[linux-block.git] / fs / nfs / nfs4file.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  linux/fs/nfs/file.c
 *
 *  Copyright (C) 1992  Rick Sladkey
 */
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/falloc.h>
#include <linux/mount.h>
#include <linux/nfs_fs.h>
#include <linux/nfs_ssc.h>
#include <linux/splice.h>
#include "delegation.h"
#include "internal.h"
#include "iostat.h"
#include "fscache.h"
#include "pnfs.h"

#include "nfstrace.h"

#ifdef CONFIG_NFS_V4_2
#include "nfs42.h"
#endif

#define NFSDBG_FACILITY		NFSDBG_FILE

static int
nfs4_file_open(struct inode *inode, struct file *filp)
{
	struct nfs_open_context *ctx;
	struct dentry *dentry = file_dentry(filp);
	struct dentry *parent = NULL;
	struct inode *dir;
	unsigned openflags = filp->f_flags;
	struct iattr attr;
	int err;

	/*
	 * If no cached dentry exists or if it's negative, NFSv4 handled the
	 * opens in ->lookup() or ->create().
	 *
	 * We only get this far for a cached positive dentry.  We skipped
	 * revalidation, so handle it here by dropping the dentry and returning
	 * -EOPENSTALE.  The VFS will retry the lookup/create/open.
	 */

	dprintk("NFS: open file(%pd2)\n", dentry);

	err = nfs_check_flags(openflags);
	if (err)
		return err;

	/* We can't create new files here */
	openflags &= ~(O_CREAT|O_EXCL);

	parent = dget_parent(dentry);
	dir = d_inode(parent);

	ctx = alloc_nfs_open_context(file_dentry(filp),
				     flags_to_mode(openflags), filp);
	err = PTR_ERR(ctx);
	if (IS_ERR(ctx))
		goto out;

	attr.ia_valid = ATTR_OPEN;
	if (openflags & O_TRUNC) {
		attr.ia_valid |= ATTR_SIZE;
		attr.ia_size = 0;
		filemap_write_and_wait(inode->i_mapping);
	}

	inode = NFS_PROTO(dir)->open_context(dir, ctx, openflags, &attr, NULL);
	if (IS_ERR(inode)) {
		err = PTR_ERR(inode);
		switch (err) {
		default:
			goto out_put_ctx;
		case -ENOENT:
		case -ESTALE:
		case -EISDIR:
		case -ENOTDIR:
		case -ELOOP:
			goto out_drop;
		}
	}
	if (inode != d_inode(dentry))
		goto out_drop;

	nfs_file_set_open_context(filp, ctx);
	nfs_fscache_open_file(inode, filp);
	err = 0;
	filp->f_mode |= FMODE_CAN_ODIRECT;

out_put_ctx:
	put_nfs_open_context(ctx);
out:
	dput(parent);
	return err;

out_drop:
	d_drop(dentry);
	err = -EOPENSTALE;
	goto out_put_ctx;
}

/*
 * Flush all dirty pages, and check for write errors.
 */
static int
nfs4_file_flush(struct file *file, fl_owner_t id)
{
	struct inode	*inode = file_inode(file);
	errseq_t since;

	dprintk("NFS: flush(%pD2)\n", file);

	nfs_inc_stats(inode, NFSIOS_VFSFLUSH);
	if ((file->f_mode & FMODE_WRITE) == 0)
		return 0;

	/*
	 * If we're holding a write delegation, then check if we're required
	 * to flush the i/o on close. If not, then just start the i/o now.
	 */
	if (!nfs4_delegation_flush_on_close(inode))
		return filemap_fdatawrite(file->f_mapping);

	/* Flush writes to the server and return any errors */
	since = filemap_sample_wb_err(file->f_mapping);
	nfs_wb_all(inode);
	return filemap_check_wb_err(file->f_mapping, since);
}

#ifdef CONFIG_NFS_V4_2
static ssize_t __nfs4_copy_file_range(struct file *file_in, loff_t pos_in,
				      struct file *file_out, loff_t pos_out,
				      size_t count, unsigned int flags)
{
	struct nfs42_copy_notify_res *cn_resp = NULL;
	struct nl4_server *nss = NULL;
	nfs4_stateid *cnrs = NULL;
	ssize_t ret;
	bool sync = false;

