Sender: Guenter Roeck <groeck7@gmail.com>
Message-ID: <c5f4fe2d-fc32-cb2f-669c-c9c5f6e3b46e@roeck-us.net>
Date:   Mon, 13 Mar 2023 15:49:26 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.7.1
Content-Language: en-US
To:     Linus Torvalds <torvalds@linux-foundation.org>
Cc:     "Paul E. McKenney" <paulmck@kernel.org>,
        Frederic Weisbecker <frederic@kernel.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
References: <CAHk-=wii6BZtVKYfvQCQqbE3+t1_yAb-ea80-3PcJ4KxgpfHkA@mail.gmail.com>
 <d915df60-d06b-47d4-8b47-8aa1bbc2aac7@roeck-us.net>
 <CAHk-=wi8xQw6eTRncbJfWWYHHi0UBG2B_AfHrzZbPr=k0V_WYQ@mail.gmail.com>
 <fce02d50-65a7-4aa7-8949-6a82321da292@roeck-us.net>
 <CAHk-=wgVCTJTQPwoLrscmOa+n9X8+bc5ieCD+7PXw35Ke7i6ng@mail.gmail.com>
From:   Guenter Roeck <linux@roeck-us.net>
Subject: Re: Linux 6.3-rc2
In-Reply-To: <CAHk-=wgVCTJTQPwoLrscmOa+n9X8+bc5ieCD+7PXw35Ke7i6ng@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 3/13/23 15:16, Linus Torvalds wrote:
> On Mon, Mar 13, 2023 at 1:30 PM Guenter Roeck <linux@roeck-us.net> wrote:
>>
>> It gets weird. Bisect log below. Reverting the identified patch does
>> indeed seem to fix the problem, only I have no clue why this might
>> be the case. The patch looks completely innocent to me. Yet, I can
>> reliably reproduce the problem with v6.3-rc2, but at least so far I
>> have not been able to reproduce it with commit f3dd0c53370 reverted
>> (and I am trying on five different servers in parallel).
> 
> Yeah, that commit looks very innocuous. I'm surprised it would even
> change any code generation, but it's very possible that it ends up
> affecting some code layout or something almost by mistake.
> 
> I'd be inclined to think that the problem is very timing-sensitive,
> and has probably been there for a while, and some random moon phase
> just made it happen now.
> 

Maybe, but it seems unlikely. I can reliably reproduce the problem
with v6.3-rc2. Typically it takes some 20-30 boot attempts to see it. With
f3dd0c53370 reverted I tried more than 1,000 boot attempts and did not see
the problem. I also don't recall ever having seen it before.

Anyway, I'll probably just disable CONFIG_DEBUG_LOCK_ALLOC for my
arm boot tests to avoid the message. I don't want it to create noise
and hide other problems.

Thanks,
Guenter