Re: [nacked] mm-oom-avoid-printk-iteration-under-rcu.patch removed from -mm tree

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2020/04/24 9:56, Yafang Shao wrote:
> I suggest to set KERN_NO_CONSOLES by default but the user can tune it
> back to the original behavior.

Yes, I'd like to set KERN_NO_CONSOLES by default.

> I'm not a fan of sysctl, but if there's no better chioce, enhancing
> vm.oom_dump_tasks seems a possible solution.

I don't think vm.oom_dump_tasks is the right variable to enhance. We can allow users
to suppress not only dump_tasks() output but also e.g. dump_stack() / show_mem()
output, for OOM-killer messages and memory allocation failure messages can take long
time (if printed to consoles) is an unhappy thing for OOM context and atomic context.
Since Dmitry Safonov is working on adding loglevel argument to show_stack(), we will
be able to control dump_stack() in near future. I think that we should introduce a new
sysctl variable under vm directory in order to implement bitmask for "which OOM and
memory allocation failure related messages should be printed to consoles". I think
that there is no need to print to consoles (i.e. saving to log files via userspace
syslog daemon is sufficient) for most systems.





[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux