2005-04-16 18:20:36 -04:00
|
|
|
/*
|
|
|
|
* Kernel Probes (KProbes)
|
|
|
|
* kernel/kprobes.c
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
|
|
|
|
*
|
|
|
|
* Copyright (C) IBM Corporation, 2002, 2004
|
|
|
|
*
|
|
|
|
* 2002-Oct Created by Vamsi Krishna S <vamsi_krishna@in.ibm.com> Kernel
|
|
|
|
* Probes initial implementation (includes suggestions from
|
|
|
|
* Rusty Russell).
|
|
|
|
* 2004-Aug Updated by Prasanna S Panchamukhi <prasanna@in.ibm.com> with
|
|
|
|
* hlists and exceptions notifier as suggested by Andi Kleen.
|
|
|
|
* 2004-July Suparna Bhattacharya <suparna@in.ibm.com> added jumper probes
|
|
|
|
* interface to access function arguments.
|
|
|
|
* 2004-Sep Prasanna S Panchamukhi <prasanna@in.ibm.com> Changed Kprobes
|
|
|
|
* exceptions notifier to be first on the priority list.
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
* 2005-May Hien Nguyen <hien@us.ibm.com>, Jim Keniston
|
|
|
|
* <jkenisto@us.ibm.com> and Prasanna S Panchamukhi
|
|
|
|
* <prasanna@in.ibm.com> added function-return probes.
|
2005-04-16 18:20:36 -04:00
|
|
|
*/
|
|
|
|
#include <linux/kprobes.h>
|
|
|
|
#include <linux/hash.h>
|
|
|
|
#include <linux/init.h>
|
2005-10-30 18:03:48 -05:00
|
|
|
#include <linux/slab.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <linux/module.h>
|
2005-06-27 18:17:01 -04:00
|
|
|
#include <linux/moduleloader.h>
|
2005-09-06 18:19:26 -04:00
|
|
|
#include <asm-generic/sections.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <asm/cacheflush.h>
|
|
|
|
#include <asm/errno.h>
|
|
|
|
#include <asm/kdebug.h>
|
|
|
|
|
|
|
|
#define KPROBE_HASH_BITS 6
|
|
|
|
#define KPROBE_TABLE_SIZE (1 << KPROBE_HASH_BITS)
|
|
|
|
|
|
|
|
static struct hlist_head kprobe_table[KPROBE_TABLE_SIZE];
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
static struct hlist_head kretprobe_inst_table[KPROBE_TABLE_SIZE];
|
2006-06-26 03:25:29 -04:00
|
|
|
static atomic_t kprobe_count;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2006-03-23 06:00:35 -05:00
|
|
|
DEFINE_MUTEX(kprobe_mutex); /* Protects kprobe_table */
|
2005-11-07 04:00:13 -05:00
|
|
|
DEFINE_SPINLOCK(kretprobe_lock); /* Protects kretprobe_inst_table */
|
2005-11-07 04:00:07 -05:00
|
|
|
static DEFINE_PER_CPU(struct kprobe *, kprobe_instance) = NULL;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2006-06-26 03:25:29 -04:00
|
|
|
static struct notifier_block kprobe_page_fault_nb = {
|
|
|
|
.notifier_call = kprobe_exceptions_notify,
|
|
|
|
.priority = 0x7fffffff /* we need to notified first */
|
|
|
|
};
|
|
|
|
|
2006-01-09 23:52:41 -05:00
|
|
|
#ifdef __ARCH_WANT_KPROBES_INSN_SLOT
|
2005-06-27 18:17:01 -04:00
|
|
|
/*
|
|
|
|
* kprobe->ainsn.insn points to the copy of the instruction to be
|
|
|
|
* single-stepped. x86_64, POWER4 and above have no-exec support and
|
|
|
|
* 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 hlist_node hlist;
|
|
|
|
kprobe_opcode_t *insns; /* Page of instruction slots */
|
|
|
|
char slot_used[INSNS_PER_PAGE];
|
|
|
|
int nused;
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct hlist_head kprobe_insn_pages;
|
|
|
|
|
|
|
|
/**
|
|
|
|
* 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.
|
|
|
|
*/
|
2005-09-06 18:19:26 -04:00
|
|
|
kprobe_opcode_t __kprobes *get_insn_slot(void)
|
2005-06-27 18:17:01 -04:00
|
|
|
{
|
|
|
|
struct kprobe_insn_page *kip;
|
|
|
|
struct hlist_node *pos;
|
|
|
|
|
|
|
|
hlist_for_each(pos, &kprobe_insn_pages) {
|
|
|
|
kip = hlist_entry(pos, struct kprobe_insn_page, hlist);
|
|
|
|
if (kip->nused < INSNS_PER_PAGE) {
|
|
|
|
int i;
|
|
|
|
for (i = 0; i < INSNS_PER_PAGE; i++) {
|
|
|
|
if (!kip->slot_used[i]) {
|
|
|
|
kip->slot_used[i] = 1;
|
|
|
|
kip->nused++;
|
|
|
|
return kip->insns + (i * MAX_INSN_SIZE);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
/* Surprise! No unused slots. Fix kip->nused. */
|
|
|
|
kip->nused = INSNS_PER_PAGE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* All out of space. Need to allocate a new page. Use slot 0.*/
|
|
|
|
kip = kmalloc(sizeof(struct kprobe_insn_page), GFP_KERNEL);
|
|
|
|
if (!kip) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Use module_alloc so this page is within +/- 2GB of where the
|
|
|
|
* kernel image and loaded module images reside. This is required
|
|
|
|
* so x86_64 can correctly handle the %rip-relative fixups.
|
|
|
|
*/
|
|
|
|
kip->insns = module_alloc(PAGE_SIZE);
|
|
|
|
if (!kip->insns) {
|
|
|
|
kfree(kip);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
INIT_HLIST_NODE(&kip->hlist);
|
|
|
|
hlist_add_head(&kip->hlist, &kprobe_insn_pages);
|
|
|
|
memset(kip->slot_used, 0, INSNS_PER_PAGE);
|
|
|
|
kip->slot_used[0] = 1;
|
|
|
|
kip->nused = 1;
|
|
|
|
return kip->insns;
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
void __kprobes free_insn_slot(kprobe_opcode_t *slot)
|
2005-06-27 18:17:01 -04:00
|
|
|
{
|
|
|
|
struct kprobe_insn_page *kip;
|
|
|
|
struct hlist_node *pos;
|
|
|
|
|
|
|
|
hlist_for_each(pos, &kprobe_insn_pages) {
|
|
|
|
kip = hlist_entry(pos, struct kprobe_insn_page, hlist);
|
|
|
|
if (kip->insns <= slot &&
|
|
|
|
slot < kip->insns + (INSNS_PER_PAGE * MAX_INSN_SIZE)) {
|
|
|
|
int i = (slot - kip->insns) / MAX_INSN_SIZE;
|
|
|
|
kip->slot_used[i] = 0;
|
|
|
|
kip->nused--;
|
|
|
|
if (kip->nused == 0) {
|
|
|
|
/*
|
|
|
|
* Page is no longer in use. Free it unless
|
|
|
|
* it's the last one. We keep the last one
|
|
|
|
* so as not to have to set it up again the
|
|
|
|
* next time somebody inserts a probe.
