diff options
Diffstat (limited to 'fs/nfsd/nfs4recover.c')
| -rw-r--r-- | fs/nfsd/nfs4recover.c | 1241 |
1 files changed, 1128 insertions, 113 deletions
diff --git a/fs/nfsd/nfs4recover.c b/fs/nfsd/nfs4recover.c index 7e26caab2a2..9c271f42604 100644 --- a/fs/nfsd/nfs4recover.c +++ b/fs/nfsd/nfs4recover.c @@ -1,5 +1,6 @@ /* * Copyright (c) 2004 The Regents of the University of Michigan. +* Copyright (c) 2012 Jeff Layton <jlayton@redhat.com> * All rights reserved. * * Andy Adamson <andros@citi.umich.edu> @@ -36,15 +37,32 @@ #include <linux/namei.h> #include <linux/crypto.h> #include <linux/sched.h> +#include <linux/fs.h> +#include <linux/module.h> +#include <net/net_namespace.h> +#include <linux/sunrpc/rpc_pipe_fs.h> +#include <linux/sunrpc/clnt.h> +#include <linux/nfsd/cld.h> #include "nfsd.h" #include "state.h" #include "vfs.h" +#include "netns.h" #define NFSDDBG_FACILITY NFSDDBG_PROC +/* Declarations */ +struct nfsd4_client_tracking_ops { + int (*init)(struct net *); + void (*exit)(struct net *); + void (*create)(struct nfs4_client *); + void (*remove)(struct nfs4_client *); + int (*check)(struct nfs4_client *); + void (*grace_done)(struct nfsd_net *, time_t); +}; + /* Globals */ -static struct file *rec_file; +static char user_recovery_dirname[PATH_MAX] = "/var/lib/nfs/v4recovery"; static int nfs4_save_creds(const struct cred **original_creds) @@ -55,8 +73,8 @@ nfs4_save_creds(const struct cred **original_creds) if (!new) return -ENOMEM; - new->fsuid = 0; - new->fsgid = 0; + new->fsuid = GLOBAL_ROOT_UID; + new->fsgid = GLOBAL_ROOT_GID; *original_creds = override_creds(new); put_cred(new); return 0; @@ -82,33 +100,39 @@ md5_to_hex(char *out, char *md5) *out = '\0'; } -__be32 -nfs4_make_rec_clidname(char *dname, struct xdr_netobj *clname) +static int +nfs4_make_rec_clidname(char *dname, const struct xdr_netobj *clname) { struct xdr_netobj cksum; struct hash_desc desc; struct scatterlist sg; - __be32 status = nfserr_resource; + int status; dprintk("NFSD: nfs4_make_rec_clidname for %.*s\n", clname->len, clname->data); desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP; desc.tfm = crypto_alloc_hash("md5", 0, CRYPTO_ALG_ASYNC); - if (IS_ERR(desc.tfm)) + if (IS_ERR(desc.tfm)) { + status = PTR_ERR(desc.tfm); goto out_no_tfm; + } + cksum.len = crypto_hash_digestsize(desc.tfm); cksum.data = kmalloc(cksum.len, GFP_KERNEL); - if (cksum.data == NULL) + if (cksum.data == NULL) { + status = -ENOMEM; goto out; + } sg_init_one(&sg, clname->data, clname->len); - if (crypto_hash_digest(&desc, &sg, sg.length, cksum.data)) + status = crypto_hash_digest(&desc, &sg, sg.length, cksum.data); + if (status) goto out; md5_to_hex(dname, cksum.data); - status = nfs_ok; + status = 0; out: kfree(cksum.data); crypto_free_hash(desc.tfm); @@ -116,24 +140,57 @@ out_no_tfm: return status; } -int +/* + * If we had an error generating the recdir name for the legacy tracker + * then warn the admin. If the error doesn't appear to be transient, + * then disable recovery tracking. + */ +static void +legacy_recdir_name_error(struct nfs4_client *clp, int error) +{ + printk(KERN_ERR "NFSD: unable to generate recoverydir " + "name (%d).\n", error); + + /* + * if the algorithm just doesn't exist, then disable the recovery + * tracker altogether. The crypto libs will generally return this if + * FIPS is enabled as well. + */ + if (error == -ENOENT) { + printk(KERN_ERR "NFSD: disabling legacy clientid tracking. " + "Reboot recovery will not function correctly!\n"); + nfsd4_client_tracking_exit(clp->net); + } +} + +static void nfsd4_create_clid_dir(struct nfs4_client *clp) { const struct cred *original_cred; - char *dname = clp->cl_recdir; + char dname[HEXDIR_LEN]; struct dentry *dir, *dentry; + struct nfs4_client_reclaim *crp; int status; + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); - dprintk("NFSD: nfsd4_create_clid_dir for \"%s\"\n", dname); + if (test_and_set_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags)) + return; + if (!nn->rec_file) + return; - if (!rec_file || clp->cl_firststate) - return 0; + status = nfs4_make_rec_clidname(dname, &clp->cl_name); + if (status) + return legacy_recdir_name_error(clp, status); status = nfs4_save_creds(&original_cred); if (status < 0) - return status; + return; - dir = rec_file->f_path.dentry; + status = mnt_want_write_file(nn->rec_file); + if (status) + return; + + dir = nn->rec_file->f_path.dentry; /* lock the parent */ mutex_lock(&dir->d_inode->i_mutex); @@ -142,41 +199,55 @@ nfsd4_create_clid_dir(struct nfs4_client *clp) status = PTR_ERR(dentry); goto out_unlock; } - status = -EEXIST; - if (dentry->d_inode) { - dprintk("NFSD: nfsd4_create_clid_dir: DIRECTORY EXISTS\n"); - goto out_put; - } - status = mnt_want_write(rec_file->f_path.mnt); - if (status) + if (dentry->d_inode) + /* + * In the 4.1 case, where we're called from + * reclaim_complete(), records from the previous reboot + * may still be left, so this is OK. + * + * In the 4.0 case, we should never get here; but we may + * as well be forgiving and just succeed silently. + */ goto out_put; status = vfs_mkdir(dir->d_inode, dentry, S_IRWXU); - mnt_drop_write(rec_file->f_path.mnt); out_put: dput(dentry); out_unlock: mutex_unlock(&dir->d_inode->i_mutex); if (status == 0) { - clp->cl_firststate = 1; - vfs_fsync(rec_file, 0); + if (nn->in_grace) { + crp = nfs4_client_to_reclaim(dname, nn); + if (crp) + crp->cr_clp = clp; + } + vfs_fsync(nn->rec_file, 0); + } else { + printk(KERN_ERR "NFSD: failed to write recovery record" + " (err %d); please check that %s exists" + " and is writeable", status, + user_recovery_dirname); } + mnt_drop_write_file(nn->rec_file); nfs4_reset_creds(original_cred); - dprintk("NFSD: nfsd4_create_clid_dir returns %d\n", status); - return status; } -typedef int (recdir_func)(struct dentry *, struct dentry *); +typedef int (recdir_func)(struct dentry *, struct dentry *, struct nfsd_net *); struct name_list { char name[HEXDIR_LEN]; struct list_head list; }; +struct nfs4_dir_ctx { + struct dir_context ctx; + struct list_head names; +}; + static int nfsd4_build_namelist(void *arg, const char *name, int namlen, loff_t offset, u64 ino, unsigned int d_type) { - struct list_head *names = arg; + struct nfs4_dir_ctx *ctx = arg; struct name_list *entry; if (namlen != HEXDIR_LEN - 1) @@ -186,70 +257,63 @@ nfsd4_build_namelist(void *arg, const char *name, int namlen, return -ENOMEM; memcpy(entry->name, name, HEXDIR_LEN - 1); entry->name[HEXDIR_LEN - 1] = '\0'; - list_add(&entry->list, names); + list_add(&entry->list, &ctx->names); return 0; } static int -nfsd4_list_rec_dir(struct dentry *dir, recdir_func *f) +nfsd4_list_rec_dir(recdir_func *f, struct nfsd_net *nn) { const struct cred *original_cred; - struct file *filp; - LIST_HEAD(names); - struct name_list *entry; - struct dentry *dentry; + struct dentry *dir = nn->rec_file->f_path.dentry; + struct nfs4_dir_ctx ctx = { + .ctx.actor = nfsd4_build_namelist, + .names = LIST_HEAD_INIT(ctx.names) + }; int status; - if (!rec_file) - return 0; - status = nfs4_save_creds(&original_cred); if (status < 0) return status; - filp = dentry_open(dget(dir), mntget(rec_file->f_path.mnt), O_RDONLY, - current_cred()); - status = PTR_ERR(filp); - if (IS_ERR(filp)) - goto out; - status = vfs_readdir(filp, nfsd4_build_namelist, &names); - fput(filp); - mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT); - while (!list_empty(&names)) { - entry = list_entry(names.next, struct name_list, list); + status = vfs_llseek(nn->rec_file, 0, SEEK_SET); + if (status < 0) { + nfs4_reset_creds(original_cred); + return status; + } - dentry = lookup_one_len(entry->name, dir, HEXDIR_LEN-1); - if (IS_ERR(dentry)) { - status = PTR_ERR(dentry); - break; + status = iterate_dir(nn->rec_file, &ctx.ctx); + mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT); + while (!list_empty(&ctx.names)) { + struct name_list *entry; + entry = list_entry(ctx.names.next, struct name_list, list); + if (!status) { + struct dentry *dentry; + dentry = lookup_one_len(entry->name, dir, HEXDIR_LEN-1); + if (IS_ERR(dentry)) { + status = PTR_ERR(dentry); + break; + } + status = f(dir, dentry, nn); + dput(dentry); } - status = f(dir, dentry); - dput(dentry); - if (status) - break; list_del(&entry->list); kfree(entry); } mutex_unlock(&dir->d_inode->i_mutex); -out: - while (!list_empty(&names)) { - entry = list_entry(names.next, struct name_list, list); - list_del(&entry->list); - kfree(entry); - } nfs4_reset_creds(original_cred); return status; } static int -nfsd4_unlink_clid_dir(char *name, int namlen) +nfsd4_unlink_clid_dir(char *name, int namlen, struct nfsd_net *nn) { struct dentry *dir, *dentry; int status; dprintk("NFSD: nfsd4_unlink_clid_dir. name %.