+ Fixes 2 cases caught by might_sleep testing.
+ Replace sleep_on with wait_event.
+ MOD_INC_USE_COUNT to prevent module unload vs smbiod thread exit race.
Please apply.
/Urban
diff -urN -X exclude linux-2.5.39-orig/fs/smbfs/Makefile linux-2.5.39-smbfs/fs/smbfs/Makefile
--- linux-2.5.39-orig/fs/smbfs/Makefile Tue Sep 17 21:39:11 2002
+++ linux-2.5.39-smbfs/fs/smbfs/Makefile Sun Sep 29 19:53:13 2002
@@ -15,7 +15,7 @@
#EXTRA_CFLAGS += -DSMBFS_DEBUG_VERBOSE
#EXTRA_CFLAGS += -DDEBUG_SMB_MALLOC
#EXTRA_CFLAGS += -DDEBUG_SMB_TIMESTAMP
-EXTRA_CFLAGS += -Werror
+#EXTRA_CFLAGS += -Werror
include $(TOPDIR)/Rules.make
diff -urN -X exclude linux-2.5.39-orig/fs/smbfs/inode.c linux-2.5.39-smbfs/fs/smbfs/inode.c
--- linux-2.5.39-orig/fs/smbfs/inode.c Sat Sep 21 15:48:18 2002
+++ linux-2.5.39-smbfs/fs/smbfs/inode.c Sun Sep 29 20:51:21 2002
@@ -434,10 +434,9 @@
{
struct smb_sb_info *server = SMB_SB(sb);
- smbiod_unregister_server(server);
-
smb_lock_server(server);
server->state = CONN_INVALID;
+ smbiod_unregister_server(server);
smb_close_socket(server);
diff -urN -X exclude linux-2.5.39-orig/fs/smbfs/proc.c linux-2.5.39-smbfs/fs/smbfs/proc.c
--- linux-2.5.39-orig/fs/smbfs/proc.c Sat Sep 21 15:48:18 2002
+++ linux-2.5.39-smbfs/fs/smbfs/proc.c Sun Sep 29 19:39:31 2002
@@ -1078,6 +1078,7 @@
VERBOSE("%s/%s R/W failed, error=%d, retrying R/O\n",
DENTRY_PATH(dentry), res);
mode = read_only;
+ req->rq_flags = 0;
goto retry;
}
goto out_free;
@@ -1649,6 +1650,7 @@
result = smb_set_rw(dentry,server);
if (result == 0) {
flag = 1;
+ req->rq_flags = 0;
goto retry;
}
}
diff -urN -X exclude linux-2.5.39-orig/fs/smbfs/request.c linux-2.5.39-smbfs/fs/smbfs/request.c
--- linux-2.5.39-orig/fs/smbfs/request.c Tue Sep 17 20:05:35 2002
+++ linux-2.5.39-smbfs/fs/smbfs/request.c Sun Sep 29 19:42:36 2002
@@ -332,8 +332,8 @@
smbiod_wake_up();
- /* FIXME: replace with a timeout-able wake_event_interruptible */
- timeleft = interruptible_sleep_on_timeout(&req->rq_wait, 30*HZ);
+ timeleft = wait_event_interruptible_timeout(req->rq_wait,
+ req->rq_flags & SMB_REQ_RECEIVED, 30*HZ);
if (!timeleft || signal_pending(current)) {
/*
* On timeout or on interrupt we want to try and remove the
@@ -346,7 +346,7 @@
}
smb_unlock_server(server);
}
-
+
if (!timeleft) {
PARANOIA("request [%p, mid=%d] timed out!\n",
req, req->rq_mid);
@@ -777,7 +777,7 @@
/*
* Response completely read. Drop any extra bytes sent by the server.
