Re: [PATCH 1/2] memcg: use seq_buf_do_printk() with mem_cgroup_print_oom_meminfo()

[Sergey Senozhatsky]

On (23/04/26 17:24), Michal Hocko wrote:
> No objection from me but is it possible that more printk calls (one per
> line with this change correct?) would add a contention on the printk
> path?

It probably will have opposite effect: console->write of longer lines
keep local IRQs disabled longer and keep console_waiter printk spinning
(in console_trylock_spinning()) longer.