openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
linux/arch/x86_64/kernel/kprobes.c

694 lines
20 KiB
C
Raw Normal View History

Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Kernel Probes (KProbes)
* arch/x86_64/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 contributions from
* Rusty Russell).
* 2004-July Suparna Bhattacharya <suparna@in.ibm.com> added jumper probes
* interface to access function arguments.
* 2004-Oct Jim Keniston <kenistoj@us.ibm.com> and Prasanna S Panchamukhi
* <prasanna@in.ibm.com> adapted for x86_64
* 2005-Mar Roland McGrath <roland@redhat.com>
* Fixed to handle %rip-relative addressing mode correctly.
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
* 2005-May Rusty Lynch <rusty.lynch@intel.com>
* Added function return probes functionality
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*/
#include <linux/config.h>
#include <linux/kprobes.h>
#include <linux/ptrace.h>
#include <linux/spinlock.h>
#include <linux/string.h>
#include <linux/slab.h>
#include <linux/preempt.h>
[PATCH] kprobes: fix single-step out of line - take2 Now that PPC64 has no-execute support, here is a second try to fix the single step out of line during kprobe execution. Kprobes on x86_64 already solved this problem by allocating an executable page and using it as the scratch area for stepping out of line. Reuse that. Signed-off-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-28 06:17:01 +08:00
[PATCH] Move kprobe [dis]arming into arch specific code The architecture independent code of the current kprobes implementation is arming and disarming kprobes at registration time. The problem is that the code is assuming that arming and disarming is a just done by a simple write of some magic value to an address. This is problematic for ia64 where our instructions look more like structures, and we can not insert break points by just doing something like: *p->addr = BREAKPOINT_INSTRUCTION; The following patch to 2.6.12-rc4-mm2 adds two new architecture dependent functions: * void arch_arm_kprobe(struct kprobe *p) * void arch_disarm_kprobe(struct kprobe *p) and then adds the new functions for each of the architectures that already implement kprobes (spar64/ppc64/i386/x86_64). I thought arch_[dis]arm_kprobe was the most descriptive of what was really happening, but each of the architectures already had a disarm_kprobe() function that was really a "disarm and do some other clean-up items as needed when you stumble across a recursive kprobe." So... I took the liberty of changing the code that was calling disarm_kprobe() to call arch_disarm_kprobe(), and then do the cleanup in the block of code dealing with the recursive kprobe case. So far this patch as been tested on i386, x86_64, and ppc64, but still needs to be tested in sparc64. Signed-off-by: Rusty Lynch <rusty.lynch@intel.com> Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:25 +08:00
#include <asm/cacheflush.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
#include <asm/pgtable.h>
#include <asm/kdebug.h>
static DECLARE_MUTEX(kprobe_mutex);
static struct kprobe *current_kprobe;
static unsigned long kprobe_status, kprobe_old_rflags, kprobe_saved_rflags;
[PATCH] kprobes: Temporary disarming of reentrant probe for x86_64 This patch includes x86_64 architecture specific changes to support temporary disarming on reentrancy of probes. Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:37 +08:00
static struct kprobe *kprobe_prev;
static unsigned long kprobe_status_prev, kprobe_old_rflags_prev, kprobe_saved_rflags_prev;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
static struct pt_regs jprobe_saved_regs;
static long *jprobe_saved_rsp;
void jprobe_return_end(void);
/* copy of the kernel stack at the probe fire time */
static kprobe_opcode_t jprobes_stack[MAX_STACK_SIZE];
/*
* returns non-zero if opcode modifies the interrupt flag.
*/
static inline int is_IF_modifier(kprobe_opcode_t *insn)
{
switch (*insn) {
case 0xfa: /* cli */
case 0xfb: /* sti */
case 0xcf: /* iret/iretd */
case 0x9d: /* popf/popfd */
return 1;
}
if (*insn >= 0x40 && *insn <= 0x4f && *++insn == 0xcf)
return 1;
return 0;
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes arch_prepare_kprobe(struct kprobe *p)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
/* insn: must be on special executable page on x86_64. */
up(&kprobe_mutex);
p->ainsn.insn = get_insn_slot();
down(&kprobe_mutex);
if (!p->ainsn.insn) {
return -ENOMEM;
}
return 0;
}
/*
* Determine if the instruction uses the %rip-relative addressing mode.
* If it does, return the address of the 32-bit displacement word.
* If not, return null.
*/
static inline s32 *is_riprel(u8 *insn)
{
#define W(row,b0,b1,b2,b3,b4,b5,b6,b7,b8,b9,ba,bb,bc,bd,be,bf) \
(((b0##UL << 0x0)|(b1##UL << 0x1)|(b2##UL << 0x2)|(b3##UL << 0x3) | \
(b4##UL << 0x4)|(b5##UL << 0x5)|(b6##UL << 0x6)|(b7##UL << 0x7) | \
(b8##UL << 0x8)|(b9##UL << 0x9)|(ba##UL << 0xa)|(bb##UL << 0xb) | \
(bc##UL << 0xc)|(bd##UL << 0xd)|(be##UL << 0xe)|(bf##UL << 0xf)) \
<< (row % 64))
static const u64 onebyte_has_modrm[256 / 64] = {
/* 0 1 2 3 4 5 6 7 8 9 a b c d e f */
/* ------------------------------- */
W(0x00, 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0)| /* 00 */
W(0x10, 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0)| /* 10 */
W(0x20, 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0)| /* 20 */
W(0x30, 1,1,1,1,0,0,0,0,1,1,1,1,0,0,0,0), /* 30 */
W(0x40, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)| /* 40 */
W(0x50, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)| /* 50 */
W(0x60, 0,0,1,1,0,0,0,0,0,1,0,1,0,0,0,0)| /* 60 */
W(0x70, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0), /* 70 */
W(0x80, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)| /* 80 */
W(0x90, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)| /* 90 */
W(0xa0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)| /* a0 */
W(0xb0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0), /* b0 */
W(0xc0, 1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0)| /* c0 */
W(0xd0, 1,1,1,1,0,0,0,0,1,1,1,1,1,1,1,1)| /* d0 */
W(0xe0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)| /* e0 */
W(0xf0, 0,0,0,0,0,0,1,1,0,0,0,0,0,0,1,1) /* f0 */
/* ------------------------------- */
/* 0 1 2 3 4 5 6 7 8 9 a b c d e f */
};
static const u64 twobyte_has_modrm[256 / 64] = {
/* 0 1 2 3 4 5 6 7 8 9 a b c d e f */
/* ------------------------------- */
W(0x00, 1,1,1,1,0,0,0,0,0,0,0,0,0,1,0,1)| /* 0f */
W(0x10, 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0)| /* 1f */
W(0x20, 1,1,1,1,1,0,1,0,1,1,1,1,1,1,1,1)| /* 2f */
W(0x30, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0), /* 3f */
W(0x40, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)| /* 4f */
W(0x50, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)| /* 5f */
W(0x60, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)| /* 6f */
W(0x70, 1,1,1,1,1,1,1,0,0,0,0,0,1,1,1,1), /* 7f */
W(0x80, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0)| /* 8f */
W(0x90, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)| /* 9f */
W(0xa0, 0,0,0,1,1,1,1,1,0,0,0,1,1,1,1,1)| /* af */
W(0xb0, 1,1,1,1,1,1,1,1,0,0,1,1,1,1,1,1), /* bf */
W(0xc0, 1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0)| /* cf */
W(0xd0, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)| /* df */
W(0xe0, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1)| /* ef */
W(0xf0, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0) /* ff */
/* ------------------------------- */
/* 0 1 2 3 4 5 6 7 8 9 a b c d e f */
};
#undef W
int need_modrm;
/* Skip legacy instruction prefixes. */
while (1) {
switch (*insn) {
case 0x66:
case 0x67:
case 0x2e:
case 0x3e:
case 0x26:
case 0x64:
case 0x65:
case 0x36:
case 0xf0:
case 0xf3:
case 0xf2:
++insn;
continue;
}
break;
}
/* Skip REX instruction prefix. */
if ((*insn & 0xf0) == 0x40)
++insn;
if (*insn == 0x0f) { /* Two-byte opcode. */
++insn;
need_modrm = test_bit(*insn, twobyte_has_modrm);
} else { /* One-byte opcode. */
need_modrm = test_bit(*insn, onebyte_has_modrm);
}
if (need_modrm) {
u8 modrm = *++insn;
if ((modrm & 0xc7) == 0x05) { /* %rip+disp32 addressing mode */
/* Displacement follows ModRM byte. */
return (s32 *) ++insn;
}
}
/* No %rip-relative addressing mode here. */
return NULL;
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
void __kprobes arch_copy_kprobe(struct kprobe *p)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
s32 *ripdisp;
memcpy(p->ainsn.insn, p->addr, MAX_INSN_SIZE);
ripdisp = is_riprel(p->ainsn.insn);
if (ripdisp) {
/*
* The copied instruction uses the %rip-relative
* addressing mode. Adjust the displacement for the
* difference between the original location of this
* instruction and the location of the copy that will
* actually be run. The tricky bit here is making sure
* that the sign extension happens correctly in this
* calculation, since we need a signed 32-bit result to
* be sign-extended to 64 bits when it's added to the
* %rip value and yield the same 64-bit result that the
* sign-extension of the original signed 32-bit
* displacement would have given.