- * (Yes, servers sometimes add extra bytes to requests)
+ * (Yes, servers sometimes add extra bytes to responses)
*/
VERBOSE("smb_len: %d smb_read: %d\n",
server->smb_len, server->smb_read);
diff -urN -X exclude linux-2.5.39-orig/fs/smbfs/smbiod.c linux-2.5.39-smbfs/fs/smbfs/smbiod.c
--- linux-2.5.39-orig/fs/smbfs/smbiod.c Tue Sep 17 20:05:46 2002
+++ linux-2.5.39-smbfs/fs/smbfs/smbiod.c Sun Sep 29 20:59:17 2002
@@ -18,6 +18,7 @@
#include <linux/file.h>
#include <linux/dcache.h>
#include <linux/smp_lock.h>
+#include <linux/module.h>
#include <net/ip.h>
#include <linux/smb_fs.h>
@@ -31,7 +32,14 @@
#include "request.h"
#include "proto.h"
-static int smbiod_pid = -1;
+enum smbiod_state {
+ SMBIOD_DEAD,
+ SMBIOD_STARTING,
+ SMBIOD_RUNNING,
+};
+
+static enum smbiod_state smbiod_state = SMBIOD_DEAD;
+static pid_t smbiod_pid;
static DECLARE_WAIT_QUEUE_HEAD(smbiod_wait);
static LIST_HEAD(smb_servers);
static spinlock_t servers_lock = SPIN_LOCK_UNLOCKED;
@@ -41,14 +49,13 @@
static int smbiod(void *);
static void smbiod_start(void);
-static void smbiod_stop(void);
/*
* called when there's work for us to do
*/
void smbiod_wake_up()
{
- if (smbiod_pid == -1)
+ if (smbiod_state == SMBIOD_DEAD)
return;
set_bit(SMBIOD_DATA_READY, &smbiod_flags);
wake_up_interruptible(&smbiod_wait);
@@ -59,18 +66,16 @@
*/
static void smbiod_start()
{
- if (smbiod_pid != -1)
+ pid_t pid;
+ if (smbiod_state != SMBIOD_DEAD)
return;
- smbiod_pid = kernel_thread(smbiod, NULL, 0);
-}
+ smbiod_state = SMBIOD_STARTING;
+ spin_unlock(&servers_lock);
+ pid = kernel_thread(smbiod, NULL, 0);
-/*
- * stop smbiod if there are no open connections
- */
-static void smbiod_stop()
-{
- if (smbiod_pid != -1 && list_empty(&smb_servers))
- kill_proc(smbiod_pid, SIGKILL, 1);
+ spin_lock(&servers_lock);
+ smbiod_state = SMBIOD_RUNNING;
+ smbiod_pid = pid;
}
/*
@@ -86,19 +91,18 @@
}
/*
- * unregister a server & stop smbiod if necessary
+ * Unregister a server
+ * Must be called with the server lock held.
*/
void smbiod_unregister_server(struct smb_sb_info *server)
{
spin_lock(&servers_lock);
list_del_init(&server->entry);
VERBOSE("%p\n", server);
- smbiod_stop();
spin_unlock(&servers_lock);
- smb_lock_server(server);
+ smbiod_wake_up();
smbiod_flush(server);
- smb_unlock_server(server);
}
void smbiod_flush(struct smb_sb_info *server)
@@ -277,6 +281,7 @@
*/
static int smbiod(void *unused)
{
+ MOD_INC_USE_COUNT;
daemonize();
spin_lock_irq(¤t->sigmask_lock);
@@ -295,32 +300,40 @@
/* FIXME: Use poll? */
wait_event_interruptible(smbiod_wait,
test_bit(SMBIOD_DATA_READY, &smbiod_flags));
- if (signal_pending(current))
+ if (signal_pending(current)) {
+ spin_lock(&servers_lock);
+ smbiod_state = SMBIOD_DEAD;
+ spin_unlock(&servers_lock);
break;
+ }
clear_bit(SMBIOD_DATA_READY, &smbiod_flags);
- /*
- * We must hold the servers_lock while looking for servers
- * to check or else we have a race with put_super.
- */
spin_lock(&servers_lock);
+ if (list_empty(&smb_servers)) {
+ smbiod_state = SMBIOD_DEAD;
+ spin_unlock(&servers_lock);
+ break;
+ }
+
list_for_each_safe(pos, n, &smb_servers) {
server = list_entry(pos, struct smb_sb_info, entry);
VERBOSE("checking server %p\n", server);
- smb_lock_server(server);
- spin_unlock(&servers_lock);
- smbiod_doio(server);
+ if (server->state == CONN_VALID) {
+ spin_unlock(&servers_lock);
- smb_unlock_server(server);
- spin_lock(&servers_lock);
+ smb_lock_server(server);
+ smbiod_doio(server);
+ smb_unlock_server(server);
+
+ spin_lock(&servers_lock);
+ }
}
spin_unlock(&servers_lock);
}
VERBOSE("SMB Kernel thread exiting (%d) ...\n", current->pid);
- smbiod_pid = -1;
-
+ MOD_DEC_USE_COUNT;
return 0;
}
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Sep 30 2002 - 22:00:42 EST