[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 "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);

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>
	13
	14	#include "internal.h"
	15
	16	/**
	17	* backing_file_open - open a backing file for kernel internal use
	18	* @user_path: path that the user reuqested to open
	19	* @flags: open flags
	20	* @real_path: path of the backing file
	21	* @cred: credentials for open
	22	*
	23	* Open a backing file for a stackable filesystem (e.g., overlayfs).
	24	* @user_path may be on the stackable filesystem and @real_path on the
	25	* underlying filesystem. In this case, we want to be able to return the
	26	* @user_path of the stackable filesystem. This is done by embedding the
	27	* returned file into a container structure that also stores the stacked
	28	* file's path, which can be retrieved using backing_file_user_path().
	29	*/
	30	struct file backing_file_open(const struct path user_path, int flags,
	31	const struct path *real_path,
	32	const struct cred *cred)
	33	{
	34	struct file *f;
	35	int error;
	36
	37	f = alloc_empty_backing_file(flags, cred);
	38	if (IS_ERR(f))
	39	return f;
	40
	41	path_get(user_path);
	42	backing_file_user_path(f) = user_path;
	43	error = vfs_open(real_path, f);
	44	if (error) {
	45	fput(f);
	46	f = ERR_PTR(error);
	47	}
	48
	49	return f;
	50	}
	51	EXPORT_SYMBOL_GPL(backing_file_open);
a6293b3e AG	52
	53	struct backing_aio {
	54	struct kiocb iocb;
	55	refcount_t ref;
	56	struct kiocb *orig_iocb;
	57	/* used for aio completion */
	58	void (end_write)(struct file );
	59	struct work_struct work;
	60	long res;
	61	};
	62
	63	static struct kmem_cache *backing_aio_cachep;
	64
	65	#define BACKING_IOCB_MASK \
	66	(IOCB_NOWAIT \| IOCB_HIPRI \| IOCB_DSYNC \| IOCB_SYNC \| IOCB_APPEND)
	67
	68	static rwf_t iocb_to_rw_flags(int flags)
	69	{
	70	return (__force rwf_t)(flags & BACKING_IOCB_MASK);
	71	}
	72
	73	static void backing_aio_put(struct backing_aio *aio)
	74	{
	75	if (refcount_dec_and_test(&aio->ref)) {
	76	fput(aio->iocb.ki_filp);
	77	kmem_cache_free(backing_aio_cachep, aio);
	78	}
	79	}
	80
	81	static void backing_aio_cleanup(struct backing_aio *aio, long res)
	82	{
	83	struct kiocb *iocb = &aio->iocb;
	84	struct kiocb *orig_iocb = aio->orig_iocb;
	85
	86	if (aio->end_write)
	87	aio->end_write(orig_iocb->ki_filp);
	88
	89	orig_iocb->ki_pos = iocb->ki_pos;
	90	backing_aio_put(aio);
	91	}
	92
	93	static void backing_aio_rw_complete(struct kiocb *iocb, long res)
	94	{
	95	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);
	96	struct kiocb *orig_iocb = aio->orig_iocb;
	97
	98	if (iocb->ki_flags & IOCB_WRITE)
	99	kiocb_end_write(iocb);
	100
	101	backing_aio_cleanup(aio, res);
	102	orig_iocb->ki_complete(orig_iocb, res);
	103	}
	104
	105	static void backing_aio_complete_work(struct work_struct *work)
	106	{
	107	struct backing_aio *aio = container_of(work, struct backing_aio, work);
	108
	109	backing_aio_rw_complete(&aio->iocb, aio->res);
	110	}
	111
	112	static void backing_aio_queue_completion(struct kiocb *iocb, long res)
	113	{
	114	struct backing_aio *aio = container_of(iocb, struct backing_aio, iocb);
	115
116	/*
117	* Punt to a work queue to serialize updates of mtime/size.
