d3d12: force 4-byte aligntment for vertex-buffer stride

In D3D12, we require 4-byte alignment when using half-float vertext
formats. It's not clear to me why, but the validator complains.

Sadly, we don't have a smaller hammer, so let's just always force 4-byte
alignment istead. It's not ideal, but it works. It makes us fallback to
u_vbuf.c all the time, which has some performance overhead.
23 jobs for !264 with d3d12-force_4byte_aligned_vb_stride in 10 minutes and 4 seconds (queued for 5 seconds)
latest detached
Status Job ID Name Coverage
  Container
passed #3933865
aarch64
arm_build

00:00:19

passed #3933866
aarch64
arm_test

00:00:20

passed #3933867
mesa 1809 shell windows
windows_build_vs2019

00:00:07

passed #3933861
x86_build

00:00:25

passed #3933864
x86_build_old

00:00:20

passed #3933862
x86_test-gl

00:00:21

passed #3933863
x86_test-vk

00:00:14

 
  Meson X86 64
passed #3933871
meson-clang

00:02:17

passed #3933870
meson-classic

00:01:45

passed #3933872
meson-clover

00:08:06

passed #3933873
meson-clover-old-llvm

00:05:11

passed #3933869
meson-gallium

00:07:14

passed #3933868
meson-testing

00:06:34

passed #3933874
meson-vulkan

00:03:05

 
  Scons
passed #3933875
scons-win64

00:03:42

 
  Meson Misc
passed #3933877
aarch64
meson-arm64

00:01:46

passed #3933878
aarch64
meson-arm64-build-test

00:01:47

passed #3933876
aarch64
meson-armhf

00:02:10

passed #3933880
meson-i386

00:02:07

passed #3933883
meson-mingw32-x86_64

00:01:34

passed #3933882
kvm
meson-ppc64el

00:02:21

passed #3933881
kvm
meson-s390x

00:02:51

passed #3933879
mesa 1809 docker windows
meson-windows-vs2019

00:09:48