Hi,
Here are the patchset of the kprobes jump optimization v8
(a.k.a. Djprobe). This version is just moving onto
2.6.33-rc4-tip. Ingo, I assume its a good timing to
push this code onto -tip tree (maybe developing branch?),
since people can test it with perf-probe.
I've decided to make a separated series of patches of
jump optimization with text_poke_smp() which is
'officially' supported on Intel's processors.
So, this version of patches are just updated against
the latest tip/master, no other updates are included.
I know that int3-bypassing method (text_poke_fixup())
is currently unofficially believed as safe. But we
need to get more official answers from x86 vendors.
Moreover, we need to tweak entry_*.S for preventing
recursive NMI, because int3 inside NMI handler will
unblock NMI blocking. I'd like to push it after this
series of patches are merged.
Anyway, thanks Mathieu and Peter, for helping me to
implement it and organizing discussion points about
int3-bypass XMC!
These patches can be applied on the latest -tip.
Changes in v8:
- Update patches against the latest tip/master.
- Drop text_poke_fixup() related patches.
- Update benchmark results and add jprobes and kprobe(post-handler)
results.
And kprobe stress test didn't found any regressions - from kprobes,
under kvm/x86.
TODO:
- Support NMI-safe int3-bypassing text_poke.
- Support preemptive kernel (by stack unwinding and checking address).
Jump Optimized Kprobes
======================
o Concept
Kprobes uses the int3 breakpoint instruction on x86 for instrumenting
probes into running kernel. Jump optimization allows kprobes to replace
breakpoint with a jump instruction for reducing probing overhead drastically.
o Performance
An optimized kprobe 5 times faster than a kprobe.
Optimizing probes gains its performance. Usually, a kprobe hit takes
0.5 to 1.0 microseconds to process. On the other hand, a jump optimized
probe hit takes less than 0.1 microseconds (actual number depends on the
processor). Here is a sample overheads.
Intel(R) Xeon(R) CPU E5410 @ 2.33GHz
(without debugging options, with text_poke_smp patch, 2.6.33-rc4-tip+)
x86-32 x86-64
kprobe: 0.80us 0.99us
kprobe+booster: 0.33us 0.43us
kprobe+optimized: 0.05us 0.06us
kprobe(post-handler): 0.81us 1.00us
kretprobe : 1.10us 1.24us
kretprobe+booster: 0.61us 0.68us
kretprobe+optimized: 0.33us 0.30us
jprobe: 1.37us 1.67us
jprobe+booster: 0.80us 1.10us
(booster skips single-stepping, kprobe with post handler
isn't boosted/optimized, and jprobe isn't optimized.)
Note that jump optimization also consumes more memory, but not so much.
It just uses ~200 bytes, so, even if you use ~10,000 probes, it just
consumes a few MB.
o Usage
Set CONFIG_OPTPROBES=y when building a kernel, then all *probes will be
optimized if possible.
Kprobes decodes probed function and checks whether the target instructions
can be optimized(replaced with a jump) safely. If it can't be, Kprobes just
doesn't optimize it.
o Optimization
Before preparing optimization, Kprobes inserts original(user-defined)
kprobe on the specified address. So, even if the kprobe is not
possible to be optimized, it just uses a normal kprobe.
- Safety check
First, Kprobes gets the address of probed function and checks whether the
optimized region, which will be replaced by a jump instruction, does NOT
straddle the function boundary, because if the optimized region reaches the
next function, its caller causes unexpected results.
Next, Kprobes decodes whole body of probed function and checks there is
NO indirect jump, NO instruction which will cause exception by checking
exception_tables (this will jump to fixup code and fixup code jumps into
same function body) and NO near jump which jumps into the optimized region
(except the 1st byte of jump), because if some jump instruction jumps
into the middle of another instruction, it causes unexpected results too.
Kprobes also measures the length of instructions which will be replaced
by a jump instruction, because a jump instruction is longer than 1 byte,
it may replaces multiple instructions, and it checks whether those
instructions can be executed out-of-line.
- Preparing detour code
Then, Kprobes prepares "detour" buffer, which contains exception emulating
code (push/pop registers, call handler), copied instructions(Kprobes copies
instructions which will be replaced by a jump, to the detour buffer), and
a jump which jumps back to the original execution path.
- Pre-optimization
After preparing detour code, Kprobes enqueues the kprobe to optimizing list
and kicks kprobe-optimizer workqueue to optimize it. To wait other optimized
probes, kprobe-optimizer will delay to work.
When the optimized-kprobe is hit before optimization, its handler
changes IP(instruction pointer) to copied code and exits. So, those
copied instructions are executed on the detour buffer.
- Optimization
Kprobe-optimizer doesn't start instruction-replacing soon, it waits
synchronize_sched for safety, because some processors are possible to be
interrupted on the middle of instruction series (2nd or Nth instruction)
which will be replaced by a jump instruction(*).
As you know, synchronize_sched() can ensure that all interruptions which
were executed when synchronize_sched() was called are done, only if
CONFIG_PREEMPT=n. So, this version supports only the kernel with
CONFIG_PREEMPT=n.(**)
After that, kprobe-optimizer calls stop_machine() to replace probed-
instructions with a jump instruction by using text_poke_smp().
- Unoptimization
When unregistering, disabling kprobe or being blocked by other kprobe,
an optimized-kprobe will be unoptimized. Before kprobe-optimizer runs,
the kprobe just be dequeued from the optimized list. When the optimization
has been done, it replaces a jump with int3 breakpoint and original code
by using text_poke_smp().
(*)Please imagine that 2nd instruction is interrupted and
optimizer replaces the 2nd instruction with jump *address*
while the interrupt handler is running. When the interrupt
returns to original address, there is no valid instructions
and it causes unexpected result.
(**)This optimization-safety checking may be replaced with stop-machine
method which ksplice is done for supporting CONFIG_PREEMPT=y kernel.
Thank you,
---
Masami Hiramatsu (9):
kprobes: Add documents of jump optimization
kprobes/x86: Support kprobes jump optimization on x86
x86: Add text_poke_smp for SMP cross modifying code
kprobes/x86: Cleanup save/restore registers
kprobes/x86: Boost probes when reentering
kprobes: Jump optimization sysctl interface
kprobes: Introduce kprobes jump optimization
kprobes: Introduce generic insn_slot framework
kprobes/x86: Cleanup RELATIVEJUMP_INSTRUCTION to RELATIVEJUMP_OPCODE
Documentation/kprobes.txt | 191 ++++++++++-
arch/Kconfig | 13 +
arch/x86/Kconfig | 1
arch/x86/include/asm/alternative.h | 4
arch/x86/include/asm/kprobes.h | 31 ++
arch/x86/kernel/alternative.c | 60 +++
arch/x86/kernel/kprobes.c | 596 ++++++++++++++++++++++++++++------
include/linux/kprobes.h | 44 +++
kernel/kprobes.c | 626 +++++++++++++++++++++++++++++++-----
kernel/sysctl.c | 12 +
10 files changed, 1373 insertions(+), 205 deletions(-)
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Change RELATIVEJUMP_INSTRUCTION macro to RELATIVEJUMP_OPCODE since it
represents just the opcode byte.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
arch/x86/include/asm/kprobes.h | 2 +-
arch/x86/kernel/kprobes.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/x86/include/asm/kprobes.h b/arch/x86/include/asm/kprobes.h
index 4fe681d..eaec8ea 100644
--- a/arch/x86/include/asm/kprobes.h
+++ b/arch/x86/include/asm/kprobes.h
@@ -32,7 +32,7 @@ struct kprobe;
typedef u8 kprobe_opcode_t;
#define BREAKPOINT_INSTRUCTION 0xcc
-#define RELATIVEJUMP_INSTRUCTION 0xe9
+#define RELATIVEJUMP_OPCODE 0xe9
#define MAX_INSN_SIZE 16
#define MAX_STACK_SIZE 64
#define MIN_STACK_SIZE(ADDR) \
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index 5b8c750..7039e6e 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -115,7 +115,7 @@ static void __kprobes set_jmp_op(void *from, void *to)
} __attribute__((packed)) * jop;
jop = (struct __arch_jmp_op *)from;
jop->raddr = (s32)((long)(to) - ((long)(from) + 5));
- jop->op = RELATIVEJUMP_INSTRUCTION;
+ jop->op = RELATIVEJUMP_OPCODE;
}
/*
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Introduce kprobes jump optimization arch-independent parts.
Kprobes uses breakpoint instruction for interrupting execution flow, on
some architectures, it can be replaced by a jump instruction and
interruption emulation code. This gains kprobs' performance drastically.
To enable this feature, set CONFIG_OPTPROBES=y (default y if the arch
supports OPTPROBE).
Changes in v6:
- Cleanup coding style for readability.
- Add comments around get/put_online_cpus().
Changes in v5:
- Use get_online_cpus()/put_online_cpus() for avoiding text_mutex
deadlock.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
arch/Kconfig | 13 +
include/linux/kprobes.h | 36 ++++
kernel/kprobes.c | 439 ++++++++++++++++++++++++++++++++++++++++++-----
3 files changed, 442 insertions(+), 46 deletions(-)
diff --git a/arch/Kconfig b/arch/Kconfig
index 32e8b68..2622a9d 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -57,6 +57,17 @@ config KPROBES
for kernel debugging, non-intrusive instrumentation and testing.
If in doubt, say "N".
+config OPTPROBES
+ bool "Kprobes jump optimization support (EXPERIMENTAL)"
+ default y
+ depends on KPROBES
+ depends on !PREEMPT
+ depends on HAVE_OPTPROBES
+ select KALLSYMS_ALL
+ help
+ This option will allow kprobes to optimize breakpoint to
+ a jump for reducing its overhead.
+
config HAVE_EFFICIENT_UNALIGNED_ACCESS
bool
help
@@ -99,6 +110,8 @@ config HAVE_KPROBES
config HAVE_KRETPROBES
bool
+config HAVE_OPTPROBES
+ bool
#
# An arch should select this if it provides all these things:
#
diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
index 1b672f7..aed1f95 100644
--- a/include/linux/kprobes.h
+++ b/include/linux/kprobes.h
@@ -122,6 +122,11 @@ struct kprobe {
/* Kprobe status flags */
#define KPROBE_FLAG_GONE 1 /* breakpoint has already gone */
#define KPROBE_FLAG_DISABLED 2 /* probe is temporarily disabled */
+#define KPROBE_FLAG_OPTIMIZED 4 /*
+ * probe is really optimized.
+ * NOTE:
+ * this flag is only for optimized_kprobe.
+ */
/* Has this kprobe gone ? */
static inline int kprobe_gone(struct kprobe *p)
@@ -134,6 +139,12 @@ static inline int kprobe_disabled(struct kprobe *p)
{
return p->flags & (KPROBE_FLAG_DISABLED | KPROBE_FLAG_GONE);
}
+
+/* Is this kprobe really running optimized path ? */
+static inline int kprobe_optimized(struct kprobe *p)
+{
+ return p->flags & KPROBE_FLAG_OPTIMIZED;
+}
/*
* Special probe type that uses setjmp-longjmp type tricks to resume
* execution at a specified entry with a matching prototype corresponding
@@ -249,6 +260,31 @@ extern kprobe_opcode_t *get_insn_slot(void);
extern void free_insn_slot(kprobe_opcode_t *slot, int dirty);
extern void kprobes_inc_nmissed_count(struct kprobe *p);
+#ifdef CONFIG_OPTPROBES
+/*
+ * Internal structure for direct jump optimized probe
+ */
+struct optimized_kprobe {
+ struct kprobe kp;
+ struct list_head list; /* list for optimizing queue */
+ struct arch_optimized_insn optinsn;
+};
+
+/* Architecture dependent functions for direct jump optimization */
+extern int arch_prepared_optinsn(struct arch_optimized_insn *optinsn);
+extern int arch_check_optimized_kprobe(struct optimized_kprobe *op);
+extern int arch_prepare_optimized_kprobe(struct optimized_kprobe *op);
+extern void arch_remove_optimized_kprobe(struct optimized_kprobe *op);
+extern int arch_optimize_kprobe(struct optimized_kprobe *op);
+extern void arch_unoptimize_kprobe(struct optimized_kprobe *op);
+extern kprobe_opcode_t *get_optinsn_slot(void);
+extern void free_optinsn_slot(kprobe_opcode_t *slot, int dirty);
+extern int arch_within_optimized_kprobe(struct optimized_kprobe *op,
+ unsigned long addr);
+
+extern void opt_pre_handler(struct kprobe *p, struct pt_regs *regs);
+#endif /* CONFIG_OPTPROBES */
+
/* Get the kprobe at this addr (if any) - called with preemption disabled */
struct kprobe *get_kprobe(void *addr);
void kretprobe_hash_lock(struct task_struct *tsk,
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 57a2a23..759bb8e 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -44,6 +44,7 @@
#include <linux/debugfs.h>
#include <linux/kdebug.h>
#include <linux/memory.h>
+#include <linux/cpu.h>
#include <asm-generic/sections.h>
#include <asm/cacheflush.h>
@@ -303,6 +304,33 @@ void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty)
__free_insn_slot(&kprobe_insn_slots, slot, dirty);
mutex_unlock(&kprobe_insn_mutex);
}
+#ifdef CONFIG_OPTPROBES
+/* For optimized_kprobe buffer */
+static DEFINE_MUTEX(kprobe_optinsn_mutex); /* Protects kprobe_optinsn_slots */
+static struct kprobe_insn_cache kprobe_optinsn_slots = {
+ .pages = LIST_HEAD_INIT(kprobe_optinsn_slots.pages),
+ /* .insn_size is initialized later */
+ .nr_garbage = 0,
+};
+/* Get a slot for optimized_kprobe buffer */
+kprobe_opcode_t __kprobes *get_optinsn_slot(void)
+{
+ kprobe_opcode_t *ret = NULL;
+
+ mutex_lock(&kprobe_optinsn_mutex);
+ ret = __get_insn_slot(&kprobe_optinsn_slots);
+ mutex_unlock(&kprobe_optinsn_mutex);
+
+ return ret;
+}
+
+void __kprobes free_optinsn_slot(kprobe_opcode_t * slot, int dirty)
+{
+ mutex_lock(&kprobe_optinsn_mutex);
+ __free_insn_slot(&kprobe_optinsn_slots, slot, dirty);
+ mutex_unlock(&kprobe_optinsn_mutex);
+}
+#endif
#endif
/* We have preemption disabled.. so it is safe to use __ versions */
@@ -333,23 +361,303 @@ struct kprobe __kprobes *get_kprobe(void *addr)
if (p->addr == addr)
return p;
}
+
+ return NULL;
+}
+
+static int __kprobes aggr_pre_handler(struct kprobe *p, struct pt_regs *regs);
+
+/* Return true if the kprobe is an aggregator */
+static inline int kprobe_aggrprobe(struct kprobe *p)
+{
+ return p->pre_handler == aggr_pre_handler;
+}
+
+/*
+ * Keep all fields in the kprobe consistent
+ */
+static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p)
+{
+ memcpy(&p->opcode, &old_p->opcode, sizeof(kprobe_opcode_t));
+ memcpy(&p->ainsn, &old_p->ainsn, sizeof(struct arch_specific_insn));
+}
+
+#ifdef CONFIG_OPTPROBES
+/*
+ * Call all pre_handler on the list, but ignores its return value.