118	*/
119	aio->res = res;
120	INIT_WORK(&aio->work, backing_aio_complete_work);
121	queue_work(file_inode(aio->orig_iocb->ki_filp)->i_sb->s_dio_done_wq,
122	&aio->work);
123	}
124
125	static int backing_aio_init_wq(struct kiocb *iocb)
126	{
127	struct super_block *sb = file_inode(iocb->ki_filp)->i_sb;
128
129	if (sb->s_dio_done_wq)
130	return 0;
131
132	return sb_init_dio_done_wq(sb);
133	}
134
135
136	ssize_t backing_file_read_iter(struct file file, struct iov_iter iter,
137	struct kiocb *iocb, int flags,
138	struct backing_file_ctx *ctx)
139	{
140	struct backing_aio *aio = NULL;
141	const struct cred *old_cred;
142	ssize_t ret;
143
144	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
145	return -EIO;
146
147	if (!iov_iter_count(iter))
148	return 0;
149
150	if (iocb->ki_flags & IOCB_DIRECT &&
151	!(file->f_mode & FMODE_CAN_ODIRECT))
152	return -EINVAL;
153
154	old_cred = override_creds(ctx->cred);
155	if (is_sync_kiocb(iocb)) {
156	rwf_t rwf = iocb_to_rw_flags(flags);
157
158	ret = vfs_iter_read(file, iter, &iocb->ki_pos, rwf);
159	} else {
160	ret = -ENOMEM;
161	aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
162	if (!aio)
163	goto out;
164
165	aio->orig_iocb = iocb;
166	kiocb_clone(&aio->iocb, iocb, get_file(file));
167	aio->iocb.ki_complete = backing_aio_rw_complete;
168	refcount_set(&aio->ref, 2);
169	ret = vfs_iocb_iter_read(file, &aio->iocb, iter);
170	backing_aio_put(aio);
171	if (ret != -EIOCBQUEUED)
172	backing_aio_cleanup(aio, ret);
173	}
174	out:
175	revert_creds(old_cred);
176
177	if (ctx->accessed)
178	ctx->accessed(ctx->user_file);
179
180	return ret;
181	}
182	EXPORT_SYMBOL_GPL(backing_file_read_iter);
183
184	ssize_t backing_file_write_iter(struct file file, struct iov_iter iter,
185	struct kiocb *iocb, int flags,
186	struct backing_file_ctx *ctx)
187	{
188	const struct cred *old_cred;
189	ssize_t ret;
190
191	if (WARN_ON_ONCE(!(file->f_mode & FMODE_BACKING)))
192	return -EIO;
193
194	if (!iov_iter_count(iter))
195	return 0;
196
197	ret = file_remove_privs(ctx->user_file);
198	if (ret)
199	return ret;
200
201	if (iocb->ki_flags & IOCB_DIRECT &&
202	!(file->f_mode & FMODE_CAN_ODIRECT))
203	return -EINVAL;
204
205	/*
206	* Stacked filesystems don't support deferred completions, don't copy
207	* this property in case it is set by the issuer.
208	*/
209	flags &= ~IOCB_DIO_CALLER_COMP;
210
211	old_cred = override_creds(ctx->cred);
212	if (is_sync_kiocb(iocb)) {
213	rwf_t rwf = iocb_to_rw_flags(flags);
214
215	ret = vfs_iter_write(file, iter, &iocb->ki_pos, rwf);
216	if (ctx->end_write)
217	ctx->end_write(ctx->user_file);
218	} else {
219	struct backing_aio *aio;
220
221	ret = backing_aio_init_wq(iocb);
222	if (ret)
223	goto out;
224
225	ret = -ENOMEM;
226	aio = kmem_cache_zalloc(backing_aio_cachep, GFP_KERNEL);
227	if (!aio)
228	goto out;
229
230	aio->orig_iocb = iocb;
231	aio->end_write = ctx->end_write;
232	kiocb_clone(&aio->iocb, iocb, get_file(file));
233	aio->iocb.ki_flags = flags;
234	aio->iocb.ki_complete = backing_aio_queue_completion;
235	refcount_set(&aio->ref, 2);
236	ret = vfs_iocb_iter_write(file, &aio->iocb, iter);
237	backing_aio_put(aio);
238	if (ret != -EIOCBQUEUED)
239	backing_aio_cleanup(aio, ret);
240	}
241	out:
242	revert_creds(old_cred);
243
244	return ret;
245	}
246	EXPORT_SYMBOL_GPL(backing_file_write_iter);
247
248	static int __init backing_aio_init(void)
249	{
250	backing_aio_cachep = kmem_cache_create("backing_aio",
251	sizeof(struct backing_aio),
252	0, SLAB_HWCACHE_ALIGN, NULL);
253	if (!backing_aio_cachep)
254	return -ENOMEM;
255
256	return 0;
257	}
258	fs_initcall(backing_aio_init);