/*
* Create a block and initialize it.
*
* Note: we explicitly set the cookie of the grant reply to that of
* the blocked lock request. The spec explicitly mentions that the client
* should _not_ rely on the callback containing the same cookie as the
* request, but (as I found out later) that's because some implementations
* do just this. Never mind the standards comittees, they support our
* logging industries.
*/
static inline struct nlm_block *
nlmsvc_create_block(struct svc_rqst *rqstp, struct nlm_file *file,
struct nlm_lock *lock, struct nlm_cookie *cookie)
{
struct nlm_block *block;
struct nlm_host *host;
struct nlm_rqst *call;
/* Create host handle for callback */
host = nlmclnt_lookup_host(&rqstp->rq_addr,
rqstp->rq_prot, rqstp->rq_vers);
if (host == NULL)
return NULL;
/* Allocate memory for block, and initialize arguments */
if (!(block = (struct nlm_block *) kmalloc(sizeof(*block), GFP_KERNEL)))
goto failed;
memset(block, 0, sizeof(*block));
locks_init_lock(&block->b_call.a_args.lock.fl);
locks_init_lock(&block->b_call.a_res.lock.fl);
block->b_host = nlmsvc_lookup_host(rqstp);
if (block->b_host == NULL) {
goto failed_free;
}
if (!nlmclnt_setgrantargs(&block->b_call, lock))
goto failed_free;
/* Set notifier function for VFS, and init args */
block->b_call.a_args.lock.fl.fl_notify = nlmsvc_notify_blocked;
block->b_call.a_args.cookie = *cookie; /* see above */
dprintk("lockd: created block %p...\n", block);
/* Create and initialize the block */
block->b_daemon = rqstp->rq_server;
block->b_file = file;
/* Add to file's list of blocks */
block->b_fnext = file->f_blocks;
file->f_blocks = block;
/* Set up RPC arguments for callback */
call = &block->b_call;
call->a_host = host;
call->a_flags = RPC_TASK_ASYNC;
return block;
failed_free:
kfree(block);
failed:
nlm_release_host(host);
return NULL;
}
static __be32 *
nlm_decode_lock(__be32 *p, struct nlm_lock *lock)
{
struct file_lock *fl = &lock->fl;
s32 start, len, end;
if (!(p = xdr_decode_string_inplace(p, &lock->caller,
&lock->len,
NLM_MAXSTRLEN))
|| !(p = nlm_decode_fh(p, &lock->fh))
|| !(p = nlm_decode_oh(p, &lock->oh)))
return NULL;
lock->svid = ntohl(*p++);
locks_init_lock(fl);
fl->fl_owner = current->files;
fl->fl_pid = (pid_t)lock->svid;
fl->fl_flags = FL_POSIX;
fl->fl_type = F_RDLCK; /* as good as anything else */
start = ntohl(*p++);
len = ntohl(*p++);
end = start + len - 1;
fl->fl_start = s32_to_loff_t(start);
if (len == 0 || end < 0)
fl->fl_end = OFFSET_MAX;
else
fl->fl_end = s32_to_loff_t(end);
return p;
}
static int
nlm4clt_decode_testres(struct rpc_rqst *req, __be32 *p, struct nlm_res *resp)
{
if (!(p = nlm4_decode_cookie(p, &resp->cookie)))
return -EIO;
resp->status = *p++;
if (resp->status == nlm_lck_denied) {
struct file_lock *fl = &resp->lock.fl;
u32 excl;
__u64 start, len;
__s64 end;
memset(&resp->lock, 0, sizeof(resp->lock));
locks_init_lock(fl);
excl = ntohl(*p++);
resp->lock.svid = ntohl(*p++);
fl->fl_pid = (pid_t)resp->lock.svid;
if (!(p = nlm4_decode_oh(p, &resp->lock.oh)))
return -EIO;
fl->fl_flags = FL_POSIX;
fl->fl_type = excl? F_WRLCK : F_RDLCK;
p = xdr_decode_hyper(p, &start);
p = xdr_decode_hyper(p, &len);
end = start + len - 1;
fl->fl_start = s64_to_loff_t(start);
if (len == 0 || end < 0)
fl->fl_end = OFFSET_MAX;
else
fl->fl_end = s64_to_loff_t(end);
}
return 0;
}
/*
* Initialises the fields of the file lock which are invariant for
* free file_locks.
