[MGPipe] cover MGPSubData's four appended extent words in PipeFields.def

40c5443f appended LevelWidth/LevelHeight/LevelDepth/LevelPad to MGPSubData and raised
MGP_ASSERT_POD to 88, but left MGP_FIELDS_MGPSubData at its old eight rows.
gen_pipe.py's check_field_lists_cover_struct_members runs in both modes and in
pipe-gates, so "member(s) with no F(...) in PipeFields.def: LevelWidth, LevelHeight,
LevelDepth, LevelPad" reddened both 'Regenerate the MGPipe interface' and
'--self-test'.

The four rows go in declaration order. LevelWidth/Height/Depth are ordinary members
(the record's full staged mip extent, which the ResourceSubData codec's
"ResourceSubData.LevelExtent" arm bounds the dirty box against). LevelPad is padding
by role only: PADDING_MEMBER_RE is ^Pad\d*$, so a member named LevelPad is a member
to the cover check and MUST have a row. That row risks no false positive - every
producer builds the record as `MGPSubData{}` and the payload rides the wire as one
POD memcpy of all 88 bytes, so both sides of a comparison read the same zero.

Field tables only describe the existing struct, so no generated file changes:
--check stays byte-identical.

MobileGL/MG_Pipe/PipeFields.def

@@ -185,9 +185,21 @@
 #define MGP_FIELDS_MGPSubRegion(F) \
     F(X) F(Y) F(Z) F(W) F(H) F(D) F(SrcOffset) F(SrcRowStride) F(SrcSliceStride)
 
+// 40c5443f's four appended words, in declaration order. LevelWidth/Height/Depth carry the FULL
+// staged mip extent - "mutable GL images may define a lone mip or a noncanonical mip size", so
+// the dirty box below is what CHANGED, not what exists - which is what
+// MGPWireOp::ResourceSubData's "ResourceSubData.LevelExtent" arm holds the box inside
+// (PipeWireCodec.cpp) and what StagedTextureStore derives the server row/slice strides from.
+//
+// LevelPad IS PADDING BY ROLE AND NOT BY NAME, and that is the whole reason it has a row here:
+// the cover check excuses only the members NAMED Pad<n>, so this one left out is a pipe-gates
+// failure. The row costs no false positive, because every producer builds the record as
+// `MGPSubData{}` (so the word is a deterministic zero, not an uninitialised byte) and the
+// payload rides the wire as one POD memcpy - MGP_ASSERT_POD(MGPSubData, 88) is the whole
+// record, with no untransferred tail - so both sides of a comparison read the same zero.
 #define MGP_FIELDS_MGPSubData(F) \
     F(Res) F(Target) F(Level) F(SourceIsVerbatimLevelShadow) F(HasLiveHostWrites) F(UnionBox) \
-    F(RegionCount) F(Blob)
+    F(RegionCount) F(Blob) F(LevelWidth) F(LevelHeight) F(LevelDepth) F(LevelPad)
 
 #define MGP_FIELDS_MGPSubDataComplete(F) \
     F(Res) F(Target) F(FirstLevel) F(LevelCount) F(PullSerial)