Fix SIMD data overallocation (dotnet#71043)

The JIT was allocating 16 bytes for TYP_SIMD8 static data loads,
and marking the data section as requiring 16 byte alignment as
well. Fix this to not allocate more than necessary, or over align.

src/coreclr/jit/emit.cpp

@@ -7275,7 +7275,12 @@ UNATIVE_OFFSET emitter::emitDataGenBeg(unsigned size, unsigned alignment, var_ty
     }
 
     assert((secOffs % alignment) == 0);
-    emitConsDsc.alignment = max(emitConsDsc.alignment, alignment);
+    if (emitConsDsc.alignment < alignment)
+    {
+        JITDUMP("Increasing data section alignment from %u to %u for type %s\n", emitConsDsc.alignment, alignment,
+                varTypeName(dataType));
+        emitConsDsc.alignment = alignment;
+    }
 
     /* Advance the current offset */
     emitConsDsc.dsdOffs += size;
@@ -7673,7 +7678,7 @@ void emitter::emitOutputDataSec(dataSecDsc* sec, BYTE* dst)
 
     if (emitComp->opts.disAsm)
     {
-        emitDispDataSec(sec);
+        emitDispDataSec(sec, dst);
     }
 
     unsigned secNum = 0;
@@ -7794,25 +7799,32 @@ void emitter::emitOutputDataSec(dataSecDsc* sec, BYTE* dst)
 //
 // Arguments:
 //    section - the data section description
+//    dst     - address of the data section
 //
 // Notes:
 //    The output format attempts to mirror typical assembler syntax.
 //    Data section entries lack type information so float/double entries
 //    are displayed as if they are integers/longs.
 //
-void emitter::emitDispDataSec(dataSecDsc* section)
+void emitter::emitDispDataSec(dataSecDsc* section, BYTE* dst)
 {
     printf("\n");
 
     unsigned offset = 0;
 
     for (dataSection* data = section->dsdList; data != nullptr; data = data->dsNext)
     {
+        if (emitComp->opts.disAddr)
+        {
+            printf("; @" FMT_ADDR "\n", DBG_ADDR(dst));
+        }
+
         const char* labelFormat = "%-7s";
         char        label[64];
         sprintf_s(label, ArrLen(label), "RWD%02u", offset);
         printf(labelFormat, label);
         offset += data->dsSize;
+        dst += data->dsSize;
 
         if ((data->dsType == dataSection::blockRelative32) || (data->dsType == dataSection::blockAbsoluteAddr))
         {

src/coreclr/jit/emit.h

@@ -2555,7 +2555,7 @@ class emitter
 
     void emitOutputDataSec(dataSecDsc* sec, BYTE* dst);
 #ifdef DEBUG
-    void emitDispDataSec(dataSecDsc* section);
+    void emitDispDataSec(dataSecDsc* section, BYTE* dst);
 #endif
 
     /************************************************************************/

src/coreclr/jit/lower.cpp

@@ -7384,6 +7384,7 @@ void Lowering::LowerSIMD(GenTreeSIMD* simdNode)
 
             if (arg->IsCnsFltOrDbl())
             {
+                noway_assert(constArgCount < ArrLen(constArgValues));
                 constArgValues[constArgCount] = static_cast<float>(arg->AsDblCon()->gtDconVal);
                 constArgCount++;
             }
@@ -7396,10 +7397,14 @@ void Lowering::LowerSIMD(GenTreeSIMD* simdNode)
                 BlockRange().Remove(arg);
             }
 
-            assert(sizeof(constArgValues) == 16);
+            // For SIMD12, even though there might be 12 bytes of constants, we need to store 16 bytes of data
+            // since we've bashed the node the TYP_SIMD16 and do a 16-byte indirection.
+            assert(varTypeIsSIMD(simdNode));
+            const unsigned cnsSize = genTypeSize(simdNode);
+            assert(cnsSize <= sizeof(constArgValues));
 
-            unsigned cnsSize  = sizeof(constArgValues);
-            unsigned cnsAlign = (comp->compCodeOpt() != Compiler::SMALL_CODE) ? cnsSize : 1;
+            const unsigned cnsAlign =
+                (comp->compCodeOpt() != Compiler::SMALL_CODE) ? cnsSize : emitter::dataSection::MIN_DATA_ALIGN;
 
             CORINFO_FIELD_HANDLE hnd =
                 comp->GetEmitter()->emitBlkConst(constArgValues, cnsSize, cnsAlign, simdNode->GetSimdBaseType());