linux-brain/include/linux/jump_label.h
Daniel Bristot de Oliveira c2ba8a15f3 jump_label: Batch updates if arch supports it
If the architecture supports the batching of jump label updates, use it!

An easy way to see the benefits of this patch is switching the
schedstats on and off. For instance:

-------------------------- %< ----------------------------
  #!/bin/sh
  while [ true ]; do
      sysctl -w kernel.sched_schedstats=1
      sleep 2
      sysctl -w kernel.sched_schedstats=0
      sleep 2
  done
-------------------------- >% ----------------------------

while watching the IPI count:

-------------------------- %< ----------------------------
  # watch -n1 "cat /proc/interrupts | grep Function"
-------------------------- >% ----------------------------

With the current mode, it is possible to see +- 168 IPIs each 2 seconds,
while with this patch the number of IPIs goes to 3 each 2 seconds.

Regarding the performance impact of this patch set, I made two measurements:

    The time to update a key (the task that is causing the change)
    The time to run the int3 handler (the side effect on a thread that
                                      hits the code being changed)

The schedstats static key was chosen as the key to being switched on and off.
The reason being is that it is used in more than 56 places, in a hot path. The
change in the schedstats static key will be done with the following command:

while [ true ]; do
    sysctl -w kernel.sched_schedstats=1
    usleep 500000
    sysctl -w kernel.sched_schedstats=0
    usleep 500000
done

In this way, they key will be updated twice per second. To force the hit of the
int3 handler, the system will also run a kernel compilation with two jobs per
CPU. The test machine is a two nodes/24 CPUs box with an Intel Xeon processor
@2.27GHz.

Regarding the update part, on average, the regular kernel takes 57 ms to update
the schedstats key, while the kernel with the batch updates takes just 1.4 ms
on average. Although it seems to be too good to be true, it makes sense: the
schedstats key is used in 56 places, so it was expected that it would take
around 56 times to update the keys with the current implementation, as the
IPIs are the most expensive part of the update.

Regarding the int3 handler, the non-batch handler takes 45 ns on average, while
the batch version takes around 180 ns. At first glance, it seems to be a high
value. But it is not, considering that it is doing 56 updates, rather than one!
It is taking four times more, only. This gain is possible because the patch
uses a binary search in the vector: log2(56)=5.8. So, it was expected to have
an overhead within four times.

(voice of tv propaganda) But, that is not all! As the int3 handler keeps on for
a shorter period (because the update part is on for a shorter time), the number
of hits in the int3 handler decreased by 10%.

The question then is: Is it worth paying the price of "135 ns" more in the int3
handler?

Considering that, in this test case, we are saving the handling of 53 IPIs,
that takes more than these 135 ns, it seems to be a meager price to be paid.
Moreover, the test case was forcing the hit of the int3, in practice, it
does not take that often. While the IPI takes place on all CPUs, hitting
the int3 handler or not!

For instance, in an isolated CPU with a process running in user-space
(nohz_full use-case), the chances of hitting the int3 handler is barely zero,
while there is no way to avoid the IPIs. By bounding the IPIs, we are improving
a lot this scenario.

Signed-off-by: Daniel Bristot de Oliveira <bristot@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Chris von Recklinghausen <crecklin@redhat.com>
Cc: Clark Williams <williams@redhat.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Jason Baron <jbaron@akamai.com>
Cc: Jiri Kosina <jkosina@suse.cz>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Scott Wood <swood@redhat.com>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/acc891dbc2dbc9fd616dd680529a2337b1d1274c.1560325897.git.bristot@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2019-06-17 12:09:22 +02:00

