dir.c 36.2 KB
Newer Older
1
#include <linux/ceph/ceph_debug.h>
S
Sage Weil 已提交
2 3 4 5

#include <linux/spinlock.h>
#include <linux/fs_struct.h>
#include <linux/namei.h>
6
#include <linux/slab.h>
S
Sage Weil 已提交
7 8 9
#include <linux/sched.h>

#include "super.h"
10
#include "mds_client.h"
S
Sage Weil 已提交
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

/*
 * Directory operations: readdir, lookup, create, link, unlink,
 * rename, etc.
 */

/*
 * Ceph MDS operations are specified in terms of a base ino and
 * relative path.  Thus, the client can specify an operation on a
 * specific inode (e.g., a getattr due to fstat(2)), or as a path
 * relative to, say, the root directory.
 *
 * Normally, we limit ourselves to strict inode ops (no path component)
 * or dentry operations (a single path component relative to an ino).  The
 * exception to this is open_root_dentry(), which will open the mount
 * point by name.
 */

const struct inode_operations ceph_dir_iops;
const struct file_operations ceph_dir_fops;
S
Sage Weil 已提交
31
const struct dentry_operations ceph_dentry_ops;
S
Sage Weil 已提交
32 33 34 35 36 37 38 39 40 41 42

/*
 * Initialize ceph dentry state.
 */
int ceph_init_dentry(struct dentry *dentry)
{
	struct ceph_dentry_info *di;

	if (dentry->d_fsdata)
		return 0;

43
	di = kmem_cache_alloc(ceph_dentry_cachep, GFP_NOFS | __GFP_ZERO);
S
Sage Weil 已提交
44 45 46 47
	if (!di)
		return -ENOMEM;          /* oh well */

	spin_lock(&dentry->d_lock);
48 49 50
	if (dentry->d_fsdata) {
		/* lost a race */
		kmem_cache_free(ceph_dentry_cachep, di);
S
Sage Weil 已提交
51
		goto out_unlock;
52
	}
53

54
	if (ceph_snap(dentry->d_parent->d_inode) == CEPH_NOSNAP)
55 56 57 58 59 60
		d_set_d_op(dentry, &ceph_dentry_ops);
	else if (ceph_snap(dentry->d_parent->d_inode) == CEPH_SNAPDIR)
		d_set_d_op(dentry, &ceph_snapdir_dentry_ops);
	else
		d_set_d_op(dentry, &ceph_snap_dentry_ops);

S
Sage Weil 已提交
61 62 63
	di->dentry = dentry;
	di->lease_session = NULL;
	dentry->d_time = jiffies;
64 65 66
	/* avoid reordering d_fsdata setup so that the check above is safe */
	smp_mb();
	dentry->d_fsdata = di;
S
Sage Weil 已提交
67 68 69 70 71 72
	ceph_dentry_lru_add(dentry);
out_unlock:
	spin_unlock(&dentry->d_lock);
	return 0;
}

S
Sage Weil 已提交
73 74 75 76 77 78 79 80
struct inode *ceph_get_dentry_parent_inode(struct dentry *dentry)
{
	struct inode *inode = NULL;

	if (!dentry)
		return NULL;

	spin_lock(&dentry->d_lock);
81
	if (!IS_ROOT(dentry)) {
S
Sage Weil 已提交
82 83 84 85 86 87
		inode = dentry->d_parent->d_inode;
		ihold(inode);
	}
	spin_unlock(&dentry->d_lock);
	return inode;
}
S
Sage Weil 已提交
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102


/*
 * for readdir, we encode the directory frag and offset within that
 * frag into f_pos.
 */
static unsigned fpos_frag(loff_t p)
{
	return p >> 32;
}
static unsigned fpos_off(loff_t p)
{
	return p & 0xffffffff;
}

Y
Yan, Zheng 已提交
103 104 105 106 107 108 109 110
static int fpos_cmp(loff_t l, loff_t r)
{
	int v = ceph_frag_compare(fpos_frag(l), fpos_frag(r));
	if (v)
		return v;
	return (int)(fpos_off(l) - fpos_off(r));
}

S
Sage Weil 已提交
111 112 113 114 115 116 117
/*
 * When possible, we try to satisfy a readdir by peeking at the
 * dcache.  We make this work by carefully ordering dentries on
 * d_u.d_child when we initially get results back from the MDS, and
 * falling back to a "normal" sync readdir if any dentries in the dir
 * are dropped.
 *
118
 * Complete dir indicates that we have all dentries in the dir.  It is
S
Sage Weil 已提交
119 120 121
 * defined IFF we hold CEPH_CAP_FILE_SHARED (which will be revoked by
 * the MDS if/when the directory is modified).
 */
122 123
static int __dcache_readdir(struct file *file,  struct dir_context *ctx,
			    u32 shared_gen)
S
Sage Weil 已提交
124
{
A
Al Viro 已提交
125 126
	struct ceph_file_info *fi = file->private_data;
	struct dentry *parent = file->f_dentry;
S
Sage Weil 已提交
127 128 129 130 131 132 133 134 135 136
	struct inode *dir = parent->d_inode;
	struct list_head *p;
	struct dentry *dentry, *last;
	struct ceph_dentry_info *di;
	int err = 0;

	/* claim ref on last dentry we returned */
	last = fi->dentry;
	fi->dentry = NULL;

137 138
	dout("__dcache_readdir %p v%u at %llu (last %p)\n",
	     dir, shared_gen, ctx->pos, last);
S
Sage Weil 已提交
139

N
Nick Piggin 已提交
140
	spin_lock(&parent->d_lock);
S
Sage Weil 已提交
141 142

	/* start at beginning? */
A
Al Viro 已提交
143
	if (ctx->pos == 2 || last == NULL ||
144
	    fpos_cmp(ctx->pos, ceph_dentry(last)->offset) < 0) {
S
Sage Weil 已提交
145 146 147 148 149 150 151 152 153 154 155 156
		if (list_empty(&parent->d_subdirs))
			goto out_unlock;
		p = parent->d_subdirs.prev;
		dout(" initial p %p/%p\n", p->prev, p->next);
	} else {
		p = last->d_u.d_child.prev;
	}

more:
	dentry = list_entry(p, struct dentry, d_u.d_child);
	di = ceph_dentry(dentry);
	while (1) {
S
Sage Weil 已提交
157 158
		dout(" p %p/%p %s d_subdirs %p/%p\n", p->prev, p->next,
		     d_unhashed(dentry) ? "!hashed" : "hashed",
S
Sage Weil 已提交
159 160
		     parent->d_subdirs.prev, parent->d_subdirs.next);
		if (p == &parent->d_subdirs) {
161
			fi->flags |= CEPH_F_ATEND;
S
Sage Weil 已提交
162 163
			goto out_unlock;
		}
N
Nick Piggin 已提交
164
		spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
165 166
		if (di->lease_shared_gen == shared_gen &&
		    !d_unhashed(dentry) && dentry->d_inode &&
167
		    ceph_snap(dentry->d_inode) != CEPH_SNAPDIR &&
S
Sage Weil 已提交
168
		    ceph_ino(dentry->d_inode) != CEPH_INO_CEPH &&
Y
Yan, Zheng 已提交
169
		    fpos_cmp(ctx->pos, di->offset) <= 0)
S
Sage Weil 已提交
170 171 172
			break;
		dout(" skipping %p %.*s at %llu (%llu)%s%s\n", dentry,
		     dentry->d_name.len, dentry->d_name.name, di->offset,
A
Al Viro 已提交
173
		     ctx->pos, d_unhashed(dentry) ? " unhashed" : "",
S
Sage Weil 已提交
174
		     !dentry->d_inode ? " null" : "");
N
Nick Piggin 已提交
175
		spin_unlock(&dentry->d_lock);
S
Sage Weil 已提交
176 177 178 179 180
		p = p->prev;
		dentry = list_entry(p, struct dentry, d_u.d_child);
		di = ceph_dentry(dentry);
	}

