i965,anv,iris: Only set VectorMaskEnable when needed

For cases with lots of very small primitives, this may improve
performance because we're not executing those dead channels all the
time.

Shader-db reports no instruction or cycle-count changes.  However, by
hacking up the driver to report when this optimization triggers, it
appears to affect about 10% of shader-db.
13 jobs for !1054 with review/intel-conservative-helper-invoc in 11 minutes and 51 seconds (queued for 3 seconds)
latest detached
Status Job ID Name Coverage
  Containers Build
passed #442059
debian

00:00:28

 
  Build+Test
passed #442068
meson-arm64

00:02:28

passed #442067
meson-armhf

00:02:37

passed #442061
meson-clang

00:06:05

passed #442065
meson-clover

00:08:46

passed #442069
meson-i386

00:02:16

passed #442064
meson-main

00:04:21

passed #442060
meson-swr-glvnd

00:02:21

passed #442066
meson-vulkan

00:02:54

passed #442071
scons-llvm

00:03:18

passed #442070
scons-nollvm

00:03:03

passed #442062
scons-swr

00:05:44

passed #442063
scons-win64

00:05:55