Re: [PATCH 1/2] x86: dumpstack, use pr_cont

[Jiri Slaby <jslaby@suse.cz>]

On 10/06/2015, 06:00 PM, Rasmus Villemoes wrote:
>> --- a/arch/x86/kernel/dumpstack.c
>> +++ b/arch/x86/kernel/dumpstack.c
>> @@ -260,18 +260,18 @@ int __die(const char *str, struct pt_regs *regs, long err)
>>  	printk(KERN_DEFAULT
>>  	       "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter);
>>  #ifdef CONFIG_PREEMPT
>> -	printk("PREEMPT ");
>> +	pr_cont("PREEMPT ");
>>  #endif
>>  #ifdef CONFIG_SMP
>> -	printk("SMP ");
>> +	pr_cont("SMP ");
>>  #endif
>>  #ifdef CONFIG_DEBUG_PAGEALLOC
>> -	printk("DEBUG_PAGEALLOC ");
>> +	pr_cont("DEBUG_PAGEALLOC");
> 
> cosmetic: this lost a space.

Oh, good catch. This was left from the times when DEBUG_PAGEALLOC was
last. Now KASAN is, but I added a space to both of them locally for the
time being. Anyway:

> May I suggest 
> 
> diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c
> index 9c30acfadae2..b473a47d1851 100644
> --- a/arch/x86/kernel/dumpstack.c
> +++ b/arch/x86/kernel/dumpstack.c
> @@ -257,21 +257,23 @@ int __die(const char *str, struct pt_regs *regs, long err)
>  	unsigned short ss;
>  	unsigned long sp;
>  #endif
> -	printk(KERN_DEFAULT
> -	       "%s: %04lx [#%d] ", str, err & 0xffff, ++die_counter);
> +	static const char build_flags[] = ""
>  #ifdef CONFIG_PREEMPT
> -	printk("PREEMPT ");
> +	" PREEMPT"
>  #endif
>  #ifdef CONFIG_SMP
> -	printk("SMP ");
> +	" SMP"
>  #endif
>  #ifdef CONFIG_DEBUG_PAGEALLOC
> -	printk("DEBUG_PAGEALLOC ");
> +	" DEBUG_PAGEALLOC"
>  #endif
>  #ifdef CONFIG_KASAN
> -	printk("KASAN");
> +	" KASAN"
>  #endif
> -	printk("\n");
> +		;
> +	printk(KERN_DEFAULT
> +	       "%s: %04lx [#%d]%s\n", str, err & 0xffff, ++die_counter,
> +		build_flags);
>  	if (notify_die(DIE_OOPS, str, regs, err,
>  			current->thread.trap_nr, SIGSEGV) == NOTIFY_STOP)
>  		return 1;
> 
> instead, so that there's only one printk call and the pr_cont issue goes
> away?

I like this. So if you resend as a proper patch, I will definitely ack
it. But it's up to the maintainers which one they prefer.

thanks,
-- 
js
suse labs