AFS: implement file locking
authorDavid Howells <dhowells@redhat.com>
Mon, 16 Jul 2007 06:40:12 +0000 (23:40 -0700)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Mon, 16 Jul 2007 16:05:43 +0000 (09:05 -0700)
Implement file locking for AFS.

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 files changed:
fs/afs/Makefile
fs/afs/afs.h
fs/afs/afs_fs.h
fs/afs/callback.c
fs/afs/dir.c
fs/afs/file.c
fs/afs/flock.c [new file with mode: 0644]
fs/afs/fsclient.c
fs/afs/internal.h
fs/afs/main.c
fs/afs/misc.c
fs/afs/super.c
fs/afs/vnode.c
include/linux/fs.h

index 73ce561f3ea0c2ea5cd901156a0c3cbf730ed200..a66671082cfbb98d37277e73e3255c70d8415946 100644 (file)
@@ -8,6 +8,7 @@ kafs-objs := \
        cmservice.o \
        dir.o \
        file.o \
+       flock.o \
        fsclient.o \
        inode.o \
        main.o \
index 2452579481404fe702aa852b39ab404254773c5b..c548aa346f0dd6f9a7aed6664abb32bcbb6ddd79 100644 (file)
@@ -37,6 +37,13 @@ typedef enum {
        AFS_FTYPE_SYMLINK       = 3,
 } afs_file_type_t;
 
+typedef enum {
+       AFS_LOCK_READ           = 0,    /* read lock request */
+       AFS_LOCK_WRITE          = 1,    /* write lock request */
+} afs_lock_type_t;
+
+#define AFS_LOCKWAIT           (5 * 60) /* time until a lock times out (seconds) */
+
 /*
  * AFS file identifier
  */
@@ -120,6 +127,7 @@ struct afs_file_status {
        struct afs_fid          parent;         /* parent dir ID for non-dirs only */
        time_t                  mtime_client;   /* last time client changed data */
        time_t                  mtime_server;   /* last time server changed data */
+       s32                     lock_count;     /* file lock count (0=UNLK -1=WRLCK +ve=#RDLCK */
 };
 
 /*
index a18c374ebe080d9cbbc9544bd2e4dbdbf42ec62a..eb647323d8f03a3f6634d4f83d0aa648f9093201 100644 (file)
@@ -31,6 +31,9 @@ enum AFS_FS_Operations {
        FSGETVOLUMEINFO         = 148,  /* AFS Get information about a volume */
        FSGETVOLUMESTATUS       = 149,  /* AFS Get volume status information */
        FSGETROOTVOLUME         = 151,  /* AFS Get root volume name */
+       FSSETLOCK               = 156,  /* AFS Request a file lock */
+       FSEXTENDLOCK            = 157,  /* AFS Extend a file lock */
+       FSRELEASELOCK           = 158,  /* AFS Release a file lock */
        FSLOOKUP                = 161,  /* AFS lookup file in directory */
        FSFETCHDATA64           = 65537, /* AFS Fetch file data */
        FSSTOREDATA64           = 65538, /* AFS Store file data */
index bacf518c6fa80d891f24e1168a82abf1a48d2d62..b8243945818de3abd394435953acb8ed8e88eeef 100644 (file)
@@ -125,6 +125,9 @@ static void afs_break_callback(struct afs_server *server,
                spin_unlock(&server->cb_lock);
 
                queue_work(afs_callback_update_worker, &vnode->cb_broken_work);
+               if (list_empty(&vnode->granted_locks) &&
+                   !list_empty(&vnode->pending_locks))
+                       afs_lock_may_be_available(vnode);
                spin_unlock(&vnode->lock);
        }
 }
index 546c59522eb16e0307070102df305cc05766f419..33fe39ad4e0327a91e925c10ad54629c5f7c763a 100644 (file)
@@ -44,6 +44,7 @@ const struct file_operations afs_dir_file_operations = {
        .open           = afs_dir_open,
        .release        = afs_release,
        .readdir        = afs_readdir,
+       .lock           = afs_lock,
 };
 
 const struct inode_operations afs_dir_inode_operations = {
index aede7eb66dd4f970a74021b41487a5a0a70c0616..525f7c56e068d877e95dedea3a15dafc116fdb41 100644 (file)
@@ -34,6 +34,8 @@ const struct file_operations afs_file_operations = {
        .mmap           = generic_file_readonly_mmap,
        .splice_read    = generic_file_splice_read,
        .fsync          = afs_fsync,
+       .lock           = afs_lock,
+       .flock          = afs_flock,
 };
 
