[linux-block.git] / mm / process_vm_access.c

// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * linux/mm/process_vm_access.c
 *
 * Copyright (C) 2010-2011 Christopher Yeoh <cyeoh@au1.ibm.com>, IBM Corp.
 */

#include <linux/compat.h>
#include <linux/mm.h>
#include <linux/uio.h>
#include <linux/sched.h>
#include <linux/sched/mm.h>
#include <linux/highmem.h>
#include <linux/ptrace.h>
#include <linux/slab.h>
#include <linux/syscalls.h>

/**
 * process_vm_rw_pages - read/write pages from task specified
 * @pages: array of pointers to pages we want to copy
 * @offset: offset in page to start copying from/to
 * @len: number of bytes to copy
 * @iter: where to copy to/from locally
 * @vm_write: 0 means copy from, 1 means copy to
 * Returns 0 on success, error code otherwise
 */
static int process_vm_rw_pages(struct page **pages,
			       unsigned offset,
			       size_t len,
			       struct iov_iter *iter,
			       int vm_write)
{
	/* Do the copy for each page */
	while (len && iov_iter_count(iter)) {
		struct page *page = *pages++;
		size_t copy = PAGE_SIZE - offset;
		size_t copied;

		if (copy > len)
			copy = len;

		if (vm_write)
			copied = copy_page_from_iter(page, offset, copy, iter);
		else
			copied = copy_page_to_iter(page, offset, copy, iter);

		len -= copied;
		if (copied < copy && iov_iter_count(iter))
			return -EFAULT;
		offset = 0;
	}
	return 0;
}

/* Maximum number of pages kmalloc'd to hold struct page's during copy */
#define PVM_MAX_KMALLOC_PAGES (PAGE_SIZE * 2)

/**
 * process_vm_rw_single_vec - read/write pages from task specified
 * @addr: start memory address of target process
 * @len: size of area to copy to/from
 * @iter: where to copy to/from locally
 * @process_pages: struct pages area that can store at least
 *  nr_pages_to_copy struct page pointers
 * @mm: mm for task
 * @task: task to read/write from
 * @vm_write: 0 means copy from, 1 means copy to
 * Returns 0 on success or on failure error code
 */
static int process_vm_rw_single_vec(unsigned long addr,
				    unsigned long len,
				    struct iov_iter *iter,
				    struct page **process_pages,
				    struct mm_struct *mm,
				    struct task_struct *task,
				    int vm_write)
{
	unsigned long pa = addr & PAGE_MASK;
	unsigned long start_offset = addr - pa;
	unsigned long nr_pages;
	ssize_t rc = 0;
	unsigned long max_pages_per_loop = PVM_MAX_KMALLOC_PAGES
		/ sizeof(struct pages *);
	unsigned int flags = 0;

	/* Work out address and page range required */
	if (len == 0)
		return 0;
	nr_pages = (addr + len - 1) / PAGE_SIZE - addr / PAGE_SIZE + 1;

	if (vm_write)
		flags |= FOLL_WRITE;

	while (!rc && nr_pages && iov_iter_count(iter)) {
		int pinned_pages = min(nr_pages, max_pages_per_loop);
		int locked = 1;
		size_t bytes;

		/*
		 * Get the pages we're interested in.  We must
		 * access remotely because task/mm might not
		 * current/current->mm
		 */
		mmap_read_lock(mm);
		pinned_pages = pin_user_pages_remote(mm, pa, pinned_pages,
						     flags, process_pages,
						     NULL, &locked);
		if (locked)
			mmap_read_unlock(mm);
		if (pinned_pages <= 0)
			return -EFAULT;

		bytes = pinned_pages * PAGE_SIZE - start_offset;
		if (bytes > len)
			bytes = len;

		rc = process_vm_rw_pages(process_pages,
					 start_offset, bytes, iter,
					 vm_write);
		len -= bytes;
		start_offset = 0;
		nr_pages -= pinned_pages;
		pa += pinned_pages * PAGE_SIZE;

