[AArch64] Add an endian_lane_rtx helper routine

Later patches turn the number of vector units into a poly_int.
We deliberately don't support applying GEN_INT to those (except
in target code that doesn't distinguish between poly_ints and normal
constants); gen_int_mode needs to be used instead.

This patch therefore replaces instances of:

  GEN_INT (ENDIAN_LANE_N (builtin_mode, INTVAL (op[opc])))

with uses of a new endian_lane_rtx function.

2017-11-06  Richard Sandiford  <richard.sandiford@linaro.org>
	    Alan Hayward  <alan.hayward@arm.com>
	    David Sherwood  <david.sherwood@arm.com>

gcc/
	* config/aarch64/aarch64-protos.h (aarch64_endian_lane_rtx): Declare.
	* config/aarch64/aarch64.c (aarch64_endian_lane_rtx): New function.
	* config/aarch64/aarch64.h (ENDIAN_LANE_N): Take the number
	of units rather than the mode.
	* config/aarch64/iterators.md (nunits): New mode attribute.
	* config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
	Use aarch64_endian_lane_rtx instead of GEN_INT (ENDIAN_LANE_N ...).
	* config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>)
	(aarch64_dup_lane_<vswap_width_name><mode>, *aarch64_mul3_elt<mode>)
	(*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
	(*aarch64_mul3_elt_to_64v2df, *aarch64_mla_elt<mode>): Likewise.
	(*aarch64_mla_elt_<vswap_width_name><mode>, *aarch64_mls_elt<mode>)
	(*aarch64_mls_elt_<vswap_width_name><mode>, *aarch64_fma4_elt<mode>)
	(*aarch64_fma4_elt_<vswap_width_name><mode>):: Likewise.
	(*aarch64_fma4_elt_to_64v2df, *aarch64_fnma4_elt<mode>): Likewise.
	(*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
	(*aarch64_fnma4_elt_to_64v2df, reduc_plus_scal_<mode>): Likewise.
	(reduc_plus_scal_v4sf, reduc_<maxmin_uns>_scal_<mode>): Likewise.
	(reduc_<maxmin_uns>_scal_<mode>): Likewise.
	(*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
	(*aarch64_get_lane_zero_extendsi<mode>): Likewise.
	(aarch64_get_lane<mode>, *aarch64_mulx_elt_<vswap_width_name><mode>)
	(*aarch64_mulx_elt<mode>, *aarch64_vgetfmulx<mode>): Likewise.
	(aarch64_sq<r>dmulh_lane<mode>, aarch64_sq<r>dmulh_laneq<mode>)
	(aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_lane<mode>): Likewise.
	(aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_laneq<mode>): Likewise.
	(aarch64_sqdml<SBINQOPS:as>l_lane<mode>): Likewise.
	(aarch64_sqdml<SBINQOPS:as>l_laneq<mode>): Likewise.
	(aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
	(aarch64_sqdml<SBINQOPS:as>l2_laneq<mode>_internal): Likewise.
	(aarch64_sqdmull_lane<mode>, aarch64_sqdmull_laneq<mode>): Likewise.
	(aarch64_sqdmull2_lane<mode>_internal): Likewise.
	(aarch64_sqdmull2_laneq<mode>_internal): Likewise.
	(aarch64_vec_load_lanesoi_lane<mode>): Likewise.
	(aarch64_vec_store_lanesoi_lane<mode>): Likewise.
	(aarch64_vec_load_lanesci_lane<mode>): Likewise.
	(aarch64_vec_store_lanesci_lane<mode>): Likewise.
	(aarch64_vec_load_lanesxi_lane<mode>): Likewise.
	(aarch64_vec_store_lanesxi_lane<mode>): Likewise.
	(aarch64_simd_vec_set<mode>): Update use of ENDIAN_LANE_N.
	(aarch64_simd_vec_setv2di): Likewise.

