2022-04-05 02:38:08

by Steven Rostedt

[permalink] [raw]
Subject: Re: Corrupted function tracer in 5.17.0-rc1

On Mon, 4 Apr 2022 15:55:36 +0100
Mark Rutland <[email protected]> wrote:

> From the title, this is v5.17-rc1, which is known broken due to the mcount sort
> issue that was subsequently fixed in commit:

Ah, yes. I was thinking of 5.18-rc1 (which has just been released)

>
> 4ed308c445a1e3ab ("ftrace: Have architectures opt-in for mcount build time sorting")
>
> Is it possible to try with the final v5.17? Or at least try with that patch
> cherry-picked?

Right, 5.17-rc1 is known broken. Please try the latest, 5.17 and let us
know if it is still an issue.

Thanks,

-- Steve


2022-04-06 13:01:44

by [email protected]

[permalink] [raw]
Subject: Re: Corrupted function tracer in 5.17.0-rc1

Hey Steven & Mark,

On 04.04.2022 11:06, Steven Rostedt wrote:
>On Mon, 4 Apr 2022 15:55:36 +0100
>Mark Rutland <[email protected]> wrote:
>
>> From the title, this is v5.17-rc1, which is known broken due to the mcount sort
>> issue that was subsequently fixed in commit:
>
>Ah, yes. I was thinking of 5.18-rc1 (which has just been released)
>
>>
>> 4ed308c445a1e3ab ("ftrace: Have architectures opt-in for mcount build time sorting")
>>
>> Is it possible to try with the final v5.17? Or at least try with that patch
>> cherry-picked?
>
>Right, 5.17-rc1 is known broken. Please try the latest, 5.17 and let us
>know if it is still an issue.

Using 5.18-rc1 it works well, thanks for the quick help.
>
>Thanks,
>
>-- Steve

Greetings,
Sebastian