[linux.git] / fs / backing-file.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 * Common helpers for stackable filesystems and backing files.
 *
 * Forked from fs/overlayfs/file.c.
 *
 * Copyright (C) 2017 Red Hat, Inc.
 * Copyright (C) 2023 CTERA Networks.
 */

#include <linux/fs.h>
#include <linux/backing-file.h>
#include <linux/splice.h>
#include <linux/mm.h>

#include "internal.h"

/**
 * backing_file_open - open a backing file for kernel internal use
 * @user_path:	path that the user reuqested to open
 * @flags:	open flags
 * @real_path:	path of the backing file
 * @cred:	credentials for open
 *
 * Open a backing file for a stackable filesystem (e.g., overlayfs).
 * @user_path may be on the stackable filesystem and @real_path on the
 * underlying filesystem.  In this case, we want to be able to return the
 * @user_path of the stackable filesystem. This is done by embedding the
 * returned file into a container structure that also stores the stacked
 * file's path, which can be retrieved using backing_file_user_path().
 */
struct file *backing_file_open(const struct path *user_path, int flags,
			       const struct path *real_path,
			       const struct cred *cred)
{
	struct file *f;
	int error;

	f = alloc_empty_backing_file(flags, cred);
	if (IS_ERR(f))
		return f;

	path_get(user_path);
	*backing_file_user_path(f) = *user_path;
	error = vfs_open(real_path, f);
	if (error) {
		fput(f);
		f = ERR_PTR(error);
	}

	return f;
}
EXPORT_SYMBOL_GPL(backing_file_open);

struct backing_aio {
	struct kiocb iocb;
	refcount_t ref;
	struct kiocb *orig_iocb;
	/* used for aio completion */
	void (*end_write)(struct file *);
	struct work_struct work;
	long res;
};

static struct kmem_cache *backing_aio_cachep;

#define BACKING_IOCB_MASK \
	(IOCB_NOWAIT | IOCB_HIPRI | IOCB_DSYNC | IOCB_SYNC | IOCB_APPEND)

static rwf_t iocb_to_rw_flags(int flags)
{
	return (__force rwf_t)(flags & BACKING_IOCB_MASK);
}

static void backing_aio_put(struct backing_aio *aio)
{
	if (refcount_dec_and_test(&aio->ref)) {
		fput(aio->iocb.ki_filp);
		kmem_cache_free(backing_aio_cachep, aio);
	}
}

static void backing_aio_cleanup(struct backing_aio *aio, long res)
{
	struct kiocb *iocb = &aio->iocb;
	struct kiocb *orig_iocb = aio->orig_iocb;

	if (aio->end_write)
		aio->end_write(orig_iocb->ki_filp);

	orig_iocb->ki_pos = iocb->ki_pos;
	backing_aio_put(aio);
}

static void backing_aio_rw_complete(struct kiocb *iocb, long res)
{
	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);
	struct kiocb *orig_iocb = aio->orig_iocb;

	if (iocb->ki_flags & IOCB_WRITE)
		kiocb_end_write(iocb);

	backing_aio_cleanup(aio, res);
	orig_iocb->ki_complete(orig_iocb, res);
}

static void backing_aio_complete_work(struct work_struct *work)
{
	struct backing_aio *aio = container_of(work, struct backing_aio, work);

	backing_aio_rw_complete(&aio->iocb, aio->res);
}

static void backing_aio_queue_completion(struct kiocb *iocb, long res)
{
	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);

	/*
	 * Punt to a work queue to serialize updates of mtime/size.
	 */
	aio->res = res;
	INIT_WORK(&aio->work, backing_aio_complete_work);
	queue_work(file_inode(aio->orig_iocb->ki_filp)->i_sb->s_dio_done_wq,
		   &aio->work);
}

static int backing_aio_init_wq(struct kiocb *iocb)
{
	struct super_block *sb = file_inode(iocb->ki_filp)->i_sb;

	if (sb->s_dio_done_wq)
		return 0;

	return sb_init_dio_done_wq(sb);
}


ssize_t backing_file_read_iter(struct file *file, struct iov_iter *iter,
			       struct kiocb *iocb, int flags,
			       struct backing_file_ctx *ctx)
{
	struct backing_aio *aio = NULL;
	const struct cred *old_cred;
	ssize_t ret;

