diff options
author | Greg Farnum <gregf@hq.newdream.net> | 2010-08-02 15:34:23 -0700 |
---|---|---|
committer | Sage Weil <sage@newdream.net> | 2010-08-02 16:10:53 -0700 |
commit | 40819f6fb227c1832935b775ac22aef10aa6f6dd (patch) | |
tree | 14f3e73f9585cc157ab2daab487c7b3964394fe2 | |
parent | fbaad9797a761c2d5ff6e755bbb4c046207a1ca2 (diff) | |
download | kernel_samsung_espresso10-40819f6fb227c1832935b775ac22aef10aa6f6dd.zip kernel_samsung_espresso10-40819f6fb227c1832935b775ac22aef10aa6f6dd.tar.gz kernel_samsung_espresso10-40819f6fb227c1832935b775ac22aef10aa6f6dd.tar.bz2 |
ceph: add flock/fcntl lock support
Implement flock inode operation to support advisory file locking. All
lock/unlock operations are synchronous with the MDS. Lock state is
sent when reconnecting to a recovering MDS to restore the shared lock
state.
Signed-off-by: Greg Farnum <gregf@hq.newdream.net>
Signed-off-by: Sage Weil <sage@newdream.net>
-rw-r--r-- | fs/ceph/Makefile | 2 | ||||
-rw-r--r-- | fs/ceph/file.c | 2 | ||||
-rw-r--r-- | fs/ceph/locks.c | 256 | ||||
-rw-r--r-- | fs/ceph/mds_client.c | 18 | ||||
-rw-r--r-- | fs/ceph/super.h | 8 |
5 files changed, 284 insertions, 2 deletions
diff --git a/fs/ceph/Makefile b/fs/ceph/Makefile index 6a660e6..278e117 100644 --- a/fs/ceph/Makefile +++ b/fs/ceph/Makefile @@ -6,7 +6,7 @@ ifneq ($(KERNELRELEASE),) obj-$(CONFIG_CEPH_FS) += ceph.o -ceph-objs := super.o inode.o dir.o file.o addr.o ioctl.o \ +ceph-objs := super.o inode.o dir.o file.o locks.o addr.o ioctl.o \ export.o caps.o snap.o xattr.o \ messenger.o msgpool.o buffer.o pagelist.o \ mds_client.o mdsmap.o \ diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 5fe50ff..e850e63 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -938,6 +938,8 @@ const struct file_operations ceph_file_fops = { .aio_write = ceph_aio_write, .mmap = ceph_mmap, .fsync = ceph_fsync, + .lock = ceph_lock, + .flock = ceph_flock, .splice_read = generic_file_splice_read, .splice_write = generic_file_splice_write, .unlocked_ioctl = ceph_ioctl, diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c new file mode 100644 index 0000000..ae85af0 --- /dev/null +++ b/fs/ceph/locks.c @@ -0,0 +1,256 @@ +#include "ceph_debug.h" + +#include <linux/file.h> +#include <linux/namei.h> + +#include "super.h" +#include "mds_client.h" +#include "pagelist.h" + +/** + * Implement fcntl and flock locking functions. + */ +static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, + u64 pid, u64 pid_ns, + int cmd, u64 start, u64 length, u8 wait) +{ + struct inode *inode = file->f_dentry->d_inode; + struct ceph_mds_client *mdsc = + &ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_request *req; + int err; + + req = ceph_mdsc_create_request(mdsc, operation, USE_AUTH_MDS); + if (IS_ERR(req)) + return PTR_ERR(req); + req->r_inode = igrab(inode); + + dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, " + "length: %llu, wait: %d, type`: %d", (int)lock_type, + (int)operation, pid, start, length, wait, cmd); + + req->r_args.filelock_change.rule = lock_type; + req->r_args.filelock_change.type = cmd; + req->r_args.filelock_change.pid = cpu_to_le64(pid); + /* This should be adjusted, but I'm not sure if + namespaces actually get id numbers*/ + req->r_args.filelock_change.pid_namespace = + cpu_to_le64((u64)pid_ns); + req->r_args.filelock_change.start = cpu_to_le64(start); + req->r_args.filelock_change.length = cpu_to_le64(length); + req->r_args.filelock_change.wait = wait; + + err = ceph_mdsc_do_request(mdsc, inode, req); + ceph_mdsc_put_request(req); + dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, " + "length: %llu, wait: %d, type`: %d err code %d", (int)lock_type, + (int)operation, pid, start, length, wait, cmd, err); + return err; +} + +/** + * Attempt to set an fcntl lock. + * For now, this just goes away to the server. Later it may be more awesome. + */ +int ceph_lock(struct file *file, int cmd, struct file_lock *fl) +{ + u64 length; + u8 lock_cmd; + int err; + u8 wait = 0; + u16 op = CEPH_MDS_OP_SETFILELOCK; + + fl->fl_nspid = get_pid(task_tgid(current)); + dout("ceph_lock, fl_pid:%d", fl->fl_pid); + + /* set wait bit as appropriate, then make command as Ceph expects it*/ + if (F_SETLKW == cmd) + wait = 1; + if (F_GETLK == cmd) + op = CEPH_MDS_OP_GETFILELOCK; + + if (F_RDLCK == fl->fl_type) + lock_cmd = CEPH_LOCK_SHARED; + else if (F_WRLCK == fl->fl_type) + lock_cmd = CEPH_LOCK_EXCL; + else + lock_cmd = CEPH_LOCK_UNLOCK; + + if (LLONG_MAX == fl->fl_end) + length = 0; + else + length = fl->fl_end - fl->fl_start + 1; + + err = ceph_lock_message(CEPH_LOCK_FCNTL, op, file, + (u64)fl->fl_pid, (u64)fl->fl_nspid, + lock_cmd, fl->fl_start, + length, wait); + if (!err) { + dout("mds locked, locking locally"); + err = posix_lock_file(file, fl, NULL); + if (err && (CEPH_MDS_OP_SETFILELOCK == op)) { + /* undo! This should only happen if the kernel detects + * local deadlock. */ + ceph_lock_message(CEPH_LOCK_FCNTL, op, file, + (u64)fl->fl_pid, (u64)fl->fl_nspid, + CEPH_LOCK_UNLOCK, fl->fl_start, + length, 0); + dout("got %d on posix_lock_file, undid lock", err); + } + } else { + dout("mds returned error code %d", err); + } + return err; +} + +int ceph_flock(struct file *file, int cmd, struct file_lock *fl) +{ + u64 length; + u8 lock_cmd; + int err; + u8 wait = 1; + + fl->fl_nspid = get_pid(task_tgid(current)); + dout("ceph_flock, fl_pid:%d", fl->fl_pid); + + /* set wait bit, then clear it out of cmd*/ + if (cmd & LOCK_NB) + wait = 0; + cmd = cmd & (LOCK_SH | LOCK_EX | LOCK_UN); + /* set command sequence that Ceph wants to see: + shared lock, exclusive lock, or unlock */ + if (LOCK_SH == cmd) + lock_cmd = CEPH_LOCK_SHARED; + else if (LOCK_EX == cmd) + lock_cmd = CEPH_LOCK_EXCL; + else + lock_cmd = CEPH_LOCK_UNLOCK; + /* mds requires start and length rather than start and end */ + if (LLONG_MAX == fl->fl_end) + length = 0; + else + length = fl->fl_end - fl->fl_start + 1; + + err = ceph_lock_message(CEPH_LOCK_FLOCK, CEPH_MDS_OP_SETFILELOCK, + file, (u64)fl->fl_pid, (u64)fl->fl_nspid, + lock_cmd, fl->fl_start, + length, wait); + if (!err) { + err = flock_lock_file_wait(file, fl); + if (err) { + ceph_lock_message(CEPH_LOCK_FLOCK, + CEPH_MDS_OP_SETFILELOCK, + file, (u64)fl->fl_pid, + (u64)fl->fl_nspid, + CEPH_LOCK_UNLOCK, fl->fl_start, + length, 0); + dout("got %d on flock_lock_file_wait, undid lock", err); + } + } else { + dout("mds error code %d", err); + } + return err; +} + +/** + * Must be called with BKL already held. Fills in the passed + * counter variables, so you can prepare pagelist metadata before calling + * ceph_encode_locks. + */ +void ceph_count_locks(struct inode *inode, int *fcntl_count, int *flock_count) +{ + struct file_lock *lock; + + *fcntl_count = 0; + *flock_count = 0; + + for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { + if (lock->fl_flags & FL_POSIX) + ++(*fcntl_count); + else if (lock->fl_flags & FL_FLOCK) + ++(*flock_count); + } + dout("counted %d flock locks and %d fcntl locks", + *flock_count, *fcntl_count); +} + +/** + * Encode the flock and fcntl locks for the given inode into the pagelist. + * Format is: #fcntl locks, sequential fcntl locks, #flock locks, + * sequential flock locks. + * Must be called with BLK already held, and the lock numbers should have + * been