x86_64: prepare shared kernel/kprobes.c
authorThomas Gleixner <tglx@linutronix.de>
Thu, 11 Oct 2007 09:15:12 +0000 (11:15 +0200)
committerThomas Gleixner <tglx@linutronix.de>
Thu, 11 Oct 2007 09:15:12 +0000 (11:15 +0200)
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
arch/x86_64/kernel/Makefile
arch/x86_64/kernel/kprobes.c [deleted file]
arch/x86_64/kernel/kprobes_64.c [new file with mode: 0644]

index 976ea5e7222bdbde3a86eac3e66ce52c0ef68dd0..fd19c73a8da0a7fc66ff7e199561b14b62d318df 100644 (file)
@@ -32,7 +32,7 @@ obj-$(CONFIG_EARLY_PRINTK)    += early_printk.o
 obj-$(CONFIG_IOMMU)            += pci-gart.o aperture_64.o
 obj-$(CONFIG_CALGARY_IOMMU)    += pci-calgary_64.o tce_64.o
 obj-$(CONFIG_SWIOTLB)          += pci-swiotlb_64.o
-obj-$(CONFIG_KPROBES)          += kprobes.o
+obj-$(CONFIG_KPROBES)          += kprobes_64.o
 obj-$(CONFIG_X86_PM_TIMER)     += pmtimer_64.o
 obj-$(CONFIG_X86_VSMP)         += vsmp_64.o
 obj-$(CONFIG_K8_NB)            += k8.o
diff --git a/arch/x86_64/kernel/kprobes.c b/arch/x86_64/kernel/kprobes.c
deleted file mode 100644 (file)
index a30e004..0000000
+++ /dev/null
@@ -1,749 +0,0 @@
-/*
- *  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.
- * 2005-May     Rusty Lynch <rusty.lynch@intel.com>
- *              Added function return probes functionality
- */
-
-#include <linux/kprobes.h>
-#include <linux/ptrace.h>
-#include <linux/string.h>
-#include <linux/slab.h>
-#include <linux/preempt.h>
-#include <linux/module.h>
-#include <linux/kdebug.h>
-
-#include <asm/pgtable.h>
-#include <asm/uaccess.h>
-#include <asm/alternative.h>
-
-void jprobe_return_end(void);
-static void __kprobes arch_copy_kprobe(struct kprobe *p);
-
-DEFINE_PER_CPU(struct kprobe *, current_kprobe) = NULL;
-DEFINE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk);
-
-/*
- * returns non-zero if opcode modifies the interrupt flag.
- */
-static __always_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;
-}
-
-int __kprobes arch_prepare_kprobe(struct kprobe *p)
-{
-       /* insn: must be on special executable page on x86_64. */
-       p->ainsn.insn = get_insn_slot();
-       if (!p->ainsn.insn) {
-               return -ENOMEM;
-       }
-       arch_copy_kprobe(p);
-       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 s32 __kprobes *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;
-}
-
-static void __kprobes arch_copy_kprobe(struct kprobe *p)
-{
-       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;
-       }
-       p->opcode = *p->addr;
-}
-
-void __kprobes arch_arm_kprobe(struct kprobe *p)
-{
-       text_poke(p->addr, ((unsigned char []){BREAKPOINT_INSTRUCTION}), 1);
-}
-
-void __kprobes arch_disarm_kprobe(struct kprobe *p)
-{
-       text_poke(p->addr, &p->opcode, 1);
-}
-
-void __kprobes arch_remove_kprobe(struct kprobe *p)
-{
-       mutex_lock(&kprobe_mutex);
-       free_insn_slot(p->ainsn.insn, 0);
-       mutex_unlock(&kprobe_mutex);
-}
-
-static void __kprobes save_previous_kprobe(struct kprobe_ctlblk *kcb)
-{
-       kcb->prev_kprobe.kp = kprobe_running();
-       kcb->prev_kprobe.status = kcb->kprobe_status;
-       kcb->prev_kprobe.old_rflags = kcb->kprobe_old_rflags;
-       kcb->prev_kprobe.saved_rflags = kcb->kprobe_saved_rflags;
-}
-
-static void __kprobes restore_previous_kprobe(struct kprobe_ctlblk *kcb)
-{
-       __get_cpu_var(current_kprobe) = kcb->prev_kprobe.kp;
-       kcb->kprobe_status = kcb->prev_kprobe.status;
-       kcb->kprobe_old_rflags = kcb->prev_kprobe.old_rflags;