Reviewed-by: James Greenhalgh <james.greenhalgh@arm.com>

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>

From-SVN: r254466

7 files changed, 126 insertions, 66 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 889e36a..04f4d13 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,49 @@
+2017-11-06  Richard Sandiford  <richard.sandiford@linaro.org>
+	    Alan Hayward  <alan.hayward@arm.com>
+	    David Sherwood  <david.sherwood@arm.com>
+
+	* config/aarch64/aarch64-protos.h (aarch64_endian_lane_rtx): Declare.
+	* config/aarch64/aarch64.c (aarch64_endian_lane_rtx): New function.
+	* config/aarch64/aarch64.h (ENDIAN_LANE_N): Take the number
+	of units rather than the mode.
+	* config/aarch64/iterators.md (nunits): New mode attribute.
+	* config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
+	Use aarch64_endian_lane_rtx instead of GEN_INT (ENDIAN_LANE_N ...).
+	* config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>)
+	(aarch64_dup_lane_<vswap_width_name><mode>, *aarch64_mul3_elt<mode>)
+	(*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
+	(*aarch64_mul3_elt_to_64v2df, *aarch64_mla_elt<mode>): Likewise.
+	(*aarch64_mla_elt_<vswap_width_name><mode>, *aarch64_mls_elt<mode>)
+	(*aarch64_mls_elt_<vswap_width_name><mode>, *aarch64_fma4_elt<mode>)
+	(*aarch64_fma4_elt_<vswap_width_name><mode>):: Likewise.
+	(*aarch64_fma4_elt_to_64v2df, *aarch64_fnma4_elt<mode>): Likewise.
+	(*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
+	(*aarch64_fnma4_elt_to_64v2df, reduc_plus_scal_<mode>): Likewise.
+	(reduc_plus_scal_v4sf, reduc_<maxmin_uns>_scal_<mode>): Likewise.
+	(reduc_<maxmin_uns>_scal_<mode>): Likewise.
+	(*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
+	(*aarch64_get_lane_zero_extendsi<mode>): Likewise.
+	(aarch64_get_lane<mode>, *aarch64_mulx_elt_<vswap_width_name><mode>)
+	(*aarch64_mulx_elt<mode>, *aarch64_vgetfmulx<mode>): Likewise.
+	(aarch64_sq<r>dmulh_lane<mode>, aarch64_sq<r>dmulh_laneq<mode>)
+	(aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_lane<mode>): Likewise.
+	(aarch64_sqrdml<SQRDMLH_AS:rdma_as>h_laneq<mode>): Likewise.
+	(aarch64_sqdml<SBINQOPS:as>l_lane<mode>): Likewise.
+	(aarch64_sqdml<SBINQOPS:as>l_laneq<mode>): Likewise.
+	(aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
+	(aarch64_sqdml<SBINQOPS:as>l2_laneq<mode>_internal): Likewise.
+	(aarch64_sqdmull_lane<mode>, aarch64_sqdmull_laneq<mode>): Likewise.
+	(aarch64_sqdmull2_lane<mode>_internal): Likewise.
+	(aarch64_sqdmull2_laneq<mode>_internal): Likewise.
+	(aarch64_vec_load_lanesoi_lane<mode>): Likewise.
+	(aarch64_vec_store_lanesoi_lane<mode>): Likewise.
+	(aarch64_vec_load_lanesci_lane<mode>): Likewise.
+	(aarch64_vec_store_lanesci_lane<mode>): Likewise.
+	(aarch64_vec_load_lanesxi_lane<mode>): Likewise.
+	(aarch64_vec_store_lanesxi_lane<mode>): Likewise.
+	(aarch64_simd_vec_set<mode>): Update use of ENDIAN_LANE_N.
+	(aarch64_simd_vec_setv2di): Likewise.
+
 2017-11-06  Carl Love  <cel@us.ibm.com>
 
