From 0cfb6aee70bddbef6ec796b255f588ce0e126766 Mon Sep 17 00:00:00 2001 From: Guillaume Knispel Date: Fri, 8 Sep 2017 16:17:55 -0700 Subject: ipc: optimize semget/shmget/msgget for lots of keys 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 Reviewed-by: Marc Pardo Cc: Davidlohr Bueso Cc: Kees Cook Cc: Manfred Spraul Cc: Alexey Dobriyan Cc: "Eric W. Biederman" Cc: "Peter Zijlstra (Intel)" Cc: Ingo Molnar Cc: Sebastian Andrzej Siewior Cc: Serge Hallyn Cc: Andrey Vagin Cc: Guillaume Knispel Cc: Marc Pardo Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- ipc/util.h | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'ipc/util.h') diff --git a/ipc/util.h b/ipc/util.h index c692010e6f0a..80c9f51c3f07 100644 --- a/ipc/util.h +++ b/ipc/util.h @@ -15,8 +15,8 @@ #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; @@ -30,17 +30,17 @@ static inline void mq_put_mnt(struct ipc_namespace *ns) { } #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) { } @@ -79,7 +79,7 @@ struct ipc_ops { 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 *)); @@ -104,6 +104,9 @@ int ipc_get_maxid(struct ipc_ids *); /* 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); -- cgit v1.2.3-55-g7522