-       kcb->kprobe_saved_rflags = kcb->prev_kprobe.saved_rflags;
-}
-
-static void __kprobes set_current_kprobe(struct kprobe *p, struct pt_regs *regs,
-                               struct kprobe_ctlblk *kcb)
-{
-       __get_cpu_var(current_kprobe) = p;
-       kcb->kprobe_saved_rflags = kcb->kprobe_old_rflags
-               = (regs->eflags & (TF_MASK | IF_MASK));
-       if (is_IF_modifier(p->ainsn.insn))
-               kcb->kprobe_saved_rflags &= ~IF_MASK;
-}
-
-static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs)
-{
-       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;
-}
-
-/* Called with kretprobe_lock held */
-void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
-                                     struct pt_regs *regs)
-{
-       unsigned long *sara = (unsigned long *)regs->rsp;
-
-       ri->ret_addr = (kprobe_opcode_t *) *sara;
-       /* Replace the return addr with trampoline addr */
-       *sara = (unsigned long) &kretprobe_trampoline;
-}
-
-int __kprobes kprobe_handler(struct pt_regs *regs)
-{
-       struct kprobe *p;
-       int ret = 0;
-       kprobe_opcode_t *addr = (kprobe_opcode_t *)(regs->rip - sizeof(kprobe_opcode_t));
-       struct kprobe_ctlblk *kcb;
-
-       /*
-        * We don't want to be preempted for the entire
-        * duration of kprobe processing
-        */
-       preempt_disable();
-       kcb = get_kprobe_ctlblk();
-
-       /* Check we're not actually recursing */
-       if (kprobe_running()) {
-               p = get_kprobe(addr);
-               if (p) {
-                       if (kcb->kprobe_status == KPROBE_HIT_SS &&
-                               *p->ainsn.insn == BREAKPOINT_INSTRUCTION) {
-                               regs->eflags &= ~TF_MASK;
-                               regs->eflags |= kcb->kprobe_saved_rflags;
-                               goto no_kprobe;
-                       } else if (kcb->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;
-                               reset_current_kprobe();
-                               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(kcb);
-                               set_current_kprobe(p, regs, kcb);
-                               kprobes_inc_nmissed_count(p);
-                               prepare_singlestep(p, regs);
-                               kcb->kprobe_status = KPROBE_REENTER;
-                               return 1;
-                       }
-               } else {
-                       if (*addr != BREAKPOINT_INSTRUCTION) {
-                       /* The breakpoint instruction was removed by
-                        * another cpu right after we hit, no further
-                        * handling of this interrupt is appropriate
-                        */
-                               regs->rip = (unsigned long)addr;
-                               ret = 1;
-                               goto no_kprobe;
-                       }
-                       p = __get_cpu_var(current_kprobe);
-                       if (p->break_handler && p->break_handler(p, regs)) {
-                               goto ss_probe;
-                       }
-               }
-               goto no_kprobe;
-       }
-
-       p = get_kprobe(addr);
-       if (!p) {
-               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.
-                        * Back up over the (now missing) int3 and run
-                        * the original instruction.