	/* Only offload copy if superblock is the same */
	if (file_in->f_op != &nfs4_file_operations)
		return -EXDEV;
	if (!nfs_server_capable(file_inode(file_out), NFS_CAP_COPY) ||
	    !nfs_server_capable(file_inode(file_in), NFS_CAP_COPY))
		return -EOPNOTSUPP;
	if (file_inode(file_in) == file_inode(file_out))
		return -EOPNOTSUPP;
	/* if the copy size if smaller than 2 RPC payloads, make it
	 * synchronous
	 */
	if (count <= 2 * NFS_SERVER(file_inode(file_in))->rsize)
		sync = true;
retry:
	if (!nfs42_files_from_same_server(file_in, file_out)) {
		/*
		 * for inter copy, if copy size is too small
		 * then fallback to generic copy.
		 */
		if (sync)
			return -EOPNOTSUPP;
		cn_resp = kzalloc(sizeof(struct nfs42_copy_notify_res),
				  GFP_KERNEL);
		if (unlikely(cn_resp == NULL))
			return -ENOMEM;

		ret = nfs42_proc_copy_notify(file_in, file_out, cn_resp);
		if (ret) {
			ret = -EOPNOTSUPP;
			goto out;
		}
		nss = &cn_resp->cnr_src;
		cnrs = &cn_resp->cnr_stateid;
	}
	ret = nfs42_proc_copy(file_in, pos_in, file_out, pos_out, count,
				nss, cnrs, sync);
out:
	kfree(cn_resp);

	if (ret == -EAGAIN)
		goto retry;
	return ret;
}

static ssize_t nfs4_copy_file_range(struct file *file_in, loff_t pos_in,
				    struct file *file_out, loff_t pos_out,
				    size_t count, unsigned int flags)
{
	ssize_t ret;

	ret = __nfs4_copy_file_range(file_in, pos_in, file_out, pos_out, count,
				     flags);
	if (ret == -EOPNOTSUPP || ret == -EXDEV)
		ret = splice_copy_file_range(file_in, pos_in, file_out,
					     pos_out, count);
	return ret;
}

static loff_t nfs4_file_llseek(struct file *filep, loff_t offset, int whence)
{
	loff_t ret;

	switch (whence) {
	case SEEK_HOLE:
	case SEEK_DATA:
		ret = nfs42_proc_llseek(filep, offset, whence);
		if (ret != -EOPNOTSUPP)
			return ret;
		fallthrough;
	default:
		return nfs_file_llseek(filep, offset, whence);
	}
}

static long nfs42_fallocate(struct file *filep, int mode, loff_t offset, loff_t len)
{
	struct inode *inode = file_inode(filep);
	long ret;

	if (!S_ISREG(inode->i_mode))
		return -EOPNOTSUPP;

	if ((mode != 0) && (mode != (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE)))
		return -EOPNOTSUPP;

	ret = inode_newsize_ok(inode, offset + len);
	if (ret < 0)
		return ret;

	if (mode & FALLOC_FL_PUNCH_HOLE)
		return nfs42_proc_deallocate(filep, offset, len);
	return nfs42_proc_allocate(filep, offset, len);
}

static loff_t nfs42_remap_file_range(struct file *src_file, loff_t src_off,
		struct file *dst_file, loff_t dst_off, loff_t count,
		unsigned int remap_flags)
{
	struct inode *dst_inode = file_inode(dst_file);
	struct nfs_server *server = NFS_SERVER(dst_inode);
	struct inode *src_inode = file_inode(src_file);
	unsigned int bs = server->clone_blksize;
	bool same_inode = false;
	int ret;

	/* NFS does not support deduplication. */
	if (remap_flags & REMAP_FILE_DEDUP)
		return -EOPNOTSUPP;

	if (remap_flags & ~REMAP_FILE_ADVISORY)
		return -EINVAL;

	if (IS_SWAPFILE(dst_inode) || IS_SWAPFILE(src_inode))
		return -ETXTBSY;

	/* check alignment w.r.t. clone_blksize */
	ret = -EINVAL;
	if (bs) {
		if (!IS_ALIGNED(src_off, bs) || !IS_ALIGNED(dst_off, bs))
			goto out;
		if (!IS_ALIGNED(count, bs) && i_size_read(src_inode) != (src_off + count))
			goto out;
	}

	if (src_inode == dst_inode)
		same_inode = true;

	/* XXX: do we lock at all? what if server needs CB_RECALL_LAYOUT? */
	if (same_inode) {
		inode_lock(src_inode);
	} else if (dst_inode < src_inode) {
		inode_lock_nested(dst_inode, I_MUTEX_PARENT);
		inode_lock_nested(src_inode, I_MUTEX_CHILD);
	} else {
		inode_lock_nested(src_inode, I_MUTEX_PARENT);
		inode_lock_nested(dst_inode, I_MUTEX_CHILD);
	}

	/* flush all pending writes on both src and dst so that server
	 * has the latest data */
	ret = nfs_sync_inode(src_inode);
	if (ret)
		goto out_unlock;
	ret = nfs_sync_inode(dst_inode);
	if (ret)
		goto out_unlock;

	ret = nfs42_proc_clone(src_file, dst_file, src_off, dst_off, count);