506 lines
15 KiB
C

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _LINUX_JUMP_LABEL_H
#define _LINUX_JUMP_LABEL_H
/*
* Jump label support
*
* Copyright (C) 2009-2012 Jason Baron <jbaron@redhat.com>
* Copyright (C) 2011-2012 Red Hat, Inc., Peter Zijlstra
*
* DEPRECATED API:
*
* The use of 'struct static_key' directly, is now DEPRECATED. In addition
* static_key_{true,false}() is also DEPRECATED. IE DO NOT use the following:
*
* struct static_key false = STATIC_KEY_INIT_FALSE;
* struct static_key true = STATIC_KEY_INIT_TRUE;
* static_key_true()
* static_key_false()
*
* The updated API replacements are:
*
* DEFINE_STATIC_KEY_TRUE(key);
* DEFINE_STATIC_KEY_FALSE(key);
* DEFINE_STATIC_KEY_ARRAY_TRUE(keys, count);
* DEFINE_STATIC_KEY_ARRAY_FALSE(keys, count);
* static_branch_likely()
* static_branch_unlikely()
*
* Jump labels provide an interface to generate dynamic branches using
* self-modifying code. Assuming toolchain and architecture support, if we
* define a "key" that is initially false via "DEFINE_STATIC_KEY_FALSE(key)",
* an "if (static_branch_unlikely(&key))" statement is an unconditional branch
* (which defaults to false - and the true block is placed out of line).
* Similarly, we can define an initially true key via
* "DEFINE_STATIC_KEY_TRUE(key)", and use it in the same
* "if (static_branch_unlikely(&key))", in which case we will generate an
* unconditional branch to the out-of-line true branch. Keys that are
* initially true or false can be using in both static_branch_unlikely()
* and static_branch_likely() statements.
*
* At runtime we can change the branch target by setting the key
* to true via a call to static_branch_enable(), or false using
* static_branch_disable(). If the direction of the branch is switched by
* these calls then we run-time modify the branch target via a
* no-op -> jump or jump -> no-op conversion. For example, for an
* initially false key that is used in an "if (static_branch_unlikely(&key))"
* statement, setting the key to true requires us to patch in a jump
* to the out-of-line of true branch.
*
* In addition to static_branch_{enable,disable}, we can also reference count
* the key or branch direction via static_branch_{inc,dec}. Thus,
* static_branch_inc() can be thought of as a 'make more true' and
* static_branch_dec() as a 'make more false'.
*
* Since this relies on modifying code, the branch modifying functions
* must be considered absolute slow paths (machine wide synchronization etc.).
* OTOH, since the affected branches are unconditional, their runtime overhead
* will be absolutely minimal, esp. in the default (off) case where the total
* effect is a single NOP of appropriate size. The on case will patch in a jump
* to the out-of-line block.
*
* When the control is directly exposed to userspace, it is prudent to delay the
* decrement to avoid high frequency code modifications which can (and do)
* cause significant performance degradation. Struct static_key_deferred and
* static_key_slow_dec_deferred() provide for this.
*
* Lacking toolchain and or architecture support, static keys fall back to a
* simple conditional branch.
*
* Additional babbling in: Documentation/static-keys.txt
*/
#ifndef __ASSEMBLY__
#include <linux/types.h>
#include <linux/compiler.h>
extern bool static_key_initialized;
#define STATIC_KEY_CHECK_USE(key) WARN(!static_key_initialized, \
"%s(): static key '%pS' used before call to jump_label_init()", \
__func__, (key))
#ifdef CONFIG_JUMP_LABEL
struct static_key {
atomic_t enabled;
/*
* Note:
* To make anonymous unions work with old compilers, the static