|
|
|
|
*/
|
|
|
|
hlist_del(&kip->hlist);
|
|
|
|
if (hlist_empty(&kprobe_insn_pages)) {
|
|
|
|
INIT_HLIST_NODE(&kip->hlist);
|
|
|
|
hlist_add_head(&kip->hlist,
|
|
|
|
&kprobe_insn_pages);
|
|
|
|
} else {
|
|
|
|
module_free(NULL, kip->insns);
|
|
|
|
kfree(kip);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2006-01-09 23:52:41 -05:00
|
|
|
#endif
|
2005-06-27 18:17:01 -04:00
|
|
|
|
2005-11-07 04:00:07 -05:00
|
|
|
/* We have preemption disabled.. so it is safe to use __ versions */
|
|
|
|
static inline void set_kprobe_instance(struct kprobe *kp)
|
|
|
|
{
|
|
|
|
__get_cpu_var(kprobe_instance) = kp;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void reset_kprobe_instance(void)
|
|
|
|
{
|
|
|
|
__get_cpu_var(kprobe_instance) = NULL;
|
|
|
|
}
|
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
/*
|
|
|
|
* This routine is called either:
|
2006-01-09 23:52:43 -05:00
|
|
|
* - under the kprobe_mutex - during kprobe_[un]register()
|
2005-11-07 04:00:13 -05:00
|
|
|
* OR
|
2005-11-07 04:00:14 -05:00
|
|
|
* - with preemption disabled - from arch/xxx/kernel/kprobes.c
|
2005-11-07 04:00:13 -05:00
|
|
|
*/
|
2005-09-06 18:19:26 -04:00
|
|
|
struct kprobe __kprobes *get_kprobe(void *addr)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
struct hlist_head *head;
|
|
|
|
struct hlist_node *node;
|
2005-11-07 04:00:13 -05:00
|
|
|
struct kprobe *p;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
head = &kprobe_table[hash_ptr(addr, KPROBE_HASH_BITS)];
|
2005-11-07 04:00:13 -05:00
|
|
|
hlist_for_each_entry_rcu(p, node, head, hlist) {
|
2005-04-16 18:20:36 -04:00
|
|
|
if (p->addr == addr)
|
|
|
|
return p;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-05-05 19:15:42 -04:00
|
|
|
/*
|
|
|
|
* Aggregate handlers for multiple kprobes support - these handlers
|
|
|
|
* take care of invoking the individual kprobe handlers on p->list
|
|
|
|
*/
|
2005-09-06 18:19:26 -04:00
|
|
|
static int __kprobes aggr_pre_handler(struct kprobe *p, struct pt_regs *regs)
|
2005-05-05 19:15:42 -04:00
|
|
|
{
|
|
|
|
struct kprobe *kp;
|
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
list_for_each_entry_rcu(kp, &p->list, list) {
|
2005-05-05 19:15:42 -04:00
|
|
|
if (kp->pre_handler) {
|
2005-11-07 04:00:07 -05:00
|
|
|
set_kprobe_instance(kp);
|
2005-06-23 03:09:41 -04:00
|
|
|
if (kp->pre_handler(kp, regs))
|
|
|
|
return 1;
|
2005-05-05 19:15:42 -04:00
|
|
|
}
|
2005-11-07 04:00:07 -05:00
|
|
|
reset_kprobe_instance();
|
2005-05-05 19:15:42 -04:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
static void __kprobes aggr_post_handler(struct kprobe *p, struct pt_regs *regs,
|
|
|
|
unsigned long flags)
|
2005-05-05 19:15:42 -04:00
|
|
|
{
|
|
|
|
struct kprobe *kp;
|
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
list_for_each_entry_rcu(kp, &p->list, list) {
|
2005-05-05 19:15:42 -04:00
|
|
|
if (kp->post_handler) {
|
2005-11-07 04:00:07 -05:00
|
|
|
set_kprobe_instance(kp);
|
2005-05-05 19:15:42 -04:00
|
|
|
kp->post_handler(kp, regs, flags);
|
2005-11-07 04:00:07 -05:00
|
|
|
reset_kprobe_instance();
|
2005-05-05 19:15:42 -04:00
|
|
|
}
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
static int __kprobes aggr_fault_handler(struct kprobe *p, struct pt_regs *regs,
|
|
|
|
int trapnr)
|
2005-05-05 19:15:42 -04:00
|
|
|
{
|
2005-11-07 04:00:07 -05:00
|
|
|
struct kprobe *cur = __get_cpu_var(kprobe_instance);
|
|
|
|
|
2005-05-05 19:15:42 -04:00
|
|
|
/*
|
|
|
|
* if we faulted "during" the execution of a user specified
|
|
|
|
* probe handler, invoke just that probe's fault handler
|
|
|
|
*/
|
2005-11-07 04:00:07 -05:00
|
|
|
if (cur && cur->fault_handler) {
|
|
|
|
if (cur->fault_handler(cur, regs, trapnr))
|
2005-05-05 19:15:42 -04:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
static int __kprobes aggr_break_handler(struct kprobe *p, struct pt_regs *regs)
|
2005-06-23 03:09:41 -04:00
|
|
|
{
|
2005-11-07 04:00:07 -05:00
|
|
|
struct kprobe *cur = __get_cpu_var(kprobe_instance);
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
if (cur && cur->break_handler) {
|
|
|
|
if (cur->break_handler(cur, regs))
|
|
|
|
ret = 1;
|
2005-06-23 03:09:41 -04:00
|
|
|
}
|
2005-11-07 04:00:07 -05:00
|
|
|
reset_kprobe_instance();
|
|
|
|
return ret;
|
2005-06-23 03:09:41 -04:00
|
|
|
}
|
|
|
|
|
2005-12-12 03:37:34 -05:00
|
|
|
/* Walks the list and increments nmissed count for multiprobe case */
|
|
|
|
void __kprobes kprobes_inc_nmissed_count(struct kprobe *p)
|
|
|
|
{
|
|
|
|
struct kprobe *kp;
|
|
|
|
if (p->pre_handler != aggr_pre_handler) {
|
|
|
|
p->nmissed++;
|
|
|
|
} else {
|
|
|
|
list_for_each_entry_rcu(kp, &p->list, list)
|
|
|
|
kp->nmissed++;
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
/* Called with kretprobe_lock held */
|
2005-09-06 18:19:26 -04:00
|
|
|
struct kretprobe_instance __kprobes *get_free_rp_inst(struct kretprobe *rp)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
struct hlist_node *node;
|
|
|
|
struct kretprobe_instance *ri;
|
|
|
|
hlist_for_each_entry(ri, node, &rp->free_instances, uflist)
|
|
|
|
return ri;
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
/* Called with kretprobe_lock held */
|
2005-09-06 18:19:26 -04:00
|
|
|
static struct kretprobe_instance __kprobes *get_used_rp_inst(struct kretprobe
|
|
|
|
*rp)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
struct hlist_node *node;
|
|
|
|
struct kretprobe_instance *ri;
|
|
|
|
hlist_for_each_entry(ri, node, &rp->used_instances, uflist)
|
|
|
|
return ri;
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
/* Called with kretprobe_lock held */
|
2005-09-06 18:19:26 -04:00
|
|
|
void __kprobes add_rp_inst(struct kretprobe_instance *ri)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Remove rp inst off the free list -
|
|
|
|
* Add it back when probed function returns
|
|
|
|
*/
|
|
|
|
hlist_del(&ri->uflist);
|
[PATCH] Return probe redesign: architecture independent changes
The following is the second version of the function return probe patches
I sent out earlier this week. Changes since my last submission include:
* Fix in ppc64 code removing an unneeded call to re-enable preemption
* Fix a build problem in ia64 when kprobes was turned off
* Added another BUG_ON check to each of the architecture trampoline
handlers
My initial patch description ==>
From my experiences with adding return probes to x86_64 and ia64, and the
feedback on LKML to those patches, I think we can simplify the design
for return probes.
The following patch tweaks the original design such that:
* Instead of storing the stack address in the return probe instance, the
task pointer is stored. This gives us all we need in order to:
- find the correct return probe instance when we enter the trampoline
(even if we are recursing)
- find all left-over return probe instances when the task is going away
This has the side effect of simplifying the implementation since more
work can be done in kernel/kprobes.c since architecture specific knowledge
of the stack layout is no longer required. Specifically, we no longer have:
- arch_get_kprobe_task()
- arch_kprobe_flush_task()
- get_rp_inst_tsk()
- get_rp_inst()
- trampoline_post_handler() <see next bullet>
* Instead of splitting the return probe handling and cleanup logic across
the pre and post trampoline handlers, all the work is pushed into the
pre function (trampoline_probe_handler), and then we skip single stepping
the original function. In this case the original instruction to be single
stepped was just a NOP, and we can do without the extra interruption.
The new flow of events to having a return probe handler execute when a target
function exits is:
* At system initialization time, a kprobe is inserted at the beginning of
kretprobe_trampoline. kernel/kprobes.c use to handle this on it's own,
but ia64 needed to do this a little differently (i.e. a function pointer
is really a pointer to a structure containing the instruction pointer and
a global pointer), so I added the notion of arch_init(), so that
kernel/kprobes.c:init_kprobes() now allows architecture specific
initialization by calling arch_init() before exiting. Each architecture
now registers a kprobe on it's own trampoline function.