+ * This must be called from arch-dep optimized caller.
+ */
+void __kprobes opt_pre_handler(struct kprobe *p, struct pt_regs *regs)
+{
+ struct kprobe *kp;
+
+ list_for_each_entry_rcu(kp, &p->list, list) {
+ if (kp->pre_handler && likely(!kprobe_disabled(kp))) {
+ set_kprobe_instance(kp);
+ kp->pre_handler(kp, regs);
+ }
+ reset_kprobe_instance();
+ }
+}
+
+/* Return true(!0) if the kprobe is ready for optimization. */
+static inline int kprobe_optready(struct kprobe *p)
+{
+ struct optimized_kprobe *op;
+
+ if (kprobe_aggrprobe(p)) {
+ op = container_of(p, struct optimized_kprobe, kp);
+ return arch_prepared_optinsn(&op->optinsn);
+ }
+
+ return 0;
+}
+
+/* Return an optimized kprobe which replaces instructions including addr. */
+struct kprobe *__kprobes get_optimized_kprobe(unsigned long addr)
+{
+ int i;
+ struct kprobe *p = NULL;
+ struct optimized_kprobe *op;
+
+ for (i = 0; !p && i < MAX_OPTIMIZED_LENGTH; i++)
+ p = get_kprobe((void *)(addr - i));
+
+ if (p && kprobe_optready(p)) {
+ op = container_of(p, struct optimized_kprobe, kp);
+ if (arch_within_optimized_kprobe(op, addr))
+ return p;
+ }
+
return NULL;
}
+/* Optimization staging list, protected by kprobe_mutex */
+static LIST_HEAD(optimizing_list);
+
+static void kprobe_optimizer(struct work_struct *work);
+static DECLARE_DELAYED_WORK(optimizing_work, kprobe_optimizer);
+#define OPTIMIZE_DELAY 5
+
+/* Kprobe jump optimizer */
+static __kprobes void kprobe_optimizer(struct work_struct *work)
+{
+ struct optimized_kprobe *op, *tmp;
+
+ /* Lock modules while optimizing kprobes */
+ mutex_lock(&module_mutex);
+ mutex_lock(&kprobe_mutex);
+ if (kprobes_all_disarmed)
+ goto end;
+
+ /*
+ * Wait for quiesence period to ensure all running interrupts
+ * are done. Because optprobe may modify multiple instructions
+ * there is a chance that Nth instruction is interrupted. In that
+ * case, running interrupt can return to 2nd-Nth byte of jump
+ * instruction. This wait is for avoiding it.
+ */
+ synchronize_sched();
+
+ /*
+ * The optimization/unoptimization refers online_cpus via
+ * stop_machine() and cpu-hotplug modifies online_cpus.
+ * And same time, text_mutex will be held in cpu-hotplug and here.
+ * This combination can cause a deadlock (cpu-hotplug try to lock
+ * text_mutex but stop_machine can not be done because online_cpus
+ * has been changed)
+ * To avoid this deadlock, we need to call get_online_cpus()
+ * for preventing cpu-hotplug outside of text_mutex locking.
+ */
+ get_online_cpus();
+ mutex_lock(&text_mutex);
+ list_for_each_entry_safe(op, tmp, &optimizing_list, list) {
+ WARN_ON(kprobe_disabled(&op->kp));
+ if (arch_optimize_kprobe(op) < 0)
+ op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED;
+ list_del_init(&op->list);
+ }
+ mutex_unlock(&text_mutex);
+ put_online_cpus();
+end:
+ mutex_unlock(&kprobe_mutex);
+ mutex_unlock(&module_mutex);
+}
+
+/* Optimize kprobe if p is ready to be optimized */
+static __kprobes void optimize_kprobe(struct kprobe *p)
+{
+ struct optimized_kprobe *op;
+
+ /* Check if the kprobe is disabled or not ready for optimization. */
+ if (!kprobe_optready(p) ||
+ (kprobe_disabled(p) || kprobes_all_disarmed))
+ return;
+
+ /* Both of break_handler and post_handler are not supported. */
+ if (p->break_handler || p->post_handler)
+ return;
+
+ op = container_of(p, struct optimized_kprobe, kp);
+
+ /* Check there is no other kprobes at the optimized instructions */
+ if (arch_check_optimized_kprobe(op) < 0)
+ return;
+
+ /* Check if it is already optimized. */
+ if (op->kp.flags & KPROBE_FLAG_OPTIMIZED)
+ return;
+
+ op->kp.flags |= KPROBE_FLAG_OPTIMIZED;
+ list_add(&op->list, &optimizing_list);
+ if (!delayed_work_pending(&optimizing_work))
+ schedule_delayed_work(&optimizing_work, OPTIMIZE_DELAY);
+}
+
+/* Unoptimize a kprobe if p is optimized */
+static __kprobes void unoptimize_kprobe(struct kprobe *p)
+{
+ struct optimized_kprobe *op;
+
+ if ((p->flags & KPROBE_FLAG_OPTIMIZED) && kprobe_aggrprobe(p)) {
+ op = container_of(p, struct optimized_kprobe, kp);
+ if (!list_empty(&op->list))
+ /* Dequeue from the optimization queue */
+ list_del_init(&op->list);
+ else
+ /* Replace jump with break */
+ arch_unoptimize_kprobe(op);
+ op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED;
+ }
+}
+
+/* Remove optimized instructions */
+static void __kprobes kill_optimized_kprobe(struct kprobe *p)
+{
+ struct optimized_kprobe *op;
+
+ op = container_of(p, struct optimized_kprobe, kp);
+ if (!list_empty(&op->list)) {
+ /* Dequeue from the optimization queue */
+ list_del_init(&op->list);
+ op->kp.flags &= ~KPROBE_FLAG_OPTIMIZED;
+ }
+ /* Don't unoptimize, because the target code will be freed. */
+ arch_remove_optimized_kprobe(op);
+}
+
+/* Try to prepare optimized instructions */
+static __kprobes void prepare_optimized_kprobe(struct kprobe *p)
+{
+ struct optimized_kprobe *op;
+
+ op = container_of(p, struct optimized_kprobe, kp);
+ arch_prepare_optimized_kprobe(op);
+}
+
+/* Free optimized instructions and optimized_kprobe */
+static __kprobes void free_aggr_kprobe(struct kprobe *p)
+{
+ struct optimized_kprobe *op;
+
+ op = container_of(p, struct optimized_kprobe, kp);
+ arch_remove_optimized_kprobe(op);
+ kfree(op);
+}
+
+/* Allocate new optimized_kprobe and try to prepare optimized instructions */
+static __kprobes struct kprobe *alloc_aggr_kprobe(struct kprobe *p)
+{
+ struct optimized_kprobe *op;
+
+ op = kzalloc(sizeof(struct optimized_kprobe), GFP_KERNEL);
+ if (!op)
+ return NULL;
+
+ INIT_LIST_HEAD(&op->list);
+ op->kp.addr = p->addr;
+ arch_prepare_optimized_kprobe(op);
+
+ return &op->kp;
+}
+
+static void __kprobes init_aggr_kprobe(struct kprobe *ap, struct kprobe *p);
+
+/*
+ * Prepare an optimized_kprobe and optimize it
+ * NOTE: p must be a normal registered kprobe
+ */
+static __kprobes void try_to_optimize_kprobe(struct kprobe *p)
+{
+ struct kprobe *ap;
+ struct optimized_kprobe *op;
+
+ ap = alloc_aggr_kprobe(p);
+ if (!ap)
+ return;
+
+ op = container_of(ap, struct optimized_kprobe, kp);
+ if (!arch_prepared_optinsn(&op->optinsn)) {
+ /* If failed to setup optimizing, fallback to kprobe */
+ free_aggr_kprobe(ap);
+ return;
+ }
+
+ init_aggr_kprobe(ap, p);
+ optimize_kprobe(ap);
+}
+#else /* !CONFIG_OPTPROBES */
+#define get_optimized_kprobe(addr) (NULL)
+#define optimize_kprobe(p) do {} while (0)
+#define unoptimize_kprobe(p) do {} while (0)
+#define kill_optimized_kprobe(p) do {} while (0)
+#define prepare_optimized_kprobe(p) do {} while (0)
+#define try_to_optimize_kprobe(p) do {} while (0)
+
+static __kprobes void free_aggr_kprobe(struct kprobe *p)
+{
+ kfree(p);
+}
+
+static __kprobes struct kprobe *alloc_aggr_kprobe(struct kprobe *p)
+{
+ return kzalloc(sizeof(struct kprobe), GFP_KERNEL);
+}
+#endif /* CONFIG_OPTPROBES */
+
+static void __kprobes __arm_kprobe(struct kprobe *kp)
+{
+ arch_arm_kprobe(kp);
+ optimize_kprobe(kp); /* Try to optimize (add kprobe to a list) */
+}
+
+static void __kprobes __disarm_kprobe(struct kprobe *kp)
+{
+ unoptimize_kprobe(kp); /* Try to unoptimize */
+ arch_disarm_kprobe(kp);
+}
+
/* Arm a kprobe with text_mutex */
static void __kprobes arm_kprobe(struct kprobe *kp)
{
+ /*
+ * Here, since __arm_kprobe() doesn't use stop_machine(),
+ * this doesn't cause deadlock on text_mutex. So, we don't
+ * need get_online_cpus().
+ */
mutex_lock(&text_mutex);
- arch_arm_kprobe(kp);
+ __arm_kprobe(kp);
mutex_unlock(&text_mutex);
}
/* Disarm a kprobe with text_mutex */
static void __kprobes disarm_kprobe(struct kprobe *kp)
{
+ get_online_cpus(); /* For avoiding text_mutex deadlock */
mutex_lock(&text_mutex);
- arch_disarm_kprobe(kp);
+ __disarm_kprobe(kp);
mutex_unlock(&text_mutex);
+ put_online_cpus();
}
/*
@@ -418,7 +726,7 @@ static int __kprobes aggr_break_handler(struct kprobe *p, struct pt_regs *regs)
void __kprobes kprobes_inc_nmissed_count(struct kprobe *p)
{
struct kprobe *kp;
- if (p->pre_handler != aggr_pre_handler) {
+ if (!kprobe_aggrprobe(p)) {
p->nmissed++;
} else {
list_for_each_entry_rcu(kp, &p->list, list)
@@ -542,21 +850,16 @@ static void __kprobes cleanup_rp_inst(struct kretprobe *rp)
}
/*
- * Keep all fields in the kprobe consistent
- */
-static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p)
-{
- memcpy(&p->opcode, &old_p->opcode, sizeof(kprobe_opcode_t));
- memcpy(&p->ainsn, &old_p->ainsn, sizeof(struct arch_specific_insn));
-}
-
-/*
* Add the new probe to ap->list. Fail if this is the
* second jprobe at the address - two jprobes can't coexist
*/
static int __kprobes add_new_kprobe(struct kprobe *ap, struct kprobe *p)
{
BUG_ON(kprobe_gone(ap) || kprobe_gone(p));
+
+ if (p->break_handler || p->post_handler)
+ unoptimize_kprobe(ap); /* Fall back to normal kprobe */
+
if (p->break_handler) {
if (ap->break_handler)
return -EEXIST;
@@ -571,7 +874,7 @@ static int __kprobes add_new_kprobe(struct kprobe *ap, struct kprobe *p)
ap->flags &= ~KPROBE_FLAG_DISABLED;
if (!kprobes_all_disarmed)
/* Arm the breakpoint again. */
- arm_kprobe(ap);
+ __arm_kprobe(ap);
}
return 0;
}
@@ -580,12 +883,13 @@ static int __kprobes add_new_kprobe(struct kprobe *ap, struct kprobe *p)
* Fill in the required fields of the "manager kprobe". Replace the
* earlier kprobe in the hlist with the manager kprobe
*/
-static inline void add_aggr_kprobe(struct kprobe *ap, struct kprobe *p)
+static void __kprobes init_aggr_kprobe(struct kprobe *ap, struct kprobe *p)
{
+ /* Copy p's insn slot to ap */
copy_kprobe(p, ap);
flush_insn_slot(ap);
ap->addr = p->addr;
- ap->flags = p->flags;
+ ap->flags = p->flags & ~KPROBE_FLAG_OPTIMIZED;
ap->pre_handler = aggr_pre_handler;
ap->fault_handler = aggr_fault_handler;
/* We don't care the kprobe which has gone. */
@@ -595,8 +899,9 @@ static inline void add_aggr_kprobe(struct kprobe *ap, struct kprobe *p)
ap->break_handler = aggr_break_handler;
INIT_LIST_HEAD(&ap->list);
- list_add_rcu(&p->list, &ap->list);
+ INIT_HLIST_NODE(&ap->hlist);
+ list_add_rcu(&p->list, &ap->list);
hlist_replace_rcu(&p->hlist, &ap->hlist);
}
@@ -610,12 +915,12 @@ static int __kprobes register_aggr_kprobe(struct kprobe *old_p,
int ret = 0;
struct kprobe *ap = old_p;
- if (old_p->pre_handler != aggr_pre_handler) {
- /* If old_p is not an aggr_probe, create new aggr_kprobe. */
- ap = kzalloc(sizeof(struct kprobe), GFP_KERNEL);
+ if (!kprobe_aggrprobe(old_p)) {
+ /* If old_p is not an aggr_kprobe, create new aggr_kprobe. */
+ ap = alloc_aggr_kprobe(old_p);
if (!ap)
return -ENOMEM;
- add_aggr_kprobe(ap, old_p);
+ init_aggr_kprobe(ap, old_p);
}
if (kprobe_gone(ap)) {
@@ -634,6 +939,9 @@ static int __kprobes register_aggr_kprobe(struct kprobe *old_p,
*/
return ret;
+ /* Prepare optimized instructions if possible. */
+ prepare_optimized_kprobe(ap);
+
/*
* Clear gone flag to prevent allocating new slot again, and
* set disabled flag because it is not armed yet.
