[PATCH][ARM] FreeBSD arm support, EABI, v2

[PATCH][ARM] FreeBSD arm support, EABI, v2

[Andreas Tobler]

[-- Attachment #1: Type: text/plain, Size: 2138 bytes --]

Hi all,

this is the second attempt.

I reworked the issues Richard mentioned in the previous review.
I also found one issue which will break build/bootstrap if I pass
--enable-gnu-indirect-function, also fixed.

One thing which came up is the way we generate code for the
armv6*-*-freebsd* triplet versus the arm-*-freebsd* triplet.

I think the thing which confuses is the fact that we have only two fixed 
triplets where we build a complete OS with. Means the whole OS is built 
with the same optimization, not only the kernel or one binary.

For the armv6* we want to benefit from the cpu's functionality by 
default. We build all __ARM_ARCH >= 6 with TARGET_CPU_arm1176jzs,
on the other hand all __ARM_ARCH <=5 will be built with TARGET_CPU_arm9.

Now who becomes arm-*-freebsd* and who becomes armv6*-*-freebsd*?

As tried above, we only know two triplets, so __ARM_ARCH >= 6 becomes
armv6*-*-freebsd* and __ARM_ARCH <=5 becomes arm-*-freebsd*.

armv8 is not yet in the portfolio and it will become something 
different, either arm64 or aarch64, I do not know.

I'd like to keep this since our system compilers, clang and gcc-4.2.1 
behave the same.
If we have to change here, we would confuse people quite a lot.

The whole thing is FreeBSD specific and does not touch others.

As usual, bootstrapped, cross compiled, tested.

Ok for trunk?

TIA,
Andreas

toplevel:

	* configure.ac: Don't add ${libgcj} for arm*-*-freebsd*.
	* configure: Regenerate.
gcc:
	* config.gcc (arm*-*-freebsd*): New configuration.
	* config/arm/freebsd.h: New file.
	* config.host: Add extra components for arm*-*-freebsd*.
	* config/arm/arm.h: Introduce MAX_SYNC_LIBFUNC_SIZE.
	* config/arm/arm.c (arm_init_libfuncs): Use MAX_SYNC_LIBFUNC_SIZE.

libgcc:

	* config.host (arm*-*-freebsd*): Add new configuration for
	arm*-*-freebsd*.
	* config/arm/freebsd-atomic.c: New file.
	* config/arm/t-freebsd: Likewise.
	* config/arm/unwind-arm.h: Add __FreeBSD__ to the list of
	'PC-relative indirect' OS's.

libatomic:

	* configure.tgt: Exclude arm*-*-freebsd* from try_ifunc.

libstdc++-v3:

	* configure.host: Add arm*-*-freebsd* port_specific_symbol_files.

[-- Attachment #2: gcc5_arm_fbsd_20141127-1.diff --]
[-- Type: text/plain, Size: 22654 bytes --]

Index: configure
===================================================================
--- configure	(revision 218129)
+++ configure	(working copy)
@@ -3427,6 +3427,9 @@
   alpha*-*-*vms*)
     noconfigdirs="$noconfigdirs ${libgcj}"
     ;;
+  arm*-*-freebsd*)
+    noconfigdirs="$noconfigdirs ${libgcj}"
+    ;;
   arm-wince-pe)
     noconfigdirs="$noconfigdirs ${libgcj}"
     ;;
Index: configure.ac
===================================================================
--- configure.ac	(revision 218129)
+++ configure.ac	(working copy)
@@ -781,6 +781,9 @@
   alpha*-*-*vms*)
     noconfigdirs="$noconfigdirs ${libgcj}"
     ;;
+  arm*-*-freebsd*)
+    noconfigdirs="$noconfigdirs ${libgcj}"
+    ;;
   arm-wince-pe)
     noconfigdirs="$noconfigdirs ${libgcj}"
     ;;
Index: gcc/config/arm/arm.c
===================================================================
--- gcc/config/arm/arm.c	(revision 218129)
+++ gcc/config/arm/arm.c	(working copy)
@@ -2160,7 +2160,7 @@
 {
   /* For Linux, we have access to kernel support for atomic operations.  */
   if (arm_abi == ARM_ABI_AAPCS_LINUX)
-    init_sync_libfuncs (2 * UNITS_PER_WORD);
+    init_sync_libfuncs (MAX_SYNC_LIBFUNC_SIZE);
 
   /* There are no special library functions unless we are using the
      ARM BPABI.  */
Index: gcc/config/arm/arm.h
===================================================================
--- gcc/config/arm/arm.h	(revision 218129)
+++ gcc/config/arm/arm.h	(working copy)
@@ -766,6 +766,11 @@
 #define PCC_BITFIELD_TYPE_MATTERS TARGET_AAPCS_BASED
 #endif
 
+/* The maximum size of the sync library functions supported.  */
+#ifndef MAX_SYNC_LIBFUNC_SIZE
+#define MAX_SYNC_LIBFUNC_SIZE (2 * UNITS_PER_WORD);
+#endif
+
 \f
 /* Standard register usage.  */
 