*/
s64 disp = (u8 *) p->addr + *ripdisp - (u8 *) p->ainsn.insn;
BUG_ON((s64) (s32) disp != disp); /* Sanity check. */
*ripdisp = disp;
}
[PATCH] Move kprobe [dis]arming into arch specific code The architecture independent code of the current kprobes implementation is arming and disarming kprobes at registration time. The problem is that the code is assuming that arming and disarming is a just done by a simple write of some magic value to an address. This is problematic for ia64 where our instructions look more like structures, and we can not insert break points by just doing something like: *p->addr = BREAKPOINT_INSTRUCTION; The following patch to 2.6.12-rc4-mm2 adds two new architecture dependent functions: * void arch_arm_kprobe(struct kprobe *p) * void arch_disarm_kprobe(struct kprobe *p) and then adds the new functions for each of the architectures that already implement kprobes (spar64/ppc64/i386/x86_64). I thought arch_[dis]arm_kprobe was the most descriptive of what was really happening, but each of the architectures already had a disarm_kprobe() function that was really a "disarm and do some other clean-up items as needed when you stumble across a recursive kprobe." So... I took the liberty of changing the code that was calling disarm_kprobe() to call arch_disarm_kprobe(), and then do the cleanup in the block of code dealing with the recursive kprobe case. So far this patch as been tested on i386, x86_64, and ppc64, but still needs to be tested in sparc64. Signed-off-by: Rusty Lynch <rusty.lynch@intel.com> Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:25 +08:00
p->opcode = *p->addr;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
void __kprobes arch_arm_kprobe(struct kprobe *p)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
[PATCH] Move kprobe [dis]arming into arch specific code The architecture independent code of the current kprobes implementation is arming and disarming kprobes at registration time. The problem is that the code is assuming that arming and disarming is a just done by a simple write of some magic value to an address. This is problematic for ia64 where our instructions look more like structures, and we can not insert break points by just doing something like: *p->addr = BREAKPOINT_INSTRUCTION; The following patch to 2.6.12-rc4-mm2 adds two new architecture dependent functions: * void arch_arm_kprobe(struct kprobe *p) * void arch_disarm_kprobe(struct kprobe *p) and then adds the new functions for each of the architectures that already implement kprobes (spar64/ppc64/i386/x86_64). I thought arch_[dis]arm_kprobe was the most descriptive of what was really happening, but each of the architectures already had a disarm_kprobe() function that was really a "disarm and do some other clean-up items as needed when you stumble across a recursive kprobe." So... I took the liberty of changing the code that was calling disarm_kprobe() to call arch_disarm_kprobe(), and then do the cleanup in the block of code dealing with the recursive kprobe case. So far this patch as been tested on i386, x86_64, and ppc64, but still needs to be tested in sparc64. Signed-off-by: Rusty Lynch <rusty.lynch@intel.com> Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:25 +08:00
*p->addr = BREAKPOINT_INSTRUCTION;
flush_icache_range((unsigned long) p->addr,
(unsigned long) p->addr + sizeof(kprobe_opcode_t));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
void __kprobes arch_disarm_kprobe(struct kprobe *p)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
*p->addr = p->opcode;
[PATCH] Move kprobe [dis]arming into arch specific code The architecture independent code of the current kprobes implementation is arming and disarming kprobes at registration time. The problem is that the code is assuming that arming and disarming is a just done by a simple write of some magic value to an address. This is problematic for ia64 where our instructions look more like structures, and we can not insert break points by just doing something like: *p->addr = BREAKPOINT_INSTRUCTION; The following patch to 2.6.12-rc4-mm2 adds two new architecture dependent functions: * void arch_arm_kprobe(struct kprobe *p) * void arch_disarm_kprobe(struct kprobe *p) and then adds the new functions for each of the architectures that already implement kprobes (spar64/ppc64/i386/x86_64). I thought arch_[dis]arm_kprobe was the most descriptive of what was really happening, but each of the architectures already had a disarm_kprobe() function that was really a "disarm and do some other clean-up items as needed when you stumble across a recursive kprobe." So... I took the liberty of changing the code that was calling disarm_kprobe() to call arch_disarm_kprobe(), and then do the cleanup in the block of code dealing with the recursive kprobe case. So far this patch as been tested on i386, x86_64, and ppc64, but still needs to be tested in sparc64. Signed-off-by: Rusty Lynch <rusty.lynch@intel.com> Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:25 +08:00