	/* truncate inode page cache of the dst range so that future reads can fetch
	 * new data from server */
	if (!ret)
		truncate_inode_pages_range(&dst_inode->i_data, dst_off, dst_off + count - 1);

out_unlock:
	if (same_inode) {
		inode_unlock(src_inode);
	} else if (dst_inode < src_inode) {
		inode_unlock(src_inode);
		inode_unlock(dst_inode);
	} else {
		inode_unlock(dst_inode);
		inode_unlock(src_inode);
	}
out:
	return ret < 0 ? ret : count;
}

static int read_name_gen = 1;
#define SSC_READ_NAME_BODY "ssc_read_%d"

static struct file *__nfs42_ssc_open(struct vfsmount *ss_mnt,
		struct nfs_fh *src_fh, nfs4_stateid *stateid)
{
	struct nfs_fattr *fattr = nfs_alloc_fattr();
	struct file *filep, *res;
	struct nfs_server *server;
	struct inode *r_ino = NULL;
	struct nfs_open_context *ctx;
	struct nfs4_state_owner *sp;
	char *read_name = NULL;
	int len, status = 0;

	server = NFS_SB(ss_mnt->mnt_sb);

	if (!fattr)
		return ERR_PTR(-ENOMEM);

	status = nfs4_proc_getattr(server, src_fh, fattr, NULL);
	if (status < 0) {
		res = ERR_PTR(status);
		goto out;
	}

	if (!S_ISREG(fattr->mode)) {
		res = ERR_PTR(-EBADF);
		goto out;
	}

	res = ERR_PTR(-ENOMEM);
	len = strlen(SSC_READ_NAME_BODY) + 16;
	read_name = kzalloc(len, GFP_KERNEL);
	if (read_name == NULL)
		goto out;
	snprintf(read_name, len, SSC_READ_NAME_BODY, read_name_gen++);

	r_ino = nfs_fhget(ss_mnt->mnt_sb, src_fh, fattr);
	if (IS_ERR(r_ino)) {
		res = ERR_CAST(r_ino);
		goto out_free_name;
	}

	filep = alloc_file_pseudo(r_ino, ss_mnt, read_name, O_RDONLY,
				     r_ino->i_fop);
	if (IS_ERR(filep)) {
		res = ERR_CAST(filep);
		iput(r_ino);
		goto out_free_name;
	}

	ctx = alloc_nfs_open_context(filep->f_path.dentry,
				     flags_to_mode(filep->f_flags), filep);
	if (IS_ERR(ctx)) {
		res = ERR_CAST(ctx);
		goto out_filep;
	}

	res = ERR_PTR(-EINVAL);
	sp = nfs4_get_state_owner(server, ctx->cred, GFP_KERNEL);
	if (sp == NULL)
		goto out_ctx;

	ctx->state = nfs4_get_open_state(r_ino, sp);
	if (ctx->state == NULL)
		goto out_stateowner;

	set_bit(NFS_SRV_SSC_COPY_STATE, &ctx->state->flags);
	memcpy(&ctx->state->open_stateid.other, &stateid->other,
	       NFS4_STATEID_OTHER_SIZE);
	update_open_stateid(ctx->state, stateid, NULL, filep->f_mode);
	set_bit(NFS_OPEN_STATE, &ctx->state->flags);

	nfs_file_set_open_context(filep, ctx);
	put_nfs_open_context(ctx);

	file_ra_state_init(&filep->f_ra, filep->f_mapping->host->i_mapping);
	res = filep;
out_free_name:
	kfree(read_name);
out:
	nfs_free_fattr(fattr);
	return res;
out_stateowner:
	nfs4_put_state_owner(sp);
out_ctx:
	put_nfs_open_context(ctx);
out_filep:
	fput(filep);
	goto out_free_name;
}

static void __nfs42_ssc_close(struct file *filep)
{
	struct nfs_open_context *ctx = nfs_file_open_context(filep);

	ctx->state->flags = 0;
}

static const struct nfs4_ssc_client_ops nfs4_ssc_clnt_ops_tbl = {
	.sco_open = __nfs42_ssc_open,
	.sco_close = __nfs42_ssc_close,
};

/**
 * nfs42_ssc_register_ops - Wrapper to register NFS_V4 ops in nfs_common
 *
 * Return values:
 *   None
 */
void nfs42_ssc_register_ops(void)
{
	nfs42_ssc_register(&nfs4_ssc_clnt_ops_tbl);
}