* register_kretprobe() will insert a kprobe at the beginning of the targeted
function with the kprobe pre_handler set to arch_prepare_kretprobe
(still no change)
* When the target function is entered, the kprobe is fired, calling
arch_prepare_kretprobe (still no change)
* In arch_prepare_kretprobe() we try to get a free instance and if one is
available then we fill out the instance with a pointer to the return probe,
the original return address, and a pointer to the task structure (instead
of the stack address.) Just like before we change the return address
to the trampoline function and mark the instance as used.
If multiple return probes are registered for a given target function,
then arch_prepare_kretprobe() will get called multiple times for the same
task (since our kprobe implementation is able to handle multiple kprobes
at the same address.) Past the first call to arch_prepare_kretprobe,
we end up with the original address stored in the return probe instance
pointing to our trampoline function. (This is a significant difference
from the original arch_prepare_kretprobe design.)
* Target function executes like normal and then returns to kretprobe_trampoline.
* kprobe inserted on the first instruction of kretprobe_trampoline is fired
and calls trampoline_probe_handler() (no change here)
* trampoline_probe_handler() consumes each of the instances associated with
the current task by calling the registered handler function and marking
the instance as unused until an instance is found that has a return address
different then the trampoline function.
(change similar to my previous ia64 RFC)
* If the task is killed with some left-over return probe instances (meaning
that a target function was entered, but never returned), then we just
free any instances associated with the task. (Not much different other
then we can handle this without calling architecture specific functions.)
There is a known problem that this patch does not yet solve where
registering a return probe flush_old_exec or flush_thread will put us
in a bad state. Most likely the best way to handle this is to not allow
registering return probes on these two functions.
(Significant change)
This patch series applies to the 2.6.12-rc6-mm1 kernel, and provides:
* kernel/kprobes.c changes
* i386 patch of existing return probes implementation
* x86_64 patch of existing return probe implementation
* ia64 implementation
* ppc64 implementation (provided by Ananth)
This patch implements the architecture independant changes for a reworking
of the kprobes based function return probes design. Changes include:
* Removing functions for querying a return probe instance off a stack address
* Removing the stack_addr field from the kretprobe_instance definition,
and adding a task pointer
* Adding architecture specific initialization via arch_init()
* Removing extern definitions for the architecture trampoline functions
(this isn't needed anymore since the architecture handles the
initialization of the kprobe in the return probe trampoline function.)
Signed-off-by: Rusty Lynch <rusty.lynch@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-27 18:17:08 -04:00
|
|
|
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
/* Add rp inst onto table */
|
|
|
|
INIT_HLIST_NODE(&ri->hlist);
|
|
|
|
hlist_add_head(&ri->hlist,
|
[PATCH] Return probe redesign: architecture independent changes
The following is the second version of the function return probe patches
I sent out earlier this week. Changes since my last submission include:
* Fix in ppc64 code removing an unneeded call to re-enable preemption
* Fix a build problem in ia64 when kprobes was turned off
* Added another BUG_ON check to each of the architecture trampoline
handlers
My initial patch description ==>
From my experiences with adding return probes to x86_64 and ia64, and the
feedback on LKML to those patches, I think we can simplify the design
for return probes.
The following patch tweaks the original design such that:
* Instead of storing the stack address in the return probe instance, the
task pointer is stored. This gives us all we need in order to:
- find the correct return probe instance when we enter the trampoline
(even if we are recursing)
- find all left-over return probe instances when the task is going away
This has the side effect of simplifying the implementation since more
work can be done in kernel/kprobes.c since architecture specific knowledge
of the stack layout is no longer required. Specifically, we no longer have:
- arch_get_kprobe_task()
- arch_kprobe_flush_task()
- get_rp_inst_tsk()
- get_rp_inst()
- trampoline_post_handler() <see next bullet>
* Instead of splitting the return probe handling and cleanup logic across
the pre and post trampoline handlers, all the work is pushed into the
pre function (trampoline_probe_handler), and then we skip single stepping
the original function. In this case the original instruction to be single
stepped was just a NOP, and we can do without the extra interruption.
The new flow of events to having a return probe handler execute when a target
function exits is:
* At system initialization time, a kprobe is inserted at the beginning of
kretprobe_trampoline. kernel/kprobes.c use to handle this on it's own,
but ia64 needed to do this a little differently (i.e. a function pointer
is really a pointer to a structure containing the instruction pointer and
a global pointer), so I added the notion of arch_init(), so that
kernel/kprobes.c:init_kprobes() now allows architecture specific
initialization by calling arch_init() before exiting. Each architecture
now registers a kprobe on it's own trampoline function.
* register_kretprobe() will insert a kprobe at the beginning of the targeted
function with the kprobe pre_handler set to arch_prepare_kretprobe
(still no change)
* When the target function is entered, the kprobe is fired, calling
arch_prepare_kretprobe (still no change)
* In arch_prepare_kretprobe() we try to get a free instance and if one is
available then we fill out the instance with a pointer to the return probe,
the original return address, and a pointer to the task structure (instead
of the stack address.) Just like before we change the return address
to the trampoline function and mark the instance as used.
If multiple return probes are registered for a given target function,
then arch_prepare_kretprobe() will get called multiple times for the same
task (since our kprobe implementation is able to handle multiple kprobes
at the same address.) Past the first call to arch_prepare_kretprobe,
we end up with the original address stored in the return probe instance
pointing to our trampoline function. (This is a significant difference
from the original arch_prepare_kretprobe design.)
* Target function executes like normal and then returns to kretprobe_trampoline.
* kprobe inserted on the first instruction of kretprobe_trampoline is fired
and calls trampoline_probe_handler() (no change here)
* trampoline_probe_handler() consumes each of the instances associated with
the current task by calling the registered handler function and marking
the instance as unused until an instance is found that has a return address
different then the trampoline function.
(change similar to my previous ia64 RFC)
* If the task is killed with some left-over return probe instances (meaning
that a target function was entered, but never returned), then we just
free any instances associated with the task. (Not much different other
then we can handle this without calling architecture specific functions.)
There is a known problem that this patch does not yet solve where
registering a return probe flush_old_exec or flush_thread will put us
in a bad state. Most likely the best way to handle this is to not allow
registering return probes on these two functions.
(Significant change)
This patch series applies to the 2.6.12-rc6-mm1 kernel, and provides:
* kernel/kprobes.c changes
* i386 patch of existing return probes implementation
* x86_64 patch of existing return probe implementation
* ia64 implementation
* ppc64 implementation (provided by Ananth)
This patch implements the architecture independant changes for a reworking
of the kprobes based function return probes design. Changes include:
* Removing functions for querying a return probe instance off a stack address
* Removing the stack_addr field from the kretprobe_instance definition,
and adding a task pointer
* Adding architecture specific initialization via arch_init()
* Removing extern definitions for the architecture trampoline functions
(this isn't needed anymore since the architecture handles the
initialization of the kprobe in the return probe trampoline function.)
Signed-off-by: Rusty Lynch <rusty.lynch@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-27 18:17:08 -04:00
|
|
|
&kretprobe_inst_table[hash_ptr(ri->task, KPROBE_HASH_BITS)]);
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
|
|
|
|
/* Also add this rp inst to the used list. */
|
|
|
|
INIT_HLIST_NODE(&ri->uflist);
|
|
|
|
hlist_add_head(&ri->uflist, &ri->rp->used_instances);
|
|
|
|
}
|
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
/* Called with kretprobe_lock held */
|
2005-09-06 18:19:26 -04:00
|
|
|
void __kprobes recycle_rp_inst(struct kretprobe_instance *ri)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
/* remove rp inst off the rprobe_inst_table */
|
|
|
|
hlist_del(&ri->hlist);
|
|
|
|
if (ri->rp) {
|
|
|
|
/* remove rp inst off the used list */
|
|
|
|
hlist_del(&ri->uflist);
|
|
|
|
/* put rp inst back onto the free list */
|
|
|
|
INIT_HLIST_NODE(&ri->uflist);
|
|
|
|
hlist_add_head(&ri->uflist, &ri->rp->free_instances);
|
|
|
|
} else
|
|
|
|
/* Unregistering */
|
|
|
|
kfree(ri);
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
struct hlist_head __kprobes *kretprobe_inst_table_head(struct task_struct *tsk)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
return &kretprobe_inst_table[hash_ptr(tsk, KPROBE_HASH_BITS)];
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2006-03-26 04:38:20 -05:00
|
|
|
* This function is called from finish_task_switch when task tk becomes dead,
|
|
|
|
* so that we can recycle any function-return probe instances associated
|
|
|
|
* with this task. These left over instances represent probed functions
|
|
|
|
* that have been called but will never return.