gathered under the same lock holding window. + */ +int ceph_encode_locks(struct inode *inode, struct ceph_pagelist *pagelist, + int num_fcntl_locks, int num_flock_locks) +{ + struct file_lock *lock; + struct ceph_filelock cephlock; + int err = 0; + + dout("encoding %d flock and %d fcntl locks", num_flock_locks, + num_fcntl_locks); + err = ceph_pagelist_append(pagelist, &num_fcntl_locks, sizeof(u32)); + if (err) + goto fail; + for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { + if (lock->fl_flags & FL_POSIX) { + err = lock_to_ceph_filelock(lock, &cephlock); + if (err) + goto fail; + err = ceph_pagelist_append(pagelist, &cephlock, + sizeof(struct ceph_filelock)); + } + if (err) + goto fail; + } + + err = ceph_pagelist_append(pagelist, &num_flock_locks, sizeof(u32)); + if (err) + goto fail; + for (lock = inode->i_flock; lock != NULL; lock = lock->fl_next) { + if (lock->fl_flags & FL_FLOCK) { + err = lock_to_ceph_filelock(lock, &cephlock); + if (err) + goto fail; + err = ceph_pagelist_append(pagelist, &cephlock, + sizeof(struct ceph_filelock)); + } + if (err) + goto fail; + } +fail: + return err; +} + +/* + * Given a pointer to a lock, convert it to a ceph filelock + */ +int lock_to_ceph_filelock(struct file_lock *lock, + struct ceph_filelock *cephlock) +{ + int err = 0; + + cephlock->start = cpu_to_le64(lock->fl_start); + cephlock->length = cpu_to_le64(lock->fl_end - lock->fl_start + 1); + cephlock->client = cpu_to_le64(0); + cephlock->pid = cpu_to_le64(lock->fl_pid); + cephlock->pid_namespace = cpu_to_le64((u64)lock->fl_nspid); + + switch (lock->fl_type) { + case F_RDLCK: + cephlock->type = CEPH_LOCK_SHARED; + break; + case F_WRLCK: + cephlock->type = CEPH_LOCK_EXCL; + break; + case F_UNLCK: + cephlock->type = CEPH_LOCK_UNLOCK; + break; + default: + dout("Have unknown lock type %d", lock->fl_type); + err = -EINVAL; + } + + return err; +} diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 615f720..9f0833e 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -3,6 +3,7 @@ #include <linux/wait.h> #include <linux/slab.h> #include <linux/sched.h> +#include <linux/smp_lock.h> #include "mds_client.h" #include "mon_client.h" @@ -2335,7 +2336,22 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap, } spin_unlock(&inode->i_lock); - err = ceph_pagelist_append(pagelist, &rec, reclen); + if (recon_state->flock) { + int num_fcntl_locks, num_flock_locks; + + lock_kernel(); + ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks); + rec.v2.flock_len = (2*sizeof(u32) + + (num_fcntl_locks+num_flock_locks) * + sizeof(struct ceph_filelock)); + + err = ceph_pagelist_append(pagelist, &rec, reclen); + if (!err) + err = ceph_encode_locks(inode, pagelist, + num_fcntl_locks, + num_flock_locks); + unlock_kernel(); + } out: kfree(path); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index 83f7cc5..95adce4 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -892,6 +892,14 @@ extern void ceph_debugfs_cleanup(void); extern int ceph_debugfs_client_init(struct ceph_client *client); extern void ceph_debugfs_client_cleanup(struct ceph_client *client); +/* locks.c */ +extern int ceph_lock(struct file *file, int cmd, struct file_lock *fl); +extern int ceph_flock(struct file *file, int cmd, struct file_lock *fl); +extern void ceph_count_locks(struct inode *inode, int *p_num, int *f_num); +extern int ceph_encode_locks(struct inode *i, struct ceph_pagelist *p, + int p_locks, int f_locks); +extern int lock_to_ceph_filelock(struct file_lock *fl, struct ceph_filelock *c); + static inline struct inode *get_dentry_parent_inode(struct dentry *dentry) { if (dentry && dentry->d_parent) |