Index: gcc/config/arm/freebsd.h
===================================================================
--- gcc/config/arm/freebsd.h	(revision 0)
+++ gcc/config/arm/freebsd.h	(working copy)
@@ -0,0 +1,180 @@
+/* Definitions of target machine for GNU compiler, FreeBSD/arm version.
+   Copyright (C) 2002-2014 Free Software Foundation, Inc.
+   Contributed by Wasabi Systems, Inc.
+
+   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.
+
+   Under Section 7 of GPL version 3, you are granted additional
+   permissions described in the GCC Runtime Library Exception, version
+   3.1, as published by the Free Software Foundation.
+
+   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/>.  */
+
+#undef  SUBTARGET_CPP_SPEC
+#define SUBTARGET_CPP_SPEC FBSD_CPP_SPEC
+
+#undef  SUBTARGET_EXTRA_SPECS
+#define SUBTARGET_EXTRA_SPECS					\
+  { "subtarget_extra_asm_spec",	SUBTARGET_EXTRA_ASM_SPEC },	\
+  { "subtarget_asm_float_spec", SUBTARGET_ASM_FLOAT_SPEC }, 	\
+  { "fbsd_dynamic_linker", FBSD_DYNAMIC_LINKER }
+
+#undef SUBTARGET_EXTRA_ASM_SPEC
+#define SUBTARGET_EXTRA_ASM_SPEC \
+  "%{mabi=apcs-gnu|mabi=atpcs:-meabi=gnu;:-meabi=5} " TARGET_FIX_V4BX_SPEC " \
+  %{fpic|fpie:-k} %{fPIC|fPIE:-k}"
+
+#undef SUBTARGET_ASM_FLOAT_SPEC
+#ifdef TARGET_FREEBSD_ARM_HARD_FLOAT
+/* Default to full vfp if we build for arm*hf.  */
+#define SUBTARGET_ASM_FLOAT_SPEC "%{!mfpu=*:-mfpu=vfp}"
+#else
+#define SUBTARGET_ASM_FLOAT_SPEC "%{!mfpu=*:-mfpu=softvfp}"
+#endif
+
+#undef	LINK_SPEC
+#define LINK_SPEC "							\
+  %{p:%nconsider using `-pg' instead of `-p' with gprof (1) }		\
+  %{v:-V}								\
+  %{assert*} %{R*} %{rpath*} %{defsym*}					\
+  %{shared:-Bshareable %{h*} %{soname*}}				\
+  %{!shared:								\
+    %{!static:								\
+      %{rdynamic:-export-dynamic}					\
+      %{!dynamic-linker:-dynamic-linker %(fbsd_dynamic_linker) }}	\
+    %{static:-Bstatic}}							\
+  %{!static:--hash-style=both --enable-new-dtags}			\
+  %{symbolic:-Bsymbolic}						\
+  -X %{mbig-endian:-EB} %{mlittle-endian:-EL}"
+
+/* TARGET_BIG_ENDIAN_DEFAULT is set in
+   config.gcc for big endian configurations.  */
+#if TARGET_BIG_ENDIAN_DEFAULT
+#define TARGET_ENDIAN_DEFAULT    MASK_BIG_END
+#define TARGET_ENDIAN_OPTION     "mbig-endian"
+#define TARGET_LINKER_EMULATION  "armelfb_fbsd"
+#else
+#define TARGET_ENDIAN_DEFAULT    0
+#define TARGET_ENDIAN_OPTION     "mlittle-endian"
+#define TARGET_LINKER_EMULATION  "armelf_fbsd"
+#endif
+
+#undef	SUBTARGET_EXTRA_LINK_SPEC
+#define SUBTARGET_EXTRA_LINK_SPEC " -m " TARGET_LINKER_EMULATION " -p"
+
+#undef  TARGET_OS_CPP_BUILTINS
+#define TARGET_OS_CPP_BUILTINS() 		\
+  do						\
+    {						\
+	FBSD_TARGET_OS_CPP_BUILTINS ();		\
+	TARGET_BPABI_CPP_BUILTINS ();		\
+    }						\
+  while (false)
+
+/* We default to a soft-float ABI so that binaries can run on all
+   target hardware.  */
+#undef TARGET_DEFAULT_FLOAT_ABI
+#ifdef TARGET_FREEBSD_ARM_HARD_FLOAT
+#define TARGET_DEFAULT_FLOAT_ABI ARM_FLOAT_ABI_HARD
+#else
+#define TARGET_DEFAULT_FLOAT_ABI ARM_FLOAT_ABI_SOFT
+#endif
+
+#undef ARM_DEFAULT_ABI
+
+/* AACPS_LINUX has access to kernel atomic ops while we don't.
+   But AACPS defaults to short_enums.  */
+#define ARM_DEFAULT_ABI ARM_ABI_AAPCS_LINUX
+
+#undef TARGET_DEFAULT
+#define TARGET_DEFAULT (MASK_INTERWORK | TARGET_ENDIAN_DEFAULT)
+
+/* We do not have any MULTILIB_OPTIONS specified, so there are no
+   MULTILIB_DEFAULTS.  */
+#undef  MULTILIB_DEFAULTS
+
+/*  Use the AAPCS type for wchar_t, override the one from config/freebsd.h.  */
+#undef  WCHAR_TYPE
+#define WCHAR_TYPE  "unsigned int"
+
+#undef  WCHAR_TYPE_SIZE
+#define WCHAR_TYPE_SIZE BITS_PER_WORD
+
+#if defined (TARGET_FREEBSD_ARMv6)
+#undef  SUBTARGET_CPU_DEFAULT
+#define SUBTARGET_CPU_DEFAULT TARGET_CPU_arm1176jzs
+#else
+#undef  SUBTARGET_CPU_DEFAULT
+#define SUBTARGET_CPU_DEFAULT   TARGET_CPU_arm9
+#endif
+
+#define SUBTARGET_OVERRIDE_OPTIONS		\
+do {						\
+    if (unaligned_access)			\
+	unaligned_access = 0;			\
+} while (0)
+
+#undef MAX_SYNC_LIBFUNC_SIZE
+#define MAX_SYNC_LIBFUNC_SIZE 4 /* UNITS_PER_WORD not defined yet.  */
+
+/* FreeBSD does its profiling differently to the Acorn compiler.  We
+   don't need a word following the mcount call; and to skip it
+   requires either an assembly stub or use of fomit-frame-pointer when
+   compiling the profiling functions.  Since we break Acorn CC
+   compatibility below a little more won't hurt.  */
+
+#undef ARM_FUNCTION_PROFILER
+#define ARM_FUNCTION_PROFILER(STREAM,LABELNO)		\
+{							\
+  asm_fprintf (STREAM, "\tmov\t%Rip, %Rlr\n");		\
+  asm_fprintf (STREAM, "\tbl\t__mcount%s\n",		\
+	       (TARGET_ARM && NEED_PLT_RELOC)		\
+	       ? "(PLT)" : "");				\
+}
+
+/* Clear the instruction cache from `BEG' to `END'.  This makes a
+   call to the ARM_SYNC_ICACHE architecture specific syscall.  */
+#define CLEAR_INSN_CACHE(BEG, END)					\
+do									\
+  {									\
+    extern int sysarch (int number, void *args);			\
+    struct								\
+      {									\
+	unsigned int addr;						\
+	int          len;						\
+      } s;								\
+    s.addr = (unsigned int) (BEG);					\
+    s.len = (END) - (BEG);						\
+    (void) sysarch (0, &s);						\
+  }									\
+while (0)
+
+/* This is how we tell the assembler that two symbols have the same value.  */
+#define ASM_OUTPUT_DEF(FILE, NAME1, NAME2) \
+  do					   \
+    {					   \
+     assemble_name (FILE, NAME1); 	   \
+     fputs (" = ", FILE);		   \
+     assemble_name (FILE, NAME2);	   \
+     fputc ('\n', FILE);		   \
+    }					   \
+  while (0)
+
+/* Add  .note.GNU-stack.  */
+#undef NEED_INDICATE_EXEC_STACK
+#define NEED_INDICATE_EXEC_STACK 1
+
+#define ARM_TARGET2_DWARF_FORMAT (DW_EH_PE_pcrel | DW_EH_PE_indirect)
+
Index: gcc/config.gcc
===================================================================
--- gcc/config.gcc	(revision 218129)
+++ gcc/config.gcc	(working copy)
@@ -259,7 +259,6 @@
    arm*-wince-pe*			\
  | arm*-*-ecos-elf			\
  | arm*-*-elf				\
- | arm*-*-freebsd*			\
  | arm*-*-linux*			\
  | arm*-*-uclinux*			\
  | i[34567]86-go32-*			\
@@ -1009,6 +1008,27 @@
 	extra_options="${extra_options} arm/vxworks.opt"
 	tmake_file="${tmake_file} arm/t-arm arm/t-vxworks"
 	;;
+arm*-*-freebsd*)                # ARM FreeBSD EABI
+	tm_file="dbxelf.h elfos.h ${fbsd_tm_file} arm/elf.h"
+	case $target in
+	arm*b-*-freebsd*)
+	    tm_defines="${tm_defines} TARGET_BIG_ENDIAN_DEFAULT=1"
+	    ;;
+	esac
+	tmake_file="${tmake_file} arm/t-arm arm/t-bpabi"
+	tm_file="${tm_file} arm/bpabi.h arm/freebsd.h arm/aout.h arm/arm.h"
+	case $target in
+	armv6*-*-freebsd*)
+	    tm_defines="${tm_defines} TARGET_FREEBSD_ARMv6=1"
+	    ;;
+	esac
+	case $target in
+	arm*hf-*-freebsd*)
+	    tm_defines="${tm_defines} TARGET_FREEBSD_ARM_HARD_FLOAT=1"
+	    ;;
+	esac
+	with_tls=${with_tls:-gnu}
+	;;
 arm*-*-netbsdelf*)
 	tm_file="dbxelf.h elfos.h netbsd.h netbsd-elf.h arm/elf.h arm/aout.h ${tm_file} arm/netbsd-elf.h"
 	extra_options="${extra_options} netbsd.opt netbsd-elf.opt"
Index: gcc/config.host
===================================================================
--- gcc/config.host	(revision 218129)
+++ gcc/config.host	(working copy)
@@ -99,7 +99,7 @@
 esac
 