		/* If vm_write is set, the pages need to be made dirty: */
		unpin_user_pages_dirty_lock(process_pages, pinned_pages,
					    vm_write);
	}

	return rc;
}

/* Maximum number of entries for process pages array
   which lives on stack */
#define PVM_MAX_PP_ARRAY_COUNT 16

/**
 * process_vm_rw_core - core of reading/writing pages from task specified
 * @pid: PID of process to read/write from/to
 * @iter: where to copy to/from locally
 * @rvec: iovec array specifying where to copy to/from in the other process
 * @riovcnt: size of rvec array
 * @flags: currently unused
 * @vm_write: 0 if reading from other process, 1 if writing to other process
 *
 * Returns the number of bytes read/written or error code. May
 *  return less bytes than expected if an error occurs during the copying
 *  process.
 */
static ssize_t process_vm_rw_core(pid_t pid, struct iov_iter *iter,
				  const struct iovec *rvec,
				  unsigned long riovcnt,
				  unsigned long flags, int vm_write)
{
	struct task_struct *task;
	struct page *pp_stack[PVM_MAX_PP_ARRAY_COUNT];
	struct page **process_pages = pp_stack;
	struct mm_struct *mm;
	unsigned long i;
	ssize_t rc = 0;
	unsigned long nr_pages = 0;
	unsigned long nr_pages_iov;
	ssize_t iov_len;
	size_t total_len = iov_iter_count(iter);

	/*
	 * Work out how many pages of struct pages we're going to need
	 * when eventually calling get_user_pages
	 */
	for (i = 0; i < riovcnt; i++) {
		iov_len = rvec[i].iov_len;
		if (iov_len > 0) {
			nr_pages_iov = ((unsigned long)rvec[i].iov_base
					+ iov_len)
				/ PAGE_SIZE - (unsigned long)rvec[i].iov_base
				/ PAGE_SIZE + 1;
			nr_pages = max(nr_pages, nr_pages_iov);
		}
	}

	if (nr_pages == 0)
		return 0;

	if (nr_pages > PVM_MAX_PP_ARRAY_COUNT) {
		/* For reliability don't try to kmalloc more than
		   2 pages worth */
		process_pages = kmalloc(min_t(size_t, PVM_MAX_KMALLOC_PAGES,
					      sizeof(struct pages *)*nr_pages),
					GFP_KERNEL);

		if (!process_pages)
			return -ENOMEM;
	}

	/* Get process information */
	task = find_get_task_by_vpid(pid);
	if (!task) {
		rc = -ESRCH;
		goto free_proc_pages;
	}

	mm = mm_access(task, PTRACE_MODE_ATTACH_REALCREDS);
	if (!mm || IS_ERR(mm)) {
		rc = IS_ERR(mm) ? PTR_ERR(mm) : -ESRCH;
		/*
		 * Explicitly map EACCES to EPERM as EPERM is a more
		 * appropriate error code for process_vw_readv/writev
		 */
		if (rc == -EACCES)
			rc = -EPERM;
		goto put_task_struct;
	}

	for (i = 0; i < riovcnt && iov_iter_count(iter) && !rc; i++)
		rc = process_vm_rw_single_vec(
			(unsigned long)rvec[i].iov_base, rvec[i].iov_len,
			iter, process_pages, mm, task, vm_write);

	/* copied = space before - space after */
	total_len -= iov_iter_count(iter);

	/* If we have managed to copy any data at all then
	   we return the number of bytes copied. Otherwise
	   we return the error code */
	if (total_len)
		rc = total_len;

	mmput(mm);

put_task_struct:
	put_task_struct(task);

free_proc_pages:
	if (process_pages != pp_stack)
		kfree(process_pages);
	return rc;
}

