[thirdparty/kernel/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 file *backing_tmpfile_open(const struct path *user_path, int flags,
				  const struct path *real_parentpath,
				  umode_t mode, const struct cred *cred)
{
	struct mnt_idmap *real_idmap = mnt_idmap(real_parentpath->mnt);
	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_tmpfile(real_idmap, real_parentpath, f, mode);
	if (error) {
		fput(f);
		f = ERR_PTR(error);
	}
	return f;
}
EXPORT_SYMBOL(backing_tmpfile_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(backing_aio, SLAB_HWCACHE_ALIGN);
	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	54
9a87907d MS	55	struct file backing_tmpfile_open(const struct path user_path, int flags,
	56	const struct path *real_parentpath,
	57	umode_t mode, const struct cred *cred)
	58	{
	59	struct mnt_idmap *real_idmap = mnt_idmap(real_parentpath->mnt);
	60	struct file *f;
	61	int error;
	62
	63	f = alloc_empty_backing_file(flags, cred);
	64	if (IS_ERR(f))
	65	return f;
	66
	67	path_get(user_path);
	68	backing_file_user_path(f) = user_path;
	69	error = vfs_tmpfile(real_idmap, real_parentpath, f, mode);
	70	if (error) {
	71	fput(f);
	72	f = ERR_PTR(error);
	73	}
	74	return f;
	75	}
	76	EXPORT_SYMBOL(backing_tmpfile_open);
	77
a6293b3e AG	78	struct backing_aio {
	79	struct kiocb iocb;
	80	refcount_t ref;
	81	struct kiocb *orig_iocb;
	82	/* used for aio completion */
	83	void (end_write)(struct file );
	84	struct work_struct work;
	85	long res;
	86	};
	87
	88	static struct kmem_cache *backing_aio_cachep;
	89
	90	#define BACKING_IOCB_MASK \
	91	(IOCB_NOWAIT \| IOCB_HIPRI \| IOCB_DSYNC \| IOCB_SYNC \| IOCB_APPEND)
	92
	93	static rwf_t iocb_to_rw_flags(int flags)
	94	{
	95	return (__force rwf_t)(flags & BACKING_IOCB_MASK);
	96	}
	97
	98	static void backing_aio_put(struct backing_aio *aio)
	99	{
	100	if (refcount_dec_and_test(&aio->ref)) {
	101	fput(aio->iocb.ki_filp);
	102	kmem_cache_free(backing_aio_cachep, aio);
	103	}
	104	}
	105
	106	static void backing_aio_cleanup(struct backing_aio *aio, long res)
	107	{
	108	struct kiocb *iocb = &aio->iocb;
	109	struct kiocb *orig_iocb = aio->orig_iocb;
	110
	111	if (aio->end_write)
	112	aio->end_write(orig_iocb->ki_filp);
	113
	114	orig_iocb->ki_pos = iocb->ki_pos;
	115	backing_aio_put(aio);
	116	}
	117
	118	static void backing_aio_rw_complete(struct kiocb *iocb, long res)
	119	{
	120	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);
	121	struct kiocb *orig_iocb = aio->orig_iocb;
	122
	123	if (iocb->ki_flags & IOCB_WRITE)
	124	kiocb_end_write(iocb);
	125
	126	backing_aio_cleanup(aio, res);
	127	orig_iocb->ki_complete(orig_iocb, res);
	128	}
	129
	130	static void backing_aio_complete_work(struct work_struct *work)
	131	{
	132	struct backing_aio *aio = container_of(work, struct backing_aio, work);
	133
	134	backing_aio_rw_complete(&aio->iocb, aio->res);
	135	}
	136
	137	static void backing_aio_queue_completion(struct kiocb *iocb, long res)
	138	{
	139	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);
	140
	141	/*
142	* Punt to a work queue to serialize updates of mtime/size.
143	*/
144	aio->res = res;
145	INIT_WORK(&aio->work, backing_aio_complete_work);
146	queue_work(file_inode(aio->orig_iocb->ki_filp)->i_sb->s_dio_done_wq,
147	&aio->work);
148	}
149
150	static int backing_aio_init_wq(struct kiocb *iocb)
151	{
152	struct super_block *sb = file_inode(iocb->ki_filp)->i_sb;
153
154	if (sb->s_dio_done_wq)
155	return 0;
156
157	return sb_init_dio_done_wq(sb);
158	}
159
160
161	ssize_t backing_file_read_iter(struct file file, struct iov_iter iter,
162	struct kiocb *iocb, int flags,
163	struct backing_file_ctx *ctx)
164	{
165	struct backing_aio *aio = NULL;
166	const struct cred *old_cred;
167	ssize_t ret;
168
169	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
170	return -EIO;
171
172	if (!iov_iter_count(iter))
173	return 0;
174
175	if (iocb->ki_flags & IOCB_DIRECT &&
176	!(file->f_mode & FMODE_CAN_ODIRECT))
177	return -EINVAL;
178
179	old_cred = override_creds(ctx->cred);
180	if (is_sync_kiocb(iocb)) {
181	rwf_t rwf = iocb_to_rw_flags(flags);
182
183	ret = vfs_iter_read(file, iter, &iocb->ki_pos, rwf);
184	} else {
185	ret = -ENOMEM;
186	aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
187	if (!aio)
188	goto out;
189
190	aio->orig_iocb = iocb;
191	kiocb_clone(&aio->iocb, iocb, get_file(file));
192	aio->iocb.ki_complete = backing_aio_rw_complete;
193	refcount_set(&aio->ref, 2);
194	ret = vfs_iocb_iter_read(file, &aio->iocb, iter);
195	backing_aio_put(aio);
196	if (ret != -EIOCBQUEUED)
197	backing_aio_cleanup(aio, ret);
198	}
199	out:
200	revert_creds(old_cred);
201
202	if (ctx->accessed)
203	ctx->accessed(ctx->user_file);
204
205	return ret;
206	}
207	EXPORT_SYMBOL_GPL(backing_file_read_iter);
208
209	ssize_t backing_file_write_iter(struct file file, struct iov_iter iter,
210	struct kiocb *iocb, int flags,
211	struct backing_file_ctx *ctx)
212	{
213	const struct cred *old_cred;
214	ssize_t ret;
215
216	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
217	return -EIO;
218
219	if (!iov_iter_count(iter))
220	return 0;
221
222	ret = file_remove_privs(ctx->user_file);
223	if (ret)
224	return ret;
225
226	if (iocb->ki_flags & IOCB_DIRECT &&
227	!(file->f_mode & FMODE_CAN_ODIRECT))
228	return -EINVAL;
229
230	/*
231	* Stacked filesystems don't support deferred completions, don't copy
232	* this property in case it is set by the issuer.