 	* config/rs6000/rs6000-c.c (P8V_BUILTIN_VEC_REVB): Add power 8
diff --git a/gcc/config/aarch64/aarch64-builtins.c b/gcc/config/aarch64/aarch64-builtins.c
index 242b2e3..ff4ab60 100644
--- a/gcc/config/aarch64/aarch64-builtins.c
+++ b/gcc/config/aarch64/aarch64-builtins.c
@@ -1067,8 +1067,8 @@ aarch64_simd_expand_args (rtx target, int icode, int have_retval,
 					    GET_MODE_NUNITS (builtin_mode),
 					    exp);
 		  /* Keep to GCC-vector-extension lane indices in the RTL.  */
-		  op[opc] =
-		    GEN_INT (ENDIAN_LANE_N (builtin_mode, INTVAL (op[opc])));
+		  op[opc] = aarch64_endian_lane_rtx (builtin_mode,
+						     INTVAL (op[opc]));
 		}
 	      goto constant_arg;
 
@@ -1081,7 +1081,7 @@ aarch64_simd_expand_args (rtx target, int icode, int have_retval,
 		  aarch64_simd_lane_bounds (op[opc],
 					    0, GET_MODE_NUNITS (vmode), exp);
 		  /* Keep to GCC-vector-extension lane indices in the RTL.  */
-		  op[opc] = GEN_INT (ENDIAN_LANE_N (vmode, INTVAL (op[opc])));
+		  op[opc] = aarch64_endian_lane_rtx (vmode, INTVAL (op[opc]));
 		}
 	      /* Fall through - if the lane index isn't a constant then
 		 the next case will error.  */
diff --git a/gcc/config/aarch64/aarch64-protos.h b/gcc/config/aarch64/aarch64-protos.h
index 345bfe8..4df2ee0 100644
--- a/gcc/config/aarch64/aarch64-protos.h
+++ b/gcc/config/aarch64/aarch64-protos.h
@@ -425,6 +425,7 @@ void aarch64_simd_emit_reg_reg_move (rtx *, machine_mode, unsigned int);
 rtx aarch64_simd_expand_builtin (int, tree, rtx);
 
 void aarch64_simd_lane_bounds (rtx, HOST_WIDE_INT, HOST_WIDE_INT, const_tree);
+rtx aarch64_endian_lane_rtx (machine_mode, unsigned int);
 
 void aarch64_split_128bit_move (rtx, rtx);
 
diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index a3600b3..445503d 100644
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -80,7 +80,7 @@
           )))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "dup\\t%0.<Vtype>, %1.<Vetype>[%2]";
   }
   [(set_attr "type" "neon_dup<q>")]
@@ -95,8 +95,7 @@
           )))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<VSWAP_WIDTH>mode,
-					  INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<VSWAP_WIDTH>mode, INTVAL (operands[2]));
     return "dup\\t%0.<Vtype>, %1.<Vetype>[%2]";
   }
   [(set_attr "type" "neon_dup<q>")]
@@ -502,7 +501,7 @@
       (match_operand:VMUL 3 "register_operand" "w")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "<f>mul\\t%0.<Vtype>, %3.<Vtype>, %1.<Vetype>[%2]";
   }
   [(set_attr "type" "neon<fp>_mul_<stype>_scalar<q>")]
@@ -518,8 +517,7 @@
       (match_operand:VMUL_CHANGE_NLANES 3 "register_operand" "w")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<VSWAP_WIDTH>mode,
-					  INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<VSWAP_WIDTH>mode, INTVAL (operands[2]));
     return "<f>mul\\t%0.<Vtype>, %3.<Vtype>, %1.<Vetype>[%2]";
   }
   [(set_attr "type" "neon<fp>_mul_<Vetype>_scalar<q>")]
@@ -572,7 +570,7 @@
        (match_operand:DF 3 "register_operand" "w")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (V2DFmode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (V2DFmode, INTVAL (operands[2]));
     return "fmul\\t%0.2d, %3.2d, %1.d[%2]";
   }
   [(set_attr "type" "neon_fp_mul_d_scalar_q")]
@@ -707,7 +705,7 @@
 	    (match_operand:SI 2 "immediate_operand" "i,i,i")))]
   "TARGET_SIMD"
   {
-   int elt = ENDIAN_LANE_N (<MODE>mode, exact_log2 (INTVAL (operands[2])));
+   int elt = ENDIAN_LANE_N (<nunits>, exact_log2 (INTVAL (operands[2])));
    operands[2] = GEN_INT ((HOST_WIDE_INT) 1 << elt);
    switch (which_alternative)
      {
@@ -1073,7 +1071,7 @@
 	    (match_operand:SI 2 "immediate_operand" "i,i")))]
   "TARGET_SIMD"
   {
-    int elt = ENDIAN_LANE_N (V2DImode, exact_log2 (INTVAL (operands[2])));
+    int elt = ENDIAN_LANE_N (2, exact_log2 (INTVAL (operands[2])));
     operands[2] = GEN_INT ((HOST_WIDE_INT) 1 << elt);
     switch (which_alternative)
       {
@@ -1110,7 +1108,7 @@
 	    (match_operand:SI 2 "immediate_operand" "i")))]
   "TARGET_SIMD"
   {
-    int elt = ENDIAN_LANE_N (<MODE>mode, exact_log2 (INTVAL (operands[2])));
+    int elt = ENDIAN_LANE_N (<nunits>, exact_log2 (INTVAL (operands[2])));
 