/**
 * process_vm_rw - check iovecs before calling core routine
 * @pid: PID of process to read/write from/to
 * @lvec: iovec array specifying where to copy to/from locally
 * @liovcnt: size of lvec array
 * @rvec: iovec array specifying where to copy to/from in the other process
 * @riovcnt: size of rvec array
 * @flags: currently unused
 * @vm_write: 0 if reading from other process, 1 if writing to other process
 *
 * Returns the number of bytes read/written or error code. May
 *  return less bytes than expected if an error occurs during the copying
 *  process.
 */
static ssize_t process_vm_rw(pid_t pid,
			     const struct iovec __user *lvec,
			     unsigned long liovcnt,
			     const struct iovec __user *rvec,
			     unsigned long riovcnt,
			     unsigned long flags, int vm_write)
{
	struct iovec iovstack_l[UIO_FASTIOV];
	struct iovec iovstack_r[UIO_FASTIOV];
	struct iovec *iov_l = iovstack_l;
	struct iovec *iov_r;
	struct iov_iter iter;
	ssize_t rc;
	int dir = vm_write ? ITER_SOURCE : ITER_DEST;

	if (flags != 0)
		return -EINVAL;

	/* Check iovecs */
	rc = import_iovec(dir, lvec, liovcnt, UIO_FASTIOV, &iov_l, &iter);
	if (rc < 0)
		return rc;
	if (!iov_iter_count(&iter))
		goto free_iov_l;
	iov_r = iovec_from_user(rvec, riovcnt, UIO_FASTIOV, iovstack_r,
				in_compat_syscall());
	if (IS_ERR(iov_r)) {
		rc = PTR_ERR(iov_r);
		goto free_iov_l;
	}
	rc = process_vm_rw_core(pid, &iter, iov_r, riovcnt, flags, vm_write);
	if (iov_r != iovstack_r)
		kfree(iov_r);
free_iov_l:
	kfree(iov_l);
	return rc;
}

SYSCALL_DEFINE6(process_vm_readv, pid_t, pid, const struct iovec __user *, lvec,
		unsigned long, liovcnt, const struct iovec __user *, rvec,
		unsigned long, riovcnt,	unsigned long, flags)
{
	return process_vm_rw(pid, lvec, liovcnt, rvec, riovcnt, flags, 0);
}

