This patch fixes a missing cleanup in cases where an error occurs
after an event has already been assigned to a group.
Before an event has not been put back to the "todo" list properly
(the list_del() for the working list was missing) and also was not
put back in cases of error in dazukofs_group_read().
There is still something missing:
If a filedescriptor has already been opened for an event, it has to
be closed before it is put back to the "working list".
I will send a fix for this issue with another patch.
Geschäftsführender Gesellschafter: Tjark Auerbach
Sitz der Gesellschaft: Tettnang
Handelsregister: Amtsgericht Ulm, HRB 630992
ALLGEMEINE GESCHÄFTSBEDINGUNGEN
Es gelten unsere Allgemeinen Geschäftsbedingungen
(AGB). Sie finden sie in der jeweils gültigen Fassung
im Internet unter http://www.avira.de/agb
***************************************************
diff -Nurp dazukofs-3.1.3-patch3/event.c dazukofs-3.1.3-patch4/event.c
--- dazukofs-3.1.3-patch3/event.c 2010-07-06 17:28:05.000000000 +0200
+++ dazukofs-3.1.3-patch4/event.c 2010-07-08 18:10:41.000000000 +0200
@@ -28,50 +28,17 @@
#include <linux/pid.h>
#include <linux/slab.h>
+#include "event.h"
#include "dev.h"
#include "dazukofs_fs.h"
struct dazukofs_proc {
- struct list_head list;
+ struct list_head list;
struct pid *proc_id;
int within_list;
};
-struct dazukofs_event {
- unsigned long event_id;
- struct dentry *dentry;
- struct vfsmount *mnt;
- struct pid *proc_id;
- wait_queue_head_t queue;
-
- /* protects: deny, deprecated, assigned */
- struct mutex assigned_mutex;
-
- int deny;
- int deprecated;
- int assigned;
-};
-struct dazukofs_event_container {
- struct list_head list;
- struct dazukofs_event *event;
- struct file *file;
- int fd;
-};
-
-struct dazukofs_group {
- struct list_head list;
- char *name;
- size_t name_length;
- unsigned long group_id;
- struct dazukofs_event_container todo_list;
- wait_queue_head_t queue;
- struct dazukofs_event_container working_list;
- atomic_t use_count;
- int tracking;
- int track_count;
- int deprecated;
-};
static struct dazukofs_group group_list;
static int group_count;
@@ -899,6 +866,15 @@ out:
return ret;
}
+
+void set_event_pending(struct dazukofs_group *grp,
+ struct dazukofs_event_container *ec)
+{
+ mutex_lock(&work_mutex);
+ list_add(&ec->list, &grp->working_list.list);
+ mutex_unlock(&work_mutex);
+}
+
/**
* unclaim_event - return an event to the todo list
* @grp: group to which the event is assigned
@@ -907,7 +883,7 @@ out:
* Description: This function puts the given event container on the todo
* list and wake the group.
*/
-static void unclaim_event(struct dazukofs_group *grp,
+void unclaim_event(struct dazukofs_group *grp,
struct dazukofs_event_container *ec)
{
/* put the event on the todo list */
@@ -938,7 +914,6 @@ static struct dazukofs_event_container *
ec = list_first_entry(&grp->todo_list.list,
struct dazukofs_event_container, list);
list_del(&ec->list);
- list_add(&ec->list, &grp->working_list.list);
}
mutex_unlock(&work_mutex);
@@ -1034,7 +1009,6 @@ static int is_event_available(struct daz
return ret;
}
-/**
* dazukofs_get_event - get an event to process
* @group_id: id of the group we belong to
* @event_id: to be filled in with the new event id
@@ -1047,8 +1021,8 @@ static int is_event_available(struct daz
*
* Returns 0 on success.