	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
		return -EIO;

	if (!iov_iter_count(iter))
		return 0;

	if (iocb->ki_flags & IOCB_DIRECT &&
	    !(file->f_mode & FMODE_CAN_ODIRECT))
		return -EINVAL;

	old_cred = override_creds(ctx->cred);
	if (is_sync_kiocb(iocb)) {
		rwf_t rwf = iocb_to_rw_flags(flags);

		ret = vfs_iter_read(file, iter, &iocb->ki_pos, rwf);
	} else {
		ret = -ENOMEM;
		aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
		if (!aio)
			goto out;

		aio->orig_iocb = iocb;
		kiocb_clone(&aio->iocb, iocb, get_file(file));
		aio->iocb.ki_complete = backing_aio_rw_complete;
		refcount_set(&aio->ref, 2);
		ret = vfs_iocb_iter_read(file, &aio->iocb, iter);
		backing_aio_put(aio);
		if (ret != -EIOCBQUEUED)
			backing_aio_cleanup(aio, ret);
	}
out:
	revert_creds(old_cred);

	if (ctx->accessed)
		ctx->accessed(ctx->user_file);

	return ret;
}
EXPORT_SYMBOL_GPL(backing_file_read_iter);

ssize_t backing_file_write_iter(struct file *file, struct iov_iter *iter,
				struct kiocb *iocb, int flags,
				struct backing_file_ctx *ctx)
{
	const struct cred *old_cred;
	ssize_t ret;

	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
		return -EIO;

	if (!iov_iter_count(iter))
		return 0;

	ret = file_remove_privs(ctx->user_file);
	if (ret)
		return ret;

	if (iocb->ki_flags & IOCB_DIRECT &&
	    !(file->f_mode & FMODE_CAN_ODIRECT))
		return -EINVAL;

	/*
	 * Stacked filesystems don't support deferred completions, don't copy
	 * this property in case it is set by the issuer.
	 */
	flags &= ~IOCB_DIO_CALLER_COMP;

	old_cred = override_creds(ctx->cred);
	if (is_sync_kiocb(iocb)) {
		rwf_t rwf = iocb_to_rw_flags(flags);

		ret = vfs_iter_write(file, iter, &iocb->ki_pos, rwf);
		if (ctx->end_write)
			ctx->end_write(ctx->user_file);
	} else {
		struct backing_aio *aio;

		ret = backing_aio_init_wq(iocb);
		if (ret)
			goto out;

		ret = -ENOMEM;
		aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
		if (!aio)
			goto out;

		aio->orig_iocb = iocb;
		aio->end_write = ctx->end_write;
		kiocb_clone(&aio->iocb, iocb, get_file(file));
		aio->iocb.ki_flags = flags;
		aio->iocb.ki_complete = backing_aio_queue_completion;
		refcount_set(&aio->ref, 2);
		ret = vfs_iocb_iter_write(file, &aio->iocb, iter);
		backing_aio_put(aio);
		if (ret != -EIOCBQUEUED)
			backing_aio_cleanup(aio, ret);
	}
out:
	revert_creds(old_cred);

	return ret;
}
EXPORT_SYMBOL_GPL(backing_file_write_iter);

ssize_t backing_file_splice_read(struct file *in, loff_t *ppos,
				 struct pipe_inode_info *pipe, size_t len,
				 unsigned int flags,
				 struct backing_file_ctx *ctx)
{
	const struct cred *old_cred;
	ssize_t ret;

	if (WARN_ON_ONCE(!(in->f_mode & FMODE_BACKING)))
		return -EIO;

	old_cred = override_creds(ctx->cred);
	ret = vfs_splice_read(in, ppos, pipe, len, flags);
	revert_creds(old_cred);

	if (ctx->accessed)
		ctx->accessed(ctx->user_file);

	return ret;
}
EXPORT_SYMBOL_GPL(backing_file_splice_read);

ssize_t backing_file_splice_write(struct pipe_inode_info *pipe,
				  struct file *out, loff_t *ppos, size_t len,
				  unsigned int flags,
				  struct backing_file_ctx *ctx)
{
	const struct cred *old_cred;
	ssize_t ret;

	if (WARN_ON_ONCE(!(out->f_mode & FMODE_BACKING)))
		return -EIO;

	ret = file_remove_privs(ctx->user_file);
	if (ret)
		return ret;

	old_cred = override_creds(ctx->cred);
	file_start_write(out);
	ret = iter_file_splice_write(pipe, out, ppos, len, flags);
	file_end_write(out);
	revert_creds(old_cred);

	if (ctx->end_write)
		ctx->end_write(ctx->user_file);

	return ret;
}
EXPORT_SYMBOL_GPL(backing_file_splice_write);

int backing_file_mmap(struct file *file, struct vm_area_struct *vma,
		      struct backing_file_ctx *ctx)
{
	const struct cred *old_cred;
	int ret;

