From: Linus Torvalds Date: Thu, 23 Oct 2008 17:22:40 +0000 (-0700) Subject: Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6 X-Git-Tag: v2.6.28-rc1~24 X-Git-Url: http://www.pilppa.org/gitweb/gitweb.cgi?a=commitdiff_plain;h=5ed487bc2c44ca4e9668ef9cb54c830e2a9fac47;hp=-c;p=linux-2.6-omap-h63xx.git Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6 * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6: (46 commits) [PATCH] fs: add a sanity check in d_free [PATCH] i_version: remount support [patch] vfs: make security_inode_setattr() calling consistent [patch 1/3] FS_MBCACHE: don't needlessly make it built-in [PATCH] move executable checking into ->permission() [PATCH] fs/dcache.c: update comment of d_validate() [RFC PATCH] touch_mnt_namespace when the mount flags change [PATCH] reiserfs: add missing llseek method [PATCH] fix ->llseek for more directories [PATCH vfs-2.6 6/6] vfs: add LOOKUP_RENAME_TARGET intent [PATCH vfs-2.6 5/6] vfs: remove LOOKUP_PARENT from non LOOKUP_PARENT lookup [PATCH vfs-2.6 4/6] vfs: remove unnecessary fsnotify_d_instantiate() [PATCH vfs-2.6 3/6] vfs: add __d_instantiate() helper [PATCH vfs-2.6 2/6] vfs: add d_ancestor() [PATCH vfs-2.6 1/6] vfs: replace parent == dentry->d_parent by IS_ROOT() [PATCH] get rid of on-stack dentry in udf [PATCH 2/2] anondev: switch to IDA [PATCH 1/2] anondev: init IDR statically [JFFS2] Use d_splice_alias() not d_add() in jffs2_lookup() [PATCH] Optimise NFS readdir hack slightly. ... --- 5ed487bc2c44ca4e9668ef9cb54c830e2a9fac47 diff --combined fs/ext3/super.c index cac29ee3b14,5b7fee10566..34b6fca765d --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@@ -393,8 -393,7 +393,8 @@@ static void ext3_put_super (struct supe int i; ext3_xattr_put_super(sb); - journal_destroy(sbi->s_journal); + if (journal_destroy(sbi->s_journal) < 0) + ext3_abort(sb, __func__, "Couldn't clean up the journal"); if (!(sb->s_flags & MS_RDONLY)) { EXT3_CLEAR_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER); es->s_state = cpu_to_le16(sbi->s_mount_state); @@@ -2297,9 -2296,7 +2297,9 @@@ static void ext3_mark_recovery_complete journal_t *journal = EXT3_SB(sb)->s_journal; journal_lock_updates(journal); - journal_flush(journal); + if (journal_flush(journal) < 0) + goto out; + lock_super(sb); if (EXT3_HAS_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER) && sb->s_flags & MS_RDONLY) { @@@ -2308,8 -2305,6 +2308,8 @@@ ext3_commit_super(sb, es, 1); } unlock_super(sb); + +out: journal_unlock_updates(journal); } @@@ -2409,13 -2404,7 +2409,13 @@@ static void ext3_write_super_lockfs(str /* Now we set up the journal barrier. */ journal_lock_updates(journal); - journal_flush(journal); + + /* + * We don't want to clear needs_recovery flag when we failed + * to flush the journal. + */ + if (journal_flush(journal) < 0) + return; /* Journal blocked and flushed, clear needs_recovery flag. */ EXT3_CLEAR_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER); @@@ -2794,30 -2783,30 +2794,30 @@@ static int ext3_quota_on_mount(struct s * Standard function to be called on quota_on */ static int ext3_quota_on(struct super_block *sb, int type, int format_id, - char *path, int remount) + char *name, int remount) { int err; - struct nameidata nd; + struct path path; if (!test_opt(sb, QUOTA)) return -EINVAL; - /* When remounting, no checks are needed and in fact, path is NULL */ + /* When remounting, no checks are needed and in fact, name is NULL */ if (remount) - return vfs_quota_on(sb, type, format_id, path, remount); + return vfs_quota_on(sb, type, format_id, name, remount); - err = path_lookup(path, LOOKUP_FOLLOW, &nd); + err = kern_path(name, LOOKUP_FOLLOW, &path); if (err) return err; /* Quotafile not on the same filesystem? */ - if (nd.path.mnt->mnt_sb != sb) { - path_put(&nd.path); + if (path.mnt->mnt_sb != sb) { + path_put(&path); return -EXDEV; } /* Journaling quota? */ if (EXT3_SB(sb)->s_qf_names[type]) { /* Quotafile not of fs root? */ - if (nd.path.dentry->d_parent->d_inode != sb->s_root->d_inode) + if (path.dentry->d_parent != sb->s_root) printk(KERN_WARNING "EXT3-fs: Quota file not on filesystem root. " "Journaled quota will not work.\n"); @@@ -2827,22 -2816,18 +2827,22 @@@ * When we journal data on quota file, we have to flush journal to see * all updates to the file when we bypass pagecache... */ - if (ext3_should_journal_data(nd.path.dentry->d_inode)) { + if (ext3_should_journal_data(path.dentry->d_inode)) { /* * We don't need to lock updates but journal_flush() could * otherwise be livelocked... */ journal_lock_updates(EXT3_SB(sb)->s_journal); - journal_flush(EXT3_SB(sb)->s_journal); + err = journal_flush(EXT3_SB(sb)->s_journal); journal_unlock_updates(EXT3_SB(sb)->s_journal); + if (err) { + path_put(&nd.path); + return err; + } } - err = vfs_quota_on_path(sb, type, format_id, &nd.path); - path_put(&nd.path); + err = vfs_quota_on_path(sb, type, format_id, &path); + path_put(&path); return err; } diff --combined fs/nfsd/export.c index 5cd882b8871,676201dbdf8..5839b229cd0 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@@ -99,7 -99,7 +99,7 @@@ static int expkey_parse(struct cache_de int fsidtype; char *ep; struct svc_expkey key; - struct svc_expkey *ek; + struct svc_expkey *ek = NULL; if (mesg[mlen-1] != '\n') return -EINVAL; @@@ -107,8 -107,7 +107,8 @@@ buf = kmalloc(PAGE_SIZE, GFP_KERNEL); err = -ENOMEM; - if (!buf) goto out; + if (!buf) + goto out; err = -EINVAL; if ((len=qword_get(&mesg, buf, PAGE_SIZE)) <= 0) @@@ -152,34 -151,32 +152,32 @@@ /* now we want a pathname, or empty meaning NEGATIVE */ err = -EINVAL; - if ((len=qword_get(&mesg, buf, PAGE_SIZE)) < 0) + len = qword_get(&mesg, buf, PAGE_SIZE); + if (len < 0) goto out; dprintk("Path seems to be <%s>\n", buf); err = 0; if (len == 0) { set_bit(CACHE_NEGATIVE, &key.h.flags); ek = svc_expkey_update(&key, ek); - if (ek) - cache_put(&ek->h, &svc_expkey_cache); - else err = -ENOMEM; + if (!ek) + err = -ENOMEM; } else { - struct nameidata nd; - err = path_lookup(buf, 0, &nd); + err = kern_path(buf, 0, &key.ek_path); if (err) goto out; dprintk("Found the path %s\n", buf); - key.ek_path = nd.path; ek = svc_expkey_update(&key, ek); - if (ek) - cache_put(&ek->h, &svc_expkey_cache); - else + if (!ek) err = -ENOMEM; - path_put(&nd.path); + path_put(&key.ek_path); } cache_flush(); out: + if (ek) + cache_put(&ek->h, &svc_expkey_cache); if (dom) auth_domain_put(dom); kfree(buf); @@@ -501,35 -498,22 +499,22 @@@ static int svc_export_parse(struct cach int len; int err; struct auth_domain *dom = NULL; - struct nameidata nd; - struct svc_export exp, *expp; + struct svc_export exp = {}, *expp; int an_int; - nd.path.dentry = NULL; - exp.ex_pathname = NULL; - - /* fs locations */ - exp.ex_fslocs.locations = NULL; - exp.ex_fslocs.locations_count = 0; - exp.ex_fslocs.migrated = 0; - - exp.ex_uuid = NULL; - - /* secinfo */ - exp.ex_nflavors = 0; - if (mesg[mlen-1] != '\n') return -EINVAL; mesg[mlen-1] = 0; buf = kmalloc(PAGE_SIZE, GFP_KERNEL); - err = -ENOMEM; - if (!buf) goto out; + if (!buf) + return -ENOMEM; /* client */ - len = qword_get(&mesg, buf, PAGE_SIZE); err = -EINVAL; - if (len <= 0) goto out; + len = qword_get(&mesg, buf, PAGE_SIZE); + if (len <= 0) + goto out; err = -ENOENT; dom = auth_domain_find(buf); @@@ -538,25 -522,25 +523,25 @@@ /* path */ err = -EINVAL; - if ((len=qword_get(&mesg, buf, PAGE_SIZE)) <= 0) - goto out; - err = path_lookup(buf, 0, &nd); - if (err) goto out_no_path; + if ((len = qword_get(&mesg, buf, PAGE_SIZE)) <= 0) + goto out1; + + err = kern_path(buf, 0, &exp.ex_path); + if (err) + goto out1; - exp.h.flags = 0; exp.ex_client = dom; - exp.ex_path.mnt = nd.path.mnt; - exp.ex_path.dentry = nd.path.dentry; - exp.ex_pathname = kstrdup(buf, GFP_KERNEL); + err = -ENOMEM; + exp.ex_pathname = kstrdup(buf, GFP_KERNEL); if (!exp.ex_pathname) - goto out; + goto out2; /* expiry */ err = -EINVAL; exp.h.expiry_time = get_expiry(&mesg); if (exp.h.expiry_time == 0) - goto out; + goto out3; /* flags */ err = get_int(&mesg, &an_int); @@@ -564,22 -548,26 +549,26 @@@ err = 0; set_bit(CACHE_NEGATIVE, &exp.h.flags); } else { - if (err || an_int < 0) goto out; + if (err || an_int < 0) + goto out3; exp.ex_flags= an_int; /* anon uid */ err = get_int(&mesg, &an_int); - if (err) goto out; + if (err) + goto out3; exp.ex_anon_uid= an_int; /* anon gid */ err = get_int(&mesg, &an_int); - if (err) goto out; + if (err) + goto out3; exp.ex_anon_gid= an_int; /* fsid */ err = get_int(&mesg, &an_int); - if (err) goto out; + if (err) + goto out3; exp.ex_fsid = an_int; while ((len = qword_get(&mesg, buf, PAGE_SIZE)) > 0) { @@@ -605,12 -593,13 +594,13 @@@ */ break; if (err) - goto out; + goto out4; } - err = check_export(nd.path.dentry->d_inode, exp.ex_flags, + err = check_export(exp.ex_path.dentry->d_inode, exp.ex_flags, exp.ex_uuid); - if (err) goto out; + if (err) + goto out4; } expp = svc_export_lookup(&exp); @@@ -623,15 -612,16 +613,16 @@@ err = -ENOMEM; else exp_put(expp); - out: + out4: nfsd4_fslocs_free(&exp.ex_fslocs); kfree(exp.ex_uuid); + out3: kfree(exp.ex_pathname); - if (nd.path.dentry) - path_put(&nd.path); - out_no_path: - if (dom) - auth_domain_put(dom); + out2: + path_put(&exp.ex_path); + out1: + auth_domain_put(dom); + out: kfree(buf); return err; } @@@ -999,7 -989,7 +990,7 @@@ exp_export(struct nfsctl_export *nxp struct svc_export *exp = NULL; struct svc_export new; struct svc_expkey *fsid_key = NULL; - struct nameidata nd; + struct path path; int err; /* Consistency check */ @@@ -1022,12 -1012,12 +1013,12 @@@ /* Look up the dentry */ - err = path_lookup(nxp->ex_path, 0, &nd); + err = kern_path(nxp->ex_path, 0, &path); if (err) goto out_put_clp; err = -EINVAL; - exp = exp_get_by_name(clp, nd.path.mnt, nd.path.dentry, NULL); + exp = exp_get_by_name(clp, path.mnt, path.dentry, NULL); memset(&new, 0, sizeof(new)); @@@ -1035,8 -1025,8 +1026,8 @@@ if ((nxp->ex_flags & NFSEXP_FSID) && (!IS_ERR(fsid_key = exp_get_fsid_key(clp, nxp->ex_dev))) && fsid_key->ek_path.mnt && - (fsid_key->ek_path.mnt != nd.path.mnt || - fsid_key->ek_path.dentry != nd.path.dentry)) + (fsid_key->ek_path.mnt != path.mnt || + fsid_key->ek_path.dentry != path.dentry)) goto finish; if (!IS_ERR(exp)) { @@@ -1052,7 -1042,7 +1043,7 @@@ goto finish; } - err = check_export(nd.path.dentry->d_inode, nxp->ex_flags, NULL); + err = check_export(path.dentry->d_inode, nxp->ex_flags, NULL); if (err) goto finish; err = -ENOMEM; @@@ -1065,7 -1055,7 +1056,7 @@@ if (!new.ex_pathname) goto finish; new.ex_client = clp; - new.ex_path = nd.path; + new.ex_path = path; new.ex_flags = nxp->ex_flags; new.ex_anon_uid = nxp->ex_anon_uid; new.ex_anon_gid = nxp->ex_anon_gid; @@@ -1091,7 -1081,7 +1082,7 @@@ finish exp_put(exp); if (fsid_key && !IS_ERR(fsid_key)) cache_put(&fsid_key->h, &svc_expkey_cache); - path_put(&nd.path); + path_put(&path); out_put_clp: auth_domain_put(clp); out_unlock: @@@ -1122,7 -1112,7 +1113,7 @@@ exp_unexport(struct nfsctl_export *nxp { struct auth_domain *dom; svc_export *exp; - struct nameidata nd; + struct path path; int err; /* Consistency check */ @@@ -1139,13 -1129,13 +1130,13 @@@ goto out_unlock; } - err = path_lookup(nxp->ex_path, 0, &nd); + err = kern_path(nxp->ex_path, 0, &path); if (err) goto out_domain; err = -EINVAL; - exp = exp_get_by_name(dom, nd.path.mnt, nd.path.dentry, NULL); - path_put(&nd.path); + exp = exp_get_by_name(dom, path.mnt, path.dentry, NULL); + path_put(&path); if (IS_ERR(exp)) goto out_domain; @@@ -1167,26 -1157,26 +1158,26 @@@ out_unlock * since its harder to fool a kernel module than a user space program. */ int - exp_rootfh(svc_client *clp, char *path, struct knfsd_fh *f, int maxsize) + exp_rootfh(svc_client *clp, char *name, struct knfsd_fh *f, int maxsize) { struct svc_export *exp; - struct nameidata nd; + struct path path; struct inode *inode; struct svc_fh fh; int err; err = -EPERM; /* NB: we probably ought to check that it's NUL-terminated */ - if (path_lookup(path, 0, &nd)) { - printk("nfsd: exp_rootfh path not found %s", path); + if (kern_path(name, 0, &path)) { + printk("nfsd: exp_rootfh path not found %s", name); return err; } - inode = nd.path.dentry->d_inode; + inode = path.dentry->d_inode; dprintk("nfsd: exp_rootfh(%s [%p] %s:%s/%ld)\n", - path, nd.path.dentry, clp->name, + name, path.dentry, clp->name, inode->i_sb->s_id, inode->i_ino); - exp = exp_parent(clp, nd.path.mnt, nd.path.dentry, NULL); + exp = exp_parent(clp, path.mnt, path.dentry, NULL); if (IS_ERR(exp)) { err = PTR_ERR(exp); goto out; @@@ -1196,7 -1186,7 +1187,7 @@@ * fh must be initialized before calling fh_compose */ fh_init(&fh, maxsize); - if (fh_compose(&fh, exp, nd.path.dentry, NULL)) + if (fh_compose(&fh, exp, path.dentry, NULL)) err = -EINVAL; else err = 0; @@@ -1204,7 -1194,7 +1195,7 @@@ fh_put(&fh); exp_put(exp); out: - path_put(&nd.path); + path_put(&path); return err; } diff --combined fs/nfsd/vfs.c index 9609eb51d72,49d4b8725ca..0bc56f6d927 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@@ -410,7 -410,6 +410,7 @@@ out_nfserr static ssize_t nfsd_getxattr(struct dentry *dentry, char *key, void **buf) { ssize_t buflen; + ssize_t ret; buflen = vfs_getxattr(dentry, key, NULL, 0); if (buflen <= 0) @@@ -420,10 -419,7 +420,10 @@@ if (!*buf) return -ENOMEM; - return vfs_getxattr(dentry, key, *buf, buflen); + ret = vfs_getxattr(dentry, key, *buf, buflen); + if (ret < 0) + kfree(*buf); + return ret; } #endif @@@ -1817,6 -1813,115 +1817,115 @@@ out return err; } + /* + * We do this buffering because we must not call back into the file + * system's ->lookup() method from the filldir callback. That may well + * deadlock a number of file systems. + * + * This is based heavily on the implementation of same in XFS. + */ + struct buffered_dirent { + u64 ino; + loff_t offset; + int namlen; + unsigned int d_type; + char name[]; + }; + + struct readdir_data { + char *dirent; + size_t used; + int full; + }; + + static int nfsd_buffered_filldir(void *__buf, const char *name, int namlen, + loff_t offset, u64 ino, unsigned int d_type) + { + struct readdir_data *buf = __buf; + struct buffered_dirent *de = (void *)(buf->dirent + buf->used); + unsigned int reclen; + + reclen = ALIGN(sizeof(struct buffered_dirent) + namlen, sizeof(u64)); + if (buf->used + reclen > PAGE_SIZE) { + buf->full = 1; + return -EINVAL; + } + + de->namlen = namlen; + de->offset = offset; + de->ino = ino; + de->d_type = d_type; + memcpy(de->name, name, namlen); + buf->used += reclen; + + return 0; + } + + static int nfsd_buffered_readdir(struct file *file, filldir_t func, + struct readdir_cd *cdp, loff_t *offsetp) + { + struct readdir_data buf; + struct buffered_dirent *de; + int host_err; + int size; + loff_t offset; + + buf.dirent = (void *)__get_free_page(GFP_KERNEL); + if (!buf.dirent) + return -ENOMEM; + + offset = *offsetp; + cdp->err = nfserr_eof; /* will be cleared on successful read */ + + while (1) { + unsigned int reclen; + + buf.used = 0; + buf.full = 0; + + host_err = vfs_readdir(file, nfsd_buffered_filldir, &buf); + if (buf.full) + host_err = 0; + + if (host_err < 0) + break; + + size = buf.used; + + if (!size) + break; + + de = (struct buffered_dirent *)buf.dirent; + while (size > 0) { + offset = de->offset; + + if (func(cdp, de->name, de->namlen, de->offset, + de->ino, de->d_type)) + goto done; + + if (cdp->err != nfs_ok) + goto done; + + reclen = ALIGN(sizeof(*de) + de->namlen, + sizeof(u64)); + size -= reclen; + de = (struct buffered_dirent *)((char *)de + reclen); + } + offset = vfs_llseek(file, 0, SEEK_CUR); + if (!buf.full) + break; + } + + done: + free_page((unsigned long)(buf.dirent)); + + if (host_err) + return nfserrno(host_err); + + *offsetp = offset; + return cdp->err; + } + /* * Read entries from a directory. * The NFSv3/4 verifier we ignore for now. @@@ -1826,7 -1931,6 +1935,6 @@@ nfsd_readdir(struct svc_rqst *rqstp, st struct readdir_cd *cdp, filldir_t func) { __be32 err; - int host_err; struct file *file; loff_t offset = *offsetp; @@@ -1840,21 -1944,7 +1948,7 @@@ goto out_close; } - /* - * Read the directory entries. This silly loop is necessary because - * readdir() is not guaranteed to fill up the entire buffer, but - * may choose to do less. - */ - - do { - cdp->err = nfserr_eof; /* will be cleared on successful read */ - host_err = vfs_readdir(file, func, cdp); - } while (host_err >=0 && cdp->err == nfs_ok); - if (host_err) - err = nfserrno(host_err); - else - err = cdp->err; - *offsetp = vfs_llseek(file, 0, 1); + err = nfsd_buffered_readdir(file, func, cdp, offsetp); if (err == nfserr_eof || err == nfserr_toosmall) err = nfs_ok; /* can still be found in ->err */ diff --combined init/main.c index b038fa14204,c6a1024a27a..130d1a0eef1 --- a/init/main.c +++ b/init/main.c @@@ -52,7 -52,6 +52,7 @@@ #include #include #include +#include #include #include #include @@@ -648,7 -647,6 +648,7 @@@ asmlinkage void __init start_kernel(voi vmalloc_init(); vfs_caches_init_early(); cpuset_init_early(); + page_cgroup_init(); mem_init(); enable_debug_pagealloc(); cpu_hotplug_init(); @@@ -672,7 -670,6 +672,6 @@@ fork_init(num_physpages); proc_caches_init(); buffer_init(); - unnamed_dev_init(); key_init(); security_init(); vfs_caches_init(num_physpages); @@@ -699,7 -696,13 +698,7 @@@ } static int initcall_debug; - -static int __init initcall_debug_setup(char *str) -{ - initcall_debug = 1; - return 1; -} -__setup("initcall_debug", initcall_debug_setup); +core_param(initcall_debug, initcall_debug, bool, 0644); int do_one_initcall(initcall_t fn) { @@@ -769,6 -772,8 +768,6 @@@ static void __init do_initcalls(void static void __init do_basic_setup(void) { rcu_init_sched(); /* needed by module_init stage. */ - /* drivers will send hotplug events */ - init_workqueues(); usermodehelper_init(); driver_init(); init_irq_proc(); @@@ -852,8 -857,6 +851,8 @@@ static int __init kernel_init(void * un cad_pid = task_pid(current); + init_workqueues(); + smp_prepare_cpus(setup_max_cpus); do_pre_smp_initcalls();