 case ${host} in
-  arm*-*-linux*)
+  arm*-*-freebsd* | arm*-*-linux*)
     case ${target} in
       arm*-*-*)
 	host_extra_gcc_objs="driver-arm.o"
Index: libatomic/configure.tgt
===================================================================
--- libatomic/configure.tgt	(revision 218129)
+++ libatomic/configure.tgt	(working copy)
@@ -37,10 +37,15 @@
 
   arm*)
 	ARCH=arm
-	# ??? Detect when -march=armv7 is already enabled.
-	try_ifunc=yes
-	;;
-
+	case "${target}" in
+            arm*-*-freebsd*)
+                ;;
+            *)
+                # ??? Detect when -march=armv7 is already enabled.
+                try_ifunc=yes
+                ;;
+        esac
+        ;;
   sparc)
 	case " ${CC} ${CFLAGS} " in
 	  *" -m64 "*)
Index: libgcc/config/arm/freebsd-atomic.c
===================================================================
--- libgcc/config/arm/freebsd-atomic.c	(revision 0)
+++ libgcc/config/arm/freebsd-atomic.c	(working copy)
@@ -0,0 +1,224 @@
+/* FreeBSD specific atomic operations for ARM EABI.
+   Copyright (C) 2014 Free Software Foundation, Inc.
+
+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.
+
+Under Section 7 of GPL version 3, you are granted additional
+permissions described in the GCC Runtime Library Exception, version
+3.1, as published by the Free Software Foundation.
+
+You should have received a copy of the GNU General Public License and
+a copy of the GCC Runtime Library Exception along with this program;
+see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
+<http://www.gnu.org/licenses/>.  */
+
+#include <sys/types.h>
+
+#define HIDDEN __attribute__ ((visibility ("hidden")))
+
+#define ARM_VECTORS_HIGH 0xffff0000U
+#define ARM_TP_ADDRESS   (ARM_VECTORS_HIGH + 0x1000)
+#define ARM_RAS_START    (ARM_TP_ADDRESS + 4)
+
+void HIDDEN
+__sync_synchronize (void)
+{
+#if defined (__ARM_ARCH_6__) || defined (__ARM_ARCH_6J__)       \
+    || defined (__ARM_ARCH_6K__) || defined (__ARM_ARCH_6T2__)  \
+    || defined (__ARM_ARCH_6Z__) || defined (__ARM_ARCH_6ZK__)  \
+    || defined (__ARM_ARCH_7__) || defined (__ARM_ARCH_7A__)
+#if defined (__ARM_ARCH_7__) || defined (__ARM_ARCH_7A__)
+    __asm __volatile ("dmb" : : : "memory");
+#else
+    __asm __volatile ("mcr p15, 0, r0, c7, c10, 5" : : : "memory");
+#endif
+#else
+    __asm __volatile ("nop" : : : "memory");
+#endif
+}
+
+#if defined (__ARM_ARCH_6__) || defined (__ARM_ARCH_6J__)        \
+    || defined (__ARM_ARCH_6K__) || defined (__ARM_ARCH_6T2__)   \
+    || defined (__ARM_ARCH_6Z__) || defined (__ARM_ARCH_6ZK__)   \
+    || defined (__ARM_ARCH_7__) || defined (__ARM_ARCH_7A__)
+
+/* These systems should be supported by the compiler.  */
+
+#else /* __ARM_ARCH_5__  */
+
+#define	SYNC_LOCK_TEST_AND_SET_N(N, TYPE, LDR, STR)			\
+TYPE HIDDEN    								\
+__sync_lock_test_and_set_##N (TYPE *mem, TYPE val)			\
+{									\
+        unsigned int old, temp, ras_start;                              \
+                                                                        \
+        ras_start = ARM_RAS_START;					\
+        __asm volatile (						\
+                /* Set up Restartable Atomic Sequence.  */		\
+                "1:"							\
+                "\tadr   %2, 1b\n"					\
+                "\tstr   %2, [%5]\n"					\
+                "\tadr   %2, 2f\n"					\
+                "\tstr   %2, [%5, #4]\n"				\
+                                                                        \
+                "\t"LDR" %0, %4\n"	/* Load old value.  */		\
+                "\t"STR" %3, %1\n"	/* Store new value.  */		\
+                                                                        \
+                /* Tear down Restartable Atomic Sequence.  */		\
+                "2:"							\
+                "\tmov   %2, #0x00000000\n"				\
+                "\tstr   %2, [%5]\n"					\
+                "\tmov   %2, #0xffffffff\n"				\
+                "\tstr   %2, [%5, #4]\n"				\
+                : "=&r" (old), "=m" (*mem), "=&r" (temp)		\
+                : "r" (val), "m" (*mem), "r" (ras_start));		\
+        return (old);							\
+}
+
+#define SYNC_LOCK_RELEASE_N(N, TYPE)					\
+void HIDDEN								\
+__sync_lock_release_##N (TYPE *ptr)     				\
+{					       				\
+    /* All writes before this point must be seen before we release	\
+       the lock itself.  */						\
+    __sync_synchronize ();     						\
+    *ptr = 0;								\
+}
+
+#define	SYNC_VAL_CAS_N(N, TYPE, LDR, STREQ)             		\
+TYPE HIDDEN    								\
+__sync_val_compare_and_swap_##N (TYPE *mem, TYPE expected,		\
+    TYPE desired)							\
+{									\
+        unsigned int old, temp, ras_start;                              \
+                                                                        \
+        ras_start = ARM_RAS_START;					\
+        __asm volatile (						\
+                /* Set up Restartable Atomic Sequence.  */		\
+                "1:"							\
+                "\tadr   %2, 1b\n"					\
+                "\tstr   %2, [%6]\n"					\
+                "\tadr   %2, 2f\n"					\
+                "\tstr   %2, [%6, #4]\n"				\
+                                                                        \
+                "\t"LDR" %0, %5\n"     /* Load old value.  */		\
+                "\tcmp   %0, %3\n"     /* Compare to expected value.  */\
+                "\t"STREQ" %4, %1\n"   /* Store new value.  */		\
+                                                                        \
+                /* Tear down Restartable Atomic Sequence.  */		\
+                "2:"							\
+                "\tmov   %2, #0x00000000\n"				\
+                "\tstr   %2, [%6]\n"					\