	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)) ||
	    WARN_ON_ONCE(ctx->user_file != vma->vm_file))
		return -EIO;

	if (!file->f_op->mmap)
		return -ENODEV;

	vma_set_file(vma, file);

	old_cred = override_creds(ctx->cred);
	ret = call_mmap(vma->vm_file, vma);
	revert_creds(old_cred);

	if (ctx->accessed)
		ctx->accessed(ctx->user_file);

	return ret;
}
EXPORT_SYMBOL_GPL(backing_file_mmap);

static int __init backing_aio_init(void)
{
	backing_aio_cachep = kmem_cache_create("backing_aio",
					       sizeof(struct backing_aio),
					       0, SLAB_HWCACHE_ALIGN, NULL);
	if (!backing_aio_cachep)
		return -ENOMEM;

	return 0;
}
fs_initcall(backing_aio_init);
Commit	Line	Data
f91a704f AG	1	// SPDX-License-Identifier: GPL-2.0-only
	2	/*
	3	* Common helpers for stackable filesystems and backing files.
	4	*
a6293b3e AG	5	* Forked from fs/overlayfs/file.c.
	6	*
	7	* Copyright (C) 2017 Red Hat, Inc.
f91a704f AG	8	* Copyright (C) 2023 CTERA Networks.
	9	*/
	10
	11	#include <linux/fs.h>
	12	#include <linux/backing-file.h>
9b7e9e2f	13	#include <linux/splice.h>
f567377e	14	#include <linux/mm.h>
f91a704f AG	15
	16	#include "internal.h"
	17
	18	/**
	19	* backing_file_open - open a backing file for kernel internal use
	20	* @user_path: path that the user reuqested to open
	21	* @flags: open flags
	22	* @real_path: path of the backing file
	23	* @cred: credentials for open
	24	*
	25	* Open a backing file for a stackable filesystem (e.g., overlayfs).
	26	* @user_path may be on the stackable filesystem and @real_path on the
	27	* underlying filesystem. In this case, we want to be able to return the
	28	* @user_path of the stackable filesystem. This is done by embedding the
	29	* returned file into a container structure that also stores the stacked
	30	* file's path, which can be retrieved using backing_file_user_path().
	31	*/
	32	struct file backing_file_open(const struct path user_path, int flags,
	33	const struct path *real_path,
	34	const struct cred *cred)
	35	{
	36	struct file *f;
	37	int error;
	38
	39	f = alloc_empty_backing_file(flags, cred);
	40	if (IS_ERR(f))
	41	return f;
	42
	43	path_get(user_path);
	44	backing_file_user_path(f) = user_path;
	45	error = vfs_open(real_path, f);
	46	if (error) {
	47	fput(f);
	48	f = ERR_PTR(error);
	49	}
	50
	51	return f;
	52	}
	53	EXPORT_SYMBOL_GPL(backing_file_open);
a6293b3e AG	54
	55	struct backing_aio {
	56	struct kiocb iocb;
	57	refcount_t ref;
	58	struct kiocb *orig_iocb;
	59	/* used for aio completion */
	60	void (end_write)(struct file );
	61	struct work_struct work;
	62	long res;
	63	};
	64
	65	static struct kmem_cache *backing_aio_cachep;
	66
	67	#define BACKING_IOCB_MASK \
	68	(IOCB_NOWAIT \| IOCB_HIPRI \| IOCB_DSYNC \| IOCB_SYNC \| IOCB_APPEND)
	69
	70	static rwf_t iocb_to_rw_flags(int flags)
	71	{
	72	return (__force rwf_t)(flags & BACKING_IOCB_MASK);
	73	}
	74
	75	static void backing_aio_put(struct backing_aio *aio)
	76	{
	77	if (refcount_dec_and_test(&aio->ref)) {
	78	fput(aio->iocb.ki_filp);
	79	kmem_cache_free(backing_aio_cachep, aio);
	80	}
	81	}
	82
	83	static void backing_aio_cleanup(struct backing_aio *aio, long res)
	84	{
	85	struct kiocb *iocb = &aio->iocb;
	86	struct kiocb *orig_iocb = aio->orig_iocb;
	87
	88	if (aio->end_write)