     operands[2] = GEN_INT ((HOST_WIDE_INT)1 << elt);
     return "ins\t%0.<Vetype>[%p2], %1.<Vetype>[0]";
@@ -1155,7 +1153,7 @@
 	 (match_operand:VDQHS 4 "register_operand" "0")))]
  "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "mla\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_mla_<Vetype>_scalar<q>")]
@@ -1173,8 +1171,7 @@
 	 (match_operand:VDQHS 4 "register_operand" "0")))]
  "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<VSWAP_WIDTH>mode,
-					  INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<VSWAP_WIDTH>mode, INTVAL (operands[2]));
     return "mla\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_mla_<Vetype>_scalar<q>")]
@@ -1214,7 +1211,7 @@
 	   (match_operand:VDQHS 3 "register_operand" "w"))))]
  "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "mls\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_mla_<Vetype>_scalar<q>")]
@@ -1232,8 +1229,7 @@
 	   (match_operand:VDQHS 3 "register_operand" "w"))))]
  "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<VSWAP_WIDTH>mode,
-					  INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<VSWAP_WIDTH>mode, INTVAL (operands[2]));
     return "mls\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_mla_<Vetype>_scalar<q>")]
@@ -1803,7 +1799,7 @@
       (match_operand:VDQF 4 "register_operand" "0")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "fmla\\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_fp_mla_<Vetype>_scalar<q>")]
@@ -1820,8 +1816,7 @@
       (match_operand:VDQSF 4 "register_operand" "0")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<VSWAP_WIDTH>mode,
-					  INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<VSWAP_WIDTH>mode, INTVAL (operands[2]));
     return "fmla\\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_fp_mla_<Vetype>_scalar<q>")]
@@ -1849,7 +1844,7 @@
       (match_operand:DF 4 "register_operand" "0")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (V2DFmode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (V2DFmode, INTVAL (operands[2]));
     return "fmla\\t%0.2d, %3.2d, %1.2d[%2]";
   }
   [(set_attr "type" "neon_fp_mla_d_scalar_q")]
@@ -1879,7 +1874,7 @@
       (match_operand:VDQF 4 "register_operand" "0")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "fmls\\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_fp_mla_<Vetype>_scalar<q>")]
@@ -1897,8 +1892,7 @@
       (match_operand:VDQSF 4 "register_operand" "0")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<VSWAP_WIDTH>mode,
-					  INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<VSWAP_WIDTH>mode, INTVAL (operands[2]));
     return "fmls\\t%0.<Vtype>, %3.<Vtype>, %1.<Vtype>[%2]";
   }
   [(set_attr "type" "neon_fp_mla_<Vetype>_scalar<q>")]
@@ -1928,7 +1922,7 @@
       (match_operand:DF 4 "register_operand" "0")))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (V2DFmode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (V2DFmode, INTVAL (operands[2]));
     return "fmls\\t%0.2d, %3.2d, %1.2d[%2]";
   }
   [(set_attr "type" "neon_fp_mla_d_scalar_q")]
@@ -2261,7 +2255,7 @@
 	       UNSPEC_ADDV)]
   "TARGET_SIMD"
   {
-    rtx elt = GEN_INT (ENDIAN_LANE_N (<MODE>mode, 0));
+    rtx elt = aarch64_endian_lane_rtx (<MODE>mode, 0);
     rtx scratch = gen_reg_rtx (<MODE>mode);
     emit_insn (gen_aarch64_reduc_plus_internal<mode> (scratch, operands[1]));
     emit_insn (gen_aarch64_get_lane<mode> (operands[0], scratch, elt));
@@ -2312,7 +2306,7 @@
 		    UNSPEC_FADDV))]
  "TARGET_SIMD"
 {
-  rtx elt = GEN_INT (ENDIAN_LANE_N (V4SFmode, 0));
+  rtx elt = aarch64_endian_lane_rtx (V4SFmode, 0);
   rtx scratch = gen_reg_rtx (V4SFmode);
   emit_insn (gen_aarch64_faddpv4sf (scratch, operands[1], operands[1]));
   emit_insn (gen_aarch64_faddpv4sf (scratch, scratch, scratch));
@@ -2354,7 +2348,7 @@
 		  FMAXMINV)]
   "TARGET_SIMD"
   {
-    rtx elt = GEN_INT (ENDIAN_LANE_N (<MODE>mode, 0));
+    rtx elt = aarch64_endian_lane_rtx (<MODE>mode, 0);
     rtx scratch = gen_reg_rtx (<MODE>mode);
     emit_insn (gen_aarch64_reduc_<maxmin_uns>_internal<mode> (scratch,
 							      operands[1]));
@@ -2370,7 +2364,7 @@
 		    MAXMINV)]
   "TARGET_SIMD"
   {
-    rtx elt = GEN_INT (ENDIAN_LANE_N (<MODE>mode, 0));
+    rtx elt = aarch64_endian_lane_rtx (<MODE>mode, 0);
     rtx scratch = gen_reg_rtx (<MODE>mode);
     emit_insn (gen_aarch64_reduc_<maxmin_uns>_internal<mode> (scratch,
 							      operands[1]));
@@ -2895,7 +2889,7 @@
 	    (parallel [(match_operand:SI 2 "immediate_operand" "i")]))))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "smov\\t%<GPI:w>0, %1.<VDQQH:Vetype>[%2]";
