aarch64: Enable parsing of user-provided AArch64 CPU tuning parameters

This patch adds support for loading custom CPU tuning parameters from a JSON
file for AArch64 targets. The '-muser-provided-CPU=' flag accepts a user
provided JSON file and overrides the internal tuning parameters at GCC runtime.

This patch was bootstrapped and regtested on aarch64-linux-gnu, no regression.

Signed-off-by: Soumya AR <soumyaa@nvidia.com>

gcc/ChangeLog:

	* config.gcc: Add aarch64-json-tunings-parser.o.
	* config/aarch64/aarch64.cc (aarch64_override_options_internal): Invoke
	aarch64_load_tuning_params_from_json if -muser-provided-CPU= is
	(aarch64_json_tunings_tests): Extern aarch64_json_tunings_tests().
	(aarch64_run_selftests): Add aarch64_json_tunings_tests().
	* config/aarch64/aarch64.opt: New option.
	* config/aarch64/t-aarch64 (aarch64-json-tunings-parser.o): New define.
	* config/aarch64/aarch64-json-schema.h: New file.
	* config/aarch64/aarch64-json-tunings-parser.cc: New file.
	* config/aarch64/aarch64-json-tunings-parser.h: New file.

gcc/config.gcc

@@ -359,7 +359,7 @@ aarch64*-*-*)
 	c_target_objs="aarch64-c.o"
 	cxx_target_objs="aarch64-c.o"
 	d_target_objs="aarch64-d.o"
-	extra_objs="aarch64-builtins.o aarch-common.o aarch64-elf-metadata.o aarch64-sve-builtins.o aarch64-sve-builtins-shapes.o aarch64-sve-builtins-base.o aarch64-sve-builtins-sve2.o aarch64-sve-builtins-sme.o cortex-a57-fma-steering.o aarch64-speculation.o aarch-bti-insert.o aarch64-early-ra.o aarch64-ldp-fusion.o aarch64-sched-dispatch.o aarch64-json-tunings-printer.o"
+	extra_objs="aarch64-builtins.o aarch-common.o aarch64-elf-metadata.o aarch64-sve-builtins.o aarch64-sve-builtins-shapes.o aarch64-sve-builtins-base.o aarch64-sve-builtins-sve2.o aarch64-sve-builtins-sme.o cortex-a57-fma-steering.o aarch64-speculation.o aarch-bti-insert.o aarch64-early-ra.o aarch64-ldp-fusion.o aarch64-sched-dispatch.o aarch64-json-tunings-printer.o aarch64-json-tunings-parser.o"
 	target_gtfiles="\$(srcdir)/config/aarch64/aarch64-protos.h \$(srcdir)/config/aarch64/aarch64-builtins.h \$(srcdir)/config/aarch64/aarch64-builtins.cc \$(srcdir)/config/aarch64/aarch64-sve-builtins.h \$(srcdir)/config/aarch64/aarch64-sve-builtins.cc"
 	target_has_targetm_common=yes
 	;;