SYSCALL_DEFINE6(process_vm_writev, pid_t, pid,
		const struct iovec __user *, lvec,
		unsigned long, liovcnt, const struct iovec __user *, rvec,
		unsigned long, riovcnt,	unsigned long, flags)
{
	return process_vm_rw(pid, lvec, liovcnt, rvec, riovcnt, flags, 1);
}
Commit	Line	Data
2874c5fd	1	// SPDX-License-Identifier: GPL-2.0-or-later
fcf63409 CY	2	/*
	3	* linux/mm/process_vm_access.c
	4	*
	5	* Copyright (C) 2010-2011 Christopher Yeoh <cyeoh@au1.ibm.com>, IBM Corp.
fcf63409 CY	6	*/
fcf63409 CY	7
f78f63da	8	#include <linux/compat.h>
fcf63409 CY	9	#include <linux/mm.h>
	10	#include <linux/uio.h>
	11	#include <linux/sched.h>
6e84f315	12	#include <linux/sched/mm.h>
fcf63409 CY	13	#include <linux/highmem.h>
	14	#include <linux/ptrace.h>
	15	#include <linux/slab.h>
	16	#include <linux/syscalls.h>
	17
fcf63409 CY	18	/**
fcf63409 CY	19	* process_vm_rw_pages - read/write pages from task specified
4bafbec7	20	* @pages: array of pointers to pages we want to copy
f144c390	21	* @offset: offset in page to start copying from/to
fcf63409	22	* @len: number of bytes to copy
4bafbec7	23	* @iter: where to copy to/from locally
fcf63409	24	* @vm_write: 0 means copy from, 1 means copy to
fcf63409 CY	25	* Returns 0 on success, error code otherwise
fcf63409 CY	26	*/
70eca12d AV	27	static int process_vm_rw_pages(struct page **pages,
70eca12d AV	28	unsigned offset,
e21345f9	29	size_t len,
9f78bdfa	30	struct iov_iter *iter,
9acc1a0f	31	int vm_write)
fcf63409	32	{
fcf63409	33	/* Do the copy for each page */
9acc1a0f	34	while (len && iov_iter_count(iter)) {
70eca12d	35	struct page page = pages++;
e21345f9	36	size_t copy = PAGE_SIZE - offset;
70eca12d	37	size_t copied;
240f3905	38
e21345f9 AV	39	if (copy > len)
	40	copy = len;
	41
803e4572	42	if (vm_write)
e7c24607	43	copied = copy_page_from_iter(page, offset, copy, iter);
803e4572	44	else
70eca12d	45	copied = copy_page_to_iter(page, offset, copy, iter);
803e4572	46
70eca12d AV	47	len -= copied;
	48	if (copied < copy && iov_iter_count(iter))
	49	return -EFAULT;
	50	offset = 0;
fcf63409	51	}
70eca12d	52	return 0;
fcf63409 CY	53	}
	54
	55	/* Maximum number of pages kmalloc'd to hold struct page's during copy */
	56	#define PVM_MAX_KMALLOC_PAGES (PAGE_SIZE * 2)
	57
	58	/**
	59	* process_vm_rw_single_vec - read/write pages from task specified
	60	* @addr: start memory address of target process
	61	* @len: size of area to copy to/from
4bafbec7	62	* @iter: where to copy to/from locally
fcf63409 CY	63	* @process_pages: struct pages area that can store at least
	64	* nr_pages_to_copy struct page pointers
	65	* @mm: mm for task
	66	* @task: task to read/write from
	67	* @vm_write: 0 means copy from, 1 means copy to
fcf63409 CY	68	* Returns 0 on success or on failure error code
	69	*/
	70	static int process_vm_rw_single_vec(unsigned long addr,
	71	unsigned long len,
9f78bdfa	72	struct iov_iter *iter,
fcf63409 CY	73	struct page **process_pages,
	74	struct mm_struct *mm,
	75	struct task_struct *task,
9acc1a0f	76	int vm_write)
fcf63409 CY	77	{
	78	unsigned long pa = addr & PAGE_MASK;
	79	unsigned long start_offset = addr - pa;
	80	unsigned long nr_pages;
fcf63409	81	ssize_t rc = 0;
fcf63409 CY	82	unsigned long max_pages_per_loop = PVM_MAX_KMALLOC_PAGES
fcf63409 CY	83	/ sizeof(struct pages *);
8b7457ef	84	unsigned int flags = 0;
fcf63409	85
fcf63409 CY	86	/* Work out address and page range required */
	87	if (len == 0)
	88	return 0;
	89	nr_pages = (addr + len - 1) / PAGE_SIZE - addr / PAGE_SIZE + 1;
	90
d4944b0e LS	91	if (vm_write)
	92	flags \|= FOLL_WRITE;
	93
4bafbec7	94	while (!rc && nr_pages && iov_iter_count(iter)) {
