@@ -1823,7 +1799,7 @@ int o2net_register_hb_callbacks(void)
/* ------------------------------------------------------------ */
-static int o2net_accept_one(struct socket *sock)
+static int o2net_accept_one(struct socket *sock, int *more)
{
int ret, slen;
struct sockaddr_in sin;
@@ -1834,6 +1810,7 @@ static int o2net_accept_one(struct socket *sock)
struct o2net_node *nn;
BUG_ON(sock == NULL);
+ *more = 0;
ret = sock_create_lite(sock->sk->sk_family, sock->sk->sk_type,
sock->sk->sk_protocol, &new_sock);
if (ret)
@@ -1845,6 +1822,7 @@ static int o2net_accept_one(struct socket *sock)
if (ret < 0)
goto out;
+ *more = 1;
new_sock->sk->sk_allocation = GFP_ATOMIC;
ret = o2net_set_nodelay(new_sock);
@@ -1861,21 +1839,25 @@ static int o2net_accept_one(struct socket *sock)
node = o2nm_get_node_by_ip(sin.sin_addr.s_addr);
if (node == NULL) {
- mlog(ML_NOTICE, "attempt to connect from unknown node at %pI4:%d\n",
- &sin.sin_addr.s_addr, ntohs(sin.sin_port));
+ printk(KERN_NOTICE "o2net: Attempt to connect from unknown "
+ "node at %pI4:%d\n", &sin.sin_addr.s_addr,
+ ntohs(sin.sin_port));
ret = -EINVAL;
goto out;
}
if (o2nm_this_node() >= node->nd_num) {
local_node = o2nm_get_node_by_num(o2nm_this_node());
- mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' ("
- "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n",
- local_node->nd_name, local_node->nd_num,
- &(local_node->nd_ipv4_address),
- ntohs(local_node->nd_ipv4_port),
- node->nd_name, node->nd_num, &sin.sin_addr.s_addr,
- ntohs(sin.sin_port));
+ if (local_node)
+ printk(KERN_NOTICE "o2net: Unexpected connect attempt "
+ "seen at node '%s' (%u, %pI4:%d) from "
+ "node '%s' (%u, %pI4:%d)\n",
+ local_node->nd_name, local_node->nd_num,
+ &(local_node->nd_ipv4_address),
+ ntohs(local_node->nd_ipv4_port),
+ node->nd_name,
+ node->nd_num, &sin.sin_addr.s_addr,
+ ntohs(sin.sin_port));
ret = -EINVAL;
goto out;
}
@@ -1900,10 +1882,10 @@ static int o2net_accept_one(struct socket *sock)
ret = 0;
spin_unlock(&nn->nn_lock);
if (ret) {
- mlog(ML_NOTICE, "attempt to connect from node '%s' at "
- "%pI4:%d but it already has an open connection\n",
- node->nd_name, &sin.sin_addr.s_addr,
- ntohs(sin.sin_port));
+ printk(KERN_NOTICE "o2net: Attempt to connect from node '%s' "
+ "at %pI4:%d but it already has an open connection\n",
+ node->nd_name, &sin.sin_addr.s_addr,
+ ntohs(sin.sin_port));
goto out;
}
@@ -1939,16 +1921,41 @@ out:
return ret;
}
+/*
+ * This function is invoked in response to one or more
+ * pending accepts at softIRQ level. We must drain the
+ * entire que before returning.
+ */
+
static void o2net_accept_many(struct work_struct *work)
{
struct socket *sock = o2net_listen_sock;
- while (o2net_accept_one(sock) == 0)
+ int more;
+ int err;
+
+ /*
+ * It is critical to note that due to interrupt moderation
+ * at the network driver level, we can't assume to get a
+ * softIRQ for every single conn since tcp SYN packets
+ * can arrive back-to-back, and therefore many pending
+ * accepts may result in just 1 softIRQ. If we terminate
+ * the o2net_accept_one() loop upon seeing an err, what happens
+ * to the rest of the conns in the queue? If no new SYN
+ * arrives for hours, no softIRQ will be delivered,
+ * and the connections will just sit in the queue.
+ */
+
+ for (;;) {
+ err = o2net_accept_one(sock, &more);
+ if (!more)
+ break;
cond_resched();
+ }
}
-static void o2net_listen_data_ready(struct sock *sk, int bytes)
+static void o2net_listen_data_ready(struct sock *sk)
{
- void (*ready)(struct sock *sk, int bytes);
+ void (*ready)(struct sock *sk);
read_lock(&sk->sk_callback_lock);
ready = sk->sk_user_data;
@@ -1957,18 +1964,29 @@ static void o2net_listen_data_ready(struct sock *sk, int bytes)
goto out;
}
- /* ->sk_data_ready is also called for a newly established child socket
- * before it has been accepted and the acceptor has set up their
- * data_ready.. we only want to queue listen work for our listening
- * socket */
+ /* This callback may called twice when a new connection
+ * is being established as a child socket inherits everything
+ * from a parent LISTEN socket, including the data_ready cb of
+ * the parent. This leads to a hazard. In o2net_accept_one()
+ * we are still initializing the child socket but have not
+ * changed the inherited data_ready callback yet when
+ * data starts arriving.
+ * We avoid this hazard by checking the state.
+ * For the listening socket, the state will be TCP_LISTEN; for the new
+ * socket, will be TCP_ESTABLISHED. Also, in this case,
+ * sk->sk_user_data is not a valid function pointer.
