]> git.kernelconcepts.de Git - karo-tx-linux.git/commitdiff
ARC: kprobes support
authorVineet Gupta <vgupta@synopsys.com>
Tue, 22 Jan 2013 11:33:59 +0000 (17:03 +0530)
committerVineet Gupta <vgupta@synopsys.com>
Fri, 15 Feb 2013 17:46:05 +0000 (23:16 +0530)
Origin port done by Rajeshwar Ranga

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
Cc: Rajeshwar Ranga <rajeshwar.ranga@gmail.com>
arch/arc/Kconfig
arch/arc/include/asm/kprobes.h [new file with mode: 0644]
arch/arc/include/asm/ptrace.h
arch/arc/kernel/Makefile
arch/arc/kernel/disasm.c
arch/arc/kernel/kprobes.c [new file with mode: 0644]
arch/arc/kernel/traps.c

index 84b23fae661ce0e77b2cd98c1d1f9e9a9d36d73f..cde8d3fcec94562e157b56b77694df2c595adac0 100644 (file)
@@ -24,6 +24,8 @@ config ARC
        select GENERIC_SMP_IDLE_THREAD
        select HAVE_ARCH_TRACEHOOK
        select HAVE_GENERIC_HARDIRQS
+       select HAVE_KPROBES
+       select HAVE_KRETPROBES
        select HAVE_MEMBLOCK
        select HAVE_MOD_ARCH_SPECIFIC if ARC_DW2_UNWIND
        select HAVE_OPROFILE
diff --git a/arch/arc/include/asm/kprobes.h b/arch/arc/include/asm/kprobes.h
new file mode 100644 (file)
index 0000000..4d9c211
--- /dev/null
@@ -0,0 +1,62 @@
+/*
+ * Copyright (C) 2004, 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#ifndef _ARC_KPROBES_H
+#define _ARC_KPROBES_H
+
+#ifdef CONFIG_KPROBES
+
+typedef u16 kprobe_opcode_t;
+
+#define UNIMP_S_INSTRUCTION 0x79e0
+#define TRAP_S_2_INSTRUCTION 0x785e
+
+#define MAX_INSN_SIZE   8
+#define MAX_STACK_SIZE  64
+
+struct arch_specific_insn {
+       int is_short;
+       kprobe_opcode_t *t1_addr, *t2_addr;
+       kprobe_opcode_t t1_opcode, t2_opcode;
+};
+
+#define flush_insn_slot(p)  do {  } while (0)
+
+#define kretprobe_blacklist_size    0
+
+struct kprobe;
+
+void arch_remove_kprobe(struct kprobe *p);
+
+int kprobe_exceptions_notify(struct notifier_block *self,
+                            unsigned long val, void *data);
+
+struct prev_kprobe {
+       struct kprobe *kp;
+       unsigned long status;
+};
+
+struct kprobe_ctlblk {
+       unsigned int kprobe_status;
+       struct pt_regs jprobe_saved_regs;
+       char jprobes_stack[MAX_STACK_SIZE];
+       struct prev_kprobe prev_kprobe;
+};
+
+int kprobe_fault_handler(struct pt_regs *regs, unsigned long cause);
+void kretprobe_trampoline(void);
+void trap_is_kprobe(unsigned long cause, unsigned long address,
+                          struct pt_regs *regs);
+#else
+static void trap_is_kprobe(unsigned long cause, unsigned long address,
+                          struct pt_regs *regs)
+{
+}
+#endif
+
+#endif
index a98957a95fb3d5e5cbb545efa9644c29d5b84ff4..063ed0040ef7b464a3adf7189d2014095b1c23b5 100644 (file)
@@ -111,6 +111,11 @@ struct callee_regs {
        (struct pt_regs *)(pg_start + THREAD_SIZE - 4) - 1;     \
 })
 
+static inline long regs_return_value(struct pt_regs *regs)
+{
+       return regs->r0;
+}
+
 #endif /* !__ASSEMBLY__ */
 
