1 files changed, 237 insertions, 108 deletions

diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c
index 2b06410e584e..72ad836e4fe0 100644
--- a/net/sunrpc/cache.c
+++ b/net/sunrpc/cache.c
@@ -28,21 +28,21 @@
 #include <linux/workqueue.h>
 #include <linux/mutex.h>
 #include <linux/pagemap.h>
-#include <linux/smp_lock.h>
 #include <asm/ioctls.h>
 #include <linux/sunrpc/types.h>
 #include <linux/sunrpc/cache.h>
 #include <linux/sunrpc/stats.h>
 #include <linux/sunrpc/rpc_pipe_fs.h>
+#include "netns.h"
 
 #define  RPCDBG_FACILITY RPCDBG_CACHE
 
-static int cache_defer_req(struct cache_req *req, struct cache_head *item);
+static bool cache_defer_req(struct cache_req *req, struct cache_head *item);
 static void cache_revisit_request(struct cache_head *item);
 
 static void cache_init(struct cache_head *h)
 {
-        time_t now = get_seconds();
+        time_t now = seconds_since_boot();
         h->next = NULL;
         h->flags = 0;
         kref_init(&h->ref);
@@ -52,7 +52,7 @@ static void cache_init(struct cache_head *h)
 
 static inline int cache_is_expired(struct cache_detail *detail, struct cache_head *h)
 {
-        return  (h->expiry_time < get_seconds()) ||
+        return  (h->expiry_time < seconds_since_boot()) ||
                 (detail->flush_time > h->last_refresh);
 }
 
@@ -127,7 +127,8 @@ static void cache_dequeue(struct cache_detail *detail, struct cache_head *ch);
 static void cache_fresh_locked(struct cache_head *head, time_t expiry)
 {
         head->expiry_time = expiry;
-        head->last_refresh = get_seconds();
+        head->last_refresh = seconds_since_boot();
+        smp_wmb(); /* paired with smp_rmb() in cache_is_valid() */
         set_bit(CACHE_VALID, &head->flags);
 }
 
@@ -208,11 +209,36 @@ static inline int cache_is_valid(struct cache_detail *detail, struct cache_head
                 /* entry is valid */
                 if (test_bit(CACHE_NEGATIVE, &h->flags))
                         return -ENOENT;
-                else
+                else {
+                        /*
+                         * In combination with write barrier in
+                         * sunrpc_cache_update, ensures that anyone
+                         * using the cache entry after this sees the
+                         * updated contents:
+                         */
+                        smp_rmb();
                         return 0;
+                }
         }
 }
 
+static int try_to_negate_entry(struct cache_detail *detail, struct cache_head *h)
+{
+        int rv;
+
+        write_lock(&detail->hash_lock);
+        rv = cache_is_valid(detail, h);
+        if (rv != -EAGAIN) {
+                write_unlock(&detail->hash_lock);
+                return rv;
+        }
+        set_bit(CACHE_NEGATIVE, &h->flags);
+        cache_fresh_locked(h, seconds_since_boot()+CACHE_NEW_EXPIRY);
+        write_unlock(&detail->hash_lock);
+        cache_fresh_unlocked(h, detail);
+        return -ENOENT;
+}
+
 /*
  * This is the generic cache management routine for all
  * the authentication caches.
@@ -238,7 +264,7 @@ int cache_check(struct cache_detail *detail,
 
         /* now see if we want to start an upcall */
         refresh_age = (h->expiry_time - h->last_refresh);
-        age = get_seconds() - h->last_refresh;
+        age = seconds_since_boot() - h->last_refresh;
 
         if (rqstp == NULL) {
                 if (rv == -EAGAIN)
@@ -251,14 +277,8 @@ int cache_check(struct cache_detail *detail,
                         case -EINVAL:
                                 clear_bit(CACHE_PENDING, &h->flags);
                                 cache_revisit_request(h);
-                                if (rv == -EAGAIN) {
-                                        set_bit(CACHE_NEGATIVE, &h->flags);
-                                        cache_fresh_locked(h, get_seconds()+CACHE_NEW_EXPIRY);
-                                        cache_fresh_unlocked(h, detail);
-                                        rv = -ENOENT;
-                                }
+                                rv = try_to_negate_entry(detail, h);
                                 break;
-
                         case -EAGAIN:
                                 clear_bit(CACHE_PENDING, &h->flags);
                                 cache_revisit_request(h);
@@ -268,8 +288,11 @@ int cache_check(struct cache_detail *detail,
         }
 