* initialization of them requires brackets. This creates a dependency
* on the order of the struct with the initializers. If any fields
* are added, STATIC_KEY_INIT_TRUE and STATIC_KEY_INIT_FALSE may need
* to be modified.
*
* bit 0 => 1 if key is initially true
* 0 if initially false
* bit 1 => 1 if points to struct static_key_mod
* 0 if points to struct jump_entry
*/
union {
unsigned long type;
struct jump_entry *entries;
struct static_key_mod *next;
};
};
#else
struct static_key {
atomic_t enabled;
};
#endif /* CONFIG_JUMP_LABEL */
#endif /* __ASSEMBLY__ */
#ifdef CONFIG_JUMP_LABEL
#include <asm/jump_label.h>
#ifndef __ASSEMBLY__
#ifdef CONFIG_HAVE_ARCH_JUMP_LABEL_RELATIVE
struct jump_entry {
s32 code;
s32 target;
long key; // key may be far away from the core kernel under KASLR
};
static inline unsigned long jump_entry_code(const struct jump_entry *entry)
{
return (unsigned long)&entry->code + entry->code;
}
static inline unsigned long jump_entry_target(const struct jump_entry *entry)
{
return (unsigned long)&entry->target + entry->target;
}
static inline struct static_key *jump_entry_key(const struct jump_entry *entry)
{
long offset = entry->key & ~3L;
return (struct static_key *)((unsigned long)&entry->key + offset);
}
#else
static inline unsigned long jump_entry_code(const struct jump_entry *entry)
{
return entry->code;
}
static inline unsigned long jump_entry_target(const struct jump_entry *entry)
{
return entry->target;
}
static inline struct static_key *jump_entry_key(const struct jump_entry *entry)
{
return (struct static_key *)((unsigned long)entry->key & ~3UL);
}
#endif
static inline bool jump_entry_is_branch(const struct jump_entry *entry)
{
return (unsigned long)entry->key & 1UL;
}
static inline bool jump_entry_is_init(const struct jump_entry *entry)
{
return (unsigned long)entry->key & 2UL;
}
static inline void jump_entry_set_init(struct jump_entry *entry)
{
entry->key |= 2;
}
#endif
#endif
#ifndef __ASSEMBLY__
enum jump_label_type {
JUMP_LABEL_NOP = 0,
JUMP_LABEL_JMP,
};
struct module;
#ifdef CONFIG_JUMP_LABEL
#define JUMP_TYPE_FALSE 0UL
#define JUMP_TYPE_TRUE 1UL
#define JUMP_TYPE_LINKED 2UL
#define JUMP_TYPE_MASK 3UL
static __always_inline bool static_key_false(struct static_key *key)
{
return arch_static_branch(key, false);
}
static __always_inline bool static_key_true(struct static_key *key)
{
return !arch_static_branch(key, true);
}
extern struct jump_entry __start___jump_table[];
extern struct jump_entry __stop___jump_table[];
extern void jump_label_init(void);
extern void jump_label_lock(void);
extern void jump_label_unlock(void);
extern void arch_jump_label_transform(struct jump_entry *entry,
enum jump_label_type type);
extern void arch_jump_label_transform_static(struct jump_entry *entry,
enum jump_label_type type);
extern bool arch_jump_label_transform_queue(struct jump_entry *entry,
enum jump_label_type type);
extern void arch_jump_label_transform_apply(void);
extern int jump_label_text_reserved(void *start, void *end);
extern void static_key_slow_inc(struct static_key *key);
extern void static_key_slow_dec(struct static_key *key);
extern void static_key_slow_inc_cpuslocked(struct static_key *key);
extern void static_key_slow_dec_cpuslocked(struct static_key *key);
extern void jump_label_apply_nops(struct module *mod);
extern int static_key_count(struct static_key *key);
extern void static_key_enable(struct static_key *key);
extern void static_key_disable(struct static_key *key);
extern void static_key_enable_cpuslocked(struct static_key *key);
extern void static_key_disable_cpuslocked(struct static_key *key);