/**
 * nfs42_ssc_unregister_ops - wrapper to un-register NFS_V4 ops in nfs_common
 *
 * Return values:
 *   None.
 */
void nfs42_ssc_unregister_ops(void)
{
	nfs42_ssc_unregister(&nfs4_ssc_clnt_ops_tbl);
}
#endif /* CONFIG_NFS_V4_2 */

static int nfs4_setlease(struct file *file, int arg, struct file_lease **lease,
			 void **priv)
{
	return nfs4_proc_setlease(file, arg, lease, priv);
}

const struct file_operations nfs4_file_operations = {
	.read_iter	= nfs_file_read,
	.write_iter	= nfs_file_write,
	.mmap		= nfs_file_mmap,
	.open		= nfs4_file_open,
	.flush		= nfs4_file_flush,
	.release	= nfs_file_release,
	.fsync		= nfs_file_fsync,
	.lock		= nfs_lock,
	.flock		= nfs_flock,
	.splice_read	= nfs_file_splice_read,
	.splice_write	= iter_file_splice_write,
	.check_flags	= nfs_check_flags,
	.setlease	= nfs4_setlease,
#ifdef CONFIG_NFS_V4_2
	.copy_file_range = nfs4_copy_file_range,
	.llseek		= nfs4_file_llseek,
	.fallocate	= nfs42_fallocate,
	.remap_file_range = nfs42_remap_file_range,
#else
	.llseek		= nfs_file_llseek,
#endif
};
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
ce4ef7c0 BS	2	/*
	3	* linux/fs/nfs/file.c
	4	*
	5	* Copyright (C) 1992 Rick Sladkey
	6	*/
f4ac1674	7	#include <linux/fs.h>
bea51b30	8	#include <linux/file.h>
f4ac1674	9	#include <linux/falloc.h>
f2aedb71	10	#include <linux/mount.h>
ce4ef7c0	11	#include <linux/nfs_fs.h>
0cfcd405	12	#include <linux/nfs_ssc.h>
705bcfcb	13	#include <linux/splice.h>
5445b1fb	14	#include "delegation.h"
ce4ef7c0	15	#include "internal.h"
5445b1fb	16	#include "iostat.h"
a4ff1468	17	#include "fscache.h"
ce4ef7c0 BS	18	#include "pnfs.h"
ce4ef7c0 BS	19
81b79afb TM	20	#include "nfstrace.h"
81b79afb TM	21
1c6dcbe5 AS	22	#ifdef CONFIG_NFS_V4_2
	23	#include "nfs42.h"
	24	#endif
	25
ce4ef7c0 BS	26	#define NFSDBG_FACILITY NFSDBG_FILE
	27
	28	static int
	29	nfs4_file_open(struct inode inode, struct file filp)
	30	{
	31	struct nfs_open_context *ctx;
be62a1a8	32	struct dentry *dentry = file_dentry(filp);
ce4ef7c0 BS	33	struct dentry *parent = NULL;
	34	struct inode *dir;
	35	unsigned openflags = filp->f_flags;
	36	struct iattr attr;
	37	int err;
	38
ce4ef7c0 BS	39	/*
	40	* If no cached dentry exists or if it's negative, NFSv4 handled the
	41	* opens in ->lookup() or ->create().
	42	*
	43	* We only get this far for a cached positive dentry. We skipped
	44	* revalidation, so handle it here by dropping the dentry and returning
	45	* -EOPENSTALE. The VFS will retry the lookup/create/open.
	46	*/
	47
6de1472f	48	dprintk("NFS: open file(%pd2)\n", dentry);
ce4ef7c0	49
18a60089 BC	50	err = nfs_check_flags(openflags);
	51	if (err)
	52	return err;
	53
ce4ef7c0 BS	54	/* We can't create new files here */
	55	openflags &= ~(O_CREAT\|O_EXCL);
	56
	57	parent = dget_parent(dentry);
2b0143b5	58	dir = d_inode(parent);
ce4ef7c0	59
6f1c1d95 C	60	ctx = alloc_nfs_open_context(file_dentry(filp),
6f1c1d95 C	61	flags_to_mode(openflags), filp);
ce4ef7c0 BS	62	err = PTR_ERR(ctx);
	63	if (IS_ERR(ctx))
	64	goto out;
	65
	66	attr.ia_valid = ATTR_OPEN;
	67	if (openflags & O_TRUNC) {
	68	attr.ia_valid \|= ATTR_SIZE;
	69	attr.ia_size = 0;
8b7d9d09	70	filemap_write_and_wait(inode->i_mapping);
ce4ef7c0 BS	71	}
ce4ef7c0 BS	72
c5c3fb5f	73	inode = NFS_PROTO(dir)->open_context(dir, ctx, openflags, &attr, NULL);