 const struct inode_operations afs_file_inode_operations = {
diff --git a/fs/afs/flock.c b/fs/afs/flock.c
new file mode 100644 (file)
index 0000000..8f07f8d
--- /dev/null
@@ -0,0 +1,558 @@
+/* AFS file locking support
+ *
+ * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ *
+ * 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.
+ */
+
+#include <linux/smp_lock.h>
+#include "internal.h"
+
+#define AFS_LOCK_GRANTED       0
+#define AFS_LOCK_PENDING       1
+
+static void afs_fl_copy_lock(struct file_lock *new, struct file_lock *fl);
+static void afs_fl_release_private(struct file_lock *fl);
+
+static struct workqueue_struct *afs_lock_manager;
+
+static struct file_lock_operations afs_lock_ops = {
+       .fl_copy_lock           = afs_fl_copy_lock,
+       .fl_release_private     = afs_fl_release_private,
+};
+
+/*
+ * initialise the lock manager thread if it isn't already running
+ */
+static int afs_init_lock_manager(void)
+{
+       if (!afs_lock_manager) {
+               afs_lock_manager = create_singlethread_workqueue("kafs_lockd");
+               if (!afs_lock_manager)
+                       return -ENOMEM;
+       }
+       return 0;
+}
+
+/*
+ * destroy the lock manager thread if it's running
+ */
+void __exit afs_kill_lock_manager(void)
+{
+       if (afs_lock_manager)
+               destroy_workqueue(afs_lock_manager);
+}
+
+/*
+ * if the callback is broken on this vnode, then the lock may now be available
+ */
+void afs_lock_may_be_available(struct afs_vnode *vnode)
+{
+       _enter("{%x:%u}", vnode->fid.vid, vnode->fid.vnode);
+
+       queue_delayed_work(afs_lock_manager, &vnode->lock_work, 0);
+}
+
+/*
+ * the lock will time out in 5 minutes unless we extend it, so schedule
+ * extension in a bit less than that time
+ */
+static void afs_schedule_lock_extension(struct afs_vnode *vnode)
+{
+       queue_delayed_work(afs_lock_manager, &vnode->lock_work,
+                          AFS_LOCKWAIT * HZ / 2);
+}
+
+/*
+ * do work for a lock, including:
+ * - probing for a lock we're waiting on but didn't get immediately
+ * - extending a lock that's close to timing out
+ */
+void afs_lock_work(struct work_struct *work)
+{
+       struct afs_vnode *vnode =
+               container_of(work, struct afs_vnode, lock_work.work);
+       struct file_lock *fl;
+       afs_lock_type_t type;
+       struct key *key;
+       int ret;
+
+       _enter("{%x:%u}", vnode->fid.vid, vnode->fid.vnode);
+
+       spin_lock(&vnode->lock);
+
+       if (test_bit(AFS_VNODE_UNLOCKING, &vnode->flags)) {
+               _debug("unlock");
+               spin_unlock(&vnode->lock);
+
+               /* attempt to release the server lock; if it fails, we just
+                * wait 5 minutes and it'll time out anyway */
+               ret = afs_vnode_release_lock(vnode, vnode->unlock_key);
+               if (ret < 0)
+                       printk(KERN_WARNING "AFS:"
+                              " Failed to release lock on {%x:%x} error %d\n",
+                              vnode->fid.vid, vnode->fid.vnode, ret);
+
+               spin_lock(&vnode->lock);
+               key_put(vnode->unlock_key);
+               vnode->unlock_key = NULL;
+               clear_bit(AFS_VNODE_UNLOCKING, &vnode->flags);
+       }
+
+       /* if we've got a lock, then it must be time to extend that lock as AFS
+        * locks time out after 5 minutes */
+       if (!list_empty(&vnode->granted_locks)) {
+               _debug("extend");
+
+               if (test_and_set_bit(AFS_VNODE_LOCKING, &vnode->flags))
+                       BUG();
+               fl = list_entry(vnode->granted_locks.next,
+                               struct file_lock, fl_u.afs.link);
+               key = key_get(fl->fl_file->private_data);
+               spin_unlock(&vnode->lock);
+
+               ret = afs_vnode_extend_lock(vnode, key);
+               clear_bit(AFS_VNODE_LOCKING, &vnode->flags);
+               key_put(key);
+               switch (ret) {
+               case 0:
+                       afs_schedule_lock_extension(vnode);
+                       break;
+               default:
+                       /* ummm... we failed to extend the lock - retry
+                        * extension shortly */
+                       printk(KERN_WARNING "AFS:"
+                              " Failed to extend lock on {%x:%x} error %d\n",
+                              vnode->fid.vid, vnode->fid.vnode, ret);
+                       queue_delayed_work(afs_lock_manager, &vnode->lock_work,
+                                          HZ * 10);
+                       break;
+               }
+               _leave(" [extend]");
+               return;
+       }
+
+       /* if we don't have a granted lock, then we must've been called back by
+        * the server, and so if might be possible to get a lock we're
+        * currently waiting for */
+       if (!list_empty(&vnode->pending_locks)) {
+               _debug("get");
+
+               if (test_and_set_bit(AFS_VNODE_LOCKING, &vnode->flags))
+                       BUG();
+               fl = list_entry(vnode->pending_locks.next,
+                               struct file_lock, fl_u.afs.link);
+               key = key_get(fl->fl_file->private_data);
