radv: Reduce overhead of radv_emit_all_graphics_states

This is based on ideas from Mike's original patches.

Testing

CPU settings used for testing:

# Set all cores to use performance governor
for i in /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor; do echo performance > $i; done
# Disable CPU boost
echo 0 > /sys/devices/system/cpu/cpufreq/boost
# Disable address space randomization
echo 0 > /proc/sys/kernel/randomize_va_space

Test score is the average of 5 runs:

for i in {1..5}; do echo "--- run $i"; ./vkoverhead -test 0 -duration 15; done

commit	6850U score	3900X score	7950X score	i5-2500 score
main 882fd3c5	43084.8	31330.2	51047	20335.68
this MR	47486.8	39617.8	58162.2	23167