poly_tomont ppc64le: restore full generated body

The two prior hand-edits of the generated mlkem/ copy dropped one of
the four unrolled iterations. This restores the byte-for-byte simpasm
output; the only change from main is the frame trim (320->224) and the
matching save/restore set. No functional change to the assembly.

Author: Crypteaucajun

mlkem/src/native/ppc64le/src/poly_tomont_ppc_asm.S

@@ -200,6 +200,54 @@ MLK_ASM_FN_SYMBOL(poly_tomont_ppc_asm)
         stxvd2x 50, 9, 3
         stxvd2x 55, 10, 3
         stxvd2x 39, 11, 3
+        lxvd2x 45, 0, 3
+        addi 3, 3, 16
+        lxvd2x 50, 0, 3
+        addi 3, 3, 16
+        lxvd2x 55, 0, 3
+        addi 3, 3, 16
+        lxvd2x 39, 0, 3
+        addi 3, 3, 16
+        vmhraddshs 14, 13, 2, 3
+        vmhraddshs 19, 18, 2, 3
+        vmhraddshs 24, 23, 2, 3
+        vmhraddshs 8, 7, 2, 3
+        vmladduhm 13, 13, 0, 3
+        vmladduhm 18, 18, 0, 3
+        vmladduhm 23, 23, 0, 3
+        vmladduhm 7, 7, 0, 3
+        vmladduhm 27, 14, 5, 13
+        vmladduhm 28, 19, 5, 18
+        vmladduhm 29, 24, 5, 23
+        vmladduhm 30, 8, 5, 7
+        lxvd2x 45, 0, 3
+        addi 3, 3, 16
+        lxvd2x 50, 0, 3
+        addi 3, 3, 16
+        lxvd2x 55, 0, 3
+        addi 3, 3, 16
+        lxvd2x 39, 0, 3
+        addi 3, 3, 16
+        vmhraddshs 14, 13, 2, 3
+        vmhraddshs 19, 18, 2, 3
+        vmhraddshs 24, 23, 2, 3
+        vmhraddshs 8, 7, 2, 3
+        vmladduhm 13, 13, 0, 3
+        vmladduhm 18, 18, 0, 3
+        vmladduhm 23, 23, 0, 3
+        vmladduhm 7, 7, 0, 3
+        vmladduhm 13, 14, 5, 13
+        vmladduhm 18, 19, 5, 18
+        vmladduhm 23, 24, 5, 23
+        vmladduhm 7, 8, 5, 7
+        stxvd2x 59, 4, 3
+        stxvd2x 60, 5, 3
+        stxvd2x 61, 6, 3
+        stxvd2x 62, 7, 3
+        stxvd2x 45, 8, 3
+        stxvd2x 50, 9, 3
+        stxvd2x 55, 10, 3
+        stxvd2x 39, 11, 3
         li 6, 128
         li 7, 144
         li 8, 160