-                        */
-                       regs->rip = (unsigned long)addr;
-                       ret = 1;
-               }
-               /* Not one of ours: let kernel handle it */
-               goto no_kprobe;
-       }
-
-       set_current_kprobe(p, regs, kcb);
-       kcb->kprobe_status = KPROBE_HIT_ACTIVE;
-
-       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);
-       kcb->kprobe_status = KPROBE_HIT_SS;
-       return 1;
-
-no_kprobe:
-       preempt_enable_no_resched();
-       return ret;
-}
-
-/*
- * 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
- */
-int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
-{
-       struct kretprobe_instance *ri = NULL;
-       struct hlist_head *head, empty_rp;
-       struct hlist_node *node, *tmp;
-       unsigned long flags, orig_ret_address = 0;
-       unsigned long trampoline_address =(unsigned long)&kretprobe_trampoline;
-
-       INIT_HLIST_HEAD(&empty_rp);
-       spin_lock_irqsave(&kretprobe_lock, flags);
-       head = kretprobe_inst_table_head(current);
-
-       /*
-        * 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;
-               recycle_rp_inst(ri, &empty_rp);
-
-               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;
-       }
-
-       kretprobe_assert(ri, orig_ret_address, trampoline_address);
-       regs->rip = orig_ret_address;
-
-       reset_current_kprobe();
-       spin_unlock_irqrestore(&kretprobe_lock, flags);
-       preempt_enable_no_resched();
-
-       hlist_for_each_entry_safe(ri, node, tmp, &empty_rp, hlist) {
-               hlist_del(&ri->hlist);
-               kfree(ri);
-       }
-       /*
-        * By returning a non-zero value, we are telling
-        * kprobe_handler() that we don't want the post_handler
-        * to run (and have re-enabled preemption)
-        */
-       return 1;
-}
-
-/*
- * 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.
- */
-static void __kprobes resume_execution(struct kprobe *p,
-               struct pt_regs *regs, struct kprobe_ctlblk *kcb)
-{
-       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 |= kcb->kprobe_old_rflags;
-               break;
-       case 0xc3:              /* ret/lret */
-       case 0xcb:
-       case 0xc2:
-       case 0xca:
-               regs->eflags &= ~TF_MASK;
-               /* rip is already adjusted, no more changes required*/
-               return;
-       case 0xe8:              /* call relative - Fix return addr */
-               *tos = orig_rip + (*tos - copy_rip);
-               break;
-       case 0xff:
-               if ((insn[1] & 0x30) == 0x10) {
-                       /* call absolute, indirect */
-                       /* Fix return addr; rip is correct. */
-                       next_rip = regs->rip;
-                       *tos = orig_rip + (*tos - copy_rip);
-               } else if (((insn[1] & 0x31) == 0x20) ||        /* jmp near, absolute indirect */
-                          ((insn[1] & 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);
-       }
-}
-
-int __kprobes post_kprobe_handler(struct pt_regs *regs)
-{
-       struct kprobe *cur = kprobe_running();
-       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
-
-       if (!cur)
-               return 0;
-
-       if ((kcb->kprobe_status != KPROBE_REENTER) && cur->post_handler) {
-               kcb->kprobe_status = KPROBE_HIT_SSDONE;
-               cur->post_handler(cur, regs, 0);
-       }
-
-       resume_execution(cur, regs, kcb);
-       regs->eflags |= kcb->kprobe_saved_rflags;
-
-       /* Restore the original saved kprobes variables and continue. */
-       if (kcb->kprobe_status == KPROBE_REENTER) {
-               restore_previous_kprobe(kcb);
-               goto out;
-       }
-       reset_current_kprobe();
-out:
-       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;
-}
-
-int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
-{
-       struct kprobe *cur = kprobe_running();
-       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
-       const struct exception_table_entry *fixup;
-
-       switch(kcb->kprobe_status) {
-       case KPROBE_HIT_SS:
-       case KPROBE_REENTER:
-               /*
-                * We are here because the instruction being single
-                * stepped caused a page fault. We reset the current
-                * kprobe and the rip points back to the probe address
-                * and allow the page fault handler to continue as a
-                * normal page fault.