+                "\tmov   %2, #0xffffffff\n"				\
+                "\tstr   %2, [%6, #4]\n"				\
+                : "=&r" (old), "=m" (*mem), "=&r" (temp)		\
+                : "r" (expected), "r" (desired), "m" (*mem),		\
+                  "r" (ras_start));					\
+        return (old);							\
+}
+
+typedef unsigned char bool;
+
+#define SYNC_BOOL_CAS_N(N, TYPE)                                        \
+bool HIDDEN								\
+__sync_bool_compare_and_swap_##N (TYPE *ptr, TYPE oldval,		\
+                                  TYPE newval)                          \
+{									\
+    TYPE actual_oldval							\
+      = __sync_val_compare_and_swap_##N (ptr, oldval, newval);          \
+    return (oldval == actual_oldval);					\
+}
+
+#define	SYNC_FETCH_AND_OP_N(N, TYPE, LDR, STR, NAME, OP)		\
+TYPE HIDDEN								\
+__sync_fetch_and_##NAME##_##N (TYPE *mem, TYPE val)	       		\
+{									\
+        unsigned int old, temp, ras_start;                              \
+                                                                        \
+        ras_start = ARM_RAS_START;					\
+        __asm volatile (						\
+                /* Set up Restartable Atomic Sequence.  */		\
+                "1:"							\
+                "\tadr   %2, 1b\n"					\
+                "\tstr   %2, [%5]\n"					\
+                "\tadr   %2, 2f\n"					\
+                "\tstr   %2, [%5, #4]\n"				\
+                                                                        \
+                "\t"LDR" %0, %4\n"	/* Load old value.  */		\
+                "\t"OP"  %2, %0, %3\n"	/* Calculate new value.  */	\
+                "\t"STR" %2, %1\n"	/* Store new value.  */		\
+                                                                        \
+                /* Tear down Restartable Atomic Sequence.  */		\
+                "2:"							\
+                "\tmov   %2, #0x00000000\n"				\
+                "\tstr   %2, [%5]\n"					\
+                "\tmov   %2, #0xffffffff\n"				\
+                "\tstr   %2, [%5, #4]\n"				\
+                : "=&r" (old), "=m" (*mem), "=&r" (temp)		\
+                : "r" (val), "m" (*mem), "r" (ras_start));		\
+        return (old);							\
+}
+
+#define	SYNC_OP_AND_FETCH_N(N, TYPE, LDR, STR, NAME, OP)		\
+TYPE HIDDEN   								\
+__sync_##NAME##_and_fetch_##N (TYPE *mem, TYPE val)	       		\
+{									\
+        unsigned int old, temp, ras_start;                              \
+                                                                        \
+        ras_start = ARM_RAS_START;					\
+        __asm volatile (						\
+                /* Set up Restartable Atomic Sequence.  */		\
+                "1:"							\
+                "\tadr   %2, 1b\n"					\
+                "\tstr   %2, [%5]\n"					\
+                "\tadr   %2, 2f\n"					\
+                "\tstr   %2, [%5, #4]\n"				\
+                                                                        \
+                "\t"LDR" %0, %4\n"	/* Load old value.  */		\
+                "\t"OP"  %2, %0, %3\n"	/* Calculate new value.  */	\
+                "\t"STR" %2, %1\n"	/* Store new value.  */		\
+                                                                        \
+                /* Tear down Restartable Atomic Sequence.  */		\
+                "2:"							\
+                "\tmov   %2, #0x00000000\n"				\
+                "\tstr   %2, [%5]\n"					\
+                "\tmov   %2, #0xffffffff\n"				\
+                "\tstr   %2, [%5, #4]\n"				\
+                : "=&r" (old), "=m" (*mem), "=&r" (temp)		\
+                : "r" (val), "m" (*mem), "r" (ras_start));		\
+        return (old);							\
+}
+
+#define	EMIT_ALL_OPS_N(N, TYPE, LDR, STR, STREQ)			\
+SYNC_LOCK_TEST_AND_SET_N (N, TYPE, LDR, STR)				\
+SYNC_LOCK_RELEASE_N (N, TYPE)                                           \
+SYNC_VAL_CAS_N (N, TYPE, LDR, STREQ)	                		\
+SYNC_BOOL_CAS_N (N, TYPE)                                               \
+SYNC_FETCH_AND_OP_N (N, TYPE, LDR, STR, add, "add")		        \
+SYNC_FETCH_AND_OP_N (N, TYPE, LDR, STR, and, "and")		        \
+SYNC_FETCH_AND_OP_N (N, TYPE, LDR, STR, or, "orr")		        \
+SYNC_FETCH_AND_OP_N (N, TYPE, LDR, STR, sub, "sub")		        \
+SYNC_FETCH_AND_OP_N (N, TYPE, LDR, STR, xor, "eor")                     \
+SYNC_OP_AND_FETCH_N (N, TYPE, LDR, STR, add, "add")		        \
+SYNC_OP_AND_FETCH_N (N, TYPE, LDR, STR, and, "and")		        \
+SYNC_OP_AND_FETCH_N (N, TYPE, LDR, STR, or, "orr")		        \
+SYNC_OP_AND_FETCH_N (N, TYPE, LDR, STR, sub, "sub")		        \
+SYNC_OP_AND_FETCH_N (N, TYPE, LDR, STR, xor, "eor")
+
+
+
+EMIT_ALL_OPS_N (1, unsigned char, "ldrb", "strb", "streqb")
+EMIT_ALL_OPS_N (2, unsigned short, "ldrh", "strh", "streqh")
+EMIT_ALL_OPS_N (4, unsigned int, "ldr", "str", "streq")
+
+#endif
Index: libgcc/config/arm/t-freebsd
===================================================================
--- libgcc/config/arm/t-freebsd	(revision 0)
+++ libgcc/config/arm/t-freebsd	(working copy)
@@ -0,0 +1,9 @@
+# Just for these, we omit the frame pointer since it makes such a big
+# difference.  It is then pointless adding debugging.
+HOST_LIBGCC2_CFLAGS += -fomit-frame-pointer
+
+LIB2ADD_ST += $(srcdir)/config/arm/freebsd-atomic.c
+
+# Use a version of div0 which raises SIGFPE.
+LIB1ASMFUNCS := $(filter-out _dvmd_tls,$(LIB1ASMFUNCS)) _dvmd_lnx
+
Index: libgcc/config/arm/unwind-arm.h
===================================================================
--- libgcc/config/arm/unwind-arm.h	(revision 218129)
+++ libgcc/config/arm/unwind-arm.h	(working copy)
@@ -48,7 +48,8 @@
       if (!tmp)
 	return 0;
 
