On Tue, 09 Feb 2016 17:07:42 +0100 Denys Vlasenko <[email protected]> wrote:
> On 02/09/2016 04:50 PM, Steven Rostedt wrote: > > One thing is to find the spamming code and fix that. > > We can't rely that there won't be never-ending concurrent printks, > right? For one, in many setups user can cause printk flood. No, that would be a bug. This is why we have printk_ratelimit for. > > I think we must ensure that printk does not livelock. printk is a kernel utility. The users of printk must ensure this. -- Steve
