Slides

Slides

The slides may be found here.

I am glad that neither Jon nor Jake injured either themselves or their keyboards during those two leisurely strolls through the slides. ;–)

There were several hours worth of discussion in the cited email thread, so the purpose of the slides was for me to quickly figure out what people wanted to focus on.

Those wanting to know more about the possibility of hazard pointers in the Linux kernel might find something of interest here. Boqun Feng is working on a prototype, in which a number of people expressed interest during subsequent hallway-track sessions. TL;DR: Situations in which performance and/or scalability is being limited by a not-zero atomic increment of a reference count might be well served by hazard pointers. (In addition to the memory-footprint motivation called out in the article.)