/*
* We should be using ATOMIC_INIT() for initializing .enabled, but
* the inclusion of atomic.h is problematic for inclusion of jump_label.h
* in 'low-level' headers. Thus, we are initializing .enabled with a
* raw value, but have added a BUILD_BUG_ON() to catch any issues in
* jump_label_init() see: kernel/jump_label.c.
*/
#define STATIC_KEY_INIT_TRUE \
{ .enabled = { 1 }, \
{ .entries = (void *)JUMP_TYPE_TRUE } }
#define STATIC_KEY_INIT_FALSE \
{ .enabled = { 0 }, \
{ .entries = (void *)JUMP_TYPE_FALSE } }
#else /* !CONFIG_JUMP_LABEL */
#include <linux/atomic.h>
#include <linux/bug.h>
static inline int static_key_count(struct static_key *key)
{
return atomic_read(&key->enabled);
}
static __always_inline void jump_label_init(void)
{
static_key_initialized = true;
}
static __always_inline bool static_key_false(struct static_key *key)
{
if (unlikely(static_key_count(key) > 0))
return true;
return false;
}
static __always_inline bool static_key_true(struct static_key *key)
{
if (likely(static_key_count(key) > 0))
return true;
return false;
}
static inline void static_key_slow_inc(struct static_key *key)
{
STATIC_KEY_CHECK_USE(key);
atomic_inc(&key->enabled);
}
static inline void static_key_slow_dec(struct static_key *key)
{
STATIC_KEY_CHECK_USE(key);
atomic_dec(&key->enabled);
}
#define static_key_slow_inc_cpuslocked(key) static_key_slow_inc(key)
#define static_key_slow_dec_cpuslocked(key) static_key_slow_dec(key)
static inline int jump_label_text_reserved(void *start, void *end)
{
return 0;
}
static inline void jump_label_lock(void) {}
static inline void jump_label_unlock(void) {}
static inline int jump_label_apply_nops(struct module *mod)
{
return 0;
}
static inline void static_key_enable(struct static_key *key)
{
STATIC_KEY_CHECK_USE(key);
if (atomic_read(&key->enabled) != 0) {
WARN_ON_ONCE(atomic_read(&key->enabled) != 1);
return;
}
atomic_set(&key->enabled, 1);
}
static inline void static_key_disable(struct static_key *key)
{
STATIC_KEY_CHECK_USE(key);
if (atomic_read(&key->enabled) != 1) {
WARN_ON_ONCE(atomic_read(&key->enabled) != 0);
return;
}
atomic_set(&key->enabled, 0);
}
#define static_key_enable_cpuslocked(k) static_key_enable((k))
#define static_key_disable_cpuslocked(k) static_key_disable((k))
#define STATIC_KEY_INIT_TRUE { .enabled = ATOMIC_INIT(1) }
#define STATIC_KEY_INIT_FALSE { .enabled = ATOMIC_INIT(0) }
#endif /* CONFIG_JUMP_LABEL */
#define STATIC_KEY_INIT STATIC_KEY_INIT_FALSE
#define jump_label_enabled static_key_enabled
/* -------------------------------------------------------------------------- */
/*
* Two type wrappers around static_key, such that we can use compile time
* type differentiation to emit the right code.
*
* All the below code is macros in order to play type games.
*/
struct static_key_true {
struct static_key key;
};
struct static_key_false {
struct static_key key;
};
#define STATIC_KEY_TRUE_INIT (struct static_key_true) { .key = STATIC_KEY_INIT_TRUE, }
#define STATIC_KEY_FALSE_INIT (struct static_key_false){ .key = STATIC_KEY_INIT_FALSE, }
#define DEFINE_STATIC_KEY_TRUE(name) \
struct static_key_true name = STATIC_KEY_TRUE_INIT
#define DEFINE_STATIC_KEY_TRUE_RO(name) \
struct static_key_true name __ro_after_init = STATIC_KEY_TRUE_INIT
#define DECLARE_STATIC_KEY_TRUE(name) \
extern struct static_key_true name
#define DEFINE_STATIC_KEY_FALSE(name) \
struct static_key_false name = STATIC_KEY_FALSE_INIT
#define DEFINE_STATIC_KEY_FALSE_RO(name) \
struct static_key_false name __ro_after_init = STATIC_KEY_FALSE_INIT
#define DECLARE_STATIC_KEY_FALSE(name) \