         if (rv == -EAGAIN) {
-                if (cache_defer_req(rqstp, h) < 0) {
-                        /* Request is not deferred */
+                if (!cache_defer_req(rqstp, h)) {
+                        /*
+                         * Request was not deferred; handle it as best
+                         * we can ourselves:
+                         */
                         rv = cache_is_valid(detail, h);
                         if (rv == -EAGAIN)
                                 rv = -ETIMEDOUT;
@@ -388,11 +411,11 @@ static int cache_clean(void)
                         return -1;
                 }
                 current_detail = list_entry(next, struct cache_detail, others);
-                if (current_detail->nextcheck > get_seconds())
+                if (current_detail->nextcheck > seconds_since_boot())
                         current_index = current_detail->hash_size;
                 else {
                         current_index = 0;
-                        current_detail->nextcheck = get_seconds()+30*60;
+                        current_detail->nextcheck = seconds_since_boot()+30*60;
                 }
         }
 
@@ -477,7 +500,7 @@ EXPORT_SYMBOL_GPL(cache_flush);
 void cache_purge(struct cache_detail *detail)
 {
         detail->flush_time = LONG_MAX;
-        detail->nextcheck = get_seconds();
+        detail->nextcheck = seconds_since_boot();
         cache_flush();
         detail->flush_time = 1;
 }
@@ -506,81 +529,157 @@ EXPORT_SYMBOL_GPL(cache_purge);
 
 static DEFINE_SPINLOCK(cache_defer_lock);
 static LIST_HEAD(cache_defer_list);
-static struct list_head cache_defer_hash[DFR_HASHSIZE];
+static struct hlist_head cache_defer_hash[DFR_HASHSIZE];
 static int cache_defer_cnt;
 
-static int cache_defer_req(struct cache_req *req, struct cache_head *item)
+static void __unhash_deferred_req(struct cache_deferred_req *dreq)
+{
+        hlist_del_init(&dreq->hash);
+        if (!list_empty(&dreq->recent)) {
+                list_del_init(&dreq->recent);
+                cache_defer_cnt--;
+        }
+}
+
+static void __hash_deferred_req(struct cache_deferred_req *dreq, struct cache_head *item)
 {
-        struct cache_deferred_req *dreq, *discard;
         int hash = DFR_HASH(item);
 
-        if (cache_defer_cnt >= DFR_MAX) {
-                /* too much in the cache, randomly drop this one,
-                 * or continue and drop the oldest below
-                 */
-                if (net_random()&1)
-                        return -ENOMEM;
-        }
-        dreq = req->defer(req);
-        if (dreq == NULL)
-                return -ENOMEM;
+        INIT_LIST_HEAD(&dreq->recent);
+        hlist_add_head(&dreq->hash, &cache_defer_hash[hash]);
+}
+
+static void setup_deferral(struct cache_deferred_req *dreq,
+                           struct cache_head *item,
+                           int count_me)
+{
 
         dreq->item = item;
 
         spin_lock(&cache_defer_lock);
 
-        list_add(&dreq->recent, &cache_defer_list);
+        __hash_deferred_req(dreq, item);
 
-        if (cache_defer_hash[hash].next == NULL)
-                INIT_LIST_HEAD(&cache_defer_hash[hash]);
-        list_add(&dreq->hash, &cache_defer_hash[hash]);
-
-        /* it is in, now maybe clean up */
-        discard = NULL;
-        if (++cache_defer_cnt > DFR_MAX) {
-                discard = list_entry(cache_defer_list.prev,
-                                     struct cache_deferred_req, recent);
-                list_del_init(&discard->recent);
-                list_del_init(&discard->hash);
-                cache_defer_cnt--;
+        if (count_me) {
+                cache_defer_cnt++;
+                list_add(&dreq->recent, &cache_defer_list);
         }
+
         spin_unlock(&cache_defer_lock);
 
