fetchTransformedBilinearARGB32PM fails to compile in embedded linux

08:28:11 painting/qdrawhelper.cpp: In function 'const uint* fetchTransformedBilinearARGB32PM(uint*, const Operator*, const QSpanData*, int, int, int)':
08:28:11 painting/qdrawhelper.cpp:1857:68: note: use -flax-vector-conversions to permit conversions between vectors with differing element types or numbers of subparts
08:28:11                          int16x8_t v_disty_ = vshlq_n_s16(v_disty, 4);
08:28:11                                                                     ^
08:28:11 painting/qdrawhelper.cpp:1857:68: error: cannot convert 'int32x4_t {aka __vector(4) __builtin_neon_si}' to 'int16x8_t {aka __vector(8) __builtin_neon_hi}' for argument '1' to 'int16x8_t vshlq_n_s16(int16x8_t, int)'
08:28:11 painting/qdrawhelper.cpp:1258:50: error: cannot convert 'int32x4_t {aka __vector(4) __builtin_neon_si}' to 'int16x8_t {aka __vector(8) __builtin_neon_hi}' for argument '2' to 'int16x8_t vmulq_s16(int16x8_t, int16x8_t)'
08:28:11      const int16x8_t dxdy = vmulq_s16(distx, disty); \
08:28:11                                                   ^
08:28:11 painting/qdrawhelper.cpp:1859:25: note: in expansion of macro 'interpolate_4_pixels_16_neon'
08:28:11                          interpolate_4_pixels_16_neon(vreinterpretq_s16_s32(tl.vect), vreinterpretq_s16_s32(tr.vect), vreinterpretq_s16_s32(bl.vect), vreinterpretq_s16_s32(br.vect), vreinterpretq_s16_s32(v_distx), v_disty, v_disty_, colorMask, invColorMask, v_256, b);
08:28:11                          ^
08:28:12 make[3]: *** [.obj/qdrawhelper.o] Error 1
08:28:12 make[3]: Leaving directory `/work/build/qt5-src/qtbase/src/gui'