diff options
author | Linus Torvalds | 2019-05-07 20:51:58 -0700 |
---|---|---|
committer | Linus Torvalds | 2019-05-07 20:51:58 -0700 |
commit | e5fef2a9732580c5bd30c0097f5e9091a3d58ce5 (patch) | |
tree | cb91b5a51f7e451300e06e203a50cbed8aabbc1c | |
parent | 149e703cb8bfcbdae46140b108bb6f7d2407df8f (diff) | |
parent | f5e4546347bc847be30b3cf904db5fc874b3c5dc (diff) |
Merge tag 'afs-next-20190507' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs
Pull AFS updates from David Howells:
"A set of fix and development patches for AFS for 5.2.
Summary:
- Fix the AFS file locking so that sqlite can run on an AFS mount and
also so that firefox and gnome can use a homedir that's mounted
through AFS.
This required emulation of fine-grained locking when the server
will only support whole-file locks and no upgrade/downgrade. Four
modes are provided, settable by mount parameter:
"flock=local" - No reference to the server
"flock=openafs" - Fine-grained locks are local-only, whole-file
locks require sufficient server locks
"flock=strict" - All locks require sufficient server locks
"flock=write" - Always get an exclusive server lock
If the volume is a read-only or backup volume, then flock=local for
that volume.
- Log extra information for a couple of cases where the client mucks
up somehow: AFS vnode with undefined type and dir check failure -
in both cases we seem to end up with unfilled data, but the issues
happen infrequently and are difficult to reproduce at will.
- Implement silly rename for unlink() and rename().
- Set i_blocks so that du can get some information about usage.
- Fix xattr handlers to return the right amount of data and to not
overflow buffers.
- Implement getting/setting raw AFS and YFS ACLs as xattrs"
* tag 'afs-next-20190507' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
afs: Implement YFS ACL setting
afs: Get YFS ACLs and information through xattrs
afs: implement acl setting
afs: Get an AFS3 ACL as an xattr
afs: Fix getting the afs.fid xattr
afs: Fix the afs.cell and afs.volume xattr handlers
afs: Calculate i_blocks based on file size
afs: Log more information for "kAFS: AFS vnode with undefined type\n"
afs: Provide mount-time configurable byte-range file locking emulation
afs: Add more tracepoints
afs: Implement sillyrename for unlink and rename
afs: Add directory reload tracepoint
afs: Handle lock rpc ops failing on a file that got deleted
afs: Improve dir check failure reports
afs: Add file locking tracepoints
afs: Further fix file locking
afs: Fix AFS file locking to allow fine grained locks
afs: Calculate lock extend timer from set/extend reply reception
afs: Split wait from afs_make_call()
-rw-r--r-- | fs/afs/Makefile | 1 | ||||
-rw-r--r-- | fs/afs/afs_fs.h | 2 | ||||
-rw-r--r-- | fs/afs/dir.c | 167 | ||||
-rw-r--r-- | fs/afs/dir_silly.c | 239 | ||||
-rw-r--r-- | fs/afs/flock.c | 616 | ||||
-rw-r--r-- | fs/afs/fs_probe.c | 13 | ||||
-rw-r--r-- | fs/afs/fsclient.c | 277 | ||||
-rw-r--r-- | fs/afs/inode.c | 43 | ||||
-rw-r--r-- | fs/afs/internal.h | 64 | ||||
-rw-r--r-- | fs/afs/protocol_yfs.h | 6 | ||||
-rw-r--r-- | fs/afs/rxrpc.c | 33 | ||||
-rw-r--r-- | fs/afs/super.c | 34 | ||||
-rw-r--r-- | fs/afs/vl_probe.c | 14 | ||||
-rw-r--r-- | fs/afs/vlclient.c | 26 | ||||
-rw-r--r-- | fs/afs/xattr.c | 270 | ||||
-rw-r--r-- | fs/afs/yfsclient.c | 329 | ||||
-rw-r--r-- | include/linux/fs.h | 1 | ||||
-rw-r--r-- | include/trace/events/afs.h | 385 |
18 files changed, 2142 insertions, 378 deletions
diff --git a/fs/afs/Makefile b/fs/afs/Makefile index 0738e2bf5193..cbf31f6cd177 100644 --- a/fs/afs/Makefile +++ b/fs/afs/Makefile @@ -13,6 +13,7 @@ kafs-y := \ cmservice.o \ dir.o \ dir_edit.o \ + dir_silly.o \ dynroot.o \ file.o \ flock.o \ diff --git a/fs/afs/afs_fs.h b/fs/afs/afs_fs.h index ddfa88a7a9c0..18a54ca422f8 100644 --- a/fs/afs/afs_fs.h +++ b/fs/afs/afs_fs.h @@ -17,8 +17,10 @@ enum AFS_FS_Operations { FSFETCHDATA = 130, /* AFS Fetch file data */ + FSFETCHACL = 131, /* AFS Fetch file ACL */ FSFETCHSTATUS = 132, /* AFS Fetch file status */ FSSTOREDATA = 133, /* AFS Store file data */ + FSSTOREACL = 134, /* AFS Store file ACL */ FSSTORESTATUS = 135, /* AFS Store file status */ FSREMOVEFILE = 136, /* AFS Remove a file */ FSCREATEFILE = 137, /* AFS Create a file */ diff --git a/fs/afs/dir.c b/fs/afs/dir.c index 8a2562e3a316..9a466be583d2 100644 --- a/fs/afs/dir.c +++ b/fs/afs/dir.c @@ -26,6 +26,7 @@ static int afs_dir_open(struct inode *inode, struct file *file); static int afs_readdir(struct file *file, struct dir_context *ctx); static int afs_d_revalidate(struct dentry *dentry, unsigned int flags); static int afs_d_delete(const struct dentry *dentry); +static void afs_d_iput(struct dentry *dentry, struct inode *inode); static int afs_lookup_one_filldir(struct dir_context *ctx, const char *name, int nlen, loff_t fpos, u64 ino, unsigned dtype); static int afs_lookup_filldir(struct dir_context *ctx, const char *name, int nlen, @@ -85,6 +86,7 @@ const struct dentry_operations afs_fs_dentry_operations = { .d_delete = afs_d_delete, .d_release = afs_d_release, .d_automount = afs_d_automount, + .d_iput = afs_d_iput, }; struct afs_lookup_one_cookie { @@ -160,6 +162,38 @@ error: } /* + * Check the contents of a directory that we've just read. + */ +static bool afs_dir_check_pages(struct afs_vnode *dvnode, struct afs_read *req) +{ + struct afs_xdr_dir_page *dbuf; + unsigned int i, j, qty = PAGE_SIZE / sizeof(union afs_xdr_dir_block); + + for (i = 0; i < req->nr_pages; i++) + if (!afs_dir_check_page(dvnode, req->pages[i], req->actual_len)) + goto bad; + return true; + +bad: + pr_warn("DIR %llx:%llx f=%llx l=%llx al=%llx r=%llx\n", + dvnode->fid.vid, dvnode->fid.vnode, + req->file_size, req->len, req->actual_len, req->remain); + pr_warn("DIR %llx %x %x %x\n", + req->pos, req->index, req->nr_pages, req->offset); + + for (i = 0; i < req->nr_pages; i++) { + dbuf = kmap(req->pages[i]); + for (j = 0; j < qty; j++) { + union afs_xdr_dir_block *block = &dbuf->blocks[j]; + + pr_warn("[%02x] %32phN\n", i * qty + j, block); + } + kunmap(req->pages[i]); + } + return false; +} + +/* * open an AFS directory file */ static int afs_dir_open(struct inode *inode, struct file *file) @@ -277,6 +311,7 @@ retry: goto error; if (!test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags)) { + trace_afs_reload_dir(dvnode); ret = afs_fetch_data(dvnode, key, req); if (ret < 0) goto error_unlock; @@ -288,10 +323,8 @@ retry: /* Validate the data we just read. */ ret = -EIO; - for (i = 0; i < req->nr_pages; i++) - if (!afs_dir_check_page(dvnode, req->pages[i], - req->actual_len)) - goto error_unlock; + if (!afs_dir_check_pages(dvnode, req)) + goto error_unlock; // TODO: Trim excess pages @@ -743,7 +776,7 @@ success: ti = afs_iget(dir->i_sb, key, &cookie->fids[i], &cookie->statuses[i], &cookie->callbacks[i], - cbi); + cbi, dvnode); if (i == 0) { inode = ti; } else { @@ -875,8 +908,14 @@ static struct dentry *afs_lookup(struct inode *dir, struct dentry *dentry, (void *)(unsigned long)dvnode->status.data_version; } d = d_splice_alias(inode, dentry); - if (!IS_ERR_OR_NULL(d)) + if (!IS_ERR_OR_NULL(d)) { d->d_fsdata = dentry->d_fsdata; + trace_afs_lookup(dvnode, &d->d_name, + inode ? AFS_FS_I(inode) : NULL); + } else { + trace_afs_lookup(dvnode, &dentry->d_name, + inode ? AFS_FS_I(inode) : NULL); + } return d; } @@ -1053,6 +1092,16 @@ zap: } /* + * Clean up sillyrename files on dentry removal. + */ +static void afs_d_iput(struct dentry *dentry, struct inode *inode) +{ + if (dentry->d_flags & DCACHE_NFSFS_RENAMED) + afs_silly_iput(dentry, inode); + iput(inode); +} + +/* * handle dentry release */ void afs_d_release(struct dentry *dentry) @@ -1076,7 +1125,7 @@ static void afs_vnode_new_inode(struct afs_fs_cursor *fc, return; inode = afs_iget(fc->vnode->vfs_inode.i_sb, fc->key, - newfid, newstatus, newcb, fc->cbi); + newfid, newstatus, newcb, fc->cbi, fc->vnode); if (IS_ERR(inode)) { /* ENOMEM or EINTR at a really inconvenient time - just abandon * the new directory on the server. @@ -1194,6 +1243,12 @@ static int afs_rmdir(struct inode *dir, struct dentry *dentry) goto error_key; } + if (vnode) { + ret = down_write_killable(&vnode->rmdir_lock); + if (ret < 0) + goto error_key; + } + ret = -ERESTARTSYS; if (afs_begin_vnode_operation(&fc, dvnode, key)) { while (afs_select_fileserver(&fc)) { @@ -1212,6 +1267,8 @@ static int afs_rmdir(struct inode *dir, struct dentry *dentry) } } + if (vnode) + up_write(&vnode->rmdir_lock); error_key: key_put(key); error: @@ -1228,9 +1285,9 @@ error: * However, if we didn't have a callback promise outstanding, or it was * outstanding on a different server, then it won't break it either... */ -static int afs_dir_remove_link(struct dentry *dentry, struct key *key, - unsigned long d_version_before, - unsigned long d_version_after) +int afs_dir_remove_link(struct dentry *dentry, struct key *key, + unsigned long d_version_before, + unsigned long d_version_after) { bool dir_valid; int ret = 0; @@ -1277,6 +1334,7 @@ static int afs_unlink(struct inode *dir, struct dentry *dentry) struct afs_vnode *dvnode = AFS_FS_I(dir), *vnode = NULL; struct key *key; unsigned long d_version = (unsigned long)dentry->d_fsdata; + bool need_rehash = false; u64 data_version = dvnode->status.data_version; int ret; @@ -1300,6 +1358,21 @@ static int afs_unlink(struct inode *dir, struct dentry *dentry) goto error_key; } + spin_lock(&dentry->d_lock); + if (vnode && d_count(dentry) > 1) { + spin_unlock(&dentry->d_lock); + /* Start asynchronous writeout of the inode */ + write_inode_now(d_inode(dentry), 0); + ret = afs_sillyrename(dvnode, vnode, dentry, key); + goto error_key; + } + if (!d_unhashed(dentry)) { + /* Prevent a race with RCU lookup. */ + __d_drop(dentry); + need_rehash = true; + } + spin_unlock(&dentry->d_lock); + ret = -ERESTARTSYS; if (afs_begin_vnode_operation(&fc, dvnode, key)) { while (afs_select_fileserver(&fc)) { @@ -1331,6 +1404,9 @@ static int afs_unlink(struct inode *dir, struct dentry *dentry) afs_edit_dir_for_unlink); } + if (need_rehash && ret < 0 && ret != -ENOENT) + d_rehash(dentry); + error_key: key_put(key); error: @@ -1551,6 +1627,8 @@ static int afs_rename(struct inode *old_dir, struct dentry *old_dentry, { struct afs_fs_cursor fc; struct afs_vnode *orig_dvnode, *new_dvnode, *vnode; + struct dentry *tmp = NULL, *rehash = NULL; + struct inode *new_inode; struct key *key; u64 orig_data_version, new_data_version; bool new_negative = d_is_negative(new_dentry); @@ -1559,6 +1637,10 @@ static int afs_rename(struct inode *old_dir, struct dentry *old_dentry, if (flags) return -EINVAL; + /* Don't allow silly-rename files be moved around. */ + if (old_dentry->d_flags & DCACHE_NFSFS_RENAMED) + return -EINVAL; + vnode = AFS_FS_I(d_inode(old_dentry)); orig_dvnode = AFS_FS_I(old_dir); new_dvnode = AFS_FS_I(new_dir); @@ -1577,12 +1659,48 @@ static int afs_rename(struct inode *old_dir, struct dentry *old_dentry, goto error; } + /* For non-directories, check whether the target is busy and if so, + * make a copy of the dentry and then do a silly-rename. If the + * silly-rename succeeds, the copied dentry is hashed and becomes the + * new target. + */ + if (d_is_positive(new_dentry) && !d_is_dir(new_dentry)) { + /* To prevent any new references to the target during the + * rename, we unhash the dentry in advance. + */ + if (!d_unhashed(new_dentry)) { + d_drop(new_dentry); + rehash = new_dentry; + } + + if (d_count(new_dentry) > 2) { + /* copy the target dentry's name */ + ret = -ENOMEM; + tmp = d_alloc(new_dentry->d_parent, + &new_dentry->d_name); + if (!tmp) + goto error_rehash; + + ret = afs_sillyrename(new_dvnode, + AFS_FS_I(d_inode(new_dentry)), + new_dentry, key); + if (ret) + goto error_rehash; + + new_dentry = tmp; + rehash = NULL; + new_negative = true; + orig_data_version = orig_dvnode->status.data_version; + new_data_version = new_dvnode->status.data_version; + } + } + ret = -ERESTARTSYS; if (afs_begin_vnode_operation(&fc, orig_dvnode, key)) { if (orig_dvnode != new_dvnode) { if (mutex_lock_interruptible_nested(&new_dvnode->io_lock, 1) < 0) { afs_end_vnode_operation(&fc); - goto error_key; + goto error_rehash; } } while (afs_select_fileserver(&fc)) { @@ -1599,25 +1717,42 @@ static int afs_rename(struct inode *old_dir, struct dentry *old_dentry, mutex_unlock(&new_dvnode->io_lock); ret = afs_end_vnode_operation(&fc); if (ret < 0) - goto error_key; + goto error_rehash; } if (ret == 0) { + if (rehash) + d_rehash(rehash); if (test_bit(AFS_VNODE_DIR_VALID, &orig_dvnode->flags)) afs_edit_dir_remove(orig_dvnode, &old_dentry->d_name, - afs_edit_dir_for_rename); + afs_edit_dir_for_rename_0); if (!new_negative && test_bit(AFS_VNODE_DIR_VALID, &new_dvnode->flags)) afs_edit_dir_remove(new_dvnode, &new_dentry->d_name, - afs_edit_dir_for_rename); + afs_edit_dir_for_rename_1); if (test_bit(AFS_VNODE_DIR_VALID, &new_dvnode->flags)) afs_edit_dir_add(new_dvnode, &new_dentry->d_name, - &vnode->fid, afs_edit_dir_for_rename); + &vnode->fid, afs_edit_dir_for_rename_2); + + new_inode = d_inode(new_dentry); + if (new_inode) { + spin_lock(&new_inode->i_lock); + if (new_inode->i_nlink > 0) + drop_nlink(new_inode); + spin_unlock(&new_inode->i_lock); + } + d_move(old_dentry, new_dentry); + goto error_tmp; } -error_key: +error_rehash: + if (rehash) + d_rehash(rehash); +error_tmp: + if (tmp) + dput(tmp); key_put(key); error: _leave(" = %d", ret); diff --git a/fs/afs/dir_silly.c b/fs/afs/dir_silly.c new file mode 100644 index 000000000000..f6f89fdab6b2 --- /dev/null +++ b/fs/afs/dir_silly.c @@ -0,0 +1,239 @@ +/* AFS silly rename handling + * + * Copyright (C) 2019 Red Hat, Inc. All Rights Reserved. + * Written by David Howells (dhowells@redhat.com) + * - Derived from NFS's sillyrename. