xref: /linux/tools/include/uapi/linux/fs.h (revision a1ff5a7d78a036d6c2178ee5acd6ba4946243800)
1e2019472SJohn Hubbard /* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2*77179b6fSAndrii Nakryiko #ifndef _UAPI_LINUX_FS_H
3*77179b6fSAndrii Nakryiko #define _UAPI_LINUX_FS_H
4e2019472SJohn Hubbard 
5e2019472SJohn Hubbard /*
6e2019472SJohn Hubbard  * This file has definitions for some important file table structures
7e2019472SJohn Hubbard  * and constants and structures used by various generic file system
8e2019472SJohn Hubbard  * ioctl's.  Please do not make any changes in this file before
9e2019472SJohn Hubbard  * sending patches for review to linux-fsdevel@vger.kernel.org and
10e2019472SJohn Hubbard  * linux-api@vger.kernel.org.
11e2019472SJohn Hubbard  */
12e2019472SJohn Hubbard 
13e2019472SJohn Hubbard #include <linux/limits.h>
14e2019472SJohn Hubbard #include <linux/ioctl.h>
15e2019472SJohn Hubbard #include <linux/types.h>
16*77179b6fSAndrii Nakryiko #ifndef __KERNEL__
17e2019472SJohn Hubbard #include <linux/fscrypt.h>
18*77179b6fSAndrii Nakryiko #endif
19e2019472SJohn Hubbard 
20e2019472SJohn Hubbard /* Use of MS_* flags within the kernel is restricted to core mount(2) code. */
21*77179b6fSAndrii Nakryiko #if !defined(__KERNEL__)
22e2019472SJohn Hubbard #include <linux/mount.h>
23*77179b6fSAndrii Nakryiko #endif
24e2019472SJohn Hubbard 
25e2019472SJohn Hubbard /*
26e2019472SJohn Hubbard  * It's silly to have NR_OPEN bigger than NR_FILE, but you can change
27e2019472SJohn Hubbard  * the file limit at runtime and only root can increase the per-process
28e2019472SJohn Hubbard  * nr_file rlimit, so it's safe to set up a ridiculously high absolute
29e2019472SJohn Hubbard  * upper limit on files-per-process.
30e2019472SJohn Hubbard  *
31e2019472SJohn Hubbard  * Some programs (notably those using select()) may have to be
32e2019472SJohn Hubbard  * recompiled to take full advantage of the new limits..
33e2019472SJohn Hubbard  */
34e2019472SJohn Hubbard 
35e2019472SJohn Hubbard /* Fixed constants first: */
36e2019472SJohn Hubbard #undef NR_OPEN
37e2019472SJohn Hubbard #define INR_OPEN_CUR 1024	/* Initial setting for nfile rlimits */
38e2019472SJohn Hubbard #define INR_OPEN_MAX 4096	/* Hard limit for nfile rlimits */
39e2019472SJohn Hubbard 
40e2019472SJohn Hubbard #define BLOCK_SIZE_BITS 10
41e2019472SJohn Hubbard #define BLOCK_SIZE (1<<BLOCK_SIZE_BITS)
42e2019472SJohn Hubbard 
43e2019472SJohn Hubbard #define SEEK_SET	0	/* seek relative to beginning of file */
44e2019472SJohn Hubbard #define SEEK_CUR	1	/* seek relative to current file position */
45e2019472SJohn Hubbard #define SEEK_END	2	/* seek relative to end of file */
46e2019472SJohn Hubbard #define SEEK_DATA	3	/* seek to the next data */
47e2019472SJohn Hubbard #define SEEK_HOLE	4	/* seek to the next hole */
48e2019472SJohn Hubbard #define SEEK_MAX	SEEK_HOLE
49e2019472SJohn Hubbard 
50e2019472SJohn Hubbard #define RENAME_NOREPLACE	(1 << 0)	/* Don't overwrite target */
51e2019472SJohn Hubbard #define RENAME_EXCHANGE		(1 << 1)	/* Exchange source and dest */
52e2019472SJohn Hubbard #define RENAME_WHITEOUT		(1 << 2)	/* Whiteout source */
53e2019472SJohn Hubbard 
54e2019472SJohn Hubbard struct file_clone_range {
55e2019472SJohn Hubbard 	__s64 src_fd;
56e2019472SJohn Hubbard 	__u64 src_offset;
57e2019472SJohn Hubbard 	__u64 src_length;
58e2019472SJohn Hubbard 	__u64 dest_offset;
59e2019472SJohn Hubbard };
60e2019472SJohn Hubbard 
61e2019472SJohn Hubbard struct fstrim_range {
62e2019472SJohn Hubbard 	__u64 start;
63e2019472SJohn Hubbard 	__u64 len;
64e2019472SJohn Hubbard 	__u64 minlen;
65e2019472SJohn Hubbard };
66e2019472SJohn Hubbard 
67e2019472SJohn Hubbard /*
68e2019472SJohn Hubbard  * We include a length field because some filesystems (vfat) have an identifier
69e2019472SJohn Hubbard  * that we do want to expose as a UUID, but doesn't have the standard length.
70e2019472SJohn Hubbard  *
71e2019472SJohn Hubbard  * We use a fixed size buffer beacuse this interface will, by fiat, never
72e2019472SJohn Hubbard  * support "UUIDs" longer than 16 bytes; we don't want to force all downstream
73e2019472SJohn Hubbard  * users to have to deal with that.