@@ -642,6 +950,7 @@ static int __kprobes register_aggr_kprobe(struct kprobe *old_p,
| KPROBE_FLAG_DISABLED;
}
+ /* Copy ap's insn slot to p */
copy_kprobe(ap, p);
return add_new_kprobe(ap, p);
}
@@ -791,16 +1100,25 @@ int __kprobes register_kprobe(struct kprobe *p)
p->nmissed = 0;
INIT_LIST_HEAD(&p->list);
mutex_lock(&kprobe_mutex);
+
+ get_online_cpus(); /* For avoiding text_mutex deadlock. */
+ mutex_lock(&text_mutex);
+
old_p = get_kprobe(p->addr);
if (old_p) {
+ /* Since this may unoptimize old_p, locking text_mutex. */
ret = register_aggr_kprobe(old_p, p);
goto out;
}
- mutex_lock(&text_mutex);
+ /* Check collision with other optimized kprobes */
+ old_p = get_optimized_kprobe((unsigned long)p->addr);
+ if (unlikely(old_p))
+ unoptimize_kprobe(old_p); /* Fallback to unoptimized kprobe */
+
ret = arch_prepare_kprobe(p);
if (ret)
- goto out_unlock_text;
+ goto out;
INIT_HLIST_NODE(&p->hlist);
hlist_add_head_rcu(&p->hlist,
@@ -809,9 +1127,12 @@ int __kprobes register_kprobe(struct kprobe *p)
if (!kprobes_all_disarmed && !kprobe_disabled(p))
arch_arm_kprobe(p);
-out_unlock_text:
- mutex_unlock(&text_mutex);
+ /* Try to optimize kprobe */
+ try_to_optimize_kprobe(p);
+
out:
+ mutex_unlock(&text_mutex);
+ put_online_cpus();
mutex_unlock(&kprobe_mutex);
if (probed_mod)
@@ -833,7 +1154,7 @@ static int __kprobes __unregister_kprobe_top(struct kprobe *p)
return -EINVAL;
if (old_p == p ||
- (old_p->pre_handler == aggr_pre_handler &&
+ (kprobe_aggrprobe(old_p) &&
list_is_singular(&old_p->list))) {
/*
* Only probe on the hash list. Disarm only if kprobes are
@@ -841,8 +1162,13 @@ static int __kprobes __unregister_kprobe_top(struct kprobe *p)
* already have been removed. We save on flushing icache.
*/
if (!kprobes_all_disarmed && !kprobe_disabled(old_p))
- disarm_kprobe(p);
+ disarm_kprobe(old_p);
hlist_del_rcu(&old_p->hlist);
+
+ /* If another kprobe was blocked, optimize it. */
+ old_p = get_optimized_kprobe((unsigned long)p->addr);
+ if (unlikely(old_p))
+ optimize_kprobe(old_p);
} else {
if (p->break_handler && !kprobe_gone(p))
old_p->break_handler = NULL;
@@ -857,8 +1183,13 @@ noclean:
list_del_rcu(&p->list);
if (!kprobe_disabled(old_p)) {
try_to_disable_aggr_kprobe(old_p);
- if (!kprobes_all_disarmed && kprobe_disabled(old_p))
- disarm_kprobe(old_p);
+ if (!kprobes_all_disarmed) {
+ if (kprobe_disabled(old_p))
+ disarm_kprobe(old_p);
+ else
+ /* Try to optimize this probe again */
+ optimize_kprobe(old_p);
+ }
}
}
return 0;
@@ -875,7 +1206,7 @@ static void __kprobes __unregister_kprobe_bottom(struct kprobe *p)
old_p = list_entry(p->list.next, struct kprobe, list);
list_del(&p->list);
arch_remove_kprobe(old_p);
- kfree(old_p);
+ free_aggr_kprobe(old_p);
}
}
@@ -1171,7 +1502,7 @@ static void __kprobes kill_kprobe(struct kprobe *p)
struct kprobe *kp;
p->flags |= KPROBE_FLAG_GONE;
- if (p->pre_handler == aggr_pre_handler) {
+ if (kprobe_aggrprobe(p)) {
/*
* If this is an aggr_kprobe, we have to list all the
* chained probes and mark them GONE.
@@ -1180,6 +1511,7 @@ static void __kprobes kill_kprobe(struct kprobe *p)
kp->flags |= KPROBE_FLAG_GONE;
p->post_handler = NULL;
p->break_handler = NULL;
+ kill_optimized_kprobe(p);
}
/*
* Here, we can remove insn_slot safely, because no thread calls
@@ -1289,6 +1621,11 @@ static int __init init_kprobes(void)
}
}
+#if defined(CONFIG_OPTPROBES) && defined(__ARCH_WANT_KPROBES_INSN_SLOT)
+ /* Init kprobe_optinsn_slots */
+ kprobe_optinsn_slots.insn_size = MAX_OPTINSN_SIZE;
+#endif
+
/* By default, kprobes are armed */
kprobes_all_disarmed = false;
@@ -1307,7 +1644,7 @@ static int __init init_kprobes(void)
#ifdef CONFIG_DEBUG_FS
static void __kprobes report_probe(struct seq_file *pi, struct kprobe *p,
- const char *sym, int offset,char *modname)
+ const char *sym, int offset, char *modname, struct kprobe *pp)
{
char *kprobe_type;
@@ -1317,19 +1654,21 @@ static void __kprobes report_probe(struct seq_file *pi, struct kprobe *p,
kprobe_type = "j";
else
kprobe_type = "k";
+
if (sym)
- seq_printf(pi, "%p %s %s+0x%x %s %s%s\n",
+ seq_printf(pi, "%p %s %s+0x%x %s ",
p->addr, kprobe_type, sym, offset,
- (modname ? modname : " "),
- (kprobe_gone(p) ? "[GONE]" : ""),
- ((kprobe_disabled(p) && !kprobe_gone(p)) ?
- "[DISABLED]" : ""));
+ (modname ? modname : " "));
else
- seq_printf(pi, "%p %s %p %s%s\n",
- p->addr, kprobe_type, p->addr,
- (kprobe_gone(p) ? "[GONE]" : ""),
- ((kprobe_disabled(p) && !kprobe_gone(p)) ?
- "[DISABLED]" : ""));
+ seq_printf(pi, "%p %s %p ",
+ p->addr, kprobe_type, p->addr);
+
+ if (!pp)
+ pp = p;
+ seq_printf(pi, "%s%s%s\n",
+ (kprobe_gone(p) ? "[GONE]" : ""),
+ ((kprobe_disabled(p) && !kprobe_gone(p)) ? "[DISABLED]" : ""),
+ (kprobe_optimized(pp) ? "[OPTIMIZED]" : ""));
}
static void __kprobes *kprobe_seq_start(struct seq_file *f, loff_t *pos)
@@ -1365,11 +1704,11 @@ static int __kprobes show_kprobe_addr(struct seq_file *pi, void *v)
hlist_for_each_entry_rcu(p, node, head, hlist) {
sym = kallsyms_lookup((unsigned long)p->addr, NULL,
&offset, &modname, namebuf);
- if (p->pre_handler == aggr_pre_handler) {
+ if (kprobe_aggrprobe(p)) {
list_for_each_entry_rcu(kp, &p->list, list)
- report_probe(pi, kp, sym, offset, modname);
+ report_probe(pi, kp, sym, offset, modname, p);
} else
- report_probe(pi, p, sym, offset, modname);
+ report_probe(pi, p, sym, offset, modname, NULL);
}
preempt_enable();
return 0;
@@ -1472,12 +1811,13 @@ static void __kprobes arm_all_kprobes(void)
if (!kprobes_all_disarmed)
goto already_enabled;
+ /* Arming kprobes doesn't optimize kprobe itself */
mutex_lock(&text_mutex);
for (i = 0; i < KPROBE_TABLE_SIZE; i++) {
head = &kprobe_table[i];
hlist_for_each_entry_rcu(p, node, head, hlist)
if (!kprobe_disabled(p))
- arch_arm_kprobe(p);
+ __arm_kprobe(p);
}
mutex_unlock(&text_mutex);
@@ -1504,16 +1844,23 @@ static void __kprobes disarm_all_kprobes(void)
kprobes_all_disarmed = true;
printk(KERN_INFO "Kprobes globally disabled\n");
+
+ /*
+ * Here we call get_online_cpus() for avoiding text_mutex deadlock,
+ * because disarming may also unoptimize kprobes.
+ */
+ get_online_cpus();
mutex_lock(&text_mutex);
for (i = 0; i < KPROBE_TABLE_SIZE; i++) {
head = &kprobe_table[i];
hlist_for_each_entry_rcu(p, node, head, hlist) {
if (!arch_trampoline_kprobe(p) && !kprobe_disabled(p))
- arch_disarm_kprobe(p);
+ __disarm_kprobe(p);
}
}
mutex_unlock(&text_mutex);
+ put_online_cpus();
mutex_unlock(&kprobe_mutex);
/* Allow all currently running kprobes to complete */
synchronize_sched();
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Add /proc/sys/debug/kprobes-optimization sysctl which enables and disables
kprobes jump optimization on the fly for debugging.
Changes in v7:
- Remove ctl_name = CTL_UNNUMBERED for upstream compatibility.
Changes in v6:
- Update comments and coding style.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
include/linux/kprobes.h | 8 ++++
kernel/kprobes.c | 89 +++++++++++++++++++++++++++++++++++++++++++++--
kernel/sysctl.c | 12 ++++++
3 files changed, 106 insertions(+), 3 deletions(-)
diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
index aed1f95..e7d1b2e 100644
--- a/include/linux/kprobes.h
+++ b/include/linux/kprobes.h
@@ -283,6 +283,14 @@ extern int arch_within_optimized_kprobe(struct optimized_kprobe *op,
unsigned long addr);
extern void opt_pre_handler(struct kprobe *p, struct pt_regs *regs);
+
+#ifdef CONFIG_SYSCTL
+extern int sysctl_kprobes_optimization;
+extern int proc_kprobes_optimization_handler(struct ctl_table *table,
+ int write, void __user *buffer,
+ size_t *length, loff_t *ppos);
+#endif
+
#endif /* CONFIG_OPTPROBES */
/* Get the kprobe at this addr (if any) - called with preemption disabled */
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 759bb8e..7f30d2f 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -42,6 +42,7 @@
#include <linux/freezer.h>
#include <linux/seq_file.h>
#include <linux/debugfs.h>
+#include <linux/sysctl.h>
#include <linux/kdebug.h>
#include <linux/memory.h>
#include <linux/cpu.h>
@@ -383,6 +384,9 @@ static inline void copy_kprobe(struct kprobe *old_p, struct kprobe *p)
}
#ifdef CONFIG_OPTPROBES
+/* NOTE: change this value only with kprobe_mutex held */
+static bool kprobes_allow_optimization;
+
/*
* Call all pre_handler on the list, but ignores its return value.
* This must be called from arch-dep optimized caller.