ce4ef7c0 BS	74	if (IS_ERR(inode)) {
	75	err = PTR_ERR(inode);
	76	switch (err) {
ce4ef7c0	77	default:
90cf500e TM	78	goto out_put_ctx;
	79	case -ENOENT:
	80	case -ESTALE:
	81	case -EISDIR:
	82	case -ENOTDIR:
	83	case -ELOOP:
ce4ef7c0 BS	84	goto out_drop;
	85	}
	86	}
2b0143b5	87	if (inode != d_inode(dentry))
ce4ef7c0 BS	88	goto out_drop;
ce4ef7c0 BS	89
ce4ef7c0	90	nfs_file_set_open_context(filp, ctx);
f1fe29b4	91	nfs_fscache_open_file(inode, filp);
ce4ef7c0	92	err = 0;
5ee3d10f	93	filp->f_mode \|= FMODE_CAN_ODIRECT;
ce4ef7c0 BS	94
	95	out_put_ctx:
	96	put_nfs_open_context(ctx);
	97	out:
	98	dput(parent);
	99	return err;
	100
	101	out_drop:
	102	d_drop(dentry);
	103	err = -EOPENSTALE;
	104	goto out_put_ctx;
	105	}
	106
5445b1fb TM	107	/*
	108	* Flush all dirty pages, and check for write errors.
	109	*/
	110	static int
	111	nfs4_file_flush(struct file *file, fl_owner_t id)
	112	{
	113	struct inode *inode = file_inode(file);
67dd23f9	114	errseq_t since;
5445b1fb TM	115
	116	dprintk("NFS: flush(%pD2)\n", file);
	117
	118	nfs_inc_stats(inode, NFSIOS_VFSFLUSH);
	119	if ((file->f_mode & FMODE_WRITE) == 0)
	120	return 0;
	121
	122	/*
	123	* If we're holding a write delegation, then check if we're required
	124	* to flush the i/o on close. If not, then just start the i/o now.
	125	*/
	126	if (!nfs4_delegation_flush_on_close(inode))
	127	return filemap_fdatawrite(file->f_mapping);
	128
	129	/* Flush writes to the server and return any errors */
67dd23f9 SM	130	since = filemap_sample_wb_err(file->f_mapping);
	131	nfs_wb_all(inode);
	132	return filemap_check_wb_err(file->f_mapping, since);
5445b1fb TM	133	}
5445b1fb TM	134
1c6dcbe5	135	#ifdef CONFIG_NFS_V4_2
64bf5ff5 DC	136	static ssize_t __nfs4_copy_file_range(struct file *file_in, loff_t pos_in,
	137	struct file *file_out, loff_t pos_out,
	138	size_t count, unsigned int flags)
2e72448b	139	{
0491567b	140	struct nfs42_copy_notify_res *cn_resp = NULL;
1d38f3f0 OK	141	struct nl4_server *nss = NULL;
1d38f3f0 OK	142	nfs4_stateid *cnrs = NULL;
0491567b	143	ssize_t ret;
12751010	144	bool sync = false;
0491567b	145
5dae222a	146	/* Only offload copy if superblock is the same */
8dff1df5	147	if (file_in->f_op != &nfs4_file_operations)
5dae222a	148	return -EXDEV;
d8a6ad91 DW	149	if (!nfs_server_capable(file_inode(file_out), NFS_CAP_COPY) \|\|
d8a6ad91 DW	150	!nfs_server_capable(file_inode(file_in), NFS_CAP_COPY))
0769663b	151	return -EOPNOTSUPP;
837bb1d7	152	if (file_inode(file_in) == file_inode(file_out))
0769663b	153	return -EOPNOTSUPP;
12751010 OK	154	/* if the copy size if smaller than 2 RPC payloads, make it
	155	* synchronous
	156	*/
	157	if (count <= 2 * NFS_SERVER(file_inode(file_in))->rsize)
	158	sync = true;
0e65a32c	159	retry:
0491567b	160	if (!nfs42_files_from_same_server(file_in, file_out)) {
ca7d1d1a DN	161	/*
	162	* for inter copy, if copy size is too small
	163	* then fallback to generic copy.
12751010	164	*/
ca7d1d1a	165	if (sync)
12751010	166	return -EOPNOTSUPP;
0491567b	167	cn_resp = kzalloc(sizeof(struct nfs42_copy_notify_res),
4fb547be	168	GFP_KERNEL);
0491567b OK	169	if (unlikely(cn_resp == NULL))
	170	return -ENOMEM;
	171
	172	ret = nfs42_proc_copy_notify(file_in, file_out, cn_resp);
	173	if (ret) {
	174	ret = -EOPNOTSUPP;
