cairooverlay: Optimize premultiplication/unpremultiplication loops

Pull in video frame fields into local variables. Without this the
compiler must assume that they could've changed on every use and read
them from memory again.

This reduces the inner loop from 6 memory reads per pixels to 4, and the
number of writes stays at 3.
4 jobs for cairooverlay-premultiply in 5 minutes and 31 seconds (queued for 3 seconds)
Status Job ID Name Coverage
  Preparation
passed #59024
gst indent

00:00:14

passed #59023
manifest

00:00:17

 
  Build
passed #59025
build fedora x86_64

00:01:23

 
  Full Builds
passed #59026
build android arm64 api28

00:03:51