 #define orig_r8_IS_SCALL               0x0001
index 6ae4dd4ecea0b8da235f4c14ea704de66a38b4da..d331bd7047a18ce29c9d8d866abca5b3ebe8d0bc 100644 (file)
@@ -15,6 +15,7 @@ obj-y += devtree.o
 obj-$(CONFIG_MODULES)                  += arcksyms.o module.o
 obj-$(CONFIG_SMP)                      += smp.o
 obj-$(CONFIG_ARC_DW2_UNWIND)           += unwind.o
+obj-$(CONFIG_KPROBES)                  += kprobes.o
 
 obj-$(CONFIG_ARC_FPU_SAVE_RESTORE)     += fpu.o
 CFLAGS_fpu.o   += -mdpfp
index 254d11a4f4958062e21fe42f96c5712fb9b4ebaf..51bad8ff373b4a054c7d846682d2b716b981e565 100644 (file)
@@ -497,9 +497,8 @@ void __kprobes set_reg(int reg, long val, struct pt_regs *regs,
  * @pc +2/4/6 (ARCompact ISA allows free intermixing of 16/32 bit insns).
  *
  * If @pc is a branch
- *     -@tgt_if_br is set to branch target.
- *     -If branch has delay slot, @next_pc updated with actual next PC.
- *
+ *     -@tgt_if_br is set to branch target.
+ *     -If branch has delay slot, @next_pc updated with actual next PC.
  */
 int __kprobes disasm_next_pc(unsigned long pc, struct pt_regs *regs,
                             struct callee_regs *cregs,
diff --git a/arch/arc/kernel/kprobes.c b/arch/arc/kernel/kprobes.c
new file mode 100644 (file)
index 0000000..47e42f2
--- /dev/null
@@ -0,0 +1,525 @@
+/*
+ * Copyright (C) 2004, 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/types.h>
+#include <linux/kprobes.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/kprobes.h>
+#include <linux/kdebug.h>
+#include <linux/sched.h>
+#include <linux/uaccess.h>
+#include <asm/cacheflush.h>
+#include <asm/current.h>
+#include <asm/disasm.h>
+
+#define MIN_STACK_SIZE(addr)   min((unsigned long)MAX_STACK_SIZE, \
+               (unsigned long)current_thread_info() + THREAD_SIZE - (addr))
+
+DEFINE_PER_CPU(struct kprobe *, current_kprobe) = NULL;
+DEFINE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk);
+
+int __kprobes arch_prepare_kprobe(struct kprobe *p)
+{
+       /* Attempt to probe at unaligned address */
+       if ((unsigned long)p->addr & 0x01)
+               return -EINVAL;
+
+       /* Address should not be in exception handling code */
+
+       p->ainsn.is_short = is_short_instr((unsigned long)p->addr);
+       p->opcode = *p->addr;
+
+       return 0;
+}
+
+void __kprobes arch_arm_kprobe(struct kprobe *p)
+{
+       *p->addr = UNIMP_S_INSTRUCTION;
+
+       flush_icache_range((unsigned long)p->addr,
+                          (unsigned long)p->addr + sizeof(kprobe_opcode_t));
+}
+
+void __kprobes arch_disarm_kprobe(struct kprobe *p)
+{
+       *p->addr = p->opcode;
+
+       flush_icache_range((unsigned long)p->addr,
+                          (unsigned long)p->addr + sizeof(kprobe_opcode_t));
+}
+
+void __kprobes arch_remove_kprobe(struct kprobe *p)
+{
+       arch_disarm_kprobe(p);
+
+       /* Can we remove the kprobe in the middle of kprobe handling? */
+       if (p->ainsn.t1_addr) {
+               *(p->ainsn.t1_addr) = p->ainsn.t1_opcode;
+
+               flush_icache_range((unsigned long)p->ainsn.t1_addr,
+                                  (unsigned long)p->ainsn.t1_addr +
+                                  sizeof(kprobe_opcode_t));
+
+               p->ainsn.t1_addr = NULL;
+       }
+
+       if (p->ainsn.t2_addr) {
+               *(p->ainsn.t2_addr) = p->ainsn.t2_opcode;
+
+               flush_icache_range((unsigned long)p->ainsn.t2_addr,
+                                  (unsigned long)p->ainsn.t2_addr +
+                                  sizeof(kprobe_opcode_t));
+
+               p->ainsn.t2_addr = NULL;
+       }
+}
+
+static void __kprobes save_previous_kprobe(struct kprobe_ctlblk *kcb)
+{
+       kcb->prev_kprobe.kp = kprobe_running();
+       kcb->prev_kprobe.status = kcb->kprobe_status;
+}
+
+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;
+}
+
+static inline void __kprobes set_current_kprobe(struct kprobe *p)
+{
+       __get_cpu_var(current_kprobe) = p;
+}
+
+static void __kprobes resume_execution(struct kprobe *p, unsigned long addr,
+                                      struct pt_regs *regs)
+{
+       /* Remove the trap instructions inserted for single step and
+        * restore the original instructions
+        */
+       if (p->ainsn.t1_addr) {
+               *(p->ainsn.t1_addr) = p->ainsn.t1_opcode;
+
+               flush_icache_range((unsigned long)p->ainsn.t1_addr,
+                                  (unsigned long)p->ainsn.t1_addr +
+                                  sizeof(kprobe_opcode_t));
+
+               p->ainsn.t1_addr = NULL;
+       }
+
+       if (p->ainsn.t2_addr) {
+               *(p->ainsn.t2_addr) = p->ainsn.t2_opcode;
+
+               flush_icache_range((unsigned long)p->ainsn.t2_addr,
+                                  (unsigned long)p->ainsn.t2_addr +
+                                  sizeof(kprobe_opcode_t));
+
+               p->ainsn.t2_addr = NULL;
+       }
+
+       return;
+}
+
+static void __kprobes setup_singlestep(struct kprobe *p, struct pt_regs *regs)
+{
+       unsigned long next_pc;
+       unsigned long tgt_if_br = 0;
+       int is_branch;
+       unsigned long bta;