gcc/config/aarch64/aarch64-json-schema.h

@@ -0,0 +1,264 @@
+/* Raw JSON schema for the AArch64 tuning parameters.
+   Copyright The GNU Toolchain Authors.
+
+   This file is part of GCC.
+
+   GCC is free software; you can redistribute it and/or modify it
+   under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3, or (at your option)
+   any later version.
+
+   GCC is distributed in the hope that it will be useful, but
+   WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with GCC; see the file COPYING3.  If not see
+   <http://www.gnu.org/licenses/>.  */
+
+#ifndef AARCH64_JSON_SCHEMA_H
+#define AARCH64_JSON_SCHEMA_H
+
+static const char *schema_json = R"json(
+{
+  "metadata": {
+    "gcc_version": "int"
+  },
+  "tune_params": {
+    "insn_extra_cost": {
+      "alu": {
+	"arith": "int",
+	"logical": "int",
+	"shift": "int",
+	"shift_reg": "int",
+	"arith_shift": "int",
+	"arith_shift_reg": "int",
+	"log_shift": "int",
+	"log_shift_reg": "int",
+	"extend": "int",
+	"extend_arith": "int",
+	"bfi": "int",
+	"bfx": "int",
+	"clz": "int",
+	"rev": "int",
+	"non_exec": "int",
+	"non_exec_costs_exec": "boolean"
+      },
+      "mult": [
+	{
+	  "simple": "int",
+	  "flag_setting": "int",
+	  "extend": "int",
+	  "add": "int",
+	  "extend_add": "int",
+	  "idiv": "int"
+	},
+	{
+	  "simple": "int",
+	  "flag_setting": "int",
+	  "extend": "int",
+	  "add": "int",
+	  "extend_add": "int",
+	  "idiv": "int"
+	}
+      ],
+      "ldst": {
+	"load": "int",
+	"load_sign_extend": "int",
+	"ldrd": "int",
+	"ldm_1st": "int",
+	"ldm_regs_per_insn_1st": "int",
+	"ldm_regs_per_insn_subsequent": "int",
+	"loadf": "int",
+	"loadd": "int",
+	"load_unaligned": "int",
+	"store": "int",
+	"strd": "int",
+	"stm_1st": "int",
+	"stm_regs_per_insn_1st": "int",
+	"stm_regs_per_insn_subsequent": "int",
+	"storef": "int",
+	"stored": "int",
+	"store_unaligned": "int",
+	"loadv": "int",
+	"storev": "int"
+      },
+      "fp": [
+	{
+	  "div": "int",
+	  "mult": "int",
+	  "mult_addsub": "int",
+	  "fma": "int",
+	  "addsub": "int",
+	  "fpconst": "int",
+	  "neg": "int",
+	  "compare": "int",
+	  "widen": "int",
+	  "narrow": "int",
+	  "toint": "int",
+	  "fromint": "int",
+	  "roundint": "int"
+	},
+	{
+	  "div": "int",
+	  "mult": "int",
+	  "mult_addsub": "int",
+	  "fma": "int",
+	  "addsub": "int",
+	  "fpconst": "int",
+	  "neg": "int",
+	  "compare": "int",
+	  "widen": "int",
+	  "narrow": "int",
+	  "toint": "int",
+	  "fromint": "int",
+	  "roundint": "int"
+	}
+      ],
+      "vect": {
+	"alu": "int",
+	"mult": "int",
+	"movi": "int",
+	"dup": "int",
+	"extract": "int"
+      }
+    },
+    "addr_cost": {
+      "addr_scale_costs": {
+	"hi": "int",
+	"si": "int",
+	"di": "int",
+	"ti": "int"
+      },
+      "pre_modify": "int",
+      "post_modify": "int",
+      "post_modify_ld3_st3": "int",
+      "post_modify_ld4_st4": "int",
+      "register_offset": "int",
+      "register_sextend": "int",
+      "register_zextend": "int",
+      "imm_offset": "int"
+    },
+    "regmove_cost": {
+      "GP2GP": "int",
+      "GP2FP": "int",
+      "FP2GP": "int",
+      "FP2FP": "int"
+    },
+    "vec_costs": {
+      "scalar_int_stmt_cost": "int",
+      "scalar_fp_stmt_cost": "int",
+      "scalar_load_cost": "int",