74e2019472SJohn Hubbard  */
75e2019472SJohn Hubbard struct fsuuid2 {
76e2019472SJohn Hubbard 	__u8	len;
77e2019472SJohn Hubbard 	__u8	uuid[16];
78e2019472SJohn Hubbard };
79e2019472SJohn Hubbard 
80e2019472SJohn Hubbard struct fs_sysfs_path {
81e2019472SJohn Hubbard 	__u8			len;
82e2019472SJohn Hubbard 	__u8			name[128];
83e2019472SJohn Hubbard };
84e2019472SJohn Hubbard 
85e2019472SJohn Hubbard /* extent-same (dedupe) ioctls; these MUST match the btrfs ioctl definitions */
86e2019472SJohn Hubbard #define FILE_DEDUPE_RANGE_SAME		0
87e2019472SJohn Hubbard #define FILE_DEDUPE_RANGE_DIFFERS	1
88e2019472SJohn Hubbard 
89e2019472SJohn Hubbard /* from struct btrfs_ioctl_file_extent_same_info */
90e2019472SJohn Hubbard struct file_dedupe_range_info {
91e2019472SJohn Hubbard 	__s64 dest_fd;		/* in - destination file */
92e2019472SJohn Hubbard 	__u64 dest_offset;	/* in - start of extent in destination */
93e2019472SJohn Hubbard 	__u64 bytes_deduped;	/* out - total # of bytes we were able
94e2019472SJohn Hubbard 				 * to dedupe from this file. */
95e2019472SJohn Hubbard 	/* status of this dedupe operation:
96e2019472SJohn Hubbard 	 * < 0 for error
97e2019472SJohn Hubbard 	 * == FILE_DEDUPE_RANGE_SAME if dedupe succeeds
98e2019472SJohn Hubbard 	 * == FILE_DEDUPE_RANGE_DIFFERS if data differs
99e2019472SJohn Hubbard 	 */
100e2019472SJohn Hubbard 	__s32 status;		/* out - see above description */
101e2019472SJohn Hubbard 	__u32 reserved;		/* must be zero */
102e2019472SJohn Hubbard };
103e2019472SJohn Hubbard 
104e2019472SJohn Hubbard /* from struct btrfs_ioctl_file_extent_same_args */
105e2019472SJohn Hubbard struct file_dedupe_range {
106e2019472SJohn Hubbard 	__u64 src_offset;	/* in - start of extent in source */
107e2019472SJohn Hubbard 	__u64 src_length;	/* in - length of extent */
108e2019472SJohn Hubbard 	__u16 dest_count;	/* in - total elements in info array */
109e2019472SJohn Hubbard 	__u16 reserved1;	/* must be zero */
110e2019472SJohn Hubbard 	__u32 reserved2;	/* must be zero */
111e2019472SJohn Hubbard 	struct file_dedupe_range_info info[];
112e2019472SJohn Hubbard };
113e2019472SJohn Hubbard 
114e2019472SJohn Hubbard /* And dynamically-tunable limits and defaults: */
115e2019472SJohn Hubbard struct files_stat_struct {
116e2019472SJohn Hubbard 	unsigned long nr_files;		/* read only */
117e2019472SJohn Hubbard 	unsigned long nr_free_files;	/* read only */
118e2019472SJohn Hubbard 	unsigned long max_files;		/* tunable */
119e2019472SJohn Hubbard };
120e2019472SJohn Hubbard 
121e2019472SJohn Hubbard struct inodes_stat_t {
122e2019472SJohn Hubbard 	long nr_inodes;
123e2019472SJohn Hubbard 	long nr_unused;
124e2019472SJohn Hubbard 	long dummy[5];		/* padding for sysctl ABI compatibility */
125e2019472SJohn Hubbard };
126e2019472SJohn Hubbard 
127e2019472SJohn Hubbard 
128e2019472SJohn Hubbard #define NR_FILE  8192	/* this can well be larger on a larger system */
129e2019472SJohn Hubbard 
130e2019472SJohn Hubbard /*
131e2019472SJohn Hubbard  * Structure for FS_IOC_FSGETXATTR[A] and FS_IOC_FSSETXATTR.
132e2019472SJohn Hubbard  */
133e2019472SJohn Hubbard struct fsxattr {
134e2019472SJohn Hubbard 	__u32		fsx_xflags;	/* xflags field value (get/set) */
135e2019472SJohn Hubbard 	__u32		fsx_extsize;	/* extsize field value (get/set)*/
136e2019472SJohn Hubbard 	__u32		fsx_nextents;	/* nextents field value (get)	*/
137e2019472SJohn Hubbard 	__u32		fsx_projid;	/* project identifier (get/set) */
138e2019472SJohn Hubbard 	__u32		fsx_cowextsize;	/* CoW extsize field value (get/set)*/
139e2019472SJohn Hubbard 	unsigned char	fsx_pad[8];
140e2019472SJohn Hubbard };
141e2019472SJohn Hubbard 
142e2019472SJohn Hubbard /*
143e2019472SJohn Hubbard  * Flags for the fsx_xflags field
144e2019472SJohn Hubbard  */
145e2019472SJohn Hubbard #define FS_XFLAG_REALTIME	0x00000001	/* data in realtime volume */
146e2019472SJohn Hubbard #define FS_XFLAG_PREALLOC	0x00000002	/* preallocated file extents */
147e2019472SJohn Hubbard #define FS_XFLAG_IMMUTABLE	0x00000008	/* file cannot be modified */
148e2019472SJohn Hubbard #define FS_XFLAG_APPEND		0x00000010	/* all writes append */
149e2019472SJohn Hubbard #define FS_XFLAG_SYNC		0x00000020	/* all writes synchronous */
