From: Davidlohr Bueso <dave@stgolabs.net>
To: akpm@linux-foundation.org
Cc: manfred@colorfullife.com, dave@stgolabs.net,
linux-kernel@vger.kernel.org, Davidlohr Bueso <dbueso@suse.de>
Subject: [PATCH 3/4] sysv,ipc: cacheline align kern_ipc_perm
Date: Thu, 9 Feb 2017 12:53:01 -0800 [thread overview]
Message-ID: <1486673582-6979-4-git-send-email-dave@stgolabs.net> (raw)
In-Reply-To: <1486673582-6979-1-git-send-email-dave@stgolabs.net>
Assign 'struct kern_ipc_perm' its own cacheline to avoid false
sharing with sysv ipc calls. While the structure itself is rather
read-mostly throughout the lifespan of ipc, the spinlock causes
most of the invalidations. One example is 31a7c4746e9 (ipc/sem.c:
cacheline align the ipc spinlock for semaphores). Therefore,
extend this to all ipc.
The effect of cacheline alignment on sems can be seen in sembench,
which deals mostly with semtimedop wait/wakes is seen to improve raw
throughput (worker loops) between 8 to 12% on a 24-core x86 with
over 4 threads.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
include/linux/ipc.h | 7 +++----
include/linux/sem.h | 3 +--
2 files changed, 4 insertions(+), 6 deletions(-)
diff --git a/include/linux/ipc.h b/include/linux/ipc.h
index 9d84942ae2e5..71fd92d81b26 100644
--- a/include/linux/ipc.h
+++ b/include/linux/ipc.h
@@ -8,8 +8,7 @@
#define IPCMNI 32768 /* <= MAX_INT limit for ipc arrays (including sysctl changes) */
/* used by in-kernel data structures */
-struct kern_ipc_perm
-{
+struct kern_ipc_perm {
spinlock_t lock;
bool deleted;
int id;
@@ -18,9 +17,9 @@ struct kern_ipc_perm
kgid_t gid;
kuid_t cuid;
kgid_t cgid;
- umode_t mode;
+ umode_t mode;
unsigned long seq;
void *security;
-};
+} ____cacheline_aligned_in_smp;
#endif /* _LINUX_IPC_H */
diff --git a/include/linux/sem.h b/include/linux/sem.h
index 4fc222f8755d..9edec926e9d9 100644
--- a/include/linux/sem.h
+++ b/include/linux/sem.h
@@ -10,8 +10,7 @@ struct task_struct;
/* One sem_array data structure for each set of semaphores in the system. */
struct sem_array {
- struct kern_ipc_perm ____cacheline_aligned_in_smp
- sem_perm; /* permissions .. see ipc.h */
+ struct kern_ipc_perm sem_perm; /* permissions .. see ipc.h */
time_t sem_ctime; /* last change time */
struct sem *sem_base; /* ptr to first semaphore in array */
struct list_head pending_alter; /* pending operations */
--
2.6.6
next prev parent reply other threads:[~2017-02-09 20:53 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-02-09 20:52 [PATCH -tip 0/4] ipc/shm: moar updates for v4.11 Davidlohr Bueso
2017-02-09 20:52 ` [PATCH 1/4] ipc/shm: do not check for MAP_POPULATE Davidlohr Bueso
2017-02-10 2:15 ` Hugh Dickins
2017-02-13 18:14 ` Davidlohr Bueso
2017-02-09 20:53 ` [PATCH 2/4] ipc/shm: some shmat cleanups Davidlohr Bueso
2017-02-09 20:53 ` Davidlohr Bueso [this message]
2017-02-09 20:53 ` [PATCH 4/4] mm,hugetlb: compute page_size_log properly Davidlohr Bueso
2017-02-10 10:20 ` Michal Hocko
2017-02-10 16:51 ` Davidlohr Bueso
2017-02-20 16:11 ` Michal Hocko
2017-02-22 16:03 ` Davidlohr Bueso
2017-02-22 16:11 ` Michal Hocko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1486673582-6979-4-git-send-email-dave@stgolabs.net \
--to=dave@stgolabs.net \
--cc=akpm@linux-foundation.org \
--cc=dbueso@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=manfred@colorfullife.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).