flush_icache_range((unsigned long) p->addr,
(unsigned long) p->addr + sizeof(kprobe_opcode_t));
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
void __kprobes arch_remove_kprobe(struct kprobe *p)
[PATCH] Move kprobe [dis]arming into arch specific code The architecture independent code of the current kprobes implementation is arming and disarming kprobes at registration time. The problem is that the code is assuming that arming and disarming is a just done by a simple write of some magic value to an address. This is problematic for ia64 where our instructions look more like structures, and we can not insert break points by just doing something like: *p->addr = BREAKPOINT_INSTRUCTION; The following patch to 2.6.12-rc4-mm2 adds two new architecture dependent functions: * void arch_arm_kprobe(struct kprobe *p) * void arch_disarm_kprobe(struct kprobe *p) and then adds the new functions for each of the architectures that already implement kprobes (spar64/ppc64/i386/x86_64). I thought arch_[dis]arm_kprobe was the most descriptive of what was really happening, but each of the architectures already had a disarm_kprobe() function that was really a "disarm and do some other clean-up items as needed when you stumble across a recursive kprobe." So... I took the liberty of changing the code that was calling disarm_kprobe() to call arch_disarm_kprobe(), and then do the cleanup in the block of code dealing with the recursive kprobe case. So far this patch as been tested on i386, x86_64, and ppc64, but still needs to be tested in sparc64. Signed-off-by: Rusty Lynch <rusty.lynch@intel.com> Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:25 +08:00
{
up(&kprobe_mutex);
free_insn_slot(p->ainsn.insn);
down(&kprobe_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
[PATCH] kprobes: Temporary disarming of reentrant probe for x86_64 This patch includes x86_64 architecture specific changes to support temporary disarming on reentrancy of probes. Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:37 +08:00
static inline void save_previous_kprobe(void)
{
kprobe_prev = current_kprobe;
kprobe_status_prev = kprobe_status;
kprobe_old_rflags_prev = kprobe_old_rflags;
kprobe_saved_rflags_prev = kprobe_saved_rflags;
}
static inline void restore_previous_kprobe(void)
{
current_kprobe = kprobe_prev;
kprobe_status = kprobe_status_prev;
kprobe_old_rflags = kprobe_old_rflags_prev;
kprobe_saved_rflags = kprobe_saved_rflags_prev;
}
static inline void set_current_kprobe(struct kprobe *p, struct pt_regs *regs)
{
current_kprobe = p;
kprobe_saved_rflags = kprobe_old_rflags
= (regs->eflags & (TF_MASK | IF_MASK));
if (is_IF_modifier(p->ainsn.insn))
kprobe_saved_rflags &= ~IF_MASK;
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
regs->eflags |= TF_MASK;
regs->eflags &= ~IF_MASK;
/*single step inline if the instruction is an int3*/
if (p->opcode == BREAKPOINT_INSTRUCTION)
regs->rip = (unsigned long)p->addr;
else
regs->rip = (unsigned long)p->ainsn.insn;
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
void __kprobes arch_prepare_kretprobe(struct kretprobe *rp,
struct pt_regs *regs)
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
{
unsigned long *sara = (unsigned long *)regs->rsp;
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
struct kretprobe_instance *ri;
if ((ri = get_free_rp_inst(rp)) != NULL) {
ri->rp = rp;
ri->task = current;
ri->ret_addr = (kprobe_opcode_t *) *sara;
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
/* Replace the return addr with trampoline addr */
*sara = (unsigned long) &kretprobe_trampoline;
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
add_rp_inst(ri);
} else {
rp->nmissed++;
}
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Interrupts are disabled on entry as trap3 is an interrupt gate and they
* remain disabled thorough out this function.
*/
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes kprobe_handler(struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct kprobe *p;
int ret = 0;
kprobe_opcode_t *addr = (kprobe_opcode_t *)(regs->rip - sizeof(kprobe_opcode_t));
/* We're in an interrupt, but this is clear and BUG()-safe. */
preempt_disable();
/* Check we're not actually recursing */
if (kprobe_running()) {
/* We *are* holding lock here, so this is safe.
