ipc_findkey() used to scan all objects to look for the wanted key. This
is slow when using a high number of keys. This change adds an rhashtable
of kern_ipc_perm objects in ipc_ids, so that one lookup cease to be O(n).
This change gives a 865% improvement of benchmark reaim.jobs_per_min on a
56 threads Intel(R) Xeon(R) CPU E5-2695 v3 @ 2.30GHz with 256G memory [1]
Other (more micro) benchmark results, by the author: On an i5 laptop, the
following loop executed right after a reboot took, without and with this
change:
for (int i = 0, k=0x424242; i < KEYS; ++i)
semget(k++, 1, IPC_CREAT | 0600);
total total max single max single
KEYS without with call without call with
1 3.5 4.9 µs 3.5 4.9
10 7.6 8.6 µs 3.7 4.7
32 16.2 15.9 µs 4.3 5.3
100 72.9 41.8 µs 3.7 4.7
1000 5,630.0 502.0 µs * *
10000 1,340,000.0 7,240.0 µs * *
31900 17,600,000.0 22,200.0 µs * *
*: unreliable measure: high variance
The duration for a lookup-only usage was obtained by the same loop once
the keys are present:
total total max single max single
KEYS without with call without call with
1 2.1 2.5 µs 2.1 2.5
10 4.5 4.8 µs 2.2 2.3
32 13.0 10.8 µs 2.3 2.8
100 82.9 25.1 µs * 2.3
1000 5,780.0 217.0 µs * *
10000 1,470,000.0 2,520.0 µs * *
31900 17,400,000.0 7,810.0 µs * *
Finally, executing each semget() in a new process gave, when still
summing only the durations of these syscalls:
creation:
total total
KEYS without with
1 3.7 5.0 µs
10 32.9 36.7 µs
32 125.0 109.0 µs
100 523.0 353.0 µs
1000 20,300.0 3,280.0 µs
10000 2,470,000.0 46,700.0 µs
31900 27,800,000.0 219,000.0 µs
lookup-only:
total total
KEYS without with
1 2.5 2.7 µs
10 25.4 24.4 µs
32 106.0 72.6 µs
100 591.0 352.0 µs
1000 22,400.0 2,250.0 µs
10000 2,510,000.0 25,700.0 µs
31900 28,200,000.0 115,000.0 µs
[1] http://lkml.kernel.org/r/
20170814060507.GE23258@yexl-desktop
Link: http://lkml.kernel.org/r/20170815194954.ck32ta2z35yuzpwp@debix
Signed-off-by: Guillaume Knispel <guillaume.knispel@supersonicimagine.com>
Reviewed-by: Marc Pardo <marc.pardo@supersonicimagine.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Kees Cook <keescook@chromium.org>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Serge Hallyn <serge@hallyn.com>
Cc: Andrey Vagin <avagin@openvz.org>
Cc: Guillaume Knispel <guillaume.knispel@supersonicimagine.com>
Cc: Marc Pardo <marc.pardo@supersonicimagine.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
#include <linux/spinlock.h>
#include <linux/uidgid.h>
+#include <linux/rhashtable.h>
#include <uapi/linux/ipc.h>
#include <linux/refcount.h>
unsigned long seq;
void *security;
+ struct rhash_head khtnode;
+
struct rcu_head rcu;
refcount_t refcount;
} ____cacheline_aligned_in_smp __randomize_layout;
#include <linux/nsproxy.h>
#include <linux/ns_common.h>
#include <linux/refcount.h>
+#include <linux/rhashtable.h>
struct user_namespace;
struct ipc_ids {
int in_use;
unsigned short seq;
+ bool tables_initialized;
struct rw_semaphore rwsem;
struct idr ipcs_idr;
int next_id;
+ struct rhashtable key_ht;
};
struct ipc_namespace {