+               type = (fl->fl_type == F_RDLCK) ?
+                       AFS_LOCK_READ : AFS_LOCK_WRITE;
+               spin_unlock(&vnode->lock);
+
+               ret = afs_vnode_set_lock(vnode, key, type);
+               clear_bit(AFS_VNODE_LOCKING, &vnode->flags);
+               switch (ret) {
+               case -EWOULDBLOCK:
+                       _debug("blocked");
+                       break;
+               case 0:
+                       _debug("acquired");
+                       if (type == AFS_LOCK_READ)
+                               set_bit(AFS_VNODE_READLOCKED, &vnode->flags);
+                       else
+                               set_bit(AFS_VNODE_WRITELOCKED, &vnode->flags);
+                       ret = AFS_LOCK_GRANTED;
+               default:
+                       spin_lock(&vnode->lock);
+                       /* the pending lock may have been withdrawn due to a
+                        * signal */
+                       if (list_entry(vnode->pending_locks.next,
+                                      struct file_lock, fl_u.afs.link) == fl) {
+                               fl->fl_u.afs.state = ret;
+                               if (ret == AFS_LOCK_GRANTED)
+                                       list_move_tail(&fl->fl_u.afs.link,
+                                                      &vnode->granted_locks);
+                               else
+                                       list_del_init(&fl->fl_u.afs.link);
+                               wake_up(&fl->fl_wait);
+                               spin_unlock(&vnode->lock);
+                       } else {
+                               _debug("withdrawn");
+                               clear_bit(AFS_VNODE_READLOCKED, &vnode->flags);
+                               clear_bit(AFS_VNODE_WRITELOCKED, &vnode->flags);
+                               spin_unlock(&vnode->lock);
+                               afs_vnode_release_lock(vnode, key);
+                               if (!list_empty(&vnode->pending_locks))
+                                       afs_lock_may_be_available(vnode);
+                       }
+                       break;
+               }
+               key_put(key);
+               _leave(" [pend]");
+               return;
+       }
+
+       /* looks like the lock request was withdrawn on a signal */
+       spin_unlock(&vnode->lock);
+       _leave(" [no locks]");
+}
+
+/*
+ * pass responsibility for the unlocking of a vnode on the server to the
+ * manager thread, lest a pending signal in the calling thread interrupt
+ * AF_RXRPC
+ * - the caller must hold the vnode lock
+ */
+static void afs_defer_unlock(struct afs_vnode *vnode, struct key *key)
+{
+       cancel_delayed_work(&vnode->lock_work);
+       if (!test_and_clear_bit(AFS_VNODE_READLOCKED, &vnode->flags) &&
+           !test_and_clear_bit(AFS_VNODE_WRITELOCKED, &vnode->flags))
+               BUG();
+       if (test_and_set_bit(AFS_VNODE_UNLOCKING, &vnode->flags))
+               BUG();
+       vnode->unlock_key = key_get(key);
+       afs_lock_may_be_available(vnode);
+}
+
+/*
+ * request a lock on a file on the server
+ */
+static int afs_do_setlk(struct file *file, struct file_lock *fl)
+{
+       struct afs_vnode *vnode = AFS_FS_I(file->f_mapping->host);
+       afs_lock_type_t type;
+       struct key *key = file->private_data;
+       int ret;
+
+       _enter("{%x:%u},%u", vnode->fid.vid, vnode->fid.vnode, fl->fl_type);
+
+       /* only whole-file locks are supported */
+       if (fl->fl_start != 0 || fl->fl_end != OFFSET_MAX)
+               return -EINVAL;
+
+       ret = afs_init_lock_manager();
+       if (ret < 0)
+               return ret;
+
+       fl->fl_ops = &afs_lock_ops;
+       INIT_LIST_HEAD(&fl->fl_u.afs.link);
+       fl->fl_u.afs.state = AFS_LOCK_PENDING;
+
+       type = (fl->fl_type == F_RDLCK) ? AFS_LOCK_READ : AFS_LOCK_WRITE;
+
+       lock_kernel();
+
+       /* make sure we've got a callback on this file and that our view of the
+        * data version is up to date */
+       ret = afs_vnode_fetch_status(vnode, NULL, key);
+       if (ret < 0)
+               goto error;
+
+       if (vnode->status.lock_count != 0 && !(fl->fl_flags & FL_SLEEP)) {
+               ret = -EAGAIN;
+               goto error;
+       }
+
+       spin_lock(&vnode->lock);
+
+       if (list_empty(&vnode->pending_locks)) {
+               /* if there's no-one else with a lock on this vnode, then we
+                * need to ask the server for a lock */
+               if (list_empty(&vnode->granted_locks)) {
+                       _debug("not locked");
+                       ASSERTCMP(vnode->flags &
+                                 ((1 << AFS_VNODE_LOCKING) |
+                                  (1 << AFS_VNODE_READLOCKED) |
+                                  (1 << AFS_VNODE_WRITELOCKED)), ==, 0);
+                       list_add_tail(&fl->fl_u.afs.link, &vnode->pending_locks);
+                       set_bit(AFS_VNODE_LOCKING, &vnode->flags);
+                       spin_unlock(&vnode->lock);
+
+                       ret = afs_vnode_set_lock(vnode, key, type);
+                       clear_bit(AFS_VNODE_LOCKING, &vnode->flags);
+                       switch (ret) {