+ */
+
if (sk->sk_state == TCP_LISTEN) {
- mlog(ML_TCP, "bytes: %d\n", bytes);
queue_work(o2net_wq, &o2net_listen_work);
+ } else {
+ ready = NULL;
}
out:
read_unlock(&sk->sk_callback_lock);
- ready(sk, bytes);
+ if (ready != NULL)
+ ready(sk);
}
static int o2net_open_listening_sock(__be32 addr, __be16 port)
@@ -1983,7 +2001,7 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port)
ret = sock_create(PF_INET, SOCK_STREAM, IPPROTO_TCP, &sock);
if (ret < 0) {
- mlog(ML_ERROR, "unable to create socket, ret=%d\n", ret);
+ printk(KERN_ERR "o2net: Error %d while creating socket\n", ret);
goto out;
}
@@ -1997,19 +2015,18 @@ static int o2net_open_listening_sock(__be32 addr, __be16 port)
o2net_listen_sock = sock;
INIT_WORK(&o2net_listen_work, o2net_accept_many);
- sock->sk->sk_reuse = 1;
+ sock->sk->sk_reuse = SK_CAN_REUSE;
ret = sock->ops->bind(sock, (struct sockaddr *)&sin, sizeof(sin));
if (ret < 0) {
- mlog(ML_ERROR, "unable to bind socket at %pI4:%u, "
- "ret=%d\n", &addr, ntohs(port), ret);
+ printk(KERN_ERR "o2net: Error %d while binding socket at "
+ "%pI4:%u\n", ret, &addr, ntohs(port));
goto out;
}
ret = sock->ops->listen(sock, 64);
- if (ret < 0) {
- mlog(ML_ERROR, "unable to listen on %pI4:%u, ret=%d\n",
- &addr, ntohs(port), ret);
- }
+ if (ret < 0)
+ printk(KERN_ERR "o2net: Error %d while listening on %pI4:%u\n",
+ ret, &addr, ntohs(port));
out:
if (ret) {
diff --git a/fs/ocfs2/cluster/tcp.h b/fs/ocfs2/cluster/tcp.h
index fd6179eb26d..5bada2a69b5 100644
--- a/fs/ocfs2/cluster/tcp.h
+++ b/fs/ocfs2/cluster/tcp.h
@@ -106,6 +106,8 @@ int o2net_register_handler(u32 msg_type, u32 key, u32 max_len,
struct list_head *unreg_list);
void o2net_unregister_handler_list(struct list_head *list);
+void o2net_fill_node_map(unsigned long *map, unsigned bytes);
+
struct o2nm_node;
int o2net_register_hb_callbacks(void);
void o2net_unregister_hb_callbacks(void);
diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h
index 4cbcb65784a..dc024367110 100644
--- a/fs/ocfs2/cluster/tcp_internal.h
+++ b/fs/ocfs2/cluster/tcp_internal.h
@@ -165,7 +165,7 @@ struct o2net_sock_container {
/* original handlers for the sockets */
void (*sc_state_change)(struct sock *sk);
- void (*sc_data_ready)(struct sock *sk, int bytes);
+ void (*sc_data_ready)(struct sock *sk);
u32 sc_msg_key;
u16 sc_msg_type;
diff --git a/fs/ocfs2/cluster/ver.c b/fs/ocfs2/cluster/ver.c
deleted file mode 100644
index a56eee6abad..00000000000
--- a/fs/ocfs2/cluster/ver.c
+++ /dev/null
@@ -1,42 +0,0 @@
-/* -*- mode: c; c-basic-offset: 8; -*-
- * vim: noexpandtab sw=8 ts=8 sts=0:
- *
- * ver.c
- *
- * version string
- *
- * Copyright (C) 2002, 2005 Oracle. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public
- * License as published by the Free Software Foundation; either
- * version 2 of the License, or (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public
- * License along with this program; if not, write to the
- * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
- * Boston, MA 021110-1307, USA.
- */
-
-#include <linux/module.h>
-#include <linux/kernel.h>
-
-#include "ver.h"
-
-#define CLUSTER_BUILD_VERSION "1.5.0"
-
-#define VERSION_STR "OCFS2 Node Manager " CLUSTER_BUILD_VERSION
-
-void cluster_print_version(void)
-{
- printk(KERN_INFO "%s\n", VERSION_STR);
-}
-
-MODULE_DESCRIPTION(VERSION_STR);
-
-MODULE_VERSION(CLUSTER_BUILD_VERSION);
diff --git a/fs/ocfs2/cluster/ver.h b/fs/ocfs2/cluster/ver.h
deleted file mode 100644
index 32554c3382c..00000000000
--- a/fs/ocfs2/cluster/ver.h
+++ /dev/null
@@ -1,31 +0,0 @@
-/* -*- mode: c; c-basic-offset: 8; -*-
- * vim: noexpandtab sw=8 ts=8 sts=0:
- *
- * ver.h
- *
- * Function prototypes
- *
- * Copyright (C) 2005 Oracle. All rights reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public
- * License as published by the Free Software Foundation; either
- * version 2 of the License, or (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public
- * License along with this program; if not, write to the
- * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
- * Boston, MA 021110-1307, USA.
- */
-
-#ifndef O2CLUSTER_VER_H
-#define O2CLUSTER_VER_H
-
-void cluster_print_version(void);
-
-#endif /* O2CLUSTER_VER_H */
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-11-10 07:32:49 +0000