-#if (defined(linux) && !defined(__uClinux__)) || defined(__NetBSD__)
+#if (defined(linux) && !defined(__uClinux__)) || defined(__NetBSD__) \
+    || defined(__FreeBSD__)
       /* Pc-relative indirect.  */
 #define _GLIBCXX_OVERRIDE_TTYPE_ENCODING (DW_EH_PE_pcrel | DW_EH_PE_indirect)
       tmp += ptr;
Index: libgcc/config.host
===================================================================
--- libgcc/config.host	(revision 218129)
+++ libgcc/config.host	(working copy)
@@ -366,6 +366,13 @@
 	tmake_file="$tmake_file arm/t-arm arm/t-elf t-softfp-sfdf t-softfp-excl arm/t-softfp t-softfp"
 	extra_parts="$extra_parts crti.o crtn.o"
 	;;
+arm*-*-freebsd*)                # ARM FreeBSD EABI
+	tmake_file="${tmake_file} arm/t-arm t-fixedpoint-gnu-prefix arm/t-elf"
+	tmake_file="${tmake_file} arm/t-bpabi arm/t-freebsd t-slibgcc-libgcc"
+	tm_file="${tm_file} arm/bpabi-lib.h"
+	unwind_header=config/arm/unwind-arm.h
+	tmake_file="${tmake_file} t-softfp-sfdf t-softfp-excl arm/t-softfp t-softfp"
+	;;
 arm*-*-netbsdelf*)
 	tmake_file="$tmake_file arm/t-arm arm/t-netbsd t-slibgcc-gld-nover"
 	;;
Index: libstdc++-v3/configure.host
===================================================================
--- libstdc++-v3/configure.host	(revision 218129)
+++ libstdc++-v3/configure.host	(working copy)
@@ -366,6 +366,9 @@
 	;;
     esac
     ;;
+  arm*-*-freebsd*)
+     port_specific_symbol_files="\$(srcdir)/../config/os/gnu-linux/arm-eabi-extra.ver"
+     ;;
   powerpc*-*-darwin*)
     port_specific_symbol_files="\$(srcdir)/../config/os/bsd/darwin/ppc-extra.ver"
     ;;