@@ -447,7 +451,7 @@ static __kprobes void kprobe_optimizer(struct work_struct *work)
/* Lock modules while optimizing kprobes */
mutex_lock(&module_mutex);
mutex_lock(&kprobe_mutex);
- if (kprobes_all_disarmed)
+ if (kprobes_all_disarmed || !kprobes_allow_optimization)
goto end;
/*
@@ -490,7 +494,7 @@ static __kprobes void optimize_kprobe(struct kprobe *p)
struct optimized_kprobe *op;
/* Check if the kprobe is disabled or not ready for optimization. */
- if (!kprobe_optready(p) ||
+ if (!kprobe_optready(p) || !kprobes_allow_optimization ||
(kprobe_disabled(p) || kprobes_all_disarmed))
return;
@@ -606,6 +610,81 @@ static __kprobes void try_to_optimize_kprobe(struct kprobe *p)
init_aggr_kprobe(ap, p);
optimize_kprobe(ap);
}
+
+#ifdef CONFIG_SYSCTL
+static void __kprobes optimize_all_kprobes(void)
+{
+ struct hlist_head *head;
+ struct hlist_node *node;
+ struct kprobe *p;
+ unsigned int i;
+
+ /* If optimization is already allowed, just return */
+ if (kprobes_allow_optimization)
+ return;
+
+ kprobes_allow_optimization = true;
+ mutex_lock(&text_mutex);
+ for (i = 0; i < KPROBE_TABLE_SIZE; i++) {
+ head = &kprobe_table[i];
+ hlist_for_each_entry_rcu(p, node, head, hlist)
+ if (!kprobe_disabled(p))
+ optimize_kprobe(p);
+ }
+ mutex_unlock(&text_mutex);
+ printk(KERN_INFO "Kprobes globally optimized\n");
+}
+
+static void __kprobes unoptimize_all_kprobes(void)
+{
+ struct hlist_head *head;
+ struct hlist_node *node;
+ struct kprobe *p;
+ unsigned int i;
+
+ /* If optimization is already prohibited, just return */
+ if (!kprobes_allow_optimization)
+ return;
+
+ kprobes_allow_optimization = false;
+ printk(KERN_INFO "Kprobes globally unoptimized\n");
+ get_online_cpus(); /* For avoiding text_mutex deadlock */
+ mutex_lock(&text_mutex);
+ for (i = 0; i < KPROBE_TABLE_SIZE; i++) {
+ head = &kprobe_table[i];
+ hlist_for_each_entry_rcu(p, node, head, hlist) {
+ if (!kprobe_disabled(p))
+ unoptimize_kprobe(p);
+ }
+ }
+
+ mutex_unlock(&text_mutex);
+ put_online_cpus();
+ /* Allow all currently running kprobes to complete */
+ synchronize_sched();
+}
+
+int sysctl_kprobes_optimization;
+int proc_kprobes_optimization_handler(struct ctl_table *table, int write,
+ void __user *buffer, size_t *length,
+ loff_t *ppos)
+{
+ int ret;
+
+ mutex_lock(&kprobe_mutex);
+ sysctl_kprobes_optimization = kprobes_allow_optimization ? 1 : 0;
+ ret = proc_dointvec_minmax(table, write, buffer, length, ppos);
+
+ if (sysctl_kprobes_optimization)
+ optimize_all_kprobes();
+ else
+ unoptimize_all_kprobes();
+ mutex_unlock(&kprobe_mutex);
+
+ return ret;
+}
+#endif /* CONFIG_SYSCTL */
+
#else /* !CONFIG_OPTPROBES */
#define get_optimized_kprobe(addr) (NULL)
#define optimize_kprobe(p) do {} while (0)
@@ -1621,10 +1700,14 @@ static int __init init_kprobes(void)
}
}
-#if defined(CONFIG_OPTPROBES) && defined(__ARCH_WANT_KPROBES_INSN_SLOT)
+#if defined(CONFIG_OPTPROBES)
+#if defined(__ARCH_WANT_KPROBES_INSN_SLOT)
/* Init kprobe_optinsn_slots */
kprobe_optinsn_slots.insn_size = MAX_OPTINSN_SIZE;
#endif
+ /* By default, kprobes can be optimized */
+ kprobes_allow_optimization = true;
+#endif
/* By default, kprobes are armed */
kprobes_all_disarmed = false;
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 9b459f0..7480e80 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -50,6 +50,7 @@
#include <linux/ftrace.h>
#include <linux/slow-work.h>
#include <linux/perf_event.h>
+#include <linux/kprobes.h>
#include <asm/uaccess.h>
#include <asm/processor.h>
@@ -1459,6 +1460,17 @@ static struct ctl_table debug_table[] = {
.proc_handler = proc_dointvec
},
#endif
+#if defined(CONFIG_OPTPROBES)
+ {
+ .procname = "kprobes-optimization",
+ .data = &sysctl_kprobes_optimization,
+ .maxlen = sizeof(int),
+ .mode = 0644,
+ .proc_handler = proc_kprobes_optimization_handler,
+ .extra1 = &zero,
+ .extra2 = &one,
+ },
+#endif
{ }
};
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Integrate prepare_singlestep() into setup_singlestep() to boost up reenter
probes, if possible.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
arch/x86/kernel/kprobes.c | 48 ++++++++++++++++++++++++---------------------
1 files changed, 26 insertions(+), 22 deletions(-)
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index 7039e6e..57de6ca 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -403,18 +403,6 @@ static void __kprobes restore_btf(void)
update_debugctlmsr(current->thread.debugctlmsr);
}
-static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs)
-{
- clear_btf();
- regs->flags |= X86_EFLAGS_TF;
- regs->flags &= ~X86_EFLAGS_IF;
- /* single step inline if the instruction is an int3 */
- if (p->opcode == BREAKPOINT_INSTRUCTION)
- regs->ip = (unsigned long)p->addr;
- else
- regs->ip = (unsigned long)p->ainsn.insn;
-}
-
void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
struct pt_regs *regs)
{
@@ -427,19 +415,38 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
}
static void __kprobes setup_singlestep(struct kprobe *p, struct pt_regs *regs,
- struct kprobe_ctlblk *kcb)
+ struct kprobe_ctlblk *kcb, int reenter)
{
#if !defined(CONFIG_PREEMPT) || defined(CONFIG_FREEZER)
if (p->ainsn.boostable == 1 && !p->post_handler) {
/* Boost up -- we can execute copied instructions directly */
- reset_current_kprobe();
+ if (!reenter)
+ reset_current_kprobe();
+ /*
+ * Reentering boosted probe doesn't reset current_kprobe,
+ * nor set current_kprobe, because it doesn't use single
+ * stepping.
+ */
regs->ip = (unsigned long)p->ainsn.insn;
preempt_enable_no_resched();
return;
}
#endif
- prepare_singlestep(p, regs);
- kcb->kprobe_status = KPROBE_HIT_SS;
+ if (reenter) {
+ save_previous_kprobe(kcb);
+ set_current_kprobe(p, regs, kcb);
+ kcb->kprobe_status = KPROBE_REENTER;
+ } else
+ kcb->kprobe_status = KPROBE_HIT_SS;
+ /* Prepare real single stepping */
+ clear_btf();
+ regs->flags |= X86_EFLAGS_TF;
+ regs->flags &= ~X86_EFLAGS_IF;
+ /* single step inline if the instruction is an int3 */
+ if (p->opcode == BREAKPOINT_INSTRUCTION)
+ regs->ip = (unsigned long)p->addr;
+ else
+ regs->ip = (unsigned long)p->ainsn.insn;
}
/*
@@ -453,11 +460,8 @@ static int __kprobes reenter_kprobe(struct kprobe *p, struct pt_regs *regs,
switch (kcb->kprobe_status) {
case KPROBE_HIT_SSDONE:
case KPROBE_HIT_ACTIVE:
- save_previous_kprobe(kcb);
- set_current_kprobe(p, regs, kcb);
kprobes_inc_nmissed_count(p);
- prepare_singlestep(p, regs);
- kcb->kprobe_status = KPROBE_REENTER;
+ setup_singlestep(p, regs, kcb, 1);
break;
case KPROBE_HIT_SS:
/* A probe has been hit in the codepath leading up to, or just
@@ -532,13 +536,13 @@ static int __kprobes kprobe_handler(struct pt_regs *regs)
* more here.
*/
if (!p->pre_handler || !p->pre_handler(p, regs))
- setup_singlestep(p, regs, kcb);
+ setup_singlestep(p, regs, kcb, 0);
return 1;
}
} else if (kprobe_running()) {
p = __get_cpu_var(current_kprobe);
if (p->break_handler && p->break_handler(p, regs)) {
- setup_singlestep(p, regs, kcb);
+ setup_singlestep(p, regs, kcb, 0);
return 1;
}
} /* else: not a kprobe fault; let the kernel handle it */
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Introduce SAVE/RESOTRE_REGS_STRING for cleanup kretprobe-trampoline asm
code. These macros will be used for emulating interruption.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
arch/x86/kernel/kprobes.c | 128 ++++++++++++++++++++++++---------------------
1 files changed, 67 insertions(+), 61 deletions(-)
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index 57de6ca..f848984 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -551,6 +551,69 @@ static int __kprobes kprobe_handler(struct pt_regs *regs)
return 0;
}
+#ifdef CONFIG_X86_64
+#define SAVE_REGS_STRING \
+ /* Skip cs, ip, orig_ax. */ \
+ " subq $24, %rsp\n" \
+ " pushq %rdi\n" \
+ " pushq %rsi\n" \
+ " pushq %rdx\n" \
+ " pushq %rcx\n" \
+ " pushq %rax\n" \
+ " pushq %r8\n" \
+ " pushq %r9\n" \
+ " pushq %r10\n" \
+ " pushq %r11\n" \
+ " pushq %rbx\n" \
+ " pushq %rbp\n" \
+ " pushq %r12\n" \
+ " pushq %r13\n" \
+ " pushq %r14\n" \
+ " pushq %r15\n"
+#define RESTORE_REGS_STRING \
+ " popq %r15\n" \
+ " popq %r14\n" \
+ " popq %r13\n" \
+ " popq %r12\n" \
+ " popq %rbp\n" \
+ " popq %rbx\n" \
+ " popq %r11\n" \
+ " popq %r10\n" \
+ " popq %r9\n" \
+ " popq %r8\n" \
+ " popq %rax\n" \
+ " popq %rcx\n" \
+ " popq %rdx\n" \
+ " popq %rsi\n" \
+ " popq %rdi\n" \
+ /* Skip orig_ax, ip, cs */ \
+ " addq $24, %rsp\n"
+#else
+#define SAVE_REGS_STRING \
+ /* Skip cs, ip, orig_ax and gs. */ \
+ " subl $16, %esp\n" \
+ " pushl %fs\n" \
+ " pushl %ds\n" \
+ " pushl %es\n" \
+ " pushl %eax\n" \
+ " pushl %ebp\n" \
+ " pushl %edi\n" \
+ " pushl %esi\n" \
+ " pushl %edx\n" \
+ " pushl %ecx\n" \
+ " pushl %ebx\n"
+#define RESTORE_REGS_STRING \
+ " popl %ebx\n" \
+ " popl %ecx\n" \
+ " popl %edx\n" \
+ " popl %esi\n" \
+ " popl %edi\n" \
+ " popl %ebp\n" \
+ " popl %eax\n" \
+ /* Skip ds, es, fs, gs, orig_ax, and ip. Note: don't pop cs here*/\
+ " addl $24, %esp\n"
+#endif
+
/*
* When a retprobed function returns, this code saves registers and
* calls trampoline_handler() runs, which calls the kretprobe's handler.
@@ -564,65 +627,16 @@ static void __used __kprobes kretprobe_trampoline_holder(void)
/* We don't bother saving the ss register */
" pushq %rsp\n"
" pushfq\n"
- /*
- * Skip cs, ip, orig_ax.
- * trampoline_handler() will plug in these values
- */
- " subq $24, %rsp\n"
- " pushq %rdi\n"
- " pushq %rsi\n"
- " pushq %rdx\n"
- " pushq %rcx\n"
- " pushq %rax\n"
- " pushq %r8\n"
- " pushq %r9\n"
- " pushq %r10\n"
- " pushq %r11\n"
- " pushq %rbx\n"
- " pushq %rbp\n"
- " pushq %r12\n"
- " pushq %r13\n"
- " pushq %r14\n"
- " pushq %r15\n"
+ SAVE_REGS_STRING
" movq %rsp, %rdi\n"
" call trampoline_handler\n"
/* Replace saved sp with true return address. */
" movq %rax, 152(%rsp)\n"
- " popq %r15\n"
- " popq %r14\n"
- " popq %r13\n"
- " popq %r12\n"
- " popq %rbp\n"
- " popq %rbx\n"
- " popq %r11\n"
- " popq %r10\n"
- " popq %r9\n"
- " popq %r8\n"
- " popq %rax\n"
- " popq %rcx\n"
- " popq %rdx\n"
- " popq %rsi\n"
- " popq %rdi\n"
- /* Skip orig_ax, ip, cs */
- " addq $24, %rsp\n"
+ RESTORE_REGS_STRING
" popfq\n"
#else
" pushf\n"
- /*
- * Skip cs, ip, orig_ax and gs.
- * trampoline_handler() will plug in these values
- */
- " subl $16, %esp\n"
- " pushl %fs\n"
- " pushl %es\n"
- " pushl %ds\n"
- " pushl %eax\n"
- " pushl %ebp\n"
- " pushl %edi\n"
- " pushl %esi\n"
- " pushl %edx\n"
- " pushl %ecx\n"
- " pushl %ebx\n"
+ SAVE_REGS_STRING
" movl %esp, %eax\n"
" call trampoline_handler\n"
/* Move flags to cs */
@@ -630,15 +644,7 @@ static void __used __kprobes kretprobe_trampoline_holder(void)
" movl %edx, 52(%esp)\n"
/* Replace saved flags with true return address. */
" movl %eax, 56(%esp)\n"
- " popl %ebx\n"
- " popl %ecx\n"
- " popl %edx\n"
- " popl %esi\n"
- " popl %edi\n"
- " popl %ebp\n"
- " popl %eax\n"
- /* Skip ds, es, fs, gs, orig_ax and ip */
- " addl $24, %esp\n"
+ RESTORE_REGS_STRING
" popf\n"
#endif
" ret\n");
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Add generic text_poke_smp for SMP which uses stop_machine()
to synchronize modifying code.
This stop_machine() method is officially described at "7.1.3
Handling Self- and Cross-Modifying Code" on the intel's
software developer's manual 3A.
Since stop_machine() can't protect code against NMI/MCE, this
function can not modify those handlers. And also, this function
is basically for modifying multibyte-single-instruction. For
modifying multibyte-multi-instructions, we need another special
trap & detour code.
This code originaly comes from immediate values with stop_machine()
version. Thanks Mathieu!
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
---
arch/x86/include/asm/alternative.h | 4 ++
arch/x86/kernel/alternative.c | 60 ++++++++++++++++++++++++++++++++++++
2 files changed, 63 insertions(+), 1 deletions(-)
diff --git a/arch/x86/include/asm/alternative.h b/arch/x86/include/asm/alternative.h
index 3b5b828..09388c6 100644
--- a/arch/x86/include/asm/alternative.h
+++ b/arch/x86/include/asm/alternative.h
@@ -160,10 +160,12 @@ static inline void apply_paravirt(struct paravirt_patch_site *start,
* invalid instruction possible) or if the instructions are changed from a
* consistent state to another consistent state atomically.
* More care must be taken when modifying code in the SMP case because of
- * Intel's errata.
+ * Intel's errata. text_poke_smp() takes care that errata, but still
+ * doesn't support NMI/MCE handler code modifying.
* On the local CPU you need to be protected again NMI or MCE handlers seeing an
* inconsistent instruction while you patch.