+
+       /* Copy the opcode back to the kprobe location and execute the
+        * instruction. Because of this we will not be able to get into the
+        * same kprobe until this kprobe is done
+        */
+       *(p->addr) = p->opcode;
+
+       flush_icache_range((unsigned long)p->addr,
+                          (unsigned long)p->addr + sizeof(kprobe_opcode_t));
+
+       /* Now we insert the trap at the next location after this instruction to
+        * single step. If it is a branch we insert the trap at possible branch
+        * targets
+        */
+
+       bta = regs->bta;
+
+       if (regs->status32 & 0x40) {
+               /* We are in a delay slot with the branch taken */
+
+               next_pc = bta & ~0x01;
+
+               if (!p->ainsn.is_short) {
+                       if (bta & 0x01)
+                               regs->blink += 2;
+                       else {
+                               /* Branch not taken */
+                               next_pc += 2;
+
+                               /* next pc is taken from bta after executing the
+                                * delay slot instruction
+                                */
+                               regs->bta += 2;
+                       }
+               }
+
+               is_branch = 0;
+       } else
+               is_branch =
+                   disasm_next_pc((unsigned long)p->addr, regs,
+                       (struct callee_regs *) current->thread.callee_reg,
+                       &next_pc, &tgt_if_br);
+
+       p->ainsn.t1_addr = (kprobe_opcode_t *) next_pc;
+       p->ainsn.t1_opcode = *(p->ainsn.t1_addr);
+       *(p->ainsn.t1_addr) = TRAP_S_2_INSTRUCTION;
+
+       flush_icache_range((unsigned long)p->ainsn.t1_addr,
+                          (unsigned long)p->ainsn.t1_addr +
+                          sizeof(kprobe_opcode_t));
+
+       if (is_branch) {
+               p->ainsn.t2_addr = (kprobe_opcode_t *) tgt_if_br;
+               p->ainsn.t2_opcode = *(p->ainsn.t2_addr);
+               *(p->ainsn.t2_addr) = TRAP_S_2_INSTRUCTION;
+
+               flush_icache_range((unsigned long)p->ainsn.t2_addr,
+                                  (unsigned long)p->ainsn.t2_addr +
+                                  sizeof(kprobe_opcode_t));
+       }
+}
+
+int __kprobes arc_kprobe_handler(unsigned long addr, struct pt_regs *regs)
+{
+       struct kprobe *p;
+       struct kprobe_ctlblk *kcb;
+
+       preempt_disable();
+
+       kcb = get_kprobe_ctlblk();
+       p = get_kprobe((unsigned long *)addr);
+
+       if (p) {
+               /*
+                * We have reentered the kprobe_handler, since another kprobe
+                * was hit while within the handler, we save the original
+                * kprobes and single step on the instruction of the new probe
+                * without calling any user handlers to avoid recursive
+                * kprobes.
+                */
+               if (kprobe_running()) {
+                       save_previous_kprobe(kcb);
+                       set_current_kprobe(p);
+                       kprobes_inc_nmissed_count(p);
+                       setup_singlestep(p, regs);
+                       kcb->kprobe_status = KPROBE_REENTER;
+                       return 1;
+               }
+
+               set_current_kprobe(p);
+               kcb->kprobe_status = KPROBE_HIT_ACTIVE;
+
+               /* If we have no pre-handler or it returned 0, we continue with
+                * normal processing. If we have a pre-handler and it returned
+                * non-zero - which is expected from setjmp_pre_handler for
+                * jprobe, we return without single stepping and leave that to
+                * the break-handler which is invoked by a kprobe from
+                * jprobe_return
+                */
+               if (!p->pre_handler || !p->pre_handler(p, regs)) {
+                       setup_singlestep(p, regs);
+                       kcb->kprobe_status = KPROBE_HIT_SS;
+               }
+
+               return 1;
+       } else if (kprobe_running()) {
+               p = __get_cpu_var(current_kprobe);
+               if (p->break_handler && p->break_handler(p, regs)) {
+                       setup_singlestep(p, regs);
+                       kcb->kprobe_status = KPROBE_HIT_SS;
+                       return 1;
+               }
+       }
+
+       /* no_kprobe: */
+       preempt_enable_no_resched();
+       return 0;
+}
+
+static int __kprobes arc_post_kprobe_handler(unsigned long addr,
+                                        struct pt_regs *regs)
+{
+       struct kprobe *cur = kprobe_running();
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+
+       if (!cur)
+               return 0;
+
+       resume_execution(cur, addr, regs);
+
+       /* Rearm the kprobe */
+       arch_arm_kprobe(cur);
+
+       /*
+        * When we return from trap instruction we go to the next instruction
+        * We restored the actual instruction in resume_exectuiont and we to
+        * return to the same address and execute it
+        */
+       regs->ret = addr;