	175	goto out;
	176	}
1d38f3f0 OK	177	nss = &cn_resp->cnr_src;
1d38f3f0 OK	178	cnrs = &cn_resp->cnr_stateid;
0491567b	179	}
1d38f3f0	180	ret = nfs42_proc_copy(file_in, pos_in, file_out, pos_out, count,
12751010	181	nss, cnrs, sync);
0491567b	182	out:
98c27f27 TR	183	kfree(cn_resp);
98c27f27 TR	184
0e65a32c OK	185	if (ret == -EAGAIN)
0e65a32c OK	186	goto retry;
0491567b	187	return ret;
2e72448b AS	188	}
2e72448b AS	189
64bf5ff5 DC	190	static ssize_t nfs4_copy_file_range(struct file *file_in, loff_t pos_in,
	191	struct file *file_out, loff_t pos_out,
	192	size_t count, unsigned int flags)
	193	{
	194	ssize_t ret;
	195
	196	ret = __nfs4_copy_file_range(file_in, pos_in, file_out, pos_out, count,
	197	flags);
5dae222a	198	if (ret == -EOPNOTSUPP \|\| ret == -EXDEV)
705bcfcb AG	199	ret = splice_copy_file_range(file_in, pos_in, file_out,
705bcfcb AG	200	pos_out, count);
64bf5ff5 DC	201	return ret;
	202	}
	203
1c6dcbe5 AS	204	static loff_t nfs4_file_llseek(struct file *filep, loff_t offset, int whence)
	205	{
	206	loff_t ret;
	207
	208	switch (whence) {
	209	case SEEK_HOLE:
	210	case SEEK_DATA:
	211	ret = nfs42_proc_llseek(filep, offset, whence);
e67afa7e	212	if (ret != -EOPNOTSUPP)
1c6dcbe5	213	return ret;
df561f66	214	fallthrough;
1c6dcbe5 AS	215	default:
	216	return nfs_file_llseek(filep, offset, whence);
	217	}
	218	}
f4ac1674 AS	219
	220	static long nfs42_fallocate(struct file *filep, int mode, loff_t offset, loff_t len)
	221	{
	222	struct inode *inode = file_inode(filep);
	223	long ret;
	224
	225	if (!S_ISREG(inode->i_mode))
	226	return -EOPNOTSUPP;
	227
624bd5b7	228	if ((mode != 0) && (mode != (FALLOC_FL_PUNCH_HOLE \| FALLOC_FL_KEEP_SIZE)))
f4ac1674 AS	229	return -EOPNOTSUPP;
	230
	231	ret = inode_newsize_ok(inode, offset + len);
	232	if (ret < 0)
	233	return ret;
	234
624bd5b7	235	if (mode & FALLOC_FL_PUNCH_HOLE)
f830f7dd AS	236	return nfs42_proc_deallocate(filep, offset, len);
f830f7dd AS	237	return nfs42_proc_allocate(filep, offset, len);
f4ac1674	238	}
bea51b30	239
42ec3d4c DW	240	static loff_t nfs42_remap_file_range(struct file *src_file, loff_t src_off,
42ec3d4c DW	241	struct file *dst_file, loff_t dst_off, loff_t count,
2e5dfc99	242	unsigned int remap_flags)
bea51b30 PT	243	{
bea51b30 PT	244	struct inode *dst_inode = file_inode(dst_file);
811b7b85	245	struct nfs_server *server = NFS_SERVER(dst_inode);
04b38d60	246	struct inode *src_inode = file_inode(src_file);
811b7b85	247	unsigned int bs = server->clone_blksize;
21fad313	248	bool same_inode = false;
bea51b30 PT	249	int ret;
bea51b30 PT	250
9026b3a9 DW	251	/* NFS does not support deduplication. */
	252	if (remap_flags & REMAP_FILE_DEDUP)
	253	return -EOPNOTSUPP;
	254
	255	if (remap_flags & ~REMAP_FILE_ADVISORY)
2e5dfc99 DW	256	return -EINVAL;
2e5dfc99 DW	257
f5fdf124 MZ	258	if (IS_SWAPFILE(dst_inode) \|\| IS_SWAPFILE(src_inode))
	259	return -ETXTBSY;
	260
811b7b85 PT	261	/* check alignment w.r.t. clone_blksize */
	262	ret = -EINVAL;
	263	if (bs) {
	264	if (!IS_ALIGNED(src_off, bs) \|\| !IS_ALIGNED(dst_off, bs))
04b38d60	265	goto out;
811b7b85	266	if (!IS_ALIGNED(count, bs) && i_size_read(src_inode) != (src_off + count))
04b38d60	267	goto out;
811b7b85 PT	268	}
811b7b85 PT	269
04b38d60 CH	270	if (src_inode == dst_inode)
04b38d60 CH	271	same_inode = true;