*/
extern void *text_poke(void *addr, const void *opcode, size_t len);
+extern void *text_poke_smp(void *addr, const void *opcode, size_t len);
#endif /* _ASM_X86_ALTERNATIVE_H */
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 2589ea4..aff169e 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -7,6 +7,7 @@
#include <linux/mm.h>
#include <linux/vmalloc.h>
#include <linux/memory.h>
+#include <linux/stop_machine.h>
#include <asm/alternative.h>
#include <asm/sections.h>
#include <asm/pgtable.h>
@@ -554,3 +555,62 @@ void *__kprobes text_poke(void *addr, const void *opcode, size_t len)
local_irq_restore(flags);
return addr;
}
+
+/*
+ * Cross-modifying kernel text with stop_machine().
+ * This code originally comes from immediate value.
+ */
+static atomic_t stop_machine_first;
+static int wrote_text;
+
+struct text_poke_params {
+ void *addr;
+ const void *opcode;
+ size_t len;
+};
+
+static int __kprobes stop_machine_text_poke(void *data)
+{
+ struct text_poke_params *tpp = data;
+
+ if (atomic_dec_and_test(&stop_machine_first)) {
+ text_poke(tpp->addr, tpp->opcode, tpp->len);
+ smp_wmb(); /* Make sure other cpus see that this has run */
+ wrote_text = 1;
+ } else {
+ while (!wrote_text)
+ smp_rmb();
+ sync_core();
+ }
+
+ flush_icache_range((unsigned long)tpp->addr,
+ (unsigned long)tpp->addr + tpp->len);
+ return 0;
+}
+
+/**
+ * text_poke_smp - Update instructions on a live kernel on SMP
+ * @addr: address to modify
+ * @opcode: source of the copy
+ * @len: length to copy
+ *
+ * Modify multi-byte instruction by using stop_machine() on SMP. This allows
+ * user to poke/set multi-byte text on SMP. Only non-NMI/MCE code modifying
+ * should be allowed, since stop_machine() does _not_ protect code against
+ * NMI and MCE.
+ *
+ * Note: Must be called under get_online_cpus() and text_mutex.
+ */
+void *__kprobes text_poke_smp(void *addr, const void *opcode, size_t len)
+{
+ struct text_poke_params tpp;
+
+ tpp.addr = addr;
+ tpp.opcode = opcode;
+ tpp.len = len;
+ atomic_set(&stop_machine_first, 1);
+ wrote_text = 0;
+ stop_machine(stop_machine_text_poke, (void *)&tpp, NULL);
+ return addr;
+}
+
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Introduce x86 arch-specific optimization code, which supports both of
x86-32 and x86-64.
This code also supports safety checking, which decodes whole of a function
in which probe is inserted, and checks following conditions before
optimization:
- The optimized instructions which will be replaced by a jump instruction
don't straddle the function boundary.
- There is no indirect jump instruction, because it will jumps into
the address range which is replaced by jump operand.
- There is no jump/loop instruction which jumps into the address range
which is replaced by jump operand.
- Don't optimize kprobes if it is in functions into which fixup code will
jumps.
This uses text_poke_multibyte() which doesn't support modifying code on
NMI/MCE handler. However, since kprobes itself doesn't support NMI/MCE
code probing, it's not a problem.
Changes in v6:
- Split stop_machine-based jump patching code.
- Update comments and coding style.
Changes in v5:
- Introduce stop_machine-based jump replacing.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
arch/x86/Kconfig | 1
arch/x86/include/asm/kprobes.h | 29 +++
arch/x86/kernel/kprobes.c | 420 ++++++++++++++++++++++++++++++++++++++--
3 files changed, 428 insertions(+), 22 deletions(-)
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index cbcbfde..e6f5a98 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -31,6 +31,7 @@ config X86
select ARCH_WANT_FRAME_POINTERS
select HAVE_DMA_ATTRS
select HAVE_KRETPROBES
+ select HAVE_OPTPROBES
select HAVE_FTRACE_MCOUNT_RECORD
select HAVE_DYNAMIC_FTRACE
select HAVE_FUNCTION_TRACER
diff --git a/arch/x86/include/asm/kprobes.h b/arch/x86/include/asm/kprobes.h
index eaec8ea..4ffa345 100644
--- a/arch/x86/include/asm/kprobes.h
+++ b/arch/x86/include/asm/kprobes.h
@@ -33,6 +33,9 @@ struct kprobe;
typedef u8 kprobe_opcode_t;
#define BREAKPOINT_INSTRUCTION 0xcc
#define RELATIVEJUMP_OPCODE 0xe9
+#define RELATIVEJUMP_SIZE 5
+#define RELATIVECALL_OPCODE 0xe8
+#define RELATIVE_ADDR_SIZE 4
#define MAX_INSN_SIZE 16
#define MAX_STACK_SIZE 64
#define MIN_STACK_SIZE(ADDR) \
@@ -44,6 +47,17 @@ typedef u8 kprobe_opcode_t;
#define flush_insn_slot(p) do { } while (0)
+/* optinsn template addresses */
+extern kprobe_opcode_t optprobe_template_entry;
+extern kprobe_opcode_t optprobe_template_val;
+extern kprobe_opcode_t optprobe_template_call;
+extern kprobe_opcode_t optprobe_template_end;
+#define MAX_OPTIMIZED_LENGTH (MAX_INSN_SIZE + RELATIVE_ADDR_SIZE)
+#define MAX_OPTINSN_SIZE \
+ (((unsigned long)&optprobe_template_end - \
+ (unsigned long)&optprobe_template_entry) + \
+ MAX_OPTIMIZED_LENGTH + RELATIVEJUMP_SIZE)
+
extern const int kretprobe_blacklist_size;
void arch_remove_kprobe(struct kprobe *p);
@@ -64,6 +78,21 @@ struct arch_specific_insn {
int boostable;
};
+struct arch_optimized_insn {
+ /* copy of the original instructions */
+ kprobe_opcode_t copied_insn[RELATIVE_ADDR_SIZE];
+ /* detour code buffer */
+ kprobe_opcode_t *insn;
+ /* the size of instructions copied to detour code buffer */
+ size_t size;
+};
+
+/* Return true (!0) if optinsn is prepared for optimization. */
+static inline int arch_prepared_optinsn(struct arch_optimized_insn *optinsn)
+{
+ return optinsn->size;
+}
+
struct prev_kprobe {
struct kprobe *kp;
unsigned long status;
diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c
index f848984..9d2c873 100644
--- a/arch/x86/kernel/kprobes.c
+++ b/arch/x86/kernel/kprobes.c
@@ -106,16 +106,22 @@ struct kretprobe_blackpoint kretprobe_blacklist[] = {
};
const int kretprobe_blacklist_size = ARRAY_SIZE(kretprobe_blacklist);
-/* Insert a jump instruction at address 'from', which jumps to address 'to'.*/
-static void __kprobes set_jmp_op(void *from, void *to)
+static void __kprobes __synthesize_relative_insn(void *from, void *to, u8 op)
{
- struct __arch_jmp_op {
- char op;
+ struct __arch_relative_insn {
+ u8 op;
s32 raddr;
- } __attribute__((packed)) * jop;
- jop = (struct __arch_jmp_op *)from;
- jop->raddr = (s32)((long)(to) - ((long)(from) + 5));
- jop->op = RELATIVEJUMP_OPCODE;
+ } __attribute__((packed)) *insn;
+
+ insn = (struct __arch_relative_insn *)from;
+ insn->raddr = (s32)((long)(to) - ((long)(from) + 5));
+ insn->op = op;
+}
+
+/* Insert a jump instruction at address 'from', which jumps to address 'to'.*/
+static void __kprobes synthesize_reljump(void *from, void *to)
+{
+ __synthesize_relative_insn(from, to, RELATIVEJUMP_OPCODE);
}
/*
@@ -202,7 +208,7 @@ static int recover_probed_instruction(kprobe_opcode_t *buf, unsigned long addr)
/*
* Basically, kp->ainsn.insn has an original instruction.
* However, RIP-relative instruction can not do single-stepping
- * at different place, fix_riprel() tweaks the displacement of
+ * at different place, __copy_instruction() tweaks the displacement of
* that instruction. In that case, we can't recover the instruction
* from the kp->ainsn.insn.
*
@@ -284,21 +290,37 @@ static int __kprobes is_IF_modifier(kprobe_opcode_t *insn)
}
/*
- * Adjust the displacement if the instruction uses the %rip-relative
- * addressing mode.
+ * Copy an instruction and adjust the displacement if the instruction
+ * uses the %rip-relative addressing mode.
* If it does, Return the address of the 32-bit displacement word.
* If not, return null.
* Only applicable to 64-bit x86.
*/
-static void __kprobes fix_riprel(struct kprobe *p)
+static int __kprobes __copy_instruction(u8 *dest, u8 *src, int recover)
{
-#ifdef CONFIG_X86_64
struct insn insn;
- kernel_insn_init(&insn, p->ainsn.insn);
+ int ret;
+ kprobe_opcode_t buf[MAX_INSN_SIZE];
+ kernel_insn_init(&insn, src);
+ if (recover) {
+ insn_get_opcode(&insn);
+ if (insn.opcode.bytes[0] == BREAKPOINT_INSTRUCTION) {
+ ret = recover_probed_instruction(buf,
+ (unsigned long)src);
+ if (ret)
+ return 0;
+ kernel_insn_init(&insn, buf);
+ }
+ }
+ insn_get_length(&insn);
+ memcpy(dest, insn.kaddr, insn.length);
+
+#ifdef CONFIG_X86_64
if (insn_rip_relative(&insn)) {
s64 newdisp;
u8 *disp;
+ kernel_insn_init(&insn, dest);
insn_get_displacement(&insn);
/*
* The copied instruction uses the %rip-relative addressing
@@ -312,20 +334,23 @@ static void __kprobes fix_riprel(struct kprobe *p)
* extension of the original signed 32-bit displacement would
* have given.
*/
- newdisp = (u8 *) p->addr + (s64) insn.displacement.value -
- (u8 *) p->ainsn.insn;
+ newdisp = (u8 *) src + (s64) insn.displacement.value -
+ (u8 *) dest;
BUG_ON((s64) (s32) newdisp != newdisp); /* Sanity check. */
- disp = (u8 *) p->ainsn.insn + insn_offset_displacement(&insn);
+ disp = (u8 *) dest + insn_offset_displacement(&insn);
*(s32 *) disp = (s32) newdisp;
}
#endif
+ return insn.length;
}
static void __kprobes arch_copy_kprobe(struct kprobe *p)
{
- memcpy(p->ainsn.insn, p->addr, MAX_INSN_SIZE * sizeof(kprobe_opcode_t));
-
- fix_riprel(p);
+ /*
+ * Copy an instruction without recovering int3, because it will be
+ * put by another subsystem.
+ */
+ __copy_instruction(p->ainsn.insn, p->addr, 0);
if (can_boost(p->addr))
p->ainsn.boostable = 0;
@@ -414,9 +439,20 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
*sara = (unsigned long) &kretprobe_trampoline;
}
+#ifdef CONFIG_OPTPROBES
+static int __kprobes setup_detour_execution(struct kprobe *p,
+ struct pt_regs *regs,
+ int reenter);
+#else
+#define setup_detour_execution(p, regs, reenter) (0)
+#endif
+
static void __kprobes setup_singlestep(struct kprobe *p, struct pt_regs *regs,
struct kprobe_ctlblk *kcb, int reenter)
{
+ if (setup_detour_execution(p, regs, reenter))
+ return;
+
#if !defined(CONFIG_PREEMPT) || defined(CONFIG_FREEZER)
if (p->ainsn.boostable == 1 && !p->post_handler) {
/* Boost up -- we can execute copied instructions directly */
@@ -812,8 +848,8 @@ static void __kprobes resume_execution(struct kprobe *p,
* These instructions can be executed directly if it
* jumps back to correct address.