N
Nick Piggin 已提交
181
	dget_dlock(dentry);
N
Nick Piggin 已提交
182
	spin_unlock(&dentry->d_lock);
N
Nick Piggin 已提交
183
	spin_unlock(&parent->d_lock);
S
Sage Weil 已提交
184

185 186 187 188 189 190 191 192
	/* make sure a dentry wasn't dropped while we didn't have parent lock */
	if (!ceph_dir_is_complete(dir)) {
		dout(" lost dir complete on %p; falling back to mds\n", dir);
		dput(dentry);
		err = -EAGAIN;
		goto out;
	}

A
Al Viro 已提交
193
	dout(" %llu (%llu) dentry %p %.*s %p\n", di->offset, ctx->pos,
S
Sage Weil 已提交
194
	     dentry, dentry->d_name.len, dentry->d_name.name, dentry->d_inode);
A
Al Viro 已提交
195 196
	if (!dir_emit(ctx, dentry->d_name.name,
		      dentry->d_name.len,
Y
Yehuda Sadeh 已提交
197
		      ceph_translate_ino(dentry->d_sb, dentry->d_inode->i_ino),
A
Al Viro 已提交
198 199
		      dentry->d_inode->i_mode >> 12)) {
		if (last) {
S
Sage Weil 已提交
200 201
			/* remember our position */
			fi->dentry = last;
Y
Yan, Zheng 已提交
202
			fi->next_offset = fpos_off(di->offset);
S
Sage Weil 已提交
203
		}
A
Al Viro 已提交
204 205
		dput(dentry);
		return 0;
S
Sage Weil 已提交
206
	}
207

208 209
	ctx->pos = di->offset + 1;

A
Al Viro 已提交
210 211 212
	if (last)
		dput(last);
	last = dentry;
S
Sage Weil 已提交
213

N
Nick Piggin 已提交
214
	spin_lock(&parent->d_lock);
215 216
	p = p->prev;	/* advance to next dentry */
	goto more;
S
Sage Weil 已提交
217 218

out_unlock:
N
Nick Piggin 已提交
219
	spin_unlock(&parent->d_lock);
220 221
out:
	if (last)
S
Sage Weil 已提交
222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244
		dput(last);
	return err;
}

/*
 * make note of the last dentry we read, so we can
 * continue at the same lexicographical point,
 * regardless of what dir changes take place on the
 * server.
 */
static int note_last_dentry(struct ceph_file_info *fi, const char *name,
			    int len)
{
	kfree(fi->last_name);
	fi->last_name = kmalloc(len+1, GFP_NOFS);
	if (!fi->last_name)
		return -ENOMEM;
	memcpy(fi->last_name, name, len);
	fi->last_name[len] = 0;
	dout("note_last_dentry '%s'\n", fi->last_name);
	return 0;
}

A
Al Viro 已提交
245
static int ceph_readdir(struct file *file, struct dir_context *ctx)
S
Sage Weil 已提交
246
{
A
Al Viro 已提交
247 248
	struct ceph_file_info *fi = file->private_data;
	struct inode *inode = file_inode(file);
S
Sage Weil 已提交
249
	struct ceph_inode_info *ci = ceph_inode(inode);
250 251
	struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
	struct ceph_mds_client *mdsc = fsc->mdsc;
A
Al Viro 已提交
252 253
	unsigned frag = fpos_frag(ctx->pos);
	int off = fpos_off(ctx->pos);
S
Sage Weil 已提交
254 255 256 257
	int err;
	u32 ftype;
	struct ceph_mds_reply_info_parsed *rinfo;

A
Al Viro 已提交
258
	dout("readdir %p file %p frag %u off %u\n", inode, file, frag, off);
259
	if (fi->flags & CEPH_F_ATEND)
S
Sage Weil 已提交
260 261 262
		return 0;

	/* always start with . and .. */
A
Al Viro 已提交
263
	if (ctx->pos == 0) {
S
Sage Weil 已提交
264 265
		/* note dir version at start of readdir so we can tell
		 * if any dentries get dropped */
266
		fi->dir_release_count = atomic_read(&ci->i_release_count);
S
Sage Weil 已提交
267 268

		dout("readdir off 0 -> '.'\n");
A
Al Viro 已提交
269
		if (!dir_emit(ctx, ".", 1, 
Y
Yehuda Sadeh 已提交
270
			    ceph_translate_ino(inode->i_sb, inode->i_ino),
A
Al Viro 已提交
271
			    inode->i_mode >> 12))
S
Sage Weil 已提交
272
			return 0;
A
Al Viro 已提交
273
		ctx->pos = 1;
S
Sage Weil 已提交
274 275
		off = 1;
	}
A
Al Viro 已提交
276 277
	if (ctx->pos == 1) {
		ino_t ino = parent_ino(file->f_dentry);
S
Sage Weil 已提交
278
		dout("readdir off 1 -> '..'\n");
A
Al Viro 已提交
279
		if (!dir_emit(ctx, "..", 2,
Y
Yehuda Sadeh 已提交
280
			    ceph_translate_ino(inode->i_sb, ino),
A
Al Viro 已提交
281
			    inode->i_mode >> 12))
S
Sage Weil 已提交
282
			return 0;
A
Al Viro 已提交
283
		ctx->pos = 2;
S
Sage Weil 已提交
284 285 286 287
		off = 2;
	}

	/* can we use the dcache? */
288
	spin_lock(&ci->i_ceph_lock);
A
Al Viro 已提交
289
	if ((ctx->pos == 2 || fi->dentry) &&
290
	    !ceph_test_mount_opt(fsc, NOASYNCREADDIR) &&
291
	    ceph_snap(inode) != CEPH_SNAPDIR &&
292
	    __ceph_dir_is_complete(ci) &&
S
Sage Weil 已提交
293
	    __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) {
294
		u32 shared_gen = ci->i_shared_gen;
295
		spin_unlock(&ci->i_ceph_lock);
296
		err = __dcache_readdir(file, ctx, shared_gen);
297
		if (err != -EAGAIN)
S
Sage Weil 已提交
298
			return err;
299 300
		frag = fpos_frag(ctx->pos);
		off = fpos_off(ctx->pos);
301
	} else {
302
		spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
	}
	if (fi->dentry) {
		err = note_last_dentry(fi, fi->dentry->d_name.name,
				       fi->dentry->d_name.len);
		if (err)
			return err;
		dput(fi->dentry);
		fi->dentry = NULL;
	}

	/* proceed with a normal readdir */

more:
	/* do we have the correct frag content buffered? */
	if (fi->frag != frag || fi->last_readdir == NULL) {
		struct ceph_mds_request *req;
		int op = ceph_snap(inode) == CEPH_SNAPDIR ?
			CEPH_MDS_OP_LSSNAP : CEPH_MDS_OP_READDIR;