   }
   [(set_attr "type" "neon_to_gp<q>")]
@@ -2909,7 +2903,7 @@
 	    (parallel [(match_operand:SI 2 "immediate_operand" "i")]))))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "umov\\t%w0, %1.<Vetype>[%2]";
   }
   [(set_attr "type" "neon_to_gp<q>")]
@@ -2925,7 +2919,7 @@
 	  (parallel [(match_operand:SI 2 "immediate_operand" "i, i, i")])))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     switch (which_alternative)
       {
 	case 0:
@@ -3301,8 +3295,7 @@
 	 UNSPEC_FMULX))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<VSWAP_WIDTH>mode,
-					  INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<VSWAP_WIDTH>mode, INTVAL (operands[3]));
     return "fmulx\t%<v>0<Vmtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_fp_mul_<Vetype>_scalar<q>")]
@@ -3321,7 +3314,7 @@
 	 UNSPEC_FMULX))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[3]));
     return "fmulx\t%<v>0<Vmtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_fp_mul_<Vetype><q>")]
@@ -3355,7 +3348,7 @@
 	 UNSPEC_FMULX))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[3]));
     return "fmulx\t%<Vetype>0, %<Vetype>1, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "fmul<Vetype>")]
@@ -3441,7 +3434,7 @@
 	 VQDMULH))]
   "TARGET_SIMD"
   "*
-   operands[3] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[3])));
+   operands[3] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[3]));
    return \"sq<r>dmulh\\t%0.<Vtype>, %1.<Vtype>, %2.<Vetype>[%3]\";"
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar<q>")]
 )
@@ -3456,7 +3449,7 @@
 	 VQDMULH))]
   "TARGET_SIMD"
   "*
-   operands[3] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[3])));
+   operands[3] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[3]));
    return \"sq<r>dmulh\\t%0.<Vtype>, %1.<Vtype>, %2.<Vetype>[%3]\";"
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar<q>")]
 )
@@ -3471,7 +3464,7 @@
 	 VQDMULH))]
   "TARGET_SIMD"
   "*
-   operands[3] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[3])));
+   operands[3] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[3]));
    return \"sq<r>dmulh\\t%<v>0, %<v>1, %2.<v>[%3]\";"
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar<q>")]
 )
@@ -3486,7 +3479,7 @@
 	 VQDMULH))]
   "TARGET_SIMD"
   "*
-   operands[3] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[3])));
+   operands[3] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[3]));
    return \"sq<r>dmulh\\t%<v>0, %<v>1, %2.<v>[%3]\";"
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar<q>")]
 )
@@ -3518,7 +3511,7 @@
 	  SQRDMLH_AS))]
    "TARGET_SIMD_RDMA"
    {
-     operands[4] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[4])));
+     operands[4] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[4]));
      return
       "sqrdml<SQRDMLH_AS:rdma_as>h\\t%0.<Vtype>, %2.<Vtype>, %3.<Vetype>[%4]";
    }
@@ -3536,7 +3529,7 @@
 	  SQRDMLH_AS))]
    "TARGET_SIMD_RDMA"
    {
-     operands[4] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[4])));
+     operands[4] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[4]));
      return
       "sqrdml<SQRDMLH_AS:rdma_as>h\\t%<v>0, %<v>2, %3.<Vetype>[%4]";
    }
@@ -3556,7 +3549,7 @@
 	  SQRDMLH_AS))]
    "TARGET_SIMD_RDMA"
    {
-     operands[4] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[4])));
+     operands[4] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[4]));
      return
       "sqrdml<SQRDMLH_AS:rdma_as>h\\t%0.<Vtype>, %2.<Vtype>, %3.<Vetype>[%4]";
    }
@@ -3574,7 +3567,7 @@
 	  SQRDMLH_AS))]
    "TARGET_SIMD_RDMA"
    {
-     operands[4] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[4])));
+     operands[4] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[4]));
      return
       "sqrdml<SQRDMLH_AS:rdma_as>h\\t%<v>0, %<v>2, %3.<v>[%4]";
    }
@@ -3618,7 +3611,7 @@
 	    (const_int 1))))]
   "TARGET_SIMD"
   {
-    operands[4] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[4])));
+    operands[4] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[4]));
     return
       "sqdml<SBINQOPS:as>l\\t%<vw2>0<Vmwtype>, %<v>2<Vmtype>, %3.<Vetype>[%4]";
   }
@@ -3642,7 +3635,7 @@
 	    (const_int 1))))]
   "TARGET_SIMD"
   {
-    operands[4] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[4])));
+    operands[4] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[4]));
     return
       "sqdml<SBINQOPS:as>l\\t%<vw2>0<Vmwtype>, %<v>2<Vmtype>, %3.<Vetype>[%4]";
   }