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
*/
|
2005-09-06 18:19:26 -04:00
|
|
|
void __kprobes kprobe_flush_task(struct task_struct *tk)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
[PATCH] Return probe redesign: architecture independent changes
The following is the second version of the function return probe patches
I sent out earlier this week. Changes since my last submission include:
* Fix in ppc64 code removing an unneeded call to re-enable preemption
* Fix a build problem in ia64 when kprobes was turned off
* Added another BUG_ON check to each of the architecture trampoline
handlers
My initial patch description ==>
From my experiences with adding return probes to x86_64 and ia64, and the
feedback on LKML to those patches, I think we can simplify the design
for return probes.
The following patch tweaks the original design such that:
* Instead of storing the stack address in the return probe instance, the
task pointer is stored. This gives us all we need in order to:
- find the correct return probe instance when we enter the trampoline
(even if we are recursing)
- find all left-over return probe instances when the task is going away
This has the side effect of simplifying the implementation since more
work can be done in kernel/kprobes.c since architecture specific knowledge
of the stack layout is no longer required. Specifically, we no longer have:
- arch_get_kprobe_task()
- arch_kprobe_flush_task()
- get_rp_inst_tsk()
- get_rp_inst()
- trampoline_post_handler() <see next bullet>
* Instead of splitting the return probe handling and cleanup logic across
the pre and post trampoline handlers, all the work is pushed into the
pre function (trampoline_probe_handler), and then we skip single stepping
the original function. In this case the original instruction to be single
stepped was just a NOP, and we can do without the extra interruption.
The new flow of events to having a return probe handler execute when a target
function exits is:
* At system initialization time, a kprobe is inserted at the beginning of
kretprobe_trampoline. kernel/kprobes.c use to handle this on it's own,
but ia64 needed to do this a little differently (i.e. a function pointer
is really a pointer to a structure containing the instruction pointer and
a global pointer), so I added the notion of arch_init(), so that
kernel/kprobes.c:init_kprobes() now allows architecture specific
initialization by calling arch_init() before exiting. Each architecture
now registers a kprobe on it's own trampoline function.
* register_kretprobe() will insert a kprobe at the beginning of the targeted
function with the kprobe pre_handler set to arch_prepare_kretprobe
(still no change)
* When the target function is entered, the kprobe is fired, calling
arch_prepare_kretprobe (still no change)
* In arch_prepare_kretprobe() we try to get a free instance and if one is
available then we fill out the instance with a pointer to the return probe,
the original return address, and a pointer to the task structure (instead
of the stack address.) Just like before we change the return address
to the trampoline function and mark the instance as used.
If multiple return probes are registered for a given target function,
then arch_prepare_kretprobe() will get called multiple times for the same
task (since our kprobe implementation is able to handle multiple kprobes
at the same address.) Past the first call to arch_prepare_kretprobe,
we end up with the original address stored in the return probe instance
pointing to our trampoline function. (This is a significant difference
from the original arch_prepare_kretprobe design.)
* Target function executes like normal and then returns to kretprobe_trampoline.
* kprobe inserted on the first instruction of kretprobe_trampoline is fired
and calls trampoline_probe_handler() (no change here)
* trampoline_probe_handler() consumes each of the instances associated with
the current task by calling the registered handler function and marking
the instance as unused until an instance is found that has a return address
different then the trampoline function.
(change similar to my previous ia64 RFC)
* If the task is killed with some left-over return probe instances (meaning
that a target function was entered, but never returned), then we just
free any instances associated with the task. (Not much different other
then we can handle this without calling architecture specific functions.)
There is a known problem that this patch does not yet solve where
registering a return probe flush_old_exec or flush_thread will put us
in a bad state. Most likely the best way to handle this is to not allow
registering return probes on these two functions.
(Significant change)
This patch series applies to the 2.6.12-rc6-mm1 kernel, and provides:
* kernel/kprobes.c changes
* i386 patch of existing return probes implementation
* x86_64 patch of existing return probe implementation
* ia64 implementation
* ppc64 implementation (provided by Ananth)
This patch implements the architecture independant changes for a reworking
of the kprobes based function return probes design. Changes include:
* Removing functions for querying a return probe instance off a stack address
* Removing the stack_addr field from the kretprobe_instance definition,
and adding a task pointer
* Adding architecture specific initialization via arch_init()
* Removing extern definitions for the architecture trampoline functions
(this isn't needed anymore since the architecture handles the
initialization of the kprobe in the return probe trampoline function.)
Signed-off-by: Rusty Lynch <rusty.lynch@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-27 18:17:08 -04:00
|
|
|
struct kretprobe_instance *ri;
|
|
|
|
struct hlist_head *head;
|
|
|
|
struct hlist_node *node, *tmp;
|
2005-06-23 03:09:26 -04:00
|
|
|
unsigned long flags = 0;
|
[PATCH] Return probe redesign: architecture independent changes
The following is the second version of the function return probe patches
I sent out earlier this week. Changes since my last submission include:
* Fix in ppc64 code removing an unneeded call to re-enable preemption
* Fix a build problem in ia64 when kprobes was turned off
* Added another BUG_ON check to each of the architecture trampoline
handlers
My initial patch description ==>
From my experiences with adding return probes to x86_64 and ia64, and the
feedback on LKML to those patches, I think we can simplify the design
for return probes.
The following patch tweaks the original design such that:
* Instead of storing the stack address in the return probe instance, the
task pointer is stored. This gives us all we need in order to:
- find the correct return probe instance when we enter the trampoline
(even if we are recursing)
- find all left-over return probe instances when the task is going away
This has the side effect of simplifying the implementation since more
work can be done in kernel/kprobes.c since architecture specific knowledge
of the stack layout is no longer required. Specifically, we no longer have:
- arch_get_kprobe_task()
- arch_kprobe_flush_task()
- get_rp_inst_tsk()
- get_rp_inst()
- trampoline_post_handler() <see next bullet>
* Instead of splitting the return probe handling and cleanup logic across
the pre and post trampoline handlers, all the work is pushed into the
pre function (trampoline_probe_handler), and then we skip single stepping
the original function. In this case the original instruction to be single
stepped was just a NOP, and we can do without the extra interruption.
The new flow of events to having a return probe handler execute when a target
function exits is:
* At system initialization time, a kprobe is inserted at the beginning of
kretprobe_trampoline. kernel/kprobes.c use to handle this on it's own,
but ia64 needed to do this a little differently (i.e. a function pointer
is really a pointer to a structure containing the instruction pointer and
a global pointer), so I added the notion of arch_init(), so that
kernel/kprobes.c:init_kprobes() now allows architecture specific
initialization by calling arch_init() before exiting. Each architecture
now registers a kprobe on it's own trampoline function.
* register_kretprobe() will insert a kprobe at the beginning of the targeted
function with the kprobe pre_handler set to arch_prepare_kretprobe
(still no change)
* When the target function is entered, the kprobe is fired, calling
arch_prepare_kretprobe (still no change)
* In arch_prepare_kretprobe() we try to get a free instance and if one is
available then we fill out the instance with a pointer to the return probe,
the original return address, and a pointer to the task structure (instead
of the stack address.) Just like before we change the return address
to the trampoline function and mark the instance as used.
If multiple return probes are registered for a given target function,
then arch_prepare_kretprobe() will get called multiple times for the same
task (since our kprobe implementation is able to handle multiple kprobes
at the same address.) Past the first call to arch_prepare_kretprobe,
we end up with the original address stored in the return probe instance
pointing to our trampoline function. (This is a significant difference
from the original arch_prepare_kretprobe design.)
* Target function executes like normal and then returns to kretprobe_trampoline.
