diff options
| author | Andi Kleen <ak@suse.de> | 2005-04-16 15:25:13 -0700 |
|---|---|---|
| committer | Linus Torvalds <torvalds@ppc970.osdl.org> | 2005-04-16 15:25:13 -0700 |
| commit | daeeafecf029a05a304fa299706b3ecc90e16175 (patch) | |
| tree | 7b17fd7408b47892e391604c6791674f03109e1c /arch/x86_64 | |
| parent | ece90303ec7e2e8d8610eb03ee7648755dd8efe7 (diff) | |
[PATCH] x86_64: Keep only a single debug notifier chain
Calling a notifier three times in the debug handler does not make much sense,
because a debugger can figure out the various conditions by itself. Remove
the additional calls to DIE_DEBUG and DIE_DEBUGSTEP completely.
This matches what i386 does now.
This also makes sure interrupts are always still disabled when calling a
debugger, which prevents:
BUG: using smp_processor_id() in preemptible [00000001] code: tpopf/1470
caller is post_kprobe_handler+0x9/0x70
Call Trace:<ffffffff8024f10f>{smp_processor_id+191} <ffffffff80120e69>{post_kpro
be_handler+9}
<ffffffff80120f7a>{kprobe_exceptions_notify+58}
<ffffffff80144fc0>{notifier_call_chain+32} <ffffffff80110daf>{do_debug+335}
<ffffffff8010f513>{debug+127} <EOE>
on preemptible debug kernels with kprobes when single stepping in user space.
This was probably a bug even on non preempt kernels, this function was
supposed to be running with interrupts off according to a comment there.
Note to third part debugger maintainers: please double check your debugger can
still single step.
Cc: <prasanna@in.ibm.com>
Cc: <jbeulich@novell.com>
Cc: <kaos@sgi.com>
Cc: <jim.houston@ccur.com>
Cc: <jfv@bluesong.net>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'arch/x86_64')
| -rw-r--r-- | arch/x86_64/kernel/traps.c | 14 |
1 files changed, 3 insertions, 11 deletions
diff --git a/arch/x86_64/kernel/traps.c b/arch/x86_64/kernel/traps.c index 2f42fd5c075f..f6ccf155e3e5 100644 --- a/arch/x86_64/kernel/traps.c +++ b/arch/x86_64/kernel/traps.c @@ -659,9 +659,9 @@ asmlinkage void do_debug(struct pt_regs * regs, unsigned long error_code) asm("movq %%db6,%0" : "=r" (condition)); if (notify_die(DIE_DEBUG, "debug", regs, condition, error_code, - SIGTRAP) == NOTIFY_STOP) { + SIGTRAP) == NOTIFY_STOP) return; - } + conditional_sti(regs); /* Mask out spurious debug traps due to lazy DR7 setting */ @@ -674,9 +674,7 @@ asmlinkage void do_debug(struct pt_regs * regs, unsigned long error_code) tsk->thread.debugreg6 = condition; /* Mask out spurious TF errors due to lazy TF clearing */ - if ((condition & DR_STEP) && - (notify_die(DIE_DEBUGSTEP, "debugstep", regs, condition, - 1, SIGTRAP) != NOTIFY_STOP)) { + if (condition & DR_STEP) { /* * The TF error should be masked out only if the current * process is not traced and if the TRAP flag has been set @@ -711,16 +709,10 @@ asmlinkage void do_debug(struct pt_regs * regs, unsigned long error_code) force_sig_info(SIGTRAP, &info, tsk); clear_dr7: asm volatile("movq %0,%%db7"::"r"(0UL)); - notify_die(DIE_DEBUG, "debug", regs, condition, 1, SIGTRAP); return; clear_TF_reenable: set_tsk_thread_flag(tsk, TIF_SINGLESTEP); - -clear_TF: - /* RED-PEN could cause spurious errors */ - if (notify_die(DIE_DEBUG, "debug2", regs, condition, 1, SIGTRAP) - != NOTIFY_STOP) regs->eflags &= ~TF_MASK; } |