		/* discard old result, if any */
323
		if (fi->last_readdir) {
S
Sage Weil 已提交
324
			ceph_mdsc_put_request(fi->last_readdir);
325 326
			fi->last_readdir = NULL;
		}
S
Sage Weil 已提交
327 328 329 330 331 332

		dout("readdir fetching %llx.%llx frag %x offset '%s'\n",
		     ceph_vinop(inode), frag, fi->last_name);
		req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS);
		if (IS_ERR(req))
			return PTR_ERR(req);
333 334 335 336 337
		err = ceph_alloc_readdir_reply_buffer(req, inode);
		if (err) {
			ceph_mdsc_put_request(req);
			return err;
		}
338 339
		req->r_inode = inode;
		ihold(inode);
A
Al Viro 已提交
340
		req->r_dentry = dget(file->f_dentry);
S
Sage Weil 已提交
341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
		/* hints to request -> mds selection code */
		req->r_direct_mode = USE_AUTH_MDS;
		req->r_direct_hash = ceph_frag_value(frag);
		req->r_direct_is_hash = true;
		req->r_path2 = kstrdup(fi->last_name, GFP_NOFS);
		req->r_readdir_offset = fi->next_offset;
		req->r_args.readdir.frag = cpu_to_le32(frag);
		err = ceph_mdsc_do_request(mdsc, NULL, req);
		if (err < 0) {
			ceph_mdsc_put_request(req);
			return err;
		}
		dout("readdir got and parsed readdir result=%d"
		     " on frag %x, end=%d, complete=%d\n", err, frag,
		     (int)req->r_reply_info.dir_end,
		     (int)req->r_reply_info.dir_complete);

		if (!req->r_did_prepopulate) {
			dout("readdir !did_prepopulate");
360 361
			/* preclude from marking dir complete */
			fi->dir_release_count--;
S
Sage Weil 已提交
362 363 364
		}

		/* note next offset and last dentry name */
365 366 367 368 369 370 371 372 373
		rinfo = &req->r_reply_info;
		if (le32_to_cpu(rinfo->dir_dir->frag) != frag) {
			frag = le32_to_cpu(rinfo->dir_dir->frag);
			if (ceph_frag_is_leftmost(frag))
				fi->next_offset = 2;
			else
				fi->next_offset = 0;
			off = fi->next_offset;
		}
Y
Yan, Zheng 已提交
374
		fi->frag = frag;
S
Sage Weil 已提交
375 376 377 378 379 380
		fi->offset = fi->next_offset;
		fi->last_readdir = req;

		if (req->r_reply_info.dir_end) {
			kfree(fi->last_name);
			fi->last_name = NULL;
381 382 383 384
			if (ceph_frag_is_rightmost(frag))
				fi->next_offset = 2;
			else
				fi->next_offset = 0;
S
Sage Weil 已提交
385 386 387 388 389 390 391 392 393 394 395 396 397
		} else {
			err = note_last_dentry(fi,
				       rinfo->dir_dname[rinfo->dir_nr-1],
				       rinfo->dir_dname_len[rinfo->dir_nr-1]);
			if (err)
				return err;
			fi->next_offset += rinfo->dir_nr;
		}
	}

	rinfo = &fi->last_readdir->r_reply_info;
	dout("readdir frag %x num %d off %d chunkoff %d\n", frag,
	     rinfo->dir_nr, off, fi->offset);
A
Al Viro 已提交
398 399

	ctx->pos = ceph_make_fpos(frag, off);
400
	while (off >= fi->offset && off - fi->offset < rinfo->dir_nr) {
S
Sage Weil 已提交
401 402
		struct ceph_mds_reply_inode *in =
			rinfo->dir_in[off - fi->offset].in;
403 404 405
		struct ceph_vino vino;
		ino_t ino;

S
Sage Weil 已提交
406
		dout("readdir off %d (%d/%d) -> %lld '%.*s' %p\n",
A
Al Viro 已提交
407
		     off, off - fi->offset, rinfo->dir_nr, ctx->pos,
S
Sage Weil 已提交
408 409 410 411
		     rinfo->dir_dname_len[off - fi->offset],
		     rinfo->dir_dname[off - fi->offset], in);
		BUG_ON(!in);
		ftype = le32_to_cpu(in->mode) >> 12;
412 413 414
		vino.ino = le64_to_cpu(in->ino);
		vino.snap = le64_to_cpu(in->snapid);
		ino = ceph_vino_to_ino(vino);
A
Al Viro 已提交
415
		if (!dir_emit(ctx,
S
Sage Weil 已提交
416 417
			    rinfo->dir_dname[off - fi->offset],
			    rinfo->dir_dname_len[off - fi->offset],
A
Al Viro 已提交
418
			    ceph_translate_ino(inode->i_sb, ino), ftype)) {
S
Sage Weil 已提交
419 420 421 422
			dout("filldir stopping us...\n");
			return 0;
		}
		off++;
A
Al Viro 已提交
423
		ctx->pos++;
S
Sage Weil 已提交
424 425 426 427 428 429 430 431 432 433 434 435
	}

	if (fi->last_name) {
		ceph_mdsc_put_request(fi->last_readdir);
		fi->last_readdir = NULL;
		goto more;
	}

	/* more frags? */
	if (!ceph_frag_is_rightmost(frag)) {
		frag = ceph_frag_next(frag);
		off = 0;
A
Al Viro 已提交
436
		ctx->pos = ceph_make_fpos(frag, off);
S
Sage Weil 已提交
437 438 439
		dout("readdir next frag is %x\n", frag);
		goto more;
	}
440
	fi->flags |= CEPH_F_ATEND;
S
Sage Weil 已提交
441 442 443 444 445 446

	/*
	 * if dir_release_count still matches the dir, no dentries
	 * were released during the whole readdir, and we should have
	 * the complete dir contents in our cache.
	 */
447
	spin_lock(&ci->i_ceph_lock);
448
	if (atomic_read(&ci->i_release_count) == fi->dir_release_count) {
449
		dout(" marking %p complete\n", inode);
450
		__ceph_dir_set_complete(ci, fi->dir_release_count);
S
Sage Weil 已提交
451
	}
452
	spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
453

A
Al Viro 已提交
454
	dout("readdir %p file %p done.\n", inode, file);
S
Sage Weil 已提交
455 456 457
	return 0;
}

Y
Yan, Zheng 已提交
458
static void reset_readdir(struct ceph_file_info *fi, unsigned frag)
S
Sage Weil 已提交
459 460 461 462 463 464
{
	if (fi->last_readdir) {
		ceph_mdsc_put_request(fi->last_readdir);
		fi->last_readdir = NULL;
	}
	kfree(fi->last_name);
S
Sage Weil 已提交
465
	fi->last_name = NULL;
Y
Yan, Zheng 已提交
466 467 468 469
	if (ceph_frag_is_leftmost(frag))
		fi->next_offset = 2;  /* compensate for . and .. */
	else
		fi->next_offset = 0;
S
Sage Weil 已提交
470 471 472 473
	if (fi->dentry) {
		dput(fi->dentry);
		fi->dentry = NULL;
	}
474
	fi->flags &= ~CEPH_F_ATEND;
S
Sage Weil 已提交
475 476
}