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public Licence + * as published by the Free Software Foundation; either version + * 2 of the Licence, or (at your option) any later version. + */ + +#include <linux/kernel.h> +#include <linux/fs.h> +#include <linux/namei.h> +#include <linux/fsnotify.h> +#include "internal.h" + +/* + * Actually perform the silly rename step. + */ +static int afs_do_silly_rename(struct afs_vnode *dvnode, struct afs_vnode *vnode, + struct dentry *old, struct dentry *new, + struct key *key) +{ + struct afs_fs_cursor fc; + u64 dir_data_version = dvnode->status.data_version; + int ret = -ERESTARTSYS; + + _enter("%pd,%pd", old, new); + + trace_afs_silly_rename(vnode, false); + if (afs_begin_vnode_operation(&fc, dvnode, key)) { + while (afs_select_fileserver(&fc)) { + fc.cb_break = afs_calc_vnode_cb_break(dvnode); + afs_fs_rename(&fc, old->d_name.name, + dvnode, new->d_name.name, + dir_data_version, dir_data_version); + } + + afs_vnode_commit_status(&fc, dvnode, fc.cb_break); + ret = afs_end_vnode_operation(&fc); + } + + if (ret == 0) { + spin_lock(&old->d_lock); + old->d_flags |= DCACHE_NFSFS_RENAMED; + spin_unlock(&old->d_lock); + if (dvnode->silly_key != key) { + key_put(dvnode->silly_key); + dvnode->silly_key = key_get(key); + } + + if (test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags)) + afs_edit_dir_remove(dvnode, &old->d_name, + afs_edit_dir_for_silly_0); + if (test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags)) + afs_edit_dir_add(dvnode, &new->d_name, + &vnode->fid, afs_edit_dir_for_silly_1); + + /* vfs_unlink and the like do not issue this when a file is + * sillyrenamed, so do it here. + */ + fsnotify_nameremove(old, 0); + } + + _leave(" = %d", ret); + return ret; +} + +/** + * afs_sillyrename - Perform a silly-rename of a dentry + * + * AFS is stateless and the server doesn't know when the client is holding a + * file open. To prevent application problems when a file is unlinked while + * it's still open, the client performs a "silly-rename". That is, it renames + * the file to a hidden file in the same directory, and only performs the + * unlink once the last reference to it is put. + * + * The final cleanup is done during dentry_iput. + */ +int afs_sillyrename(struct afs_vnode *dvnode, struct afs_vnode *vnode, + struct dentry *dentry, struct key *key) +{ + static unsigned int sillycounter; + struct dentry *sdentry = NULL; + unsigned char silly[16]; + int ret = -EBUSY; + + _enter(""); + + /* We don't allow a dentry to be silly-renamed twice. */ + if (dentry->d_flags & DCACHE_NFSFS_RENAMED) + return -EBUSY; + + sdentry = NULL; + do { + int slen; + + dput(sdentry); + sillycounter++; + + /* Create a silly name. Note that the ".__afs" prefix is + * understood by the salvager and must not be changed. + */ + slen = scnprintf(silly, sizeof(silly), ".__afs%04X", sillycounter); + sdentry = lookup_one_len(silly, dentry->d_parent, slen); + + /* N.B. Better to return EBUSY here ... it could be dangerous + * to delete the file while it's in use. + */ + if (IS_ERR(sdentry)) + goto out; + } while (!d_is_negative(sdentry)); + + ihold(&vnode->vfs_inode); + + ret = afs_do_silly_rename(dvnode, vnode, dentry, sdentry, key); + switch (ret) { + case 0: + /* The rename succeeded. */ + d_move(dentry, sdentry); + break; + case -ERESTARTSYS: + /* The result of the rename is unknown. Play it safe by forcing + * a new lookup. + */ + d_drop(dentry); + d_drop(sdentry); + } + + iput(&vnode->vfs_inode); + dput(sdentry); +out: + _leave(" = %d", ret); + return ret; +} + +/* + * Tell the server to remove a sillyrename file. + */ +static int afs_do_silly_unlink(struct afs_vnode *dvnode, struct afs_vnode *vnode, + struct dentry *dentry, struct key *key) +{ + struct afs_fs_cursor fc; + u64 dir_data_version = dvnode->status.data_version; + int ret = -ERESTARTSYS; + + _enter(""); + + trace_afs_silly_rename(vnode, true); + if (afs_begin_vnode_operation(&fc, dvnode, key)) { + while (afs_select_fileserver(&fc)) { + fc.cb_break = afs_calc_vnode_cb_break(dvnode); + + if (test_bit(AFS_SERVER_FL_IS_YFS, &fc.cbi->server->flags) && + !test_bit(AFS_SERVER_FL_NO_RM2, &fc.cbi->server->flags)) { + yfs_fs_remove_file2(&fc, vnode, dentry->d_name.name, + dir_data_version); + if (fc.ac.error != -ECONNABORTED || + fc.ac.abort_code != RXGEN_OPCODE) + continue; + set_bit(AFS_SERVER_FL_NO_RM2, &fc.cbi->server->flags); + } + + afs_fs_remove(&fc, vnode, dentry->d_name.name, false, + dir_data_version); + } + + afs_vnode_commit_status(&fc, dvnode, fc.cb_break); + ret = afs_end_vnode_operation(&fc); + if (ret == 0) { + drop_nlink(&vnode->vfs_inode); + if (vnode->vfs_inode.i_nlink == 0) { + set_bit(AFS_VNODE_DELETED, &vnode->flags); + clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags); + } + } + if (ret == 0 && + test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags)) + afs_edit_dir_remove(dvnode, &dentry->d_name, + afs_edit_dir_for_unlink); + } + + _leave(" = %d", ret); + return ret; +} + +/* + * Remove sillyrename file on iput. + */ +int afs_silly_iput(struct dentry *dentry, struct inode *inode) +{ + struct afs_vnode *dvnode = AFS_FS_I(d_inode(dentry->d_parent)); + struct afs_vnode *vnode = AFS_FS_I(inode); + struct dentry *alias; + int ret; + + DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); + + _enter("%p{%pd},%llx", dentry, dentry, vnode->fid.vnode); + + down_read(&dvnode->rmdir_lock); + + alias = d_alloc_parallel(dentry->d_parent, &dentry->d_name, &wq); + if (IS_ERR(alias)) { + up_read(&dvnode->rmdir_lock); + return 0; + } + + if (!d_in_lookup(alias)) { + /* We raced with lookup... See if we need to transfer the + * sillyrename information to the aliased dentry. + */ + ret = 0; + spin_lock(&alias->d_lock); + if (d_really_is_positive(alias) && + !(alias->d_flags & DCACHE_NFSFS_RENAMED)) { + alias->d_flags |= DCACHE_NFSFS_RENAMED; + ret = 1; + } + spin_unlock(&alias->d_lock); + up_read(&dvnode->rmdir_lock); + dput(alias); + return ret; + } + + /* Stop lock-release from complaining. */ + spin_lock(&vnode->lock); + vnode->lock_state = AFS_VNODE_LOCK_DELETED; + trace_afs_flock_ev(vnode, NULL, afs_flock_silly_delete, 0); + spin_unlock(&vnode->lock); + + afs_do_silly_unlink(dvnode, vnode, dentry, dvnode->silly_key); + up_read(&dvnode->rmdir_lock); + d_lookup_done(alias); + dput(alias); + return 1; +} diff --git a/fs/afs/flock.c b/fs/afs/flock.c index 6a0174258382..adc88eff7849 100644 --- a/fs/afs/flock.c +++ b/fs/afs/flock.c @@ -13,9 +13,11 @@ #define AFS_LOCK_GRANTED 0 #define AFS_LOCK_PENDING 1 +#define AFS_LOCK_YOUR_TRY 2 struct workqueue_struct *afs_lock_manager; +static void afs_next_locker(struct afs_vnode *vnode, int error); static void afs_fl_copy_lock(struct file_lock *new, struct file_lock *fl); static void afs_fl_release_private(struct file_lock *fl); @@ -24,6 +26,14 @@ static const struct file_lock_operations afs_lock_ops = { .fl_release_private = afs_fl_release_private, }; +static inline void afs_set_lock_state(struct afs_vnode *vnode, enum afs_lock_state state) +{ + _debug("STATE %u -> %u", vnode->lock_state, state); + vnode->lock_state = state; +} + +static atomic_t afs_file_lock_debug_id; + /* * if the callback is broken on this vnode, then the lock may now be available */ @@ -31,7 +41,14 @@ void afs_lock_may_be_available(struct afs_vnode *vnode) { _enter("{%llx:%llu}", vnode->fid.vid, vnode->fid.vnode); - queue_delayed_work(afs_lock_manager, &vnode->lock_work, 0); + if (vnode->lock_state != AFS_VNODE_LOCK_WAITING_FOR_CB) + return; + + spin_lock(&vnode->lock); + if (vnode->lock_state == AFS_VNODE_LOCK_WAITING_FOR_CB) + afs_next_locker(vnode, 0); + trace_afs_flock_ev(vnode, NULL, afs_flock_callback_break, 0); + spin_unlock(&vnode->lock); } /* @@ -40,8 +57,35 @@ void afs_lock_may_be_available(struct afs_vnode *vnode) */ static void afs_schedule_lock_extension(struct afs_vnode *vnode) { - queue_delayed_work(afs_lock_manager, &vnode->lock_work, - AFS_LOCKWAIT * HZ / 2); + ktime_t expires_at, now, duration; + u64 duration_j; + + expires_at = ktime_add_ms(vnode->locked_at, AFS_LOCKWAIT * 1000 / 2); + now = ktime_get_real(); + duration = ktime_sub(expires_at, now); + if (duration <= 0) + duration_j = 0; + else + duration_j = nsecs_to_jiffies(ktime_to_ns(duration)); + + queue_delayed_work(afs_lock_manager, &vnode->lock_work, duration_j); +} + +/* + * In the case of successful completion of a lock operation, record the time + * the reply appeared and start the lock extension timer. + */ +void afs_lock_op_done(struct afs_call *call) +{ + struct afs_vnode *vnode = call->reply[0]; + + if (call->error == 0) { + spin_lock(&vnode->lock); + trace_afs_flock_ev(vnode, NULL, afs_flock_timestamp, 0); + vnode->locked_at = call->reply_time; + afs_schedule_lock_extension(vnode); + spin_unlock(&vnode->lock); + } } /* @@ -49,22 +93,90 @@ static void afs_schedule_lock_extension(struct afs_vnode *vnode) * first lock in the queue is itself a readlock) * - the caller must hold the vnode lock */ -static void afs_grant_locks(struct afs_vnode *vnode, struct file_lock *fl) +static void afs_grant_locks(struct afs_vnode *vnode) { struct file_lock *p, *_p; + bool exclusive = (vnode->lock_type == AFS_LOCK_WRITE); - list_move_tail(&fl->fl_u.afs.link, &vnode->granted_locks); - if (fl->fl_type == F_RDLCK) { - list_for_each_entry_safe(p, _p, &vnode->pending_locks, - fl_u.afs.link) { - if (p->fl_type == F_RDLCK) { - p->fl_u.afs.state = AFS_LOCK_GRANTED; - list_move_tail(&p->fl_u.afs.link, - &vnode->granted_locks); - wake_up(&p->fl_wait); - } + list_for_each_entry_safe(p, _p, &vnode->pending_locks, fl_u.afs.link) { + if (!exclusive && p->fl_type == F_WRLCK) + continue; + + list_move_tail(&p->fl_u.afs.link, &vnode->granted_locks); + p->fl_u.afs.state = AFS_LOCK_GRANTED; + trace_afs_flock_op(vnode, p, afs_flock_op_grant); + wake_up(&p->fl_wait); + } +} + +/* + * If an error is specified, reject every pending lock that matches the + * authentication and type of the lock we failed to get. If there are any + * remaining lockers, try to wake up one of them to have a go. + */ +static void afs_next_locker(struct afs_vnode *vnode, int error) +{ + struct file_lock *p, *_p, *next = NULL; + struct key *key = vnode->lock_key; + unsigned int fl_type = F_RDLCK; + + _enter(""); + + if (vnode->lock_type == AFS_LOCK_WRITE) + fl_type = F_WRLCK; + + list_for_each_entry_safe(p, _p, &vnode->pending_locks, fl_u.afs.link) { + if (error && + p->fl_type == fl_type && + afs_file_key(p->fl_file) == key) { + list_del_init(&p->fl_u.afs.link); + p->fl_u.afs.state = error; + wake_up(&p->fl_wait); } + + /* Select the next locker to hand off to. */ + if (next && + (next->fl_type == F_WRLCK || p->fl_type == F_RDLCK)) + continue; + next = p; } + + vnode->lock_key = NULL; + key_put(key); + + if (next) { + afs_set_lock_state(vnode, AFS_VNODE_LOCK_SETTING); + next->fl_u.afs.state = AFS_LOCK_YOUR_TRY; + trace_afs_flock_op(vnode, next, afs_flock_op_wake); + wake_up(&next->fl_wait); + } else { + afs_set_lock_state(vnode, AFS_VNODE_LOCK_NONE); + trace_afs_flock_ev(vnode, NULL, afs_flock_no_lockers, 0); + } + + _leave(""); +} + +/* + * Kill off all waiters in the the pending lock queue due to the vnode being + * deleted. + */ +static void afs_kill_lockers_enoent(struct afs_vnode *vnode) +{ + struct file_lock *p; + + afs_set_lock_state(vnode, AFS_VNODE_LOCK_DELETED); + + while (!list_empty(&vnode->pending_locks)) { + p = list_entry(vnode->pending_locks.next, + struct file_lock, fl_u.afs.link); + list_del_init(&p->fl_u.afs.link); + p->fl_u.afs.state = -ENOENT; + wake_up(&p->fl_wait); + } + + key_put(vnode->lock_key); + vnode->lock_key = NULL; } /* @@ -170,8 +282,6 @@ void afs_lock_work(struct work_struct *work) { struct afs_vnode *vnode = container_of(work, struct afs_vnode, lock_work.work); - struct file_lock *fl, *next; - afs_lock_type_t type; struct key *key; int ret; @@ -183,35 +293,28 @@ again: _debug("wstate %u for %p", vnode->lock_state, vnode); switch (vnode->lock_state) { case AFS_VNODE_LOCK_NEED_UNLOCK: - _debug("unlock"); - vnode->lock_state = AFS_VNODE_LOCK_UNLOCKING; + afs_set_lock_state(vnode, AFS_VNODE_LOCK_UNLOCKING); + trace_afs_flock_ev(vnode, NULL, afs_flock_work_unlocking, 0); spin_unlock(&vnode->lock); /* attempt to release the server lock; if it fails, we just * wait 5 minutes and it'll expire anyway */ ret = afs_release_lock(vnode, vnode->lock_key); - if (ret < 0) + if (ret < 0 && vnode->lock_state != AFS_VNODE_LOCK_DELETED) { + trace_afs_flock_ev(vnode, NULL, afs_flock_release_fail, + ret); printk(KERN_WARNING "AFS:" " Failed to release lock on {%llx:%llx} error %d\n", vnode->fid.vid, vnode->fid.vnode, ret); - - spin_lock(&vnode->lock); - key_put(vnode->lock_key); - vnode->lock_key = NULL; - vnode->lock_state = AFS_VNODE_LOCK_NONE; - - if (list_empty(&vnode->pending_locks)) { - spin_unlock(&vnode->lock); - return; } - /* The new front of the queue now owns the state variables. */ - next = list_entry(vnode->pending_locks.next, - struct file_lock, fl_u.afs.link); - vnode->lock_key = key_get(afs_file_key(next->fl_file)); - vnode->lock_type = (next->fl_type == F_RDLCK) ? AFS_LOCK_READ : AFS_LOCK_WRITE; - vnode->lock_state = AFS_VNODE_LOCK_WAITING_FOR_CB; - goto again; + spin_lock(&vnode->lock); + if (ret == -ENOENT) + afs_kill_lockers_enoent(vnode); + else + afs_next_locker(vnode, 0); + spin_unlock(&vnode->lock); + return; /* If we've already got a lock, then it must be time to extend that * lock as AFS locks time out after 5 minutes. @@ -222,86 +325,55 @@ again: ASSERT(!list_empty(&vnode->granted_locks)); key = key_get(vnode->lock_key); - vnode->lock_state = AFS_VNODE_LOCK_EXTENDING; + afs_set_lock_state(vnode, AFS_VNODE_LOCK_EXTENDING); + trace_afs_flock_ev(vnode, NULL, afs_flock_work_extending, 0); spin_unlock(&vnode->lock); ret = afs_extend_lock(vnode, key); /* RPC */ key_put(key); - if (ret < 0) + if (ret < 0) { + trace_afs_flock_ev(vnode, NULL, afs_flock_extend_fail, + ret); pr_warning("AFS: