mirror of
https://github.com/torvalds/linux.git
synced 2024-12-03 17:41:22 +00:00
450cbdd012
MFENCE appears to be way slower than a locked instruction - let's use
LOCK ADD unconditionally, as we always did on old 32-bit.
Performance testing results:
perf stat -r 10 -- ./virtio_ring_0_9 --sleep --host-affinity 0 --guest-affinity 0
Before:
0.922565990 seconds time elapsed ( +- 1.15% )
After:
0.578667024 seconds time elapsed ( +- 1.21% )
i.e. about ~60% faster.
Just poking at SP would be the most natural, but if we then read the
value from SP, we get a false dependency which will slow us down.
This was noted in this article:
http://shipilev.net/blog/2014/on-the-fence-with-dependencies/
And is easy to reproduce by sticking a barrier in a small non-inline
function.
So let's use a negative offset - which avoids this problem since we
build with the red zone disabled.
For userspace, use an address just below the redzone.
The one difference between LOCK ADD and MFENCE is that LOCK ADD does
not affect CLFLUSH, previous patches converted all uses of CLFLUSH to
call mb(), such that changes to smp_mb() won't affect it.
Update mb/rmb/wmb() on 32-bit to use the negative offset, too, for
consistency.
As a follow-up, it might be worth considering switching users
of CLFLUSH to another API (e.g. clflush_mb()?) - we will
then be able to convert mb() to smp_mb() again.
Also arguably, GCC should switch to use LOCK ADD for __sync_synchronize().
This might be worth pursuing separately.
Suggested-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: qemu-devel@nongnu.org
Cc: virtualization@lists.linux-foundation.org
Link: http://lkml.kernel.org/r/1509118355-4890-1-git-send-email-mst@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
138 lines
2.7 KiB
C
138 lines
2.7 KiB
C
/*
|
|
* Copyright (C) 2016 Red Hat, Inc.
|
|
* Author: Michael S. Tsirkin <mst@redhat.com>
|
|
* This work is licensed under the terms of the GNU GPL, version 2.
|
|
*
|
|
* Common macros and functions for ring benchmarking.
|
|
*/
|
|
#ifndef MAIN_H
|
|
#define MAIN_H
|
|
|
|
#include <stdbool.h>
|
|
|
|
extern int param;
|
|
|
|
extern bool do_exit;
|
|
|
|
#if defined(__x86_64__) || defined(__i386__)
|
|
#include "x86intrin.h"
|
|
|
|
static inline void wait_cycles(unsigned long long cycles)
|
|
{
|
|
unsigned long long t;
|
|
|
|
t = __rdtsc();
|
|
while (__rdtsc() - t < cycles) {}
|
|
}
|
|
|
|
#define VMEXIT_CYCLES 500
|
|
#define VMENTRY_CYCLES 500
|
|
|
|
#elif defined(__s390x__)
|
|
static inline void wait_cycles(unsigned long long cycles)
|
|
{
|
|
asm volatile("0: brctg %0,0b" : : "d" (cycles));
|
|
}
|
|
|
|
/* tweak me */
|
|
#define VMEXIT_CYCLES 200
|
|
#define VMENTRY_CYCLES 200
|
|
|
|
#else
|
|
static inline void wait_cycles(unsigned long long cycles)
|
|
{
|
|
_Exit(5);
|
|
}
|
|
#define VMEXIT_CYCLES 0
|
|
#define VMENTRY_CYCLES 0
|
|
#endif
|
|
|
|
static inline void vmexit(void)
|
|
{
|
|
if (!do_exit)
|
|
return;
|
|
|
|
wait_cycles(VMEXIT_CYCLES);
|
|
}
|
|
static inline void vmentry(void)
|
|
{
|
|
if (!do_exit)
|
|
return;
|
|
|
|
wait_cycles(VMENTRY_CYCLES);
|
|
}
|
|
|
|
/* implemented by ring */
|
|
void alloc_ring(void);
|
|
/* guest side */
|
|
int add_inbuf(unsigned, void *, void *);
|
|
void *get_buf(unsigned *, void **);
|
|
void disable_call();
|
|
bool used_empty();
|
|
bool enable_call();
|
|
void kick_available();
|
|
/* host side */
|
|
void disable_kick();
|
|
bool avail_empty();
|
|
bool enable_kick();
|
|
bool use_buf(unsigned *, void **);
|
|
void call_used();
|
|
|
|
/* implemented by main */
|
|
extern bool do_sleep;
|
|
void kick(void);
|
|
void wait_for_kick(void);
|
|
void call(void);
|
|
void wait_for_call(void);
|
|
|
|
extern unsigned ring_size;
|
|
|
|
/* Compiler barrier - similar to what Linux uses */
|
|
#define barrier() asm volatile("" ::: "memory")
|
|
|
|
/* Is there a portable way to do this? */
|
|
#if defined(__x86_64__) || defined(__i386__)
|
|
#define cpu_relax() asm ("rep; nop" ::: "memory")
|
|
#elif defined(__s390x__)
|
|
#define cpu_relax() barrier()
|
|
#else
|
|
#define cpu_relax() assert(0)
|
|
#endif
|
|
|
|
extern bool do_relax;
|
|
|
|
static inline void busy_wait(void)
|
|
{
|
|
if (do_relax)
|
|
cpu_relax();
|
|
else
|
|
/* prevent compiler from removing busy loops */
|
|
barrier();
|
|
}
|
|
|
|
#if defined(__x86_64__) || defined(__i386__)
|
|
#define smp_mb() asm volatile("lock; addl $0,-128(%%rsp)" ::: "memory", "cc")
|
|
#else
|
|
/*
|
|
* Not using __ATOMIC_SEQ_CST since gcc docs say they are only synchronized
|
|
* with other __ATOMIC_SEQ_CST calls.
|
|
*/
|
|
#define smp_mb() __sync_synchronize()
|
|
#endif
|
|
|
|
/*
|
|
* This abuses the atomic builtins for thread fences, and
|
|
* adds a compiler barrier.
|
|
*/
|
|
#define smp_release() do { \
|
|
barrier(); \
|
|
__atomic_thread_fence(__ATOMIC_RELEASE); \
|
|
} while (0)
|
|
|
|
#define smp_acquire() do { \
|
|
__atomic_thread_fence(__ATOMIC_ACQUIRE); \
|
|
barrier(); \
|
|
} while (0)
|
|
|
|
#endif
|