477
static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int whence)
S
Sage Weil 已提交
478 479 480
{
	struct ceph_file_info *fi = file->private_data;
	struct inode *inode = file->f_mapping->host;
Y
Yan, Zheng 已提交
481
	loff_t old_offset = ceph_make_fpos(fi->frag, fi->next_offset);
S
Sage Weil 已提交
482 483 484
	loff_t retval;

	mutex_lock(&inode->i_mutex);
485
	retval = -EINVAL;
486
	switch (whence) {
S
Sage Weil 已提交
487 488 489 490 491
	case SEEK_END:
		offset += inode->i_size + 2;   /* FIXME */
		break;
	case SEEK_CUR:
		offset += file->f_pos;
492 493 494 495
	case SEEK_SET:
		break;
	default:
		goto out;
S
Sage Weil 已提交
496
	}
497

Y
Yan, Zheng 已提交
498
	if (offset >= 0) {
S
Sage Weil 已提交
499 500 501
		if (offset != file->f_pos) {
			file->f_pos = offset;
			file->f_version = 0;
502
			fi->flags &= ~CEPH_F_ATEND;
S
Sage Weil 已提交
503 504 505 506 507 508 509 510
		}
		retval = offset;

		/*
		 * discard buffered readdir content on seekdir(0), or
		 * seek to new frag, or seek prior to current chunk.
		 */
		if (offset == 0 ||
Y
Yan, Zheng 已提交
511
		    fpos_frag(offset) != fi->frag ||
S
Sage Weil 已提交
512 513
		    fpos_off(offset) < fi->offset) {
			dout("dir_llseek dropping %p content\n", file);
Y
Yan, Zheng 已提交
514
			reset_readdir(fi, fpos_frag(offset));
S
Sage Weil 已提交
515 516 517
		}

		/* bump dir_release_count if we did a forward seek */
Y
Yan, Zheng 已提交
518
		if (fpos_cmp(offset, old_offset) > 0)
S
Sage Weil 已提交
519 520
			fi->dir_release_count--;
	}
521
out:
S
Sage Weil 已提交
522 523 524 525 526
	mutex_unlock(&inode->i_mutex);
	return retval;
}

/*
527
 * Handle lookups for the hidden .snap directory.
S
Sage Weil 已提交
528
 */
529 530
int ceph_handle_snapdir(struct ceph_mds_request *req,
			struct dentry *dentry, int err)
S
Sage Weil 已提交
531
{
532
	struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb);
533
	struct inode *parent = dentry->d_parent->d_inode; /* we hold i_mutex */
S
Sage Weil 已提交
534 535 536

	/* .snap dir? */
	if (err == -ENOENT &&
537
	    ceph_snap(parent) == CEPH_NOSNAP &&
538
	    strcmp(dentry->d_name.name,
539
		   fsc->mount_options->snapdir_name) == 0) {
S
Sage Weil 已提交
540 541 542
		struct inode *inode = ceph_get_snapdir(parent);
		dout("ENOENT on snapdir %p '%.*s', linking to snapdir %p\n",
		     dentry, dentry->d_name.len, dentry->d_name.name, inode);
543
		BUG_ON(!d_unhashed(dentry));
S
Sage Weil 已提交
544 545 546
		d_add(dentry, inode);
		err = 0;
	}
547 548
	return err;
}
S
Sage Weil 已提交
549

550 551 552 553 554 555 556 557 558 559 560 561 562 563
/*
 * Figure out final result of a lookup/open request.
 *
 * Mainly, make sure we return the final req->r_dentry (if it already
 * existed) in place of the original VFS-provided dentry when they
 * differ.
 *
 * Gracefully handle the case where the MDS replies with -ENOENT and
 * no trace (which it may do, at its discretion, e.g., if it doesn't
 * care to issue a lease on the negative dentry).
 */
struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
				  struct dentry *dentry, int err)
{
S
Sage Weil 已提交
564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586
	if (err == -ENOENT) {
		/* no trace? */
		err = 0;
		if (!req->r_reply_info.head->is_dentry) {
			dout("ENOENT and no trace, dentry %p inode %p\n",
			     dentry, dentry->d_inode);
			if (dentry->d_inode) {
				d_drop(dentry);
				err = -ENOENT;
			} else {
				d_add(dentry, NULL);
			}
		}
	}
	if (err)
		dentry = ERR_PTR(err);
	else if (dentry != req->r_dentry)
		dentry = dget(req->r_dentry);   /* we got spliced */
	else
		dentry = NULL;
	return dentry;
}

S
Sage Weil 已提交
587 588 589 590 591 592
static int is_root_ceph_dentry(struct inode *inode, struct dentry *dentry)
{
	return ceph_ino(inode) == CEPH_INO_ROOT &&
		strncmp(dentry->d_name.name, ".ceph", 5) == 0;
}

S
Sage Weil 已提交
593 594 595 596 597
/*
 * Look up a single dir entry.  If there is a lookup intent, inform
 * the MDS so that it gets our 'caps wanted' value in a single op.
 */
static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
A
Al Viro 已提交
598
				  unsigned int flags)
S
Sage Weil 已提交
599
{
600 601
	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620
	struct ceph_mds_request *req;
	int op;
	int err;

	dout("lookup %p dentry %p '%.*s'\n",
	     dir, dentry, dentry->d_name.len, dentry->d_name.name);

	if (dentry->d_name.len > NAME_MAX)
		return ERR_PTR(-ENAMETOOLONG);

	err = ceph_init_dentry(dentry);
	if (err < 0)
		return ERR_PTR(err);

	/* can we conclude ENOENT locally? */
	if (dentry->d_inode == NULL) {
		struct ceph_inode_info *ci = ceph_inode(dir);
		struct ceph_dentry_info *di = ceph_dentry(dentry);

621
		spin_lock(&ci->i_ceph_lock);
S
Sage Weil 已提交
622 623
		dout(" dir %p flags are %d\n", dir, ci->i_ceph_flags);
		if (strncmp(dentry->d_name.name,
624
			    fsc->mount_options->snapdir_name,
S
Sage Weil 已提交
625
			    dentry->d_name.len) &&
S
Sage Weil 已提交
626
		    !is_root_ceph_dentry(dir, dentry) &&
627
		    __ceph_dir_is_complete(ci) &&
S
Sage Weil 已提交
628
		    (__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1))) {
629
			spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
630 631 632 633 634
			dout(" dir %p complete, -ENOENT\n", dir);
			d_add(dentry, NULL);
			di->lease_shared_gen = ci->i_shared_gen;
			return NULL;
		}
635
		spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
636 637 638 639 640 641
	}

	op = ceph_snap(dir) == CEPH_SNAPDIR ?
		CEPH_MDS_OP_LOOKUPSNAP : CEPH_MDS_OP_LOOKUP;
	req = ceph_mdsc_create_request(mdsc, op, USE_ANY_MDS);
	if (IS_ERR(req))
J
Julia Lawall 已提交
642
		return ERR_CAST(req);
S
Sage Weil 已提交
643 644 645 646 647 648
	req->r_dentry = dget(dentry);
	req->r_num_caps = 2;
	/* we only need inode linkage */
	req->r_args.getattr.mask = cpu_to_le32(CEPH_STAT_CAP_INODE);
	req->r_locked_dir = dir;
	err = ceph_mdsc_do_request(mdsc, NULL, req);
649
	err = ceph_handle_snapdir(req, dentry, err);
S
Sage Weil 已提交
650 651 652 653 654 655 656 657 658 659 660 661
	dentry = ceph_finish_lookup(req, dentry, err);
	ceph_mdsc_put_request(req);  /* will dput(dentry) */
	dout("lookup result=%p\n", dentry);
	return dentry;
}