Failed to extend lock on {%llx:%llx} error %d\n", vnode->fid.vid, vnode->fid.vnode, ret); + } spin_lock(&vnode->lock); + if (ret == -ENOENT) { + afs_kill_lockers_enoent(vnode); + spin_unlock(&vnode->lock); + return; + } + if (vnode->lock_state != AFS_VNODE_LOCK_EXTENDING) goto again; - vnode->lock_state = AFS_VNODE_LOCK_GRANTED; + afs_set_lock_state(vnode, AFS_VNODE_LOCK_GRANTED); - if (ret == 0) - afs_schedule_lock_extension(vnode); - else + if (ret != 0) queue_delayed_work(afs_lock_manager, &vnode->lock_work, HZ * 10); spin_unlock(&vnode->lock); _leave(" [ext]"); return; - /* If we don't have a granted lock, then we must've been called - * back by the server, and so if might be possible to get a - * lock we're currently waiting for. - */ + /* If we're waiting for a callback to indicate lock release, we can't + * actually rely on this, so need to recheck at regular intervals. The + * problem is that the server might not notify us if the lock just + * expires (say because a client died) rather than being explicitly + * released. + */ case AFS_VNODE_LOCK_WAITING_FOR_CB: - _debug("get"); - - key = key_get(vnode->lock_key); - type = vnode->lock_type; - vnode->lock_state = AFS_VNODE_LOCK_SETTING; + _debug("retry"); + afs_next_locker(vnode, 0); spin_unlock(&vnode->lock); + return; - ret = afs_set_lock(vnode, key, type); /* RPC */ - key_put(key); - - spin_lock(&vnode->lock); - switch (ret) { - case -EWOULDBLOCK: - _debug("blocked"); - break; - case 0: - _debug("acquired"); - vnode->lock_state = AFS_VNODE_LOCK_GRANTED; - /* Fall through */ - default: - /* Pass the lock or the error onto the first locker in - * the list - if they're looking for this type of lock. - * If they're not, we assume that whoever asked for it - * took a signal. - */ - if (list_empty(&vnode->pending_locks)) { - _debug("withdrawn"); - vnode->lock_state = AFS_VNODE_LOCK_NEED_UNLOCK; - goto again; - } - - fl = list_entry(vnode->pending_locks.next, - struct file_lock, fl_u.afs.link); - type = (fl->fl_type == F_RDLCK) ? AFS_LOCK_READ : AFS_LOCK_WRITE; - if (vnode->lock_type != type) { - _debug("changed"); - vnode->lock_state = AFS_VNODE_LOCK_NEED_UNLOCK; - goto again; - } - - fl->fl_u.afs.state = ret; - if (ret == 0) - afs_grant_locks(vnode, fl); - else - list_del_init(&fl->fl_u.afs.link); - wake_up(&fl->fl_wait); - spin_unlock(&vnode->lock); - _leave(" [granted]"); - return; - } + case AFS_VNODE_LOCK_DELETED: + afs_kill_lockers_enoent(vnode); + spin_unlock(&vnode->lock); + return; /* Fall through */ default: @@ -320,14 +392,16 @@ again: */ static void afs_defer_unlock(struct afs_vnode *vnode) { - _enter(""); + _enter("%u", vnode->lock_state); - if (vnode->lock_state == AFS_VNODE_LOCK_GRANTED || - vnode->lock_state == AFS_VNODE_LOCK_EXTENDING) { + if (list_empty(&vnode->granted_locks) && + (vnode->lock_state == AFS_VNODE_LOCK_GRANTED || + vnode->lock_state == AFS_VNODE_LOCK_EXTENDING)) { cancel_delayed_work(&vnode->lock_work); - vnode->lock_state = AFS_VNODE_LOCK_NEED_UNLOCK; - afs_lock_may_be_available(vnode); + afs_set_lock_state(vnode, AFS_VNODE_LOCK_NEED_UNLOCK); + trace_afs_flock_ev(vnode, NULL, afs_flock_defer_unlock, 0); + queue_delayed_work(afs_lock_manager, &vnode->lock_work, 0); } } @@ -336,7 +410,7 @@ static void afs_defer_unlock(struct afs_vnode *vnode) * whether we think that we have a locking permit. */ static int afs_do_setlk_check(struct afs_vnode *vnode, struct key *key, - afs_lock_type_t type, bool can_sleep) + enum afs_flock_mode mode, afs_lock_type_t type) { afs_access_t access; int ret; @@ -364,160 +438,177 @@ static int afs_do_setlk_check(struct afs_vnode *vnode, struct key *key, if (type == AFS_LOCK_READ) { if (!(access & (AFS_ACE_INSERT | AFS_ACE_WRITE | AFS_ACE_LOCK))) return -EACCES; - if (vnode->status.lock_count == -1 && !can_sleep) - return -EAGAIN; /* Write locked */ } else { if (!(access & (AFS_ACE_INSERT | AFS_ACE_WRITE))) return -EACCES; - if (vnode->status.lock_count != 0 && !can_sleep) - return -EAGAIN; /* Locked */ } return 0; } /* - * Remove the front runner from the pending queue. - * - The caller must hold vnode->lock. - */ -static void afs_dequeue_lock(struct afs_vnode *vnode, struct file_lock *fl) -{ - struct file_lock *next; - - _enter(""); - - /* ->lock_type, ->lock_key and ->lock_state only belong to this - * file_lock if we're at the front of the pending queue or if we have - * the lock granted or if the lock_state is NEED_UNLOCK or UNLOCKING. - */ - if (vnode->granted_locks.next == &fl->fl_u.afs.link && - vnode->granted_locks.prev == &fl->fl_u.afs.link) { - list_del_init(&fl->fl_u.afs.link); - afs_defer_unlock(vnode); - return; - } - - if (!list_empty(&vnode->granted_locks) || - vnode->pending_locks.next != &fl->fl_u.afs.link) { - list_del_init(&fl->fl_u.afs.link); - return; - } - - list_del_init(&fl->fl_u.afs.link); - key_put(vnode->lock_key); - vnode->lock_key = NULL; - vnode->lock_state = AFS_VNODE_LOCK_NONE; - - if (list_empty(&vnode->pending_locks)) - return; - - /* The new front of the queue now owns the state variables. */ - next = list_entry(vnode->pending_locks.next, - struct file_lock, fl_u.afs.link); - vnode->lock_key = key_get(afs_file_key(next->fl_file)); - vnode->lock_type = (next->fl_type == F_RDLCK) ? AFS_LOCK_READ : AFS_LOCK_WRITE; - vnode->lock_state = AFS_VNODE_LOCK_WAITING_FOR_CB; - afs_lock_may_be_available(vnode); -} - -/* * request a lock on a file on the server */ static int afs_do_setlk(struct file *file, struct file_lock *fl) { struct inode *inode = locks_inode(file); struct afs_vnode *vnode = AFS_FS_I(inode); + enum afs_flock_mode mode = AFS_FS_S(inode->i_sb)->flock_mode; afs_lock_type_t type; struct key *key = afs_file_key(file); + bool partial, no_server_lock = false; int ret; - _enter("{%llx:%llu},%u", vnode->fid.vid, vnode->fid.vnode, fl->fl_type); + if (mode == afs_flock_mode_unset) + mode = afs_flock_mode_openafs; - /* only whole-file locks are supported */ - if (fl->fl_start != 0 || fl->fl_end != OFFSET_MAX) - return -EINVAL; + _enter("{%llx:%llu},%llu-%llu,%u,%u", + vnode->fid.vid, vnode->fid.vnode, + fl->fl_start, fl->fl_end, fl->fl_type, mode); fl->fl_ops = &afs_lock_ops; INIT_LIST_HEAD(&fl->fl_u.afs.link); fl->fl_u.afs.state = AFS_LOCK_PENDING; + partial = (fl->fl_start != 0 || fl->fl_end != OFFSET_MAX); type = (fl->fl_type == F_RDLCK) ? AFS_LOCK_READ : AFS_LOCK_WRITE; + if (mode == afs_flock_mode_write && partial) + type = AFS_LOCK_WRITE; - ret = afs_do_setlk_check(vnode, key, type, fl->fl_flags & FL_SLEEP); + ret = afs_do_setlk_check(vnode, key, mode, type); if (ret < 0) return ret; - spin_lock(&vnode->lock); + trace_afs_flock_op(vnode, fl, afs_flock_op_set_lock); - /* If we've already got a readlock on the server then we instantly - * grant another readlock, irrespective of whether there are any - * pending writelocks. + /* AFS3 protocol only supports full-file locks and doesn't provide any + * method of upgrade/downgrade, so we need to emulate for partial-file + * locks. + * + * The OpenAFS client only gets a server lock for a full-file lock and + * keeps partial-file locks local. Allow this behaviour to be emulated + * (as the default). */ - if (type == AFS_LOCK_READ && - vnode->lock_state == AFS_VNODE_LOCK_GRANTED && - vnode->lock_type == AFS_LOCK_READ) { - _debug("instant readlock"); - ASSERT(!list_empty(&vnode->granted_locks)); - goto share_existing_lock; + if (mode == afs_flock_mode_local || + (partial && mode == afs_flock_mode_openafs)) { + no_server_lock = true; + goto skip_server_lock; } + spin_lock(&vnode->lock); list_add_tail(&fl->fl_u.afs.link, &vnode->pending_locks); + ret = -ENOENT; + if (vnode->lock_state == AFS_VNODE_LOCK_DELETED) + goto error_unlock; + + /* If we've already got a lock on the server then try to move to having + * the VFS grant the requested lock. Note that this means that other + * clients may get starved out. + */ + _debug("try %u", vnode->lock_state); + if (vnode->lock_state == AFS_VNODE_LOCK_GRANTED) { + if (type == AFS_LOCK_READ) { + _debug("instant readlock"); + list_move_tail(&fl->fl_u.afs.link, &vnode->granted_locks); + fl->fl_u.afs.state = AFS_LOCK_GRANTED; + goto vnode_is_locked_u; + } + + if (vnode->lock_type == AFS_LOCK_WRITE) { + _debug("instant writelock"); + list_move_tail(&fl->fl_u.afs.link, &vnode->granted_locks); + fl->fl_u.afs.state = AFS_LOCK_GRANTED; + goto vnode_is_locked_u; + } + } + + if (vnode->lock_state == AFS_VNODE_LOCK_NONE && + !(fl->fl_flags & FL_SLEEP)) { + ret = -EAGAIN; + if (type == AFS_LOCK_READ) { + if (vnode->status.lock_count == -1) + goto lock_is_contended; /* Write locked */ + } else { + if (vnode->status.lock_count != 0) + goto lock_is_contended; /* Locked */ + } + } + if (vnode->lock_state != AFS_VNODE_LOCK_NONE) goto need_to_wait; +try_to_lock: /* We don't have a lock on this vnode and we aren't currently waiting * for one either, so ask the server for a lock. * * Note that we need to be careful if we get interrupted by a signal * after dispatching the request as we may still get the lock, even * though we don't wait for the reply (it's not too bad a problem - the - * lock will expire in 10 mins anyway). + * lock will expire in 5 mins anyway). */ - _debug("not locked"); + trace_afs_flock_ev(vnode, fl, afs_flock_try_to_lock, 0); vnode->lock_key = key_get(key); vnode->lock_type = type; - vnode->lock_state = AFS_VNODE_LOCK_SETTING; + afs_set_lock_state(vnode, AFS_VNODE_LOCK_SETTING); spin_unlock(&vnode->lock); ret = afs_set_lock(vnode, key, type); /* RPC */ spin_lock(&vnode->lock); switch (ret) { + case -EKEYREJECTED: + case -EKEYEXPIRED: + case -EKEYREVOKED: + case -EPERM: + case -EACCES: + fl->fl_u.afs.state = ret; + trace_afs_flock_ev(vnode, fl, afs_flock_fail_perm, ret); + list_del_init(&fl->fl_u.afs.link); + afs_next_locker(vnode, ret); + goto error_unlock; + + case -ENOENT: + fl->fl_u.afs.state = ret; + trace_afs_flock_ev(vnode, fl, afs_flock_fail_other, ret); + list_del_init(&fl->fl_u.afs.link); + afs_kill_lockers_enoent(vnode); + goto error_unlock; + default: - goto abort_attempt; + fl->fl_u.afs.state = ret; + trace_afs_flock_ev(vnode, fl, afs_flock_fail_other, ret); + list_del_init(&fl->fl_u.afs.link); + afs_next_locker(vnode, 0); + goto error_unlock; case -EWOULDBLOCK: /* The server doesn't have a lock-waiting queue, so the client * will have to retry. The server will break the outstanding * callbacks on a file when a lock is released. */ - _debug("would block"); ASSERT(list_empty(&vnode->granted_locks)); ASSERTCMP(vnode->pending_locks.next, ==, &fl->fl_u.afs.link); - vnode->lock_state = AFS_VNODE_LOCK_WAITING_FOR_CB; - goto need_to_wait; + goto lock_is_contended; case 0: - _debug("acquired"); - break; + afs_set_lock_state(vnode, AFS_VNODE_LOCK_GRANTED); + trace_afs_flock_ev(vnode, fl, afs_flock_acquired, type); + afs_grant_locks(vnode); + goto vnode_is_locked_u; } - /* we've acquired a server lock, but it needs to be renewed after 5 - * mins */ - vnode->lock_state = AFS_VNODE_LOCK_GRANTED; - afs_schedule_lock_extension(vnode); - -share_existing_lock: - /* the lock has been granted as far as we're concerned... */ - fl->fl_u.afs.state = AFS_LOCK_GRANTED; - list_move_tail(&fl->fl_u.afs.link, &vnode->granted_locks); - -given_lock: - /* ... but we do still need to get the VFS's blessing */ +vnode_is_locked_u: spin_unlock(&vnode->lock); - - ret = posix_lock_file(file, fl, NULL); +vnode_is_locked: + /* the lock has been granted by the server... */ + ASSERTCMP(fl->fl_u.afs.state, ==, AFS_LOCK_GRANTED); + +skip_server_lock: + /* ... but the VFS still needs to distribute access on this client. */ + trace_afs_flock_ev(vnode, fl, afs_flock_vfs_locking, 0); + ret = locks_lock_file_wait(file, fl); + trace_afs_flock_ev(vnode, fl, afs_flock_vfs_lock, ret); if (ret < 0) goto vfs_rejected_lock; @@ -529,38 +620,62 @@ given_lock: _leave(" = 0"); return 0; +lock_is_contended: + if (!