-                */
-               regs->rip = (unsigned long)cur->addr;
-               regs->eflags |= kcb->kprobe_old_rflags;
-               if (kcb->kprobe_status == KPROBE_REENTER)
-                       restore_previous_kprobe(kcb);
-               else
-                       reset_current_kprobe();
-               preempt_enable_no_resched();
-               break;
-       case KPROBE_HIT_ACTIVE:
-       case KPROBE_HIT_SSDONE:
-               /*
-                * We increment the nmissed count for accounting,
-                * we can also use npre/npostfault count for accouting
-                * these specific fault cases.
-                */
-               kprobes_inc_nmissed_count(cur);
-
-               /*
-                * We come here because instructions in the pre/post
-                * handler caused the page_fault, this could happen
-                * if handler tries to access user space by
-                * copy_from_user(), get_user() etc. Let the
-                * user-specified handler try to fix it first.
-                */
-               if (cur->fault_handler && cur->fault_handler(cur, regs, trapnr))
-                       return 1;
-
-               /*
-                * In case the user-specified fault handler returned
-                * zero, try to fix up.
-                */
-               fixup = search_exception_tables(regs->rip);
-               if (fixup) {
-                       regs->rip = fixup->fixup;
-                       return 1;
-               }
-
-               /*
-                * fixup() could not handle it,
-                * Let do_page_fault() fix it.
-                */
-               break;
-       default:
-               break;
-       }
-       return 0;
-}
-
-/*
- * Wrapper routine for handling exceptions.
- */
-int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
-                                      unsigned long val, void *data)
-{
-       struct die_args *args = (struct die_args *)data;
-       int ret = NOTIFY_DONE;
-
-       if (args->regs && user_mode(args->regs))
-               return ret;
-
-       switch (val) {
-       case DIE_INT3:
-               if (kprobe_handler(args->regs))
-                       ret = NOTIFY_STOP;
-               break;
-       case DIE_DEBUG:
-               if (post_kprobe_handler(args->regs))
-                       ret = NOTIFY_STOP;
-               break;
-       case DIE_GPF:
-       case DIE_PAGE_FAULT:
-               /* kprobe_running() needs smp_processor_id() */
-               preempt_disable();
-               if (kprobe_running() &&
-                   kprobe_fault_handler(args->regs, args->trapnr))
-                       ret = NOTIFY_STOP;
-               preempt_enable();
-               break;
-       default:
-               break;
-       }
-       return ret;
-}
-
-int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs)
-{
-       struct jprobe *jp = container_of(p, struct jprobe, kp);
-       unsigned long addr;
-       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
-
-       kcb->jprobe_saved_regs = *regs;
-       kcb->jprobe_saved_rsp = (long *) regs->rsp;
-       addr = (unsigned long)(kcb->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(kcb->jprobes_stack, (kprobe_opcode_t *)addr,
-                       MIN_STACK_SIZE(addr));
-       regs->eflags &= ~IF_MASK;
-       regs->rip = (unsigned long)(jp->entry);
-       return 1;
-}
-
-void __kprobes jprobe_return(void)
-{
-       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
-
-       asm volatile ("       xchg   %%rbx,%%rsp     \n"
-                     "       int3                      \n"
-                     "       .globl jprobe_return_end  \n"
-                     "       jprobe_return_end:        \n"
-                     "       nop                       \n"::"b"
-                     (kcb->jprobe_saved_rsp):"memory");
-}
-
-int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
-{
-       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
-       u8 *addr = (u8 *) (regs->rip - 1);
-       unsigned long stack_addr = (unsigned long)(kcb->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 != kcb->jprobe_saved_rsp) {
-                       struct pt_regs *saved_regs =
-                           container_of(kcb->jprobe_saved_rsp,
-                                           struct pt_regs, rsp);
-                       printk("current rsp %p does not match saved rsp %p\n",
-                              (long *)regs->rsp, kcb->jprobe_saved_rsp);
-                       printk("Saved registers for jprobe %p\n", jp);
-                       show_registers(saved_regs);
-                       printk("Current registers\n");
-                       show_registers(regs);
-                       BUG();
-               }
-               *regs = kcb->jprobe_saved_regs;
-               memcpy((kprobe_opcode_t *) stack_addr, kcb->jprobes_stack,
-                      MIN_STACK_SIZE(stack_addr));
-               preempt_enable_no_resched();
-               return 1;
-       }
-       return 0;
-}
-
-static struct kprobe trampoline_p = {
-       .addr = (kprobe_opcode_t *) &kretprobe_trampoline,
-       .pre_handler = trampoline_probe_handler
-};
-
-int __init arch_init_kprobes(void)
-{
-       return register_kprobe(&trampoline_p);
-}
-
-int __kprobes arch_trampoline_kprobe(struct kprobe *p)
-{
-       if (p->addr == (kprobe_opcode_t *)&kretprobe_trampoline)
-               return 1;
-
-       return 0;
-}
diff --git a/arch/x86_64/kernel/kprobes_64.c b/arch/x86_64/kernel/kprobes_64.c
new file mode 100644 (file)
index 0000000..a30e004
--- /dev/null
@@ -0,0 +1,749 @@
+/*
+ *  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.