*/
-int dazukofs_get_event(unsigned long group_id, unsigned long *event_id,
- int *fd, pid_t *pid)
+int dazukofs_get_event(unsigned long group_id, struct dazukofs_group **pgrp,
+ struct dazukofs_event_container **pec)
{
struct dazukofs_group *grp = NULL;
struct dazukofs_event_container *ec;
@@ -1067,10 +1041,8 @@ int dazukofs_get_event(unsigned long gro
}
mutex_unlock(&work_mutex);
- if (!found) {
- ret = -EFAULT;
- goto out;
- }
+ if (!found)
+ return -EINVAL;
while (1) {
ret = wait_event_freezable(grp->queue,
@@ -1087,14 +1059,9 @@ int dazukofs_get_event(unsigned long gro
ec = claim_event(grp);
if (ec) {
ret = open_file(ec);
- if (ret == 0) {
- *event_id = ec->event->event_id;
- *fd = ec->fd;
-
- /* set to 0 if not within namespace */
- *pid = pid_vnr(ec->event->proc_id);
+ if (ret == 0)
break;
- } else {
+ else {
unclaim_event(grp, ec);
if (ret == -ENFILE) {
/* The registered process may not open
@@ -1103,8 +1070,16 @@ int dazukofs_get_event(unsigned long gro
}
}
}
+ /* we did not get the event */
+ atomic_dec(&grp->use_count);
}
- atomic_dec(&grp->use_count);
-out:
- return ret;
+
+ if (ret) {
+ atomic_dec(&grp->use_count);
+ return ret;
+ }
+
+ *pgrp = grp;
+ *pec = ec;
+ return 0;
}
diff -Nurp dazukofs-3.1.3-patch3/event.h dazukofs-3.1.3-patch4/event.h
--- dazukofs-3.1.3-patch3/event.h 2010-07-06 17:28:05.000000000 +0200
+++ dazukofs-3.1.3-patch4/event.h 2010-07-08 16:39:46.000000000 +0200
@@ -21,11 +21,56 @@
#ifndef __EVENT_H
#define __EVENT_H
+#include <linux/list.h>
+#include <linux/wait.h>
+
+
+struct dazukofs_event {
+ unsigned long event_id;
+ struct dentry *dentry;
+ struct vfsmount *mnt;
+ struct pid *proc_id;
+ wait_queue_head_t queue;
+
+ /* protects: deny, deprecated, assigned */
+ struct mutex assigned_mutex;
+
+ int deny;
+ int deprecated;
+ int assigned;
+};
+
+struct dazukofs_event_container {
+ struct list_head list;
+ struct dazukofs_event *event;
+ struct file *file;
+ int fd;
+};
+
+struct dazukofs_group {
+ struct list_head list;
+ char *name;
+ size_t name_length;
+ unsigned long group_id;
+ struct dazukofs_event_container todo_list;
+ wait_queue_head_t queue;
+ struct dazukofs_event_container working_list;
+ atomic_t use_count;
+ int tracking;
+ int track_count;
+ int deprecated;
+};
+
extern int dazukofs_init_events(void);
extern void dazukofs_destroy_events(void);
-extern int dazukofs_get_event(unsigned long group_id,
- unsigned long *event_id, int *fd, pid_t *pid);
+extern void set_event_pending(struct dazukofs_group *grp,
+ struct dazukofs_event_container *ec);
+extern void unclaim_event(struct dazukofs_group *grp,
+ struct dazukofs_event_container *ec);
+extern int dazukofs_get_event(unsigned long group_id,
+ struct dazukofs_group **pgrp,
+ struct dazukofs_event_container **pec);
extern int dazukofs_return_event(unsigned long group_id,
unsigned long event_id, int deny);
diff -Nurp dazukofs-3.1.3-patch3/group_dev.c dazukofs-3.1.3-patch4/group_dev.c
--- dazukofs-3.1.3-patch3/group_dev.c 2010-07-06 17:40:52.000000000 +0200
+++ dazukofs-3.1.3-patch4/group_dev.c 2010-07-08 15:30:19.000000000 +0200
@@ -58,33 +58,41 @@ static ssize_t dazukofs_group_read(int g
{
#define DAZUKOFS_MIN_READ_BUFFER 43
char tmp[DAZUKOFS_MIN_READ_BUFFER];
- ssize_t tmp_used;
- pid_t pid;
- int fd;
- int err;
- unsigned long event_id;
+ struct dazukofs_event_container *ec;
+ struct dazukofs_group *grp;
+ ssize_t rv;
if (length < DAZUKOFS_MIN_READ_BUFFER)
return -EINVAL;
- err = dazukofs_get_event(group_id, &event_id, &fd, &pid);
- if (err) {
+ rv = dazukofs_get_event(group_id, &grp, &ec);
+ if (rv) {
/* convert some errors to acceptable read(2) errno values */
- if (err == -ENFILE)
- return -EIO;
- return err;
+ if (rv == -ENFILE)
+ rv = -EIO;
+ return rv;
}
- tmp_used = snprintf(tmp, sizeof(tmp)-1, "id=%lu\nfd=%d\npid=%d\n",
- event_id, fd, pid);
- if (tmp_used >= sizeof(tmp))
- return -EINVAL;
+ rv = snprintf(tmp, sizeof(tmp)-1, "id=%lu\nfd=%d\npid=%d\n",
+ ec->event->event_id, ec->fd,
+ pid_vnr(ec->event->proc_id));
+ if (rv >= sizeof(tmp)) {
+ unclaim_event(grp, ec);
+ rv = -EINVAL;
+ goto err;
+ }
- if (copy_to_user(buffer, tmp, tmp_used))
- return -EFAULT;
+ if (copy_to_user(buffer, tmp, rv)) {
+ unclaim_event(grp, ec);
+ rv = -EFAULT;
+ goto err;
+ }
+ set_event_pending(grp, ec);
- return tmp_used;
+err:
+ atomic_dec(&grp->use_count);
+ return rv;
}
static ssize_t dazukofs_group_write(int group_id, struct file *file,
_______________________________________________
Dazuko-devel mailing list
Dazuko-devel@nongnu.org
http://lists.nongnu.org/mailman/listinfo/dazuko-devel