Re: [PATCH][ARM] FreeBSD arm support, EABI, v2

[Andreas Tobler]

Ping?!

Thanks,
Andreas

On 27.11.14 21:56, Andreas Tobler wrote:
> Hi all,
>
> this is the second attempt.
>
> I reworked the issues Richard mentioned in the previous review.
> I also found one issue which will break build/bootstrap if I pass
> --enable-gnu-indirect-function, also fixed.
>
> One thing which came up is the way we generate code for the
> armv6*-*-freebsd* triplet versus the arm-*-freebsd* triplet.
>
> I think the thing which confuses is the fact that we have only two fixed
> triplets where we build a complete OS with. Means the whole OS is built
> with the same optimization, not only the kernel or one binary.
>
> For the armv6* we want to benefit from the cpu's functionality by
> default. We build all __ARM_ARCH >= 6 with TARGET_CPU_arm1176jzs,
> on the other hand all __ARM_ARCH <=5 will be built with TARGET_CPU_arm9.
>
> Now who becomes arm-*-freebsd* and who becomes armv6*-*-freebsd*?
>
> As tried above, we only know two triplets, so __ARM_ARCH >= 6 becomes
> armv6*-*-freebsd* and __ARM_ARCH <=5 becomes arm-*-freebsd*.
>
> armv8 is not yet in the portfolio and it will become something
> different, either arm64 or aarch64, I do not know.
>
> I'd like to keep this since our system compilers, clang and gcc-4.2.1
> behave the same.
> If we have to change here, we would confuse people quite a lot.
>
> The whole thing is FreeBSD specific and does not touch others.
>
> As usual, bootstrapped, cross compiled, tested.
>
> Ok for trunk?
>
> TIA,
> Andreas
>
> toplevel:
>
> 	* configure.ac: Don't add ${libgcj} for arm*-*-freebsd*.
> 	* configure: Regenerate.
> gcc:
> 	* config.gcc (arm*-*-freebsd*): New configuration.
> 	* config/arm/freebsd.h: New file.
> 	* config.host: Add extra components for arm*-*-freebsd*.
> 	* config/arm/arm.h: Introduce MAX_SYNC_LIBFUNC_SIZE.
> 	* config/arm/arm.c (arm_init_libfuncs): Use MAX_SYNC_LIBFUNC_SIZE.
>
> libgcc:
>
> 	* config.host (arm*-*-freebsd*): Add new configuration for
> 	arm*-*-freebsd*.
> 	* config/arm/freebsd-atomic.c: New file.
> 	* config/arm/t-freebsd: Likewise.
> 	* config/arm/unwind-arm.h: Add __FreeBSD__ to the list of
> 	'PC-relative indirect' OS's.
>
> libatomic:
>
> 	* configure.tgt: Exclude arm*-*-freebsd* from try_ifunc.
>
> libstdc++-v3:
>
> 	* configure.host: Add arm*-*-freebsd* port_specific_symbol_files.
>