+ * 2005-May     Rusty Lynch <rusty.lynch@intel.com>
+ *              Added function return probes functionality
+ */
+
+#include <linux/kprobes.h>
+#include <linux/ptrace.h>
+#include <linux/string.h>
+#include <linux/slab.h>
+#include <linux/preempt.h>
+#include <linux/module.h>
+#include <linux/kdebug.h>
+
+#include <asm/pgtable.h>
+#include <asm/uaccess.h>
+#include <asm/alternative.h>
+
+void jprobe_return_end(void);
+static void __kprobes arch_copy_kprobe(struct kprobe *p);
+
+DEFINE_PER_CPU(struct kprobe *, current_kprobe) = NULL;
+DEFINE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk);
+
+/*
+ * returns non-zero if opcode modifies the interrupt flag.
+ */
+static __always_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;
+}
+
+int __kprobes arch_prepare_kprobe(struct kprobe *p)
+{
+       /* insn: must be on special executable page on x86_64. */
+       p->ainsn.insn = get_insn_slot();
+       if (!p->ainsn.insn) {
+               return -ENOMEM;
+       }
+       arch_copy_kprobe(p);
+       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 s32 __kprobes *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;
+}
+
+static void __kprobes arch_copy_kprobe(struct kprobe *p)
+{
+       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;
+       }
+       p->opcode = *p->addr;
+}
+
+void __kprobes arch_arm_kprobe(struct kprobe *p)
+{
+       text_poke(p->addr, ((unsigned char []){BREAKPOINT_INSTRUCTION}), 1);
+}
+
+void __kprobes arch_disarm_kprobe(struct kprobe *p)
+{
+       text_poke(p->addr, &p->opcode, 1);
+}
+
+void __kprobes arch_remove_kprobe(struct kprobe *p)
+{
+       mutex_lock(&kprobe_mutex);
+       free_insn_slot(p->ainsn.insn, 0);
+       mutex_unlock(&kprobe_mutex);
+}
+
+static void __kprobes save_previous_kprobe(struct kprobe_ctlblk *kcb)
+{
+       kcb->prev_kprobe.kp = kprobe_running();
+       kcb->prev_kprobe.status = kcb->kprobe_status;
+       kcb->prev_kprobe.old_rflags = kcb->kprobe_old_rflags;
+       kcb->prev_kprobe.saved_rflags = kcb->kprobe_saved_rflags;
+}
+
+static void __kprobes restore_previous_kprobe(struct kprobe_ctlblk *kcb)
+{
+       __get_cpu_var(current_kprobe) = kcb->prev_kprobe.kp;
+       kcb->kprobe_status = kcb->prev_kprobe.status;
+       kcb->kprobe_old_rflags = kcb->prev_kprobe.old_rflags;
+       kcb->kprobe_saved_rflags = kcb->prev_kprobe.saved_rflags;
+}
+
+static void __kprobes set_current_kprobe(struct kprobe *p, struct pt_regs *regs,
+                               struct kprobe_ctlblk *kcb)
+{
+       __get_cpu_var(current_kprobe) = p;
+       kcb->kprobe_saved_rflags = kcb->kprobe_old_rflags
+               = (regs->eflags & (TF_MASK | IF_MASK));
+       if (is_IF_modifier(p->ainsn.insn))
+               kcb->kprobe_saved_rflags &= ~IF_MASK;
+}
+
+static void __kprobes prepare_singlestep(struct kprobe *p, struct pt_regs *regs)
+{
+       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;
+}
+
+/* Called with kretprobe_lock held */
+void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
+                                     struct pt_regs *regs)
+{
+       unsigned long *sara = (unsigned long *)regs->rsp;
+
+       ri->ret_addr = (kprobe_opcode_t *) *sara;
+       /* Replace the return addr with trampoline addr */
+       *sara = (unsigned long) &kretprobe_trampoline;
+}
+
+int __kprobes kprobe_handler(struct pt_regs *regs)
+{
+       struct kprobe *p;
+       int ret = 0;
+       kprobe_opcode_t *addr = (kprobe_opcode_t *)(regs->rip - sizeof(kprobe_opcode_t));
+       struct kprobe_ctlblk *kcb;
+
+       /*
+        * We don't want to be preempted for the entire
+        * duration of kprobe processing
+        */
+       preempt_disable();
+       kcb = get_kprobe_ctlblk();
+
+       /* Check we're not actually recursing */
+       if (kprobe_running()) {
+               p = get_kprobe(addr);