+
+       if ((kcb->kprobe_status != KPROBE_REENTER) && cur->post_handler) {
+               kcb->kprobe_status = KPROBE_HIT_SSDONE;
+               cur->post_handler(cur, regs, 0);
+       }
+
+       if (kcb->kprobe_status == KPROBE_REENTER) {
+               restore_previous_kprobe(kcb);
+               goto out;
+       }
+
+       reset_current_kprobe();
+
+out:
+       preempt_enable_no_resched();
+       return 1;
+}
+
+/*
+ * Fault can be for the instruction being single stepped or for the
+ * pre/post handlers in the module.
+ * This is applicable for applications like user probes, where we have the
+ * probe in user space and the handlers in the kernel
+ */
+
+int __kprobes kprobe_fault_handler(struct pt_regs *regs, unsigned long trapnr)
+{
+       struct kprobe *cur = kprobe_running();
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+
+       switch (kcb->kprobe_status) {
+       case KPROBE_HIT_SS:
+       case KPROBE_REENTER:
+               /*
+                * We are here because the instruction being single stepped
+                * caused the fault. We reset the current kprobe and allow the
+                * exception handler as if it is regular exception. In our
+                * case it doesn't matter because the system will be halted
+                */
+               resume_execution(cur, (unsigned long)cur->addr, regs);
+
+               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 are here because the instructions in the pre/post handler
+                * caused the fault.
+                */
+
+               /* 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.
+                */
+               if (fixup_exception(regs))
+                       return 1;
+
+               /*
+                * fixup_exception() could not handle it,
+                * Let do_page_fault() fix it.
+                */
+               break;
+
+       default:
+               break;
+       }
+       return 0;
+}
+
+int __kprobes kprobe_exceptions_notify(struct notifier_block *self,
+                                      unsigned long val, void *data)
+{
+       struct die_args *args = data;
+       unsigned long addr = args->err;
+       int ret = NOTIFY_DONE;
+
+       switch (val) {
+       case DIE_IERR:
+               if (arc_kprobe_handler(addr, args->regs))
+                       return NOTIFY_STOP;
+               break;
+
+       case DIE_TRAP:
+               if (arc_post_kprobe_handler(addr, args->regs))
+                       return NOTIFY_STOP;
+               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);
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+       unsigned long sp_addr = regs->sp;
+
+       kcb->jprobe_saved_regs = *regs;
+       memcpy(kcb->jprobes_stack, (void *)sp_addr, MIN_STACK_SIZE(sp_addr));
+       regs->ret = (unsigned long)(jp->entry);
+
+       return 1;
+}
+
+void __kprobes jprobe_return(void)
+{
+       __asm__ __volatile__("unimp_s");
+       return;
+}
+
+int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
+{
+       struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+       unsigned long sp_addr;
+
+       *regs = kcb->jprobe_saved_regs;
+       sp_addr = regs->sp;
+       memcpy((void *)sp_addr, kcb->jprobes_stack, MIN_STACK_SIZE(sp_addr));
+       preempt_enable_no_resched();
+
+       return 1;
+}
+
+static void __used kretprobe_trampoline_holder(void)
+{
+       __asm__ __volatile__(".global kretprobe_trampoline\n"
+                            "kretprobe_trampoline:\n" "nop\n");
+}
+
+void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri,
+                                     struct pt_regs *regs)
+{
+
+       ri->ret_addr = (kprobe_opcode_t *) regs->blink;
+
+       /* Replace the return addr with trampoline addr */
+       regs->blink = (unsigned long)&kretprobe_trampoline;
+}
+
+static 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);
+       kretprobe_hash_lock(current, &head, &flags);
+
+       /*
+        * 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 than 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->ret = orig_ret_address;
+
+       reset_current_kprobe();
+       kretprobe_hash_unlock(current, &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 the kprobe handler
+        * that we don't want the post_handler to run
+        */
+       return 1;
+}
+
+static struct kprobe trampoline_p = {
+       .addr = (kprobe_opcode_t *) &kretprobe_trampoline,
+       .pre_handler = trampoline_probe_handler
+};
+
+int __init arch_init_kprobes(void)
+{
+       /* Registering the trampoline code for the kret probe */
+       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;
+}
+
+void trap_is_kprobe(unsigned long cause, unsigned long address,
+                   struct pt_regs *regs)
+{
+       notify_die(DIE_TRAP, "kprobe_trap", regs, address, cause, SIGTRAP);
+}
index fd2457cec226b9e2e044a20885391eb67e455e7e..c6396b48fcd2d8ea124ac7ae1d252a2896e0b2f9 100644 (file)
@@ -15,6 +15,7 @@
 #include <linux/uaccess.h>
 #include <asm/ptrace.h>
 #include <asm/setup.h>
+#include <asm/kprobes.h>
 