}
-void msg_init_ns(struct ipc_namespace *ns)
+int msg_init_ns(struct ipc_namespace *ns)
{
ns->msg_ctlmax = MSGMAX;
ns->msg_ctlmnb = MSGMNB;
atomic_set(&ns->msg_bytes, 0);
atomic_set(&ns->msg_hdrs, 0);
- ipc_init_ids(&ns->ids[IPC_MSG_IDS]);
+ return ipc_init_ids(&ns->ids[IPC_MSG_IDS]);
}
#ifdef CONFIG_IPC_NS
{
free_ipcs(ns, &msg_ids(ns), freeque);
idr_destroy(&ns->ids[IPC_MSG_IDS].ipcs_idr);
+ rhashtable_destroy(&ns->ids[IPC_MSG_IDS].key_ht);
}
#endif
}
#endif
-void __init msg_init(void)
+int __init msg_init(void)
{
- msg_init_ns(&init_ipc_ns);
+ const int err = msg_init_ns(&init_ipc_ns);
ipc_init_proc_interface("sysvipc/msg",
" key msqid perms cbytes qnum lspid lrpid uid gid cuid cgid stime rtime ctime\n",
IPC_MSG_IDS, sysvipc_msg_proc_show);
+ return err;
}
ns->user_ns = get_user_ns(user_ns);
ns->ucounts = ucounts;
- err = mq_init_ns(ns);
+ err = sem_init_ns(ns);
if (err)
goto fail_put;
+ err = msg_init_ns(ns);
+ if (err)
+ goto fail_destroy_sem;
+ err = shm_init_ns(ns);
+ if (err)
+ goto fail_destroy_msg;
- sem_init_ns(ns);
- msg_init_ns(ns);
- shm_init_ns(ns);
+ err = mq_init_ns(ns);
+ if (err)
+ goto fail_destroy_shm;
return ns;
+fail_destroy_shm:
+ shm_exit_ns(ns);
+fail_destroy_msg:
+ msg_exit_ns(ns);
+fail_destroy_sem:
+ sem_exit_ns(ns);
fail_put:
put_user_ns(ns->user_ns);
ns_free_inum(&ns->ns);
#define sc_semopm sem_ctls[2]
#define sc_semmni sem_ctls[3]
-void sem_init_ns(struct ipc_namespace *ns)
+int sem_init_ns(struct ipc_namespace *ns)
{
ns->sc_semmsl = SEMMSL;
ns->sc_semmns = SEMMNS;
ns->sc_semopm = SEMOPM;
ns->sc_semmni = SEMMNI;
ns->used_sems = 0;
- ipc_init_ids(&ns->ids[IPC_SEM_IDS]);
+ return ipc_init_ids(&ns->ids[IPC_SEM_IDS]);
}
#ifdef CONFIG_IPC_NS
{
free_ipcs(ns, &sem_ids(ns), freeary);
idr_destroy(&ns->ids[IPC_SEM_IDS].ipcs_idr);
+ rhashtable_destroy(&ns->ids[IPC_SEM_IDS].key_ht);
}
#endif
-void __init sem_init(void)
+int __init sem_init(void)
{
- sem_init_ns(&init_ipc_ns);
+ const int err = sem_init_ns(&init_ipc_ns);
+
ipc_init_proc_interface("sysvipc/sem",
" key semid perms nsems uid gid cuid cgid otime ctime\n",
IPC_SEM_IDS, sysvipc_sem_proc_show);
+ return err;
}
/**
static int sysvipc_shm_proc_show(struct seq_file *s, void *it);
#endif
-void shm_init_ns(struct ipc_namespace *ns)
+int shm_init_ns(struct ipc_namespace *ns)
{
ns->shm_ctlmax = SHMMAX;
ns->shm_ctlall = SHMALL;
ns->shm_ctlmni = SHMMNI;
ns->shm_rmid_forced = 0;
ns->shm_tot = 0;
- ipc_init_ids(&shm_ids(ns));
+ return ipc_init_ids(&shm_ids(ns));
}
/*
if (shp->shm_nattch) {
shp->shm_perm.mode |= SHM_DEST;
/* Do not find it any more */
- shp->shm_perm.key = IPC_PRIVATE;
+ ipc_set_key_private(&shm_ids(ns), &shp->shm_perm);
shm_unlock(shp);
} else
shm_destroy(ns, shp);
{
free_ipcs(ns, &shm_ids(ns), do_shm_rmid);
idr_destroy(&ns->ids[IPC_SHM_IDS].ipcs_idr);
+ rhashtable_destroy(&ns->ids[IPC_SHM_IDS].key_ht);
}
#endif
static int __init ipc_ns_init(void)
{
- shm_init_ns(&init_ipc_ns);
- return 0;
+ const int err = shm_init_ns(&init_ipc_ns);
+ WARN(err, "ipc: sysv shm_init_ns failed: %d\n", err);