Disarm the probe we just hit, and ignore it. */
p = get_kprobe(addr);
if (p) {
[PATCH] kprobes: fix bug when probed on task and isr functions This patch fixes a race condition where in system used to hang or sometime crash within minutes when kprobes are inserted on ISR routine and a task routine. The fix has been stress tested on i386, ia64, pp64 and on x86_64. To reproduce the problem insert kprobes on schedule() and do_IRQ() functions and you should see hang or system crash. Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> Signed-off-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Acked-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:35 +08:00
if (kprobe_status == KPROBE_HIT_SS &&
*p->ainsn.insn == BREAKPOINT_INSTRUCTION) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
regs->eflags &= ~TF_MASK;
regs->eflags |= kprobe_saved_rflags;
unlock_kprobes();
goto no_kprobe;
[PATCH] kprobes: Temporary disarming of reentrant probe for x86_64 This patch includes x86_64 architecture specific changes to support temporary disarming on reentrancy of probes. Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:37 +08:00
} else if (kprobe_status == KPROBE_HIT_SSDONE) {
/* TODO: Provide re-entrancy from
* post_kprobes_handler() and avoid exception
* stack corruption while single-stepping on
* the instruction of the new probe.
*/
arch_disarm_kprobe(p);
regs->rip = (unsigned long)p->addr;
ret = 1;
} else {
/* We have reentered the kprobe_handler(), since
* another probe was hit while within the
* handler. We here save the original kprobe
* variables and just single step on instruction
* of the new probe without calling any user
* handlers.
*/
save_previous_kprobe();
set_current_kprobe(p, regs);
p->nmissed++;
prepare_singlestep(p, regs);
kprobe_status = KPROBE_REENTER;
return 1;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
}
} else {
p = current_kprobe;
if (p->break_handler && p->break_handler(p, regs)) {
goto ss_probe;
}
}
/* If it's not ours, can't be delete race, (we hold lock). */
goto no_kprobe;
}
lock_kprobes();
p = get_kprobe(addr);
if (!p) {
unlock_kprobes();
if (*addr != BREAKPOINT_INSTRUCTION) {
/*
* The breakpoint instruction was removed right
* after we hit it. Another cpu has removed
* either a probepoint or a debugger breakpoint
* at this address. In either case, no further
* handling of this interrupt is appropriate.
[PATCH] kprobes: fix handling of simultaneous probe hit/unregister This patch fixes a bug in kprobes's handling of a corner case on i386 and x86_64. On an SMP system, if one CPU unregisters a kprobe just after another CPU hits that probepoint, kprobe_handler() on the latter CPU sees that the kprobe has been unregistered, and attempts to let the CPU continue as if the probepoint hadn't been hit. The bug is that on i386 and x86_64, we were neglecting to set the IP back to the beginning of the probed instruction. This could cause an oops or crash. This bug doesn't exist on ppc64 and ia64, where a breakpoint instruction leaves the IP pointing to the beginning of the instruction. I don't know about sparc64. (Dave, could you please advise?) This fix has been tested on i386 and x86_64 SMP systems. To reproduce the problem, set one CPU to work registering and unregistering a kprobe repeatedly, and another CPU pounding the probepoint in a tight loop. Acked-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Jim Keniston <jkenisto@us.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:34 +08:00
* Back up over the (now missing) int3 and run
* the original instruction.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
*/
[PATCH] kprobes: fix handling of simultaneous probe hit/unregister This patch fixes a bug in kprobes's handling of a corner case on i386 and x86_64. On an SMP system, if one CPU unregisters a kprobe just after another CPU hits that probepoint, kprobe_handler() on the latter CPU sees that the kprobe has been unregistered, and attempts to let the CPU continue as if the probepoint hadn't been hit. The bug is that on i386 and x86_64, we were neglecting to set the IP back to the beginning of the probed instruction. This could cause an oops or crash. This bug doesn't exist on ppc64 and ia64, where a breakpoint instruction leaves the IP pointing to the beginning of the instruction. I don't know about sparc64. (Dave, could you please advise?) This fix has been tested on i386 and x86_64 SMP systems. To reproduce the problem, set one CPU to work registering and unregistering a kprobe repeatedly, and another CPU pounding the probepoint in a tight loop. Acked-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Jim Keniston <jkenisto@us.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:34 +08:00
regs->rip = (unsigned long)addr;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
ret = 1;
}
/* Not one of ours: let kernel handle it */
goto no_kprobe;
}
kprobe_status = KPROBE_HIT_ACTIVE;
[PATCH] kprobes: Temporary disarming of reentrant probe for x86_64 This patch includes x86_64 architecture specific changes to support temporary disarming on reentrancy of probes. Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:37 +08:00
set_current_kprobe(p, regs);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
if (p->pre_handler && p->pre_handler(p, regs))
/* handler has already set things up, so skip ss setup */
return 1;
ss_probe:
prepare_singlestep(p, regs);
kprobe_status = KPROBE_HIT_SS;
return 1;
no_kprobe:
preempt_enable_no_resched();
return ret;
}
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
/*
* For function-return probes, init_kprobes() establishes a probepoint
* here. When a retprobed function returns, this probe is hit and
* trampoline_probe_handler() runs, calling the kretprobe's handler.
