From: Rasmus Villemoes <linux@rasmusvillemoes.dk>
To: Andrew Morton <akpm@linux-foundation.org>,
Richard Weinberger <richard@nod.at>,
"Peter Zijlstra (Intel)" <peterz@infradead.org>,
Andy Lutomirski <luto@amacapital.net>,
Darren Hart <darren@dvhart.com>, Oleg Nesterov <oleg@redhat.com>,
Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>,
linux-kernel@vger.kernel.org
Subject: [not-a-PATCH 2/2] futex: eliminate instruction from hash_futex
Date: Mon, 22 Jun 2015 15:16:19 +0200 [thread overview]
Message-ID: <1434978981-12831-2-git-send-email-linux@rasmusvillemoes.dk> (raw)
In-Reply-To: <1434978981-12831-1-git-send-email-linux@rasmusvillemoes.dk>
After initialization, futex_hashsize is only used in hash_futex, where
we subtract 1 to get the proper mask. So store that mask instead.
Or not: Unfortunately, this causes gcc to emit an 'and with memory op'
instead of doing a load which can happen at least somewhat in parallel
with the last few instructions computing hash, as this diff shows:
-: 48 8b 15 a5 61 e4 00 mov 0xe461a5(%rip),%rdx # ffffffff81eff9f8 <__futex_data+0x8>
: 31 c8 xor %ecx,%eax
: c1 c9 08 ror $0x8,%ecx
: 29 c8 sub %ecx,%eax
-: 48 83 ea 01 sub $0x1,%rdx
-: 48 21 d0 and %rdx,%rax
+: 23 05 9f 61 e4 00 and 0xe4619f(%rip),%eax # ffffffff81eff9f8 <__futex_data+0x8>
: 48 c1 e0 06 shl $0x6,%rax
-: 48 03 05 84 61 e4 00 add 0xe46184(%rip),%rax # ffffffff81eff9f0 <__futex_data>
+: 48 03 05 8c 61 e4 00 add 0xe4618c(%rip),%rax # ffffffff81eff9f0 <__futex_data>
: c3 retq
Not-signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
---
kernel/futex.c | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/kernel/futex.c b/kernel/futex.c
index c5f33bf78293..22865bcbcae8 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -255,16 +255,16 @@ struct futex_hash_bucket {
} ____cacheline_aligned_in_smp;
/*
- * The base of the bucket array and its size are always used together
+ * The base of the bucket array and the mask are always used together
* (after initialization only in hash_futex()), so ensure that they
* reside in the same cacheline.
*/
static struct {
struct futex_hash_bucket *queues;
- unsigned long hashsize;
+ u32 mask;
} __futex_data __read_mostly __aligned(16);
#define futex_queues (__futex_data.queues)
-#define futex_hashsize (__futex_data.hashsize)
+#define futex_hashmask (__futex_data.mask)
static inline void futex_get_mm(union futex_key *key)
@@ -320,7 +320,7 @@ static struct futex_hash_bucket *hash_futex(union futex_key *key)
u32 hash = jhash2((u32*)&key->both.word,
(sizeof(key->both.word)+sizeof(key->both.ptr))/4,
key->both.offset);
- return &futex_queues[hash & (futex_hashsize - 1)];
+ return &futex_queues[hash & futex_hashmask];
}
/*
@@ -3035,6 +3035,7 @@ static int __init futex_init(void)
{
unsigned int futex_shift;
unsigned long i;
+ unsigned long futex_hashsize;
#if CONFIG_BASE_SMALL
futex_hashsize = 16;
@@ -3045,7 +3046,7 @@ static int __init futex_init(void)
futex_queues = alloc_large_system_hash("futex", sizeof(*futex_queues),
futex_hashsize, 0,
futex_hashsize < 256 ? HASH_SMALL : 0,
- &futex_shift, NULL,
+ &futex_shift, &futex_hashmask,
futex_hashsize, futex_hashsize);
futex_hashsize = 1UL << futex_shift;
--
2.1.3
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
Please read the FAQ at http://www.tux.org/lkml/
next prev parent reply other threads:[~2015-06-22 13:17 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-06-22 13:16 [PATCH 1/2] futex: eliminate cache miss from futex_hash() Rasmus Villemoes
2015-06-22 13:16 ` Rasmus Villemoes [this message]
2015-06-22 13:33 ` Peter Zijlstra
2015-06-22 13:45 ` Rasmus Villemoes
2015-06-22 13:55 ` Peter Zijlstra
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=1434978981-12831-2-git-send-email-linux@rasmusvillemoes.dk \
--to=linux@rasmusvillemoes.dk \
--cc=akpm@linux-foundation.org \
--cc=darren@dvhart.com \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=mtk.manpages@gmail.com \
--cc=oleg@redhat.com \
--cc=peterz@infradead.org \
--cc=richard@nod.at \
/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
all inboxes | Powered by JetHome®