+               if (p) {
+                       if (kcb->kprobe_status == KPROBE_HIT_SS &&
+                               *p->ainsn.insn == BREAKPOINT_INSTRUCTION) {
+                               regs->eflags &= ~TF_MASK;
+                               regs->eflags |= kcb->kprobe_saved_rflags;
+                               goto no_kprobe;
+                       } else if (kcb->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;
+                               reset_current_kprobe();
+                               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(kcb);
+                               set_current_kprobe(p, regs, kcb);
+                               kprobes_inc_nmissed_count(p);
+                               prepare_singlestep(p, regs);
+                               kcb->kprobe_status = KPROBE_REENTER;
+                               return 1;
+                       }
+               } else {
+                       if (*addr != BREAKPOINT_INSTRUCTION) {
+                       /* The breakpoint instruction was removed by
+                        * another cpu right after we hit, no further
+                        * handling of this interrupt is appropriate
+                        */
+                               regs->rip = (unsigned long)addr;
+                               ret = 1;
+                               goto no_kprobe;
+                       }
+                       p = __get_cpu_var(current_kprobe);
+                       if (p->break_handler && p->break_handler(p, regs)) {
+                               goto ss_probe;
+                       }
+               }
+               goto no_kprobe;
+       }
+
+       p = get_kprobe(addr);
+       if (!p) {
+               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.
+                        * Back up over the (now missing) int3 and run
+                        * the original instruction.
+                        */
+                       regs->rip = (unsigned long)addr;
+                       ret = 1;
+               }
+               /* Not one of ours: let kernel handle it */
+               goto no_kprobe;
+       }
+
+       set_current_kprobe(p, regs, kcb);
+       kcb->kprobe_status = KPROBE_HIT_ACTIVE;
+
+       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);
+       kcb->kprobe_status = KPROBE_HIT_SS;
+       return 1;
+
+no_kprobe:
+       preempt_enable_no_resched();
+       return ret;
+}
+
+/*
+ * 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
+ */
+int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
+{
+       struct kretprobe_instance *ri = NULL;
+       struct hlist_head *head, empty_rp;
+       struct hlist_node *node, *tmp;
+       unsigned long flags, orig_ret_address = 0;
+       unsigned long trampoline_address =(unsigned long)&kretprobe_trampoline;
+
+       INIT_HLIST_HEAD(&empty_rp);
+       spin_lock_irqsave(&kretprobe_lock, flags);
+       head = kretprobe_inst_table_head(current);
+
+       /*
+        * 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;
+               recycle_rp_inst(ri, &empty_rp);
+
+               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;
+       }
+
+       kretprobe_assert(ri, orig_ret_address, trampoline_address);
+       regs->rip = orig_ret_address;
+
+       reset_current_kprobe();
+       spin_unlock_irqrestore(&kretprobe_lock, flags);
+       preempt_enable_no_resched();
+
+       hlist_for_each_entry_safe(ri, node, tmp, &empty_rp, hlist) {
+               hlist_del(&ri->hlist);
+               kfree(ri);
+       }
+       /*
+        * By returning a non-zero value, we are telling
+        * kprobe_handler() that we don't want the post_handler
+        * to run (and have re-enabled preemption)
+        */
+       return 1;
+}
+
+/*
+ * 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.