+                       case 0:
+                               goto acquired_server_lock;
+                       case -EWOULDBLOCK:
+                               spin_lock(&vnode->lock);
+                               ASSERT(list_empty(&vnode->granted_locks));
+                               ASSERTCMP(vnode->pending_locks.next, ==,
+                                         &fl->fl_u.afs.link);
+                               goto wait;
+                       default:
+                               spin_lock(&vnode->lock);
+                               list_del_init(&fl->fl_u.afs.link);
+                               spin_unlock(&vnode->lock);
+                               goto error;
+                       }
+               }
+
+               /* if we've already got a readlock on the server and no waiting
+                * writelocks, then we might be able to instantly grant another
+                * readlock */
+               if (type == AFS_LOCK_READ &&
+                   vnode->flags & (1 << AFS_VNODE_READLOCKED)) {
+                       _debug("instant readlock");
+                       ASSERTCMP(vnode->flags &
+                                 ((1 << AFS_VNODE_LOCKING) |
+                                  (1 << AFS_VNODE_WRITELOCKED)), ==, 0);
+                       ASSERT(!list_empty(&vnode->granted_locks));
+                       goto sharing_existing_lock;
+               }
+       }
+
+       /* otherwise, we need to wait for a local lock to become available */
+       _debug("wait local");
+       list_add_tail(&fl->fl_u.afs.link, &vnode->pending_locks);
+wait:
+       if (!(fl->fl_flags & FL_SLEEP)) {
+               _debug("noblock");
+               ret = -EAGAIN;
+               goto abort_attempt;
+       }
+       spin_unlock(&vnode->lock);
+
+       /* now we need to sleep and wait for the lock manager thread to get the
+        * lock from the server */
+       _debug("sleep");
+       ret = wait_event_interruptible(fl->fl_wait,
+                                      fl->fl_u.afs.state <= AFS_LOCK_GRANTED);
+       if (fl->fl_u.afs.state <= AFS_LOCK_GRANTED) {
+               ret = fl->fl_u.afs.state;
+               if (ret < 0)
+                       goto error;
+               spin_lock(&vnode->lock);
+               goto given_lock;
+       }
+
+       /* we were interrupted, but someone may still be in the throes of
+        * giving us the lock */
+       _debug("intr");
+       ASSERTCMP(ret, ==, -ERESTARTSYS);
+
+       spin_lock(&vnode->lock);
+       if (fl->fl_u.afs.state <= AFS_LOCK_GRANTED) {
+               ret = fl->fl_u.afs.state;
+               if (ret < 0) {
+                       spin_unlock(&vnode->lock);
+                       goto error;
+               }
+               goto given_lock;
+       }
+
+abort_attempt:
+       /* we aren't going to get the lock, either because we're unwilling to
+        * wait, or because some signal happened */
+       _debug("abort");
+       if (list_empty(&vnode->granted_locks) &&
+           vnode->pending_locks.next == &fl->fl_u.afs.link) {
+               if (vnode->pending_locks.prev != &fl->fl_u.afs.link) {
+                       /* kick the next pending lock into having a go */
+                       list_del_init(&fl->fl_u.afs.link);
+                       afs_lock_may_be_available(vnode);
+               }
+       } else {
+               list_del_init(&fl->fl_u.afs.link);
+       }
+       spin_unlock(&vnode->lock);
+       goto error;
+
+acquired_server_lock:
+       /* we've acquired a server lock, but it needs to be renewed after 5
+        * mins */
+       spin_lock(&vnode->lock);
+       afs_schedule_lock_extension(vnode);
+       if (type == AFS_LOCK_READ)
+               set_bit(AFS_VNODE_READLOCKED, &vnode->flags);
+       else
+               set_bit(AFS_VNODE_WRITELOCKED, &vnode->flags);
+sharing_existing_lock:
+       /* the lock has been granted as far as we're concerned... */
+       fl->fl_u.afs.state = AFS_LOCK_GRANTED;
+       list_move_tail(&fl->fl_u.afs.link, &vnode->granted_locks);
+given_lock:
+       /* ... but we do still need to get the VFS's blessing */
+       ASSERT(!(vnode->flags & (1 << AFS_VNODE_LOCKING)));
+       ASSERT((vnode->flags & ((1 << AFS_VNODE_READLOCKED) |
+                               (1 << AFS_VNODE_WRITELOCKED))) != 0);
+       ret = posix_lock_file(file, fl, NULL);
+       if (ret < 0)
+               goto vfs_rejected_lock;
+       spin_unlock(&vnode->lock);
+
+       /* again, make sure we've got a callback on this file and, again, make
+        * sure that our view of the data version is up to date (we ignore
+        * errors incurred here and deal with the consequences elsewhere) */
+       afs_vnode_fetch_status(vnode, NULL, key);
+
+error:
+       unlock_kernel();
+       _leave(" = %d", ret);
+       return ret;
+
+vfs_rejected_lock:
+       /* the VFS rejected the lock we just obtained, so we have to discard
+        * what we just got */
+       _debug("vfs refused %d", ret);
+       list_del_init(&fl->fl_u.afs.link);
+       if (list_empty(&vnode->granted_locks))
+               afs_defer_unlock(vnode, key);
+       spin_unlock(&vnode->lock);
+       goto abort_attempt;
+}
+
+/*