150e2019472SJohn Hubbard #define FS_XFLAG_NOATIME	0x00000040	/* do not update access time */
151e2019472SJohn Hubbard #define FS_XFLAG_NODUMP		0x00000080	/* do not include in backups */
152e2019472SJohn Hubbard #define FS_XFLAG_RTINHERIT	0x00000100	/* create with rt bit set */
153e2019472SJohn Hubbard #define FS_XFLAG_PROJINHERIT	0x00000200	/* create with parents projid */
154e2019472SJohn Hubbard #define FS_XFLAG_NOSYMLINKS	0x00000400	/* disallow symlink creation */
155e2019472SJohn Hubbard #define FS_XFLAG_EXTSIZE	0x00000800	/* extent size allocator hint */
156e2019472SJohn Hubbard #define FS_XFLAG_EXTSZINHERIT	0x00001000	/* inherit inode extent size */
157e2019472SJohn Hubbard #define FS_XFLAG_NODEFRAG	0x00002000	/* do not defragment */
158e2019472SJohn Hubbard #define FS_XFLAG_FILESTREAM	0x00004000	/* use filestream allocator */
159e2019472SJohn Hubbard #define FS_XFLAG_DAX		0x00008000	/* use DAX for IO */
160e2019472SJohn Hubbard #define FS_XFLAG_COWEXTSIZE	0x00010000	/* CoW extent size allocator hint */
161e2019472SJohn Hubbard #define FS_XFLAG_HASATTR	0x80000000	/* no DIFLAG for this	*/
162e2019472SJohn Hubbard 
163e2019472SJohn Hubbard /* the read-only stuff doesn't really belong here, but any other place is
164e2019472SJohn Hubbard    probably as bad and I don't want to create yet another include file. */
165e2019472SJohn Hubbard 
166e2019472SJohn Hubbard #define BLKROSET   _IO(0x12,93)	/* set device read-only (0 = read-write) */
167e2019472SJohn Hubbard #define BLKROGET   _IO(0x12,94)	/* get read-only status (0 = read_write) */
168e2019472SJohn Hubbard #define BLKRRPART  _IO(0x12,95)	/* re-read partition table */
169e2019472SJohn Hubbard #define BLKGETSIZE _IO(0x12,96)	/* return device size /512 (long *arg) */
170e2019472SJohn Hubbard #define BLKFLSBUF  _IO(0x12,97)	/* flush buffer cache */
171e2019472SJohn Hubbard #define BLKRASET   _IO(0x12,98)	/* set read ahead for block device */
172e2019472SJohn Hubbard #define BLKRAGET   _IO(0x12,99)	/* get current read ahead setting */
173e2019472SJohn Hubbard #define BLKFRASET  _IO(0x12,100)/* set filesystem (mm/filemap.c) read-ahead */
174e2019472SJohn Hubbard #define BLKFRAGET  _IO(0x12,101)/* get filesystem (mm/filemap.c) read-ahead */
175e2019472SJohn Hubbard #define BLKSECTSET _IO(0x12,102)/* set max sectors per request (ll_rw_blk.c) */
176e2019472SJohn Hubbard #define BLKSECTGET _IO(0x12,103)/* get max sectors per request (ll_rw_blk.c) */
177e2019472SJohn Hubbard #define BLKSSZGET  _IO(0x12,104)/* get block device sector size */
178e2019472SJohn Hubbard #if 0
179e2019472SJohn Hubbard #define BLKPG      _IO(0x12,105)/* See blkpg.h */
180e2019472SJohn Hubbard 
181e2019472SJohn Hubbard /* Some people are morons.  Do not use sizeof! */
182e2019472SJohn Hubbard 
183e2019472SJohn Hubbard #define BLKELVGET  _IOR(0x12,106,size_t)/* elevator get */
184e2019472SJohn Hubbard #define BLKELVSET  _IOW(0x12,107,size_t)/* elevator set */
185e2019472SJohn Hubbard /* This was here just to show that the number is taken -
186e2019472SJohn Hubbard    probably all these _IO(0x12,*) ioctls should be moved to blkpg.h. */
187e2019472SJohn Hubbard #endif
188e2019472SJohn Hubbard /* A jump here: 108-111 have been used for various private purposes. */
189e2019472SJohn Hubbard #define BLKBSZGET  _IOR(0x12,112,size_t)
190e2019472SJohn Hubbard #define BLKBSZSET  _IOW(0x12,113,size_t)
191e2019472SJohn Hubbard #define BLKGETSIZE64 _IOR(0x12,114,size_t)	/* return device size in bytes (u64 *arg) */
192e2019472SJohn Hubbard #define BLKTRACESETUP _IOWR(0x12,115,struct blk_user_trace_setup)
193e2019472SJohn Hubbard #define BLKTRACESTART _IO(0x12,116)
194e2019472SJohn Hubbard #define BLKTRACESTOP _IO(0x12,117)
195e2019472SJohn Hubbard #define BLKTRACETEARDOWN _IO(0x12,118)
196e2019472SJohn Hubbard #define BLKDISCARD _IO(0x12,119)
197e2019472SJohn Hubbard #define BLKIOMIN _IO(0x12,120)
198e2019472SJohn Hubbard #define BLKIOOPT _IO(0x12,121)
199e2019472SJohn Hubbard #define BLKALIGNOFF _IO(0x12,122)
200e2019472SJohn Hubbard #define BLKPBSZGET _IO(0x12,123)
201e2019472SJohn Hubbard #define BLKDISCARDZEROES _IO(0x12,124)
202e2019472SJohn Hubbard #define BLKSECDISCARD _IO(0x12,125)
203e2019472SJohn Hubbard #define BLKROTATIONAL _IO(0x12,126)