* kprobe inserted on the first instruction of kretprobe_trampoline is fired
and calls trampoline_probe_handler() (no change here)
* trampoline_probe_handler() consumes each of the instances associated with
the current task by calling the registered handler function and marking
the instance as unused until an instance is found that has a return address
different then the trampoline function.
(change similar to my previous ia64 RFC)
* If the task is killed with some left-over return probe instances (meaning
that a target function was entered, but never returned), then we just
free any instances associated with the task. (Not much different other
then we can handle this without calling architecture specific functions.)
There is a known problem that this patch does not yet solve where
registering a return probe flush_old_exec or flush_thread will put us
in a bad state. Most likely the best way to handle this is to not allow
registering return probes on these two functions.
(Significant change)
This patch series applies to the 2.6.12-rc6-mm1 kernel, and provides:
* kernel/kprobes.c changes
* i386 patch of existing return probes implementation
* x86_64 patch of existing return probe implementation
* ia64 implementation
* ppc64 implementation (provided by Ananth)
This patch implements the architecture independant changes for a reworking
of the kprobes based function return probes design. Changes include:
* Removing functions for querying a return probe instance off a stack address
* Removing the stack_addr field from the kretprobe_instance definition,
and adding a task pointer
* Adding architecture specific initialization via arch_init()
* Removing extern definitions for the architecture trampoline functions
(this isn't needed anymore since the architecture handles the
initialization of the kprobe in the return probe trampoline function.)
Signed-off-by: Rusty Lynch <rusty.lynch@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-27 18:17:08 -04:00
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
spin_lock_irqsave(&kretprobe_lock, flags);
|
2006-03-26 04:38:20 -05:00
|
|
|
head = kretprobe_inst_table_head(tk);
|
[PATCH] Return probe redesign: architecture independent changes
The following is the second version of the function return probe patches
I sent out earlier this week. Changes since my last submission include:
* Fix in ppc64 code removing an unneeded call to re-enable preemption
* Fix a build problem in ia64 when kprobes was turned off
* Added another BUG_ON check to each of the architecture trampoline
handlers
My initial patch description ==>
From my experiences with adding return probes to x86_64 and ia64, and the
feedback on LKML to those patches, I think we can simplify the design
for return probes.
The following patch tweaks the original design such that:
* Instead of storing the stack address in the return probe instance, the
task pointer is stored. This gives us all we need in order to:
- find the correct return probe instance when we enter the trampoline
(even if we are recursing)
- find all left-over return probe instances when the task is going away
This has the side effect of simplifying the implementation since more
work can be done in kernel/kprobes.c since architecture specific knowledge
of the stack layout is no longer required. Specifically, we no longer have:
- arch_get_kprobe_task()
- arch_kprobe_flush_task()
- get_rp_inst_tsk()
- get_rp_inst()
- trampoline_post_handler() <see next bullet>
* Instead of splitting the return probe handling and cleanup logic across
the pre and post trampoline handlers, all the work is pushed into the
pre function (trampoline_probe_handler), and then we skip single stepping
the original function. In this case the original instruction to be single
stepped was just a NOP, and we can do without the extra interruption.
The new flow of events to having a return probe handler execute when a target
function exits is:
* At system initialization time, a kprobe is inserted at the beginning of
kretprobe_trampoline. kernel/kprobes.c use to handle this on it's own,
but ia64 needed to do this a little differently (i.e. a function pointer
is really a pointer to a structure containing the instruction pointer and
a global pointer), so I added the notion of arch_init(), so that
kernel/kprobes.c:init_kprobes() now allows architecture specific
initialization by calling arch_init() before exiting. Each architecture
now registers a kprobe on it's own trampoline function.
* register_kretprobe() will insert a kprobe at the beginning of the targeted
function with the kprobe pre_handler set to arch_prepare_kretprobe
(still no change)
* When the target function is entered, the kprobe is fired, calling
arch_prepare_kretprobe (still no change)
* In arch_prepare_kretprobe() we try to get a free instance and if one is
available then we fill out the instance with a pointer to the return probe,
the original return address, and a pointer to the task structure (instead
of the stack address.) Just like before we change the return address
to the trampoline function and mark the instance as used.
If multiple return probes are registered for a given target function,
then arch_prepare_kretprobe() will get called multiple times for the same
task (since our kprobe implementation is able to handle multiple kprobes
at the same address.) Past the first call to arch_prepare_kretprobe,
we end up with the original address stored in the return probe instance
pointing to our trampoline function. (This is a significant difference
from the original arch_prepare_kretprobe design.)
* Target function executes like normal and then returns to kretprobe_trampoline.
* kprobe inserted on the first instruction of kretprobe_trampoline is fired
and calls trampoline_probe_handler() (no change here)
* trampoline_probe_handler() consumes each of the instances associated with
the current task by calling the registered handler function and marking
the instance as unused until an instance is found that has a return address
different then the trampoline function.
(change similar to my previous ia64 RFC)
* If the task is killed with some left-over return probe instances (meaning
that a target function was entered, but never returned), then we just
free any instances associated with the task. (Not much different other
then we can handle this without calling architecture specific functions.)
There is a known problem that this patch does not yet solve where
registering a return probe flush_old_exec or flush_thread will put us
in a bad state. Most likely the best way to handle this is to not allow
registering return probes on these two functions.
(Significant change)
This patch series applies to the 2.6.12-rc6-mm1 kernel, and provides:
* kernel/kprobes.c changes
* i386 patch of existing return probes implementation
* x86_64 patch of existing return probe implementation
* ia64 implementation
* ppc64 implementation (provided by Ananth)
This patch implements the architecture independant changes for a reworking
of the kprobes based function return probes design. Changes include:
* Removing functions for querying a return probe instance off a stack address
* Removing the stack_addr field from the kretprobe_instance definition,
and adding a task pointer
* Adding architecture specific initialization via arch_init()
* Removing extern definitions for the architecture trampoline functions
(this isn't needed anymore since the architecture handles the
initialization of the kprobe in the return probe trampoline function.)