+ return err;
}
pure_initcall(ipc_ns_init);
*/
static int __init ipc_init(void)
{
- sem_init();
- msg_init();
+ int err_sem, err_msg;
+
+ err_sem = sem_init();
+ WARN(err_sem, "ipc: sysv sem_init failed: %d\n", err_sem);
+ err_msg = msg_init();
+ WARN(err_msg, "ipc: sysv msg_init failed: %d\n", err_msg);
shm_init();
- return 0;
+
+ return err_msg ? err_msg : err_sem;
}
device_initcall(ipc_init);
+static const struct rhashtable_params ipc_kht_params = {
+ .head_offset = offsetof(struct kern_ipc_perm, khtnode),
+ .key_offset = offsetof(struct kern_ipc_perm, key),
+ .key_len = FIELD_SIZEOF(struct kern_ipc_perm, key),
+ .locks_mul = 1,
+ .automatic_shrinking = true,
+};
+
/**
* ipc_init_ids - initialise ipc identifiers
* @ids: ipc identifier set
*
* Set up the sequence range to use for the ipc identifier range (limited
- * below IPCMNI) then initialise the ids idr.
+ * below IPCMNI) then initialise the keys hashtable and ids idr.
*/
-void ipc_init_ids(struct ipc_ids *ids)
+int ipc_init_ids(struct ipc_ids *ids)
{
+ int err;
ids->in_use = 0;
ids->seq = 0;
ids->next_id = -1;
init_rwsem(&ids->rwsem);
+ err = rhashtable_init(&ids->key_ht, &ipc_kht_params);
+ if (err)
+ return err;
idr_init(&ids->ipcs_idr);
+ ids->tables_initialized = true;
+ return 0;
}
#ifdef CONFIG_PROC_FS
* Returns the locked pointer to the ipc structure if found or NULL
* otherwise. If key is found ipc points to the owning ipc structure
*
- * Called with ipc_ids.rwsem held.
+ * Called with writer ipc_ids.rwsem held.
*/
static struct kern_ipc_perm *ipc_findkey(struct ipc_ids *ids, key_t key)
{
- struct kern_ipc_perm *ipc;
- int next_id;
- int total;
-
- for (total = 0, next_id = 0; total < ids->in_use; next_id++) {
- ipc = idr_find(&ids->ipcs_idr, next_id);
-
- if (ipc == NULL)
- continue;
+ struct kern_ipc_perm *ipcp = NULL;
- if (ipc->key != key) {
- total++;
- continue;
- }
+ if (likely(ids->tables_initialized))
+ ipcp = rhashtable_lookup_fast(&ids->key_ht, &key,
+ ipc_kht_params);
+ if (ipcp) {
rcu_read_lock();
- ipc_lock_object(ipc);
- return ipc;
+ ipc_lock_object(ipcp);
+ return ipcp;
}
return NULL;
{
kuid_t euid;
kgid_t egid;
- int id;
+ int id, err;
int next_id = ids->next_id;
if (size > IPCMNI)
size = IPCMNI;
- if (ids->in_use >= size)
+ if (!ids->tables_initialized || ids->in_use >= size)
return -ENOSPC;
idr_preload(GFP_KERNEL);
(next_id < 0) ? 0 : ipcid_to_idx(next_id), 0,
GFP_NOWAIT);
idr_preload_end();
+
+ if (id >= 0 && new->key != IPC_PRIVATE) {
+ err = rhashtable_insert_fast(&ids->key_ht, &new->khtnode,
+ ipc_kht_params);
+ if (err < 0) {
+ idr_remove(&ids->ipcs_idr, id);
+ id = err;
+ }
+ }
if (id < 0) {
spin_unlock(&new->lock);
rcu_read_unlock();
return err;
}
+/**
+ * ipc_kht_remove - remove an ipc from the key hashtable
+ * @ids: ipc identifier set
+ * @ipcp: ipc perm structure containing the key to remove
+ *
+ * ipc_ids.rwsem (as a writer) and the spinlock for this ID are held
+ * before this function is called, and remain locked on the exit.