204e2019472SJohn Hubbard #define BLKZEROOUT _IO(0x12,127)
205e2019472SJohn Hubbard #define BLKGETDISKSEQ _IOR(0x12,128,__u64)
206e2019472SJohn Hubbard /*
207e2019472SJohn Hubbard  * A jump here: 130-136 are reserved for zoned block devices
208e2019472SJohn Hubbard  * (see uapi/linux/blkzoned.h)
209e2019472SJohn Hubbard  */
210e2019472SJohn Hubbard 
211e2019472SJohn Hubbard #define BMAP_IOCTL 1		/* obsolete - kept for compatibility */
212e2019472SJohn Hubbard #define FIBMAP	   _IO(0x00,1)	/* bmap access */
213e2019472SJohn Hubbard #define FIGETBSZ   _IO(0x00,2)	/* get the block size used for bmap */
214e2019472SJohn Hubbard #define FIFREEZE	_IOWR('X', 119, int)	/* Freeze */
215e2019472SJohn Hubbard #define FITHAW		_IOWR('X', 120, int)	/* Thaw */
216e2019472SJohn Hubbard #define FITRIM		_IOWR('X', 121, struct fstrim_range)	/* Trim */
217e2019472SJohn Hubbard #define FICLONE		_IOW(0x94, 9, int)
218e2019472SJohn Hubbard #define FICLONERANGE	_IOW(0x94, 13, struct file_clone_range)
219e2019472SJohn Hubbard #define FIDEDUPERANGE	_IOWR(0x94, 54, struct file_dedupe_range)
220e2019472SJohn Hubbard 
221e2019472SJohn Hubbard #define FSLABEL_MAX 256	/* Max chars for the interface; each fs may differ */
222e2019472SJohn Hubbard 
223e2019472SJohn Hubbard #define	FS_IOC_GETFLAGS			_IOR('f', 1, long)
224e2019472SJohn Hubbard #define	FS_IOC_SETFLAGS			_IOW('f', 2, long)
225e2019472SJohn Hubbard #define	FS_IOC_GETVERSION		_IOR('v', 1, long)
226e2019472SJohn Hubbard #define	FS_IOC_SETVERSION		_IOW('v', 2, long)
227e2019472SJohn Hubbard #define FS_IOC_FIEMAP			_IOWR('f', 11, struct fiemap)
228e2019472SJohn Hubbard #define FS_IOC32_GETFLAGS		_IOR('f', 1, int)
229e2019472SJohn Hubbard #define FS_IOC32_SETFLAGS		_IOW('f', 2, int)
230e2019472SJohn Hubbard #define FS_IOC32_GETVERSION		_IOR('v', 1, int)
231e2019472SJohn Hubbard #define FS_IOC32_SETVERSION		_IOW('v', 2, int)
232e2019472SJohn Hubbard #define FS_IOC_FSGETXATTR		_IOR('X', 31, struct fsxattr)
233e2019472SJohn Hubbard #define FS_IOC_FSSETXATTR		_IOW('X', 32, struct fsxattr)
234e2019472SJohn Hubbard #define FS_IOC_GETFSLABEL		_IOR(0x94, 49, char[FSLABEL_MAX])
235e2019472SJohn Hubbard #define FS_IOC_SETFSLABEL		_IOW(0x94, 50, char[FSLABEL_MAX])
236e2019472SJohn Hubbard /* Returns the external filesystem UUID, the same one blkid returns */
237e2019472SJohn Hubbard #define FS_IOC_GETFSUUID		_IOR(0x15, 0, struct fsuuid2)
238e2019472SJohn Hubbard /*
239e2019472SJohn Hubbard  * Returns the path component under /sys/fs/ that refers to this filesystem;
240e2019472SJohn Hubbard  * also /sys/kernel/debug/ for filesystems with debugfs exports
241e2019472SJohn Hubbard  */
242e2019472SJohn Hubbard #define FS_IOC_GETFSSYSFSPATH		_IOR(0x15, 1, struct fs_sysfs_path)
243e2019472SJohn Hubbard 
244e2019472SJohn Hubbard /*
245e2019472SJohn Hubbard  * Inode flags (FS_IOC_GETFLAGS / FS_IOC_SETFLAGS)
246e2019472SJohn Hubbard  *
247e2019472SJohn Hubbard  * Note: for historical reasons, these flags were originally used and
248e2019472SJohn Hubbard  * defined for use by ext2/ext3, and then other file systems started
249e2019472SJohn Hubbard  * using these flags so they wouldn't need to write their own version
250e2019472SJohn Hubbard  * of chattr/lsattr (which was shipped as part of e2fsprogs).  You
251e2019472SJohn Hubbard  * should think twice before trying to use these flags in new
252e2019472SJohn Hubbard  * contexts, or trying to assign these flags, since they are used both
253e2019472SJohn Hubbard  * as the UAPI and the on-disk encoding for ext2/3/4.  Also, we are
254e2019472SJohn Hubbard  * almost out of 32-bit flags.  :-)
255e2019472SJohn Hubbard  *
256e2019472SJohn Hubbard  * We have recently hoisted FS_IOC_FSGETXATTR / FS_IOC_FSSETXATTR from
257e2019472SJohn Hubbard  * XFS to the generic FS level interface.  This uses a structure that
258e2019472SJohn Hubbard  * has padding and hence has more room to grow, so it may be more
259e2019472SJohn Hubbard  * appropriate for many new use cases.
260e2019472SJohn Hubbard  *
261e2019472SJohn Hubbard  * Please do not change these flags or interfaces before checking with
262e2019472SJohn Hubbard  * linux-fsdevel@vger.kernel.org and linux-api@vger.kernel.org.
