linux/arch/x86/kernel/jump_label.c
Peter Zijlstra a89dfde3dc x86: Remove dynamic NOP selection
This ensures that a NOP is a NOP and not a random other instruction that
is also a NOP. It allows simplification of dynamic code patching that
wants to verify existing code before writing new instructions (ftrace,
jump_label, static_call, etc..).

Differentiating on NOPs is not a feature.

This pessimises 32bit (DONTCARE) and 32bit on 64bit CPUs (CARELESS).
32bit is not a performance target.

Everything x86_64 since AMD K10 (2007) and Intel IvyBridge (2012) is
fine with using NOPL (as opposed to prefix NOP). And per FEATURE_NOPL
being required for x86_64, all x86_64 CPUs can use NOPL. So stop
caring about NOPs, simplify things and get on with life.

[ The problem seems to be that some uarchs can only decode NOPL on a
single front-end port while others have severe decode penalties for
excessive prefixes. All modern uarchs can handle both, except Atom,
which has prefix penalties. ]

[ Also, much doubt you can actually measure any of this on normal
workloads. ]

After this, FEATURE_NOPL is unused except for required-features for
x86_64. FEATURE_K8 is only used for PTI.

 [ bp: Kernel build measurements showed ~0.3s slowdown on Sandybridge
   which is hardly a slowdown. Get rid of X86_FEATURE_K7, while at it. ]

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Alexei Starovoitov <alexei.starovoitov@gmail.com> # bpf
Acked-by: Linus Torvalds <torvalds@linuxfoundation.org>
Link: https://lkml.kernel.org/r/20210312115749.065275711@infradead.org
2021-03-15 16:24:59 +01:00

138 lines
3.4 KiB
C

// SPDX-License-Identifier: GPL-2.0
/*
* jump label x86 support
*
* Copyright (C) 2009 Jason Baron <jbaron@redhat.com>
*
*/
#include <linux/jump_label.h>
#include <linux/memory.h>
#include <linux/uaccess.h>
#include <linux/module.h>
#include <linux/list.h>
#include <linux/jhash.h>
#include <linux/cpu.h>
#include <asm/kprobes.h>
#include <asm/alternative.h>
#include <asm/text-patching.h>
static void bug_at(const void *ip, int line)
{
/*
* The location is not an op that we were expecting.
* Something went wrong. Crash the box, as something could be
* corrupting the kernel.
*/
pr_crit("jump_label: Fatal kernel bug, unexpected op at %pS [%p] (%5ph) %d\n", ip, ip, ip, line);
BUG();
}
static const void *
__jump_label_set_jump_code(struct jump_entry *entry, enum jump_label_type type)
{
const void *expect, *code;
const void *addr, *dest;
int line;
addr = (void *)jump_entry_code(entry);
dest = (void *)jump_entry_target(entry);
code = text_gen_insn(JMP32_INSN_OPCODE, addr, dest);
if (type == JUMP_LABEL_JMP) {
expect = x86_nops[5]; line = __LINE__;
} else {
expect = code; line = __LINE__;
}
if (memcmp(addr, expect, JUMP_LABEL_NOP_SIZE))
bug_at(addr, line);
if (type == JUMP_LABEL_NOP)
code = x86_nops[5];
return code;
}
static inline void __jump_label_transform(struct jump_entry *entry,
enum jump_label_type type,
int init)
{
const void *opcode = __jump_label_set_jump_code(entry, type);
/*
* As long as only a single processor is running and the code is still
* not marked as RO, text_poke_early() can be used; Checking that
* system_state is SYSTEM_BOOTING guarantees it. It will be set to
* SYSTEM_SCHEDULING before other cores are awaken and before the
* code is write-protected.
*
* At the time the change is being done, just ignore whether we
* are doing nop -> jump or jump -> nop transition, and assume
* always nop being the 'currently valid' instruction
*/
if (init || system_state == SYSTEM_BOOTING) {
text_poke_early((void *)jump_entry_code(entry), opcode,
JUMP_LABEL_NOP_SIZE);
return;
}
text_poke_bp((void *)jump_entry_code(entry), opcode, JUMP_LABEL_NOP_SIZE, NULL);
}
static void __ref jump_label_transform(struct jump_entry *entry,
enum jump_label_type type,
int init)
{
mutex_lock(&text_mutex);
__jump_label_transform(entry, type, init);
mutex_unlock(&text_mutex);
}
void arch_jump_label_transform(struct jump_entry *entry,
enum jump_label_type type)
{
jump_label_transform(entry, type, 0);
}
bool arch_jump_label_transform_queue(struct jump_entry *entry,
enum jump_label_type type)
{
const void *opcode;
if (system_state == SYSTEM_BOOTING) {
/*
* Fallback to the non-batching mode.
*/
arch_jump_label_transform(entry, type);
return true;
}
mutex_lock(&text_mutex);
opcode = __jump_label_set_jump_code(entry, type);
text_poke_queue((void *)jump_entry_code(entry),
opcode, JUMP_LABEL_NOP_SIZE, NULL);
mutex_unlock(&text_mutex);
return true;
}
void arch_jump_label_transform_apply(void)
{
mutex_lock(&text_mutex);
text_poke_finish();
mutex_unlock(&text_mutex);
}
static enum {
JL_STATE_START,
JL_STATE_NO_UPDATE,
JL_STATE_UPDATE,
} jlstate __initdata_or_module = JL_STATE_START;
__init_or_module void arch_jump_label_transform_static(struct jump_entry *entry,
enum jump_label_type type)
{
if (jlstate == JL_STATE_UPDATE)
jump_label_transform(entry, type, 1);
}