+ * unlock on a file on the server
+ */
+static int afs_do_unlk(struct file *file, struct file_lock *fl)
+{
+       struct afs_vnode *vnode = AFS_FS_I(file->f_mapping->host);
+       struct key *key = file->private_data;
+       int ret;
+
+       _enter("{%x:%u},%u", vnode->fid.vid, vnode->fid.vnode, fl->fl_type);
+
+       /* only whole-file unlocks are supported */
+       if (fl->fl_start != 0 || fl->fl_end != OFFSET_MAX)
+               return -EINVAL;
+
+       fl->fl_ops = &afs_lock_ops;
+       INIT_LIST_HEAD(&fl->fl_u.afs.link);
+       fl->fl_u.afs.state = AFS_LOCK_PENDING;
+
+       spin_lock(&vnode->lock);
+       ret = posix_lock_file(file, fl, NULL);
+       if (ret < 0) {
+               spin_unlock(&vnode->lock);
+               _leave(" = %d [vfs]", ret);
+               return ret;
+       }
+
+       /* discard the server lock only if all granted locks are gone */
+       if (list_empty(&vnode->granted_locks))
+               afs_defer_unlock(vnode, key);
+       spin_unlock(&vnode->lock);
+       _leave(" = 0");
+       return 0;
+}
+
+/*
+ * return information about a lock we currently hold, if indeed we hold one
+ */
+static int afs_do_getlk(struct file *file, struct file_lock *fl)
+{
+       struct afs_vnode *vnode = AFS_FS_I(file->f_mapping->host);
+       struct key *key = file->private_data;
+       int ret, lock_count;
+
+       _enter("");
+
+       fl->fl_type = F_UNLCK;
+
+       mutex_lock(&vnode->vfs_inode.i_mutex);
+
+       /* check local lock records first */
+       ret = 0;
+       if (posix_test_lock(file, fl) == 0) {
+               /* no local locks; consult the server */
+               ret = afs_vnode_fetch_status(vnode, NULL, key);
+               if (ret < 0)
+                       goto error;
+               lock_count = vnode->status.lock_count;
+               if (lock_count) {
+                       if (lock_count > 0)
+                               fl->fl_type = F_RDLCK;
+                       else
+                               fl->fl_type = F_WRLCK;
+                       fl->fl_start = 0;
+                       fl->fl_end = OFFSET_MAX;
+               }
+       }
+
+error:
+       mutex_unlock(&vnode->vfs_inode.i_mutex);
+       _leave(" = %d [%hd]", ret, fl->fl_type);
+       return ret;
+}
+
+/*
+ * manage POSIX locks on a file
+ */
+int afs_lock(struct file *file, int cmd, struct file_lock *fl)
+{
+       struct afs_vnode *vnode = AFS_FS_I(file->f_dentry->d_inode);
+
+       _enter("{%x:%u},%d,{t=%x,fl=%x,r=%Ld:%Ld}",
+              vnode->fid.vid, vnode->fid.vnode, cmd,
+              fl->fl_type, fl->fl_flags,
+              (long long) fl->fl_start, (long long) fl->fl_end);
+
+       /* AFS doesn't support mandatory locks */
+       if ((vnode->vfs_inode.i_mode & (S_ISGID | S_IXGRP)) == S_ISGID &&
+           fl->fl_type != F_UNLCK)
+               return -ENOLCK;
+
+       if (IS_GETLK(cmd))
+               return afs_do_getlk(file, fl);
+       if (fl->fl_type == F_UNLCK)
+               return afs_do_unlk(file, fl);
+       return afs_do_setlk(file, fl);
+}
+
+/*
+ * manage FLOCK locks on a file
+ */
+int afs_flock(struct file *file, int cmd, struct file_lock *fl)
+{
+       struct afs_vnode *vnode = AFS_FS_I(file->f_dentry->d_inode);
+
+       _enter("{%x:%u},%d,{t=%x,fl=%x}",
+              vnode->fid.vid, vnode->fid.vnode, cmd,
+              fl->fl_type, fl->fl_flags);
+
+       /*
+        * No BSD flocks over NFS allowed.
+        * Note: we could try to fake a POSIX lock request here by
+        * using ((u32) filp | 0x80000000) or some such as the pid.
+        * Not sure whether that would be unique, though, or whether
+        * that would break in other places.
+        */
+       if (!(fl->fl_flags & FL_FLOCK))
+               return -ENOLCK;
+
+       /* we're simulating flock() locks using posix locks on the server */
+       fl->fl_owner = (fl_owner_t) file;
+       fl->fl_start = 0;
+       fl->fl_end = OFFSET_MAX;
+
+       if (fl->fl_type == F_UNLCK)
+               return afs_do_unlk(file, fl);
+       return afs_do_setlk(file, fl);
+}
+
+/*
+ * the POSIX lock management core VFS code copies the lock record and adds the
+ * copy into its own list, so we need to add that copy to the vnode's lock
+ * queue in the same place as the original (which will be deleted shortly
+ * after)
+ */
+static void afs_fl_copy_lock(struct file_lock *new, struct file_lock *fl)
+{
+       _enter("");
+
+       list_add(&new->fl_u.afs.link, &fl->fl_u.afs.link);
+}
+
+/*
+ * need to remove this lock from the vnode queue when it's removed from the
+ * VFS's list
+ */
+static void afs_fl_release_private(struct file_lock *fl)
+{
+       _enter("");
+
+       list_del_init(&fl->fl_u.afs.link);
+}
index 5dff1308b6f0da5475a20701063036a78bf21b90..023b95b0d9d7612b41ef895c4a2989513fa6b47a 100644 (file)
@@ -67,7 +67,7 @@ static void xdr_decode_AFSFetchStatus(const __be32 **_bp,
        EXTRACT(status->group);
        bp++; /* sync counter */
        data_version |= (u64) ntohl(*bp++) << 32;
-       bp++; /* lock count */
+       EXTRACT(status->lock_count);
        size |= (u64) ntohl(*bp++) << 32;
        bp++; /* spare 4 */
        *_bp = bp;
@@ -1748,3 +1748,156 @@ int afs_fs_get_volume_status(struct afs_server *server,
 
