On 2/8/2021 10:53 AM, Borislav Petkov wrote:

On Mon, Feb 08, 2021 at 10:50:07AM -0800, Yu, Yu-cheng wrote:

quoted

I have not run into the situation.  Initially it was there because other
faults have it.

Which other faults?

exc_general_protection() and do_trap() both call show_signal(), which 
then calls printk_ratelimit().

quoted

When you asked, I went through it and put out my reasoning.

What does that mean?

I went through my patch and check if ratelimit is necessary, and then 
describe the finding.

quoted

I think it still makes sense to keep it.

Because you have a hunch or you actually have an objective reason why?

For example, if a shell script, in a loop re-starts an app when it 
exits, and the app is causing control-protection fault.  The log 
messages should be rate limited.