SUSE Labs Conference 2020 - State of printk rework - lockless ring buffer challenge

Published: 09 October 2020
on channel: SUSE Labs

222

Thanks to #bigbluebutton for hosting the conference!

printk() is the kernel mouth. It mostly works except when it does not or when it takes too long. This has been problem for everyone. There were several attemps to improve the situation but many of them ended as endless chain of hacks.

Two hallf hears ago, one brave and motivated guy, John Ogness, started preparing a massive printk() rework to make it more reliable and predictable. It was part of the final attack to upstream Real Time patchset.

First RFC has been sent 1.5 half year ago. It solved many problems but it was not acceptable as it was. The decision was to split the rework into few stages and start with lockless ringbuffer.

The lockless ringbuffer is finally in the maintainer tree and will likely go into 5.10.

Why does printk() need yet another lockless ring buffer? Why did it take so long to get it done? What are the next steps of the printk() rework? Do I understand or like lockless algorithms now?

Let's see if the talk can answer these and maybe more questions.

Speaker
Petr Mladek

Watch video SUSE Labs Conference 2020 - State of printk rework - lockless ring buffer challenge online, duration hours minute second in high quality that is uploaded to the channel SUSE Labs 09 October 2020. Share the link to the video on social media so that your subscribers and friends will also watch this video. This video clip has been viewed 222 times and liked it 5 visitors.

85,159

4.3K