/*
 * If we do a create but get no trace back from the MDS, follow up with
 * a lookup (the VFS expects us to link up the provided dentry).
 */
int ceph_handle_notrace_create(struct inode *dir, struct dentry *dentry)
{
A
Al Viro 已提交
662
	struct dentry *result = ceph_lookup(dir, dentry, 0);
S
Sage Weil 已提交
663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679

	if (result && !IS_ERR(result)) {
		/*
		 * We created the item, then did a lookup, and found
		 * it was already linked to another inode we already
		 * had in our cache (and thus got spliced).  Link our
		 * dentry to that inode, but don't hash it, just in
		 * case the VFS wants to dereference it.
		 */
		BUG_ON(!result->d_inode);
		d_instantiate(dentry, result->d_inode);
		return 0;
	}
	return PTR_ERR(result);
}

static int ceph_mknod(struct inode *dir, struct dentry *dentry,
A
Al Viro 已提交
680
		      umode_t mode, dev_t rdev)
S
Sage Weil 已提交
681
{
682 683
	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
684
	struct ceph_mds_request *req;
685
	struct ceph_acls_info acls = {};
S
Sage Weil 已提交
686 687 688 689 690
	int err;

	if (ceph_snap(dir) != CEPH_NOSNAP)
		return -EROFS;

691 692 693 694
	err = ceph_pre_init_acls(dir, &mode, &acls);
	if (err < 0)
		return err;

A
Al Viro 已提交
695
	dout("mknod in dir %p dentry %p mode 0%ho rdev %d\n",
S
Sage Weil 已提交
696 697 698
	     dir, dentry, mode, rdev);
	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_MKNOD, USE_AUTH_MDS);
	if (IS_ERR(req)) {
699 700
		err = PTR_ERR(req);
		goto out;
S
Sage Weil 已提交
701 702 703 704 705 706 707 708
	}
	req->r_dentry = dget(dentry);
	req->r_num_caps = 2;
	req->r_locked_dir = dir;
	req->r_args.mknod.mode = cpu_to_le32(mode);
	req->r_args.mknod.rdev = cpu_to_le32(rdev);
	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
709 710 711 712
	if (acls.pagelist) {
		req->r_pagelist = acls.pagelist;
		acls.pagelist = NULL;
	}
S
Sage Weil 已提交
713 714 715 716
	err = ceph_mdsc_do_request(mdsc, dir, req);
	if (!err && !req->r_reply_info.head->is_dentry)
		err = ceph_handle_notrace_create(dir, dentry);
	ceph_mdsc_put_request(req);
717
out:
G
Guangliang Zhao 已提交
718
	if (!err)
719
		ceph_init_inode_acls(dentry->d_inode, &acls);
720
	else
S
Sage Weil 已提交
721
		d_drop(dentry);
722
	ceph_release_acls_info(&acls);
S
Sage Weil 已提交
723 724 725
	return err;
}

A
Al Viro 已提交
726
static int ceph_create(struct inode *dir, struct dentry *dentry, umode_t mode,
A
Al Viro 已提交
727
		       bool excl)
S
Sage Weil 已提交
728
{
729
	return ceph_mknod(dir, dentry, mode, 0);
S
Sage Weil 已提交
730 731 732 733 734
}

static int ceph_symlink(struct inode *dir, struct dentry *dentry,
			    const char *dest)
{
735 736
	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
737 738 739 740 741 742 743 744 745
	struct ceph_mds_request *req;
	int err;

	if (ceph_snap(dir) != CEPH_NOSNAP)
		return -EROFS;

	dout("symlink in dir %p dentry %p to '%s'\n", dir, dentry, dest);
	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_SYMLINK, USE_AUTH_MDS);
	if (IS_ERR(req)) {
746 747
		err = PTR_ERR(req);
		goto out;
S
Sage Weil 已提交
748 749 750 751 752 753 754 755 756 757 758
	}
	req->r_dentry = dget(dentry);
	req->r_num_caps = 2;
	req->r_path2 = kstrdup(dest, GFP_NOFS);
	req->r_locked_dir = dir;
	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
	err = ceph_mdsc_do_request(mdsc, dir, req);
	if (!err && !req->r_reply_info.head->is_dentry)
		err = ceph_handle_notrace_create(dir, dentry);
	ceph_mdsc_put_request(req);
759 760
out:
	if (err)
S
Sage Weil 已提交
761 762 763 764
		d_drop(dentry);
	return err;
}

765
static int ceph_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
S
Sage Weil 已提交
766
{
767 768
	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
769
	struct ceph_mds_request *req;
770
	struct ceph_acls_info acls = {};
S
Sage Weil 已提交
771 772 773 774 775 776 777 778 779
	int err = -EROFS;
	int op;

	if (ceph_snap(dir) == CEPH_SNAPDIR) {
		/* mkdir .snap/foo is a MKSNAP */
		op = CEPH_MDS_OP_MKSNAP;
		dout("mksnap dir %p snap '%.*s' dn %p\n", dir,
		     dentry->d_name.len, dentry->d_name.name, dentry);
	} else if (ceph_snap(dir) == CEPH_NOSNAP) {
780
		dout("mkdir dir %p dn %p mode 0%ho\n", dir, dentry, mode);
S
Sage Weil 已提交
781 782 783 784
		op = CEPH_MDS_OP_MKDIR;
	} else {
		goto out;
	}
785 786 787 788 789 790

	mode |= S_IFDIR;
	err = ceph_pre_init_acls(dir, &mode, &acls);
	if (err < 0)
		goto out;

S
Sage Weil 已提交
791 792 793 794 795 796 797 798 799 800 801 802
	req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS);
	if (IS_ERR(req)) {
		err = PTR_ERR(req);
		goto out;
	}

	req->r_dentry = dget(dentry);
	req->r_num_caps = 2;
	req->r_locked_dir = dir;
	req->r_args.mkdir.mode = cpu_to_le32(mode);
	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
803 804 805 806
	if (acls.pagelist) {
		req->r_pagelist = acls.pagelist;
		acls.pagelist = NULL;
	}
S
Sage Weil 已提交
807 808 809 810 811
	err = ceph_mdsc_do_request(mdsc, dir, req);
	if (!err && !req->r_reply_info.head->is_dentry)
		err = ceph_handle_notrace_create(dir, dentry);
	ceph_mdsc_put_request(req);
out:
812
	if (!err)
813
		ceph_init_inode_acls(dentry->d_inode, &acls);
814
	else
S
Sage Weil 已提交
815
		d_drop(dentry);
816
	ceph_release_acls_info(&acls);
S
Sage Weil 已提交
817 818 819 820 821 822
	return err;
}

static int ceph_link(struct dentry *old_dentry, struct inode *dir,
		     struct dentry *dentry)
{
823 824
	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
825 826 827 828 829 830 831 832 833 834 835 836 837 838 839
	struct ceph_mds_request *req;
	int err;

	if (ceph_snap(dir) != CEPH_NOSNAP)
		return -EROFS;

	dout("link in dir %p old_dentry %p dentry %p\n", dir,
	     old_dentry, dentry);
	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_LINK, USE_AUTH_MDS);
	if (IS_ERR(req)) {
		d_drop(dentry);
		return PTR_ERR(req);
	}
	req->r_dentry = dget(dentry);
	req->r_num_caps = 2;
840
	req->r_old_dentry = dget(old_dentry);
S
Sage Weil 已提交
841 842 843
	req->r_locked_dir = dir;
	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
844 845
	/* release LINK_SHARED on source inode (mds will lock it) */
	req->r_old_inode_drop = CEPH_CAP_LINK_SHARED;
S
Sage Weil 已提交
846
	err = ceph_mdsc_do_request(mdsc, dir, req);
847
	if (err) {
S
Sage Weil 已提交
848
		d_drop(dentry);
849 850 851 852
	} else if (!req->r_reply_info.head->is_dentry) {
		ihold(old_dentry->d_inode);
		d_instantiate(dentry, old_dentry->d_inode);
	}
S
Sage Weil 已提交
853 854 855 856 857 858 859 860 861 862 863 864 865 866 867
	ceph_mdsc_put_request(req);
	return err;
}