Signed-off-by: Rusty Lynch <rusty.lynch@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-27 18:17:08 -04:00
|
|
|
hlist_for_each_entry_safe(ri, node, tmp, head, hlist) {
|
|
|
|
if (ri->task == tk)
|
|
|
|
recycle_rp_inst(ri);
|
|
|
|
}
|
2005-11-07 04:00:13 -05:00
|
|
|
spin_unlock_irqrestore(&kretprobe_lock, flags);
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void free_rp_inst(struct kretprobe *rp)
|
|
|
|
{
|
|
|
|
struct kretprobe_instance *ri;
|
|
|
|
while ((ri = get_free_rp_inst(rp)) != NULL) {
|
|
|
|
hlist_del(&ri->uflist);
|
|
|
|
kfree(ri);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-06-23 03:09:41 -04:00
|
|
|
/*
|
|
|
|
* 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 old_p->list. Fail if this is the
|
|
|
|
* second jprobe at the address - two jprobes can't coexist
|
|
|
|
*/
|
2005-09-06 18:19:26 -04:00
|
|
|
static int __kprobes add_new_kprobe(struct kprobe *old_p, struct kprobe *p)
|
2005-06-23 03:09:41 -04:00
|
|
|
{
|
|
|
|
if (p->break_handler) {
|
2006-06-26 03:25:22 -04:00
|
|
|
if (old_p->break_handler)
|
|
|
|
return -EEXIST;
|
2005-11-07 04:00:13 -05:00
|
|
|
list_add_tail_rcu(&p->list, &old_p->list);
|
2006-06-26 03:25:22 -04:00
|
|
|
old_p->break_handler = aggr_break_handler;
|
2005-06-23 03:09:41 -04:00
|
|
|
} else
|
2005-11-07 04:00:13 -05:00
|
|
|
list_add_rcu(&p->list, &old_p->list);
|
2006-06-26 03:25:22 -04:00
|
|
|
if (p->post_handler && !old_p->post_handler)
|
|
|
|
old_p->post_handler = aggr_post_handler;
|
2005-06-23 03:09:41 -04:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-05-05 19:15:42 -04:00
|
|
|
/*
|
|
|
|
* 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)
|
|
|
|
{
|
2005-06-23 03:09:41 -04:00
|
|
|
copy_kprobe(p, ap);
|
2006-07-30 06:03:26 -04:00
|
|
|
flush_insn_slot(ap);
|
2005-05-05 19:15:42 -04:00
|
|
|
ap->addr = p->addr;
|
|
|
|
ap->pre_handler = aggr_pre_handler;
|
|
|
|
ap->fault_handler = aggr_fault_handler;
|
2006-06-26 03:25:22 -04:00
|
|
|
if (p->post_handler)
|
|
|
|
ap->post_handler = aggr_post_handler;
|
|
|
|
if (p->break_handler)
|
|
|
|
ap->break_handler = aggr_break_handler;
|
2005-05-05 19:15:42 -04:00
|
|
|
|
|
|
|
INIT_LIST_HEAD(&ap->list);
|
2005-11-07 04:00:13 -05:00
|
|
|
list_add_rcu(&p->list, &ap->list);
|
2005-05-05 19:15:42 -04:00
|
|
|
|
2005-12-12 03:37:12 -05:00
|
|
|
hlist_replace_rcu(&p->hlist, &ap->hlist);
|
2005-05-05 19:15:42 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is the second or subsequent kprobe at the address - handle
|
|
|
|
* the intricacies
|
|
|
|
*/
|
2005-09-06 18:19:26 -04:00
|
|
|
static int __kprobes register_aggr_kprobe(struct kprobe *old_p,
|
|
|
|
struct kprobe *p)
|
2005-05-05 19:15:42 -04:00
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
struct kprobe *ap;
|
|
|
|
|
2005-06-23 03:09:41 -04:00
|
|
|
if (old_p->pre_handler == aggr_pre_handler) {
|
|
|
|
copy_kprobe(old_p, p);
|
|
|
|
ret = add_new_kprobe(old_p, p);
|
2005-05-05 19:15:42 -04:00
|
|
|
} else {
|
2006-01-09 23:52:46 -05:00
|
|
|
ap = kzalloc(sizeof(struct kprobe), GFP_KERNEL);
|
2005-05-05 19:15:42 -04:00
|
|
|
if (!ap)
|
|
|
|
return -ENOMEM;
|
|
|
|
add_aggr_kprobe(ap, old_p);
|
2005-06-23 03:09:41 -04:00
|
|
|
copy_kprobe(ap, p);
|
|
|
|
ret = add_new_kprobe(ap, p);
|
2005-05-05 19:15:42 -04:00
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
static int __kprobes in_kprobes_functions(unsigned long addr)
|
|
|
|
{
|
|
|
|
if (addr >= (unsigned long)__kprobes_text_start
|
|
|
|
&& addr < (unsigned long)__kprobes_text_end)
|
|
|
|
return -EINVAL;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-01-11 15:17:41 -05:00
|
|
|
static int __kprobes __register_kprobe(struct kprobe *p,
|
|
|
|
unsigned long called_from)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
int ret = 0;
|
2005-05-05 19:15:42 -04:00
|
|
|
struct kprobe *old_p;
|
2006-01-11 15:17:41 -05:00
|
|
|
struct module *probed_mod;
|
2005-12-12 03:37:00 -05:00
|
|
|
|
|
|
|
if ((!kernel_text_address((unsigned long) p->addr)) ||
|
|
|
|
in_kprobes_functions((unsigned long) p->addr))
|
|
|
|
return -EINVAL;
|
|
|
|
|
2006-01-11 15:17:41 -05:00
|
|
|
p->mod_refcounted = 0;
|
|
|
|
/* Check are we probing a module */
|
|
|
|
if ((probed_mod = module_text_address((unsigned long) p->addr))) {
|
|
|
|
struct module *calling_mod = module_text_address(called_from);
|
|
|
|
/* We must allow modules to probe themself and
|
|
|
|
* in this case avoid incrementing the module refcount,
|
|
|
|
* so as to allow unloading of self probing modules.
|
|
|
|
*/
|
|
|
|
if (calling_mod && (calling_mod != probed_mod)) {
|
|
|
|
if (unlikely(!try_module_get(probed_mod)))
|
|
|
|
return -EINVAL;
|
|
|
|
p->mod_refcounted = 1;
|
|
|
|
} else
|
|
|
|
probed_mod = NULL;
|
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2005-11-07 04:00:13 -05:00
|
|
|
p->nmissed = 0;
|
2006-03-23 06:00:35 -05:00
|
|
|
mutex_lock(&kprobe_mutex);
|
2005-05-05 19:15:42 -04:00
|
|
|
old_p = get_kprobe(p->addr);
|
|
|
|
if (old_p) {
|
|
|
|
ret = register_aggr_kprobe(old_p, p);
|
2006-06-26 03:25:29 -04:00
|
|
|
if (!ret)
|
|
|
|
atomic_inc(&kprobe_count);
|
2005-04-16 18:20:36 -04:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2006-01-09 23:52:43 -05:00
|
|
|
if ((ret = arch_prepare_kprobe(p)) != 0)
|
|
|
|
goto out;
|
|
|
|
|
2005-05-05 19:15:42 -04:00
|
|
|
INIT_HLIST_NODE(&p->hlist);
|
2005-11-07 04:00:13 -05:00
|
|
|
hlist_add_head_rcu(&p->hlist,
|
2005-04-16 18:20:36 -04:00
|
|
|
&kprobe_table[hash_ptr(p->addr, KPROBE_HASH_BITS)]);
|
|
|
|
|
2006-06-26 03:25:29 -04:00
|
|
|
if (atomic_add_return(1, &kprobe_count) == \
|
|
|
|
(ARCH_INACTIVE_KPROBE_COUNT + 1))
|
|
|
|
register_page_fault_notifier(&kprobe_page_fault_nb);
|
|
|
|
|
2005-06-23 03:09:25 -04:00
|
|
|
arch_arm_kprobe(p);
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
out:
|
2006-03-23 06:00:35 -05:00
|
|
|
mutex_unlock(&kprobe_mutex);
|
2006-01-09 23:52:43 -05:00
|
|
|
|
2006-01-11 15:17:41 -05:00
|
|
|
if (ret && probed_mod)
|
|
|
|
module_put(probed_mod);
|
2005-04-16 18:20:36 -04:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2006-01-11 15:17:41 -05:00
|
|
|
int __kprobes register_kprobe(struct kprobe *p)
|
|
|
|
{
|
|
|
|
return __register_kprobe(p,
|
|
|
|
(unsigned long)__builtin_return_address(0));
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
void __kprobes unregister_kprobe(struct kprobe *p)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2005-12-12 03:37:00 -05:00
|
|
|
struct module *mod;
|
2006-01-09 23:52:44 -05:00
|
|
|
struct kprobe *old_p, *list_p;
|
|
|
|
int cleanup_p;
|
2005-05-05 19:15:42 -04:00
|
|
|
|
2006-03-23 06:00:35 -05:00
|
|
|
mutex_lock(&kprobe_mutex);
|
2005-05-05 19:15:42 -04:00
|
|
|
old_p = get_kprobe(p->addr);
|
2006-01-09 23:52:43 -05:00
|
|
|
if (unlikely(!old_p)) {
|
2006-03-23 06:00:35 -05:00
|
|
|
mutex_unlock(&kprobe_mutex);
|
2006-01-09 23:52:43 -05:00
|
|
|
return;
|
|
|
|
}
|
2006-01-09 23:52:44 -05:00
|
|
|
if (p != old_p) {
|
|
|
|
list_for_each_entry_rcu(list_p, &old_p->list, list)
|
|
|
|
if (list_p == p)
|
|
|
|
/* kprobe p is a valid probe */
|
|
|
|
goto valid_p;
|
2006-03-23 06:00:35 -05:00
|
|
|
mutex_unlock(&kprobe_mutex);
|
2006-01-09 23:52:44 -05:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
valid_p:
|
|
|
|
if ((old_p == p) || ((old_p->pre_handler == aggr_pre_handler) &&
|
2006-01-09 23:52:43 -05:00
|
|
|
(p->list.next == &old_p->list) &&
|
2006-01-09 23:52:44 -05:00