233	*/
234	flags &= ~IOCB_DIO_CALLER_COMP;
235
236	old_cred = override_creds(ctx->cred);
237	if (is_sync_kiocb(iocb)) {
238	rwf_t rwf = iocb_to_rw_flags(flags);
239
240	ret = vfs_iter_write(file, iter, &iocb->ki_pos, rwf);
241	if (ctx->end_write)
242	ctx->end_write(ctx->user_file);
243	} else {
244	struct backing_aio *aio;
245
246	ret = backing_aio_init_wq(iocb);
247	if (ret)
248	goto out;
249
250	ret = -ENOMEM;
251	aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
252	if (!aio)
253	goto out;
254
255	aio->orig_iocb = iocb;
256	aio->end_write = ctx->end_write;
257	kiocb_clone(&aio->iocb, iocb, get_file(file));
258	aio->iocb.ki_flags = flags;
259	aio->iocb.ki_complete = backing_aio_queue_completion;
260	refcount_set(&aio->ref, 2);
261	ret = vfs_iocb_iter_write(file, &aio->iocb, iter);
262	backing_aio_put(aio);
263	if (ret != -EIOCBQUEUED)
264	backing_aio_cleanup(aio, ret);
265	}
266	out:
267	revert_creds(old_cred);
268
269	return ret;
270	}
271	EXPORT_SYMBOL_GPL(backing_file_write_iter);
272
9b7e9e2f AG	273	ssize_t backing_file_splice_read(struct file in, loff_t ppos,
	274	struct pipe_inode_info *pipe, size_t len,
	275	unsigned int flags,
	276	struct backing_file_ctx *ctx)
	277	{
	278	const struct cred *old_cred;
	279	ssize_t ret;
	280
	281	if (WARN_ON_ONCE(!(in->f_mode & FMODE_BACKING)))
	282	return -EIO;
	283
	284	old_cred = override_creds(ctx->cred);
	285	ret = vfs_splice_read(in, ppos, pipe, len, flags);
	286	revert_creds(old_cred);
	287
	288	if (ctx->accessed)
	289	ctx->accessed(ctx->user_file);
	290
	291	return ret;
	292	}
	293	EXPORT_SYMBOL_GPL(backing_file_splice_read);
	294
	295	ssize_t backing_file_splice_write(struct pipe_inode_info *pipe,
	296	struct file out, loff_t ppos, size_t len,
	297	unsigned int flags,
	298	struct backing_file_ctx *ctx)
	299	{
	300	const struct cred *old_cred;
	301	ssize_t ret;
	302
	303	if (WARN_ON_ONCE(!(out->f_mode & FMODE_BACKING)))
	304	return -EIO;
	305
	306	ret = file_remove_privs(ctx->user_file);
	307	if (ret)
	308	return ret;
	309
	310	old_cred = override_creds(ctx->cred);
	311	file_start_write(out);
	312	ret = iter_file_splice_write(pipe, out, ppos, len, flags);
	313	file_end_write(out);
	314	revert_creds(old_cred);
	315
	316	if (ctx->end_write)
	317	ctx->end_write(ctx->user_file);
	318
	319	return ret;
	320	}
	321	EXPORT_SYMBOL_GPL(backing_file_splice_write);
	322
f567377e AG	323	int backing_file_mmap(struct file file, struct vm_area_struct vma,
	324	struct backing_file_ctx *ctx)
	325	{
	326	const struct cred *old_cred;
	327	int ret;
	328
	329	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)) \|\|
	330	WARN_ON_ONCE(ctx->user_file != vma->vm_file))
	331	return -EIO;
	332
	333	if (!file->f_op->mmap)
	334	return -ENODEV;
	335
	336	vma_set_file(vma, file);
	337
	338	old_cred = override_creds(ctx->cred);
	339	ret = call_mmap(vma->vm_file, vma);
	340	revert_creds(old_cred);
	341
	342	if (ctx->accessed)
	343	ctx->accessed(ctx->user_file);
	344
	345	return ret;
	346	}
	347	EXPORT_SYMBOL_GPL(backing_file_mmap);
	348
a6293b3e AG	349	static int __init backing_aio_init(void)
a6293b3e AG	350	{
617fc777	351	backing_aio_cachep = KMEM_CACHE(backing_aio, SLAB_HWCACHE_ALIGN);
a6293b3e AG	352	if (!backing_aio_cachep)
	353	return -ENOMEM;
	354
	355	return 0;
	356	}
	357	fs_initcall(backing_aio_init);