        return afs_make_call(&server->addr, call, GFP_NOFS, wait_mode);
 }
+
+/*
+ * deliver reply data to an FS.SetLock, FS.ExtendLock or FS.ReleaseLock
+ */
+static int afs_deliver_fs_xxxx_lock(struct afs_call *call,
+                                   struct sk_buff *skb, bool last)
+{
+       const __be32 *bp;
+
+       _enter("{%u},{%u},%d", call->unmarshall, skb->len, last);
+
+       afs_transfer_reply(call, skb);
+       if (!last)
+               return 0;
+
+       if (call->reply_size != call->reply_max)
+               return -EBADMSG;
+
+       /* unmarshall the reply once we've received all of it */
+       bp = call->buffer;
+       /* xdr_decode_AFSVolSync(&bp, call->replyX); */
+
+       _leave(" = 0 [done]");
+       return 0;
+}
+
+/*
+ * FS.SetLock operation type
+ */
+static const struct afs_call_type afs_RXFSSetLock = {
+       .name           = "FS.SetLock",
+       .deliver        = afs_deliver_fs_xxxx_lock,
+       .abort_to_error = afs_abort_to_error,
+       .destructor     = afs_flat_call_destructor,
+};
+
+/*
+ * FS.ExtendLock operation type
+ */
+static const struct afs_call_type afs_RXFSExtendLock = {
+       .name           = "FS.ExtendLock",
+       .deliver        = afs_deliver_fs_xxxx_lock,
+       .abort_to_error = afs_abort_to_error,
+       .destructor     = afs_flat_call_destructor,
+};
+
+/*
+ * FS.ReleaseLock operation type
+ */
+static const struct afs_call_type afs_RXFSReleaseLock = {
+       .name           = "FS.ReleaseLock",
+       .deliver        = afs_deliver_fs_xxxx_lock,
+       .abort_to_error = afs_abort_to_error,
+       .destructor     = afs_flat_call_destructor,
+};
+
+/*
+ * get a lock on a file
+ */
+int afs_fs_set_lock(struct afs_server *server,
+                   struct key *key,
+                   struct afs_vnode *vnode,
+                   afs_lock_type_t type,
+                   const struct afs_wait_mode *wait_mode)
+{
+       struct afs_call *call;
+       __be32 *bp;
+
+       _enter("");
+
+       call = afs_alloc_flat_call(&afs_RXFSSetLock, 5 * 4, 6 * 4);
+       if (!call)
+               return -ENOMEM;
+
+       call->key = key;
+       call->reply = vnode;
+       call->service_id = FS_SERVICE;
+       call->port = htons(AFS_FS_PORT);
+
+       /* marshall the parameters */
+       bp = call->request;
+       *bp++ = htonl(FSSETLOCK);
+       *bp++ = htonl(vnode->fid.vid);
+       *bp++ = htonl(vnode->fid.vnode);
+       *bp++ = htonl(vnode->fid.unique);
+       *bp++ = htonl(type);
+
+       return afs_make_call(&server->addr, call, GFP_NOFS, wait_mode);
+}
+
+/*
+ * extend a lock on a file
+ */
+int afs_fs_extend_lock(struct afs_server *server,
+                      struct key *key,
+                      struct afs_vnode *vnode,
+                      const struct afs_wait_mode *wait_mode)
+{
+       struct afs_call *call;
+       __be32 *bp;
+
+       _enter("");
+
+       call = afs_alloc_flat_call(&afs_RXFSExtendLock, 4 * 4, 6 * 4);
+       if (!call)
+               return -ENOMEM;
+
+       call->key = key;
+       call->reply = vnode;
+       call->service_id = FS_SERVICE;
+       call->port = htons(AFS_FS_PORT);
+
+       /* marshall the parameters */
+       bp = call->request;
+       *bp++ = htonl(FSEXTENDLOCK);
+       *bp++ = htonl(vnode->fid.vid);
+       *bp++ = htonl(vnode->fid.vnode);
+       *bp++ = htonl(vnode->fid.unique);
+
+       return afs_make_call(&server->addr, call, GFP_NOFS, wait_mode);
+}
+
+/*
+ * release a lock on a file
+ */
+int afs_fs_release_lock(struct afs_server *server,
+                       struct key *key,
+                       struct afs_vnode *vnode,
+                       const struct afs_wait_mode *wait_mode)
+{
+       struct afs_call *call;
+       __be32 *bp;
+
+       _enter("");
+
+       call = afs_alloc_flat_call(&afs_RXFSReleaseLock, 4 * 4, 6 * 4);
+       if (!call)
+               return -ENOMEM;
+
+       call->key = key;
+       call->reply = vnode;
+       call->service_id = FS_SERVICE;
+       call->port = htons(AFS_FS_PORT);
+
+       /* marshall the parameters */
+       bp = call->request;
+       *bp++ = htonl(FSRELEASELOCK);
+       *bp++ = htonl(vnode->fid.vid);
+       *bp++ = htonl(vnode->fid.vnode);
+       *bp++ = htonl(vnode->fid.unique);
+
+       return afs_make_call(&server->addr, call, GFP_NOFS, wait_mode);
+}
index 2c55dd94a1de573c9e4169a3cbb3c993dc6422b9..6306438f331f7b14451366dbbb04182d9dae5cc5 100644 (file)
@@ -351,10 +351,18 @@ struct afs_vnode {
 #define AFS_VNODE_ZAP_DATA     3               /* set if vnode's data should be invalidated */
 #define AFS_VNODE_DELETED      4               /* set if vnode deleted on server */
 #define AFS_VNODE_MOUNTPOINT   5               /* set if vnode is a mountpoint symlink */
+#define AFS_VNODE_LOCKING      6               /* set if waiting for lock on vnode */
+#define AFS_VNODE_READLOCKED   7               /* set if vnode is read-locked on the server */
+#define AFS_VNODE_WRITELOCKED  8               /* set if vnode is write-locked on the server */
+#define AFS_VNODE_UNLOCKING    9               /* set if vnode is being unlocked on the server */
 