*/
static void init_once(void *foo, kmem_cache_t *cache, unsigned long flags)
{
struct file_lock *lock = (struct file_lock *) foo;
if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) !=
SLAB_CTOR_CONSTRUCTOR)
return;
locks_init_lock(lock);
}
/*
* Allocate an NLM RPC call struct
*/
struct nlm_rqst *
nlmclnt_alloc_call(void)
{
struct nlm_rqst *call;
while (!signalled()) {
call = (struct nlm_rqst *) kmalloc(sizeof(struct nlm_rqst), GFP_KERNEL);
if (call) {
memset(call, 0, sizeof(*call));
locks_init_lock(&call->a_args.lock.fl);
locks_init_lock(&call->a_res.lock.fl);
return call;
}
printk("nlmclnt_alloc_call: failed, waiting for memory\n");
current->state = TASK_INTERRUPTIBLE;
schedule_timeout(5*HZ);
}
return NULL;
}
/*
* RECLAIM: Try to reclaim a lock
*/
int
nlmclnt_reclaim(struct nlm_host *host, struct file_lock *fl)
{
struct nlm_rqst reqst, *req;
int status;
req = &reqst;
memset(req, 0, sizeof(*req));
locks_init_lock(&req->a_args.lock.fl);
locks_init_lock(&req->a_res.lock.fl);
req->a_host = host;
req->a_flags = 0;
/* Set up the argument struct */
nlmclnt_setlockargs(req, fl);
req->a_args.reclaim = 1;
if ((status = nlmclnt_call(req, NLMPROC_LOCK)) >= 0
&& req->a_res.status == NLM_LCK_GRANTED)
return 0;
printk(KERN_WARNING "lockd: failed to reclaim lock for pid %d "
"(errno %d, status %d)\n", fl->fl_pid,
status, req->a_res.status);
/*
* FIXME: This is a serious failure. We can
*
* a. Ignore the problem
* b. Send the owning process some signal (Linux doesn't have
* SIGLOST, though...)
* c. Retry the operation
*
* Until someone comes up with a simple implementation
* for b or c, I'll choose option a.
*/
return -ENOLCK;
}
static int decode_nlm4_holder(struct xdr_stream *xdr, struct nlm_res *result)
{
struct nlm_lock *lock = &result->lock;
struct file_lock *fl = &lock->fl;
u64 l_offset, l_len;
u32 exclusive;
int error;
__be32 *p;
s32 end;
memset(lock, 0, sizeof(*lock));
locks_init_lock(fl);
p = xdr_inline_decode(xdr, 4 + 4);
if (unlikely(p == NULL))
goto out_overflow;
exclusive = be32_to_cpup(p++);
lock->svid = be32_to_cpup(p);
fl->fl_pid = (pid_t)lock->svid;
error = decode_netobj(xdr, &lock->oh);
if (unlikely(error))
goto out;
p = xdr_inline_decode(xdr, 8 + 8);
if (unlikely(p == NULL))
goto out_overflow;
fl->fl_flags = FL_POSIX;
fl->fl_type = exclusive != 0 ? F_WRLCK : F_RDLCK;
p = xdr_decode_hyper(p, &l_offset);
xdr_decode_hyper(p, &l_len);
end = l_offset + l_len - 1;
fl->fl_start = (loff_t)l_offset;
if (l_len == 0 || end < 0)
fl->fl_end = OFFSET_MAX;
else
fl->fl_end = (loff_t)end;
error = 0;
out:
return error;
out_overflow:
print_overflow_msg(__func__, xdr);
return -EIO;
}
int
nlm4svc_decode_shareargs(struct svc_rqst *rqstp, u32 *p, nlm_args *argp)
{
struct nlm_lock *lock = &argp->lock;
memset(lock, 0, sizeof(*lock));
locks_init_lock(&lock->fl);
lock->fl.fl_pid = ~(u32) 0;
if (!(p = nlm4_decode_cookie(p, &argp->cookie))
|| !(p = xdr_decode_string_inplace(p, &lock->caller,
&lock->len, NLM_MAXSTRLEN))
|| !(p = nlm4_decode_fh(p, &lock->fh))
|| !(p = nlm4_decode_oh(p, &lock->oh)))
return 0;
argp->fsm_mode = ntohl(*p++);
argp->fsm_access = ntohl(*p++);
return xdr_argsize_check(rqstp, p);
}
static int posix_lock_file(struct file *filp, struct file_lock *caller,
unsigned int wait)
{
struct file_lock *fl;
struct file_lock *new_fl, *new_fl2;
struct file_lock *left = NULL;
struct file_lock *right = NULL;
struct file_lock **before;
int error;
int added = 0;
/*
* We may need two file_lock structures for this operation,
* so we get them in advance to avoid races.