*/
void kretprobe_trampoline_holder(void)
{
asm volatile ( ".global kretprobe_trampoline\n"
"kretprobe_trampoline: \n"
"nop\n");
}
/*
* Called when we hit the probe point at kretprobe_trampoline
*/
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
{
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
struct kretprobe_instance *ri = NULL;
struct hlist_head *head;
struct hlist_node *node, *tmp;
unsigned long orig_ret_address = 0;
unsigned long trampoline_address =(unsigned long)&kretprobe_trampoline;
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
head = kretprobe_inst_table_head(current);
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
/*
* It is possible to have multiple instances associated with a given
* task either because an multiple functions in the call path
* have a return probe installed on them, and/or more then one return
* return probe was registered for a target function.
*
* We can handle this because:
* - instances are always inserted at the head of the list
* - when multiple return probes are registered for the same
* function, the first instance's ret_addr will point to the
* real return address, and all the rest will point to
* kretprobe_trampoline
*/
hlist_for_each_entry_safe(ri, node, tmp, head, hlist) {
if (ri->task != current)
/* another task is sharing our hash bucket */
continue;
if (ri->rp && ri->rp->handler)
ri->rp->handler(ri, regs);
orig_ret_address = (unsigned long)ri->ret_addr;
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
recycle_rp_inst(ri);
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
if (orig_ret_address != trampoline_address)
/*
* This is the real return address. Any other
* instances associated with this task are for
* other calls deeper on the call stack
*/
break;
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
}
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
BUG_ON(!orig_ret_address || (orig_ret_address == trampoline_address));
regs->rip = orig_ret_address;
unlock_kprobes();
preempt_enable_no_resched();
/*
* By returning a non-zero value, we are telling
* kprobe_handler() that we have handled unlocking
* and re-enabling preemption.
*/
return 1;
[PATCH] x86_64 specific function return probes The following patch adds the x86_64 architecture specific implementation for function return probes. Function return probes is a mechanism built on top of kprobes that allows a caller to register a handler to be called when a given function exits. For example, to instrument the return path of sys_mkdir: static int sys_mkdir_exit(struct kretprobe_instance *i, struct pt_regs *regs) { printk("sys_mkdir exited\n"); return 0; } static struct kretprobe return_probe = { .handler = sys_mkdir_exit, }; <inside setup function> return_probe.kp.addr = (kprobe_opcode_t *) kallsyms_lookup_name("sys_mkdir"); if (register_kretprobe(&return_probe)) { printk(KERN_DEBUG "Unable to register return probe!\n"); /* do error path */ } <inside cleanup function> unregister_kretprobe(&return_probe); The way this works is that: * At system initialization time, kernel/kprobes.c installs a kprobe on a function called kretprobe_trampoline() that is implemented in the arch/x86_64/kernel/kprobes.c (More on this later) * When a return probe is registered using register_kretprobe(), kernel/kprobes.c will install a kprobe on the first instruction of the targeted function with the pre handler set to arch_prepare_kretprobe() which is implemented in arch/x86_64/kernel/kprobes.c. * arch_prepare_kretprobe() will prepare a kretprobe instance that stores: - nodes for hanging this instance in an empty or free list - a pointer to the return probe - the original return address - a pointer to the stack address With all this stowed away, arch_prepare_kretprobe() then sets the return address for the targeted function to a special trampoline function called kretprobe_trampoline() implemented in arch/x86_64/kernel/kprobes.c * The kprobe completes as normal, with control passing back to the target function that executes as normal, and eventually returns to our trampoline function. * Since a kprobe was installed on kretprobe_trampoline() during system initialization, control passes back to kprobes via the architecture specific function trampoline_probe_handler() which will lookup the instance in an hlist maintained by kernel/kprobes.c, and then call the handler function. * When trampoline_probe_handler() is done, the kprobes infrastructure single steps the original instruction (in this case just a top), and then calls trampoline_post_handler(). trampoline_post_handler() then looks up the instance again, puts the instance back on the free list, and then makes a long jump back to the original return instruction. So to recap, to instrument the exit path of a function this implementation will cause four interruptions: - A breakpoint at the very beginning of the function allowing us to switch out the return address - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) - A breakpoint in the trampoline function where our instrumented function returned to - A single step interruption to execute the original instruction that we replaced with the break instruction (normal kprobe flow) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:23 +08:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
/*
* Called after single-stepping. p->addr is the address of the
* instruction whose first byte has been replaced by the "int 3"
* instruction. To avoid the SMP problems that can occur when we
* temporarily put back the original opcode to single-step, we
* single-stepped a copy of the instruction. The address of this
* copy is p->ainsn.insn.