/*
 * For a soon-to-be unlinked file, drop the AUTH_RDCACHE caps.  If it
 * looks like the link count will hit 0, drop any other caps (other
 * than PIN) we don't specifically want (due to the file still being
 * open).
 */
static int drop_caps_for_unlink(struct inode *inode)
{
	struct ceph_inode_info *ci = ceph_inode(inode);
	int drop = CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL;

868
	spin_lock(&ci->i_ceph_lock);
S
Sage Weil 已提交
869 870 871 872
	if (inode->i_nlink == 1) {
		drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN);
		ci->i_ceph_flags |= CEPH_I_NODELAY;
	}
873
	spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
874 875 876 877 878 879 880 881
	return drop;
}

/*
 * rmdir and unlink are differ only by the metadata op code
 */
static int ceph_unlink(struct inode *dir, struct dentry *dentry)
{
882 883
	struct ceph_fs_client *fsc = ceph_sb_to_client(dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
884 885 886 887 888 889 890 891 892 893 894 895 896
	struct inode *inode = dentry->d_inode;
	struct ceph_mds_request *req;
	int err = -EROFS;
	int op;

	if (ceph_snap(dir) == CEPH_SNAPDIR) {
		/* rmdir .snap/foo is RMSNAP */
		dout("rmsnap dir %p '%.*s' dn %p\n", dir, dentry->d_name.len,
		     dentry->d_name.name, dentry);
		op = CEPH_MDS_OP_RMSNAP;
	} else if (ceph_snap(dir) == CEPH_NOSNAP) {
		dout("unlink/rmdir dir %p dn %p inode %p\n",
		     dir, dentry, inode);
A
Al Viro 已提交
897
		op = S_ISDIR(dentry->d_inode->i_mode) ?
S
Sage Weil 已提交
898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922
			CEPH_MDS_OP_RMDIR : CEPH_MDS_OP_UNLINK;
	} else
		goto out;
	req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS);
	if (IS_ERR(req)) {
		err = PTR_ERR(req);
		goto out;
	}
	req->r_dentry = dget(dentry);
	req->r_num_caps = 2;
	req->r_locked_dir = dir;
	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
	req->r_inode_drop = drop_caps_for_unlink(inode);
	err = ceph_mdsc_do_request(mdsc, dir, req);
	if (!err && !req->r_reply_info.head->is_dentry)
		d_delete(dentry);
	ceph_mdsc_put_request(req);
out:
	return err;
}

static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry,
		       struct inode *new_dir, struct dentry *new_dentry)
{
923 924
	struct ceph_fs_client *fsc = ceph_sb_to_client(old_dir->i_sb);
	struct ceph_mds_client *mdsc = fsc->mdsc;
S
Sage Weil 已提交
925 926 927 928 929 930 931 932 933 934 935 936 937
	struct ceph_mds_request *req;
	int err;

	if (ceph_snap(old_dir) != ceph_snap(new_dir))
		return -EXDEV;
	if (ceph_snap(old_dir) != CEPH_NOSNAP ||
	    ceph_snap(new_dir) != CEPH_NOSNAP)
		return -EROFS;
	dout("rename dir %p dentry %p to dir %p dentry %p\n",
	     old_dir, old_dentry, new_dir, new_dentry);
	req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_RENAME, USE_AUTH_MDS);
	if (IS_ERR(req))
		return PTR_ERR(req);
938
	ihold(old_dir);
S
Sage Weil 已提交
939 940 941
	req->r_dentry = dget(new_dentry);
	req->r_num_caps = 2;
	req->r_old_dentry = dget(old_dentry);
942
	req->r_old_dentry_dir = old_dir;
S
Sage Weil 已提交
943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958
	req->r_locked_dir = new_dir;
	req->r_old_dentry_drop = CEPH_CAP_FILE_SHARED;
	req->r_old_dentry_unless = CEPH_CAP_FILE_EXCL;
	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
	/* release LINK_RDCACHE on source inode (mds will lock it) */
	req->r_old_inode_drop = CEPH_CAP_LINK_SHARED;
	if (new_dentry->d_inode)
		req->r_inode_drop = drop_caps_for_unlink(new_dentry->d_inode);
	err = ceph_mdsc_do_request(mdsc, old_dir, req);
	if (!err && !req->r_reply_info.head->is_dentry) {
		/*
		 * Normally d_move() is done by fill_trace (called by
		 * do_request, above).  If there is no trace, we need
		 * to do it here.
		 */
959

S
Sage Weil 已提交
960
		d_move(old_dentry, new_dentry);
961 962 963

		/* ensure target dentry is invalidated, despite
		   rehashing bug in vfs_rename_dir */
964
		ceph_invalidate_dentry_lease(new_dentry);
965 966 967 968 969

		/* d_move screws up sibling dentries' offsets */
		ceph_dir_clear_complete(old_dir);
		ceph_dir_clear_complete(new_dir);

S
Sage Weil 已提交
970 971 972 973 974
	}
	ceph_mdsc_put_request(req);
	return err;
}

975 976 977 978 979 980 981 982 983 984
/*
 * Ensure a dentry lease will no longer revalidate.
 */
void ceph_invalidate_dentry_lease(struct dentry *dentry)
{
	spin_lock(&dentry->d_lock);
	dentry->d_time = jiffies;
	ceph_dentry(dentry)->lease_shared_gen = 0;
	spin_unlock(&dentry->d_lock);
}
S
Sage Weil 已提交
985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002

/*
 * Check if dentry lease is valid.  If not, delete the lease.  Try to
 * renew if the least is more than half up.
 */
static int dentry_lease_is_valid(struct dentry *dentry)
{
	struct ceph_dentry_info *di;
	struct ceph_mds_session *s;
	int valid = 0;
	u32 gen;
	unsigned long ttl;
	struct ceph_mds_session *session = NULL;
	struct inode *dir = NULL;
	u32 seq = 0;

	spin_lock(&dentry->d_lock);
	di = ceph_dentry(dentry);
1003
	if (di->lease_session) {
S
Sage Weil 已提交
1004
		s = di->lease_session;
1005
		spin_lock(&s->s_gen_ttl_lock);
S
Sage Weil 已提交
1006 1007
		gen = s->s_cap_gen;
		ttl = s->s_cap_ttl;
1008
		spin_unlock(&s->s_gen_ttl_lock);
S
Sage Weil 已提交
1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044

		if (di->lease_gen == gen &&
		    time_before(jiffies, dentry->d_time) &&
		    time_before(jiffies, ttl)) {
			valid = 1;
			if (di->lease_renew_after &&
			    time_after(jiffies, di->lease_renew_after)) {
				/* we should renew */
				dir = dentry->d_parent->d_inode;
				session = ceph_get_mds_session(s);
				seq = di->lease_seq;
				di->lease_renew_after = 0;
				di->lease_renew_from = jiffies;
			}
		}
	}
	spin_unlock(&dentry->d_lock);

	if (session) {
		ceph_mdsc_lease_send_msg(session, dir, dentry,
					 CEPH_MDS_LEASE_RENEW, seq);
		ceph_put_mds_session(session);
	}
	dout("dentry_lease_is_valid - dentry %p = %d\n", dentry, valid);
	return valid;
}