*/
new_fl = locks_empty_lock();
new_fl2 = locks_empty_lock();
error = -ENOLCK; /* "no luck" */
if (!(new_fl && new_fl2))
goto out;
if (caller->fl_type != F_UNLCK) {
repeat:
error = -EBUSY;
if ((fl = filp->f_inode->i_flock) && (fl->fl_flags & FL_FLOCK))
goto out;
while (fl != NULL) {
if (!posix_locks_conflict(caller, fl)) {
fl = fl->fl_next;
continue;
}
error = -EAGAIN;
if (!wait)
goto out;
error = -EDEADLK;
if (posix_locks_deadlock(caller->fl_owner, fl->fl_owner))
goto out;
error = -ERESTARTSYS;
if (current->signal & ~current->blocked)
goto out;
locks_insert_block(fl, caller);
interruptible_sleep_on(&caller->fl_wait);
locks_delete_block(fl, caller);
goto repeat;
}
}
/*
* We've allocated the new locks in advance, so there are no
* errors possible (and no blocking operations) from here on.
*
* Find the first old lock with the same owner as the new lock.
*/
before = &filp->f_inode->i_flock;
error = -EBUSY;
if ((*before != NULL) && ((*before)->fl_flags & FL_FLOCK))
goto out;
/* First skip locks owned by other processes.
*/
while ((fl = *before) && (caller->fl_owner != fl->fl_owner)) {
before = &fl->fl_next;
}
/* Process locks with this owner.
*/
while ((fl = *before) && (caller->fl_owner == fl->fl_owner)) {
/* Detect adjacent or overlapping regions (if same lock type)
*/
if (caller->fl_type == fl->fl_type) {
if (fl->fl_end < caller->fl_start - 1)
goto next_lock;
/* If the next lock in the list has entirely bigger
* addresses than the new one, insert the lock here.
*/
if (fl->fl_start > caller->fl_end + 1)
break;
/* If we come here, the new and old lock are of the
* same type and adjacent or overlapping. Make one
* lock yielding from the lower start address of both
* locks to the higher end address.
*/
if (fl->fl_start > caller->fl_start)
fl->fl_start = caller->fl_start;
else
caller->fl_start = fl->fl_start;
if (fl->fl_end < caller->fl_end)
fl->fl_end = caller->fl_end;
else
caller->fl_end = fl->fl_end;
if (added) {
locks_delete_lock(before, 0);
continue;
}
caller = fl;
added = 1;
}
else {
/* Processing for different lock types is a bit
* more complex.
*/
if (fl->fl_end < caller->fl_start)
goto next_lock;
if (fl->fl_start > caller->fl_end)
break;
if (caller->fl_type == F_UNLCK)
added = 1;
if (fl->fl_start < caller->fl_start)
left = fl;
/* If the next lock in the list has a higher end
* address than the new one, insert the new one here.
*/
if (fl->fl_end > caller->fl_end) {
right = fl;
break;
}
if (fl->fl_start >= caller->fl_start) {
/* The new lock completely replaces an old
* one (This may happen several times).
*/
if (added) {
locks_delete_lock(before, 0);
continue;
}
/* Replace the old lock with the new one.
* Wake up anybody waiting for the old one,
* as the change in lock type might satisfy
* their needs.
*/
locks_wake_up_blocks(fl, 0);
fl->fl_start = caller->fl_start;
fl->fl_end = caller->fl_end;
fl->fl_type = caller->fl_type;
caller = fl;
added = 1;
}
}
/* Go on to next lock.
