[PATCH V6] panic: Move panic_print before kmsg dumpers

[Guilherme G. Piccoli]

On 21/02/2022 22:45, Sergey Senozhatsky wrote:
> [...]
> Yeah, if Petr is fine with that then I'm OK. But at the same time,
> we have `panic_print' which is a bit mask of what panic_print_sys_info()
> should do. And now we also have a boolean `console_flush` flag that tells
> panic_print_sys_info() to ignore some (one as of now) bits of `panic_print'.
> 
> So _maybe_ panic_print_sys_info() can just accept panic_print as
> its parameter and then we can do something like this (as an example)
> 
> 	panic_print_sys_info(panic_print & ~PANIC_PRINT_ALL_PRINTK_MSG);
> 
> 
>>  	if (panic_print & PANIC_PRINT_ALL_CPU_BT)
>>  		trigger_all_cpu_backtrace();
>> @@ -286,6 +289,8 @@ void panic(const char *fmt, ...)
>>  	 */
>>  	atomic_notifier_call_chain(&panic_notifier_list, 0, buf);
> 
> [..]
> 
>> +	panic_print_sys_info(false);
> 
> Merely because `panic_print_sys_info(false);` doesn't tell much to a reader.
> Like what is print sys info false?
> 
> Or did you already discuss this?

Hi Sergey, thanks for your feedback. So, personally I prefer having the
flag - for me it's clear, it's just a matter of reading the prototype -
either we print the info _or_ we console_flush.

But let's see if others have a preference - if the preference is to use
the bitmask as you suggest, we can do it in a next version.

Cheers,


Guilherme