+}
+
+struct thread_deferred_req {
+        struct cache_deferred_req handle;
+        struct completion completion;
+};
+
+static void cache_restart_thread(struct cache_deferred_req *dreq, int too_many)
+{
+        struct thread_deferred_req *dr =
+                container_of(dreq, struct thread_deferred_req, handle);
+        complete(&dr->completion);
+}
+
+static void cache_wait_req(struct cache_req *req, struct cache_head *item)
+{
+        struct thread_deferred_req sleeper;
+        struct cache_deferred_req *dreq = &sleeper.handle;
+
+        sleeper.completion = COMPLETION_INITIALIZER_ONSTACK(sleeper.completion);
+        dreq->revisit = cache_restart_thread;
+
+        setup_deferral(dreq, item, 0);
+
+        if (!test_bit(CACHE_PENDING, &item->flags) ||
+            wait_for_completion_interruptible_timeout(
+                    &sleeper.completion, req->thread_wait) <= 0) {
+                /* The completion wasn't completed, so we need
+                 * to clean up
+                 */
+                spin_lock(&cache_defer_lock);
+                if (!hlist_unhashed(&sleeper.handle.hash)) {
+                        __unhash_deferred_req(&sleeper.handle);
+                        spin_unlock(&cache_defer_lock);
+                } else {
+                        /* cache_revisit_request already removed
+                         * this from the hash table, but hasn't
+                         * called ->revisit yet.  It will very soon
+                         * and we need to wait for it.
+                         */
+                        spin_unlock(&cache_defer_lock);
+                        wait_for_completion(&sleeper.completion);
+                }
+        }
+}
+
+static void cache_limit_defers(void)
+{
+        /* Make sure we haven't exceed the limit of allowed deferred
+         * requests.
+         */
+        struct cache_deferred_req *discard = NULL;
+
+        if (cache_defer_cnt <= DFR_MAX)
+                return;
+
+        spin_lock(&cache_defer_lock);
+
+        /* Consider removing either the first or the last */
+        if (cache_defer_cnt > DFR_MAX) {
+                if (net_random() & 1)
+                        discard = list_entry(cache_defer_list.next,
+                                             struct cache_deferred_req, recent);
+                else
+                        discard = list_entry(cache_defer_list.prev,
+                                             struct cache_deferred_req, recent);
+                __unhash_deferred_req(discard);
+        }
+        spin_unlock(&cache_defer_lock);
         if (discard)
-                /* there was one too many */
                 discard->revisit(discard, 1);
+}
 
-        if (!test_bit(CACHE_PENDING, &item->flags)) {
-                /* must have just been validated... */
-                cache_revisit_request(item);
-                return -EAGAIN;
+/* Return true if and only if a deferred request is queued. */
+static bool cache_defer_req(struct cache_req *req, struct cache_head *item)
+{
+        struct cache_deferred_req *dreq;
+
+        if (req->thread_wait) {
+                cache_wait_req(req, item);
+                if (!test_bit(CACHE_PENDING, &item->flags))
+                        return false;
         }
-        return 0;
+        dreq = req->defer(req);
+        if (dreq == NULL)
+                return false;
+        setup_deferral(dreq, item, 1);
+        if (!test_bit(CACHE_PENDING, &item->flags))
+                /* Bit could have been cleared before we managed to
+                 * set up the deferral, so need to revisit just in case
+                 */
+                cache_revisit_request(item);
+
+        cache_limit_defers();
+        return true;
 }
 
 static void cache_revisit_request(struct cache_head *item)
 {
         struct cache_deferred_req *dreq;
         struct list_head pending;
-
-        struct list_head *lp;
+        struct hlist_node *lp, *tmp;
         int hash = DFR_HASH(item);
 
         INIT_LIST_HEAD(&pending);
         spin_lock(&cache_defer_lock);
 
-        lp = cache_defer_hash[hash].next;
-        if (lp) {
-                while (lp != &cache_defer_hash[hash]) {
-                        dreq = list_entry(lp, struct cache_deferred_req, hash);
-                        lp = lp->next;
-                        if (dreq->item == item) {
-                                list_del_init(&dreq->hash);
-                                list_move(&dreq->recent, &pending);
-                                cache_defer_cnt--;
-                        }
+        hlist_for_each_entry_safe(dreq, lp, tmp, &cache_defer_hash[hash], hash)
+                if (dreq->item == item) {
+                        __unhash_deferred_req(dreq);
+                        list_add(&dreq->recent, &pending);
                 }
-        }
+
         spin_unlock(&cache_defer_lock);
 
         while (!list_empty(&pending)) {
@@ -601,9 +700,8 @@ void cache_clean_deferred(void *owner)
 
         list_for_each_entry_safe(dreq, tmp, &cache_defer_list, recent) {
                 if (dreq->owner == owner) {
-                        list_del_init(&dreq->hash);
-                        list_move(&dreq->recent, &pending);
-                        cache_defer_cnt--;
+                        __unhash_deferred_req(dreq);
+                        list_add(&dreq->recent, &pending);
                 }
         }
         spin_unlock(&cache_defer_lock);
@@ -902,7 +1000,7 @@ static int cache_release(struct inode *inode, struct file *filp,
                 filp->private_data = NULL;
                 kfree(rp);
 
-                cd->last_close = get_seconds();
+                cd->last_close = seconds_since_boot();
                 atomic_dec(&cd->readers);
         }
         module_put(cd->owner);
@@ -1015,6 +1113,23 @@ static void warn_no_listener(struct cache_detail *detail)
         }
 }
 