 void __init trap_init(void)
 {
@@ -90,6 +91,7 @@ void do_machine_check_fault(unsigned long cause, unsigned long address,
        die("Machine Check Exception", regs, address, cause);
 }
 
+
 /*
  * Entry point for traps induced by ARCompact TRAP_S <n> insn
  * This is same family as TRAP0/SWI insn (use the same vector).
@@ -109,6 +111,10 @@ void do_non_swi_trap(unsigned long cause, unsigned long address,
                trap_is_brkpt(cause, address, regs);
                break;
 
+       case 2:
+               trap_is_kprobe(param, address, regs);
+               break;
+
        default:
                break;
        }
@@ -116,10 +122,17 @@ void do_non_swi_trap(unsigned long cause, unsigned long address,
 
 /*
  * Entry point for Instruction Error Exception
+ *  -For a corner case, ARC kprobes implementation resorts to using
+ *   this exception, hence the check
  */
 void do_insterror_or_kprobe(unsigned long cause,
                                       unsigned long address,
                                       struct pt_regs *regs)
 {
+       /* Check if this exception is caused by kprobes */
+       if (notify_die(DIE_IERR, "kprobe_ierr", regs, address,
+                      cause, SIGILL) == NOTIFY_STOP)
+               return;
+
        insterror_is_error(cause, address, regs);
 }