263e2019472SJohn Hubbard  */
264e2019472SJohn Hubbard #define	FS_SECRM_FL			0x00000001 /* Secure deletion */
265e2019472SJohn Hubbard #define	FS_UNRM_FL			0x00000002 /* Undelete */
266e2019472SJohn Hubbard #define	FS_COMPR_FL			0x00000004 /* Compress file */
267e2019472SJohn Hubbard #define FS_SYNC_FL			0x00000008 /* Synchronous updates */
268e2019472SJohn Hubbard #define FS_IMMUTABLE_FL			0x00000010 /* Immutable file */
269e2019472SJohn Hubbard #define FS_APPEND_FL			0x00000020 /* writes to file may only append */
270e2019472SJohn Hubbard #define FS_NODUMP_FL			0x00000040 /* do not dump file */
271e2019472SJohn Hubbard #define FS_NOATIME_FL			0x00000080 /* do not update atime */
272e2019472SJohn Hubbard /* Reserved for compression usage... */
273e2019472SJohn Hubbard #define FS_DIRTY_FL			0x00000100
274e2019472SJohn Hubbard #define FS_COMPRBLK_FL			0x00000200 /* One or more compressed clusters */
275e2019472SJohn Hubbard #define FS_NOCOMP_FL			0x00000400 /* Don't compress */
276e2019472SJohn Hubbard /* End compression flags --- maybe not all used */
277e2019472SJohn Hubbard #define FS_ENCRYPT_FL			0x00000800 /* Encrypted file */
278e2019472SJohn Hubbard #define FS_BTREE_FL			0x00001000 /* btree format dir */
279e2019472SJohn Hubbard #define FS_INDEX_FL			0x00001000 /* hash-indexed directory */
280e2019472SJohn Hubbard #define FS_IMAGIC_FL			0x00002000 /* AFS directory */
281e2019472SJohn Hubbard #define FS_JOURNAL_DATA_FL		0x00004000 /* Reserved for ext3 */
282e2019472SJohn Hubbard #define FS_NOTAIL_FL			0x00008000 /* file tail should not be merged */
283e2019472SJohn Hubbard #define FS_DIRSYNC_FL			0x00010000 /* dirsync behaviour (directories only) */
284e2019472SJohn Hubbard #define FS_TOPDIR_FL			0x00020000 /* Top of directory hierarchies*/
285e2019472SJohn Hubbard #define FS_HUGE_FILE_FL			0x00040000 /* Reserved for ext4 */
286e2019472SJohn Hubbard #define FS_EXTENT_FL			0x00080000 /* Extents */
287e2019472SJohn Hubbard #define FS_VERITY_FL			0x00100000 /* Verity protected inode */
288e2019472SJohn Hubbard #define FS_EA_INODE_FL			0x00200000 /* Inode used for large EA */
289e2019472SJohn Hubbard #define FS_EOFBLOCKS_FL			0x00400000 /* Reserved for ext4 */
290e2019472SJohn Hubbard #define FS_NOCOW_FL			0x00800000 /* Do not cow file */
291e2019472SJohn Hubbard #define FS_DAX_FL			0x02000000 /* Inode is DAX */
292e2019472SJohn Hubbard #define FS_INLINE_DATA_FL		0x10000000 /* Reserved for ext4 */
293e2019472SJohn Hubbard #define FS_PROJINHERIT_FL		0x20000000 /* Create with parents projid */
294e2019472SJohn Hubbard #define FS_CASEFOLD_FL			0x40000000 /* Folder is case insensitive */
295e2019472SJohn Hubbard #define FS_RESERVED_FL			0x80000000 /* reserved for ext2 lib */
296e2019472SJohn Hubbard 
297e2019472SJohn Hubbard #define FS_FL_USER_VISIBLE		0x0003DFFF /* User visible flags */
298e2019472SJohn Hubbard #define FS_FL_USER_MODIFIABLE		0x000380FF /* User modifiable flags */
299e2019472SJohn Hubbard 
300e2019472SJohn Hubbard 
301e2019472SJohn Hubbard #define SYNC_FILE_RANGE_WAIT_BEFORE	1
302e2019472SJohn Hubbard #define SYNC_FILE_RANGE_WRITE		2
303e2019472SJohn Hubbard #define SYNC_FILE_RANGE_WAIT_AFTER	4
304e2019472SJohn Hubbard #define SYNC_FILE_RANGE_WRITE_AND_WAIT	(SYNC_FILE_RANGE_WRITE | \
305e2019472SJohn Hubbard 					 SYNC_FILE_RANGE_WAIT_BEFORE | \
306e2019472SJohn Hubbard 					 SYNC_FILE_RANGE_WAIT_AFTER)
307e2019472SJohn Hubbard 
308e2019472SJohn Hubbard /*
309e2019472SJohn Hubbard  * Flags for preadv2/pwritev2:
310e2019472SJohn Hubbard  */
311e2019472SJohn Hubbard 
312e2019472SJohn Hubbard typedef int __bitwise __kernel_rwf_t;
313e2019472SJohn Hubbard 
314e2019472SJohn Hubbard /* high priority request, poll if possible */
315*77179b6fSAndrii Nakryiko #define RWF_HIPRI	((__force __kernel_rwf_t)0x00000001)
316e2019472SJohn Hubbard 
317e2019472SJohn Hubbard /* per-IO O_DSYNC */
318*77179b6fSAndrii Nakryiko #define RWF_DSYNC	((__force __kernel_rwf_t)0x00000002)
319e2019472SJohn Hubbard 
320e2019472SJohn Hubbard /* per-IO O_SYNC */
321*77179b6fSAndrii Nakryiko #define RWF_SYNC	((__force __kernel_rwf_t)0x00000004)
322e2019472SJohn Hubbard 
323e2019472SJohn Hubbard /* per-IO, return -EAGAIN if operation would block */
324*77179b6fSAndrii Nakryiko #define RWF_NOWAIT	((__force __kernel_rwf_t)0x00000008)
325e2019472SJohn Hubbard 
326e2019472SJohn Hubbard /* per-IO O_APPEND */
327*77179b6fSAndrii Nakryiko #define RWF_APPEND	((__force __kernel_rwf_t)0x00000010)
328e2019472SJohn Hubbard 
329e2019472SJohn Hubbard /* per-IO negation of O_APPEND */
330*77179b6fSAndrii Nakryiko #define RWF_NOAPPEND	((__force __kernel_rwf_t)0x00000020)
331e2019472SJohn Hubbard 
332e2019472SJohn Hubbard /* mask of flags supported by the kernel */
333e2019472SJohn Hubbard #define RWF_SUPPORTED	(RWF_HIPRI | RWF_DSYNC | RWF_SYNC | RWF_NOWAIT |\
334e2019472SJohn Hubbard 			 RWF_APPEND | RWF_NOAPPEND)
335e2019472SJohn Hubbard 
336*77179b6fSAndrii Nakryiko #define PROCFS_IOCTL_MAGIC 'f'
337*77179b6fSAndrii Nakryiko 
338e2019472SJohn Hubbard /* Pagemap ioctl */