@@ -3665,7 +3658,7 @@
 	    (const_int 1))))]
   "TARGET_SIMD"
   {
-    operands[4] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[4])));
+    operands[4] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[4]));
     return
       "sqdml<SBINQOPS:as>l\\t%<vw2>0<Vmwtype>, %<v>2<Vmtype>, %3.<Vetype>[%4]";
   }
@@ -3688,7 +3681,7 @@
 	    (const_int 1))))]
   "TARGET_SIMD"
   {
-    operands[4] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[4])));
+    operands[4] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[4]));
     return
       "sqdml<SBINQOPS:as>l\\t%<vw2>0<Vmwtype>, %<v>2<Vmtype>, %3.<Vetype>[%4]";
   }
@@ -3783,7 +3776,7 @@
 	      (const_int 1))))]
   "TARGET_SIMD"
   {
-    operands[4] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[4])));
+    operands[4] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[4]));
     return
      "sqdml<SBINQOPS:as>l2\\t%<vw2>0<Vmwtype>, %<v>2<Vmtype>, %3.<Vetype>[%4]";
   }
@@ -3809,7 +3802,7 @@
 	      (const_int 1))))]
   "TARGET_SIMD"
   {
-    operands[4] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[4])));
+    operands[4] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[4]));
     return
      "sqdml<SBINQOPS:as>l2\\t%<vw2>0<Vmwtype>, %<v>2<Vmtype>, %3.<Vetype>[%4]";
   }
@@ -3956,7 +3949,7 @@
 	     (const_int 1)))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[3]));
     return "sqdmull\\t%<vw2>0<Vmwtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar_long")]
@@ -3977,7 +3970,7 @@
 	     (const_int 1)))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[3]));
     return "sqdmull\\t%<vw2>0<Vmwtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar_long")]
@@ -3997,7 +3990,7 @@
 	     (const_int 1)))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[3]));
     return "sqdmull\\t%<vw2>0<Vmwtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar_long")]
@@ -4017,7 +4010,7 @@
 	     (const_int 1)))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[3]));
     return "sqdmull\\t%<vw2>0<Vmwtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar_long")]
@@ -4095,7 +4088,7 @@
 	     (const_int 1)))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<VCOND>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<VCOND>mode, INTVAL (operands[3]));
     return "sqdmull2\\t%<vw2>0<Vmwtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar_long")]
@@ -4118,7 +4111,7 @@
 	     (const_int 1)))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<VCONQ>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<VCONQ>mode, INTVAL (operands[3]));
     return "sqdmull2\\t%<vw2>0<Vmwtype>, %<v>1<Vmtype>, %2.<Vetype>[%3]";
   }
   [(set_attr "type" "neon_sat_mul_<Vetype>_scalar_long")]
@@ -4624,7 +4617,7 @@
 		   UNSPEC_LD2_LANE))]
   "TARGET_SIMD"
   {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[3]));
     return "ld2\\t{%S0.<Vetype> - %T0.<Vetype>}[%3], %1";
   }
   [(set_attr "type" "neon_load2_one_lane")]
@@ -4668,7 +4661,7 @@
 		   UNSPEC_ST2_LANE))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "st2\\t{%S1.<Vetype> - %T1.<Vetype>}[%2], %0";
   }
   [(set_attr "type" "neon_store2_one_lane<q>")]
@@ -4722,7 +4715,7 @@
 		   UNSPEC_LD3_LANE))]
   "TARGET_SIMD"
 {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[3]));
     return "ld3\\t{%S0.<Vetype> - %U0.<Vetype>}[%3], %1";
 }
   [(set_attr "type" "neon_load3_one_lane")]
@@ -4766,7 +4759,7 @@
 		    UNSPEC_ST3_LANE))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "st3\\t{%S1.<Vetype> - %U1.<Vetype>}[%2], %0";
   }
   [(set_attr "type" "neon_store3_one_lane<q>")]
@@ -4820,7 +4813,7 @@
 		   UNSPEC_LD4_LANE))]
   "TARGET_SIMD"
 {
-    operands[3] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[3])));
+    operands[3] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[3]));
     return "ld4\\t{%S0.<Vetype> - %V0.<Vetype>}[%3], %1";
 }
   [(set_attr "type" "neon_load4_one_lane")]
@@ -4864,7 +4857,7 @@
 		    UNSPEC_ST4_LANE))]
   "TARGET_SIMD"
   {
-    operands[2] = GEN_INT (ENDIAN_LANE_N (<MODE>mode, INTVAL (operands[2])));
+    operands[2] = aarch64_endian_lane_rtx (<MODE>mode, INTVAL (operands[2]));
     return "st4\\t{%S1.<Vetype> - %V1.<Vetype>}[%2], %0";
   }
   [(set_attr "type" "neon_store4_one_lane<q>")]
diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
index 71c3693..ffcca32 100644
--- a/gcc/config/aarch64/aarch64.c
+++ b/gcc/config/aarch64/aarch64.c
@@ -11833,6 +11833,15 @@ aarch64_simd_lane_bounds (rtx operand, HOST_WIDE_INT low, HOST_WIDE_INT high,
   }
 }
 