*/
next_lock:
before = &fl->fl_next;
}
error = 0;
if (!added) {
if (caller->fl_type == F_UNLCK)
goto out;
locks_init_lock(new_fl, caller);
locks_insert_lock(before, new_fl);
new_fl = NULL;
}
if (right) {
if (left == right) {
/* The new lock breaks the old one in two pieces,
* so we have to use the second new lock (in this
* case, even F_UNLCK may fail!).
*/
left = locks_init_lock(new_fl2, right);
locks_insert_lock(before, left);
new_fl2 = NULL;
}
right->fl_start = caller->fl_end + 1;
locks_wake_up_blocks(right, 0);
}
if (left) {
left->fl_end = caller->fl_start - 1;
locks_wake_up_blocks(left, 0);
}
out:
/*
* Free any unused locks. (They haven't
* ever been used, so we use kfree().)
*/
if (new_fl)
kfree(new_fl);
if (new_fl2)
kfree(new_fl2);
return error;
}
/* Allocate a new lock, and initialize its fields from fl.
* The lock is not inserted into any lists until locks_insert_lock() or
* locks_insert_block() are called.
*/
static inline struct file_lock *locks_alloc_lock(struct file_lock *fl)
{
return locks_init_lock(locks_empty_lock(), fl);
}
int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file,
int cmd, struct file_lock *fl)
{
struct dlm_ls *ls;
struct plock_op *op;
struct plock_xop *xop;
int rv;
ls = dlm_find_lockspace_local(lockspace);
if (!ls)
return -EINVAL;
xop = kzalloc(sizeof(*xop), GFP_KERNEL);
if (!xop) {
rv = -ENOMEM;
goto out;
}
op = &xop->xop;
op->info.optype = DLM_PLOCK_OP_LOCK;
op->info.pid = fl->fl_pid;
op->info.ex = (fl->fl_type == F_WRLCK);
op->info.wait = IS_SETLKW(cmd);
op->info.fsid = ls->ls_global_id;
op->info.number = number;
op->info.start = fl->fl_start;
op->info.end = fl->fl_end;
if (fl->fl_lmops && fl->fl_lmops->fl_grant) {
/* fl_owner is lockd which doesn't distinguish
processes on the nfs client */
op->info.owner = (__u64) fl->fl_pid;
xop->callback = fl->fl_lmops->fl_grant;
locks_init_lock(&xop->flc);
locks_copy_lock(&xop->flc, fl);
xop->fl = fl;
xop->file = file;
} else {
op->info.owner = (__u64)(long) fl->fl_owner;
xop->callback = NULL;
}
send_op(op);
if (xop->callback == NULL)
wait_event(recv_wq, (op->done != 0));
else {
rv = FILE_LOCK_DEFERRED;
goto out;
}
spin_lock(&ops_lock);
if (!list_empty(&op->list)) {
log_error(ls, "dlm_posix_lock: op on list %llx",
(unsigned long long)number);
list_del(&op->list);
}
spin_unlock(&ops_lock);
rv = op->info.rv;
if (!rv) {
if (posix_lock_file_wait(file, fl) < 0)
log_error(ls, "dlm_posix_lock: vfs lock error %llx",
(unsigned long long)number);
}
kfree(xop);
out:
dlm_put_lockspace(ls);
return rv;
}
int dlm_posix_get(dlm_lockspace_t *lockspace, u64 number, struct file *file,
struct file_lock *fl)
{
struct dlm_ls *ls;
struct plock_op *op;
int rv;
ls = dlm_find_lockspace_local(lockspace);
if (!ls)
return -EINVAL;
op = kzalloc(sizeof(*op), GFP_KERNEL);
if (!op) {
rv = -ENOMEM;
goto out;
}
op->info.optype = DLM_PLOCK_OP_GET;
op->info.pid = fl->fl_pid;
op->info.ex = (fl->fl_type == F_WRLCK);
op->info.fsid = ls->ls_global_id;
op->info.number = number;
op->info.start = fl->fl_start;
op->info.end = fl->fl_end;
if (fl->fl_lmops && fl->fl_lmops->fl_grant)
op->info.owner = (__u64) fl->fl_pid;
else
op->info.owner = (__u64)(long) fl->fl_owner;
send_op(op);
wait_event(recv_wq, (op->done != 0));
spin_lock(&ops_lock);
if (!list_empty(&op->list)) {
log_error(ls, "dlm_posix_get: op on list %llx",
(unsigned long long)number);
list_del(&op->list);
}
spin_unlock(&ops_lock);
/* info.rv from userspace is 1 for conflict, 0 for no-conflict,
-ENOENT if there are no locks on the file */
rv = op->info.rv;
fl->fl_type = F_UNLCK;
if (rv == -ENOENT)
rv = 0;
else if (rv > 0) {
locks_init_lock(fl);
fl->fl_type = (op->info.ex) ? F_WRLCK : F_RDLCK;
fl->fl_flags = FL_POSIX;
fl->fl_pid = op->info.pid;
fl->fl_start = op->info.start;
fl->fl_end = op->info.end;
rv = 0;
}
kfree(op);
out:
dlm_put_lockspace(ls);
return rv;
}
/*
* This is the main entry point for the NLM client.