+ */
+static void __kprobes resume_execution(struct kprobe *p,
+               struct pt_regs *regs, struct kprobe_ctlblk *kcb)
+{
+       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 |= kcb->kprobe_old_rflags;
+               break;
+       case 0xc3:              /* ret/lret */
+       case 0xcb:
+       case 0xc2:
+       case 0xca:
+               regs->eflags &= ~TF_MASK;
+               /* rip is already adjusted, no more changes required*/
+               return;
+       case 0xe8:              /* call relative - Fix return addr */
+               *tos = orig_rip + (*tos - copy_rip);
+               break;
+       case 0xff:
+               if ((insn[1] & 0x30) == 0x10) {
+                       /* call absolute, indirect */
+                       /* Fix return addr; rip is correct. */
+                       next_rip = regs->rip;
+                       *tos = orig_rip + (*tos - copy_rip);
+               } else if (((insn[1] & 0x31) == 0x20) ||        /* jmp near, absolute indirect */
+                          ((insn[1] & 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);
+       }
+}
+
+int __kprobes post_kprobe_handler(struct pt_regs *regs)
+{
+       struct kprobe *cur = kprobe_running();
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+
+       if (!cur)
+               return 0;
+
+       if ((kcb->kprobe_status != KPROBE_REENTER) && cur->post_handler) {
+               kcb->kprobe_status = KPROBE_HIT_SSDONE;
+               cur->post_handler(cur, regs, 0);
+       }
+
+       resume_execution(cur, regs, kcb);
+       regs->eflags |= kcb->kprobe_saved_rflags;
+
+       /* Restore the original saved kprobes variables and continue. */
+       if (kcb->kprobe_status == KPROBE_REENTER) {
+               restore_previous_kprobe(kcb);
+               goto out;
+       }
+       reset_current_kprobe();
+out:
+       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;
+}
+
+int __kprobes kprobe_fault_handler(struct pt_regs *regs, int trapnr)
+{
+       struct kprobe *cur = kprobe_running();
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+       const struct exception_table_entry *fixup;
+
+       switch(kcb->kprobe_status) {
+       case KPROBE_HIT_SS:
+       case KPROBE_REENTER:
+               /*
+                * We are here because the instruction being single
+                * stepped caused a page fault. We reset the current
+                * kprobe and the rip points back to the probe address
+                * and allow the page fault handler to continue as a
+                * normal page fault.
+                */
+               regs->rip = (unsigned long)cur->addr;
+               regs->eflags |= kcb->kprobe_old_rflags;
+               if (kcb->kprobe_status == KPROBE_REENTER)
+                       restore_previous_kprobe(kcb);
+               else
+                       reset_current_kprobe();
+               preempt_enable_no_resched();
+               break;
+       case KPROBE_HIT_ACTIVE:
+       case KPROBE_HIT_SSDONE:
+               /*
+                * We increment the nmissed count for accounting,
+                * we can also use npre/npostfault count for accouting
+                * these specific fault cases.
+                */
+               kprobes_inc_nmissed_count(cur);
+
+               /*
+                * We come here because instructions in the pre/post
+                * handler caused the page_fault, this could happen
+                * if handler tries to access user space by
+                * copy_from_user(), get_user() etc. Let the
+                * user-specified handler try to fix it first.
