microsoft/compiler: avoid zero-sized texture-lods

We need to ensure we don't end up with a zero-sized texture due to a
non-uniformly sized texture using a high lod. The texture minification
rules clearly say that we should never actually reach zero here.
23 jobs for !295 with d3d12-lower_int_lod_fixes in 12 minutes and 3 seconds (queued for 9 seconds)
detached
Status Job ID Name Coverage
  Container
passed #4349717
aarch64
arm_build

00:00:21

passed #4349719
aarch64
arm_test

00:00:22

passed #4349721
mesa 1809 shell windows
windows_build_vs2019

00:00:09

passed #4349710
x86_build

00:00:14

passed #4349715
x86_build_old

00:00:13

passed #4349712
x86_test-gl

00:00:14

passed #4349713
x86_test-vk

00:00:13

 
  Meson X86 64
passed #4349726
meson-clang

00:02:56

passed #4349725
meson-classic

00:02:33

passed #4349727
meson-clover

00:05:11

passed #4349728
meson-clover-old-llvm

00:03:18

passed #4349724
meson-gallium

00:02:43

passed #4349723
meson-testing

00:06:28

passed #4349729
meson-vulkan

00:01:18

 
  Scons
passed #4349730
scons-win64

00:06:01

 
  Meson Misc
passed #4349732
aarch64
meson-arm64

00:11:35

passed #4349733
aarch64
meson-arm64-build-test

00:02:23

passed #4349731
aarch64
meson-armhf

00:02:25

passed #4349735
meson-i386

00:02:15

passed #4349738
meson-mingw32-x86_64

00:02:02

passed #4349737
kvm
meson-ppc64el

00:01:56

passed #4349736
kvm
meson-s390x

00:02:22

passed #4349734
mesa 1809 docker windows
meson-windows-vs2019

00:04:35