        long                    acl_order;      /* ACL check count (callback break count) */
 
        struct list_head        writebacks;     /* alterations in pagecache that need writing */
+       struct list_head        pending_locks;  /* locks waiting to be granted */
+       struct list_head        granted_locks;  /* locks granted on this file */
+       struct delayed_work     lock_work;      /* work to be done in locking */
+       struct key              *unlock_key;    /* key to be used in unlocking */
 
        /* outstanding callback notification on this file */
        struct rb_node          server_rb;      /* link in server->fs_vnodes */
@@ -473,6 +481,15 @@ extern const struct file_operations afs_file_operations;
 extern int afs_open(struct inode *, struct file *);
 extern int afs_release(struct inode *, struct file *);
 
+/*
+ * flock.c
+ */
+extern void __exit afs_kill_lock_manager(void);
+extern void afs_lock_work(struct work_struct *);
+extern void afs_lock_may_be_available(struct afs_vnode *);
+extern int afs_lock(struct file *, int, struct file_lock *);
+extern int afs_flock(struct file *, int, struct file_lock *);
+
 /*
  * fsclient.c
  */
@@ -513,6 +530,15 @@ extern int afs_fs_get_volume_status(struct afs_server *, struct key *,
                                    struct afs_vnode *,
                                    struct afs_volume_status *,
                                    const struct afs_wait_mode *);
+extern int afs_fs_set_lock(struct afs_server *, struct key *,
+                          struct afs_vnode *, afs_lock_type_t,
+                          const struct afs_wait_mode *);
+extern int afs_fs_extend_lock(struct afs_server *, struct key *,
+                             struct afs_vnode *,
+                             const struct afs_wait_mode *);
+extern int afs_fs_release_lock(struct afs_server *, struct key *,
+                              struct afs_vnode *,
+                              const struct afs_wait_mode *);
 
 /*
  * inode.c
@@ -681,6 +707,10 @@ extern int afs_vnode_store_data(struct afs_writeback *, pgoff_t, pgoff_t,
 extern int afs_vnode_setattr(struct afs_vnode *, struct key *, struct iattr *);
 extern int afs_vnode_get_volume_status(struct afs_vnode *, struct key *,
                                       struct afs_volume_status *);
+extern int afs_vnode_set_lock(struct afs_vnode *, struct key *,
+                             afs_lock_type_t);
+extern int afs_vnode_extend_lock(struct afs_vnode *, struct key *);
+extern int afs_vnode_release_lock(struct afs_vnode *, struct key *);
 
 /*
  * volume.c
index cd21195bbb24bd6dfe260dfa298d578e895d17d4..0f60f6b35769acfda6cd49522a770db006a08758 100644 (file)
@@ -168,6 +168,7 @@ static void __exit afs_exit(void)
        printk(KERN_INFO "kAFS: Red Hat AFS client v0.1 unregistering.\n");
 
        afs_fs_exit();
+       afs_kill_lock_manager();
        afs_close_socket();
        afs_purge_servers();
        afs_callback_update_kill();
index d1a889c4074292b8dbf6d4f251497e681067cd1a..2d33a5f7d218ad75cf26ec973f10e5958fc5a57d 100644 (file)
@@ -35,6 +35,7 @@ int afs_abort_to_error(u32 abort_code)
        case VOVERQUOTA:        return -EDQUOT;
        case VBUSY:             return -EBUSY;
        case VMOVED:            return -ENXIO;
+       case 0x2f6df0a:         return -EWOULDBLOCK;
        case 0x2f6df0c:         return -EACCES;
        case 0x2f6df0f:         return -EBUSY;
        case 0x2f6df10:         return -EEXIST;
index 2e8496ba1205386ed2e5bc959947460688714496..993cdf1cce3a494b40a7df32166813a17756bd4d 100644 (file)
@@ -460,6 +460,9 @@ static void afs_i_init_once(void *_vnode, struct kmem_cache *cachep,
        spin_lock_init(&vnode->writeback_lock);
        spin_lock_init(&vnode->lock);
        INIT_LIST_HEAD(&vnode->writebacks);
+       INIT_LIST_HEAD(&vnode->pending_locks);
+       INIT_LIST_HEAD(&vnode->granted_locks);
+       INIT_DELAYED_WORK(&vnode->lock_work, afs_lock_work);
        INIT_WORK(&vnode->cb_broken_work, afs_broken_callback_work);
 }
 
index 232c55dc245df2dd25a273b24644961394f0fe7b..ac811e445b2ddf8269f0e851abf851ae6536bab3 100644 (file)
@@ -887,11 +887,6 @@ int afs_vnode_get_volume_status(struct afs_vnode *vnode, struct key *key,
               vnode->fid.unique,
               key_serial(key));
 