Re: [PATCH][ARM] FreeBSD arm support, EABI, v2

[Andreas Tobler]

On 08.01.15 17:22, Richard Earnshaw wrote:
> On 27/11/14 20:56, Andreas Tobler wrote:
>> Hi all,
>>
>> this is the second attempt.
>>
>> I reworked the issues Richard mentioned in the previous review.
>> I also found one issue which will break build/bootstrap if I pass
>> --enable-gnu-indirect-function, also fixed.
>>
>> One thing which came up is the way we generate code for the
>> armv6*-*-freebsd* triplet versus the arm-*-freebsd* triplet.
>>
>> I think the thing which confuses is the fact that we have only two fixed
>> triplets where we build a complete OS with. Means the whole OS is built
>> with the same optimization, not only the kernel or one binary.
>>
>> For the armv6* we want to benefit from the cpu's functionality by
>> default. We build all __ARM_ARCH >= 6 with TARGET_CPU_arm1176jzs,
>> on the other hand all __ARM_ARCH <=5 will be built with TARGET_CPU_arm9.
>>
>> Now who becomes arm-*-freebsd* and who becomes armv6*-*-freebsd*?
>>
>> As tried above, we only know two triplets, so __ARM_ARCH >= 6 becomes
>> armv6*-*-freebsd* and __ARM_ARCH <=5 becomes arm-*-freebsd*.
>>
>> armv8 is not yet in the portfolio and it will become something
>> different, either arm64 or aarch64, I do not know.
>>
>> I'd like to keep this since our system compilers, clang and gcc-4.2.1
>> behave the same.
>> If we have to change here, we would confuse people quite a lot.
>>
>> The whole thing is FreeBSD specific and does not touch others.
>>
>> As usual, bootstrapped, cross compiled, tested.
>>
>> Ok for trunk?
>>
>> TIA,
>> Andreas
>>
>> toplevel:
>>
>> 	* configure.ac: Don't add ${libgcj} for arm*-*-freebsd*.
>> 	* configure: Regenerate.
>> gcc:
>> 	* config.gcc (arm*-*-freebsd*): New configuration.
>> 	* config/arm/freebsd.h: New file.
>> 	* config.host: Add extra components for arm*-*-freebsd*.
>> 	* config/arm/arm.h: Introduce MAX_SYNC_LIBFUNC_SIZE.
>> 	* config/arm/arm.c (arm_init_libfuncs): Use MAX_SYNC_LIBFUNC_SIZE.
>>
>> libgcc:
>>
>> 	* config.host (arm*-*-freebsd*): Add new configuration for
>> 	arm*-*-freebsd*.
>> 	* config/arm/freebsd-atomic.c: New file.
>> 	* config/arm/t-freebsd: Likewise.
>> 	* config/arm/unwind-arm.h: Add __FreeBSD__ to the list of
>> 	'PC-relative indirect' OS's.
>>
>> libatomic:
>>
>> 	* configure.tgt: Exclude arm*-*-freebsd* from try_ifunc.
>>
>> libstdc++-v3:
>>
>> 	* configure.host: Add arm*-*-freebsd* port_specific_symbol_files.
>>
>>
>


> Sorry for the delay in responding, I've been OoO quite a bit over the
> last month or so and not had much time for patch reviewing.

Thank you for the review. Also being very busy, I know what you're 
talking about.

> This is OK to go in.  Most of the patch is now in OS specific files and
> thus not really relevant to the CPU maintenance.
>
> One item I do think you may need to think about is the definition of
> SUBTARGET_EXTRA_LINK_SPEC.  This picks the linker emulation setting.
> But it's a build-time choice and seems to ignore a run-time -mbig-endian
> or -mlittle-endian choice on the command line.  Is this selection even
> necessary?  I don't remember ever having to do it on other OS ports -
> the linker just picks the correct emulation based on the first object
> file in the link list.

I don't think this is necessary. Certainly not yet. Might become a 
future issue.

I also got your second mail.

If you're ok with, I'll commit the first hunk and send the undwind 
update as separate patch again for review?