803e4572	95	int pinned_pages = min(nr_pages, max_pages_per_loop);
8b7457ef	96	int locked = 1;
4bafbec7	97	size_t bytes;
fcf63409	98
1e987790 DH	99	/*
1e987790 DH	100	* Get the pages we're interested in. We must
8b7457ef	101	* access remotely because task/mm might not
1e987790 DH	102	* current/current->mm
1e987790 DH	103	*/
d8ed45c5	104	mmap_read_lock(mm);
64019a2e	105	pinned_pages = pin_user_pages_remote(mm, pa, pinned_pages,
803e4572 JH	106	flags, process_pages,
803e4572 JH	107	NULL, &locked);
8b7457ef	108	if (locked)
d8ed45c5	109	mmap_read_unlock(mm);
803e4572	110	if (pinned_pages <= 0)
70eca12d AV	111	return -EFAULT;
70eca12d AV	112
803e4572	113	bytes = pinned_pages * PAGE_SIZE - start_offset;
4bafbec7 AV	114	if (bytes > len)
4bafbec7 AV	115	bytes = len;
e21345f9	116
70eca12d	117	rc = process_vm_rw_pages(process_pages,
4bafbec7	118	start_offset, bytes, iter,
9acc1a0f	119	vm_write);
4bafbec7	120	len -= bytes;
fcf63409	121	start_offset = 0;
803e4572 JH	122	nr_pages -= pinned_pages;
	123	pa += pinned_pages * PAGE_SIZE;
	124
	125	/* If vm_write is set, the pages need to be made dirty: */
f1f6a7dd JH	126	unpin_user_pages_dirty_lock(process_pages, pinned_pages,
f1f6a7dd JH	127	vm_write);
fcf63409 CY	128	}
	129
	130	return rc;
	131	}
	132
	133	/* Maximum number of entries for process pages array
	134	which lives on stack */
	135	#define PVM_MAX_PP_ARRAY_COUNT 16
	136
	137	/**
	138	* process_vm_rw_core - core of reading/writing pages from task specified
	139	* @pid: PID of process to read/write from/to
4bafbec7	140	* @iter: where to copy to/from locally
fcf63409 CY	141	* @rvec: iovec array specifying where to copy to/from in the other process
	142	* @riovcnt: size of rvec array
	143	* @flags: currently unused
	144	* @vm_write: 0 if reading from other process, 1 if writing to other process
a5d09bed	145	*
fcf63409 CY	146	* Returns the number of bytes read/written or error code. May
	147	* return less bytes than expected if an error occurs during the copying
	148	* process.
	149	*/
9f78bdfa	150	static ssize_t process_vm_rw_core(pid_t pid, struct iov_iter *iter,
fcf63409 CY	151	const struct iovec *rvec,
	152	unsigned long riovcnt,
	153	unsigned long flags, int vm_write)
	154	{
	155	struct task_struct *task;
	156	struct page *pp_stack[PVM_MAX_PP_ARRAY_COUNT];
	157	struct page **process_pages = pp_stack;
	158	struct mm_struct *mm;
	159	unsigned long i;
	160	ssize_t rc = 0;
fcf63409 CY	161	unsigned long nr_pages = 0;
fcf63409 CY	162	unsigned long nr_pages_iov;
fcf63409	163	ssize_t iov_len;
9acc1a0f	164	size_t total_len = iov_iter_count(iter);
fcf63409 CY	165
	166	/*
	167	* Work out how many pages of struct pages we're going to need
	168	* when eventually calling get_user_pages
	169	*/
	170	for (i = 0; i < riovcnt; i++) {
	171	iov_len = rvec[i].iov_len;
	172	if (iov_len > 0) {
	173	nr_pages_iov = ((unsigned long)rvec[i].iov_base
	174	+ iov_len)
	175	/ PAGE_SIZE - (unsigned long)rvec[i].iov_base
	176	/ PAGE_SIZE + 1;
	177	nr_pages = max(nr_pages, nr_pages_iov);
	178	}
	179	}
	180
	181	if (nr_pages == 0)
	182	return 0;
	183
	184	if (nr_pages > PVM_MAX_PP_ARRAY_COUNT) {
	185	/* For reliability don't try to kmalloc more than
	186	2 pages worth */
	187	process_pages = kmalloc(min_t(size_t, PVM_MAX_KMALLOC_PAGES,
	188	sizeof(struct pages )nr_pages),
	189	GFP_KERNEL);
	190
	191	if (!process_pages)
	192	return -ENOMEM;
	193	}
	194
	195	/* Get process information */
2ee08260	196	task = find_get_task_by_vpid(pid);
fcf63409 CY	197	if (!task) {
	198	rc = -ESRCH;
	199	goto free_proc_pages;
	200	}
	201