339*77179b6fSAndrii Nakryiko #define PAGEMAP_SCAN	_IOWR(PROCFS_IOCTL_MAGIC, 16, struct pm_scan_arg)
340e2019472SJohn Hubbard 
341e2019472SJohn Hubbard /* Bitmasks provided in pm_scan_args masks and reported in page_region.categories. */
342e2019472SJohn Hubbard #define PAGE_IS_WPALLOWED	(1 << 0)
343e2019472SJohn Hubbard #define PAGE_IS_WRITTEN		(1 << 1)
344e2019472SJohn Hubbard #define PAGE_IS_FILE		(1 << 2)
345e2019472SJohn Hubbard #define PAGE_IS_PRESENT		(1 << 3)
346e2019472SJohn Hubbard #define PAGE_IS_SWAPPED		(1 << 4)
347e2019472SJohn Hubbard #define PAGE_IS_PFNZERO		(1 << 5)
348e2019472SJohn Hubbard #define PAGE_IS_HUGE		(1 << 6)
349e2019472SJohn Hubbard #define PAGE_IS_SOFT_DIRTY	(1 << 7)
350e2019472SJohn Hubbard 
351e2019472SJohn Hubbard /*
352e2019472SJohn Hubbard  * struct page_region - Page region with flags
353e2019472SJohn Hubbard  * @start:	Start of the region
354e2019472SJohn Hubbard  * @end:	End of the region (exclusive)
355e2019472SJohn Hubbard  * @categories:	PAGE_IS_* category bitmask for the region
356e2019472SJohn Hubbard  */
357e2019472SJohn Hubbard struct page_region {
358e2019472SJohn Hubbard 	__u64 start;
359e2019472SJohn Hubbard 	__u64 end;
360e2019472SJohn Hubbard 	__u64 categories;
361e2019472SJohn Hubbard };
362e2019472SJohn Hubbard 
363e2019472SJohn Hubbard /* Flags for PAGEMAP_SCAN ioctl */
364e2019472SJohn Hubbard #define PM_SCAN_WP_MATCHING	(1 << 0)	/* Write protect the pages matched. */
365e2019472SJohn Hubbard #define PM_SCAN_CHECK_WPASYNC	(1 << 1)	/* Abort the scan when a non-WP-enabled page is found. */
366e2019472SJohn Hubbard 
367e2019472SJohn Hubbard /*
368e2019472SJohn Hubbard  * struct pm_scan_arg - Pagemap ioctl argument
369e2019472SJohn Hubbard  * @size:		Size of the structure
370e2019472SJohn Hubbard  * @flags:		Flags for the IOCTL
371e2019472SJohn Hubbard  * @start:		Starting address of the region
372e2019472SJohn Hubbard  * @end:		Ending address of the region
373e2019472SJohn Hubbard  * @walk_end		Address where the scan stopped (written by kernel).
374e2019472SJohn Hubbard  *			walk_end == end (address tags cleared) informs that the scan completed on entire range.
375e2019472SJohn Hubbard  * @vec:		Address of page_region struct array for output
376e2019472SJohn Hubbard  * @vec_len:		Length of the page_region struct array
377e2019472SJohn Hubbard  * @max_pages:		Optional limit for number of returned pages (0 = disabled)
378e2019472SJohn Hubbard  * @category_inverted:	PAGE_IS_* categories which values match if 0 instead of 1
379e2019472SJohn Hubbard  * @category_mask:	Skip pages for which any category doesn't match
380e2019472SJohn Hubbard  * @category_anyof_mask: Skip pages for which no category matches
381e2019472SJohn Hubbard  * @return_mask:	PAGE_IS_* categories that are to be reported in `page_region`s returned
382e2019472SJohn Hubbard  */
383e2019472SJohn Hubbard struct pm_scan_arg {
384e2019472SJohn Hubbard 	__u64 size;
385e2019472SJohn Hubbard 	__u64 flags;
386e2019472SJohn Hubbard 	__u64 start;
387e2019472SJohn Hubbard 	__u64 end;
388e2019472SJohn Hubbard 	__u64 walk_end;
389e2019472SJohn Hubbard 	__u64 vec;
390e2019472SJohn Hubbard 	__u64 vec_len;
391e2019472SJohn Hubbard 	__u64 max_pages;
392e2019472SJohn Hubbard 	__u64 category_inverted;
393e2019472SJohn Hubbard 	__u64 category_mask;
394e2019472SJohn Hubbard 	__u64 category_anyof_mask;
395e2019472SJohn Hubbard 	__u64 return_mask;
396e2019472SJohn Hubbard };
397e2019472SJohn Hubbard 
398*77179b6fSAndrii Nakryiko /* /proc/<pid>/maps ioctl */
399*77179b6fSAndrii Nakryiko #define PROCMAP_QUERY	_IOWR(PROCFS_IOCTL_MAGIC, 17, struct procmap_query)
400*77179b6fSAndrii Nakryiko 
401*77179b6fSAndrii Nakryiko enum procmap_query_flags {
402*77179b6fSAndrii Nakryiko 	/*
403*77179b6fSAndrii Nakryiko 	 * VMA permission flags.
404*77179b6fSAndrii Nakryiko 	 *
405*77179b6fSAndrii Nakryiko 	 * Can be used as part of procmap_query.query_flags field to look up
406*77179b6fSAndrii Nakryiko 	 * only VMAs satisfying specified subset of permissions. E.g., specifying
407*77179b6fSAndrii Nakryiko 	 * PROCMAP_QUERY_VMA_READABLE only will return both readable and read/write VMAs,
408*77179b6fSAndrii Nakryiko 	 * while having PROCMAP_QUERY_VMA_READABLE | PROCMAP_QUERY_VMA_WRITABLE will only
409*77179b6fSAndrii Nakryiko 	 * return read/write VMAs, though both executable/non-executable and
410*77179b6fSAndrii Nakryiko 	 * private/shared will be ignored.
411*77179b6fSAndrii Nakryiko 	 *
412*77179b6fSAndrii Nakryiko 	 * PROCMAP_QUERY_VMA_* flags are also returned in procmap_query.vma_flags
413*77179b6fSAndrii Nakryiko 	 * field to specify actual VMA permissions.
414*77179b6fSAndrii Nakryiko 	 */
415*77179b6fSAndrii Nakryiko 	PROCMAP_QUERY_VMA_READABLE		= 0x01,
416*77179b6fSAndrii Nakryiko 	PROCMAP_QUERY_VMA_WRITABLE		= 0x02,
417*77179b6fSAndrii Nakryiko 	PROCMAP_QUERY_VMA_EXECUTABLE		= 0x04,
418*77179b6fSAndrii Nakryiko 	PROCMAP_QUERY_VMA_SHARED		= 0x08,
419*77179b6fSAndrii Nakryiko 	/*
420*77179b6fSAndrii Nakryiko 	 * Query modifier flags.