*/
- set_jmp_op((void *)regs->ip,
- (void *)orig_ip + (regs->ip - copy_ip));
+ synthesize_reljump((void *)regs->ip,
+ (void *)orig_ip + (regs->ip - copy_ip));
p->ainsn.boostable = 1;
} else {
p->ainsn.boostable = -1;
@@ -1040,6 +1076,346 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
return 0;
}
+
+#ifdef CONFIG_OPTPROBES
+
+/* Insert a call instruction at address 'from', which calls address 'to'.*/
+static void __kprobes synthesize_relcall(void *from, void *to)
+{
+ __synthesize_relative_insn(from, to, RELATIVECALL_OPCODE);
+}
+
+/* Insert a move instruction which sets a pointer to eax/rdi (1st arg). */
+static void __kprobes synthesize_set_arg1(kprobe_opcode_t *addr,
+ unsigned long val)
+{
+#ifdef CONFIG_X86_64
+ *addr++ = 0x48;
+ *addr++ = 0xbf;
+#else
+ *addr++ = 0xb8;
+#endif
+ *(unsigned long *)addr = val;
+}
+
+void __kprobes kprobes_optinsn_template_holder(void)
+{
+ asm volatile (
+ ".global optprobe_template_entry\n"
+ "optprobe_template_entry: \n"
+#ifdef CONFIG_X86_64
+ /* We don't bother saving the ss register */
+ " pushq %rsp\n"
+ " pushfq\n"
+ SAVE_REGS_STRING
+ " movq %rsp, %rsi\n"
+ ".global optprobe_template_val\n"
+ "optprobe_template_val: \n"
+ ASM_NOP5
+ ASM_NOP5
+ ".global optprobe_template_call\n"
+ "optprobe_template_call: \n"
+ ASM_NOP5
+ /* Move flags to rsp */
+ " movq 144(%rsp), %rdx\n"
+ " movq %rdx, 152(%rsp)\n"
+ RESTORE_REGS_STRING
+ /* Skip flags entry */
+ " addq $8, %rsp\n"
+ " popfq\n"
+#else /* CONFIG_X86_32 */
+ " pushf\n"
+ SAVE_REGS_STRING
+ " movl %esp, %edx\n"
+ ".global optprobe_template_val\n"
+ "optprobe_template_val: \n"
+ ASM_NOP5
+ ".global optprobe_template_call\n"
+ "optprobe_template_call: \n"
+ ASM_NOP5
+ RESTORE_REGS_STRING
+ " addl $4, %esp\n" /* skip cs */
+ " popf\n"
+#endif
+ ".global optprobe_template_end\n"
+ "optprobe_template_end: \n");
+}
+
+#define TMPL_MOVE_IDX \
+ ((long)&optprobe_template_val - (long)&optprobe_template_entry)
+#define TMPL_CALL_IDX \
+ ((long)&optprobe_template_call - (long)&optprobe_template_entry)
+#define TMPL_END_IDX \
+ ((long)&optprobe_template_end - (long)&optprobe_template_entry)
+
+#define INT3_SIZE sizeof(kprobe_opcode_t)
+
+/* Optimized kprobe call back function: called from optinsn */
+static void __kprobes optimized_callback(struct optimized_kprobe *op,
+ struct pt_regs *regs)
+{
+ struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+
+ preempt_disable();
+ if (kprobe_running()) {
+ kprobes_inc_nmissed_count(&op->kp);
+ } else {
+ /* Save skipped registers */
+#ifdef CONFIG_X86_64
+ regs->cs = __KERNEL_CS;
+#else
+ regs->cs = __KERNEL_CS | get_kernel_rpl();
+ regs->gs = 0;
+#endif
+ regs->ip = (unsigned long)op->kp.addr + INT3_SIZE;
+ regs->orig_ax = ~0UL;
+
+ __get_cpu_var(current_kprobe) = &op->kp;
+ kcb->kprobe_status = KPROBE_HIT_ACTIVE;
+ opt_pre_handler(&op->kp, regs);
+ __get_cpu_var(current_kprobe) = NULL;
+ }
+ preempt_enable_no_resched();
+}
+
+static int __kprobes copy_optimized_instructions(u8 *dest, u8 *src)
+{
+ int len = 0, ret;
+
+ while (len < RELATIVEJUMP_SIZE) {
+ ret = __copy_instruction(dest + len, src + len, 1);
+ if (!ret || !can_boost(dest + len))
+ return -EINVAL;
+ len += ret;
+ }
+
+ return len;
+}
+
+/* Check whether insn is indirect jump */
+static int __kprobes insn_is_indirect_jump(struct insn *insn)
+{
+ return (insn->opcode.bytes[0] == 0xff ||
+ insn->opcode.bytes[0] == 0xea);
+}
+
+/* Check whether insn jumps into specified address range */
+static int insn_jump_into_range(struct insn *insn, unsigned long start, int len)
+{
+ unsigned long target = 0;
+
+ switch (insn->opcode.bytes[0]) {
+ case 0xe0: /* loopne */
+ case 0xe1: /* loope */
+ case 0xe2: /* loop */
+ case 0xe3: /* jcxz */
+ case 0xe9: /* near relative jump */
+ case 0xeb: /* short relative jump */
+ break;
+ case 0x0f:
+ if ((insn->opcode.bytes[1] & 0xf0) == 0x80) /* jcc near */
+ break;
+ return 0;
+ default:
+ if ((insn->opcode.bytes[0] & 0xf0) == 0x70) /* jcc short */
+ break;
+ return 0;
+ }
+ target = (unsigned long)insn->next_byte + insn->immediate.value;
+
+ return (start <= target && target <= start + len);
+}
+
+/* Decode whole function to ensure any instructions don't jump into target */
+static int __kprobes can_optimize(unsigned long paddr)
+{
+ int ret;
+ unsigned long addr, size = 0, offset = 0;
+ struct insn insn;
+ kprobe_opcode_t buf[MAX_INSN_SIZE];
+ /* Dummy buffers for lookup_symbol_attrs */
+ static char __dummy_buf[KSYM_NAME_LEN];
+
+ /* Lookup symbol including addr */
+ if (!kallsyms_lookup(paddr, &size, &offset, NULL, __dummy_buf))
+ return 0;
+
+ /* Check there is enough space for a relative jump. */
+ if (size - offset < RELATIVEJUMP_SIZE)
+ return 0;
+
+ /* Decode instructions */
+ addr = paddr - offset;
+ while (addr < paddr - offset + size) { /* Decode until function end */
+ if (search_exception_tables(addr))
+ /*
+ * Since some fixup code will jumps into this function,
+ * we can't optimize kprobe in this function.
+ */
+ return 0;
+ kernel_insn_init(&insn, (void *)addr);
+ insn_get_opcode(&insn);
+ if (insn.opcode.bytes[0] == BREAKPOINT_INSTRUCTION) {
+ ret = recover_probed_instruction(buf, addr);
+ if (ret)
+ return 0;
+ kernel_insn_init(&insn, buf);
+ }
+ insn_get_length(&insn);
+ /* Recover address */
+ insn.kaddr = (void *)addr;
+ insn.next_byte = (void *)(addr + insn.length);
+ /* Check any instructions don't jump into target */
+ if (insn_is_indirect_jump(&insn) ||
+ insn_jump_into_range(&insn, paddr + INT3_SIZE,
+ RELATIVE_ADDR_SIZE))
+ return 0;
+ addr += insn.length;
+ }
+
+ return 1;
+}
+
+/* Check optimized_kprobe can actually be optimized. */
+int __kprobes arch_check_optimized_kprobe(struct optimized_kprobe *op)
+{
+ int i;
+
+ for (i = 1; i < op->optinsn.size; i++)
+ if (get_kprobe(op->kp.addr + i))
+ return -EEXIST;
+
+ return 0;
+}
+
+/* Check the addr is within the optimized instructions. */
+int __kprobes arch_within_optimized_kprobe(struct optimized_kprobe *op,
+ unsigned long addr)
+{
+ return ((unsigned long)op->kp.addr <= addr &&
+ (unsigned long)op->kp.addr + op->optinsn.size > addr);
+}
+
+/* Free optimized instruction slot */
+static __kprobes
+void __arch_remove_optimized_kprobe(struct optimized_kprobe *op, int dirty)
+{
+ if (op->optinsn.insn) {
+ free_optinsn_slot(op->optinsn.insn, dirty);
+ op->optinsn.insn = NULL;
+ op->optinsn.size = 0;
+ }
+}
+
+void __kprobes arch_remove_optimized_kprobe(struct optimized_kprobe *op)
+{
+ __arch_remove_optimized_kprobe(op, 1);
+}
+
+/*
+ * Copy replacing target instructions
+ * Target instructions MUST be relocatable (checked inside)
+ */
+int __kprobes arch_prepare_optimized_kprobe(struct optimized_kprobe *op)
+{
+ u8 *buf;
+ int ret;
+
+ if (!can_optimize((unsigned long)op->kp.addr))
+ return -EILSEQ;
+
+ op->optinsn.insn = get_optinsn_slot();
+ if (!op->optinsn.insn)
+ return -ENOMEM;
+
+ buf = (u8 *)op->optinsn.insn;
+
+ /* Copy instructions into the out-of-line buffer */
+ ret = copy_optimized_instructions(buf + TMPL_END_IDX, op->kp.addr);
+ if (ret < 0) {
+ __arch_remove_optimized_kprobe(op, 0);
+ return ret;
+ }
+ op->optinsn.size = ret;
+
+ /* Backup instructions which will be replaced by jump address */
+ memcpy(op->optinsn.copied_insn, op->kp.addr + INT3_SIZE,
+ RELATIVE_ADDR_SIZE);
+
+ /* Copy arch-dep-instance from template */
+ memcpy(buf, &optprobe_template_entry, TMPL_END_IDX);
+
+ /* Set probe information */
+ synthesize_set_arg1(buf + TMPL_MOVE_IDX, (unsigned long)op);
+
+ /* Set probe function call */
+ synthesize_relcall(buf + TMPL_CALL_IDX, optimized_callback);
+
+ /* Set returning jmp instruction at the tail of out-of-line buffer */
+ synthesize_reljump(buf + TMPL_END_IDX + op->optinsn.size,
+ (u8 *)op->kp.addr + op->optinsn.size);
+
+ flush_icache_range((unsigned long) buf,
+ (unsigned long) buf + TMPL_END_IDX +
+ op->optinsn.size + RELATIVEJUMP_SIZE);
+ return 0;
+}
+
+/* Replace a breakpoint (int3) with a relative jump. */
+int __kprobes arch_optimize_kprobe(struct optimized_kprobe *op)
+{
+ unsigned char jmp_code[RELATIVEJUMP_SIZE];
+ s32 rel = (s32)((long)op->optinsn.insn -
+ ((long)op->kp.addr + RELATIVEJUMP_SIZE));
+
+ /* Check if the address gap is in 2GB range. */
+ if ((long)op->kp.addr + RELATIVEJUMP_SIZE + rel !=
+ (long)op->optinsn.insn)
+ return -EINVAL;
+
+ jmp_code[0] = RELATIVEJUMP_OPCODE;
+ *(s32 *)(&jmp_code[1]) = rel;
+
+ /*
+ * text_poke_smp doesn't support NMI/MCE code modifying.
+ * However, since kprobes itself also doesn't support NMI/MCE
+ * code probing, it's not a problem.
+ */
+ text_poke_smp(op->kp.addr, jmp_code, RELATIVEJUMP_SIZE);
+ return 0;
+}
+
+/* Replace a relative jump with a breakpoint (int3). */
+void __kprobes arch_unoptimize_kprobe(struct optimized_kprobe *op)
+{
+ u8 buf[RELATIVEJUMP_SIZE];
+
+ /* Set int3 to first byte for kprobes */
+ buf[0] = BREAKPOINT_INSTRUCTION;
+ memcpy(buf + 1, op->optinsn.copied_insn, RELATIVE_ADDR_SIZE);
+ text_poke_smp(op->kp.addr, buf, RELATIVEJUMP_SIZE);
+}
+
+static int __kprobes setup_detour_execution(struct kprobe *p,
+ struct pt_regs *regs,
+ int reenter)
+{
+ struct optimized_kprobe *op;
+
+ if (p->flags & KPROBE_FLAG_OPTIMIZED) {
+ /* This kprobe is really able to run optimized path. */
+ op = container_of(p, struct optimized_kprobe, kp);
+ /* Detour through copied instructions */
+ regs->ip = (unsigned long)op->optinsn.insn + TMPL_END_IDX;
+ if (!reenter)
+ reset_current_kprobe();
+ preempt_enable_no_resched();
+ return 1;
+ }
+ return 0;
+}
+#endif
+
int __init arch_init_kprobes(void)
{
return 0;
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Add documentations about kprobe jump optimization to Documentation/kprobes.txt.
Changes in v8:
- Update documentation and benchmark results.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
Documentation/kprobes.txt | 191 ++++++++++++++++++++++++++++++++++++++++++---
1 files changed, 178 insertions(+), 13 deletions(-)
diff --git a/Documentation/kprobes.txt b/Documentation/kprobes.txt
index 053037a..48af218 100644
--- a/Documentation/kprobes.txt
+++ b/Documentation/kprobes.txt
@@ -1,6 +1,7 @@
Title : Kernel Probes (Kprobes)
Authors : Jim Keniston <[email protected]>
: Prasanna S Panchamukhi <[email protected]>
+ : Masami Hiramatsu <[email protected]>
CONTENTS
@@ -14,6 +15,7 @@ CONTENTS
8. Kprobes Example
9. Jprobes Example
10. Kretprobes Example
+11. Optimization Example
Appendix A: The kprobes debugfs interface
1. Concepts: Kprobes, Jprobes, Return Probes
@@ -42,13 +44,13 @@ registration/unregistration of a group of *probes. These functions
can speed up unregistration process when you have to unregister
a lot of probes at once.
-The next three subsections explain how the different types of
-probes work. They explain certain things that you'll need to
-know in order to make the best use of Kprobes -- e.g., the
-difference between a pre_handler and a post_handler, and how
-to use the maxactive and nmissed fields of a kretprobe. But
-if you're in a hurry to start using Kprobes, you can skip ahead
-to section 2.
+The next four subsections explain how the different types of
+probes work and how the optimization works. They explain certain
+things that you'll need to know in order to make the best use of
+Kprobes -- e.g., the difference between a pre_handler and
+a post_handler, and how to use the maxactive and nmissed fields of
+a kretprobe. But if you're in a hurry to start using Kprobes, you
+can skip ahead to section 2.
1.1 How Does a Kprobe Work?
@@ -161,13 +163,109 @@ In case probed function is entered but there is no kretprobe_instance
object available, then in addition to incrementing the nmissed count,
the user entry_handler invocation is also skipped.
+1.4 How Does the Optimization Work?
+
+ If you configured kernel with CONFIG_OPTPROBES=y (currently this option is
+supported on x86/x86-64, non-preemptive kernel) and
+"debug.kprobes_optimization" sysctl sets 1, kprobes tries to use a
+jump instruction instead of breakpoint instruction automatically.
+
+1.4.1 Init a Kprobe
+
+ Before preparing optimization, Kprobes inserts original(user-defined)
+kprobe on the specified address. So, even if the kprobe is not
+possible to be optimized, it just uses a normal kprobe.
+
+1.4.2 Safety check
+
+ First, Kprobes gets the address of probed function and checks whether the
+optimized region, which will be replaced by a jump instruction, does NOT
+straddle the function boundary, because if the optimized region reaches the
+next function, its caller causes unexpected results.
+ Next, Kprobes decodes whole body of probed function and checks there is
+NO indirect jump, NO instruction which will cause exception by checking
+exception_tables (this will jump to fixup code and fixup code jumps into
+same function body) and NO near jump which jumps into the optimized region
+(except the 1st byte of jump), because if some jump instruction jumps
+into the middle of another instruction, it causes unexpected results too.
+ Kprobes also measures the length of instructions which will be replaced
+by a jump instruction, because a jump instruction is longer than 1 byte,
+it may replaces multiple instructions, and it checks whether those
+instructions can be executed out-of-line.
+
+1.4.3 Preparing detour buffer
+
+ Then, Kprobes prepares "detour" buffer, which contains exception emulating
+code (push/pop registers, call handler), copied instructions(Kprobes copies
+instructions which will be replaced by a jump, to the detour buffer), and
+a jump which jumps back to the original execution path.
+
+1.4.4 Pre-optimization
+
+ After preparing detour buffer, Kprobes checks that the probe is *NOT* in
+the below cases;
+ - The probe has either break_handler or post_handler.