+      "scalar_store_cost": "int",
+      "cond_taken_branch_cost": "int",
+      "cond_not_taken_branch_cost": "int",
+      "advsimd": {
+	"int_stmt_cost": "int",
+	"fp_stmt_cost": "int",
+	"ld2_st2_permute_cost": "int",
+	"ld3_st3_permute_cost": "int",
+	"ld4_st4_permute_cost": "int",
+	"permute_cost": "int",
+	"reduc_i8_cost": "int",
+	"reduc_i16_cost": "int",
+	"reduc_i32_cost": "int",
+	"reduc_i64_cost": "int",
+	"reduc_f16_cost": "int",
+	"reduc_f32_cost": "int",
+	"reduc_f64_cost": "int",
+	"store_elt_extra_cost": "int",
+	"vec_to_scalar_cost": "int",
+	"scalar_to_vec_cost": "int",
+	"align_load_cost": "int",
+	"unalign_load_cost": "int",
+	"unalign_store_cost": "int",
+	"store_cost": "int"
+      },
+      "sve": {
+	"clast_cost": "int",
+	"fadda_f16_cost": "int",
+	"fadda_f32_cost": "int",
+	"fadda_f64_cost": "int",
+	"gather_load_x32_cost": "uint",
+	"gather_load_x64_cost": "uint",
+	"gather_load_x32_init_cost": "int",
+	"gather_load_x64_init_cost": "int",
+	"scatter_store_elt_cost": "int"
+      },
+      "issue_info": {
+	"scalar": {
+	  "loads_stores_per_cycle": "uint",
+	  "stores_per_cycle": "uint",
+	  "general_ops_per_cycle": "uint",
+	  "fp_simd_load_general_ops": "uint",
+	  "fp_simd_store_general_ops": "uint"
+	},
+	"advsimd": {
+	  "loads_stores_per_cycle": "uint",
+	  "stores_per_cycle": "uint",
+	  "general_ops_per_cycle": "uint",
+	  "fp_simd_load_general_ops": "uint",
+	  "fp_simd_store_general_ops": "uint",
+	  "ld2_st2_general_ops": "uint",
+	  "ld3_st3_general_ops": "uint",
+	  "ld4_st4_general_ops": "uint"
+	},
+	"sve": {
+	  "loads_stores_per_cycle": "uint",
+	  "stores_per_cycle": "uint",
+	  "general_ops_per_cycle": "uint",
+	  "fp_simd_load_general_ops": "uint",
+	  "fp_simd_store_general_ops": "uint",
+	  "ld2_st2_general_ops": "uint",
+	  "ld3_st3_general_ops": "uint",
+	  "ld4_st4_general_ops": "uint",
+	  "pred_ops_per_cycle": "uint",
+	  "while_pred_ops": "uint",
+	  "int_cmp_pred_ops": "uint",
+	  "fp_cmp_pred_ops": "uint",
+	  "gather_scatter_pair_general_ops": "uint",
+	  "gather_scatter_pair_pred_ops": "uint"
+	}
+      }
+    },
+    "branch_costs": { "predictable": "int", "unpredictable": "int" },
+    "approx_modes": { "division": "int", "sqrt": "int", "recip_sqrt": "int" },
+    "sve_width": "uint",
+    "memmov_cost": {
+      "load_int": "int",
+      "store_int": "int",
+      "load_fp": "int",
+      "store_fp": "int",
+      "load_pred": "int",
+      "store_pred": "int"
+    },
+    "issue_rate": "int",
+    "fusible_ops": "uint",
+    "function_align": "string",
+    "jump_align": "string",
+    "loop_align": "string",
+    "int_reassoc_width": "int",
+    "fp_reassoc_width": "int",
+    "fma_reassoc_width": "int",
+    "vec_reassoc_width": "int",
+    "min_div_recip_mul_sf": "int",
+    "min_div_recip_mul_df": "int",
+    "max_case_values": "uint",
+    "autoprefetcher_model": "enum",
+    "extra_tuning_flags": "uint",
+    "prefetch": {
+      "num_slots": "int",
+      "l1_cache_size": "int",
+      "l1_cache_line_size": "int",
+      "l2_cache_size": "int",
+      "prefetch_dynamic_strides": "boolean",
+      "minimum_stride": "int",
+      "default_opt_level": "int"
+    },
+    "ldp_policy_model": "enum",
+    "stp_policy_model": "enum"
+  }
+})json";
+
+#endif