*/
int
nlmclnt_proc(struct inode *inode, int cmd, struct file_lock *fl)
{
struct nfs_server *nfssrv = NFS_SERVER(inode);
struct nlm_host *host;
struct nlm_rqst reqst, *call = &reqst;
sigset_t oldset;
unsigned long flags;
int status, proto, vers;
vers = (NFS_PROTO(inode)->version == 3) ? 4 : 1;
if (NFS_PROTO(inode)->version > 3) {
printk(KERN_NOTICE "NFSv4 file locking not implemented!\n");
return -ENOLCK;
}
/* Retrieve transport protocol from NFS client */
proto = NFS_CLIENT(inode)->cl_xprt->prot;
if (!(host = nlmclnt_lookup_host(NFS_ADDR(inode), proto, vers)))
return -ENOLCK;
/* Create RPC client handle if not there, and copy soft
* and intr flags from NFS client. */
if (host->h_rpcclnt == NULL) {
struct rpc_clnt *clnt;
/* Bind an rpc client to this host handle (does not
* perform a portmapper lookup) */
if (!(clnt = nlm_bind_host(host))) {
status = -ENOLCK;
goto done;
}
clnt->cl_softrtry = nfssrv->client->cl_softrtry;
clnt->cl_intr = nfssrv->client->cl_intr;
clnt->cl_chatty = nfssrv->client->cl_chatty;
}
/* Keep the old signal mask */
spin_lock_irqsave(¤t->sighand->siglock, flags);
oldset = current->blocked;
/* If we're cleaning up locks because the process is exiting,
* perform the RPC call asynchronously. */
if ((IS_SETLK(cmd) || IS_SETLKW(cmd))
&& fl->fl_type == F_UNLCK
&& (current->flags & PF_EXITING)) {
sigfillset(¤t->blocked); /* Mask all signals */
recalc_sigpending();
spin_unlock_irqrestore(¤t->sighand->siglock, flags);
call = nlmclnt_alloc_call();
if (!call) {
status = -ENOMEM;
goto out_restore;
}
call->a_flags = RPC_TASK_ASYNC;
} else {
spin_unlock_irqrestore(¤t->sighand->siglock, flags);
memset(call, 0, sizeof(*call));
locks_init_lock(&call->a_args.lock.fl);
locks_init_lock(&call->a_res.lock.fl);
}
call->a_host = host;
/* Set up the argument struct */
nlmclnt_setlockargs(call, fl);
if (IS_SETLK(cmd) || IS_SETLKW(cmd)) {
if (fl->fl_type != F_UNLCK) {
call->a_args.block = IS_SETLKW(cmd) ? 1 : 0;
status = nlmclnt_lock(call, fl);
} else
status = nlmclnt_unlock(call, fl);
} else if (IS_GETLK(cmd))
status = nlmclnt_test(call, fl);
else
status = -EINVAL;
if (status < 0 && (call->a_flags & RPC_TASK_ASYNC))
kfree(call);
out_restore:
spin_lock_irqsave(¤t->sighand->siglock, flags);
current->blocked = oldset;
recalc_sigpending();
spin_unlock_irqrestore(¤t->sighand->siglock, flags);
done:
dprintk("lockd: clnt proc returns %d\n", status);
nlm_release_host(host);
return status;
}
/*
* Initialises the fields of the file lock which are invariant for
* free file_locks.
*/
static void init_once(void *foo)
{
struct file_lock *lock = (struct file_lock *) foo;
locks_init_lock(lock);
}