+                */
+               if (cur->fault_handler && cur->fault_handler(cur, regs, trapnr))
+                       return 1;
+
+               /*
+                * In case the user-specified fault handler returned
+                * zero, try to fix up.
+                */
+               fixup = search_exception_tables(regs->rip);
+               if (fixup) {
+                       regs->rip = fixup->fixup;
+                       return 1;
+               }
+
+               /*
+                * fixup() could not handle it,
+                * Let do_page_fault() fix it.
+                */
+               break;
+       default:
+               break;
+       }
+       return 0;
+}
+
+/*
+ * Wrapper routine for handling exceptions.
+ */
+int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
+                                      unsigned long val, void *data)
+{
+       struct die_args *args = (struct die_args *)data;
+       int ret = NOTIFY_DONE;
+
+       if (args->regs && user_mode(args->regs))
+               return ret;
+
+       switch (val) {
+       case DIE_INT3:
+               if (kprobe_handler(args->regs))
+                       ret = NOTIFY_STOP;
+               break;
+       case DIE_DEBUG:
+               if (post_kprobe_handler(args->regs))
+                       ret = NOTIFY_STOP;
+               break;
+       case DIE_GPF:
+       case DIE_PAGE_FAULT:
+               /* kprobe_running() needs smp_processor_id() */
+               preempt_disable();
+               if (kprobe_running() &&
+                   kprobe_fault_handler(args->regs, args->trapnr))
+                       ret = NOTIFY_STOP;
+               preempt_enable();
+               break;
+       default:
+               break;
+       }
+       return ret;
+}
+
+int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs)
+{
+       struct jprobe *jp = container_of(p, struct jprobe, kp);
+       unsigned long addr;
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+
+       kcb->jprobe_saved_regs = *regs;
+       kcb->jprobe_saved_rsp = (long *) regs->rsp;
+       addr = (unsigned long)(kcb->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(kcb->jprobes_stack, (kprobe_opcode_t *)addr,
+                       MIN_STACK_SIZE(addr));
+       regs->eflags &= ~IF_MASK;
+       regs->rip = (unsigned long)(jp->entry);
+       return 1;
+}
+
+void __kprobes jprobe_return(void)
+{
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+
+       asm volatile ("       xchg   %%rbx,%%rsp     \n"
+                     "       int3                      \n"
+                     "       .globl jprobe_return_end  \n"
+                     "       jprobe_return_end:        \n"
+                     "       nop                       \n"::"b"
+                     (kcb->jprobe_saved_rsp):"memory");
+}
+
+int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
+{
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+       u8 *addr = (u8 *) (regs->rip - 1);
+       unsigned long stack_addr = (unsigned long)(kcb->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 != kcb->jprobe_saved_rsp) {
+                       struct pt_regs *saved_regs =
+                           container_of(kcb->jprobe_saved_rsp,
+                                           struct pt_regs, rsp);
+                       printk("current rsp %p does not match saved rsp %p\n",
+                              (long *)regs->rsp, kcb->jprobe_saved_rsp);
+                       printk("Saved registers for jprobe %p\n", jp);
+                       show_registers(saved_regs);
+                       printk("Current registers\n");
+                       show_registers(regs);
+                       BUG();
+               }
+               *regs = kcb->jprobe_saved_regs;
+               memcpy((kprobe_opcode_t *) stack_addr, kcb->jprobes_stack,
+                      MIN_STACK_SIZE(stack_addr));
+               preempt_enable_no_resched();
+               return 1;
+       }
+       return 0;
+}
+
+static struct kprobe trampoline_p = {
+       .addr = (kprobe_opcode_t *) &kretprobe_trampoline,
+       .pre_handler = trampoline_probe_handler
+};
+
+int __init arch_init_kprobes(void)
+{
+       return register_kprobe(&trampoline_p);
+}
+
+int __kprobes arch_trampoline_kprobe(struct kprobe *p)
+{
+       if (p->addr == (kprobe_opcode_t *)&kretprobe_trampoline)
+               return 1;
+
+       return 0;
+}