[PATCH v3 0/7] arm64: ftrace: fix incorrect output from stack tracer

[takahiro.akashi@linaro.org (AKASHI Takahiro)]

On 10/29/2015 12:23 AM, Will Deacon wrote:
> On Thu, Oct 08, 2015 at 07:01:37PM +0900, AKASHI Takahiro wrote:
>> This is the third patch series for fixing stack tracer on arm64.
>> The original issue was reported by Jungseok[1], and then I found more
>> issues[2].
>> (Steven, Jungseok, sorry for not replying to your comments directly.)
>>
>> I address here all the issues and implement fixes described in [2] except
>> for interrupt-triggered problems, ie. II-3). Recent discussions[3] about
>> introducing a dedicated interrupt stack suggests that we may avoid walking
>> through from an interrupt stack to a process stack.
>> (So interrupt-stack patch is a prerequisite.)
>>
>> Basically,
>> patch1 corresponds to the original issue.
>> patch2 is a proactive improvement of function_graph tracer.
>> patch3 corresponds to II-4(functions under function_graph tracer).
>> patch4 corresponds to II-5(leaf function).
>> patch5, 6 and 7 correspond to II-1(slurping stack) and II-2(differences
>> between x86 and arm64).
>>
>> Each fix can be applied independently, but if patch5, 6 and 7 are
>> acceptable, patch1 is not necessary because patch7 replaces a default
>> stack tracer.
>
> Given the comments and kbuild robot build errors, do you plan to post a
> new version of this series?

Yes, I do.
Do you have any comments that I should address before submitting a new version?
Apart from build errors, I admit that I should drop patch #4 ("arm64: ftrace:
allow for tracing leaf functions") just because I was somewhat confused.
I confirmed that "-pg" option actually disables omit-leaf-stack-frame.

Thanks,
-Takahiro AKASHI

> Will
>