+/* Peform endian correction on lane number N, which indexes a vector
+   of mode MODE, and return the result as an SImode rtx.  */
+
+rtx
+aarch64_endian_lane_rtx (machine_mode mode, unsigned int n)
+{
+  return gen_int_mode (ENDIAN_LANE_N (GET_MODE_NUNITS (mode), n), SImode);
+}
+
 /* Return TRUE if OP is a valid vector addressing mode.  */
 bool
 aarch64_simd_mem_operand_p (rtx op)
diff --git a/gcc/config/aarch64/aarch64.h b/gcc/config/aarch64/aarch64.h
index 5e2e2fc..93d29b8 100644
--- a/gcc/config/aarch64/aarch64.h
+++ b/gcc/config/aarch64/aarch64.h
@@ -913,8 +913,8 @@ extern enum aarch64_code_model aarch64_cmodel;
    || (MODE) == V4SFmode || (MODE) == V8HFmode || (MODE) == V2DImode \
    || (MODE) == V2DFmode)
 
-#define ENDIAN_LANE_N(mode, n)  \
-  (BYTES_BIG_ENDIAN ? GET_MODE_NUNITS (mode) - 1 - n : n)
+#define ENDIAN_LANE_N(NUNITS, N) \
+  (BYTES_BIG_ENDIAN ? NUNITS - 1 - N : N)
 
 /* Support for a configure-time default CPU, etc.  We currently support
    --with-arch and --with-cpu.  Both are ignored if either is specified
diff --git a/gcc/config/aarch64/iterators.md b/gcc/config/aarch64/iterators.md
index 46aa6fd..5d7b0f3 100644
--- a/gcc/config/aarch64/iterators.md
+++ b/gcc/config/aarch64/iterators.md
@@ -450,6 +450,17 @@
 (define_mode_attr rtn [(DI "d") (SI "")])
 (define_mode_attr vas [(DI "") (SI ".2s")])
 
+;; Map a vector to the number of units in it, if the size of the mode
+;; is constant.
+(define_mode_attr nunits [(V8QI "8") (V16QI "16")
+			  (V4HI "4") (V8HI "8")
+			  (V2SI "2") (V4SI "4")
+				     (V2DI "2")
+			  (V4HF "4") (V8HF "8")
+			  (V2SF "2") (V4SF "4")
+			  (V1DF "1") (V2DF "2")
+			  (DI "1") (DF "1")])
+
 ;; Map a floating point or integer mode to the appropriate register name prefix
 (define_mode_attr s [(HF "h") (SF "s") (DF "d") (SI "s") (DI "d")])