-       /* this op will fetch the status */
-       spin_lock(&vnode->lock);
-       vnode->update_cnt++;
-       spin_unlock(&vnode->lock);
-
        do {
                /* pick a server to query */
                server = afs_volume_pick_fileserver(vnode);
@@ -905,20 +900,127 @@ int afs_vnode_get_volume_status(struct afs_vnode *vnode, struct key *key,
        } while (!afs_volume_release_fileserver(vnode, server, ret));
 
        /* adjust the flags */
-       if (ret == 0) {
-               afs_vnode_finalise_status_update(vnode, server);
+       if (ret == 0)
+               afs_put_server(server);
+
+       _leave(" = %d", ret);
+       return ret;
+
+no_server:
+       return PTR_ERR(server);
+}
+
+/*
+ * get a lock on a file
+ */
+int afs_vnode_set_lock(struct afs_vnode *vnode, struct key *key,
+                      afs_lock_type_t type)
+{
+       struct afs_server *server;
+       int ret;
+
+       _enter("%s{%x:%u.%u},%x,%u",
+              vnode->volume->vlocation->vldb.name,
+              vnode->fid.vid,
+              vnode->fid.vnode,
+              vnode->fid.unique,
+              key_serial(key), type);
+
+       do {
+               /* pick a server to query */
+               server = afs_volume_pick_fileserver(vnode);
+               if (IS_ERR(server))
+                       goto no_server;
+
+               _debug("USING SERVER: %08x\n", ntohl(server->addr.s_addr));
+
+               ret = afs_fs_set_lock(server, key, vnode, type, &afs_sync_call);
+
+       } while (!afs_volume_release_fileserver(vnode, server, ret));
+
+       /* adjust the flags */
+       if (ret == 0)
+               afs_put_server(server);
+
+       _leave(" = %d", ret);
+       return ret;
+
+no_server:
+       return PTR_ERR(server);
+}
+
+/*
+ * extend a lock on a file
+ */
+int afs_vnode_extend_lock(struct afs_vnode *vnode, struct key *key)
+{
+       struct afs_server *server;
+       int ret;
+
+       _enter("%s{%x:%u.%u},%x",
+              vnode->volume->vlocation->vldb.name,
+              vnode->fid.vid,
+              vnode->fid.vnode,
+              vnode->fid.unique,
+              key_serial(key));
+
+       do {
+               /* pick a server to query */
+               server = afs_volume_pick_fileserver(vnode);
+               if (IS_ERR(server))
+                       goto no_server;
+
+               _debug("USING SERVER: %08x\n", ntohl(server->addr.s_addr));
+
+               ret = afs_fs_extend_lock(server, key, vnode, &afs_sync_call);
+
+       } while (!afs_volume_release_fileserver(vnode, server, ret));
+
+       /* adjust the flags */
+       if (ret == 0)
+               afs_put_server(server);
+
+       _leave(" = %d", ret);
+       return ret;
+
+no_server:
+       return PTR_ERR(server);
+}
+
+/*
+ * release a lock on a file
+ */
+int afs_vnode_release_lock(struct afs_vnode *vnode, struct key *key)
+{
+       struct afs_server *server;
+       int ret;
+
+       _enter("%s{%x:%u.%u},%x",
+              vnode->volume->vlocation->vldb.name,
+              vnode->fid.vid,
+              vnode->fid.vnode,
+              vnode->fid.unique,
+              key_serial(key));
+
+       do {
+               /* pick a server to query */
+               server = afs_volume_pick_fileserver(vnode);
+               if (IS_ERR(server))
+                       goto no_server;
+
+               _debug("USING SERVER: %08x\n", ntohl(server->addr.s_addr));
+
+               ret = afs_fs_release_lock(server, key, vnode, &afs_sync_call);
+
+       } while (!afs_volume_release_fileserver(vnode, server, ret));
+
+       /* adjust the flags */
+       if (ret == 0)
                afs_put_server(server);
-       } else {
-               afs_vnode_status_update_failed(vnode, ret);
-       }
 
        _leave(" = %d", ret);
        return ret;
 
 no_server:
-       spin_lock(&vnode->lock);
-       vnode->update_cnt--;
-       ASSERTCMP(vnode->update_cnt, >=, 0);
-       spin_unlock(&vnode->lock);
        return PTR_ERR(server);
 }
index 51c938a71decca6c9f6386ffbdf4256cbd0369c0..aa4530c1ff7a82fce9f0115a3e92f2a78c66a69c 100644 (file)
@@ -820,6 +820,10 @@ struct file_lock {
        union {
                struct nfs_lock_info    nfs_fl;
                struct nfs4_lock_info   nfs4_fl;
+               struct {
+                       struct list_head link;  /* link in AFS vnode's pending_locks list */
+                       int state;              /* state of grant or error if -ve */
+               } afs;
        } fl_u;
 };