(fl->fl_flags & FL_SLEEP)) { + list_del_init(&fl->fl_u.afs.link); + afs_next_locker(vnode, 0); + ret = -EAGAIN; + goto error_unlock; + } + + afs_set_lock_state(vnode, AFS_VNODE_LOCK_WAITING_FOR_CB); + trace_afs_flock_ev(vnode, fl, afs_flock_would_block, ret); + queue_delayed_work(afs_lock_manager, &vnode->lock_work, HZ * 5); + need_to_wait: /* We're going to have to wait. Either this client doesn't have a lock * on the server yet and we need to wait for a callback to occur, or - * the client does have a lock on the server, but it belongs to some - * other process(es) and is incompatible with the lock we want. + * the client does have a lock on the server, but it's shared and we + * need an exclusive lock. */ - ret = -EAGAIN; - if (fl->fl_flags & FL_SLEEP) { - spin_unlock(&vnode->lock); + spin_unlock(&vnode->lock); - _debug("sleep"); - ret = wait_event_interruptible(fl->fl_wait, - fl->fl_u.afs.state != AFS_LOCK_PENDING); + trace_afs_flock_ev(vnode, fl, afs_flock_waiting, 0); + ret = wait_event_interruptible(fl->fl_wait, + fl->fl_u.afs.state != AFS_LOCK_PENDING); + trace_afs_flock_ev(vnode, fl, afs_flock_waited, ret); + if (fl->fl_u.afs.state >= 0 && fl->fl_u.afs.state != AFS_LOCK_GRANTED) { spin_lock(&vnode->lock); - } - if (fl->fl_u.afs.state == AFS_LOCK_GRANTED) - goto given_lock; - if (fl->fl_u.afs.state < 0) - ret = fl->fl_u.afs.state; + switch (fl->fl_u.afs.state) { + case AFS_LOCK_YOUR_TRY: + fl->fl_u.afs.state = AFS_LOCK_PENDING; + goto try_to_lock; + case AFS_LOCK_PENDING: + if (ret > 0) { + /* We need to retry the lock. We may not be + * notified by the server if it just expired + * rather than being released. + */ + ASSERTCMP(vnode->lock_state, ==, AFS_VNODE_LOCK_WAITING_FOR_CB); + afs_set_lock_state(vnode, AFS_VNODE_LOCK_SETTING); + fl->fl_u.afs.state = AFS_LOCK_PENDING; + goto try_to_lock; + } + goto error_unlock; + case AFS_LOCK_GRANTED: + default: + break; + } -abort_attempt: - /* we aren't going to get the lock, either because we're unwilling to - * wait, or because some signal happened */ - _debug("abort"); - afs_dequeue_lock(vnode, fl); + spin_unlock(&vnode->lock); + } -error_unlock: - spin_unlock(&vnode->lock); - _leave(" = %d", ret); - return ret; + if (fl->fl_u.afs.state == AFS_LOCK_GRANTED) + goto vnode_is_locked; + ret = fl->fl_u.afs.state; + goto error; vfs_rejected_lock: /* The VFS rejected the lock we just obtained, so we have to discard @@ -568,11 +683,17 @@ vfs_rejected_lock: * deal with. */ _debug("vfs refused %d", ret); + if (no_server_lock) + goto error; spin_lock(&vnode->lock); list_del_init(&fl->fl_u.afs.link); - if (list_empty(&vnode->granted_locks)) - afs_defer_unlock(vnode); - goto error_unlock; + afs_defer_unlock(vnode); + +error_unlock: + spin_unlock(&vnode->lock); +error: + _leave(" = %d", ret); + return ret; } /* @@ -585,14 +706,12 @@ static int afs_do_unlk(struct file *file, struct file_lock *fl) _enter("{%llx:%llu},%u", vnode->fid.vid, vnode->fid.vnode, fl->fl_type); + trace_afs_flock_op(vnode, fl, afs_flock_op_unlock); + /* Flush all pending writes before doing anything with locks. */ vfs_fsync(file, 0); - /* only whole-file unlocks are supported */ - if (fl->fl_start != 0 || fl->fl_end != OFFSET_MAX) - return -EINVAL; - - ret = posix_lock_file(file, fl, NULL); + ret = locks_lock_file_wait(file, fl); _leave(" = %d [%u]", ret, vnode->lock_state); return ret; } @@ -608,6 +727,9 @@ static int afs_do_getlk(struct file *file, struct file_lock *fl) _enter(""); + if (vnode->lock_state == AFS_VNODE_LOCK_DELETED) + return -ENOENT; + fl->fl_type = F_UNLCK; /* check local lock records first */ @@ -619,12 +741,15 @@ static int afs_do_getlk(struct file *file, struct file_lock *fl) goto error; lock_count = READ_ONCE(vnode->status.lock_count); - if (lock_count > 0) - fl->fl_type = F_RDLCK; - else - fl->fl_type = F_WRLCK; - fl->fl_start = 0; - fl->fl_end = OFFSET_MAX; + if (lock_count != 0) { + if (lock_count > 0) + fl->fl_type = F_RDLCK; + else + fl->fl_type = F_WRLCK; + fl->fl_start = 0; + fl->fl_end = OFFSET_MAX; + fl->fl_pid = 0; + } } ret = 0; @@ -639,6 +764,8 @@ error: int afs_lock(struct file *file, int cmd, struct file_lock *fl) { struct afs_vnode *vnode = AFS_FS_I(locks_inode(file)); + enum afs_flock_operation op; + int ret; _enter("{%llx:%llu},%d,{t=%x,fl=%x,r=%Ld:%Ld}", vnode->fid.vid, vnode->fid.vnode, cmd, @@ -651,9 +778,23 @@ int afs_lock(struct file *file, int cmd, struct file_lock *fl) if (IS_GETLK(cmd)) return afs_do_getlk(file, fl); + + fl->fl_u.afs.debug_id = atomic_inc_return(&afs_file_lock_debug_id); + trace_afs_flock_op(vnode, fl, afs_flock_op_lock); + if (fl->fl_type == F_UNLCK) - return afs_do_unlk(file, fl); - return afs_do_setlk(file, fl); + ret = afs_do_unlk(file, fl); + else + ret = afs_do_setlk(file, fl); + + switch (ret) { + case 0: op = afs_flock_op_return_ok; break; + case -EAGAIN: op = afs_flock_op_return_eagain; break; + case -EDEADLK: op = afs_flock_op_return_edeadlk; break; + default: op = afs_flock_op_return_error; break; + } + trace_afs_flock_op(vnode, fl, op); + return ret; } /* @@ -662,6 +803,8 @@ int afs_lock(struct file *file, int cmd, struct file_lock *fl) int afs_flock(struct file *file, int cmd, struct file_lock *fl) { struct afs_vnode *vnode = AFS_FS_I(locks_inode(file)); + enum afs_flock_operation op; + int ret; _enter("{%llx:%llu},%d,{t=%x,fl=%x}", vnode->fid.vid, vnode->fid.vnode, cmd, @@ -677,10 +820,23 @@ int afs_flock(struct file *file, int cmd, struct file_lock *fl) if (!(fl->fl_flags & FL_FLOCK)) return -ENOLCK; + fl->fl_u.afs.debug_id = atomic_inc_return(&afs_file_lock_debug_id); + trace_afs_flock_op(vnode, fl, afs_flock_op_flock); + /* we're simulating flock() locks using posix locks on the server */ if (fl->fl_type == F_UNLCK) - return afs_do_unlk(file, fl); - return afs_do_setlk(file, fl); + ret = afs_do_unlk(file, fl); + else + ret = afs_do_setlk(file, fl); + + switch (ret) { + case 0: op = afs_flock_op_return_ok; break; + case -EAGAIN: op = afs_flock_op_return_eagain; break; + case -EDEADLK: op = afs_flock_op_return_edeadlk; break; + default: op = afs_flock_op_return_error; break; + } + trace_afs_flock_op(vnode, fl, op); + return ret; } /* @@ -695,7 +851,10 @@ static void afs_fl_copy_lock(struct file_lock *new, struct file_lock *fl) _enter(""); + new->fl_u.afs.debug_id = atomic_inc_return(&afs_file_lock_debug_id); + spin_lock(&vnode->lock); + trace_afs_flock_op(vnode, new, afs_flock_op_copy_lock); list_add(&new->fl_u.afs.link, &fl->fl_u.afs.link); spin_unlock(&vnode->lock); } @@ -711,7 +870,12 @@ static void afs_fl_release_private(struct file_lock *fl) _enter(""); spin_lock(&vnode->lock); - afs_dequeue_lock(vnode, fl); + + trace_afs_flock_op(vnode, fl, afs_flock_op_release_lock); + list_del_init(&fl->fl_u.afs.link); + if (list_empty(&vnode->granted_locks)) + afs_defer_unlock(vnode); + _debug("state %u for %p", vnode->lock_state, vnode); spin_unlock(&vnode->lock); } diff --git a/fs/afs/fs_probe.c b/fs/afs/fs_probe.c index 3a9eaec06756..5d3abde52a0f 100644 --- a/fs/afs/fs_probe.c +++ b/fs/afs/fs_probe.c @@ -141,8 +141,8 @@ static int afs_do_probe_fileserver(struct afs_net *net, struct afs_addr_cursor ac = { .index = 0, }; + struct afs_call *call; bool in_progress = false; - int err; _enter("%pU", &server->uuid); @@ -156,12 +156,13 @@ static int afs_do_probe_fileserver(struct afs_net *net, server->probe.rtt = UINT_MAX; for (ac.index = 0; ac.index < ac.alist->nr_addrs; ac.index++) { - err = afs_fs_get_capabilities(net, server, &ac, key, server_index, - true); - if (err == -EINPROGRESS) + call = afs_fs_get_capabilities(net, server, &ac, key, server_index); + if (!IS_ERR(call)) { + afs_put_call(call); in_progress = true; - else - afs_prioritise_error(_e, err, ac.abort_code); + } else { + afs_prioritise_error(_e, PTR_ERR(call), ac.abort_code); + } } if (!in_progress) diff --git a/fs/afs/fsclient.c b/fs/afs/fsclient.c index b68471ce5c35..1296f5dc4c1e 100644 --- a/fs/afs/fsclient.c +++ b/fs/afs/fsclient.c @@ -468,7 +468,9 @@ int afs_fs_fetch_file_status(struct afs_fs_cursor *fc, struct afs_volsync *volsy call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -662,7 +664,8 @@ static int afs_fs_fetch_data64(struct afs_fs_cursor *fc, struct afs_read *req) call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -709,7 +712,8 @@ int afs_fs_fetch_data(struct afs_fs_cursor *fc, struct afs_read *req) call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -828,14 +832,16 @@ int afs_fs_create(struct afs_fs_cursor *fc, *bp++ = 0; /* segment size */ afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &vnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* - * deliver reply data to an FS.RemoveFile or FS.RemoveDir + * Deliver reply data to any operation that returns file status and volume + * sync. */ -static int afs_deliver_fs_remove(struct afs_call *call) +static int afs_deliver_fs_status_and_vol(struct afs_call *call) { struct afs_vnode *vnode = call->reply[0]; const __be32 *bp; @@ -865,14 +871,14 @@ static int afs_deliver_fs_remove(struct afs_call *call) static const struct afs_call_type afs_RXFSRemoveFile = { .name = "FS.RemoveFile", .op = afs_FS_RemoveFile, - .deliver = afs_deliver_fs_remove, + .deliver = afs_deliver_fs_status_and_vol, .destructor = afs_flat_call_destructor, }; static const struct afs_call_type afs_RXFSRemoveDir = { .name = "FS.RemoveDir", .op = afs_FS_RemoveDir, - .deliver = afs_deliver_fs_remove, + .deliver = afs_deliver_fs_status_and_vol, .destructor = afs_flat_call_destructor, }; @@ -923,8 +929,9 @@ int afs_fs_remove(struct afs_fs_cursor *fc, struct afs_vnode *vnode, } afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &dvnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &dvnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1015,8 +1022,9 @@ int afs_fs_link(struct afs_fs_cursor *fc, struct afs_vnode *vnode, *bp++ = htonl(vnode->fid.unique); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &vnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1129,8 +1137,9 @@ int afs_fs_symlink(struct afs_fs_cursor *fc, *bp++ = 0; /* segment size */ afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &vnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1247,8 +1256,9 @@ int afs_fs_rename(struct afs_fs_cursor *fc, } afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &orig_dvnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call2(call, &orig_dvnode->fid, orig_name, new_name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1352,7 +1362,8 @@ static int afs_fs_store_data64(struct afs_fs_cursor *fc, *bp++ = htonl((u32) i_size); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1428,7 +1439,8 @@ int afs_fs_store_data(struct afs_fs_cursor *fc, struct address_space *mapping, afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1526,7 +1538,8 @@ static int afs_fs_setattr_size64(struct afs_fs_cursor *fc, struct iattr *attr) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1572,7 +1585,8 @@ static int afs_fs_setattr_size(struct afs_fs_cursor *fc, struct iattr *attr) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1616,7 +1630,8 @@ int afs_fs_setattr(struct afs_fs_cursor *fc, struct iattr *attr) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1800,7 +1815,8 @@ int afs_fs_get_volume_status(struct afs_fs_cursor *fc, afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1832,6 +1848,7 @@ static const struct afs_call_type afs_RXFSSetLock = { .name = "FS.SetLock", .op = afs_FS_SetLock, .deliver = afs_deliver_fs_xxxx_lock, + .done = afs_lock_op_done, .destructor = afs_flat_call_destructor, }; @@ -1842,6 +1859,7 @@ static const struct afs_call_type afs_RXFSExtendLock = { .name = "FS.ExtendLock", .op = afs_FS_ExtendLock, .deliver = afs_deliver_fs_xxxx_lock, + .done = afs_lock_op_done, .destructor = afs_flat_call_destructor, }; @@ -1876,6 +1894,7 @@ int afs_fs_set_lock(struct afs_fs_cursor *fc, afs_lock_type_t type) call->key = fc->key; call->reply[0] = vnode; + call->want_reply_time = true; /* marshall the parameters */ bp = call->request; @@ -1886,8 +1905,9 @@ int afs_fs_set_lock(struct afs_fs_cursor *fc, afs_lock_type_t type) *bp++ = htonl(type); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_calli(call, &vnode->fid, type); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1911,6 +1931,7 @@ int afs_fs_extend_lock(struct afs_fs_cursor *fc) call->key = fc->key; call->reply[0] = vnode; + call->want_reply_time = true; /* marshall the parameters */ bp = call->request; @@ -1921,7 +1942,8 @@ int afs_fs_extend_lock(struct afs_fs_cursor *fc) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1955,7 +1977,8 @@ int afs_fs_release_lock(struct afs_fs_cursor *fc) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -2000,7 +2023,8 @@ int afs_fs_give_up_all_callbacks(struct afs_net *net, *bp++ = htonl(FSGIVEUPALLCALLBACKS); /* Can't take a ref on server */ - return afs_make_call(ac, call, GFP_NOFS, false); + afs_make_call(ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, ac); } /* @@ -2070,12 +2094,11 @@ static const struct afs_call_type afs_RXFSGetCapabilities = { * Probe a fileserver for the capabilities that it supports. This can * return up to 196 words. */ -int afs_fs_get_capabilities(struct afs_net *net, - struct afs_server *server, - struct afs_addr_cursor *ac, - struct key *key, - unsigned int server_index, - bool async) +struct afs_call *afs_fs_get_capabilities(struct afs_net *net, + struct afs_server *server, + struct afs_addr_cursor *ac, + struct key *key, + unsigned int server_index) { struct afs_call *call; __be32 *bp; @@ -2084,13 +2107,14 @@ int afs_fs_get_capabilities(struct afs_net *net, call = afs_alloc_flat_call(net, &afs_RXFSGetCapabilities, 1 * 4, 16 * 4); if (!call) - return -ENOMEM; + return ERR_PTR(-ENOMEM); call->key = key; call->reply[0] = afs_get_server(server); call->reply[1] = (void *)(long)server_index; call->upgrade = true; call->want_reply_time = true; + call->async = true; /* marshall the parameters */ bp = call->request; @@ -2098,7 +2122,8 @@ int afs_fs_get_capabilities(struct afs_net *net, /* Can't take a ref on server */ trace_afs_make_fs_call(call, NULL); - return afs_make_call(ac, call, GFP_NOFS, async); + afs_make_call(ac, call, GFP_NOFS); + return call; } /* @@ -2185,7 +2210,8 @@ int afs_fs_fetch_status(struct afs_fs_cursor *fc, call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -2370,5 +2396,180 @@ int afs_fs_inline_bulk_status(struct afs_fs_cursor *fc, call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &fids[0]); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); +} + +/* + * deliver reply data to an FS.FetchACL + */ +static int afs_deliver_fs_fetch_acl(struct afs_call *call) +{ + struct afs_vnode *vnode = call->reply[1]; + struct afs_acl *acl; + const __be32 *bp; + unsigned int size; + int ret; + + _enter("{%u}", call->unmarshall); + + switch (call->unmarshall) { + case 0: + afs_extract_to_tmp(call); + call->unmarshall++; + + /* extract the returned data length */ + case 1: + ret = afs_extract_data(call, true); + if (ret < 0) + return ret; + + size = call->count2 = ntohl(call->tmp); + size = round_up(size, 4); + + acl = kmalloc(struct_size(acl, data, size), GFP_KERNEL); + if (!acl) + return -ENOMEM; + call->reply[0] = acl; + acl->size = call->count2; + afs_extract_begin(call, acl->data, size); + call->unmarshall++; + + /* extract the returned data */ + case 2: + ret = afs_extract_data(call, true); + if (ret < 0) + return ret; + + afs_extract_to_buf(call, (21 + 6) * 4); + call->unmarshall++; + + /* extract the metadata */ + case 3: + ret = afs_extract_data(call, false); + if (ret < 0) + return ret; + + bp = call->buffer; + ret = afs_decode_status(call, &bp, &vnode->status, vnode, + &vnode->status.data_version, NULL); + if (ret < 0) + return ret; + xdr_decode_AFSVolSync(&bp, call->reply[2]); + + call->unmarshall++; + + case 