21fad313	272
bea51b30	273	/* XXX: do we lock at all? what if server needs CB_RECALL_LAYOUT? */
21fad313	274	if (same_inode) {
5955102c	275	inode_lock(src_inode);
21fad313	276	} else if (dst_inode < src_inode) {
5955102c AV	277	inode_lock_nested(dst_inode, I_MUTEX_PARENT);
5955102c AV	278	inode_lock_nested(src_inode, I_MUTEX_CHILD);
bea51b30	279	} else {
5955102c AV	280	inode_lock_nested(src_inode, I_MUTEX_PARENT);
5955102c AV	281	inode_lock_nested(dst_inode, I_MUTEX_CHILD);
bea51b30 PT	282	}
	283
	284	/* flush all pending writes on both src and dst so that server
	285	* has the latest data */
	286	ret = nfs_sync_inode(src_inode);
	287	if (ret)
	288	goto out_unlock;
	289	ret = nfs_sync_inode(dst_inode);
	290	if (ret)
	291	goto out_unlock;
	292
04b38d60	293	ret = nfs42_proc_clone(src_file, dst_file, src_off, dst_off, count);
bea51b30 PT	294
	295	/* truncate inode page cache of the dst range so that future reads can fetch
	296	* new data from server */
	297	if (!ret)
	298	truncate_inode_pages_range(&dst_inode->i_data, dst_off, dst_off + count - 1);
	299
	300	out_unlock:
21fad313	301	if (same_inode) {
5955102c	302	inode_unlock(src_inode);
21fad313	303	} else if (dst_inode < src_inode) {
5955102c AV	304	inode_unlock(src_inode);
5955102c AV	305	inode_unlock(dst_inode);
bea51b30	306	} else {
5955102c AV	307	inode_unlock(dst_inode);
5955102c AV	308	inode_unlock(src_inode);
bea51b30	309	}
04b38d60	310	out:
42ec3d4c	311	return ret < 0 ? ret : count;
bea51b30	312	}
ec4b0925 OK	313
	314	static int read_name_gen = 1;
	315	#define SSC_READ_NAME_BODY "ssc_read_%d"
	316
0cfcd405 DN	317	static struct file __nfs42_ssc_open(struct vfsmount ss_mnt,
0cfcd405 DN	318	struct nfs_fh src_fh, nfs4_stateid stateid)
ec4b0925	319	{
156cd285	320	struct nfs_fattr *fattr = nfs_alloc_fattr();
ec4b0925 OK	321	struct file filep, res;
	322	struct nfs_server *server;
	323	struct inode *r_ino = NULL;
	324	struct nfs_open_context *ctx;
	325	struct nfs4_state_owner *sp;
f751c545	326	char *read_name = NULL;
ec4b0925 OK	327	int len, status = 0;
ec4b0925 OK	328
f6957b71	329	server = NFS_SB(ss_mnt->mnt_sb);
ec4b0925	330
156cd285 TM	331	if (!fattr)
156cd285 TM	332	return ERR_PTR(-ENOMEM);
ec4b0925	333
2ef61e0e	334	status = nfs4_proc_getattr(server, src_fh, fattr, NULL);
ec4b0925 OK	335	if (status < 0) {
	336	res = ERR_PTR(status);
	337	goto out;
	338	}
	339
fcfc8be1 OK	340	if (!S_ISREG(fattr->mode)) {
	341	res = ERR_PTR(-EBADF);
	342	goto out;
	343	}
	344
ec4b0925 OK	345	res = ERR_PTR(-ENOMEM);
ec4b0925 OK	346	len = strlen(SSC_READ_NAME_BODY) + 16;
4fb547be	347	read_name = kzalloc(len, GFP_KERNEL);
ec4b0925 OK	348	if (read_name == NULL)
	349	goto out;
	350	snprintf(read_name, len, SSC_READ_NAME_BODY, read_name_gen++);
	351
f6957b71	352	r_ino = nfs_fhget(ss_mnt->mnt_sb, src_fh, fattr);
ec4b0925 OK	353	if (IS_ERR(r_ino)) {
ec4b0925 OK	354	res = ERR_CAST(r_ino);
f751c545	355	goto out_free_name;
ec4b0925 OK	356	}
ec4b0925 OK	357
aa97a3ef	358	filep = alloc_file_pseudo(r_ino, ss_mnt, read_name, O_RDONLY,
ec4b0925 OK	359	r_ino->i_fop);
	360	if (IS_ERR(filep)) {
	361	res = ERR_CAST(filep);
fcfc8be1	362	iput(r_ino);
f751c545	363	goto out_free_name;
ec4b0925	364	}
ec4b0925	365
6f1c1d95 C	366	ctx = alloc_nfs_open_context(filep->f_path.dentry,