gcc/config/aarch64/aarch64-json-tunings-parser.h

@@ -0,0 +1,29 @@
+/* Routines to parse the AArch64 tuning parameters from a JSON file.
+   Copyright The GNU Toolchain Authors.
+
+   This file is part of GCC.
+
+   GCC is free software; you can redistribute it and/or modify it
+   under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 3, or (at your option)
+   any later version.
+
+   GCC is distributed in the hope that it will be useful, but
+   WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with GCC; see the file COPYING3.  If not see
+   <http://www.gnu.org/licenses/>.  */
+
+#ifndef AARCH64_JSON_TUNINGS_PARSER_H
+#define AARCH64_JSON_TUNINGS_PARSER_H
+
+#include "aarch64-protos.h"
+
+void
+aarch64_load_tuning_params_from_json (const char *data_filename,
+				      struct tune_params *tune);
+
+#endif

gcc/config/aarch64/aarch64.cc

@@ -100,6 +100,7 @@
 #include "hash-map.h"
 #include "aarch64-sched-dispatch.h"
 #include "aarch64-json-tunings-printer.h"
+#include "aarch64-json-tunings-parser.h"
 
 /* This file should be included last.  */
 #include "target-def.h"
@@ -19170,6 +19171,21 @@ aarch64_override_options_internal (struct gcc_options *opts)
     aarch64_parse_override_string (opts->x_aarch64_override_tune_string,
 				   &aarch64_tune_params);
 
+  /* We need to parse the JSON file only once per program execution.  */
+  if (opts->x_muser_provided_CPU)
+    {
+      static bool json_parsed = false;
+      static struct tune_params aarch64_json_params;
+      if (!json_parsed)
+	{
+	  aarch64_json_params = *(tune->tune);
+	  aarch64_load_tuning_params_from_json (opts->x_muser_provided_CPU,
+						&aarch64_json_params);
+	  json_parsed = true;
+	}
+      aarch64_tune_params = aarch64_json_params;
+    }
+
   if (opts->x_aarch64_ldp_policy_param)
     aarch64_tune_params.ldp_policy_model = opts->x_aarch64_ldp_policy_param;
 
@@ -32425,6 +32441,8 @@ aarch64_test_sve_folding ()
     }
 }
 
+extern void aarch64_json_tunings_tests ();
+
 /* Run all target-specific selftests.  */
 
 static void
@@ -32434,6 +32452,7 @@ aarch64_run_selftests (void)
   aarch64_test_fractional_cost ();
   aarch64_test_sysreg_encoding_clashes ();
   aarch64_test_sve_folding ();
+  aarch64_json_tunings_tests ();
 }
 
 } // namespace selftest

gcc/config/aarch64/aarch64.opt

@@ -197,6 +197,10 @@ fdump-tuning-model=
 Target Undocumented RejectNegative Negative(fdump-tuning-model=) Joined Var(fdump_tuning_model)
 -fdump-tuning-model=<filename> Dump current tuning model to a JSON file.
 
+muser-provided-CPU=
+Target Undocumented RejectNegative Negative(muser-provided-CPU=) Joined Var(muser_provided_CPU)
+-muser-provided-CPU=<json-tuning-file> User specific CPU tunings.
+
 moverride=
 Target RejectNegative ToLower Joined Var(aarch64_override_tune_string) Save
 -moverride=<string>	Power users only! Override CPU optimization parameters.

gcc/config/aarch64/t-aarch64

@@ -219,6 +219,16 @@ aarch64-json-tunings-printer.o: $(srcdir)/config/aarch64/aarch64-json-tunings-pr
 	$(COMPILER) -c $(ALL_COMPILERFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) \
 		$(srcdir)/config/aarch64/aarch64-json-tunings-printer.cc
 
+aarch64-json-tunings-parser.o: $(srcdir)/config/aarch64/aarch64-json-tunings-parser.cc \
+    $(CONFIG_H) $(SYSTEM_H) $(CORETYPES_H) $(TM_H) $(DIAGNOSTIC_CORE_H) \
+    json-parsing.h \
+    $(srcdir)/config/aarch64/aarch64-json-schema.h \
+    $(srcdir)/config/aarch64/aarch64-json-tunings-parser.h \
+    $(srcdir)/config/aarch64/aarch64-protos.h \
+    $(srcdir)/config/arm/aarch-common-protos.h
+	$(COMPILER) -c $(ALL_COMPILERFLAGS) $(ALL_CPPFLAGS) $(INCLUDES) \
+		$(srcdir)/config/aarch64/aarch64-json-tunings-parser.cc
+
 comma=,
 MULTILIB_OPTIONS    = $(subst $(comma),/, $(patsubst %, mabi=%, $(subst $(comma),$(comma)mabi=,$(TM_MULTILIB_CONFIG))))
 MULTILIB_DIRNAMES   = $(subst $(comma), ,$(TM_MULTILIB_CONFIG))