|
|
|
(p->list.prev == &old_p->list))) {
|
|
|
|
/* Only probe on the hash list */
|
2006-01-09 23:52:43 -05:00
|
|
|
arch_disarm_kprobe(p);
|
|
|
|
hlist_del_rcu(&old_p->hlist);
|
2006-01-09 23:52:44 -05:00
|
|
|
cleanup_p = 1;
|
2006-01-09 23:52:43 -05:00
|
|
|
} else {
|
|
|
|
list_del_rcu(&p->list);
|
2006-01-09 23:52:44 -05:00
|
|
|
cleanup_p = 0;
|
2006-01-09 23:52:43 -05:00
|
|
|
}
|
2005-11-07 04:00:13 -05:00
|
|
|
|
2006-03-23 06:00:35 -05:00
|
|
|
mutex_unlock(&kprobe_mutex);
|
2005-12-12 03:37:00 -05:00
|
|
|
|
2006-01-09 23:52:43 -05:00
|
|
|
synchronize_sched();
|
2006-01-11 15:17:41 -05:00
|
|
|
if (p->mod_refcounted &&
|
|
|
|
(mod = module_text_address((unsigned long)p->addr)))
|
2006-01-09 23:52:43 -05:00
|
|
|
module_put(mod);
|
2005-12-12 03:37:00 -05:00
|
|
|
|
2006-01-09 23:52:43 -05:00
|
|
|
if (cleanup_p) {
|
2006-01-09 23:52:44 -05:00
|
|
|
if (p != old_p) {
|
2006-01-09 23:52:43 -05:00
|
|
|
list_del_rcu(&p->list);
|
2005-11-07 04:00:13 -05:00
|
|
|
kfree(old_p);
|
2006-01-09 23:52:43 -05:00
|
|
|
}
|
2006-01-09 23:52:46 -05:00
|
|
|
arch_remove_kprobe(p);
|
2006-06-26 03:25:22 -04:00
|
|
|
} else {
|
|
|
|
mutex_lock(&kprobe_mutex);
|
|
|
|
if (p->break_handler)
|
|
|
|
old_p->break_handler = NULL;
|
|
|
|
if (p->post_handler){
|
|
|
|
list_for_each_entry_rcu(list_p, &old_p->list, list){
|
|
|
|
if (list_p->post_handler){
|
|
|
|
cleanup_p = 2;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (cleanup_p == 0)
|
|
|
|
old_p->post_handler = NULL;
|
|
|
|
}
|
|
|
|
mutex_unlock(&kprobe_mutex);
|
2006-01-09 23:52:43 -05:00
|
|
|
}
|
2006-06-26 03:25:29 -04:00
|
|
|
|
|
|
|
/* Call unregister_page_fault_notifier()
|
|
|
|
* if no probes are active
|
|
|
|
*/
|
|
|
|
mutex_lock(&kprobe_mutex);
|
|
|
|
if (atomic_add_return(-1, &kprobe_count) == \
|
|
|
|
ARCH_INACTIVE_KPROBE_COUNT)
|
|
|
|
unregister_page_fault_notifier(&kprobe_page_fault_nb);
|
|
|
|
mutex_unlock(&kprobe_mutex);
|
|
|
|
return;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
static struct notifier_block kprobe_exceptions_nb = {
|
2006-06-26 03:25:28 -04:00
|
|
|
.notifier_call = kprobe_exceptions_notify,
|
|
|
|
.priority = 0x7fffffff /* we need to be notified first */
|
|
|
|
};
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
int __kprobes register_jprobe(struct jprobe *jp)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
/* Todo: Verify probepoint is a function entry point */
|
|
|
|
jp->kp.pre_handler = setjmp_pre_handler;
|
|
|
|
jp->kp.break_handler = longjmp_break_handler;
|
|
|
|
|
2006-01-11 15:17:41 -05:00
|
|
|
return __register_kprobe(&jp->kp,
|
|
|
|
(unsigned long)__builtin_return_address(0));
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
void __kprobes unregister_jprobe(struct jprobe *jp)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
|
|
|
unregister_kprobe(&jp->kp);
|
|
|
|
}
|
|
|
|
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
#ifdef ARCH_SUPPORTS_KRETPROBES
|
|
|
|
|
2006-02-03 06:03:42 -05:00
|
|
|
/*
|
|
|
|
* This kprobe pre_handler is registered with every kretprobe. When probe
|
|
|
|
* hits it will set up the return probe.
|
|
|
|
*/
|
|
|
|
static int __kprobes pre_handler_kretprobe(struct kprobe *p,
|
|
|
|
struct pt_regs *regs)
|
|
|
|
{
|
|
|
|
struct kretprobe *rp = container_of(p, struct kretprobe, kp);
|
|
|
|
unsigned long flags = 0;
|
|
|
|
|
|
|
|
/*TODO: consider to only swap the RA after the last pre_handler fired */
|
|
|
|
spin_lock_irqsave(&kretprobe_lock, flags);
|
|
|
|
arch_prepare_kretprobe(rp, regs);
|
|
|
|
spin_unlock_irqrestore(&kretprobe_lock, flags);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
int __kprobes register_kretprobe(struct kretprobe *rp)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
struct kretprobe_instance *inst;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
rp->kp.pre_handler = pre_handler_kretprobe;
|
2006-04-20 05:43:11 -04:00
|
|
|
rp->kp.post_handler = NULL;
|
|
|
|
rp->kp.fault_handler = NULL;
|
|
|
|
rp->kp.break_handler = NULL;
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
|
|
|
|
/* Pre-allocate memory for max kretprobe instances */
|
|
|
|
if (rp->maxactive <= 0) {
|
|
|
|
#ifdef CONFIG_PREEMPT
|
|
|
|
rp->maxactive = max(10, 2 * NR_CPUS);
|
|
|
|
#else
|
|
|
|
rp->maxactive = NR_CPUS;
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
INIT_HLIST_HEAD(&rp->used_instances);
|
|
|
|
INIT_HLIST_HEAD(&rp->free_instances);
|
|
|
|
for (i = 0; i < rp->maxactive; i++) {
|
|
|
|
inst = kmalloc(sizeof(struct kretprobe_instance), GFP_KERNEL);
|
|
|
|
if (inst == NULL) {
|
|
|
|
free_rp_inst(rp);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
INIT_HLIST_NODE(&inst->uflist);
|
|
|
|
hlist_add_head(&inst->uflist, &rp->free_instances);
|
|
|
|
}
|
|
|
|
|
|
|
|
rp->nmissed = 0;
|
|
|
|
/* Establish function entry probe point */
|
2006-01-11 15:17:41 -05:00
|
|
|
if ((ret = __register_kprobe(&rp->kp,
|
|
|
|
(unsigned long)__builtin_return_address(0))) != 0)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
free_rp_inst(rp);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
#else /* ARCH_SUPPORTS_KRETPROBES */
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
int __kprobes register_kretprobe(struct kretprobe *rp)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
return -ENOSYS;
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* ARCH_SUPPORTS_KRETPROBES */
|
|
|
|
|
2005-09-06 18:19:26 -04:00
|
|
|
void __kprobes unregister_kretprobe(struct kretprobe *rp)
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
struct kretprobe_instance *ri;
|
|
|
|
|
|
|
|
unregister_kprobe(&rp->kp);
|
|
|
|
/* No race here */
|
2005-11-07 04:00:13 -05:00
|
|
|
spin_lock_irqsave(&kretprobe_lock, flags);
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
while ((ri = get_used_rp_inst(rp)) != NULL) {
|
|
|
|
ri->rp = NULL;
|
|
|
|
hlist_del(&ri->uflist);
|
|
|
|
}
|
2005-11-07 04:00:13 -05:00
|
|
|
spin_unlock_irqrestore(&kretprobe_lock, flags);
|
2006-02-03 06:03:43 -05:00
|
|
|
free_rp_inst(rp);
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
}
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
static int __init init_kprobes(void)
|
|
|
|
{
|
|
|
|
int i, err = 0;
|
|
|
|
|
|
|
|
/* FIXME allocate the probe table, currently defined statically */
|
|
|
|
/* initialize all list heads */
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
for (i = 0; i < KPROBE_TABLE_SIZE; i++) {
|
2005-04-16 18:20:36 -04:00
|
|
|
INIT_HLIST_HEAD(&kprobe_table[i]);
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
INIT_HLIST_HEAD(&kretprobe_inst_table[i]);
|
|
|
|
}
|
2006-06-26 03:25:29 -04:00
|
|
|
atomic_set(&kprobe_count, 0);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2005-07-05 21:54:50 -04:00
|
|
|
err = arch_init_kprobes();
|
[PATCH] Return probe redesign: architecture independent changes
The following is the second version of the function return probe patches
I sent out earlier this week. Changes since my last submission include:
* Fix in ppc64 code removing an unneeded call to re-enable preemption
* Fix a build problem in ia64 when kprobes was turned off
* Added another BUG_ON check to each of the architecture trampoline
handlers
My initial patch description ==>
From my experiences with adding return probes to x86_64 and ia64, and the
feedback on LKML to those patches, I think we can simplify the design
for return probes.