+ - Other probes are probing the instructions which will be replaced by
+ a jump instruction.
+ - The probe is disabled.
+In above cases, Kprobes just doesn't start optimizating the probe.
+
+ If the kprobe can be optimized, Kprobes enqueues the kprobe to optimizing
+list and kicks kprobe-optimizer workqueue to optimize it. To wait other
+optimized probes, kprobe-optimizer will delay to work.
+ When the optimized-kprobe is hit before optimization, its handler changes
+IP(instruction pointer) to copied code and exits. So, the instructions which
+were copied to detour buffer are executed on the detour buffer.
+
+1.4.5 Optimization
+
+ Kprobe-optimizer doesn't start instruction-replacing soon, it waits
+ synchronize_sched for safety, because some processors are possible to be
+ interrupted on the middle of instruction series (2nd or Nth instruction)
+ which will be replaced by a jump instruction(*).
+ As you know, synchronize_sched() can ensure that all interruptions which
+ were executed when synchronize_sched() was called are done, only if
+ CONFIG_PREEMPT=n. So, this version supports only the kernel with
+ CONFIG_PREEMPT=n.(**)
+ After that, kprobe-optimizer calls stop_machine() to replace probed-
+ instructions with a jump instruction by using text_poke_smp().
+
+1.4.6 Unoptimization
+ When unregistering, disabling kprobe or being blocked by other kprobe,
+ an optimized-kprobe will be unoptimized. Before kprobe-optimizer runs,
+ the kprobe is just dequeued from the optimized list. When the optimization
+ has been done, it replaces a jump with int3 breakpoint and original code
+ by using text_poke_smp().
+
+(*)Please imagine that 2nd instruction is interrupted and
+optimizer replaces the 2nd instruction with jump *address*
+while the interrupt handler is running. When the interrupt
+returns to original address, there is no valid instructions
+and it causes unexpected result.
+
+(**)This optimization-safety checking may be replaced with stop-machine
+method which ksplice is done for supporting CONFIG_PREEMPT=y kernel.
+
+NOTE for geeks:
+The jump optimization changes the kprobe's pre_handler behavior.
+Without optimization, pre_handler can change kernel execution path by
+changing regs->ip and return 1. However, after optimizing the probe,
+that modification is ignored. Thus, if you'd like to tweak kernel
+execution path, you need to avoid optimization. In that case, you can
+choose either,
+ - Set empty function to post_handler or break_handler.
+ or
+ - Config CONFIG_OPTPROBES=n.
+ or
+ - Execute 'sysctl -w debug.kprobes_optimization=n'
+
2. Architectures Supported
Kprobes, jprobes, and return probes are implemented on the following
architectures:
-- i386
-- x86_64 (AMD-64, EM64T)
+- i386 (Supports jump optimization)
+- x86_64 (AMD-64, EM64T) (Supports jump optimization)
- ppc64
- ia64 (Does not support probes on instruction slot1.)
- sparc64 (Return probes not yet implemented.)
@@ -193,6 +291,10 @@ it useful to "Compile the kernel with debug info" (CONFIG_DEBUG_INFO),
so you can use "objdump -d -l vmlinux" to see the source-to-object
code mapping.
+If you want to reduce probing overhead, set "Kprobes jump optimization
+support" (CONFIG_OPTPROBES) to "y". You can find this option under
+"Kprobes" line.
+
4. API Reference
The Kprobes API includes a "register" function and an "unregister"
@@ -387,9 +489,12 @@ the probe which has been registered.
5. Kprobes Features and Limitations
-Kprobes allows multiple probes at the same address. Currently,
-however, there cannot be multiple jprobes on the same function at
-the same time.
+Kprobes allows multiple probes at the same address even if it is optimized.
+Currently, however, there cannot be multiple jprobes on the same function
+at the same time. And also, optimized kprobes can not invoke the
+post_handler and the break_handler. So if you attempt to install the probe
+which has the the post_handler or the break_handler at the same address of
+an optimized kprobe, the probe will be unoptimized automatically.
In general, you can install a probe anywhere in the kernel.
In particular, you can probe interrupt handlers. Known exceptions
@@ -453,6 +558,37 @@ reason, Kprobes doesn't support return probes (or kprobes or jprobes)
on the x86_64 version of __switch_to(); the registration functions
return -EINVAL.
+On x86/x86-64, since the Jump Optimization of Kprobes modifies instructions
+widely, there are some limitations for optimization. To explain it,
+we introduce some terminology. Image certain binary line which is
+constructed by 2 byte instruction, 2byte instruction and 3byte instruction.
+
+ IA
+ |
+[-2][-1][0][1][2][3][4][5][6][7]
+ [ins1][ins2][ ins3 ]
+ [<- DCR ->]
+ [<- JTPR ->]
+
+ins1: 1st Instruction
+ins2: 2nd Instruction
+ins3: 3rd Instruction
+IA: Insertion Address
+JTPR: Jump Target Prohibition Region
+DCR: Detoured Code Region
+
+The instructions in DCR are copied to the out-of-line buffer
+of the djprobe instance, because the bytes in JTPR are replaced by
+a jump instruction. So, there are several limitations.
+
+a) The instructions in DCR must be relocatable.
+b) The instructions in DCR must not include call instruction.
+c) JTPR must not be targeted by any jump or call instruction.
+d) DCR must not straddle the border betweeen functions.
+
+Anyway, these limitations are checked by in-kernel instruction decoder,
+so you don't need to care about that.
+
6. Probe Overhead
On a typical CPU in use in 2005, a kprobe hit takes 0.5 to 1.0
@@ -476,6 +612,19 @@ k = 0.49 usec; j = 0.76; r = 0.80; kr = 0.82; jr = 1.07
ppc64: POWER5 (gr), 1656 MHz (SMT disabled, 1 virtual CPU per physical CPU)
k = 0.77 usec; j = 1.31; r = 1.26; kr = 1.45; jr = 1.99
+6.1 Optimized Probe Overhead
+
+Typically, an optimized kprobe hit takes 0.07 to 0.1 microseconds to
+process. Here are sample overhead figures (in usec) for x86-64 architectures.
+k = unoptimized kprobe, b = boosted(single-step skipped), o = optimized kprobe,
+r = unoptimized kretprobe, rb = boosted kretprobe, ro = optimized kretprobe.
+
+i386: Intel(R) Xeon(R) E5410, 2.33GHz, 4656.90 bogomips
+k = 0.80 usec; b = 0.33; o = 0.05; r = 1.10; rb = 0.61; ro = 0.33
+
+x86-64: Intel(R) Xeon(R) E5410, 2.33GHz, 4656.90 bogomips
+k = 0.99 usec; b = 0.43; o = 0.06; r = 1.24; rb = 0.68; ro = 0.30
+
7. TODO
a. SystemTap (http://sourceware.org/systemtap): Provides a simplified
@@ -523,7 +672,8 @@ is also specified. Following columns show probe status. If the probe is on
a virtual address that is no longer valid (module init sections, module
virtual addresses that correspond to modules that've been unloaded),
such probes are marked with [GONE]. If the probe is temporarily disabled,
-such probes are marked with [DISABLED].
+such probes are marked with [DISABLED]. If the probe is optimized, it is
+marked with [OPTIMIZED].
/sys/kernel/debug/kprobes/enabled: Turn kprobes ON/OFF forcibly.
@@ -533,3 +683,18 @@ registered probes will be disarmed, till such time a "1" is echoed to this
file. Note that this knob just disarms and arms all kprobes and doesn't
change each probe's disabling state. This means that disabled kprobes (marked
[DISABLED]) will be not enabled if you turn ON all kprobes by this knob.
+
+
+Appendix B: The kprobes sysctl interface
+
+/proc/sys/debug/kprobes-optimization: Turn kprobes optimization ON/OFF.
+
+When CONFIG_OPTPROBES=y, this sysctl interface appears and it provides a knob
+to globally and forcibly turn the jump optimization ON or OFF. By default,
+jump optimization is allowed(ON). By echoing "0" to this file or By setting
+0 to "debug.kprobes_optimization" via sysctl, all optimized probes will be
+unoptimized. And new probes registered after that will not be optimized.
+Note that this knob *Changes* the optimized state. This means that optimized
+probes (marked [OPTIMIZED]) will be unoptimized ([OPTIMIZED] tag will be
+removed). And after the knob is turned on, it will be optimized again.
+
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
Make insn_slot framework support various size slots.
Current insn_slot just supports one-size instruction buffer slot. However,
kprobes jump optimization needs larger size buffers.
Signed-off-by: Masami Hiramatsu <[email protected]>
Cc: Ananth N Mavinakayanahalli <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jim Keniston <[email protected]>
Cc: Srikar Dronamraju <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Steven Rostedt <[email protected]>
Cc: Frederic Weisbecker <[email protected]>
Cc: H. Peter Anvin <[email protected]>
Cc: Anders Kaseorg <[email protected]>
Cc: Tim Abbott <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Jason Baron <[email protected]>
Cc: Mathieu Desnoyers <[email protected]>
---
kernel/kprobes.c | 104 ++++++++++++++++++++++++++++++++++--------------------
1 files changed, 65 insertions(+), 39 deletions(-)
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index b7df302..57a2a23 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -103,26 +103,42 @@ static struct kprobe_blackpoint kprobe_blacklist[] = {
* stepping on the instruction on a vmalloced/kmalloced/data page
* is a recipe for disaster
*/
-#define INSNS_PER_PAGE (PAGE_SIZE/(MAX_INSN_SIZE * sizeof(kprobe_opcode_t)))
-
struct kprobe_insn_page {
struct list_head list;
kprobe_opcode_t *insns; /* Page of instruction slots */
- char slot_used[INSNS_PER_PAGE];
int nused;
int ngarbage;
+ char slot_used[];
+};
+
+#define KPROBE_INSN_PAGE_SIZE(slots) \
+ (offsetof(struct kprobe_insn_page, slot_used) + \
+ (sizeof(char) * (slots)))
+
+struct kprobe_insn_cache {
+ struct list_head pages; /* list of kprobe_insn_page */
+ size_t insn_size; /* size of instruction slot */
+ int nr_garbage;
};
+static int slots_per_page(struct kprobe_insn_cache *c)
+{
+ return PAGE_SIZE/(c->insn_size * sizeof(kprobe_opcode_t));
+}
+
enum kprobe_slot_state {
SLOT_CLEAN = 0,
SLOT_DIRTY = 1,
SLOT_USED = 2,
};
-static DEFINE_MUTEX(kprobe_insn_mutex); /* Protects kprobe_insn_pages */
-static LIST_HEAD(kprobe_insn_pages);
-static int kprobe_garbage_slots;
-static int collect_garbage_slots(void);
+static DEFINE_MUTEX(kprobe_insn_mutex); /* Protects kprobe_insn_slots */
+static struct kprobe_insn_cache kprobe_insn_slots = {
+ .pages = LIST_HEAD_INIT(kprobe_insn_slots.pages),
+ .insn_size = MAX_INSN_SIZE,
+ .nr_garbage = 0,
+};
+static int __kprobes collect_garbage_slots(struct kprobe_insn_cache *c);
static int __kprobes check_safety(void)
{
@@ -152,32 +168,33 @@ loop_end:
* __get_insn_slot() - Find a slot on an executable page for an instruction.
* We allocate an executable page if there's no room on existing ones.
*/
-static kprobe_opcode_t __kprobes *__get_insn_slot(void)
+static kprobe_opcode_t __kprobes *__get_insn_slot(struct kprobe_insn_cache *c)
{
struct kprobe_insn_page *kip;
retry:
- list_for_each_entry(kip, &kprobe_insn_pages, list) {
- if (kip->nused < INSNS_PER_PAGE) {
+ list_for_each_entry(kip, &c->pages, list) {
+ if (kip->nused < slots_per_page(c)) {
int i;
- for (i = 0; i < INSNS_PER_PAGE; i++) {
+ for (i = 0; i < slots_per_page(c); i++) {
if (kip->slot_used[i] == SLOT_CLEAN) {
kip->slot_used[i] = SLOT_USED;
kip->nused++;
- return kip->insns + (i * MAX_INSN_SIZE);
+ return kip->insns + (i * c->insn_size);
}
}
- /* Surprise! No unused slots. Fix kip->nused. */
- kip->nused = INSNS_PER_PAGE;
+ /* kip->nused is broken. Fix it. */
+ kip->nused = slots_per_page(c);
+ WARN_ON(1);
}
}
/* If there are any garbage slots, collect it and try again. */
- if (kprobe_garbage_slots && collect_garbage_slots() == 0) {
+ if (c->nr_garbage && collect_garbage_slots(c) == 0)
goto retry;
- }
- /* All out of space. Need to allocate a new page. Use slot 0. */
- kip = kmalloc(sizeof(struct kprobe_insn_page), GFP_KERNEL);
+
+ /* All out of space. Need to allocate a new page. */
+ kip = kmalloc(KPROBE_INSN_PAGE_SIZE(slots_per_page(c)), GFP_KERNEL);
if (!kip)
return NULL;
@@ -192,20 +209,23 @@ static kprobe_opcode_t __kprobes *__get_insn_slot(void)
return NULL;
}
INIT_LIST_HEAD(&kip->list);
- list_add(&kip->list, &kprobe_insn_pages);
- memset(kip->slot_used, SLOT_CLEAN, INSNS_PER_PAGE);
+ memset(kip->slot_used, SLOT_CLEAN, slots_per_page(c));
kip->slot_used[0] = SLOT_USED;
kip->nused = 1;
kip->ngarbage = 0;
+ list_add(&kip->list, &c->pages);
return kip->insns;
}
+
kprobe_opcode_t __kprobes *get_insn_slot(void)
{
- kprobe_opcode_t *ret;
+ kprobe_opcode_t *ret = NULL;
+
mutex_lock(&kprobe_insn_mutex);
- ret = __get_insn_slot();
+ ret = __get_insn_slot(&kprobe_insn_slots);
mutex_unlock(&kprobe_insn_mutex);
+
return ret;
}
@@ -221,7 +241,7 @@ static int __kprobes collect_one_slot(struct kprobe_insn_page *kip, int idx)
* so as not to have to set it up again the
* next time somebody inserts a probe.
