RISC-V: Disable BSWAP optimization for NUNITS < 4

RISC-V: Disable BSWAP optimization for NUNITS < 4

When fixing bugs, I notice there is a piece odd codes look incorrect.
which probably make codegen worse.

#include <stdint.h>

typedef int8_t vnx2qi __attribute__ ((vector_size (2)));

#define MASK_2(X, Y) (Y) - 1 - (X), (Y) - 2 - (X)

#define PERMUTE(TYPE, NUNITS)                                                  \
  __attribute__ ((noipa)) void permute_##TYPE (TYPE values1, TYPE values2,     \
					       TYPE *out)                      \
  {                                                                            \
    TYPE v                                                                     \
      = __builtin_shufflevector (values1, values2, MASK_##NUNITS (0, NUNITS)); \
    *(TYPE *) out = v;                                                         \
  }

#define TEST_ALL(T)                                                            \
  T (vnx2qi, 2)

TEST_ALL (PERMUTE)

Before this patch:

        vsetivli        zero,2,e8,mf8,ta,ma
        vle8.v  v1,0(a0)
        vsetivli        zero,1,e16,mf4,ta,ma
        vsrl.vi v2,v1,8
        vsll.vi v1,v1,8
        vor.vv  v1,v2,v1
        vsetivli        zero,2,e8,mf8,ta,ma
        vse8.v  v1,0(a2)
        ret

After this patch:

        vsetivli        zero,2,e8,mf8,ta,ma
        vle8.v  v3,0(a0)
        vid.v   v1
        vrsub.vi        v1,v1,1
        vrgather.vv     v2,v3,v1
        vse8.v  v2,0(a2)
        ret

Committed as it is very obvious if during code review.

gcc/ChangeLog:

	* config/riscv/riscv-v.cc (shuffle_bswap_pattern): Disable for NUNIT < 4.

gcc/testsuite/ChangeLog:

	* gcc.target/riscv/rvv/autovec/vls-vlmax/perm-4.c: Adapt test.
	* gcc.target/riscv/rvv/autovec/vls/perm-4.c: Ditto.