caaee623	202	mm = mm_access(task, PTRACE_MODE_ATTACH_REALCREDS);
8cdb878d CY	203	if (!mm \|\| IS_ERR(mm)) {
	204	rc = IS_ERR(mm) ? PTR_ERR(mm) : -ESRCH;
	205	/*
8d09db80	206	* Explicitly map EACCES to EPERM as EPERM is a more
8cdb878d CY	207	* appropriate error code for process_vw_readv/writev
	208	*/
	209	if (rc == -EACCES)
	210	rc = -EPERM;
fcf63409 CY	211	goto put_task_struct;
	212	}
	213
9acc1a0f	214	for (i = 0; i < riovcnt && iov_iter_count(iter) && !rc; i++)
fcf63409 CY	215	rc = process_vm_rw_single_vec(
fcf63409 CY	216	(unsigned long)rvec[i].iov_base, rvec[i].iov_len,
9acc1a0f AV	217	iter, process_pages, mm, task, vm_write);
	218
	219	/* copied = space before - space after */
	220	total_len -= iov_iter_count(iter);
	221
	222	/* If we have managed to copy any data at all then
	223	we return the number of bytes copied. Otherwise
	224	we return the error code */
	225	if (total_len)
	226	rc = total_len;
fcf63409	227
fcf63409 CY	228	mmput(mm);
	229
	230	put_task_struct:
	231	put_task_struct(task);
	232
	233	free_proc_pages:
	234	if (process_pages != pp_stack)
	235	kfree(process_pages);
	236	return rc;
	237	}
	238
	239	/**
	240	* process_vm_rw - check iovecs before calling core routine
	241	* @pid: PID of process to read/write from/to
	242	* @lvec: iovec array specifying where to copy to/from locally
	243	* @liovcnt: size of lvec array
	244	* @rvec: iovec array specifying where to copy to/from in the other process
	245	* @riovcnt: size of rvec array
	246	* @flags: currently unused
	247	* @vm_write: 0 if reading from other process, 1 if writing to other process
a5d09bed	248	*
fcf63409 CY	249	* Returns the number of bytes read/written or error code. May
	250	* return less bytes than expected if an error occurs during the copying
	251	* process.
	252	*/
	253	static ssize_t process_vm_rw(pid_t pid,
	254	const struct iovec __user *lvec,
	255	unsigned long liovcnt,
	256	const struct iovec __user *rvec,
	257	unsigned long riovcnt,
	258	unsigned long flags, int vm_write)
	259	{
	260	struct iovec iovstack_l[UIO_FASTIOV];
	261	struct iovec iovstack_r[UIO_FASTIOV];
	262	struct iovec *iov_l = iovstack_l;
95c9ae14	263	struct iovec *iov_r;
9f78bdfa	264	struct iov_iter iter;
fcf63409	265	ssize_t rc;
de4eda9d	266	int dir = vm_write ? ITER_SOURCE : ITER_DEST;
fcf63409 CY	267
	268	if (flags != 0)
	269	return -EINVAL;
	270
	271	/* Check iovecs */
17d17e72 AV	272	rc = import_iovec(dir, lvec, liovcnt, UIO_FASTIOV, &iov_l, &iter);
	273	if (rc < 0)
	274	return rc;
	275	if (!iov_iter_count(&iter))
bfdc5970	276	goto free_iov_l;
38dc5079 JA	277	iov_r = iovec_from_user(rvec, riovcnt, UIO_FASTIOV, iovstack_r,
38dc5079 JA	278	in_compat_syscall());
bfdc5970 CH	279	if (IS_ERR(iov_r)) {
	280	rc = PTR_ERR(iov_r);
	281	goto free_iov_l;
	282	}
9f78bdfa	283	rc = process_vm_rw_core(pid, &iter, iov_r, riovcnt, flags, vm_write);
fcf63409 CY	284	if (iov_r != iovstack_r)
fcf63409 CY	285	kfree(iov_r);
bfdc5970	286	free_iov_l:
17d17e72	287	kfree(iov_l);
fcf63409 CY	288	return rc;
	289	}
	290
	291	SYSCALL_DEFINE6(process_vm_readv, pid_t, pid, const struct iovec __user *, lvec,
	292	unsigned long, liovcnt, const struct iovec __user *, rvec,
	293	unsigned long, riovcnt, unsigned long, flags)
	294	{
	295	return process_vm_rw(pid, lvec, liovcnt, rvec, riovcnt, flags, 0);
	296	}
	297
	298	SYSCALL_DEFINE6(process_vm_writev, pid_t, pid,
	299	const struct iovec __user *, lvec,
	300	unsigned long, liovcnt, const struct iovec __user *, rvec,
	301	unsigned long, riovcnt, unsigned long, flags)
	302	{
	303	return process_vm_rw(pid, lvec, liovcnt, rvec, riovcnt, flags, 1);
	304	}