4: + break; + } + + _leave(" = 0 [done]"); + return 0; +} + +static void afs_destroy_fs_fetch_acl(struct afs_call *call) +{ + kfree(call->reply[0]); + afs_flat_call_destructor(call); +} + +/* + * FS.FetchACL operation type + */ +static const struct afs_call_type afs_RXFSFetchACL = { + .name = "FS.FetchACL", + .op = afs_FS_FetchACL, + .deliver = afs_deliver_fs_fetch_acl, + .destructor = afs_destroy_fs_fetch_acl, +}; + +/* + * Fetch the ACL for a file. + */ +struct afs_acl *afs_fs_fetch_acl(struct afs_fs_cursor *fc) +{ + struct afs_vnode *vnode = fc->vnode; + struct afs_call *call; + struct afs_net *net = afs_v2net(vnode); + __be32 *bp; + + _enter(",%x,{%llx:%llu},,", + key_serial(fc->key), vnode->fid.vid, vnode->fid.vnode); + + call = afs_alloc_flat_call(net, &afs_RXFSFetchACL, 16, (21 + 6) * 4); + if (!call) { + fc->ac.error = -ENOMEM; + return ERR_PTR(-ENOMEM); + } + + call->key = fc->key; + call->reply[0] = NULL; + call->reply[1] = vnode; + call->reply[2] = NULL; /* volsync */ + call->ret_reply0 = true; + + /* marshall the parameters */ + bp = call->request; + bp[0] = htonl(FSFETCHACL); + bp[1] = htonl(vnode->fid.vid); + bp[2] = htonl(vnode->fid.vnode); + bp[3] = htonl(vnode->fid.unique); + + call->cb_break = fc->cb_break; + afs_use_fs_server(call, fc->cbi); + trace_afs_make_fs_call(call, &vnode->fid); + afs_make_call(&fc->ac, call, GFP_KERNEL); + return (struct afs_acl *)afs_wait_for_call_to_complete(call, &fc->ac); +} + +/* + * FS.StoreACL operation type + */ +static const struct afs_call_type afs_RXFSStoreACL = { + .name = "FS.StoreACL", + .op = afs_FS_StoreACL, + .deliver = afs_deliver_fs_status_and_vol, + .destructor = afs_flat_call_destructor, +}; + +/* + * Fetch the ACL for a file. + */ +int afs_fs_store_acl(struct afs_fs_cursor *fc, const struct afs_acl *acl) +{ + struct afs_vnode *vnode = fc->vnode; + struct afs_call *call; + struct afs_net *net = afs_v2net(vnode); + size_t size; + __be32 *bp; + + _enter(",%x,{%llx:%llu},,", + key_serial(fc->key), vnode->fid.vid, vnode->fid.vnode); + + size = round_up(acl->size, 4); + call = afs_alloc_flat_call(net, &afs_RXFSStoreACL, + 5 * 4 + size, (21 + 6) * 4); + if (!call) { + fc->ac.error = -ENOMEM; + return -ENOMEM; + } + + call->key = fc->key; + call->reply[0] = vnode; + call->reply[2] = NULL; /* volsync */ + + /* marshall the parameters */ + bp = call->request; + bp[0] = htonl(FSSTOREACL); + bp[1] = htonl(vnode->fid.vid); + bp[2] = htonl(vnode->fid.vnode); + bp[3] = htonl(vnode->fid.unique); + bp[4] = htonl(acl->size); + memcpy(&bp[5], acl->data, acl->size); + if (acl->size != size) + memset((void *)&bp[5] + acl->size, 0, size - acl->size); + + trace_afs_make_fs_call(call, &vnode->fid); + afs_make_call(&fc->ac, call, GFP_KERNEL); + return afs_wait_for_call_to_complete(call, &fc->ac); } diff --git a/fs/afs/inode.c b/fs/afs/inode.c index 9cedc3fc1b77..c4652b42d545 100644 --- a/fs/afs/inode.c +++ b/fs/afs/inode.c @@ -29,10 +29,36 @@ static const struct inode_operations afs_symlink_inode_operations = { .listxattr = afs_listxattr, }; +static noinline void dump_vnode(struct afs_vnode *vnode, struct afs_vnode *parent_vnode) +{ + static unsigned long once_only; + + pr_warn("kAFS: AFS vnode with undefined type %u\n", + vnode->status.type); + pr_warn("kAFS: A=%d m=%o s=%llx v=%llx\n", + vnode->status.abort_code, + vnode->status.mode, + vnode->status.size, + vnode->status.data_version); + pr_warn("kAFS: vnode %llx:%llx:%x\n", + vnode->fid.vid, + vnode->fid.vnode, + vnode->fid.unique); + if (parent_vnode) + pr_warn("kAFS: dir %llx:%llx:%x\n", + parent_vnode->fid.vid, + parent_vnode->fid.vnode, + parent_vnode->fid.unique); + + if (!test_and_set_bit(0, &once_only)) + dump_stack(); +} + /* * Initialise an inode from the vnode status. */ -static int afs_inode_init_from_status(struct afs_vnode *vnode, struct key *key) +static int afs_inode_init_from_status(struct afs_vnode *vnode, struct key *key, + struct afs_vnode *parent_vnode) { struct inode *inode = AFS_VNODE_TO_I(vnode); @@ -80,12 +106,16 @@ static int afs_inode_init_from_status(struct afs_vnode *vnode, struct key *key) inode_nohighmem(inode); break; default: - printk("kAFS: AFS vnode with undefined type\n"); + dump_vnode(vnode, parent_vnode); read_sequnlock_excl(&vnode->cb_lock); return afs_protocol_error(NULL, -EBADMSG, afs_eproto_file_type); } - inode->i_blocks = 0; + /* + * Estimate 512 bytes blocks used, rounded up to nearest 1K + * for consistency with other AFS clients. + */ + inode->i_blocks = ((i_size_read(inode) + 1023) >> 10) << 1; vnode->invalid_before = vnode->status.data_version; read_sequnlock_excl(&vnode->cb_lock); @@ -270,7 +300,8 @@ static void afs_get_inode_cache(struct afs_vnode *vnode) */ struct inode *afs_iget(struct super_block *sb, struct key *key, struct afs_fid *fid, struct afs_file_status *status, - struct afs_callback *cb, struct afs_cb_interest *cbi) + struct afs_callback *cb, struct afs_cb_interest *cbi, + struct afs_vnode *parent_vnode) { struct afs_iget_data data = { .fid = *fid }; struct afs_super_info *as; @@ -327,7 +358,7 @@ struct inode *afs_iget(struct super_block *sb, struct key *key, vnode->cb_expires_at += ktime_get_real_seconds(); } - ret = afs_inode_init_from_status(vnode, key); + ret = afs_inode_init_from_status(vnode, key, parent_vnode); if (ret < 0) goto bad_inode; @@ -543,6 +574,8 @@ void afs_evict_inode(struct inode *inode) #endif afs_put_permits(rcu_access_pointer(vnode->permit_cache)); + key_put(vnode->silly_key); + vnode->silly_key = NULL; key_put(vnode->lock_key); vnode->lock_key = NULL; _leave(""); diff --git a/fs/afs/internal.h b/fs/afs/internal.h index 3904ab0b9563..b3cd6e8ad59d 100644 --- a/fs/afs/internal.h +++ b/fs/afs/internal.h @@ -36,11 +36,24 @@ struct pagevec; struct afs_call; +/* + * Partial file-locking emulation mode. (The problem being that AFS3 only + * allows whole-file locks and no upgrading/downgrading). + */ +enum afs_flock_mode { + afs_flock_mode_unset, + afs_flock_mode_local, /* Local locking only */ + afs_flock_mode_openafs, /* Don't get server lock for a partial lock */ + afs_flock_mode_strict, /* Always get a server lock for a partial lock */ + afs_flock_mode_write, /* Get an exclusive server lock for a partial lock */ +}; + struct afs_fs_context { bool force; /* T to force cell type */ bool autocell; /* T if set auto mount operation */ bool dyn_root; /* T if dynamic root */ bool no_cell; /* T if the source is "none" (for dynroot) */ + enum afs_flock_mode flock_mode; /* Partial file-locking emulation mode */ afs_voltype_t type; /* type of volume requested */ unsigned int volnamesz; /* size of volume name */ const char *volname; /* name of volume to mount */ @@ -221,6 +234,7 @@ struct afs_super_info { struct net *net_ns; /* Network namespace */ struct afs_cell *cell; /* The cell in which the volume resides */ struct afs_volume *volume; /* volume record */ + enum afs_flock_mode flock_mode:8; /* File locking emulation mode */ bool dyn_root; /* True if dynamic root */ }; @@ -599,6 +613,7 @@ enum afs_lock_state { AFS_VNODE_LOCK_EXTENDING, /* We're extending a lock on the server */ AFS_VNODE_LOCK_NEED_UNLOCK, /* We need to unlock on the server */ AFS_VNODE_LOCK_UNLOCKING, /* We're telling the server to unlock */ + AFS_VNODE_LOCK_DELETED, /* The vnode has been deleted whilst we have a lock */ }; /* @@ -620,6 +635,8 @@ struct afs_vnode { struct afs_permits __rcu *permit_cache; /* cache of permits so far obtained */ struct mutex io_lock; /* Lock for serialising I/O on this mutex */ struct rw_semaphore validate_lock; /* lock for validating this vnode */ + struct rw_semaphore rmdir_lock; /* Lock for rmdir vs sillyrename */ + struct key *silly_key; /* Silly rename key */ spinlock_t wb_lock; /* lock for wb_keys */ spinlock_t lock; /* waitqueue/flags lock */ unsigned long flags; @@ -638,6 +655,7 @@ struct afs_vnode { struct list_head granted_locks; /* locks granted on this file */ struct delayed_work lock_work; /* work to be done in locking */ struct key *lock_key; /* Key to be used in lock ops */ + ktime_t locked_at; /* Time at which lock obtained */ enum afs_lock_state lock_state : 8; afs_lock_type_t lock_type : 8; @@ -864,6 +882,7 @@ extern const struct address_space_operations afs_dir_aops; extern const struct dentry_operations afs_fs_dentry_operations; extern void afs_d_release(struct dentry *); +extern int afs_dir_remove_link(struct dentry *, struct key *, unsigned long, unsigned long); /* * dir_edit.c @@ -873,6 +892,13 @@ extern void afs_edit_dir_add(struct afs_vnode *, struct qstr *, struct afs_fid * extern void afs_edit_dir_remove(struct afs_vnode *, struct qstr *, enum afs_edit_dir_reason); /* + * dir_silly.c + */ +extern int afs_sillyrename(struct afs_vnode *, struct afs_vnode *, + struct dentry *, struct key *); +extern int afs_silly_iput(struct dentry *, struct inode *); + +/* * dynroot.c */ extern const struct file_operations afs_dynroot_file_operations; @@ -905,6 +931,7 @@ extern void afs_put_read(struct afs_read *); */ extern struct workqueue_struct *afs_lock_manager; +extern void afs_lock_op_done(struct afs_call *); extern void afs_lock_work(struct work_struct *); extern void afs_lock_may_be_available(struct afs_vnode *); extern int afs_lock(struct file *, int, struct file_lock *); @@ -939,8 +966,9 @@ extern int afs_fs_extend_lock(struct afs_fs_cursor *); extern int afs_fs_release_lock(struct afs_fs_cursor *); extern int afs_fs_give_up_all_callbacks(struct afs_net *, struct afs_server *, struct afs_addr_cursor *, struct key *); -extern int afs_fs_get_capabilities(struct afs_net *, struct afs_server *, - struct afs_addr_cursor *, struct key *, unsigned int, bool); +extern struct afs_call *afs_fs_get_capabilities(struct afs_net *, struct afs_server *, + struct afs_addr_cursor *, struct key *, + unsigned int); extern int afs_fs_inline_bulk_status(struct afs_fs_cursor *, struct afs_net *, struct afs_fid *, struct afs_file_status *, struct afs_callback *, unsigned int, @@ -949,6 +977,14 @@ extern int afs_fs_fetch_status(struct afs_fs_cursor *, struct afs_net *, struct afs_fid *, struct afs_file_status *, struct afs_callback *, struct afs_volsync *); +struct afs_acl { + u32 size; + u8 data[]; +}; + +extern struct afs_acl *afs_fs_fetch_acl(struct afs_fs_cursor *); +extern int afs_fs_store_acl(struct afs_fs_cursor *, const struct afs_acl *); + /* * fs_probe.c */ @@ -965,7 +1001,8 @@ extern struct inode *afs_iget_pseudo_dir(struct super_block *, bool); extern struct inode *afs_iget(struct super_block *, struct key *, struct afs_fid *, struct afs_file_status *, struct afs_callback *, - struct afs_cb_interest *); + struct afs_cb_interest *, + struct afs_vnode *); extern void afs_zap_data(struct afs_vnode *); extern int afs_validate(struct afs_vnode *, struct key *); extern int afs_getattr(const struct path *, struct kstat *, u32, unsigned int); @@ -1073,7 +1110,8 @@ extern int __net_init afs_open_socket(struct afs_net *); extern void __net_exit afs_close_socket(struct afs_net *); extern void afs_charge_preallocation(struct work_struct *); extern void afs_put_call(struct afs_call *); -extern long afs_make_call(struct afs_addr_cursor *, struct afs_call *, gfp_t, bool); +extern void afs_make_call(struct afs_addr_cursor *, struct afs_call *, gfp_t); +extern long afs_wait_for_call_to_complete(struct afs_call *, struct afs_addr_cursor *); extern struct afs_call *afs_alloc_flat_call(struct afs_net *, const struct afs_call_type *, size_t, size_t); @@ -1218,8 +1256,8 @@ extern void afs_fs_exit(void); extern struct afs_vldb_entry *afs_vl_get_entry_by_name_u(struct afs_vl_cursor *, const char *, int); extern struct afs_addr_list *afs_vl_get_addrs_u(struct afs_vl_cursor *, const uuid_t *); -extern int afs_vl_get_capabilities(struct afs_net *, struct afs_addr_cursor *, struct key *, - struct afs_vlserver *, unsigned int, bool); +extern struct afs_call *afs_vl_get_capabilities(struct afs_net *, struct afs_addr_cursor *, + struct key *, struct afs_vlserver *, unsigned int); extern struct afs_addr_list *afs_yfsvl_get_endpoints(struct afs_vl_cursor *, const uuid_t *); /* @@ -1333,6 +1371,20 @@ extern int yfs_fs_inline_bulk_status(struct afs_fs_cursor *, struct afs_net *, struct afs_callback *, unsigned int, struct afs_volsync *); +struct yfs_acl { + struct afs_acl *acl; /* Dir/file/symlink ACL */ + struct afs_acl *vol_acl; /* Whole volume ACL */ + u32 inherit_flag; /* True if ACL is inherited from parent dir */ + u32 num_cleaned; /* Number of ACEs removed due to subject removal */ + unsigned int flags; +#define YFS_ACL_WANT_ACL 0x01 /* Set if caller wants ->acl */ +#define YFS_ACL_WANT_VOL_ACL 0x02 /* Set if caller wants ->vol_acl */ +}; + +extern void yfs_free_opaque_acl(struct yfs_acl *); +extern struct yfs_acl *yfs_fs_fetch_opaque_acl(struct afs_fs_cursor *, unsigned int); +extern int yfs_fs_store_opaque_acl2(struct afs_fs_cursor *, const struct afs_acl *); + /* * Miscellaneous inline functions. */ diff --git a/fs/afs/protocol_yfs.h b/fs/afs/protocol_yfs.h index d443e2bfa094..915b9d10cdf3 100644 --- a/fs/afs/protocol_yfs.h +++ b/fs/afs/protocol_yfs.h @@ -31,9 +31,9 @@ enum YFS_CM_Operations { }; enum YFS_FS_Operations { - YFSFETCHACL = 64131, /* YFS Fetch file ACL */ + YFSFETCHACL = 64131, /* YFS Fetch file AFS3 ACL */ YFSFETCHSTATUS = 64132, /* YFS Fetch file status */ - YFSSTOREACL = 64134, /* YFS Store file ACL */ + YFSSTOREACL = 64134, /* YFS Store file AFS3 ACL */ YFSSTORESTATUS = 64135, /* YFS Store file status */ YFSREMOVEFILE = 64136, /* YFS Remove a file */ YFSCREATEFILE = 64137, /* YFS Create a file */ @@ -49,7 +49,7 @@ enum YFS_FS_Operations { YFSRELEASELOCK = 64158, /* YFS Release a file lock */ YFSLOOKUP = 64161, /* YFS lookup file in directory */ YFSFLUSHCPS = 64165, - YFSFETCHOPAQUEACL = 64168, + YFSFETCHOPAQUEACL = 64168, /* YFS Fetch file YFS ACL */ YFSWHOAMI = 64170, YFSREMOVEACL = 64171, YFSREMOVEFILE2 = 64173, diff --git a/fs/afs/rxrpc.c b/fs/afs/rxrpc.c index 3ed2c99c58ab..a34a89c75c6a 100644 --- a/fs/afs/rxrpc.c +++ b/fs/afs/rxrpc.c @@ -21,7 +21,6 @@ struct workqueue_struct *afs_async_calls; static void afs_wake_up_call_waiter(struct sock *, struct rxrpc_call *, unsigned long); -static long afs_wait_for_call_to_complete(struct afs_call *, struct afs_addr_cursor *); static void afs_wake_up_async_call(struct sock *, struct rxrpc_call *, unsigned long); static void afs_delete_async_call(struct work_struct *); static void afs_process_async_call(struct work_struct *); @@ -361,10 +360,10 @@ static int afs_send_pages(struct afs_call *call, struct msghdr *msg) } /* - * initiate a call + * Initiate a call and synchronously queue up the parameters for dispatch. Any + * error is stored into the call struct, which the caller must check for. */ -long afs_make_call(struct afs_addr_cursor *ac, struct afs_call *call, - gfp_t gfp, bool async) +void afs_make_call(struct afs_addr_cursor *ac, struct afs_call *call, gfp_t gfp) { struct sockaddr_rxrpc *srx = &ac->alist->addrs[ac->index]; struct rxrpc_call *rxcall; @@ -382,7 +381,6 @@ long afs_make_call(struct afs_addr_cursor *ac, struct afs_call *call, call, call->type->name, key_serial(call->key), atomic_read(&call->net->nr_outstanding_calls)); - call->async = async; call->addr_ix = ac->index; call->alist = afs_get_addrlist(ac->alist); @@ -415,7 +413,7 @@ long afs_make_call(struct afs_addr_cursor *ac, struct afs_call *call, rxcall = rxrpc_kernel_begin_call(call->net->socket, srx, call->key, (unsigned long)call, tx_total_len, gfp, - (async ? + (call->async ? afs_wake_up_async_call : afs_wake_up_call_waiter), call->upgrade, @@ -453,13 +451,11 @@ long afs_make_call(struct afs_addr_cursor *ac, struct afs_call *call, /* Note that at this point, we may have received the reply or an abort * - and an asynchronous call may already have completed. + * + * afs_wait_for_call_to_complete(call, ac) + * must be called to synchronously clean up. */ - if (call->async) { - afs_put_call(call); - return -EINPROGRESS; - } - - return afs_wait_for_call_to_complete(call, ac); + return; error_do_abort: if (ret != -ECONNABORTED) { @@ -495,9 +491,7 @@ error_kill_call: ac->error = ret; call->state = AFS_CALL_COMPLETE; - afs_put_call(call); _leave(" = %d", ret); - return ret; } /* @@ -604,10 +598,10 @@ call_complete: } /* - * wait synchronously for a call to complete + * Wait synchronously for a call to complete and clean up the call struct. */ -static long afs_wait_for_call_to_complete(struct afs_call *call, - struct afs_addr_cursor *ac) +long afs_wait_for_call_to_complete(struct afs_call *call, + struct afs_addr_cursor *ac) { signed long rtt2, timeout; long ret; @@ -620,6 +614,10 @@ static long afs_wait_for_call_to_complete(struct afs_call *call, _enter(""); + ret = call->error; + if (ret < 0) + goto out; + rtt = rxrpc_kernel_get_rtt(call->net->socket, call->rxcall); rtt2 = nsecs_to_jiffies64(rtt) * 2; if (rtt2 < 2) @@ -703,6 +701,7 @@ static long afs_wait_for_call_to_complete(struct afs_call *call, break; } +out: _debug("call complete"); afs_put_call(call); _leave(" = %p", (void *)ret); diff --git a/fs/afs/super.c b/fs/afs/super.c index bab89763119b..783c68cd1a35 100644 --- a/fs/afs/super.c +++ b/fs/afs/super.c @@ -46,7 +46,7 @@ struct file_system_type afs_fs_type = { .init_fs_context = afs_init_fs_context, .parameters = &afs_fs_parameters, .kill_sb = afs_kill_super, - .fs_flags = 0, + .fs_flags = FS_RENAME_DOES_D_MOVE, }; MODULE_ALIAS_FS("afs"); @@ -69,19 +69,30 @@ static atomic_t afs_count_active_inodes; enum afs_param { Opt_autocell, Opt_dyn, + Opt_flock, Opt_source, }; static const struct fs_parameter_spec afs_param_specs[] = { fsparam_flag ("autocell", Opt_autocell), fsparam_flag ("dyn", Opt_dyn), + fsparam_enum ("flock", Opt_flock), fsparam_string("source", Opt_source), {} }; +static const struct fs_parameter_enum afs_param_enums[] = { + { Opt_flock, "local", afs_flock_mode_local }, + { Opt_flock, "openafs", afs_flock_mode_openafs }, + { Opt_flock, "strict", afs_flock_mode_strict }, + { Opt_flock, "write", afs_flock_mode_write }, + {} +}; + static const struct fs_parameter_description afs_fs_parameters = { .name = "kAFS", .specs = afs_param_specs, + .enums = afs_param_enums, }; /* @@ -184,11 +195,22 @@ static int afs_show_devname(struct seq_file *m, struct dentry *root) static int afs_show_options(struct seq_file *m, struct dentry *root) { struct afs_super_info *as = AFS_FS_S(root->d_sb); + const char *p = NULL; if (as->dyn_root) seq_puts(m, ",dyn"); if (test_bit(AFS_VNODE_AUTOCELL, &AFS_FS_I(d_inode(root))->flags)) seq_puts(m, ",autocell"); + switch (as->flock_mode) { + case afs_flock_mode_unset: break; + case afs_flock_mode_local: p = "local"; break; + case afs_flock_mode_openafs: p = "openafs"; break; + case afs_flock_mode_strict: p = "strict"; break; + case afs_flock_mode_write: p = "write"; break; + } + if (p) + seq_printf(m, ",flock=%s", p); + return 0; } @@ -317,6 +339,10 @@ static int afs_parse_param(struct fs_context *fc, struct fs_parameter *param) ctx->dyn_root = true; break; + case Opt_flock: + ctx->flock_mode = result.uint_32; + break; + default: return -EINVAL; } @@ -429,7 +455,7 @@ static int afs_fill_super(struct super_block *sb, struct afs_fs_context *ctx) fid.vnode = 1; fid.vnode_hi = 0; fid.unique = 1; - inode = afs_iget(sb, ctx->key, &fid, NULL, NULL, NULL); + inode = afs_iget(sb, ctx->key, &fid, NULL, NULL, NULL, NULL); } if (IS_ERR(inode)) @@ -468,6 +494,7 @@ static struct afs_super_info *afs_alloc_sbi(struct fs_context *fc) as = kzalloc(sizeof(struct afs_super_info), GFP_KERNEL); if (as) { as->net_ns = get_net(fc->net_ns); + as->flock_mode = ctx->flock_mode; if (ctx->dyn_root) { as->dyn_root = true; } else { @@ -552,6 +579,7 @@ static int afs_get_tree(struct fs_context *fc) } fc->root = dget(sb->s_root); + trace_afs_get_tree(as->cell, as->volume); _leave(" = 0 [%p]", sb); return 0; @@ -658,6 +686,8 @@ static struct inode *afs_alloc_inode(struct super_block *sb) vnode->cb_type = 0; vnode->lock_state = AFS_VNODE_LOCK_NONE; + init_rwsem(&vnode->rmdir_lock); + _leave(" = %p", &vnode->vfs_inode); return &vnode->vfs_inode; } diff --git a/fs/afs/vl_probe.c b/fs/afs/vl_probe.c index f402ee8171a1..b05e0de04f42 100644 --- a/fs/afs/vl_probe.c +++ b/fs/afs/vl_probe.c @@ -141,8 +141,8 @@ static bool afs_do_probe_vlserver(struct afs_net *net, struct afs_addr_cursor ac = { .index = 0, }; + struct afs_call *call; bool in_progress = false; - int err; _enter("%s", server->name); @@ -156,12 +156,14 @@ static bool afs_do_probe_vlserver(struct afs_net *net, server->probe.rtt = UINT_MAX; for (ac.index = 0; ac.index < ac.alist->nr_addrs; ac.index++) { - err = afs_vl_get_capabilities(net, &ac, key, server, - server_index, true); - if (err == -EINPROGRESS) + call = afs_vl_get_capabilities(net, &ac, key, server, + server_index); + if (!IS_ERR(call)) { + afs_put_call(call); in_progress = true; - else - afs_prioritise_error(_e, err, ac.abort_code); + } else { + afs_prioritise_error(_e, PTR_ERR(call), ac.abort_code); + } } if (!in_progress) diff --git a/fs/afs/vlclient.c b/fs/afs/vlclient.c index b0175b3ef0e8..dd9ba4e96fb3 100644 --- a/fs/afs/vlclient.c +++ b/fs/afs/vlclient.c @@ -167,7 +167,8 @@ struct afs_vldb_entry *afs_vl_get_entry_by_name_u(struct afs_vl_cursor *vc, memset((void *)bp + volnamesz, 0, padsz); trace_afs_make_vl_call(call); - return (struct afs_vldb_entry *)afs_make_call(&vc->ac, call, GFP_KERNEL, false); + afs_make_call(&vc->ac, call, GFP_KERNEL); + return (struct afs_vldb_entry *)afs_wait_for_call_to_complete(call, &vc->ac); } /* @@ -306,7 +307,8 @@ struct afs_addr_list *afs_vl_get_addrs_u(struct afs_vl_cursor *vc, r->uuid.node[i] = htonl(u->node[i]); trace_afs_make_vl_call(call); - return (struct afs_addr_list *)afs_make_call(&vc->ac, call, GFP_KERNEL, false); + afs_make_call(&vc->ac, call, GFP_KERNEL); + return (struct afs_addr_list *)afs_wait_for_call_to_complete(call, &vc->ac); } /* @@ -380,12 +382,11 @@ static const struct afs_call_type afs_RXVLGetCapabilities = { * We use this to probe for service upgrade to determine what the server at the * other end supports. */ -int afs_vl_get_capabilities(struct afs_net *net, - struct afs_addr_cursor *ac, - struct key *key, - struct afs_vlserver *server, - unsigned int server_index, - bool async) +struct afs_call *afs_vl_get_capabilities(struct afs_net *net, + struct afs_addr_cursor *ac, + struct key *key, + struct afs_vlserver *server, + unsigned int server_index) { struct afs_call *call; __be32 *bp; @@ -394,13 +395,14 @@ int afs_vl_get_capabilities(struct afs_net *net, call = afs_alloc_flat_call(net, &afs_RXVLGetCapabilities, 1 * 4, 16 * 4); if (!call) - return -ENOMEM; + return ERR_PTR(-ENOMEM); call->key = key; call->reply[0] = afs_get_vlserver(server); call->reply[1] = (void *)(long)server_index; call->upgrade = true; call->want_reply_time = true; + call->async = true; /* marshall the parameters */ bp = call->request; @@ -408,7 +410,8 @@ int afs_vl_get_capabilities(struct afs_net *net, /* Can't take a ref on server */ trace_afs_make_vl_call(call); - return afs_make_call(ac, call, GFP_KERNEL, async); + afs_make_call(ac, call, GFP_KERNEL); + return call; } /* @@ -651,5 +654,6 @@ struct afs_addr_list *afs_yfsvl_get_endpoints(struct afs_vl_cursor *vc, memcpy(bp, uuid, sizeof(*uuid)); /* Type opr_uuid */ trace_afs_make_vl_call(call); - return (struct afs_addr_list *)afs_make_call(&vc->ac, call, GFP_KERNEL, false); + afs_make_call(&vc->ac, call, GFP_KERNEL); + return (struct afs_addr_list *)afs_wait_for_call_to_complete(call, &vc->ac); } diff --git a/fs/afs/xattr.c b/fs/afs/xattr.c index a2cdf25573e2..c81f85003fc7 100644 --- a/fs/afs/xattr.c +++ b/fs/afs/xattr.c @@ -16,9 +16,14 @@ #include "internal.h" static const char afs_xattr_list[] = + "afs.acl\0" "afs.cell\0" "afs.fid\0" - "afs.volume"; + "afs.volume\0" + "afs.yfs.acl\0" + "afs.yfs.acl_inherited\0" + "afs.yfs.acl_num_cleaned\0" + "afs.yfs.vol_acl"; /* * Retrieve a list of the supported xattrs. @@ -34,6 +39,248 @@ ssize_t afs_listxattr(struct dentry *dentry, char *buffer, size_t size) } /* + * Get a file's ACL. + */ +static int afs_xattr_get_acl(const struct xattr_handler *handler, + struct dentry *dentry, + struct inode *inode, const char *name, + void *buffer, size_t size) +{ + struct afs_fs_cursor fc; + struct afs_vnode *vnode = AFS_FS_I(inode); + struct afs_acl *acl = NULL; + struct key *key; + int ret; + + key = afs_request_key(vnode->volume->cell); + if (IS_ERR(key)) + return PTR_ERR(key); + + ret = -ERESTARTSYS; + if (afs_begin_vnode_operation(&fc, vnode, key)) { + while (afs_select_fileserver(&fc)) { + fc.cb_break = afs_calc_vnode_cb_break(vnode); + acl = afs_fs_fetch_acl(&fc); + } + + afs_check_for_remote_deletion(&fc, fc.vnode); + afs_vnode_commit_status(&fc, vnode, fc.cb_break); + ret = afs_end_vnode_operation(&fc); + } + + if (ret == 0) { + ret = acl->size; + if (size > 0) { + ret = -ERANGE; + if (acl->size > size) + return -ERANGE; + memcpy(buffer, acl->data, acl->size); + ret = acl->size; + } + kfree(acl); + } + + key_put(key); + return ret; +} + +/* + * Set a file's AFS3 ACL. + */ +static int afs_xattr_set_acl(const struct xattr_handler *handler, + struct dentry *dentry, + struct inode *inode, const char *name, + const void *buffer, size_t size, int flags) +{ + struct afs_fs_cursor fc; + struct afs_vnode *vnode = AFS_FS_I(inode); + struct afs_acl *acl = NULL; + struct key *key; + int ret; + + if (flags == XATTR_CREATE) + return -EINVAL; + + key = afs_request_key(vnode->volume->cell); + if (IS_ERR(key)) + return PTR_ERR(key); + + acl = kmalloc(sizeof(*acl) + size, GFP_KERNEL); + if (!acl) { + key_put(key); + return -ENOMEM; + } + + acl->size = size; + memcpy(acl->data, buffer, size); + + ret = -ERESTARTSYS; + if (afs_begin_vnode_operation(&fc, vnode, key)) { + while (afs_select_fileserver(&fc)) { + fc.cb_break = afs_calc_vnode_cb_break(vnode); + afs_fs_store_acl(&fc, acl); + } + + afs_check_for_remote_deletion(&fc, fc.vnode); + afs_vnode_commit_status(&fc, vnode, fc.cb_break); + ret = afs_end_vnode_operation(&fc); + } + + kfree(acl); + key_put(key); + return ret; +} + +static const struct xattr_handler afs_xattr_afs_acl_handler = { + .name = "afs.acl", + .get = afs_xattr_get_acl, + .set = afs_xattr_set_acl, +}; + +/* + * Get a file's YFS ACL. + */ +static int afs_xattr_get_yfs(const struct xattr_handler *handler, + struct dentry *dentry, + struct inode *inode, const char *name, + void *buffer, size_t size) +{ + struct afs_fs_cursor fc; + struct afs_vnode *vnode = AFS_FS_I(inode); + struct yfs_acl *yacl = NULL; + struct key *key; + unsigned int flags = 0; + char buf[16], *data; + int which = 0, dsize, ret; + + if (strcmp(name, "acl") == 0) + which = 0; + else if (strcmp(name, "acl_inherited") == 0) + which = 1; + else if (strcmp(name, "acl_num_cleaned") == 0) + which = 2; + else if (strcmp(name, "vol_acl") == 0) + which = 3; + else + return -EOPNOTSUPP; + + if (which == 0) + flags |= YFS_ACL_WANT_ACL; + else if (which == 3) + flags |= YFS_ACL_WANT_VOL_ACL; + + key = afs_request_key(vnode->volume->cell); + if (IS_ERR(key)) + return PTR_ERR(key); + + ret = -ERESTARTSYS; + if (afs_begin_vnode_operation(&fc, vnode, key)) { + while (afs_select_fileserver(&fc)) { + fc.cb_break = afs_calc_vnode_cb_break(vnode); + yacl = yfs_fs_fetch_opaque_acl(&fc, flags); + } + + afs_check_for_remote_deletion(&fc, fc.vnode); + afs_vnode_commit_status(&fc, vnode, fc.cb_break); + ret = afs_end_vnode_operation(&fc); + } + + if (ret == 0) { + switch (which) { + case 0: + data = yacl->acl->data; + dsize = yacl->acl->size; + break; + case 1: + data = buf; + dsize = snprintf(buf, sizeof(buf), "%u", + yacl->inherit_flag); + break; + case 2: + data = buf; + dsize = snprintf(buf, sizeof(buf), "%u", + yacl->num_cleaned); + break; + case 3: + data = yacl->vol_acl->data; + dsize = yacl->vol_acl->size; + break; + default: + ret = -EOPNOTSUPP; + goto out; + } + + ret = dsize; + if (size > 0) { + if (dsize > size) { + ret = -ERANGE; + goto out; + } + memcpy(buffer, data, dsize); + } + } + +out: + yfs_free_opaque_acl(yacl); + key_put(key); + return ret; +} + +/* + * Set a file's YFS ACL. + */ +static int afs_xattr_set_yfs(const struct xattr_handler *handler, + struct dentry *dentry, + struct inode *inode, const char *name, + const void *buffer, size_t size, int flags) +{ + struct afs_fs_cursor fc; + struct afs_vnode *vnode = AFS_FS_I(inode); + struct afs_acl *acl = NULL; + struct key *key; + int ret; + + if (flags == XATTR_CREATE || + strcmp(name, "acl") != 0) + return -EINVAL; + + key = afs_request_key(vnode->volume->cell); + if (IS_ERR(key)) + return PTR_ERR(key); + + acl = kmalloc(sizeof(*acl) + size, GFP_KERNEL); + if (!acl) { + key_put(key); + return -ENOMEM; + } + + acl->size = size; + memcpy(acl->data, buffer, size); + + ret = -ERESTARTSYS; + if (afs_begin_vnode_operation(&fc, vnode, key)) { + while (afs_select_fileserver(&fc)) { + fc.cb_break = afs_calc_vnode_cb_break(vnode); + yfs_fs_store_opaque_acl2(&fc, acl); + } + + afs_check_for_remote_deletion(&fc, fc.vnode); + afs_vnode_commit_status(&fc, vnode, fc.cb_break); + ret = afs_end_vnode_operation(&fc); + } + + kfree(acl); + key_put(key); + return ret; +} + +static const struct xattr_handler afs_xattr_yfs_handler = { + .prefix = "afs.yfs.", + .get = afs_xattr_get_yfs, + .set = afs_xattr_set_yfs, +}; + +/* * Get the name of the cell on which a file resides. */ static int afs_xattr_get_cell(const struct xattr_handler *handler, @@ -50,7 +297,7 @@ static int afs_xattr_get_cell(const struct xattr_handler *handler, return namelen; if (namelen > size) return -ERANGE; - memcpy(buffer, cell->name, size); + memcpy(buffer, cell->name, namelen); return namelen; } @@ -69,11 +316,20 @@ static int afs_xattr_get_fid(const struct xattr_handler *handler, void *buffer, size_t size) { struct afs_vnode *vnode = AFS_FS_I(inode); - char text[8 + 1 + 8 + 1 + 8 + 1]; + char text[16 + 1 + 24 + 1 + 8 + 1]; size_t len; - len = sprintf(text, "%llx:%llx:%x", - vnode->fid.vid, vnode->fid.vnode, vnode->fid.unique); + /* The volume ID is 64-bit, the vnode ID is 96-bit and the + * uniquifier is 32-bit. + */ + len = sprintf(text, "%llx:", vnode->fid.vid); + if (vnode->fid.vnode_hi) + len += sprintf(text + len, "%x%016llx", + vnode->fid.vnode_hi, vnode->fid.vnode); + else + len += sprintf(text + len, "%llx", vnode->fid.vnode); + len += sprintf(text + len, ":%x", vnode->fid.unique); + if (size == 0) return len; if (len > size) @@ -104,7 +360,7 @@ static int afs_xattr_get_volume(const struct xattr_handler *handler, return namelen; if (namelen > size) return -ERANGE; - memcpy(buffer, volname, size); + memcpy(buffer, volname, namelen); return namelen; } @@ -114,8 +370,10 @@ static const struct xattr_handler afs_xattr_afs_volume_handler = { }; const struct xattr_handler *afs_xattr_handlers[] = { + &afs_xattr_afs_acl_handler, &afs_xattr_afs_cell_handler, &afs_xattr_afs_fid_handler, &afs_xattr_afs_volume_handler, + &afs_xattr_yfs_handler, /* afs.yfs. prefix */ NULL }; diff --git a/fs/afs/yfsclient.c b/fs/afs/yfsclient.c index 871e29f06257..6cf7d161baa1 100644 --- a/fs/afs/yfsclient.c +++ b/fs/afs/yfsclient.c @@ -519,7 +519,8 @@ int yfs_fs_fetch_file_status(struct afs_fs_cursor *fc, struct afs_volsync *volsy call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -711,7 +712,8 @@ int yfs_fs_fetch_data(struct afs_fs_cursor *fc, struct afs_read *req) call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -810,8 +812,9 @@ int yfs_fs_create_file(struct afs_fs_cursor *fc, yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &vnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } static const struct afs_call_type yfs_RXFSMakeDir = { @@ -873,8 +876,9 @@ int yfs_fs_make_dir(struct afs_fs_cursor *fc, yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &vnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -963,8 +967,9 @@ int yfs_fs_remove_file2(struct afs_fs_cursor *fc, struct afs_vnode *vnode, yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &dvnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &dvnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1050,8 +1055,9 @@ int yfs_fs_remove(struct afs_fs_cursor *fc, struct afs_vnode *vnode, yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &dvnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &dvnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1135,8 +1141,9 @@ int yfs_fs_link(struct afs_fs_cursor *fc, struct afs_vnode *vnode, yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &vnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1231,8 +1238,9 @@ int yfs_fs_symlink(struct afs_fs_cursor *fc, yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &dvnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call1(call, &dvnode->fid, name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1329,8 +1337,9 @@ int yfs_fs_rename(struct afs_fs_cursor *fc, yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &orig_dvnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_call2(call, &orig_dvnode->fid, orig_name, new_name); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1436,7 +1445,8 @@ int yfs_fs_store_data(struct afs_fs_cursor *fc, struct address_space *mapping, afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1524,7 +1534,8 @@ static int yfs_fs_setattr_size(struct afs_fs_cursor *fc, struct iattr *attr) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1567,7 +1578,8 @@ int yfs_fs_setattr(struct afs_fs_cursor *fc, struct iattr *attr) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1755,13 +1767,15 @@ int yfs_fs_get_volume_status(struct afs_fs_cursor *fc, afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* - * Deliver reply data to an YFS.SetLock, YFS.ExtendLock or YFS.ReleaseLock + * Deliver reply data to operations that just return a file status and a volume + * sync record. */ -static int yfs_deliver_fs_xxxx_lock(struct afs_call *call) +static int yfs_deliver_status_and_volsync(struct afs_call *call) { struct afs_vnode *vnode = call->reply[0]; const __be32 *bp; @@ -1791,7 +1805,8 @@ static int yfs_deliver_fs_xxxx_lock(struct afs_call *call) static const struct afs_call_type yfs_RXYFSSetLock = { .name = "YFS.SetLock", .op = yfs_FS_SetLock, - .deliver = yfs_deliver_fs_xxxx_lock, + .deliver = yfs_deliver_status_and_volsync, + .done = afs_lock_op_done, .destructor = afs_flat_call_destructor, }; @@ -1801,7 +1816,8 @@ static const struct afs_call_type yfs_RXYFSSetLock = { static const struct afs_call_type yfs_RXYFSExtendLock = { .name = "YFS.ExtendLock", .op = yfs_FS_ExtendLock, - .deliver = yfs_deliver_fs_xxxx_lock, + .deliver = yfs_deliver_status_and_volsync, + .done = afs_lock_op_done, .destructor = afs_flat_call_destructor, }; @@ -1811,7 +1827,7 @@ static const struct afs_call_type yfs_RXYFSExtendLock = { static const struct afs_call_type yfs_RXYFSReleaseLock = { .name = "YFS.ReleaseLock", .op = yfs_FS_ReleaseLock, - .deliver = yfs_deliver_fs_xxxx_lock, + .deliver = yfs_deliver_status_and_volsync, .destructor = afs_flat_call_destructor, }; @@ -1838,6 +1854,7 @@ int yfs_fs_set_lock(struct afs_fs_cursor *fc, afs_lock_type_t type) call->key = fc->key; call->reply[0] = vnode; + call->want_reply_time = true; /* marshall the parameters */ bp = call->request; @@ -1848,8 +1865,9 @@ int yfs_fs_set_lock(struct afs_fs_cursor *fc, afs_lock_type_t type) yfs_check_req(call, bp); afs_use_fs_server(call, fc->cbi); - trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + trace_afs_make_fs_calli(call, &vnode->fid, type); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1874,6 +1892,7 @@ int yfs_fs_extend_lock(struct afs_fs_cursor *fc) call->key = fc->key; call->reply[0] = vnode; + call->want_reply_time = true; /* marshall the parameters */ bp = call->request; @@ -1884,7 +1903,8 @@ int yfs_fs_extend_lock(struct afs_fs_cursor *fc) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -1919,7 +1939,8 @@ int yfs_fs_release_lock(struct afs_fs_cursor *fc) afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &vnode->fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -2007,7 +2028,8 @@ int yfs_fs_fetch_status(struct afs_fs_cursor *fc, call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, fid); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); } /* @@ -2190,5 +2212,250 @@ int yfs_fs_inline_bulk_status(struct afs_fs_cursor *fc, call->cb_break = fc->cb_break; afs_use_fs_server(call, fc->cbi); trace_afs_make_fs_call(call, &fids[0]); - return afs_make_call(&fc->ac, call, GFP_NOFS, false); + afs_make_call(&fc->ac, call, GFP_NOFS); + return afs_wait_for_call_to_complete(call, &fc->ac); +} + +/* + * Deliver reply data to an YFS.FetchOpaqueACL. + */ +static int yfs_deliver_fs_fetch_opaque_acl(struct afs_call *call) +{ + struct afs_volsync *volsync = call->reply[2]; + struct afs_vnode *vnode = call->reply[1]; + struct yfs_acl *yacl = call->reply[0]; + struct afs_acl *acl; + const __be32 *bp; + unsigned int size; + int ret; + + _enter("{%u}", call->unmarshall); + + switch (call->unmarshall) { + case 0: + afs_extract_to_tmp(call); + call->unmarshall++; + + /* Extract the file ACL length */ + case 1: + ret = afs_extract_data(call, true); + if (ret < 0) + return ret; + + size = call->count2 = ntohl(call->tmp); + size = round_up(size, 4); + + if (yacl->flags & YFS_ACL_WANT_ACL) { + acl = kmalloc(struct_size(acl, data, size), GFP_KERNEL); + if (!acl) + return -ENOMEM; + yacl->acl = acl; + acl->size = call->count2; + afs_extract_begin(call, acl->data, size); + } else { + iov_iter_discard(&call->iter, READ, size); + } + call->unmarshall++; + + /* Extract the file ACL */ + case 2: + ret = afs_extract_data(call, true); + if (ret < 0) + return ret; + + afs_extract_to_tmp(call); + call->unmarshall++; + + /* Extract the volume ACL length */ + case 3: + ret = afs_extract_data(call, true); + if (ret < 0) + return ret; + + size = call->count2 = ntohl(call->tmp); + size = round_up(size, 4); + + if (yacl->flags & YFS_ACL_WANT_VOL_ACL) { + acl = kmalloc(struct_size(acl, data, size), GFP_KERNEL); + if (!acl) + return -ENOMEM; + yacl->vol_acl = acl; + acl->size = call->count2; + afs_extract_begin(call, acl->data, size); + } else { + iov_iter_discard(&call->iter, READ, size); + } + call->unmarshall++; + + /* Extract the volume ACL */ + case 4: + ret = afs_extract_data(call, true); + if (ret < 0) + return ret; + + afs_extract_to_buf(call, + sizeof(__be32) * 2 + + sizeof(struct yfs_xdr_YFSFetchStatus) + + sizeof(struct yfs_xdr_YFSVolSync)); + call->unmarshall++; + + /* extract the metadata */ + case 5: + ret = afs_extract_data(call, false); + if (ret < 0) + return ret; + + bp = call->buffer; + yacl->inherit_flag = ntohl(*bp++); + yacl->num_cleaned = ntohl(*bp++); + ret = yfs_decode_status(call, &bp, &vnode->status, vnode, + &call->expected_version, NULL); + if (ret < 0) + return ret; + xdr_decode_YFSVolSync(&bp, volsync); + + call->unmarshall++; + + case 6: + break; + } + + _leave(" = 0 [done]"); + return 0; +} + +void yfs_free_opaque_acl(struct yfs_acl *yacl) +{ + if (yacl) { + kfree(yacl->acl); + kfree(yacl->vol_acl); + kfree(yacl); + } +} + +static void yfs_destroy_fs_fetch_opaque_acl(struct afs_call *call) +{ + yfs_free_opaque_acl(call->reply[0]); + afs_flat_call_destructor(call); +} + +/* + * YFS.FetchOpaqueACL operation type + */ +static const struct afs_call_type yfs_RXYFSFetchOpaqueACL = { + .name = "YFS.FetchOpaqueACL", + .op = yfs_FS_FetchOpaqueACL, + .deliver = yfs_deliver_fs_fetch_opaque_acl, + .destructor = yfs_destroy_fs_fetch_opaque_acl, +}; + +/* + * Fetch the YFS advanced ACLs for a file. + */ +struct yfs_acl *yfs_fs_fetch_opaque_acl(struct afs_fs_cursor *fc, + unsigned int flags) +{ + struct afs_vnode *vnode = fc->vnode; + struct afs_call *call; + struct yfs_acl *yacl; + struct afs_net *net = afs_v2net(vnode); + __be32 *bp; + + _enter(",%x,{%llx:%llu},,", + key_serial(fc->key), vnode->fid.vid, vnode->fid.vnode); + + call = afs_alloc_flat_call(net, &yfs_RXYFSFetchOpaqueACL, + sizeof(__be32) * 2 + + sizeof(struct yfs_xdr_YFSFid), + sizeof(__be32) * 2 + + sizeof(struct yfs_xdr_YFSFetchStatus) + + sizeof(struct yfs_xdr_YFSVolSync)); + if (!call) + goto nomem; + + yacl = kzalloc(sizeof(struct yfs_acl), GFP_KERNEL); + if (!yacl) + goto nomem_call; + + yacl->flags = flags; + call->key = fc->key; + call->reply[0] = yacl; + call->reply[1] = vnode; + call->reply[2] = NULL; /* volsync */ + call->ret_reply0 = true; + + /* marshall the parameters */ + bp = call->request; + bp = xdr_encode_u32(bp, YFSFETCHOPAQUEACL); + bp = xdr_encode_u32(bp, 0); /* RPC flags */ + bp = xdr_encode_YFSFid(bp, &vnode->fid); + yfs_check_req(call, bp); + + call->cb_break = fc->cb_break; + afs_use_fs_server(call, fc->cbi); + trace_afs_make_fs_call(call, &vnode->fid); + afs_make_call(&fc->ac, call, GFP_KERNEL); + return (struct yfs_acl *)afs_wait_for_call_to_complete(call, &fc->ac); + +nomem_call: + afs_put_call(call); +nomem: + fc->ac.error = -ENOMEM; + return ERR_PTR(-ENOMEM); +} + +/* + * YFS.StoreOpaqueACL2 operation type + */ +static const struct afs_call_type yfs_RXYFSStoreOpaqueACL2 = { + .name = "YFS.StoreOpaqueACL2", + .op = yfs_FS_StoreOpaqueACL2, + .deliver = yfs_deliver_status_and_volsync, + .destructor = afs_flat_call_destructor, +}; + +/* + * Fetch the YFS ACL for a file. + */ +int yfs_fs_store_opaque_acl2(struct afs_fs_cursor *fc, const struct afs_acl *acl) +{ + struct afs_vnode *vnode = fc->vnode; + struct afs_call *call; + struct afs_net *net = afs_v2net(vnode); + size_t size; + __be32 *bp; + + _enter(",%x,{%llx:%llu},,", + key_serial(fc->key), vnode->fid.vid, vnode->fid.vnode); + + size = round_up(acl->size, 4); + call = afs_alloc_flat_call(net, &yfs_RXYFSStoreStatus, + sizeof(__be32) * 2 + + sizeof(struct yfs_xdr_YFSFid) + + sizeof(__be32) + size, + sizeof(struct yfs_xdr_YFSFetchStatus) + + sizeof(struct yfs_xdr_YFSVolSync)); + if (!call) { + fc->ac.error = -ENOMEM; + return -ENOMEM; + } + + call->key = fc->key; + call->reply[0] = vnode; + call->reply[2] = NULL; /* volsync */ + + /* marshall the parameters */ + bp = call->request; + bp = xdr_encode_u32(bp, YFSSTOREOPAQUEACL2); + bp = xdr_encode_u32(bp, 0); /* RPC flags */ + bp = xdr_encode_YFSFid(bp, &vnode->fid); + bp = xdr_encode_u32(bp, acl->size); + memcpy(bp, acl->data, acl->size); + if (acl->size != size) + memset((void *)bp + acl->size, 0, size - acl->size); + yfs_check_req(call, bp); + + trace_afs_make_fs_call(call, &vnode->fid); + afs_make_call(&fc->ac, call, GFP_KERNEL); + return afs_wait_for_call_to_complete(call, &fc->ac); } diff --git a/include/linux/fs.h b/include/linux/fs.h index 49d048ea0afb..0fa5b47509d5 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1098,6 +1098,7 @@ struct file_lock { struct { struct list_head link; /* link in AFS vnode's pending_locks list */ int state; /* state of grant or error if -ve */ + unsigned int debug_id; } afs; } fl_u; } __randomize_layout; diff --git a/include/trace/events/afs.h b/include/trace/events/afs.h index e3f005eae1f7..562f854ac4bf 100644 --- a/include/trace/events/afs.h +++ b/include/trace/events/afs.h @@ -33,8 +33,10 @@ enum afs_call_trace { enum afs_fs_operation { afs_FS_FetchData = 130, /* AFS Fetch file data */ + afs_FS_FetchACL = 131, /* AFS Fetch file ACL */ afs_FS_FetchStatus = 132, /* AFS Fetch file status */ afs_FS_StoreData = 133, /* AFS Store file data */ + afs_FS_StoreACL = 134, /* AFS Store file ACL */ afs_FS_StoreStatus = 135, /* AFS Store file status */ afs_FS_RemoveFile = 136, /* AFS Remove a file */ afs_FS_CreateFile = 137, /* AFS Create a file */ @@ -108,8 +110,12 @@ enum afs_edit_dir_reason { afs_edit_dir_for_create, afs_edit_dir_for_link, afs_edit_dir_for_mkdir, - afs_edit_dir_for_rename, + afs_edit_dir_for_rename_0, + afs_edit_dir_for_rename_1, + afs_edit_dir_for_rename_2, afs_edit_dir_for_rmdir, + afs_edit_dir_for_silly_0, + afs_edit_dir_for_silly_1, afs_edit_dir_for_symlink, afs_edit_dir_for_unlink, }; @@ -152,6 +158,43 @@ enum afs_file_error { afs_file_error_writeback_fail, }; +enum afs_flock_event { + afs_flock_acquired, + afs_flock_callback_break, + afs_flock_defer_unlock, + afs_flock_extend_fail, + afs_flock_fail_other, + afs_flock_fail_perm, + afs_flock_no_lockers, + afs_flock_release_fail, + afs_flock_silly_delete, + afs_flock_timestamp, + afs_flock_try_to_lock, + afs_flock_vfs_lock, + afs_flock_vfs_locking, + afs_flock_waited, + afs_flock_waiting, + afs_flock_work_extending, + afs_flock_work_retry, + afs_flock_work_unlocking, + afs_flock_would_block, +}; + +enum afs_flock_operation { + afs_flock_op_copy_lock, + afs_flock_op_flock, + afs_flock_op_grant, + afs_flock_op_lock, + afs_flock_op_release_lock, + afs_flock_op_return_ok, + afs_flock_op_return_eagain, + afs_flock_op_return_edeadlk, + afs_flock_op_return_error, + afs_flock_op_set_lock, + afs_flock_op_unlock, + afs_flock_op_wake, +}; + #endif /* end __AFS_DECLARE_TRACE_ENUMS_ONCE_ONLY */ /* @@ -237,8 +280,12 @@ enum afs_file_error { EM(afs_edit_dir_for_create, "Create") \ EM(afs_edit_dir_for_link, "Link ") \ EM(afs_edit_dir_for_mkdir, "MkDir ") \ - EM(afs_edit_dir_for_rename, "Rename") \ + EM(afs_edit_dir_for_rename_0, "Renam0") \ + EM(afs_edit_dir_for_rename_1, "Renam1") \ + EM(afs_edit_dir_for_rename_2, "Renam2") \ EM(afs_edit_dir_for_rmdir, "RmDir ") \ + EM(afs_edit_dir_for_silly_0, "S_Ren0") \ + EM(afs_edit_dir_for_silly_1, "S_Ren1") \ EM(afs_edit_dir_for_symlink, "Symlnk") \ E_(afs_edit_dir_for_unlink, "Unlink") @@ -277,6 +324,56 @@ enum afs_file_error { EM(afs_file_error_mntpt, "MNTPT_READ_FAILED") \ E_(afs_file_error_writeback_fail, "WRITEBACK_FAILED") +#define afs_flock_types \ + EM(F_RDLCK, "RDLCK") \ + EM(F_WRLCK, "WRLCK") \ + E_(F_UNLCK, "UNLCK") + +#define afs_flock_states \ + EM(AFS_VNODE_LOCK_NONE, "NONE") \ + EM(AFS_VNODE_LOCK_WAITING_FOR_CB, "WAIT_FOR_CB") \ + EM(AFS_VNODE_LOCK_SETTING, "SETTING") \ + EM(AFS_VNODE_LOCK_GRANTED, "GRANTED") \ + EM(AFS_VNODE_LOCK_EXTENDING, "EXTENDING") \ + EM(AFS_VNODE_LOCK_NEED_UNLOCK, "NEED_UNLOCK") \ + EM(AFS_VNODE_LOCK_UNLOCKING, "UNLOCKING") \ + E_(AFS_VNODE_LOCK_DELETED, "DELETED") + +#define afs_flock_events \ + EM(afs_flock_acquired, "Acquired") \ + EM(afs_flock_callback_break, "Callback") \ + EM(afs_flock_defer_unlock, "D-Unlock") \ + EM(afs_flock_extend_fail, "Ext_Fail") \ + EM(afs_flock_fail_other, "ErrOther") \ + EM(afs_flock_fail_perm, "ErrPerm ") \ + EM(afs_flock_no_lockers, "NoLocker") \ + EM(afs_flock_release_fail, "Rel_Fail") \ + EM(afs_flock_silly_delete, "SillyDel") \ + EM(afs_flock_timestamp, "Timestmp") \ + EM(afs_flock_try_to_lock, "TryToLck") \ + EM(afs_flock_vfs_lock, "VFSLock ") \ + EM(afs_flock_vfs_locking, "VFSLking") \ + EM(afs_flock_waited, "Waited ") \ + EM(afs_flock_waiting, "Waiting ") \ + EM(afs_flock_work_extending, "Extendng") \ + EM(afs_flock_work_retry, "Retry ") \ + EM(afs_flock_work_unlocking, "Unlcking") \ + E_(afs_flock_would_block, "EWOULDBL") + +#define afs_flock_operations \ + EM(afs_flock_op_copy_lock, "COPY ") \ + EM(afs_flock_op_flock, "->flock ") \ + EM(afs_flock_op_grant, "GRANT ") \ + EM(afs_flock_op_lock, "->lock ") \ + EM(afs_flock_op_release_lock, "RELEASE ") \ + EM(afs_flock_op_return_ok, "<-OK ") \ + EM(afs_flock_op_return_edeadlk, "<-EDEADL") \ + EM(afs_flock_op_return_eagain, "<-EAGAIN") \ + EM(afs_flock_op_return_error, "<-ERROR ") \ + EM(afs_flock_op_set_lock, "SET ") \ + EM(afs_flock_op_unlock, "UNLOCK ") \ + E_(afs_flock_op_wake, "WAKE ") + /* * Export enum symbols via userspace. */ @@ -293,6 +390,8 @@ afs_edit_dir_reasons; afs_eproto_causes; afs_io_errors; afs_file_errors; +afs_flock_types; +afs_flock_operations; /* * Now redefine the EM() and E_() macros to map the enums to the strings that @@ -442,6 +541,123 @@ TRACE_EVENT(afs_make_fs_call, __print_symbolic(__entry->op, afs_fs_operations)) ); +TRACE_EVENT(afs_make_fs_calli, + TP_PROTO(struct afs_call *call, const struct afs_fid *fid, + unsigned int i), + + TP_ARGS(call, fid, i), + + TP_STRUCT__entry( + __field(unsigned int, call ) + __field(unsigned int, i ) + __field(enum afs_fs_operation, op ) + __field_struct(struct afs_fid, fid ) + ), + + TP_fast_assign( + __entry->call = call->debug_id; + __entry->i = i; + __entry->op = call->operation_ID; + if (fid) { + __entry->fid = *fid; + } else { + __entry->fid.vid = 0; + __entry->fid.vnode = 0; + __entry->fid.unique = 0; + } + ), + + TP_printk("c=%08x %06llx:%06llx:%06x %s i=%u", + __entry->call, + __entry->fid.vid, + __entry->fid.vnode, + __entry->fid.unique, + __print_symbolic(__entry->op, afs_fs_operations), + __entry->i) + ); + +TRACE_EVENT(afs_make_fs_call1, + TP_PROTO(struct afs_call *call, const struct afs_fid *fid, + const char *name), + + TP_ARGS(call, fid, name), + + TP_STRUCT__entry( + __field(unsigned int, call ) + __field(enum afs_fs_operation, op ) + __field_struct(struct afs_fid, fid ) + __array(char, name, 24 ) + ), + + TP_fast_assign( + int __len = strlen(name); + __len = min(__len, 23); + __entry->call = call->debug_id; + __entry->op = call->operation_ID; + if (fid) { + __entry->fid = *fid; + } else { + __entry->fid.vid = 0; + __entry->fid.vnode = 0; + __entry->fid.unique = 0; + } + memcpy(__entry->name, name, __len); + __entry->name[__len] = 0; + ), + + TP_printk("c=%08x %06llx:%06llx:%06x %s \"%s\"", + __entry->call, + __entry->fid.vid, + __entry->fid.vnode, + __entry->fid.unique, + __print_symbolic(__entry->op, afs_fs_operations), + __entry->name) + ); + +TRACE_EVENT(afs_make_fs_call2, + TP_PROTO(struct afs_call *call, const struct afs_fid *fid, + const char *name, const char *name2), + + TP_ARGS(call, fid, name, name2), + + TP_STRUCT__entry( + __field(unsigned int, call ) + __field(enum afs_fs_operation, op ) + __field_struct(struct afs_fid, fid ) + __array(char, name, 24 ) + __array(char, name2, 24 ) + ), + + TP_fast_assign( + int __len = strlen(name); + int __len2 = strlen(name2); + __len = min(__len, 23); + __len2 = min(__len2, 23); + __entry->call = call->debug_id; + __entry->op = call->operation_ID; + if (fid) { + __entry->fid = *fid; + } else { + __entry->fid.vid = 0; + __entry->fid.vnode = 0; + __entry->fid.unique = 0; + } + memcpy(__entry->name, name, __len); + __entry->name[__len] = 0; + memcpy(__entry->name2, name2, __len2); + __entry->name2[__len2] = 0; + ), + + TP_printk("c=%08x %06llx:%06llx:%06x %s \"%s\" \"%s\"", + __entry->call, + __entry->fid.vid, + __entry->fid.vnode, + __entry->fid.unique, + __print_symbolic(__entry->op, afs_fs_operations), + __entry->name, + __entry->name2) + ); + TRACE_EVENT(afs_make_vl_call, TP_PROTO(struct afs_call *call), @@ -639,6 +855,38 @@ TRACE_EVENT(afs_call_state, __entry->ret, __entry->abort) ); +TRACE_EVENT(afs_lookup, + TP_PROTO(struct afs_vnode *dvnode, const struct qstr *name, + struct afs_vnode *vnode), + + TP_ARGS(dvnode, name, vnode), + + TP_STRUCT__entry( + __field_struct(struct afs_fid, dfid ) + __field_struct(struct afs_fid, fid ) + __array(char, name, 24 ) + ), + + TP_fast_assign( + int __len = min_t(int, name->len, 23); + __entry->dfid = dvnode->fid; + if (vnode) { + __entry->fid = vnode->fid; + } else { + __entry->fid.vid = 0; + __entry->fid.vnode = 0; + __entry->fid.unique = 0; + } + memcpy(__entry->name, name->name, __len); + __entry->name[__len] = 0; + ), + + TP_printk("d=%llx:%llx:%x \"%s\" f=%llx:%x", + __entry->dfid.vid, __entry->dfid.vnode, __entry->dfid.unique, + __entry->name, + __entry->fid.vnode, __entry->fid.unique) + ); + TRACE_EVENT(afs_edit_dir, TP_PROTO(struct afs_vnode *dvnode, enum afs_edit_dir_reason why, @@ -660,12 +908,12 @@ TRACE_EVENT(afs_edit_dir, __field(unsigned short, slot ) __field(unsigned int, f_vnode ) __field(unsigned int, f_unique ) - __array(char, name, 18 ) + __array(char, name, 24 ) ), TP_fast_assign( int __len = strlen(name); - __len = min(__len, 17); + __len = min(__len, 23); __entry->vnode = dvnode->fid.vnode; __entry->unique = dvnode->fid.unique; __entry->why = why; @@ -678,7 +926,7 @@ TRACE_EVENT(afs_edit_dir, __entry->name[__len] = 0; ), - TP_printk("d=%x:%x %s %s %u[%u] f=%x:%x %s", + TP_printk("d=%x:%x %s %s %u[%u] f=%x:%x \"%s\"", __entry->vnode, __entry->unique, __print_symbolic(__entry->why, afs_edit_dir_reasons), __print_symbolic(__entry->op, afs_edit_dir_ops), @@ -796,6 +1044,133 @@ TRACE_EVENT(afs_cm_no_server_u, __entry->call, __entry->op_id, &__entry->uuid) ); +TRACE_EVENT(afs_flock_ev, + TP_PROTO(struct afs_vnode *vnode, struct file_lock *fl, + enum afs_flock_event event, int error), + + TP_ARGS(vnode, fl, event, error), + + TP_STRUCT__entry( + __field_struct(struct afs_fid, fid ) + __field(enum afs_flock_event, event ) + __field(enum afs_lock_state, state ) + __field(int, error ) + __field(unsigned int, debug_id ) + ), + + TP_fast_assign( + __entry->fid = vnode->fid; + __entry->event = event; + __entry->state = vnode->lock_state; + __entry->error = error; + __entry->debug_id = fl ? fl->fl_u.afs.debug_id : 0; + ), + + TP_printk("%llx:%llx:%x %04x %s s=%s e=%d", + __entry->fid.vid, __entry->fid.vnode, __entry->fid.unique, + __entry->debug_id, + __print_symbolic(__entry->event, afs_flock_events), + __print_symbolic(__entry->state, afs_flock_states), + __entry->error) + ); + +TRACE_EVENT(afs_flock_op, + TP_PROTO(struct afs_vnode *vnode, struct file_lock *fl, + enum afs_flock_operation op), + + TP_ARGS(vnode, fl, op), + + TP_STRUCT__entry( + __field_struct(struct afs_fid, fid ) + __field(loff_t, from ) + __field(loff_t, len ) + __field(enum afs_flock_operation, op ) + __field(unsigned char, type ) + __field(unsigned int, flags ) + __field(unsigned int, debug_id ) + ), + + TP_fast_assign( + __entry->fid = vnode->fid; + __entry->from = fl->fl_start; + __entry->len = fl->fl_end - fl->fl_start + 1; + __entry->op = op; + __entry->type = fl->fl_type; + __entry->flags = fl->fl_flags; + __entry->debug_id = fl->fl_u.afs.debug_id; + ), + + TP_printk("%llx:%llx:%x %04x %s t=%s R=%llx/%llx f=%x", + __entry->fid.vid, __entry->fid.vnode, __entry->fid.unique, + __entry->debug_id, + __print_symbolic(__entry->op, afs_flock_operations), + __print_symbolic(__entry->type, afs_flock_types), + __entry->from, __entry->len, __entry->flags) + ); + +TRACE_EVENT(afs_reload_dir, + TP_PROTO(struct afs_vnode *vnode), + + TP_ARGS(vnode), + + TP_STRUCT__entry( + __field_struct(struct afs_fid, fid ) + ), + + TP_fast_assign( + __entry->fid = vnode->fid; + ), + + TP_printk("%llx:%llx:%x", + __entry->fid.vid, __entry->fid.vnode, __entry->fid.unique) + ); + +TRACE_EVENT(afs_silly_rename, + TP_PROTO(struct afs_vnode *vnode, bool done), + + TP_ARGS(vnode, done), + + TP_STRUCT__entry( + __field_struct(struct afs_fid, fid ) + __field(bool, done ) + ), + + TP_fast_assign( + __entry->fid = vnode->fid; + __entry->done = done; + ), + + TP_printk("%llx:%llx:%x done=%u", + __entry->fid.vid, __entry->fid.vnode, __entry->fid.unique, + __entry->done) + ); + +TRACE_EVENT(afs_get_tree, + TP_PROTO(struct afs_cell *cell, struct afs_volume *volume), + + TP_ARGS(cell, volume), + + TP_STRUCT__entry( + __field(u64, vid ) + __array(char, cell, 24 ) + __array(char, volume, 24 ) + ), + + TP_fast_assign( + int __len; + __entry->vid = volume->vid; + __len = min_t(int, cell->name_len, 23); + memcpy(__entry->cell, cell->name, __len); + __entry->cell[__len] = 0; + __len = min_t(int, volume->name_len, 23); + memcpy(__entry->volume, volume->name, __len); + __entry->volume[__len] = 0; + ), + + TP_printk("--- MOUNT %s:%s %llx", + __entry->cell, __entry->volume, __entry->vid) + ); + #endif /* _TRACE_AFS_H */ /* This part must be outside protection */ |