*
* This function prepares to return from the post-single-step
* interrupt. We have to fix up the stack as follows:
*
* 0) Except in the case of absolute or indirect jump or call instructions,
* the new rip is relative to the copied instruction. We need to make
* it relative to the original instruction.
*
* 1) If the single-stepped instruction was pushfl, then the TF and IF
* flags are set in the just-pushed eflags, and may need to be cleared.
*
* 2) If the single-stepped instruction was a call, the return address
* that is atop the stack is the address following the copied instruction.
* We need to make it the address following the original instruction.
*/
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
static void __kprobes resume_execution(struct kprobe *p, struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
unsigned long *tos = (unsigned long *)regs->rsp;
unsigned long next_rip = 0;
unsigned long copy_rip = (unsigned long)p->ainsn.insn;
unsigned long orig_rip = (unsigned long)p->addr;
kprobe_opcode_t *insn = p->ainsn.insn;
/*skip the REX prefix*/
if (*insn >= 0x40 && *insn <= 0x4f)
insn++;
switch (*insn) {
case 0x9c: /* pushfl */
*tos &= ~(TF_MASK | IF_MASK);
*tos |= kprobe_old_rflags;
break;
[PATCH] Kprobes: Incorrect handling of probes on ret/lret instruction Kprobes could not handle the insertion of a probe on the ret/lret instruction and used to oops after single stepping since kprobes was modifying eip/rip incorrectly. Adjustment of eip/rip is not required after single stepping in case of ret/lret instruction, because eip/rip points to the correct location after execution of the ret/lret instruction. This patch fixes the above problem. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-05-06 07:15:40 +08:00
case 0xc3: /* ret/lret */
case 0xcb:
case 0xc2:
case 0xca:
regs->eflags &= ~TF_MASK;
/* rip is already adjusted, no more changes required*/
return;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
case 0xe8: /* call relative - Fix return addr */
*tos = orig_rip + (*tos - copy_rip);
break;
case 0xff:
if ((*insn & 0x30) == 0x10) {
/* call absolute, indirect */
/* Fix return addr; rip is correct. */
next_rip = regs->rip;
*tos = orig_rip + (*tos - copy_rip);
} else if (((*insn & 0x31) == 0x20) || /* jmp near, absolute indirect */
((*insn & 0x31) == 0x21)) { /* jmp far, absolute indirect */
/* rip is correct. */
next_rip = regs->rip;
}
break;
case 0xea: /* jmp absolute -- rip is correct */
next_rip = regs->rip;
break;
default:
break;
}
regs->eflags &= ~TF_MASK;
if (next_rip) {
regs->rip = next_rip;
} else {
regs->rip = orig_rip + (regs->rip - copy_rip);
}
}
/*
* Interrupts are disabled on entry as trap1 is an interrupt gate and they
* remain disabled thoroughout this function. And we hold kprobe lock.
*/
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes post_kprobe_handler(struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
if (!kprobe_running())
return 0;
[PATCH] kprobes: Temporary disarming of reentrant probe for x86_64 This patch includes x86_64 architecture specific changes to support temporary disarming on reentrancy of probes. Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:37 +08:00
if ((kprobe_status != KPROBE_REENTER) && current_kprobe->post_handler) {
kprobe_status = KPROBE_HIT_SSDONE;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
current_kprobe->post_handler(current_kprobe, regs, 0);
[PATCH] kprobes: Temporary disarming of reentrant probe for x86_64 This patch includes x86_64 architecture specific changes to support temporary disarming on reentrancy of probes. Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:37 +08:00
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
resume_execution(current_kprobe, regs);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
regs->eflags |= kprobe_saved_rflags;
[PATCH] kprobes: Temporary disarming of reentrant probe for x86_64 This patch includes x86_64 architecture specific changes to support temporary disarming on reentrancy of probes. Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-23 15:09:37 +08:00
/* Restore the original saved kprobes variables and continue. */
if (kprobe_status == KPROBE_REENTER) {
restore_previous_kprobe();
goto out;
} else {
unlock_kprobes();
}
out:
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
preempt_enable_no_resched();
/*
* if somebody else is singlestepping across a probe point, eflags
* will have TF set, in which case, continue the remaining processing
* of do_debug, as if this is not a probe hit.