+static bool cache_listeners_exist(struct cache_detail *detail)
+{
+        if (atomic_read(&detail->readers))
+                return true;
+        if (detail->last_close == 0)
+                /* This cache was never opened */
+                return false;
+        if (detail->last_close < seconds_since_boot() - 30)
+                /*
+                 * We allow for the possibility that someone might
+                 * restart a userspace daemon without restarting the
+                 * server; but after 30 seconds, we give up.
+                 */
+                 return false;
+        return true;
+}
+
 /*
  * register an upcall request to user-space and queue it up for read() by the
  * upcall daemon.
@@ -1033,10 +1148,9 @@ int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h,
         char *bp;
         int len;
 
-        if (atomic_read(&detail->readers) == 0 &&
-            detail->last_close < get_seconds() - 30) {
-                        warn_no_listener(detail);
-                        return -EINVAL;
+        if (!cache_listeners_exist(detail)) {
+                warn_no_listener(detail);
+                return -EINVAL;
         }
 
         buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
@@ -1095,13 +1209,19 @@ int qword_get(char **bpp, char *dest, int bufsize)
         if (bp[0] == '\\' && bp[1] == 'x') {
                 /* HEX STRING */
                 bp += 2;
-                while (isxdigit(bp[0]) && isxdigit(bp[1]) && len < bufsize) {
-                        int byte = isdigit(*bp) ? *bp-'0' : toupper(*bp)-'A'+10;
-                        bp++;
-                        byte <<= 4;
-                        byte |= isdigit(*bp) ? *bp-'0' : toupper(*bp)-'A'+10;
-                        *dest++ = byte;
-                        bp++;
+                while (len < bufsize) {
+                        int h, l;
+
+                        h = hex_to_bin(bp[0]);
+                        if (h < 0)
+                                break;
+
+                        l = hex_to_bin(bp[1]);
+                        if (l < 0)
+                                break;
+
+                        *dest++ = (h << 4) | l;
+                        bp += 2;
                         len++;
                 }
         } else {
@@ -1219,7 +1339,8 @@ static int c_show(struct seq_file *m, void *p)
 
         ifdebug(CACHE)
                 seq_printf(m, "# expiry=%ld refcnt=%d flags=%lx\n",
-                           cp->expiry_time, atomic_read(&cp->ref.refcount), cp->flags);
+                           convert_to_wallclock(cp->expiry_time),
+                           atomic_read(&cp->ref.refcount), cp->flags);
         cache_get(cp);
         if (cache_check(cd, cp, NULL))
                 /* cache_check does a cache_put on failure */
@@ -1285,7 +1406,7 @@ static ssize_t read_flush(struct file *file, char __user *buf,
         unsigned long p = *ppos;
         size_t len;
 
-        sprintf(tbuf, "%lu\n", cd->flush_time);
+        sprintf(tbuf, "%lu\n", convert_to_wallclock(cd->flush_time));
         len = strlen(tbuf);
         if (p >= len)
                 return 0;
@@ -1303,19 +1424,20 @@ static ssize_t write_flush(struct file *file, const char __user *buf,
                            struct cache_detail *cd)
 {
         char tbuf[20];
-        char *ep;
-        long flushtime;
+        char *bp, *ep;
+
         if (*ppos || count > sizeof(tbuf)-1)
                 return -EINVAL;
         if (copy_from_user(tbuf, buf, count))
                 return -EFAULT;
         tbuf[count] = 0;
-        flushtime = simple_strtoul(tbuf, &ep, 0);
+        simple_strtoul(tbuf, &ep, 0);
         if (*ep && *ep != '\n')
                 return -EINVAL;
 
-        cd->flush_time = flushtime;
-        cd->nextcheck = get_seconds();
+        bp = tbuf;
+        cd->flush_time = get_expiry(&bp);
+        cd->nextcheck = seconds_since_boot();
         cache_flush();
 
         *ppos += count;
@@ -1348,15 +1470,10 @@ static unsigned int cache_poll_procfs(struct file *filp, poll_table *wait)
 static long cache_ioctl_procfs(struct file *filp,
                                unsigned int cmd, unsigned long arg)
 {
-        long ret;
         struct inode *inode = filp->f_path.dentry->d_inode;
         struct cache_detail *cd = PDE(inode)->data;
 