	89	aio->end_write(orig_iocb->ki_filp);
	90
	91	orig_iocb->ki_pos = iocb->ki_pos;
	92	backing_aio_put(aio);
	93	}
	94
	95	static void backing_aio_rw_complete(struct kiocb *iocb, long res)
	96	{
	97	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);
	98	struct kiocb *orig_iocb = aio->orig_iocb;
	99
	100	if (iocb->ki_flags & IOCB_WRITE)
	101	kiocb_end_write(iocb);
	102
	103	backing_aio_cleanup(aio, res);
	104	orig_iocb->ki_complete(orig_iocb, res);
	105	}
	106
	107	static void backing_aio_complete_work(struct work_struct *work)
	108	{
	109	struct backing_aio *aio = container_of(work, struct backing_aio, work);
	110
	111	backing_aio_rw_complete(&aio->iocb, aio->res);
	112	}
	113
	114	static void backing_aio_queue_completion(struct kiocb *iocb, long res)
	115	{
	116	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);
	117
118	/*
119	* Punt to a work queue to serialize updates of mtime/size.
120	*/
121	aio->res = res;
122	INIT_WORK(&aio->work, backing_aio_complete_work);
123	queue_work(file_inode(aio->orig_iocb->ki_filp)->i_sb->s_dio_done_wq,
124	&aio->work);
125	}
126
127	static int backing_aio_init_wq(struct kiocb *iocb)
128	{
129	struct super_block *sb = file_inode(iocb->ki_filp)->i_sb;
130
131	if (sb->s_dio_done_wq)
132	return 0;
133
134	return sb_init_dio_done_wq(sb);
135	}
136
137
138	ssize_t backing_file_read_iter(struct file file, struct iov_iter iter,
139	struct kiocb *iocb, int flags,
140	struct backing_file_ctx *ctx)
141	{
142	struct backing_aio *aio = NULL;
143	const struct cred *old_cred;
144	ssize_t ret;
145
146	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
147	return -EIO;
148
149	if (!iov_iter_count(iter))
150	return 0;
151
152	if (iocb->ki_flags & IOCB_DIRECT &&
153	!(file->f_mode & FMODE_CAN_ODIRECT))
154	return -EINVAL;
155
156	old_cred = override_creds(ctx->cred);
157	if (is_sync_kiocb(iocb)) {
158	rwf_t rwf = iocb_to_rw_flags(flags);
159
160	ret = vfs_iter_read(file, iter, &iocb->ki_pos, rwf);
161	} else {
162	ret = -ENOMEM;
163	aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
164	if (!aio)
165	goto out;
166
167	aio->orig_iocb = iocb;
168	kiocb_clone(&aio->iocb, iocb, get_file(file));
169	aio->iocb.ki_complete = backing_aio_rw_complete;
170	refcount_set(&aio->ref, 2);
171	ret = vfs_iocb_iter_read(file, &aio->iocb, iter);
172	backing_aio_put(aio);
173	if (ret != -EIOCBQUEUED)
174	backing_aio_cleanup(aio, ret);
175	}
176	out:
177	revert_creds(old_cred);
178
179	if (ctx->accessed)
180	ctx->accessed(ctx->user_file);
181
182	return ret;
183	}
184	EXPORT_SYMBOL_GPL(backing_file_read_iter);
185
186	ssize_t backing_file_write_iter(struct file file, struct iov_iter iter,
187	struct kiocb *iocb, int flags,
188	struct backing_file_ctx *ctx)
189	{
190	const struct cred *old_cred;
191	ssize_t ret;
192
193	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
194	return -EIO;
195
196	if (!iov_iter_count(iter))
197	return 0;
198
199	ret = file_remove_privs(ctx->user_file);
200	if (ret)
201	return ret;
202
203	if (iocb->ki_flags & IOCB_DIRECT &&
204	!(file->f_mode & FMODE_CAN_ODIRECT))
205	return -EINVAL;
206
207	/*
208	* Stacked filesystems don't support deferred completions, don't copy
209	* this property in case it is set by the issuer.
210	*/
211	flags &= ~IOCB_DIO_CALLER_COMP;
212