+ */
+static void ipc_kht_remove(struct ipc_ids *ids, struct kern_ipc_perm *ipcp)
+{
+ if (ipcp->key != IPC_PRIVATE)
+ rhashtable_remove_fast(&ids->key_ht, &ipcp->khtnode,
+ ipc_kht_params);
+}
/**
* ipc_rmid - remove an ipc identifier
int lid = ipcid_to_idx(ipcp->id);
idr_remove(&ids->ipcs_idr, lid);
+ ipc_kht_remove(ids, ipcp);
ids->in_use--;
ipcp->deleted = true;
}
+/**
+ * ipc_set_key_private - switch the key of an existing ipc to IPC_PRIVATE
+ * @ids: ipc identifier set
+ * @ipcp: ipc perm structure containing the key to modify
+ *
+ * ipc_ids.rwsem (as a writer) and the spinlock for this ID are held
+ * before this function is called, and remain locked on the exit.
+ */
+void ipc_set_key_private(struct ipc_ids *ids, struct kern_ipc_perm *ipcp)
+{
+ ipc_kht_remove(ids, ipcp);
+ ipcp->key = IPC_PRIVATE;
+}
+
int ipc_rcu_getref(struct kern_ipc_perm *ptr)
{
return refcount_inc_not_zero(&ptr->refcount);
}
/**
- * ipc_obtain_object
+ * ipc_obtain_object_idr
* @ids: ipc identifier set
* @id: ipc id to look for
*
struct kern_ipc_perm *out;
int lid = ipcid_to_idx(id);
+ if (unlikely(!ids->tables_initialized))
+ return ERR_PTR(-EINVAL);
+
out = idr_find(&ids->ipcs_idr, lid);
if (!out)
return ERR_PTR(-EINVAL);
#define SEQ_MULTIPLIER (IPCMNI)
-void sem_init(void);
-void msg_init(void);
+int sem_init(void);
+int msg_init(void);
void shm_init(void);
struct ipc_namespace;
#endif
#ifdef CONFIG_SYSVIPC
-void sem_init_ns(struct ipc_namespace *ns);
-void msg_init_ns(struct ipc_namespace *ns);
-void shm_init_ns(struct ipc_namespace *ns);
+int sem_init_ns(struct ipc_namespace *ns);
+int msg_init_ns(struct ipc_namespace *ns);
+int shm_init_ns(struct ipc_namespace *ns);
void sem_exit_ns(struct ipc_namespace *ns);
void msg_exit_ns(struct ipc_namespace *ns);
void shm_exit_ns(struct ipc_namespace *ns);
#else
-static inline void sem_init_ns(struct ipc_namespace *ns) { }
-static inline void msg_init_ns(struct ipc_namespace *ns) { }
-static inline void shm_init_ns(struct ipc_namespace *ns) { }
+static inline int sem_init_ns(struct ipc_namespace *ns) { return 0; }
+static inline int msg_init_ns(struct ipc_namespace *ns) { return 0; }
+static inline int shm_init_ns(struct ipc_namespace *ns) { return 0; }
static inline void sem_exit_ns(struct ipc_namespace *ns) { }
static inline void msg_exit_ns(struct ipc_namespace *ns) { }
struct seq_file;
struct ipc_ids;
-void ipc_init_ids(struct ipc_ids *);
+int ipc_init_ids(struct ipc_ids *);
#ifdef CONFIG_PROC_FS
void __init ipc_init_proc_interface(const char *path, const char *header,
int ids, int (*show)(struct seq_file *, void *));
/* must be called with both locks acquired. */
void ipc_rmid(struct ipc_ids *, struct kern_ipc_perm *);
+/* must be called with both locks acquired. */
+void ipc_set_key_private(struct ipc_ids *, struct kern_ipc_perm *);
+
/* must be called with ipcp locked */
int ipcperms(struct ipc_namespace *ns, struct kern_ipc_perm *ipcp, short flg);