[SRU][Bionic][Cosmic][PATCH 1/1] powerpc/livepatch: Implement reliable stack tracing for the consistency model

[Michael Ellerman]

Joseph Salisbury <joseph.salisbury at canonical.com> writes:
> On 06/07/2018 01:37 PM, Khaled Elmously wrote:
>> On 2018-06-05 11:52:15 , Joseph Salisbury wrote:
>>> On 06/04/2018 10:11 PM, Michael Ellerman wrote:
>>>> Joseph Salisbury <joseph.salisbury at canonical.com> writes:
>>>>> diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
>>>>> index d534ed901538..26a50603177c 100644
>>>>> --- a/arch/powerpc/kernel/stacktrace.c
>>>>> +++ b/arch/powerpc/kernel/stacktrace.c
>>>>> @@ -76,3 +81,115 @@ save_stack_trace_regs(struct pt_regs *regs, struct stack_trace *trace)
>>>>>  	save_context_stack(trace, regs->gpr[1], current, 0);
>>>>>  }
>>>>>  EXPORT_SYMBOL_GPL(save_stack_trace_regs);
>>>>> +
>>>>> +#ifdef CONFIG_HAVE_RELIABLE_STACKTRACE
>>>>> +int
>>>>> +save_stack_trace_tsk_reliable(struct task_struct *tsk,
>>>>> +				struct stack_trace *trace)
>>>>> +{
>>>> ...
>>>>> +		/*
>>>>> +		 * Mark stacktraces with kretprobed functions on them
>>>>> +		 * as unreliable.
>>>>> +		 */
>>>>> +		if (ip == (unsigned long)kretprobe_trampoline)
>>>>> +			return 1;
>>>> You may also want:
>>>>
>>>> 5e3f0d15ae5f ("powerpc/livepatch: Fix build error with kprobes disabled.")
>>>>
>>> Thanks for the feedback, Michael.  Would it be possible for you to post
>>> this comment to the bug?  That way we can get feedback from the original
>>> bug reporter and IBM.
>>>
>> Joe, I did not apply this patch as it is not clear to me if a V2 is on the way or if it needs more work. Please advise.
>
> I don't plan on sending a V2.  IBM tested what is in this SRU request
> and it was confirmed to fix the bug for them.  I would say apply what is
> in this request.
>
> Michael, it's probably best to open a new bug if you believe additional
> commits are needed.

OK, it just fixes a build break under some configs, so if you haven't
hit it then you probably don't need the fix.

cheers