*s\n", namlen, name); - dir = rec_file->f_path.dentry; + dir = nn->rec_file->f_path.dentry; mutex_lock_nested(&dir->d_inode->i_mutex, I_MUTEX_PARENT); dentry = lookup_one_len(name, dir, namlen); if (IS_ERR(dentry)) { @@ -267,96 +331,113 @@ out_unlock: return status; } -void +static void nfsd4_remove_clid_dir(struct nfs4_client *clp) { const struct cred *original_cred; + struct nfs4_client_reclaim *crp; + char dname[HEXDIR_LEN]; int status; + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); - if (!rec_file || !clp->cl_firststate) + if (!nn->rec_file || !test_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags)) return; - status = mnt_want_write(rec_file->f_path.mnt); + status = nfs4_make_rec_clidname(dname, &clp->cl_name); + if (status) + return legacy_recdir_name_error(clp, status); + + status = mnt_want_write_file(nn->rec_file); if (status) goto out; - clp->cl_firststate = 0; + clear_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags); status = nfs4_save_creds(&original_cred); if (status < 0) - goto out; + goto out_drop_write; - status = nfsd4_unlink_clid_dir(clp->cl_recdir, HEXDIR_LEN-1); + status = nfsd4_unlink_clid_dir(dname, HEXDIR_LEN-1, nn); nfs4_reset_creds(original_cred); - if (status == 0) - vfs_fsync(rec_file, 0); - mnt_drop_write(rec_file->f_path.mnt); + if (status == 0) { + vfs_fsync(nn->rec_file, 0); + if (nn->in_grace) { + /* remove reclaim record */ + crp = nfsd4_find_reclaim_client(dname, nn); + if (crp) + nfs4_remove_reclaim_record(crp, nn); + } + } +out_drop_write: + mnt_drop_write_file(nn->rec_file); out: if (status) printk("NFSD: Failed to remove expired client state directory" - " %.*s\n", HEXDIR_LEN, clp->cl_recdir); - return; + " %.*s\n", HEXDIR_LEN, dname); } static int -purge_old(struct dentry *parent, struct dentry *child) +purge_old(struct dentry *parent, struct dentry *child, struct nfsd_net *nn) { int status; - /* note: we currently use this path only for minorversion 0 */ - if (nfs4_has_reclaimed_state(child->d_name.name, false)) + if (nfs4_has_reclaimed_state(child->d_name.name, nn)) return 0; status = vfs_rmdir(parent->d_inode, child); if (status) - printk("failed to remove client recovery directory %s\n", - child->d_name.name); + printk("failed to remove client recovery directory %pd\n", + child); /* Keep trying, success or failure: */ return 0; } -void -nfsd4_recdir_purge_old(void) { +static void +nfsd4_recdir_purge_old(struct nfsd_net *nn, time_t boot_time) +{ int status; - if (!rec_file) + nn->in_grace = false; + if (!nn->rec_file) return; - status = mnt_want_write(rec_file->f_path.mnt); + status = mnt_want_write_file(nn->rec_file); if (status) goto out; - status = nfsd4_list_rec_dir(rec_file->f_path.dentry, purge_old); + status = nfsd4_list_rec_dir(purge_old, nn); if (status == 0) - vfs_fsync(rec_file, 0); - mnt_drop_write(rec_file->f_path.mnt); + vfs_fsync(nn->rec_file, 0); + mnt_drop_write_file(nn->rec_file); out: + nfs4_release_reclaim(nn); if (status) printk("nfsd4: failed to purge old clients from recovery" - " directory %s\n", rec_file->f_path.dentry->d_name.name); + " directory %pD\n", nn->rec_file); } static int -load_recdir(struct dentry *parent, struct dentry *child) +load_recdir(struct dentry *parent, struct dentry *child, struct nfsd_net *nn) { if (child->d_name.len != HEXDIR_LEN - 1) { - printk("nfsd4: illegal name %s in recovery directory\n", - child->d_name.name); + printk("nfsd4: illegal name %pd in recovery directory\n", + child); /* Keep trying; maybe the others are OK: */ return 0; } - nfs4_client_to_reclaim(child->d_name.name); + nfs4_client_to_reclaim(child->d_name.name, nn); return 0; } -int -nfsd4_recdir_load(void) { +static int +nfsd4_recdir_load(struct net *net) { int status; + struct nfsd_net *nn = net_generic(net, nfsd_net_id); - if (!rec_file) + if (!nn->rec_file) return 0; - status = nfsd4_list_rec_dir(rec_file->f_path.dentry, load_recdir); + status = nfsd4_list_rec_dir(load_recdir, nn); if (status) printk("nfsd4: failed loading clients from recovery" - " directory %s\n", rec_file->f_path.dentry->d_name.name); + " directory %pD\n", nn->rec_file); return status; } @@ -364,40 +445,974 @@ nfsd4_recdir_load(void) { * Hold reference to the recovery directory. */ -void -nfsd4_init_recdir(char *rec_dirname) +static int +nfsd4_init_recdir(struct net *net) { + struct nfsd_net *nn = net_generic(net, nfsd_net_id); const struct cred *original_cred; int status; printk("NFSD: Using %s as the NFSv4 state recovery directory\n", - rec_dirname); + user_recovery_dirname); - BUG_ON(rec_file); + BUG_ON(nn->rec_file); status = nfs4_save_creds(&original_cred); if (status < 0) { printk("NFSD: Unable to change credentials to find recovery" " directory: error %d\n", status); - return; + return status; } - rec_file = filp_open(rec_dirname, O_RDONLY | O_DIRECTORY, 0); - if (IS_ERR(rec_file)) { + nn->rec_file = filp_open(user_recovery_dirname, O_RDONLY | O_DIRECTORY, 0); + if (IS_ERR(nn->rec_file)) { printk("NFSD: unable to find recovery directory %s\n", - rec_dirname); - rec_file = NULL; + user_recovery_dirname); + status = PTR_ERR(nn->rec_file); + nn->rec_file = NULL; } nfs4_reset_creds(original_cred); + if (!status) + nn->in_grace = true; + return status; } -void -nfsd4_shutdown_recdir(void) + +static int +nfs4_legacy_state_init(struct net *net) +{ + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + int i; + + nn->reclaim_str_hashtbl = kmalloc(sizeof(struct list_head) * + CLIENT_HASH_SIZE, GFP_KERNEL); + if (!nn->reclaim_str_hashtbl) + return -ENOMEM; + + for (i = 0; i < CLIENT_HASH_SIZE; i++) + INIT_LIST_HEAD(&nn->reclaim_str_hashtbl[i]); + nn->reclaim_str_hashtbl_size = 0; + + return 0; +} + +static void +nfs4_legacy_state_shutdown(struct net *net) +{ + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + + kfree(nn->reclaim_str_hashtbl); +} + +static int +nfsd4_load_reboot_recovery_data(struct net *net) +{ + int status; + + status = nfsd4_init_recdir(net); + if (!status) + status = nfsd4_recdir_load(net); + if (status) + printk(KERN_ERR "NFSD: Failure reading reboot recovery data\n"); + return status; +} + +static int +nfsd4_legacy_tracking_init(struct net *net) +{ + int status; + + /* XXX: The legacy code won't work in a container */ + if (net != &init_net) { + WARN(1, KERN_ERR "NFSD: attempt to initialize legacy client " + "tracking in a container!\n"); + return -EINVAL; + } + + status = nfs4_legacy_state_init(net); + if (status) + return status; + + status = nfsd4_load_reboot_recovery_data(net); + if (status) + goto err; + return 0; + +err: + nfs4_legacy_state_shutdown(net); + return status; +} + +static void +nfsd4_shutdown_recdir(struct nfsd_net *nn) +{ + if (!nn->rec_file) + return; + fput(nn->rec_file); + nn->rec_file = NULL; +} + +static void +nfsd4_legacy_tracking_exit(struct net *net) +{ + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + + nfs4_release_reclaim(nn); + nfsd4_shutdown_recdir(nn); + nfs4_legacy_state_shutdown(net); +} + +/* + * Change the NFSv4 recovery directory to recdir. + */ +int +nfs4_reset_recoverydir(char *recdir) +{ + int status; + struct path path; + + status = kern_path(recdir, LOOKUP_FOLLOW, &path); + if (status) + return status; + status = -ENOTDIR; + if (S_ISDIR(path.dentry->d_inode->i_mode)) { + strcpy(user_recovery_dirname, recdir); + status = 0; + } + path_put(&path); + return status; +} + +char * +nfs4_recoverydir(void) +{ + return user_recovery_dirname; +} + +static int +nfsd4_check_legacy_client(struct nfs4_client *clp) +{ + int status; + char dname[HEXDIR_LEN]; + struct nfs4_client_reclaim *crp; + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); + + /* did we already find that this client is stable? */ + if (test_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags)) + return 0; + + status = nfs4_make_rec_clidname(dname, &clp->cl_name); + if (status) { + legacy_recdir_name_error(clp, status); + return status; + } + + /* look for it in the reclaim hashtable otherwise */ + crp = nfsd4_find_reclaim_client(dname, nn); + if (crp) { + set_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags); + crp->cr_clp = clp; + return 0; + } + + return -ENOENT; +} + +static struct nfsd4_client_tracking_ops nfsd4_legacy_tracking_ops = { + .init = nfsd4_legacy_tracking_init, + .exit = nfsd4_legacy_tracking_exit, + .create = nfsd4_create_clid_dir, + .remove = nfsd4_remove_clid_dir, + .check = nfsd4_check_legacy_client, + .grace_done = nfsd4_recdir_purge_old, +}; + +/* Globals */ +#define NFSD_PIPE_DIR "nfsd" +#define NFSD_CLD_PIPE "cld" + +/* per-net-ns structure for holding cld upcall info */ +struct cld_net { + struct rpc_pipe *cn_pipe; + spinlock_t cn_lock; + struct list_head cn_list; + unsigned int cn_xid; +}; + +struct cld_upcall { + struct list_head cu_list; + struct cld_net *cu_net; + struct task_struct *cu_task; + struct cld_msg cu_msg; +}; + +static int +__cld_pipe_upcall(struct rpc_pipe *pipe, struct cld_msg *cmsg) +{ + int ret; + struct rpc_pipe_msg msg; + + memset(&msg, 0, sizeof(msg)); + msg.data = cmsg; + msg.len = sizeof(*cmsg); + + /* + * Set task state before we queue the upcall. That prevents + * wake_up_process in the downcall from racing with schedule. + */ + set_current_state(TASK_UNINTERRUPTIBLE); + ret = rpc_queue_upcall(pipe, &msg); + if (ret < 0) { + set_current_state(TASK_RUNNING); + goto out; + } + + schedule(); + set_current_state(TASK_RUNNING); + + if (msg.errno < 0) + ret = msg.errno; +out: + return ret; +} + +static int +cld_pipe_upcall(struct rpc_pipe *pipe, struct cld_msg *cmsg) +{ + int ret; + + /* + * -EAGAIN occurs when pipe is closed and reopened while there are + * upcalls queued. + */ + do { + ret = __cld_pipe_upcall(pipe, cmsg); + } while (ret == -EAGAIN); + + return ret; +} + +static ssize_t +cld_pipe_downcall(struct file *filp, const char __user *src, size_t mlen) +{ + struct cld_upcall *tmp, *cup; + struct cld_msg __user *cmsg = (struct cld_msg __user *)src; + uint32_t xid; + struct nfsd_net *nn = net_generic(filp->f_dentry->d_sb->s_fs_info, + nfsd_net_id); + struct cld_net *cn = nn->cld_net; + + if (mlen != sizeof(*cmsg)) { + dprintk("%s: got %zu bytes, expected %zu\n", __func__, mlen, + sizeof(*cmsg)); + return -EINVAL; + } + + /* copy just the xid so we can try to find that */ + if (copy_from_user(&xid, &cmsg->cm_xid, sizeof(xid)) != 0) { + dprintk("%s: error when copying xid from userspace", __func__); + return -EFAULT; + } + + /* walk the list and find corresponding xid */ + cup = NULL; + spin_lock(&cn->cn_lock); + list_for_each_entry(tmp, &cn->cn_list, cu_list) { + if (get_unaligned(&tmp->cu_msg.cm_xid) == xid) { + cup = tmp; + list_del_init(&cup->cu_list); + break; + } + } + spin_unlock(&cn->cn_lock); + + /* couldn't find upcall? */ + if (!cup) { + dprintk("%s: couldn't find upcall -- xid=%u\n", __func__, xid); + return -EINVAL; + } + + if (copy_from_user(&cup->cu_msg, src, mlen) != 0) + return -EFAULT; + + wake_up_process(cup->cu_task); + return mlen; +} + +static void +cld_pipe_destroy_msg(struct rpc_pipe_msg *msg) { - if (!rec_file) + struct cld_msg *cmsg = msg->data; + struct cld_upcall *cup = container_of(cmsg, struct cld_upcall, + cu_msg); + + /* errno >= 0 means we got a downcall */ + if (msg->errno >= 0) return; - fput(rec_file); - rec_file = NULL; + + wake_up_process(cup->cu_task); +} + +static const struct rpc_pipe_ops cld_upcall_ops = { + .upcall = rpc_pipe_generic_upcall, + .downcall = cld_pipe_downcall, + .destroy_msg = cld_pipe_destroy_msg, +}; + +static struct dentry * +nfsd4_cld_register_sb(struct super_block *sb, struct rpc_pipe *pipe) +{ + struct dentry *dir, *dentry; + + dir = rpc_d_lookup_sb(sb, NFSD_PIPE_DIR); + if (dir == NULL) + return ERR_PTR(-ENOENT); + dentry = rpc_mkpipe_dentry(dir, NFSD_CLD_PIPE, NULL, pipe); + dput(dir); + return dentry; +} + +static void +nfsd4_cld_unregister_sb(struct rpc_pipe *pipe) +{ + if (pipe->dentry) + rpc_unlink(pipe->dentry); +} + +static struct dentry * +nfsd4_cld_register_net(struct net *net, struct rpc_pipe *pipe) +{ + struct super_block *sb; + struct dentry *dentry; + + sb = rpc_get_sb_net(net); + if (!sb) + return NULL; + dentry = nfsd4_cld_register_sb(sb, pipe); + rpc_put_sb_net(net); + return dentry; +} + +static void +nfsd4_cld_unregister_net(struct net *net, struct rpc_pipe *pipe) +{ + struct super_block *sb; + + sb = rpc_get_sb_net(net); + if (sb) { + nfsd4_cld_unregister_sb(pipe); + rpc_put_sb_net(net); + } +} + +/* Initialize rpc_pipefs pipe for communication with client tracking daemon */ +static int +nfsd4_init_cld_pipe(struct net *net) +{ + int ret; + struct dentry *dentry; + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + struct cld_net *cn; + + if (nn->cld_net) + return 0; + + cn = kzalloc(sizeof(*cn), GFP_KERNEL); + if (!cn) { + ret = -ENOMEM; + goto err; + } + + cn->cn_pipe = rpc_mkpipe_data(&cld_upcall_ops, RPC_PIPE_WAIT_FOR_OPEN); + if (IS_ERR(cn->cn_pipe)) { + ret = PTR_ERR(cn->cn_pipe); + goto err; + } + spin_lock_init(&cn->cn_lock); + INIT_LIST_HEAD(&cn->cn_list); + + dentry = nfsd4_cld_register_net(net, cn->cn_pipe); + if (IS_ERR(dentry)) { + ret = PTR_ERR(dentry); + goto err_destroy_data; + } + + cn->cn_pipe->dentry = dentry; + nn->cld_net = cn; + return 0; + +err_destroy_data: + rpc_destroy_pipe_data(cn->cn_pipe); +err: + kfree(cn); + printk(KERN_ERR "NFSD: unable to create nfsdcld upcall pipe (%d)\n", + ret); + return ret; +} + +static void +nfsd4_remove_cld_pipe(struct net *net) +{ + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + struct cld_net *cn = nn->cld_net; + + nfsd4_cld_unregister_net(net, cn->cn_pipe); + rpc_destroy_pipe_data(cn->cn_pipe); + kfree(nn->cld_net); + nn->cld_net = NULL; +} + +static struct cld_upcall * +alloc_cld_upcall(struct cld_net *cn) +{ + struct cld_upcall *new, *tmp; + + new = kzalloc(sizeof(*new), GFP_KERNEL); + if (!new) + return new; + + /* FIXME: hard cap on number in flight? */ +restart_search: + spin_lock(&cn->cn_lock); + list_for_each_entry(tmp, &cn->cn_list, cu_list) { + if (tmp->cu_msg.cm_xid == cn->cn_xid) { + cn->cn_xid++; + spin_unlock(&cn->cn_lock); + goto restart_search; + } + } + new->cu_task = current; + new->cu_msg.cm_vers = CLD_UPCALL_VERSION; + put_unaligned(cn->cn_xid++, &new->cu_msg.cm_xid); + new->cu_net = cn; + list_add(&new->cu_list, &cn->cn_list); + spin_unlock(&cn->cn_lock); + + dprintk("%s: allocated xid %u\n", __func__, new->cu_msg.cm_xid); + + return new; +} + +static void +free_cld_upcall(struct cld_upcall *victim) +{ + struct cld_net *cn = victim->cu_net; + + spin_lock(&cn->cn_lock); + list_del(&victim->cu_list); + spin_unlock(&cn->cn_lock); + kfree(victim); +} + +/* Ask daemon to create a new record */ +static void +nfsd4_cld_create(struct nfs4_client *clp) +{ + int ret; + struct cld_upcall *cup; + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); + struct cld_net *cn = nn->cld_net; + + /* Don't upcall if it's already stored */ + if (test_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags)) + return; + + cup = alloc_cld_upcall(cn); + if (!cup) { + ret = -ENOMEM; + goto out_err; + } + + cup->cu_msg.cm_cmd = Cld_Create; + cup->cu_msg.cm_u.cm_name.cn_len = clp->cl_name.len; + memcpy(cup->cu_msg.cm_u.cm_name.cn_id, clp->cl_name.data, + clp->cl_name.len); + + ret = cld_pipe_upcall(cn->cn_pipe, &cup->cu_msg); + if (!ret) { + ret = cup->cu_msg.cm_status; + set_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags); + } + + free_cld_upcall(cup); +out_err: + if (ret) + printk(KERN_ERR "NFSD: Unable to create client " + "record on stable storage: %d\n", ret); +} + +/* Ask daemon to create a new record */ +static void +nfsd4_cld_remove(struct nfs4_client *clp) +{ + int ret; + struct cld_upcall *cup; + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); + struct cld_net *cn = nn->cld_net; + + /* Don't upcall if it's already removed */ + if (!test_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags)) + return; + + cup = alloc_cld_upcall(cn); + if (!cup) { + ret = -ENOMEM; + goto out_err; + } + + cup->cu_msg.cm_cmd = Cld_Remove; + cup->cu_msg.cm_u.cm_name.cn_len = clp->cl_name.len; + memcpy(cup->cu_msg.cm_u.cm_name.cn_id, clp->cl_name.data, + clp->cl_name.len); + + ret = cld_pipe_upcall(cn->cn_pipe, &cup->cu_msg); + if (!ret) { + ret = cup->cu_msg.cm_status; + clear_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags); + } + + free_cld_upcall(cup); +out_err: + if (ret) + printk(KERN_ERR "NFSD: Unable to remove client " + "record from stable storage: %d\n", ret); +} + +/* Check for presence of a record, and update its timestamp */ +static int +nfsd4_cld_check(struct nfs4_client *clp) +{ + int ret; + struct cld_upcall *cup; + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); + struct cld_net *cn = nn->cld_net; + + /* Don't upcall if one was already stored during this grace pd */ + if (test_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags)) + return 0; + + cup = alloc_cld_upcall(cn); + if (!cup) { + printk(KERN_ERR "NFSD: Unable to check client record on " + "stable storage: %d\n", -ENOMEM); + return -ENOMEM; + } + + cup->cu_msg.cm_cmd = Cld_Check; + cup->cu_msg.cm_u.cm_name.cn_len = clp->cl_name.len; + memcpy(cup->cu_msg.cm_u.cm_name.cn_id, clp->cl_name.data, + clp->cl_name.len); + + ret = cld_pipe_upcall(cn->cn_pipe, &cup->cu_msg); + if (!ret) { + ret = cup->cu_msg.cm_status; + set_bit(NFSD4_CLIENT_STABLE, &clp->cl_flags); + } + + free_cld_upcall(cup); + return ret; +} + +static void +nfsd4_cld_grace_done(struct nfsd_net *nn, time_t boot_time) +{ + int ret; + struct cld_upcall *cup; + struct cld_net *cn = nn->cld_net; + + cup = alloc_cld_upcall(cn); + if (!cup) { + ret = -ENOMEM; + goto out_err; + } + + cup->cu_msg.cm_cmd = Cld_GraceDone; + cup->cu_msg.cm_u.cm_gracetime = (int64_t)boot_time; + ret = cld_pipe_upcall(cn->cn_pipe, &cup->cu_msg); + if (!ret) + ret = cup->cu_msg.cm_status; + + free_cld_upcall(cup); +out_err: + if (ret) + printk(KERN_ERR "NFSD: Unable to end grace period: %d\n", ret); +} + +static struct nfsd4_client_tracking_ops nfsd4_cld_tracking_ops = { + .init = nfsd4_init_cld_pipe, + .exit = nfsd4_remove_cld_pipe, + .create = nfsd4_cld_create, + .remove = nfsd4_cld_remove, + .check = nfsd4_cld_check, + .grace_done = nfsd4_cld_grace_done, +}; + +/* upcall via usermodehelper */ +static char cltrack_prog[PATH_MAX] = "/sbin/nfsdcltrack"; +module_param_string(cltrack_prog, cltrack_prog, sizeof(cltrack_prog), + S_IRUGO|S_IWUSR); +MODULE_PARM_DESC(cltrack_prog, "Path to the nfsdcltrack upcall program"); + +static bool cltrack_legacy_disable; +module_param(cltrack_legacy_disable, bool, S_IRUGO|S_IWUSR); +MODULE_PARM_DESC(cltrack_legacy_disable, + "Disable legacy recoverydir conversion. Default: false"); + +#define LEGACY_TOPDIR_ENV_PREFIX "NFSDCLTRACK_LEGACY_TOPDIR=" +#define LEGACY_RECDIR_ENV_PREFIX "NFSDCLTRACK_LEGACY_RECDIR=" + +static char * +nfsd4_cltrack_legacy_topdir(void) +{ + int copied; + size_t len; + char *result; + + if (cltrack_legacy_disable) + return NULL; + + len = strlen(LEGACY_TOPDIR_ENV_PREFIX) + + strlen(nfs4_recoverydir()) + 1; + + result = kmalloc(len, GFP_KERNEL); + if (!result) + return result; + + copied = snprintf(result, len, LEGACY_TOPDIR_ENV_PREFIX "%s", + nfs4_recoverydir()); + if (copied >= len) { + /* just return nothing if output was truncated */ + kfree(result); + return NULL; + } + + return result; +} + +static char * +nfsd4_cltrack_legacy_recdir(const struct xdr_netobj *name) +{ + int copied; + size_t len; + char *result; + + if (cltrack_legacy_disable) + return NULL; + + /* +1 is for '/' between "topdir" and "recdir" */ + len = strlen(LEGACY_RECDIR_ENV_PREFIX) + + strlen(nfs4_recoverydir()) + 1 + HEXDIR_LEN; + + result = kmalloc(len, GFP_KERNEL); + if (!result) + return result; + + copied = snprintf(result, len, LEGACY_RECDIR_ENV_PREFIX "%s/", + nfs4_recoverydir()); + if (copied > (len - HEXDIR_LEN)) { + /* just return nothing if output will be truncated */ + kfree(result); + return NULL; + } + + copied = nfs4_make_rec_clidname(result + copied, name); + if (copied) { + kfree(result); + return NULL; + } + + return result; +} + +static int +nfsd4_umh_cltrack_upcall(char *cmd, char *arg, char *legacy) +{ + char *envp[2]; + char *argv[4]; + int ret; + + if (unlikely(!cltrack_prog[0])) { + dprintk("%s: cltrack_prog is disabled\n", __func__); + return -EACCES; + } + + dprintk("%s: cmd: %s\n", __func__, cmd); + dprintk("%s: arg: %s\n", __func__, arg ? arg : "(null)"); + dprintk("%s: legacy: %s\n", __func__, legacy ? legacy : "(null)"); + + envp[0] = legacy; + envp[1] = NULL; + + argv[0] = (char *)cltrack_prog; + argv[1] = cmd; + argv[2] = arg; + argv[3] = NULL; + + ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_PROC); + /* + * Disable the upcall mechanism if we're getting an ENOENT or EACCES + * error. The admin can re-enable it on the fly by using sysfs + * once the problem has been fixed. + */ + if (ret == -ENOENT || ret == -EACCES) { + dprintk("NFSD: %s was not found or isn't executable (%d). " + "Setting cltrack_prog to blank string!", + cltrack_prog, ret); + cltrack_prog[0] = '\0'; + } + dprintk("%s: %s return value: %d\n", __func__, cltrack_prog, ret); + + return ret; +} + +static char * +bin_to_hex_dup(const unsigned char *src, int srclen) +{ + int i; + char *buf, *hex; + + /* +1 for terminating NULL */ + buf = kmalloc((srclen * 2) + 1, GFP_KERNEL); + if (!buf) + return buf; + + hex = buf; + for (i = 0; i < srclen; i++) { + sprintf(hex, "%2.2x", *src++); + hex += 2; + } + return buf; +} + +static int +nfsd4_umh_cltrack_init(struct net __attribute__((unused)) *net) +{ + /* XXX: The usermode helper s not working in container yet. */ + if (net != &init_net) { + WARN(1, KERN_ERR "NFSD: attempt to initialize umh client " + "tracking in a container!\n"); + return -EINVAL; + } + return nfsd4_umh_cltrack_upcall("init", NULL, NULL); +} + +static void +nfsd4_umh_cltrack_create(struct nfs4_client *clp) +{ + char *hexid; + + hexid = bin_to_hex_dup(clp->cl_name.data, clp->cl_name.len); + if (!hexid) { + dprintk("%s: can't allocate memory for upcall!\n", __func__); + return; + } + nfsd4_umh_cltrack_upcall("create", hexid, NULL); + kfree(hexid); +} + +static void +nfsd4_umh_cltrack_remove(struct nfs4_client *clp) +{ + char *hexid; + + hexid = bin_to_hex_dup(clp->cl_name.data, clp->cl_name.len); + if (!hexid) { + dprintk("%s: can't allocate memory for upcall!\n", __func__); + return; + } + nfsd4_umh_cltrack_upcall("remove", hexid, NULL); + kfree(hexid); +} + +static int +nfsd4_umh_cltrack_check(struct nfs4_client *clp) +{ + int ret; + char *hexid, *legacy; + + hexid = bin_to_hex_dup(clp->cl_name.data, clp->cl_name.len); + if (!hexid) { + dprintk("%s: can't allocate memory for upcall!