/*
 * Check if directory-wide content lease/cap is valid.
 */
static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry)
{
	struct ceph_inode_info *ci = ceph_inode(dir);
	struct ceph_dentry_info *di = ceph_dentry(dentry);
	int valid = 0;

1045
	spin_lock(&ci->i_ceph_lock);
S
Sage Weil 已提交
1046 1047
	if (ci->i_shared_gen == di->lease_shared_gen)
		valid = __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1);
1048
	spin_unlock(&ci->i_ceph_lock);
S
Sage Weil 已提交
1049 1050 1051 1052 1053 1054 1055 1056 1057
	dout("dir_lease_is_valid dir %p v%u dentry %p v%u = %d\n",
	     dir, (unsigned)ci->i_shared_gen, dentry,
	     (unsigned)di->lease_shared_gen, valid);
	return valid;
}

/*
 * Check if cached dentry can be trusted.
 */
1058
static int ceph_d_revalidate(struct dentry *dentry, unsigned int flags)
S
Sage Weil 已提交
1059
{
1060
	int valid = 0;
1061 1062
	struct inode *dir;

1063
	if (flags & LOOKUP_RCU)
1064 1065
		return -ECHILD;

S
Sage Weil 已提交
1066 1067 1068
	dout("d_revalidate %p '%.*s' inode %p offset %lld\n", dentry,
	     dentry->d_name.len, dentry->d_name.name, dentry->d_inode,
	     ceph_dentry(dentry)->offset);
S
Sage Weil 已提交
1069

1070 1071
	dir = ceph_get_dentry_parent_inode(dentry);

S
Sage Weil 已提交
1072 1073 1074 1075
	/* always trust cached snapped dentries, snapdir dentry */
	if (ceph_snap(dir) != CEPH_NOSNAP) {
		dout("d_revalidate %p '%.*s' inode %p is SNAPPED\n", dentry,
		     dentry->d_name.len, dentry->d_name.name, dentry->d_inode);
1076 1077 1078 1079 1080 1081
		valid = 1;
	} else if (dentry->d_inode &&
		   ceph_snap(dentry->d_inode) == CEPH_SNAPDIR) {
		valid = 1;
	} else if (dentry_lease_is_valid(dentry) ||
		   dir_lease_is_valid(dir, dentry)) {
1082 1083 1084 1085
		if (dentry->d_inode)
			valid = ceph_is_any_caps(dentry->d_inode);
		else
			valid = 1;
S
Sage Weil 已提交
1086 1087
	}

1088
	dout("d_revalidate %p %s\n", dentry, valid ? "valid" : "invalid");
1089
	if (valid) {
1090
		ceph_dentry_lru_touch(dentry);
1091 1092 1093
	} else {
		ceph_dir_clear_complete(dir);
	}
1094 1095
	iput(dir);
	return valid;
S
Sage Weil 已提交
1096 1097 1098
}

/*
1099
 * Release our ceph_dentry_info.
S
Sage Weil 已提交
1100
 */
1101
static void ceph_d_release(struct dentry *dentry)
S
Sage Weil 已提交
1102 1103 1104
{
	struct ceph_dentry_info *di = ceph_dentry(dentry);

1105
	dout("d_release %p\n", dentry);
1106 1107 1108 1109 1110
	ceph_dentry_lru_del(dentry);
	if (di->lease_session)
		ceph_put_mds_session(di->lease_session);
	kmem_cache_free(ceph_dentry_cachep, di);
	dentry->d_fsdata = NULL;
S
Sage Weil 已提交
1111 1112 1113
}

static int ceph_snapdir_d_revalidate(struct dentry *dentry,
1114
					  unsigned int flags)
S
Sage Weil 已提交
1115 1116 1117 1118 1119 1120 1121 1122
{
	/*
	 * Eventually, we'll want to revalidate snapped metadata
	 * too... probably...
	 */
	return 1;
}

1123 1124 1125 1126 1127 1128 1129 1130
/*
 * When the VFS prunes a dentry from the cache, we need to clear the
 * complete flag on the parent directory.
 *
 * Called under dentry->d_lock.
 */
static void ceph_d_prune(struct dentry *dentry)
{
S
Sage Weil 已提交
1131
	dout("ceph_d_prune %p\n", dentry);
1132 1133

	/* do we have a valid parent? */
1134
	if (IS_ROOT(dentry))
1135 1136
		return;

1137
	/* if we are not hashed, we don't affect dir's completeness */
1138 1139
	if (d_unhashed(dentry))
		return;
S
Sage Weil 已提交
1140

1141 1142 1143 1144
	/*
	 * we hold d_lock, so d_parent is stable, and d_fsdata is never
	 * cleared until d_release
	 */
1145
	ceph_dir_clear_complete(dentry->d_parent->d_inode);
1146
}
S
Sage Weil 已提交
1147 1148 1149 1150 1151 1152 1153 1154 1155

/*
 * read() on a dir.  This weird interface hack only works if mounted
 * with '-o dirstat'.
 */
static ssize_t ceph_read_dir(struct file *file, char __user *buf, size_t size,
			     loff_t *ppos)
{
	struct ceph_file_info *cf = file->private_data;
A
Al Viro 已提交
1156
	struct inode *inode = file_inode(file);
S
Sage Weil 已提交
1157 1158
	struct ceph_inode_info *ci = ceph_inode(inode);
	int left;
1159
	const int bufsize = 1024;
S
Sage Weil 已提交
1160

1161
	if (!ceph_test_mount_opt(ceph_sb_to_client(inode->i_sb), DIRSTAT))
S
Sage Weil 已提交
1162 1163 1164
		return -EISDIR;

	if (!cf->dir_info) {
1165
		cf->dir_info = kmalloc(bufsize, GFP_NOFS);
S
Sage Weil 已提交
1166 1167 1168
		if (!cf->dir_info)
			return -ENOMEM;
		cf->dir_info_len =
1169
			snprintf(cf->dir_info, bufsize,
S
Sage Weil 已提交
1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
				"entries:   %20lld\n"
				" files:    %20lld\n"
				" subdirs:  %20lld\n"
				"rentries:  %20lld\n"
				" rfiles:   %20lld\n"
				" rsubdirs: %20lld\n"
				"rbytes:    %20lld\n"
				"rctime:    %10ld.%09ld\n",
				ci->i_files + ci->i_subdirs,
				ci->i_files,
				ci->i_subdirs,
				ci->i_rfiles + ci->i_rsubdirs,
				ci->i_rfiles,
				ci->i_rsubdirs,
				ci->i_rbytes,
				(long)ci->i_rctime.tv_sec,
				(long)ci->i_rctime.tv_nsec);
	}

	if (*ppos >= cf->dir_info_len)
		return 0;
	size = min_t(unsigned, size, cf->dir_info_len-*ppos);
	left = copy_to_user(buf, cf->dir_info + *ppos, size);
	if (left == size)
		return -EFAULT;
	*ppos += (size - left);
	return size - left;
}