*/
- if (!list_is_singular(&kprobe_insn_pages)) {
+ if (!list_is_singular(&kip->list)) {
list_del(&kip->list);
module_free(NULL, kip->insns);
kfree(kip);
@@ -231,7 +251,7 @@ static int __kprobes collect_one_slot(struct kprobe_insn_page *kip, int idx)
return 0;
}
-static int __kprobes collect_garbage_slots(void)
+static int __kprobes collect_garbage_slots(struct kprobe_insn_cache *c)
{
struct kprobe_insn_page *kip, *next;
@@ -239,42 +259,48 @@ static int __kprobes collect_garbage_slots(void)
if (check_safety())
return -EAGAIN;
- list_for_each_entry_safe(kip, next, &kprobe_insn_pages, list) {
+ list_for_each_entry_safe(kip, next, &c->pages, list) {
int i;
if (kip->ngarbage == 0)
continue;
kip->ngarbage = 0; /* we will collect all garbages */
- for (i = 0; i < INSNS_PER_PAGE; i++) {
+ for (i = 0; i < slots_per_page(c); i++) {
if (kip->slot_used[i] == SLOT_DIRTY &&
collect_one_slot(kip, i))
break;
}
}
- kprobe_garbage_slots = 0;
+ c->nr_garbage = 0;
return 0;
}
-void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty)
+static void __kprobes __free_insn_slot(struct kprobe_insn_cache *c,
+ kprobe_opcode_t *slot, int dirty)
{
struct kprobe_insn_page *kip;
- mutex_lock(&kprobe_insn_mutex);
- list_for_each_entry(kip, &kprobe_insn_pages, list) {
- if (kip->insns <= slot &&
- slot < kip->insns + (INSNS_PER_PAGE * MAX_INSN_SIZE)) {
- int i = (slot - kip->insns) / MAX_INSN_SIZE;
+ list_for_each_entry(kip, &c->pages, list) {
+ long idx = ((long)slot - (long)kip->insns) / c->insn_size;
+ if (idx >= 0 && idx < slots_per_page(c)) {
+ WARN_ON(kip->slot_used[idx] != SLOT_USED);
if (dirty) {
- kip->slot_used[i] = SLOT_DIRTY;
+ kip->slot_used[idx] = SLOT_DIRTY;
kip->ngarbage++;
+ if (++c->nr_garbage > slots_per_page(c))
+ collect_garbage_slots(c);
} else
- collect_one_slot(kip, i);
- break;
+ collect_one_slot(kip, idx);
+ return;
}
}
+ /* Could not free this slot. */
+ WARN_ON(1);
+}
- if (dirty && ++kprobe_garbage_slots > INSNS_PER_PAGE)
- collect_garbage_slots();
-
+void __kprobes free_insn_slot(kprobe_opcode_t * slot, int dirty)
+{
+ mutex_lock(&kprobe_insn_mutex);
+ __free_insn_slot(&kprobe_insn_slots, slot, dirty);
mutex_unlock(&kprobe_insn_mutex);
}
#endif
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]
* Masami Hiramatsu ([email protected]) wrote:
> Hi,
>
> Here are the patchset of the kprobes jump optimization v8
> (a.k.a. Djprobe). This version is just moving onto
> 2.6.33-rc4-tip. Ingo, I assume its a good timing to
> push this code onto -tip tree (maybe developing branch?),
> since people can test it with perf-probe.
>
> I've decided to make a separated series of patches of
> jump optimization with text_poke_smp() which is
> 'officially' supported on Intel's processors.
> So, this version of patches are just updated against
> the latest tip/master, no other updates are included.
>
> I know that int3-bypassing method (text_poke_fixup())
> is currently unofficially believed as safe. But we
> need to get more official answers from x86 vendors.
> Moreover, we need to tweak entry_*.S for preventing
> recursive NMI, because int3 inside NMI handler will
> unblock NMI blocking. I'd like to push it after this
> series of patches are merged.
>
> Anyway, thanks Mathieu and Peter, for helping me to
> implement it and organizing discussion points about
> int3-bypass XMC!
>
> These patches can be applied on the latest -tip.
>
> Changes in v8:
> - Update patches against the latest tip/master.
> - Drop text_poke_fixup() related patches.
> - Update benchmark results and add jprobes and kprobe(post-handler)
> results.
>
> And kprobe stress test didn't found any regressions - from kprobes,
> under kvm/x86.
>
> TODO:
> - Support NMI-safe int3-bypassing text_poke.
Please have a look at:
"x86 NMI-safe INT3 and Page Fault"
http://git.kernel.org/?p=linux/kernel/git/compudj/linux-2.6-lttng.git;a=commit;h=90516e3c718e0502f6f2eb616fad4447645ca47d
and
"x86_64 page fault NMI-safe"
http://git.kernel.org/?p=linux/kernel/git/compudj/linux-2.6-lttng.git;a=commit;h=ad1bf11a68c35a44edd8d686a0842896f408e17c
That turns this TODO into the "done" section ;)
I've been using these patches in the lttng tree for 1-2 years.
Thanks,
Mathieu
> - Support preemptive kernel (by stack unwinding and checking address).
>
>
> Jump Optimized Kprobes
> ======================
> o Concept
> Kprobes uses the int3 breakpoint instruction on x86 for instrumenting
> probes into running kernel. Jump optimization allows kprobes to replace
> breakpoint with a jump instruction for reducing probing overhead drastically.
>
> o Performance
> An optimized kprobe 5 times faster than a kprobe.
>
> Optimizing probes gains its performance. Usually, a kprobe hit takes
> 0.5 to 1.0 microseconds to process. On the other hand, a jump optimized
> probe hit takes less than 0.1 microseconds (actual number depends on the
> processor). Here is a sample overheads.
>
> Intel(R) Xeon(R) CPU E5410 @ 2.33GHz
> (without debugging options, with text_poke_smp patch, 2.6.33-rc4-tip+)
>
> x86-32 x86-64
> kprobe: 0.80us 0.99us
> kprobe+booster: 0.33us 0.43us
> kprobe+optimized: 0.05us 0.06us
> kprobe(post-handler): 0.81us 1.00us
>
> kretprobe : 1.10us 1.24us
> kretprobe+booster: 0.61us 0.68us
> kretprobe+optimized: 0.33us 0.30us
>
> jprobe: 1.37us 1.67us
> jprobe+booster: 0.80us 1.10us
>
> (booster skips single-stepping, kprobe with post handler
> isn't boosted/optimized, and jprobe isn't optimized.)
>
> Note that jump optimization also consumes more memory, but not so much.
> It just uses ~200 bytes, so, even if you use ~10,000 probes, it just
> consumes a few MB.
>
>
> o Usage
> Set CONFIG_OPTPROBES=y when building a kernel, then all *probes will be
> optimized if possible.
>
> Kprobes decodes probed function and checks whether the target instructions
> can be optimized(replaced with a jump) safely. If it can't be, Kprobes just
> doesn't optimize it.
>
>
> o Optimization
> Before preparing optimization, Kprobes inserts original(user-defined)
> kprobe on the specified address. So, even if the kprobe is not
> possible to be optimized, it just uses a normal kprobe.
>
> - Safety check
> First, Kprobes gets the address of probed function and checks whether the
> optimized region, which will be replaced by a jump instruction, does NOT
> straddle the function boundary, because if the optimized region reaches the
> next function, its caller causes unexpected results.
> Next, Kprobes decodes whole body of probed function and checks there is
> NO indirect jump, NO instruction which will cause exception by checking
> exception_tables (this will jump to fixup code and fixup code jumps into
> same function body) and NO near jump which jumps into the optimized region
> (except the 1st byte of jump), because if some jump instruction jumps
> into the middle of another instruction, it causes unexpected results too.
> Kprobes also measures the length of instructions which will be replaced
> by a jump instruction, because a jump instruction is longer than 1 byte,
> it may replaces multiple instructions, and it checks whether those
> instructions can be executed out-of-line.
>
> - Preparing detour code
> Then, Kprobes prepares "detour" buffer, which contains exception emulating
> code (push/pop registers, call handler), copied instructions(Kprobes copies
> instructions which will be replaced by a jump, to the detour buffer), and
> a jump which jumps back to the original execution path.
>
> - Pre-optimization
> After preparing detour code, Kprobes enqueues the kprobe to optimizing list
> and kicks kprobe-optimizer workqueue to optimize it. To wait other optimized
> probes, kprobe-optimizer will delay to work.
> When the optimized-kprobe is hit before optimization, its handler
> changes IP(instruction pointer) to copied code and exits. So, those
> copied instructions are executed on the detour buffer.
>
> - Optimization
> Kprobe-optimizer doesn't start instruction-replacing soon, it waits
> synchronize_sched for safety, because some processors are possible to be
> interrupted on the middle of instruction series (2nd or Nth instruction)
> which will be replaced by a jump instruction(*).
> As you know, synchronize_sched() can ensure that all interruptions which
> were executed when synchronize_sched() was called are done, only if
> CONFIG_PREEMPT=n. So, this version supports only the kernel with
> CONFIG_PREEMPT=n.(**)
> After that, kprobe-optimizer calls stop_machine() to replace probed-
> instructions with a jump instruction by using text_poke_smp().
>
> - Unoptimization
> When unregistering, disabling kprobe or being blocked by other kprobe,
> an optimized-kprobe will be unoptimized. Before kprobe-optimizer runs,
> the kprobe just be dequeued from the optimized list. When the optimization
> has been done, it replaces a jump with int3 breakpoint and original code
> by using text_poke_smp().
>
> (*)Please imagine that 2nd instruction is interrupted and
> optimizer replaces the 2nd instruction with jump *address*
> while the interrupt handler is running. When the interrupt
> returns to original address, there is no valid instructions
> and it causes unexpected result.
>
> (**)This optimization-safety checking may be replaced with stop-machine
> method which ksplice is done for supporting CONFIG_PREEMPT=y kernel.
>
>
> Thank you,
>
> ---
>
> Masami Hiramatsu (9):
> kprobes: Add documents of jump optimization
> kprobes/x86: Support kprobes jump optimization on x86
> x86: Add text_poke_smp for SMP cross modifying code
> kprobes/x86: Cleanup save/restore registers
> kprobes/x86: Boost probes when reentering
> kprobes: Jump optimization sysctl interface
> kprobes: Introduce kprobes jump optimization
> kprobes: Introduce generic insn_slot framework
> kprobes/x86: Cleanup RELATIVEJUMP_INSTRUCTION to RELATIVEJUMP_OPCODE
>
>
> Documentation/kprobes.txt | 191 ++++++++++-
> arch/Kconfig | 13 +
> arch/x86/Kconfig | 1
> arch/x86/include/asm/alternative.h | 4
> arch/x86/include/asm/kprobes.h | 31 ++
> arch/x86/kernel/alternative.c | 60 +++
> arch/x86/kernel/kprobes.c | 596 ++++++++++++++++++++++++++++------
> include/linux/kprobes.h | 44 +++
> kernel/kprobes.c | 626 +++++++++++++++++++++++++++++++-----
> kernel/sysctl.c | 12 +
> 10 files changed, 1373 insertions(+), 205 deletions(-)
>
> --
> Masami Hiramatsu
>
> Software Engineer
> Hitachi Computer Products (America), Inc.
> Software Solutions Division
>
> e-mail: [email protected]
>
--
Mathieu Desnoyers
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F BA06 3F25 A8FE 3BAE 9A68
Mathieu Desnoyers wrote:
> * Masami Hiramatsu ([email protected]) wrote:
>> Hi,
>>
>> Here are the patchset of the kprobes jump optimization v8
>> (a.k.a. Djprobe). This version is just moving onto
>> 2.6.33-rc4-tip. Ingo, I assume its a good timing to
>> push this code onto -tip tree (maybe developing branch?),
>> since people can test it with perf-probe.
>>
>> I've decided to make a separated series of patches of
>> jump optimization with text_poke_smp() which is
>> 'officially' supported on Intel's processors.
>> So, this version of patches are just updated against
>> the latest tip/master, no other updates are included.
>>
>> I know that int3-bypassing method (text_poke_fixup())
>> is currently unofficially believed as safe. But we
>> need to get more official answers from x86 vendors.
>> Moreover, we need to tweak entry_*.S for preventing
>> recursive NMI, because int3 inside NMI handler will
>> unblock NMI blocking. I'd like to push it after this
>> series of patches are merged.
>>
>> Anyway, thanks Mathieu and Peter, for helping me to
>> implement it and organizing discussion points about
>> int3-bypass XMC!
>>
>> These patches can be applied on the latest -tip.
>>
>> Changes in v8:
>> - Update patches against the latest tip/master.
>> - Drop text_poke_fixup() related patches.
>> - Update benchmark results and add jprobes and kprobe(post-handler)
>> results.
>>
>> And kprobe stress test didn't found any regressions - from kprobes,
>> under kvm/x86.
>>
>> TODO:
>> - Support NMI-safe int3-bypassing text_poke.
>
> Please have a look at:
>
> "x86 NMI-safe INT3 and Page Fault"
> http://git.kernel.org/?p=linux/kernel/git/compudj/linux-2.6-lttng.git;a=commit;h=90516e3c718e0502f6f2eb616fad4447645ca47d
>
> and
>
> "x86_64 page fault NMI-safe"
> http://git.kernel.org/?p=linux/kernel/git/compudj/linux-2.6-lttng.git;a=commit;h=ad1bf11a68c35a44edd8d686a0842896f408e17c
>
> That turns this TODO into the "done" section ;)
>
> I've been using these patches in the lttng tree for 1-2 years.
Hi Mathieu,
Sure, those looks good to me too. Thanks!
I'd just like to separate this issue from kprobe jump optimization
itself this time, because;
- text_poke_fixup() is not necessary to kprobes :-)
- the int3-bypassing text_poke method is still unofficially ensured
by intel people. and we also need to require answers from other
x86 compatible processor vendors.
Anyway, my next series of patches will contain those patches.
BTW, I think your 2nd patch seems good to send to upstream separately.
Thank you!
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: [email protected]