[PATCH] panic: drop the fake trace id from the oops end marker

[PATCH] panic: drop the fake trace id from the oops end marker

[Bradley Morgan]

print_oops_end_marker() prints "---[ end trace 0000000000000000
]---". The format pretends each oops has a trace id, but the value
has been a hardcoded 0 since the oops id machinery was removed, and
nothing prints an opening marker anymore. Nothing in tree parses the
id either, the pstore selftest greps for "---[ end trace" without it.
Kill the fake id, the format specifier and the 0ULL argument go with
it.

Signed-off-by: Bradley Morgan <brads@mainlining.org>
---
 kernel/panic.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/panic.c b/kernel/panic.c
index 213725b612aa..0436a92318ed 100644
--- a/kernel/panic.c
+++ b/kernel/panic.c
@@ -1051,7 +1051,7 @@ void oops_enter(void)
 
 static void print_oops_end_marker(void)
 {
-	pr_warn("---[ end trace %016llx ]---\n", 0ULL);
+	pr_warn("---[ end trace ]---\n");
 }
 
 /*
-- 
2.53.0

Re: [PATCH] panic: drop the fake trace id from the oops end marker

[Sebastian Andrzej Siewior]

On 2026-10-04 17:54:19 [+0000], Bradley Morgan wrote:
> print_oops_end_marker() prints "---[ end trace 0000000000000000
> ]---". The format pretends each oops has a trace id, but the value
> has been a hardcoded 0 since the oops id machinery was removed, and
> nothing prints an opening marker anymore. Nothing in tree parses the
> id either, the pstore selftest greps for "---[ end trace" without it.
> Kill the fake id, the format specifier and the 0ULL argument go with
> it.

Could you please figure out _why_ it is hardcoded to 0? Someone
requested this with a reason. If you want to remove it, you need to
ensure that reason is no longer valid.

> Signed-off-by: Bradley Morgan <brads@mainlining.org>
> ---
>  kernel/panic.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/panic.c b/kernel/panic.c
> index 213725b612aa..0436a92318ed 100644
> --- a/kernel/panic.c
> +++ b/kernel/panic.c
> @@ -1051,7 +1051,7 @@ void oops_enter(void)
>  
>  static void print_oops_end_marker(void)
>  {
> -	pr_warn("---[ end trace %016llx ]---\n", 0ULL);
> +	pr_warn("---[ end trace ]---\n");
>  }
>  
>  /*

Sebastian

Re: [PATCH] panic: drop the fake trace id from the oops end marker

[Bradley Morgan]

On 19 October 2026 15:46:28 BST, Sebastian Andrzej Siewior
<bigeasy@linutronix.de> wrote:
>On 2026-10-04 17:54:19 [+0000], Bradley Morgan wrote:
>> print_oops_end_marker() prints "---[ end trace 0000000000000000
>> ]---". The format pretends each oops has a trace id, but the value
>> has been a hardcoded 0 since the oops id machinery was removed, and
>> nothing prints an opening marker anymore. Nothing in tree parses the
>> id either, the pstore selftest greps for "---[ end trace" without it.
>> Kill the fake id, the format specifier and the 0ULL argument go with
>> it.
>
>Could you please figure out _why_ it is hardcoded to 0? Someone
>requested this with a reason. If you want to remove it, you need to
>ensure that reason is no longer valid.

Nice catch!

uhhh, well it's just always been there I'd assume. Probably for selftests? 

But what's exactly the "benefit" of this? Idk.

>
>> Signed-off-by: Bradley Morgan <brads@mainlining.org>
>> ---
>>  kernel/panic.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>> 
>> diff --git a/kernel/panic.c b/kernel/panic.c
>> index 213725b612aa..0436a92318ed 100644
>> --- a/kernel/panic.c
>> +++ b/kernel/panic.c
>> @@ -1051,7 +1051,7 @@ void oops_enter(void)
>>  
>>  static void print_oops_end_marker(void)
>>  {
>> -	pr_warn("---[ end trace %016llx ]---\n", 0ULL);
>> +	pr_warn("---[ end trace ]---\n");
>>  }
>>  
>>  /*
>
>Sebastian

--- Thanks!
"I'm not a very positive person" - Linus torvalds

Re: [PATCH] panic: drop the fake trace id from the oops end marker

[Sebastian Andrzej Siewior]

On 2026-10-19 16:07:12 [+0100], Bradley Morgan wrote:
> On 19 October 2026 15:46:28 BST, Sebastian Andrzej Siewior
> <bigeasy@linutronix.de> wrote:
> >On 2026-10-04 17:54:19 [+0000], Bradley Morgan wrote:
> >> print_oops_end_marker() prints "---[ end trace 0000000000000000
> >> ]---". The format pretends each oops has a trace id, but the value
> >> has been a hardcoded 0 since the oops id machinery was removed, and
> >> nothing prints an opening marker anymore. Nothing in tree parses the
> >> id either, the pstore selftest greps for "---[ end trace" without it.
> >> Kill the fake id, the format specifier and the 0ULL argument go with
> >> it.
> >
> >Could you please figure out _why_ it is hardcoded to 0? Someone
> >requested this with a reason. If you want to remove it, you need to
> >ensure that reason is no longer valid.
> 
> Nice catch!
> 
> uhhh, well it's just always been there I'd assume. Probably for selftests? 
> 
> But what's exactly the "benefit" of this? Idk.

Once you figured out why something is the way it is you could make
change it argue that the old reason is longer valid because of …

You did not do it. You simply said remove the 0 because we have in-tree
users. At best.

Sebastian