6f1c1d95 C	367	flags_to_mode(filep->f_flags), filep);
ec4b0925 OK	368	if (IS_ERR(ctx)) {
	369	res = ERR_CAST(ctx);
	370	goto out_filep;
	371	}
	372
	373	res = ERR_PTR(-EINVAL);
	374	sp = nfs4_get_state_owner(server, ctx->cred, GFP_KERNEL);
	375	if (sp == NULL)
	376	goto out_ctx;
	377
	378	ctx->state = nfs4_get_open_state(r_ino, sp);
	379	if (ctx->state == NULL)
	380	goto out_stateowner;
	381
0b9018b9	382	set_bit(NFS_SRV_SSC_COPY_STATE, &ctx->state->flags);
ec4b0925 OK	383	memcpy(&ctx->state->open_stateid.other, &stateid->other,
	384	NFS4_STATEID_OTHER_SIZE);
	385	update_open_stateid(ctx->state, stateid, NULL, filep->f_mode);
fe8eb820	386	set_bit(NFS_OPEN_STATE, &ctx->state->flags);
ec4b0925 OK	387
	388	nfs_file_set_open_context(filep, ctx);
	389	put_nfs_open_context(ctx);
	390
	391	file_ra_state_init(&filep->f_ra, filep->f_mapping->host->i_mapping);
	392	res = filep;
f751c545 OK	393	out_free_name:
f751c545 OK	394	kfree(read_name);
ec4b0925	395	out:
156cd285	396	nfs_free_fattr(fattr);
ec4b0925 OK	397	return res;
	398	out_stateowner:
	399	nfs4_put_state_owner(sp);
	400	out_ctx:
	401	put_nfs_open_context(ctx);
	402	out_filep:
	403	fput(filep);
f751c545	404	goto out_free_name;
ec4b0925	405	}
0cfcd405 DN	406
0cfcd405 DN	407	static void __nfs42_ssc_close(struct file *filep)
ec4b0925 OK	408	{
	409	struct nfs_open_context *ctx = nfs_file_open_context(filep);
	410
	411	ctx->state->flags = 0;
	412	}
0cfcd405 DN	413
	414	static const struct nfs4_ssc_client_ops nfs4_ssc_clnt_ops_tbl = {
	415	.sco_open = __nfs42_ssc_open,
	416	.sco_close = __nfs42_ssc_close,
	417	};
	418
	419	/**
	420	* nfs42_ssc_register_ops - Wrapper to register NFS_V4 ops in nfs_common
	421	*
	422	* Return values:
	423	* None
	424	*/
	425	void nfs42_ssc_register_ops(void)
	426	{
	427	nfs42_ssc_register(&nfs4_ssc_clnt_ops_tbl);
	428	}
	429
	430	/**
	431	* nfs42_ssc_unregister_ops - wrapper to un-register NFS_V4 ops in nfs_common
	432	*
	433	* Return values:
	434	* None.
	435	*/
	436	void nfs42_ssc_unregister_ops(void)
	437	{
	438	nfs42_ssc_unregister(&nfs4_ssc_clnt_ops_tbl);
	439	}
6b7153da	440	#endif /* CONFIG_NFS_V4_2 */
bea51b30	441
c69ff407	442	static int nfs4_setlease(struct file file, int arg, struct file_lease *lease,
e93a5e93 TM	443	void **priv)
	444	{
	445	return nfs4_proc_setlease(file, arg, lease, priv);
	446	}
	447
ce4ef7c0	448	const struct file_operations nfs4_file_operations = {
3aa2d199	449	.read_iter = nfs_file_read,
edaf4369	450	.write_iter = nfs_file_write,
ce4ef7c0 BS	451	.mmap = nfs_file_mmap,
ce4ef7c0 BS	452	.open = nfs4_file_open,
5445b1fb	453	.flush = nfs4_file_flush,
ce4ef7c0	454	.release = nfs_file_release,
4ff79bc7	455	.fsync = nfs_file_fsync,
ce4ef7c0 BS	456	.lock = nfs_lock,
ce4ef7c0 BS	457	.flock = nfs_flock,
a7db5034	458	.splice_read = nfs_file_splice_read,
4da54c21	459	.splice_write = iter_file_splice_write,
ce4ef7c0	460	.check_flags = nfs_check_flags,
e93a5e93	461	.setlease = nfs4_setlease,
6b7153da	462	#ifdef CONFIG_NFS_V4_2
2e72448b	463	.copy_file_range = nfs4_copy_file_range,
6b7153da CH	464	.llseek = nfs4_file_llseek,
6b7153da CH	465	.fallocate = nfs42_fallocate,
2e5dfc99	466	.remap_file_range = nfs42_remap_file_range,
6b7153da CH	467	#else
	468	.llseek = nfs_file_llseek,
	469	#endif
ce4ef7c0	470	};