The following patch tweaks the original design such that:
* Instead of storing the stack address in the return probe instance, the
task pointer is stored. This gives us all we need in order to:
- find the correct return probe instance when we enter the trampoline
(even if we are recursing)
- find all left-over return probe instances when the task is going away
This has the side effect of simplifying the implementation since more
work can be done in kernel/kprobes.c since architecture specific knowledge
of the stack layout is no longer required. Specifically, we no longer have:
- arch_get_kprobe_task()
- arch_kprobe_flush_task()
- get_rp_inst_tsk()
- get_rp_inst()
- trampoline_post_handler() <see next bullet>
* Instead of splitting the return probe handling and cleanup logic across
the pre and post trampoline handlers, all the work is pushed into the
pre function (trampoline_probe_handler), and then we skip single stepping
the original function. In this case the original instruction to be single
stepped was just a NOP, and we can do without the extra interruption.
The new flow of events to having a return probe handler execute when a target
function exits is:
* At system initialization time, a kprobe is inserted at the beginning of
kretprobe_trampoline. kernel/kprobes.c use to handle this on it's own,
but ia64 needed to do this a little differently (i.e. a function pointer
is really a pointer to a structure containing the instruction pointer and
a global pointer), so I added the notion of arch_init(), so that
kernel/kprobes.c:init_kprobes() now allows architecture specific
initialization by calling arch_init() before exiting. Each architecture
now registers a kprobe on it's own trampoline function.
* register_kretprobe() will insert a kprobe at the beginning of the targeted
function with the kprobe pre_handler set to arch_prepare_kretprobe
(still no change)
* When the target function is entered, the kprobe is fired, calling
arch_prepare_kretprobe (still no change)
* In arch_prepare_kretprobe() we try to get a free instance and if one is
available then we fill out the instance with a pointer to the return probe,
the original return address, and a pointer to the task structure (instead
of the stack address.) Just like before we change the return address
to the trampoline function and mark the instance as used.
If multiple return probes are registered for a given target function,
then arch_prepare_kretprobe() will get called multiple times for the same
task (since our kprobe implementation is able to handle multiple kprobes
at the same address.) Past the first call to arch_prepare_kretprobe,
we end up with the original address stored in the return probe instance
pointing to our trampoline function. (This is a significant difference
from the original arch_prepare_kretprobe design.)
* Target function executes like normal and then returns to kretprobe_trampoline.
* kprobe inserted on the first instruction of kretprobe_trampoline is fired
and calls trampoline_probe_handler() (no change here)
* trampoline_probe_handler() consumes each of the instances associated with
the current task by calling the registered handler function and marking
the instance as unused until an instance is found that has a return address
different then the trampoline function.
(change similar to my previous ia64 RFC)
* If the task is killed with some left-over return probe instances (meaning
that a target function was entered, but never returned), then we just
free any instances associated with the task. (Not much different other
then we can handle this without calling architecture specific functions.)
There is a known problem that this patch does not yet solve where
registering a return probe flush_old_exec or flush_thread will put us
in a bad state. Most likely the best way to handle this is to not allow
registering return probes on these two functions.
(Significant change)
This patch series applies to the 2.6.12-rc6-mm1 kernel, and provides:
* kernel/kprobes.c changes
* i386 patch of existing return probes implementation
* x86_64 patch of existing return probe implementation
* ia64 implementation
* ppc64 implementation (provided by Ananth)
This patch implements the architecture independant changes for a reworking
of the kprobes based function return probes design. Changes include:
* Removing functions for querying a return probe instance off a stack address
* Removing the stack_addr field from the kretprobe_instance definition,
and adding a task pointer
* Adding architecture specific initialization via arch_init()
* Removing extern definitions for the architecture trampoline functions
(this isn't needed anymore since the architecture handles the
initialization of the kprobe in the return probe trampoline function.)
Signed-off-by: Rusty Lynch <rusty.lynch@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-27 18:17:08 -04:00
|
|
|
if (!err)
|
|
|
|
err = register_die_notifier(&kprobe_exceptions_nb);
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
__initcall(init_kprobes);
|
|
|
|
|
|
|
|
EXPORT_SYMBOL_GPL(register_kprobe);
|
|
|
|
EXPORT_SYMBOL_GPL(unregister_kprobe);
|
|
|
|
EXPORT_SYMBOL_GPL(register_jprobe);
|
|
|
|
EXPORT_SYMBOL_GPL(unregister_jprobe);
|
|
|
|
EXPORT_SYMBOL_GPL(jprobe_return);
|
[PATCH] kprobes: function-return probes
This patch adds function-return probes to kprobes for the i386
architecture. This enables you to establish a handler to be run when a
function returns.
1. API
Two new functions are added to kprobes:
int register_kretprobe(struct kretprobe *rp);
void unregister_kretprobe(struct kretprobe *rp);
2. Registration and unregistration
2.1 Register
To register a function-return probe, the user populates the following
fields in a kretprobe object and calls register_kretprobe() with the
kretprobe address as an argument:
kp.addr - the function's address
handler - this function is run after the ret instruction executes, but
before control returns to the return address in the caller.
maxactive - The maximum number of instances of the probed function that
can be active concurrently. For example, if the function is non-
recursive and is called with a spinlock or mutex held, maxactive = 1
should be enough. If the function is non-recursive and can never
relinquish the CPU (e.g., via a semaphore or preemption), NR_CPUS should
be enough. maxactive is used to determine how many kretprobe_instance
objects to allocate for this particular probed function. If maxactive <=
0, it is set to a default value (if CONFIG_PREEMPT maxactive=max(10, 2 *
NR_CPUS) else maxactive=NR_CPUS)
For example:
struct kretprobe rp;
rp.kp.addr = /* entrypoint address */
rp.handler = /*return probe handler */
rp.maxactive = /* e.g., 1 or NR_CPUS or 0, see the above explanation */
register_kretprobe(&rp);
The following field may also be of interest:
nmissed - Initialized to zero when the function-return probe is
registered, and incremented every time the probed function is entered but
there is no kretprobe_instance object available for establishing the
function-return probe (i.e., because maxactive was set too low).
2.2 Unregister
To unregiter a function-return probe, the user calls
unregister_kretprobe() with the same kretprobe object as registered
previously. If a probed function is running when the return probe is
unregistered, the function will return as expected, but the handler won't
be run.
3. Limitations
3.1 This patch supports only the i386 architecture, but patches for
x86_64 and ppc64 are anticipated soon.
3.2 Return probes operates by replacing the return address in the stack
(or in a known register, such as the lr register for ppc). This may
cause __builtin_return_address(0), when invoked from the return-probed
function, to return the address of the return-probes trampoline.
3.3 This implementation uses the "Multiprobes at an address" feature in
2.6.12-rc3-mm3.
3.4 Due to a limitation in multi-probes, you cannot currently establish
a return probe and a jprobe on the same function. A patch to remove
this limitation is being tested.
This feature is required by SystemTap (http://sourceware.org/systemtap),
and reflects ideas contributed by several SystemTap developers, including
Will Cohen and Ananth Mavinakayanahalli.
Signed-off-by: Hien Nguyen <hien@us.ibm.com>
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
Signed-off-by: Frederik Deweerdt <frederik.deweerdt@laposte.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 03:09:19 -04:00
|
|
|
EXPORT_SYMBOL_GPL(register_kretprobe);
|
|
|
|
EXPORT_SYMBOL_GPL(unregister_kretprobe);
|
|
|
|
|