-        lock_kernel();
-        ret = cache_ioctl(inode, filp, cmd, arg, cd);
-        unlock_kernel();
-
-        return ret;
+        return cache_ioctl(inode, filp, cmd, arg, cd);
 }
 
 static int cache_open_procfs(struct inode *inode, struct file *filp)
@@ -1441,10 +1558,13 @@ static const struct file_operations cache_flush_operations_procfs = {
         .read           = read_flush_procfs,
         .write          = write_flush_procfs,
         .release        = release_flush_procfs,
+        .llseek         = no_llseek,
 };
 
-static void remove_cache_proc_entries(struct cache_detail *cd)
+static void remove_cache_proc_entries(struct cache_detail *cd, struct net *net)
 {
+        struct sunrpc_net *sn;
+
         if (cd->u.procfs.proc_ent == NULL)
                 return;
         if (cd->u.procfs.flush_ent)
@@ -1454,15 +1574,18 @@ static void remove_cache_proc_entries(struct cache_detail *cd)
         if (cd->u.procfs.content_ent)
                 remove_proc_entry("content", cd->u.procfs.proc_ent);
         cd->u.procfs.proc_ent = NULL;
-        remove_proc_entry(cd->name, proc_net_rpc);
+        sn = net_generic(net, sunrpc_net_id);
+        remove_proc_entry(cd->name, sn->proc_net_rpc);
 }
 
 #ifdef CONFIG_PROC_FS
-static int create_cache_proc_entries(struct cache_detail *cd)
+static int create_cache_proc_entries(struct cache_detail *cd, struct net *net)
 {
         struct proc_dir_entry *p;
+        struct sunrpc_net *sn;
 
-        cd->u.procfs.proc_ent = proc_mkdir(cd->name, proc_net_rpc);
+        sn = net_generic(net, sunrpc_net_id);
+        cd->u.procfs.proc_ent = proc_mkdir(cd->name, sn->proc_net_rpc);
         if (cd->u.procfs.proc_ent == NULL)
                 goto out_nomem;
         cd->u.procfs.channel_ent = NULL;
@@ -1493,11 +1616,11 @@ static int create_cache_proc_entries(struct cache_detail *cd)
         }
         return 0;
 out_nomem:
-        remove_cache_proc_entries(cd);
+        remove_cache_proc_entries(cd, net);
         return -ENOMEM;
 }
 #else /* CONFIG_PROC_FS */
-static int create_cache_proc_entries(struct cache_detail *cd)
+static int create_cache_proc_entries(struct cache_detail *cd, struct net *net)
 {
         return 0;
 }
@@ -1508,23 +1631,33 @@ void __init cache_initialize(void)
         INIT_DELAYED_WORK_DEFERRABLE(&cache_cleaner, do_cache_clean);
 }
 
-int cache_register(struct cache_detail *cd)
+int cache_register_net(struct cache_detail *cd, struct net *net)
 {
         int ret;
 
         sunrpc_init_cache_detail(cd);
-        ret = create_cache_proc_entries(cd);
+        ret = create_cache_proc_entries(cd, net);
         if (ret)
                 sunrpc_destroy_cache_detail(cd);
         return ret;
 }
+
+int cache_register(struct cache_detail *cd)
+{
+        return cache_register_net(cd, &init_net);
+}
 EXPORT_SYMBOL_GPL(cache_register);
 
-void cache_unregister(struct cache_detail *cd)
+void cache_unregister_net(struct cache_detail *cd, struct net *net)
 {
-        remove_cache_proc_entries(cd);
+        remove_cache_proc_entries(cd, net);
         sunrpc_destroy_cache_detail(cd);
 }
+
+void cache_unregister(struct cache_detail *cd)
+{
+        cache_unregister_net(cd, &init_net);
+}
 EXPORT_SYMBOL_GPL(cache_unregister);
 
 static ssize_t cache_read_pipefs(struct file *filp, char __user *buf,
@@ -1555,13 +1688,8 @@ static long cache_ioctl_pipefs(struct file *filp,
 {
         struct inode *inode = filp->f_dentry->d_inode;
         struct cache_detail *cd = RPC_I(inode)->private;
-        long ret;
 
-        lock_kernel();
-        ret = cache_ioctl(inode, filp, cmd, arg, cd);
-        unlock_kernel();
-
-        return ret;
+        return cache_ioctl(inode, filp, cmd, arg, cd);
 }
 
 static int cache_open_pipefs(struct inode *inode, struct file *filp)
@@ -1646,6 +1774,7 @@ const struct file_operations cache_flush_operations_pipefs = {
         .read           = read_flush_pipefs,
         .write          = write_flush_pipefs,
         .release        = release_flush_pipefs,
+        .llseek         = no_llseek,
 };
 
 int sunrpc_cache_register_pipefs(struct dentry *parent,