\n", __func__); + return -ENOMEM; + } + legacy = nfsd4_cltrack_legacy_recdir(&clp->cl_name); + ret = nfsd4_umh_cltrack_upcall("check", hexid, legacy); + kfree(legacy); + kfree(hexid); + return ret; +} + +static void +nfsd4_umh_cltrack_grace_done(struct nfsd_net __attribute__((unused)) *nn, + time_t boot_time) +{ + char *legacy; + char timestr[22]; /* FIXME: better way to determine max size? */ + + sprintf(timestr, "%ld", boot_time); + legacy = nfsd4_cltrack_legacy_topdir(); + nfsd4_umh_cltrack_upcall("gracedone", timestr, legacy); + kfree(legacy); +} + +static struct nfsd4_client_tracking_ops nfsd4_umh_tracking_ops = { + .init = nfsd4_umh_cltrack_init, + .exit = NULL, + .create = nfsd4_umh_cltrack_create, + .remove = nfsd4_umh_cltrack_remove, + .check = nfsd4_umh_cltrack_check, + .grace_done = nfsd4_umh_cltrack_grace_done, +}; + +int +nfsd4_client_tracking_init(struct net *net) +{ + int status; + struct path path; + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + + /* just run the init if it the method is already decided */ + if (nn->client_tracking_ops) + goto do_init; + + /* + * First, try a UMH upcall. It should succeed or fail quickly, so + * there's little harm in trying that first. + */ + nn->client_tracking_ops = &nfsd4_umh_tracking_ops; + status = nn->client_tracking_ops->init(net); + if (!status) + return status; + + /* + * See if the recoverydir exists and is a directory. If it is, + * then use the legacy ops. + */ + nn->client_tracking_ops = &nfsd4_legacy_tracking_ops; + status = kern_path(nfs4_recoverydir(), LOOKUP_FOLLOW, &path); + if (!status) { + status = S_ISDIR(path.dentry->d_inode->i_mode); + path_put(&path); + if (status) + goto do_init; + } + + /* Finally, try to use nfsdcld */ + nn->client_tracking_ops = &nfsd4_cld_tracking_ops; + printk(KERN_WARNING "NFSD: the nfsdcld client tracking upcall will be " + "removed in 3.10. Please transition to using " + "nfsdcltrack.\n"); +do_init: + status = nn->client_tracking_ops->init(net); + if (status) { + printk(KERN_WARNING "NFSD: Unable to initialize client " + "recovery tracking! (%d)\n", status); + nn->client_tracking_ops = NULL; + } + return status; +} + +void +nfsd4_client_tracking_exit(struct net *net) +{ + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + + if (nn->client_tracking_ops) { + if (nn->client_tracking_ops->exit) + nn->client_tracking_ops->exit(net); + nn->client_tracking_ops = NULL; + } +} + +void +nfsd4_client_record_create(struct nfs4_client *clp) +{ + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); + + if (nn->client_tracking_ops) + nn->client_tracking_ops->create(clp); +} + +void +nfsd4_client_record_remove(struct nfs4_client *clp) +{ + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); + + if (nn->client_tracking_ops) + nn->client_tracking_ops->remove(clp); +} + +int +nfsd4_client_record_check(struct nfs4_client *clp) +{ + struct nfsd_net *nn = net_generic(clp->net, nfsd_net_id); + + if (nn->client_tracking_ops) + return nn->client_tracking_ops->check(clp); + + return -EOPNOTSUPP; +} + +void +nfsd4_record_grace_done(struct nfsd_net *nn, time_t boot_time) +{ + if (nn->client_tracking_ops) + nn->client_tracking_ops->grace_done(nn, boot_time); +} + +static int +rpc_pipefs_event(struct notifier_block *nb, unsigned long event, void *ptr) +{ + struct super_block *sb = ptr; + struct net *net = sb->s_fs_info; + struct nfsd_net *nn = net_generic(net, nfsd_net_id); + struct cld_net *cn = nn->cld_net; + struct dentry *dentry; + int ret = 0; + + if (!try_module_get(THIS_MODULE)) + return 0; + + if (!cn) { + module_put(THIS_MODULE); + return 0; + } + + switch (event) { + case RPC_PIPEFS_MOUNT: + dentry = nfsd4_cld_register_sb(sb, cn->cn_pipe); + if (IS_ERR(dentry)) { + ret = PTR_ERR(dentry); + break; + } + cn->cn_pipe->dentry = dentry; + break; + case RPC_PIPEFS_UMOUNT: + if (cn->cn_pipe->dentry) + nfsd4_cld_unregister_sb(cn->cn_pipe); + break; + default: + ret = -ENOTSUPP; + break; + } + module_put(THIS_MODULE); + return ret; +} + +static struct notifier_block nfsd4_cld_block = { + .notifier_call = rpc_pipefs_event, +}; + +int +register_cld_notifier(void) +{ + return rpc_pipefs_notifier_register(&nfsd4_cld_block); +} + +void +unregister_cld_notifier(void) +{ + rpc_pipefs_notifier_unregister(&nfsd4_cld_block); } |