213	old_cred = override_creds(ctx->cred);
214	if (is_sync_kiocb(iocb)) {
215	rwf_t rwf = iocb_to_rw_flags(flags);
216
217	ret = vfs_iter_write(file, iter, &iocb->ki_pos, rwf);
218	if (ctx->end_write)
219	ctx->end_write(ctx->user_file);
220	} else {
221	struct backing_aio *aio;
222
223	ret = backing_aio_init_wq(iocb);
224	if (ret)
225	goto out;
226
227	ret = -ENOMEM;
228	aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
229	if (!aio)
230	goto out;
231
232	aio->orig_iocb = iocb;
233	aio->end_write = ctx->end_write;
234	kiocb_clone(&aio->iocb, iocb, get_file(file));
235	aio->iocb.ki_flags = flags;
236	aio->iocb.ki_complete = backing_aio_queue_completion;
237	refcount_set(&aio->ref, 2);
238	ret = vfs_iocb_iter_write(file, &aio->iocb, iter);
239	backing_aio_put(aio);
240	if (ret != -EIOCBQUEUED)
241	backing_aio_cleanup(aio, ret);
242	}
243	out:
244	revert_creds(old_cred);
245
246	return ret;
247	}
248	EXPORT_SYMBOL_GPL(backing_file_write_iter);
249
9b7e9e2f AG	250	ssize_t backing_file_splice_read(struct file in, loff_t ppos,
	251	struct pipe_inode_info *pipe, size_t len,
	252	unsigned int flags,
	253	struct backing_file_ctx *ctx)
	254	{
	255	const struct cred *old_cred;
	256	ssize_t ret;
	257
	258	if (WARN_ON_ONCE(!(in->f_mode & FMODE_BACKING)))
	259	return -EIO;
	260
	261	old_cred = override_creds(ctx->cred);
	262	ret = vfs_splice_read(in, ppos, pipe, len, flags);
	263	revert_creds(old_cred);
	264
	265	if (ctx->accessed)
	266	ctx->accessed(ctx->user_file);
	267
	268	return ret;
	269	}
	270	EXPORT_SYMBOL_GPL(backing_file_splice_read);
	271
	272	ssize_t backing_file_splice_write(struct pipe_inode_info *pipe,
	273	struct file out, loff_t ppos, size_t len,
	274	unsigned int flags,
	275	struct backing_file_ctx *ctx)
	276	{
	277	const struct cred *old_cred;
	278	ssize_t ret;
	279
	280	if (WARN_ON_ONCE(!(out->f_mode & FMODE_BACKING)))
	281	return -EIO;
	282
	283	ret = file_remove_privs(ctx->user_file);
	284	if (ret)
	285	return ret;
	286
	287	old_cred = override_creds(ctx->cred);
	288	file_start_write(out);
	289	ret = iter_file_splice_write(pipe, out, ppos, len, flags);
	290	file_end_write(out);
	291	revert_creds(old_cred);
	292
	293	if (ctx->end_write)
	294	ctx->end_write(ctx->user_file);
	295
	296	return ret;
	297	}
	298	EXPORT_SYMBOL_GPL(backing_file_splice_write);
	299
f567377e AG	300	int backing_file_mmap(struct file file, struct vm_area_struct vma,
	301	struct backing_file_ctx *ctx)
	302	{
	303	const struct cred *old_cred;
	304	int ret;
	305
	306	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)) \|\|
	307	WARN_ON_ONCE(ctx->user_file != vma->vm_file))
	308	return -EIO;
	309
	310	if (!file->f_op->mmap)
	311	return -ENODEV;
	312
	313	vma_set_file(vma, file);
	314
	315	old_cred = override_creds(ctx->cred);
	316	ret = call_mmap(vma->vm_file, vma);
	317	revert_creds(old_cred);
	318
	319	if (ctx->accessed)
	320	ctx->accessed(ctx->user_file);
	321
	322	return ret;
	323	}
	324	EXPORT_SYMBOL_GPL(backing_file_mmap);
	325
a6293b3e AG	326	static int __init backing_aio_init(void)
	327	{
	328	backing_aio_cachep = kmem_cache_create("backing_aio",
	329	sizeof(struct backing_aio),
	330	0, SLAB_HWCACHE_ALIGN, NULL);
	331	if (!backing_aio_cachep)
	332	return -ENOMEM;
	333
	334	return 0;
	335	}
	336	fs_initcall(backing_aio_init);