*/
if (regs->eflags & TF_MASK)
return 0;
return 1;
}
/* Interrupts disabled, kprobe_lock held. */
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
if (current_kprobe->fault_handler
&& current_kprobe->fault_handler(current_kprobe, regs, trapnr))
return 1;
if (kprobe_status & KPROBE_HIT_SS) {
resume_execution(current_kprobe, regs);
regs->eflags |= kprobe_old_rflags;
unlock_kprobes();
preempt_enable_no_resched();
}
return 0;
}
/*
* Wrapper routine for handling exceptions.
*/
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
unsigned long val, void *data)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct die_args *args = (struct die_args *)data;
switch (val) {
case DIE_INT3:
if (kprobe_handler(args->regs))
return NOTIFY_STOP;
break;
case DIE_DEBUG:
if (post_kprobe_handler(args->regs))
return NOTIFY_STOP;
break;
case DIE_GPF:
if (kprobe_running() &&
kprobe_fault_handler(args->regs, args->trapnr))
return NOTIFY_STOP;
break;
case DIE_PAGE_FAULT:
if (kprobe_running() &&
kprobe_fault_handler(args->regs, args->trapnr))
return NOTIFY_STOP;
break;
default:
break;
}
return NOTIFY_DONE;
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
struct jprobe *jp = container_of(p, struct jprobe, kp);
unsigned long addr;
jprobe_saved_regs = *regs;
jprobe_saved_rsp = (long *) regs->rsp;
addr = (unsigned long)jprobe_saved_rsp;
/*
* As Linus pointed out, gcc assumes that the callee
* owns the argument space and could overwrite it, e.g.
* tailcall optimization. So, to be absolutely safe
* we also save and restore enough stack bytes to cover
* the argument area.
*/
memcpy(jprobes_stack, (kprobe_opcode_t *) addr, MIN_STACK_SIZE(addr));
regs->eflags &= ~IF_MASK;
regs->rip = (unsigned long)(jp->entry);
return 1;
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
void __kprobes jprobe_return(void)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
preempt_enable_no_resched();
asm volatile (" xchg %%rbx,%%rsp \n"
" int3 \n"
" .globl jprobe_return_end \n"
" jprobe_return_end: \n"
" nop \n"::"b"
(jprobe_saved_rsp):"memory");
}
[PATCH] kprobes: prevent possible race conditions x86_64 changes This patch contains the x86_64 architecture specific changes to prevent the possible race conditions. Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-07 06:19:28 +08:00
int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-17 06:20:36 +08:00
{
u8 *addr = (u8 *) (regs->rip - 1);
unsigned long stack_addr = (unsigned long)jprobe_saved_rsp;
struct jprobe *jp = container_of(p, struct jprobe, kp);
if ((addr > (u8 *) jprobe_return) && (addr < (u8 *) jprobe_return_end)) {
if ((long *)regs->rsp != jprobe_saved_rsp) {
struct pt_regs *saved_regs =
container_of(jprobe_saved_rsp, struct pt_regs, rsp);
printk("current rsp %p does not match saved rsp %p\n",
(long *)regs->rsp, jprobe_saved_rsp);
printk("Saved registers for jprobe %p\n", jp);
show_registers(saved_regs);
printk("Current registers\n");
show_registers(regs);
BUG();
}
*regs = jprobe_saved_regs;
memcpy((kprobe_opcode_t *) stack_addr, jprobes_stack,
MIN_STACK_SIZE(stack_addr));
return 1;
}
return 0;
}
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
static struct kprobe trampoline_p = {
.addr = (kprobe_opcode_t *) &kretprobe_trampoline,
.pre_handler = trampoline_probe_handler
};
[PATCH] kprobes: fix namespace problem and sparc64 build The following renames arch_init, a kprobes function for performing any architecture specific initialization, to arch_init_kprobes in order to cleanup the namespace. Also, this patch adds arch_init_kprobes to sparc64 to fix the sparc64 kprobes build from the last return probe patch. Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-07-06 09:54:50 +08:00
int __init arch_init_kprobes(void)
[PATCH] Return probe redesign: x86_64 specific changes The following patch contains the x86_64 specific changes for the new return probe design. Changes include: * Removing the architecture specific functions for querying a return probe instance off a stack address * Complete rework onf arch_prepare_kretprobe() and trampoline_probe_handler() * Removing trampoline_post_handler() * Adding arch_init() so that now we handle registering the return probe trampoline instead of kernel/kprobes.c doing it NOTE: Note that with this new design, the dependency on calculating a pointer to the task off the stack pointer no longer exist (resolving the problem of interruption stacks as pointed out in the original feedback to this port.) 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>
2005-06-28 06:17:10 +08:00
{
return register_kprobe(&trampoline_p);
}