421*77179b6fSAndrii Nakryiko 	 *
422*77179b6fSAndrii Nakryiko 	 * By default VMA that covers provided address is returned, or -ENOENT
423*77179b6fSAndrii Nakryiko 	 * is returned. With PROCMAP_QUERY_COVERING_OR_NEXT_VMA flag set, closest
424*77179b6fSAndrii Nakryiko 	 * VMA with vma_start > addr will be returned if no covering VMA is
425*77179b6fSAndrii Nakryiko 	 * found.
426*77179b6fSAndrii Nakryiko 	 *
427*77179b6fSAndrii Nakryiko 	 * PROCMAP_QUERY_FILE_BACKED_VMA instructs query to consider only VMAs that
428*77179b6fSAndrii Nakryiko 	 * have file backing. Can be combined with PROCMAP_QUERY_COVERING_OR_NEXT_VMA
429*77179b6fSAndrii Nakryiko 	 * to iterate all VMAs with file backing.
430*77179b6fSAndrii Nakryiko 	 */
431*77179b6fSAndrii Nakryiko 	PROCMAP_QUERY_COVERING_OR_NEXT_VMA	= 0x10,
432*77179b6fSAndrii Nakryiko 	PROCMAP_QUERY_FILE_BACKED_VMA		= 0x20,
433*77179b6fSAndrii Nakryiko };
434*77179b6fSAndrii Nakryiko 
435*77179b6fSAndrii Nakryiko /*
436*77179b6fSAndrii Nakryiko  * Input/output argument structured passed into ioctl() call. It can be used
437*77179b6fSAndrii Nakryiko  * to query a set of VMAs (Virtual Memory Areas) of a process.
438*77179b6fSAndrii Nakryiko  *
439*77179b6fSAndrii Nakryiko  * Each field can be one of three kinds, marked in a short comment to the
440*77179b6fSAndrii Nakryiko  * right of the field:
441*77179b6fSAndrii Nakryiko  *   - "in", input argument, user has to provide this value, kernel doesn't modify it;
442*77179b6fSAndrii Nakryiko  *   - "out", output argument, kernel sets this field with VMA data;
443*77179b6fSAndrii Nakryiko  *   - "in/out", input and output argument; user provides initial value (used
444*77179b6fSAndrii Nakryiko  *     to specify maximum allowable buffer size), and kernel sets it to actual
445*77179b6fSAndrii Nakryiko  *     amount of data written (or zero, if there is no data).
446*77179b6fSAndrii Nakryiko  *
447*77179b6fSAndrii Nakryiko  * If matching VMA is found (according to criterias specified by
448*77179b6fSAndrii Nakryiko  * query_addr/query_flags, all the out fields are filled out, and ioctl()
449*77179b6fSAndrii Nakryiko  * returns 0. If there is no matching VMA, -ENOENT will be returned.
450*77179b6fSAndrii Nakryiko  * In case of any other error, negative error code other than -ENOENT is
451*77179b6fSAndrii Nakryiko  * returned.
452*77179b6fSAndrii Nakryiko  *
453*77179b6fSAndrii Nakryiko  * Most of the data is similar to the one returned as text in /proc/<pid>/maps
454*77179b6fSAndrii Nakryiko  * file, but procmap_query provides more querying flexibility. There are no
455*77179b6fSAndrii Nakryiko  * consistency guarantees between subsequent ioctl() calls, but data returned
456*77179b6fSAndrii Nakryiko  * for matched VMA is self-consistent.
457*77179b6fSAndrii Nakryiko  */
458*77179b6fSAndrii Nakryiko struct procmap_query {
459*77179b6fSAndrii Nakryiko 	/* Query struct size, for backwards/forward compatibility */
460*77179b6fSAndrii Nakryiko 	__u64 size;
461*77179b6fSAndrii Nakryiko 	/*
462*77179b6fSAndrii Nakryiko 	 * Query flags, a combination of enum procmap_query_flags values.
463*77179b6fSAndrii Nakryiko 	 * Defines query filtering and behavior, see enum procmap_query_flags.
464*77179b6fSAndrii Nakryiko 	 *
465*77179b6fSAndrii Nakryiko 	 * Input argument, provided by user. Kernel doesn't modify it.
466*77179b6fSAndrii Nakryiko 	 */
467*77179b6fSAndrii Nakryiko 	__u64 query_flags;		/* in */
468*77179b6fSAndrii Nakryiko 	/*
469*77179b6fSAndrii Nakryiko 	 * Query address. By default, VMA that covers this address will
470*77179b6fSAndrii Nakryiko 	 * be looked up. PROCMAP_QUERY_* flags above modify this default
471*77179b6fSAndrii Nakryiko 	 * behavior further.
472*77179b6fSAndrii Nakryiko 	 *
473*77179b6fSAndrii Nakryiko 	 * Input argument, provided by user. Kernel doesn't modify it.
474*77179b6fSAndrii Nakryiko 	 */
475*77179b6fSAndrii Nakryiko 	__u64 query_addr;		/* in */
476*77179b6fSAndrii Nakryiko 	/* VMA starting (inclusive) and ending (exclusive) address, if VMA is found. */
477*77179b6fSAndrii Nakryiko 	__u64 vma_start;		/* out */
478*77179b6fSAndrii Nakryiko 	__u64 vma_end;			/* out */
479*77179b6fSAndrii Nakryiko 	/* VMA permissions flags. A combination of PROCMAP_QUERY_VMA_* flags. */
480*77179b6fSAndrii Nakryiko 	__u64 vma_flags;		/* out */
481*77179b6fSAndrii Nakryiko 	/* VMA backing page size granularity. */
482*77179b6fSAndrii Nakryiko 	__u64 vma_page_size;		/* out */
483*77179b6fSAndrii Nakryiko 	/*
484*77179b6fSAndrii Nakryiko 	 * VMA file offset. If VMA has file backing, this specifies offset
485*77179b6fSAndrii Nakryiko 	 * within the file that VMA's start address corresponds to.
