[PATCH] powerpc/crash: Fix non-smp kexec preparation

[PATCH] powerpc/crash: Fix non-smp kexec preparation

[Eddie James]

In non-smp configurations, crash_kexec_prepare is never called in
the crash shutdown path. One result of this is that the crashing_cpu
variable is never set, preventing crash_save_cpu from storing the
NT_PRSTATUS elf note in the core dump.

Fixes: c7255058b543 ("powerpc/crash: save cpu register data in crash_smp_send_stop()")
Signed-off-by: Eddie James <eajames@linux.ibm.com>
---
 arch/powerpc/kexec/crash.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/powerpc/kexec/crash.c b/arch/powerpc/kexec/crash.c
index 9ac3266e49652..8c11cfd19e734 100644
--- a/arch/powerpc/kexec/crash.c
+++ b/arch/powerpc/kexec/crash.c
@@ -359,7 +359,11 @@ void default_machine_crash_shutdown(struct pt_regs *regs)
 	if (TRAP(regs) == INTERRUPT_SYSTEM_RESET)
 		is_via_system_reset = 1;
 
+#ifdef CONFIG_SMP
 	crash_smp_send_stop();
+#else
+	crash_kexec_prepare();
+#endif
 
 	crash_save_cpu(regs, crashing_cpu);
 
-- 
2.43.5

Re: [PATCH] powerpc/crash: Fix non-smp kexec preparation

[Christophe Leroy]

Le 10/02/2025 à 20:52, Eddie James a écrit :
> [Vous ne recevez pas souvent de courriers de eajames@linux.ibm.com. Découvrez pourquoi ceci est important à https://aka.ms/LearnAboutSenderIdentification ]
> 
> In non-smp configurations, crash_kexec_prepare is never called in
> the crash shutdown path. One result of this is that the crashing_cpu
> variable is never set, preventing crash_save_cpu from storing the
> NT_PRSTATUS elf note in the core dump.
> 
> Fixes: c7255058b543 ("powerpc/crash: save cpu register data in crash_smp_send_stop()")
> Signed-off-by: Eddie James <eajames@linux.ibm.com>
> ---
>   arch/powerpc/kexec/crash.c | 4 ++++
>   1 file changed, 4 insertions(+)
> 
> diff --git a/arch/powerpc/kexec/crash.c b/arch/powerpc/kexec/crash.c
> index 9ac3266e49652..8c11cfd19e734 100644
> --- a/arch/powerpc/kexec/crash.c
> +++ b/arch/powerpc/kexec/crash.c
> @@ -359,7 +359,11 @@ void default_machine_crash_shutdown(struct pt_regs *regs)
>          if (TRAP(regs) == INTERRUPT_SYSTEM_RESET)
>                  is_via_system_reset = 1;
> 
> +#ifdef CONFIG_SMP

I see no obvious reason to use #ifdef instead of IS_ENABLED(CONFIG_SMP), 
is there any ?

Ref 
https://docs.kernel.org/process/coding-style.html#conditional-compilation

>          crash_smp_send_stop();
> +#else
> +       crash_kexec_prepare();
> +#endif
> 
>          crash_save_cpu(regs, crashing_cpu);
> 
> --
> 2.43.5
> 

Re: [PATCH] powerpc/crash: Fix non-smp kexec preparation

[Eddie James]

On 2/11/25 01:38, Christophe Leroy wrote:
>
>
> Le 10/02/2025 à 20:52, Eddie James a écrit :
>> [Vous ne recevez pas souvent de courriers de eajames@linux.ibm.com. 
>> Découvrez pourquoi ceci est important à 
>> https://aka.ms/LearnAboutSenderIdentification ]
>>
>> In non-smp configurations, crash_kexec_prepare is never called in
>> the crash shutdown path. One result of this is that the crashing_cpu
>> variable is never set, preventing crash_save_cpu from storing the
>> NT_PRSTATUS elf note in the core dump.
>>
>> Fixes: c7255058b543 ("powerpc/crash: save cpu register data in 
>> crash_smp_send_stop()")
>> Signed-off-by: Eddie James <eajames@linux.ibm.com>
>> ---
>>   arch/powerpc/kexec/crash.c | 4 ++++
>>   1 file changed, 4 insertions(+)
>>
>> diff --git a/arch/powerpc/kexec/crash.c b/arch/powerpc/kexec/crash.c
>> index 9ac3266e49652..8c11cfd19e734 100644
>> --- a/arch/powerpc/kexec/crash.c
>> +++ b/arch/powerpc/kexec/crash.c
>> @@ -359,7 +359,11 @@ void default_machine_crash_shutdown(struct 
>> pt_regs *regs)
>>          if (TRAP(regs) == INTERRUPT_SYSTEM_RESET)
>>                  is_via_system_reset = 1;
>>
>> +#ifdef CONFIG_SMP
>
> I see no obvious reason to use #ifdef instead of 
> IS_ENABLED(CONFIG_SMP), is there any ?
>
> Ref 
> https://docs.kernel.org/process/coding-style.html#conditional-compilation


Oops, no, just my habit. I'll change it.


Thanks,

Eddie


>
>>          crash_smp_send_stop();
>> +#else
>> +       crash_kexec_prepare();
>> +#endif
>>
>>          crash_save_cpu(regs, crashing_cpu);
>>
>> -- 
>> 2.43.5
>>
>