/*
 * an fsync() on a dir will wait for any uncommitted directory
 * operations to commit.
 */
1203 1204
static int ceph_dir_fsync(struct file *file, loff_t start, loff_t end,
			  int datasync)
S
Sage Weil 已提交
1205
{
A
Al Viro 已提交
1206
	struct inode *inode = file_inode(file);
S
Sage Weil 已提交
1207 1208 1209 1210 1211 1212 1213
	struct ceph_inode_info *ci = ceph_inode(inode);
	struct list_head *head = &ci->i_unsafe_dirops;
	struct ceph_mds_request *req;
	u64 last_tid;
	int ret = 0;

	dout("dir_fsync %p\n", inode);
1214 1215 1216 1217 1218
	ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
	if (ret)
		return ret;
	mutex_lock(&inode->i_mutex);

S
Sage Weil 已提交
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
	spin_lock(&ci->i_unsafe_lock);
	if (list_empty(head))
		goto out;

	req = list_entry(head->prev,
			 struct ceph_mds_request, r_unsafe_dir_item);
	last_tid = req->r_tid;

	do {
		ceph_mdsc_get_request(req);
		spin_unlock(&ci->i_unsafe_lock);
1230

S
Sage Weil 已提交
1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244
		dout("dir_fsync %p wait on tid %llu (until %llu)\n",
		     inode, req->r_tid, last_tid);
		if (req->r_timeout) {
			ret = wait_for_completion_timeout(
				&req->r_safe_completion, req->r_timeout);
			if (ret > 0)
				ret = 0;
			else if (ret == 0)
				ret = -EIO;  /* timed out */
		} else {
			wait_for_completion(&req->r_safe_completion);
		}
		ceph_mdsc_put_request(req);

1245
		spin_lock(&ci->i_unsafe_lock);
S
Sage Weil 已提交
1246 1247 1248 1249 1250 1251 1252
		if (ret || list_empty(head))
			break;
		req = list_entry(head->next,
				 struct ceph_mds_request, r_unsafe_dir_item);
	} while (req->r_tid < last_tid);
out:
	spin_unlock(&ci->i_unsafe_lock);
1253 1254
	mutex_unlock(&inode->i_mutex);

S
Sage Weil 已提交
1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267
	return ret;
}

/*
 * We maintain a private dentry LRU.
 *
 * FIXME: this needs to be changed to a per-mds lru to be useful.
 */
void ceph_dentry_lru_add(struct dentry *dn)
{
	struct ceph_dentry_info *di = ceph_dentry(dn);
	struct ceph_mds_client *mdsc;

1268 1269
	dout("dentry_lru_add %p %p '%.*s'\n", di, dn,
	     dn->d_name.len, dn->d_name.name);
1270 1271 1272 1273 1274
	mdsc = ceph_sb_to_client(dn->d_sb)->mdsc;
	spin_lock(&mdsc->dentry_lru_lock);
	list_add_tail(&di->lru, &mdsc->dentry_lru);
	mdsc->num_dentry++;
	spin_unlock(&mdsc->dentry_lru_lock);
S
Sage Weil 已提交
1275 1276 1277 1278 1279 1280 1281
}

void ceph_dentry_lru_touch(struct dentry *dn)
{
	struct ceph_dentry_info *di = ceph_dentry(dn);
	struct ceph_mds_client *mdsc;

S
Sage Weil 已提交
1282 1283
	dout("dentry_lru_touch %p %p '%.*s' (offset %lld)\n", di, dn,
	     dn->d_name.len, dn->d_name.name, di->offset);
1284 1285 1286 1287
	mdsc = ceph_sb_to_client(dn->d_sb)->mdsc;
	spin_lock(&mdsc->dentry_lru_lock);
	list_move_tail(&di->lru, &mdsc->dentry_lru);
	spin_unlock(&mdsc->dentry_lru_lock);
S
Sage Weil 已提交
1288 1289 1290 1291 1292 1293 1294
}

void ceph_dentry_lru_del(struct dentry *dn)
{
	struct ceph_dentry_info *di = ceph_dentry(dn);
	struct ceph_mds_client *mdsc;

1295 1296
	dout("dentry_lru_del %p %p '%.*s'\n", di, dn,
	     dn->d_name.len, dn->d_name.name);
1297 1298 1299 1300 1301
	mdsc = ceph_sb_to_client(dn->d_sb)->mdsc;
	spin_lock(&mdsc->dentry_lru_lock);
	list_del_init(&di->lru);
	mdsc->num_dentry--;
	spin_unlock(&mdsc->dentry_lru_lock);
S
Sage Weil 已提交
1302 1303
}

S
Sage Weil 已提交
1304 1305 1306 1307
/*
 * Return name hash for a given dentry.  This is dependent on
 * the parent directory's hash function.
 */
1308
unsigned ceph_dentry_hash(struct inode *dir, struct dentry *dn)
S
Sage Weil 已提交
1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322
{
	struct ceph_inode_info *dci = ceph_inode(dir);

	switch (dci->i_dir_layout.dl_dir_hash) {
	case 0:	/* for backward compat */
	case CEPH_STR_HASH_LINUX:
		return dn->d_name.hash;

	default:
		return ceph_str_hash(dci->i_dir_layout.dl_dir_hash,
				     dn->d_name.name, dn->d_name.len);
	}
}

S
Sage Weil 已提交
1323 1324
const struct file_operations ceph_dir_fops = {
	.read = ceph_read_dir,
A
Al Viro 已提交
1325
	.iterate = ceph_readdir,
S
Sage Weil 已提交
1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341
	.llseek = ceph_dir_llseek,
	.open = ceph_open,
	.release = ceph_release,
	.unlocked_ioctl = ceph_ioctl,
	.fsync = ceph_dir_fsync,
};

const struct inode_operations ceph_dir_iops = {
	.lookup = ceph_lookup,
	.permission = ceph_permission,
	.getattr = ceph_getattr,
	.setattr = ceph_setattr,
	.setxattr = ceph_setxattr,
	.getxattr = ceph_getxattr,
	.listxattr = ceph_listxattr,
	.removexattr = ceph_removexattr,
G
Guangliang Zhao 已提交
1342
	.get_acl = ceph_get_acl,
S
Sage Weil 已提交
1343
	.set_acl = ceph_set_acl,
S
Sage Weil 已提交
1344 1345 1346 1347 1348 1349 1350 1351
	.mknod = ceph_mknod,
	.symlink = ceph_symlink,
	.mkdir = ceph_mkdir,
	.link = ceph_link,
	.unlink = ceph_unlink,
	.rmdir = ceph_unlink,
	.rename = ceph_rename,
	.create = ceph_create,
1352
	.atomic_open = ceph_atomic_open,
S
Sage Weil 已提交
1353 1354
};

S
Sage Weil 已提交
1355
const struct dentry_operations ceph_dentry_ops = {
S
Sage Weil 已提交
1356
	.d_revalidate = ceph_d_revalidate,
1357
	.d_release = ceph_d_release,
1358
	.d_prune = ceph_d_prune,
S
Sage Weil 已提交
1359 1360
};

S
Sage Weil 已提交
1361
const struct dentry_operations ceph_snapdir_dentry_ops = {
S
Sage Weil 已提交
1362
	.d_revalidate = ceph_snapdir_d_revalidate,
1363
	.d_release = ceph_d_release,
S
Sage Weil 已提交
1364 1365
};

S
Sage Weil 已提交
1366
const struct dentry_operations ceph_snap_dentry_ops = {
1367
	.d_release = ceph_d_release,
1368
	.d_prune = ceph_d_prune,
S
Sage Weil 已提交
1369
};