Rotating framebuffers

rotating_framebuffers.md

Goals:

• support vsynced double and triple buffering
• zero buffer copies

User space execution flow:

1. block on free_queue.size() > 0
2. lock free_queue, remove top element, unlock
3. acquire buffer handle
4. map buffer into virtual memory region
5. draw into buffer
6. unmap buffer
7. lock ready_queue, enqueue buffer, unlock
8. goto 1

Kernel execution flow:

1. scanout current foreground buffer
2. at vsync
   1. if ready_queue.size() > 0, lock the queue, get first frame, render it, then lock free_queue and add current frame to it
   2. else go back rendering current frame

Notes on buffer mapping:

• the buffer can be mapped into user space virtual memory with any cache setting
• cache flush happens on unmap
• mapping could use a huge page, so it needs fewer TLB entries
• the buffer doesn't need to be mapped while it's owned by the kernel, since we only DMA from it