extern struct static_key_false name
#define DEFINE_STATIC_KEY_ARRAY_TRUE(name, count) \
struct static_key_true name[count] = { \
[0 ... (count) - 1] = STATIC_KEY_TRUE_INIT, \
}
#define DEFINE_STATIC_KEY_ARRAY_FALSE(name, count) \
struct static_key_false name[count] = { \
[0 ... (count) - 1] = STATIC_KEY_FALSE_INIT, \
}
extern bool ____wrong_branch_error(void);
#define static_key_enabled(x) \
({ \
if (!__builtin_types_compatible_p(typeof(*x), struct static_key) && \
!__builtin_types_compatible_p(typeof(*x), struct static_key_true) &&\
!__builtin_types_compatible_p(typeof(*x), struct static_key_false)) \
____wrong_branch_error(); \
static_key_count((struct static_key *)x) > 0; \
})
#ifdef CONFIG_JUMP_LABEL
/*
* Combine the right initial value (type) with the right branch order
* to generate the desired result.
*
*
* type\branch| likely (1) | unlikely (0)
* -----------+-----------------------+------------------
* | |
* true (1) | ... | ...
* | NOP | JMP L
* | <br-stmts> | 1: ...
* | L: ... |
* | |
* | | L: <br-stmts>
* | | jmp 1b
* | |
* -----------+-----------------------+------------------
* | |
* false (0) | ... | ...
* | JMP L | NOP
* | <br-stmts> | 1: ...
* | L: ... |
* | |
* | | L: <br-stmts>
* | | jmp 1b
* | |
* -----------+-----------------------+------------------
*
* The initial value is encoded in the LSB of static_key::entries,
* type: 0 = false, 1 = true.
*
* The branch type is encoded in the LSB of jump_entry::key,
* branch: 0 = unlikely, 1 = likely.
*
* This gives the following logic table:
*
* enabled type branch instuction
* -----------------------------+-----------
* 0 0 0 | NOP
* 0 0 1 | JMP
* 0 1 0 | NOP
* 0 1 1 | JMP
*
* 1 0 0 | JMP
* 1 0 1 | NOP
* 1 1 0 | JMP
* 1 1 1 | NOP
*
* Which gives the following functions:
*
* dynamic: instruction = enabled ^ branch
* static: instruction = type ^ branch
*
* See jump_label_type() / jump_label_init_type().
*/
#define static_branch_likely(x) \
({ \
bool branch; \
if (__builtin_types_compatible_p(typeof(*x), struct static_key_true)) \
branch = !arch_static_branch(&(x)->key, true); \
else if (__builtin_types_compatible_p(typeof(*x), struct static_key_false)) \
branch = !arch_static_branch_jump(&(x)->key, true); \
else \
branch = ____wrong_branch_error(); \
likely(branch); \
})
#define static_branch_unlikely(x) \
({ \
bool branch; \
if (__builtin_types_compatible_p(typeof(*x), struct static_key_true)) \
branch = arch_static_branch_jump(&(x)->key, false); \
else if (__builtin_types_compatible_p(typeof(*x), struct static_key_false)) \
branch = arch_static_branch(&(x)->key, false); \
else \
branch = ____wrong_branch_error(); \
unlikely(branch); \
})
#else /* !CONFIG_JUMP_LABEL */
#define static_branch_likely(x) likely(static_key_enabled(&(x)->key))
#define static_branch_unlikely(x) unlikely(static_key_enabled(&(x)->key))
#endif /* CONFIG_JUMP_LABEL */
/*
* Advanced usage; refcount, branch is enabled when: count != 0
*/
#define static_branch_inc(x) static_key_slow_inc(&(x)->key)
#define static_branch_dec(x) static_key_slow_dec(&(x)->key)
#define static_branch_inc_cpuslocked(x) static_key_slow_inc_cpuslocked(&(x)->key)
#define static_branch_dec_cpuslocked(x) static_key_slow_dec_cpuslocked(&(x)->key)
/*
* Normal usage; boolean enable/disable.
*/
#define static_branch_enable(x) static_key_enable(&(x)->key)
#define static_branch_disable(x) static_key_disable(&(x)->key)
#define static_branch_enable_cpuslocked(x) static_key_enable_cpuslocked(&(x)->key)
#define static_branch_disable_cpuslocked(x) static_key_disable_cpuslocked(&(x)->key)
#endif /* __ASSEMBLY__ */
#endif /* _LINUX_JUMP_LABEL_H */