486*77179b6fSAndrii Nakryiko 	 * Is set to zero if VMA has no backing file.
487*77179b6fSAndrii Nakryiko 	 */
488*77179b6fSAndrii Nakryiko 	__u64 vma_offset;		/* out */
489*77179b6fSAndrii Nakryiko 	/* Backing file's inode number, or zero, if VMA has no backing file. */
490*77179b6fSAndrii Nakryiko 	__u64 inode;			/* out */
491*77179b6fSAndrii Nakryiko 	/* Backing file's device major/minor number, or zero, if VMA has no backing file. */
492*77179b6fSAndrii Nakryiko 	__u32 dev_major;		/* out */
493*77179b6fSAndrii Nakryiko 	__u32 dev_minor;		/* out */
494*77179b6fSAndrii Nakryiko 	/*
495*77179b6fSAndrii Nakryiko 	 * If set to non-zero value, signals the request to return VMA name
496*77179b6fSAndrii Nakryiko 	 * (i.e., VMA's backing file's absolute path, with " (deleted)" suffix
497*77179b6fSAndrii Nakryiko 	 * appended, if file was unlinked from FS) for matched VMA. VMA name
498*77179b6fSAndrii Nakryiko 	 * can also be some special name (e.g., "[heap]", "[stack]") or could
499*77179b6fSAndrii Nakryiko 	 * be even user-supplied with prctl(PR_SET_VMA, PR_SET_VMA_ANON_NAME).
500*77179b6fSAndrii Nakryiko 	 *
501*77179b6fSAndrii Nakryiko 	 * Kernel will set this field to zero, if VMA has no associated name.
502*77179b6fSAndrii Nakryiko 	 * Otherwise kernel will return actual amount of bytes filled in
503*77179b6fSAndrii Nakryiko 	 * user-supplied buffer (see vma_name_addr field below), including the
504*77179b6fSAndrii Nakryiko 	 * terminating zero.
505*77179b6fSAndrii Nakryiko 	 *
506*77179b6fSAndrii Nakryiko 	 * If VMA name is longer that user-supplied maximum buffer size,
507*77179b6fSAndrii Nakryiko 	 * -E2BIG error is returned.
508*77179b6fSAndrii Nakryiko 	 *
509*77179b6fSAndrii Nakryiko 	 * If this field is set to non-zero value, vma_name_addr should point
510*77179b6fSAndrii Nakryiko 	 * to valid user space memory buffer of at least vma_name_size bytes.
511*77179b6fSAndrii Nakryiko 	 * If set to zero, vma_name_addr should be set to zero as well
512*77179b6fSAndrii Nakryiko 	 */
513*77179b6fSAndrii Nakryiko 	__u32 vma_name_size;		/* in/out */
514*77179b6fSAndrii Nakryiko 	/*
515*77179b6fSAndrii Nakryiko 	 * If set to non-zero value, signals the request to extract and return
516*77179b6fSAndrii Nakryiko 	 * VMA's backing file's build ID, if the backing file is an ELF file
517*77179b6fSAndrii Nakryiko 	 * and it contains embedded build ID.
518*77179b6fSAndrii Nakryiko 	 *
519*77179b6fSAndrii Nakryiko 	 * Kernel will set this field to zero, if VMA has no backing file,
520*77179b6fSAndrii Nakryiko 	 * backing file is not an ELF file, or ELF file has no build ID
521*77179b6fSAndrii Nakryiko 	 * embedded.
522*77179b6fSAndrii Nakryiko 	 *
523*77179b6fSAndrii Nakryiko 	 * Build ID is a binary value (not a string). Kernel will set
524*77179b6fSAndrii Nakryiko 	 * build_id_size field to exact number of bytes used for build ID.
525*77179b6fSAndrii Nakryiko 	 * If build ID is requested and present, but needs more bytes than
526*77179b6fSAndrii Nakryiko 	 * user-supplied maximum buffer size (see build_id_addr field below),
527*77179b6fSAndrii Nakryiko 	 * -E2BIG error will be returned.
528*77179b6fSAndrii Nakryiko 	 *
529*77179b6fSAndrii Nakryiko 	 * If this field is set to non-zero value, build_id_addr should point
530*77179b6fSAndrii Nakryiko 	 * to valid user space memory buffer of at least build_id_size bytes.
531*77179b6fSAndrii Nakryiko 	 * If set to zero, build_id_addr should be set to zero as well
532*77179b6fSAndrii Nakryiko 	 */
533*77179b6fSAndrii Nakryiko 	__u32 build_id_size;		/* in/out */
534*77179b6fSAndrii Nakryiko 	/*
535*77179b6fSAndrii Nakryiko 	 * User-supplied address of a buffer of at least vma_name_size bytes
536*77179b6fSAndrii Nakryiko 	 * for kernel to fill with matched VMA's name (see vma_name_size field
537*77179b6fSAndrii Nakryiko 	 * description above for details).
538*77179b6fSAndrii Nakryiko 	 *
539*77179b6fSAndrii Nakryiko 	 * Should be set to zero if VMA name should not be returned.
540*77179b6fSAndrii Nakryiko 	 */
541*77179b6fSAndrii Nakryiko 	__u64 vma_name_addr;		/* in */
542*77179b6fSAndrii Nakryiko 	/*
543*77179b6fSAndrii Nakryiko 	 * User-supplied address of a buffer of at least build_id_size bytes
544*77179b6fSAndrii Nakryiko 	 * for kernel to fill with matched VMA's ELF build ID, if available
545*77179b6fSAndrii Nakryiko 	 * (see build_id_size field description above for details).
546*77179b6fSAndrii Nakryiko 	 *
547*77179b6fSAndrii Nakryiko 	 * Should be set to zero if build ID should not be returned.
548*77179b6fSAndrii Nakryiko 	 */
549*77179b6fSAndrii Nakryiko 	__u64 build_id_addr;		/* in */
550*77179b6fSAndrii Nakryiko };
551*77179b6fSAndrii Nakryiko 
552*77179b6fSAndrii Nakryiko #endif /* _UAPI_LINUX_FS_H */
553