Thanks,
Andreas

Re: [PATCH][ARM] FreeBSD arm support, EABI, v2

[Richard Earnshaw]

On 08/01/15 20:28, Andreas Tobler wrote:
> On 08.01.15 17:22, Richard Earnshaw wrote:
>> On 27/11/14 20:56, Andreas Tobler wrote:
>>> Hi all,
>>>
>>> this is the second attempt.
>>>
>>> I reworked the issues Richard mentioned in the previous review.
>>> I also found one issue which will break build/bootstrap if I pass
>>> --enable-gnu-indirect-function, also fixed.
>>>
>>> One thing which came up is the way we generate code for the
>>> armv6*-*-freebsd* triplet versus the arm-*-freebsd* triplet.
>>>
>>> I think the thing which confuses is the fact that we have only two fixed
>>> triplets where we build a complete OS with. Means the whole OS is built
>>> with the same optimization, not only the kernel or one binary.
>>>
>>> For the armv6* we want to benefit from the cpu's functionality by
>>> default. We build all __ARM_ARCH >= 6 with TARGET_CPU_arm1176jzs,
>>> on the other hand all __ARM_ARCH <=5 will be built with TARGET_CPU_arm9.
>>>
>>> Now who becomes arm-*-freebsd* and who becomes armv6*-*-freebsd*?
>>>
>>> As tried above, we only know two triplets, so __ARM_ARCH >= 6 becomes
>>> armv6*-*-freebsd* and __ARM_ARCH <=5 becomes arm-*-freebsd*.
>>>
>>> armv8 is not yet in the portfolio and it will become something
>>> different, either arm64 or aarch64, I do not know.
>>>
>>> I'd like to keep this since our system compilers, clang and gcc-4.2.1
>>> behave the same.
>>> If we have to change here, we would confuse people quite a lot.
>>>
>>> The whole thing is FreeBSD specific and does not touch others.
>>>
>>> As usual, bootstrapped, cross compiled, tested.
>>>
>>> Ok for trunk?
>>>
>>> TIA,
>>> Andreas
>>>
>>> toplevel:
>>>
>>> 	* configure.ac: Don't add ${libgcj} for arm*-*-freebsd*.
>>> 	* configure: Regenerate.
>>> gcc:
>>> 	* config.gcc (arm*-*-freebsd*): New configuration.
>>> 	* config/arm/freebsd.h: New file.
>>> 	* config.host: Add extra components for arm*-*-freebsd*.
>>> 	* config/arm/arm.h: Introduce MAX_SYNC_LIBFUNC_SIZE.
>>> 	* config/arm/arm.c (arm_init_libfuncs): Use MAX_SYNC_LIBFUNC_SIZE.
>>>
>>> libgcc:
>>>
>>> 	* config.host (arm*-*-freebsd*): Add new configuration for
>>> 	arm*-*-freebsd*.
>>> 	* config/arm/freebsd-atomic.c: New file.
>>> 	* config/arm/t-freebsd: Likewise.
>>> 	* config/arm/unwind-arm.h: Add __FreeBSD__ to the list of
>>> 	'PC-relative indirect' OS's.
>>>
>>> libatomic:
>>>
>>> 	* configure.tgt: Exclude arm*-*-freebsd* from try_ifunc.
>>>
>>> libstdc++-v3:
>>>
>>> 	* configure.host: Add arm*-*-freebsd* port_specific_symbol_files.
>>>
>>>
>>
> 
> 
>> Sorry for the delay in responding, I've been OoO quite a bit over the
>> last month or so and not had much time for patch reviewing.
> 
> Thank you for the review. Also being very busy, I know what you're 
> talking about.
> 
>> This is OK to go in.  Most of the patch is now in OS specific files and
>> thus not really relevant to the CPU maintenance.
>>
>> One item I do think you may need to think about is the definition of
>> SUBTARGET_EXTRA_LINK_SPEC.  This picks the linker emulation setting.
>> But it's a build-time choice and seems to ignore a run-time -mbig-endian
>> or -mlittle-endian choice on the command line.  Is this selection even
>> necessary?  I don't remember ever having to do it on other OS ports -
>> the linker just picks the correct emulation based on the first object
>> file in the link list.
> 
> I don't think this is necessary. Certainly not yet. Might become a 
> future issue.
> 
> I also got your second mail.
> 
> If you're ok with, I'll commit the first hunk and send the undwind 
> update as separate patch again for review?
> 

Yes, that's fine.

R.

> Thanks,
> Andreas
> 
> 
> 

Re: [PATCH][ARM] FreeBSD arm support, EABI, v2

[Andreas Tobler]

On 09.01.15 10:27, Richard Earnshaw wrote:
> On 08/01/15 20:28, Andreas Tobler wrote:
>> On 08.01.15 17:22, Richard Earnshaw wrote:
>>> On 27/11/14 20:56, Andreas Tobler wrote:
>>>> Hi all,
>>>>
>>>> this is the second attempt.
>>>>
>>>> I reworked the issues Richard mentioned in the previous review.
>>>> I also found one issue which will break build/bootstrap if I pass
>>>> --enable-gnu-indirect-function, also fixed.
>>>>
>>>> One thing which came up is the way we generate code for the
>>>> armv6*-*-freebsd* triplet versus the arm-*-freebsd* triplet.
>>>>
>>>> I think the thing which confuses is the fact that we have only two fixed
>>>> triplets where we build a complete OS with. Means the whole OS is built
>>>> with the same optimization, not only the kernel or one binary.
>>>>
>>>> For the armv6* we want to benefit from the cpu's functionality by
>>>> default. We build all __ARM_ARCH >= 6 with TARGET_CPU_arm1176jzs,
>>>> on the other hand all __ARM_ARCH <=5 will be built with TARGET_CPU_arm9.
>>>>
>>>> Now who becomes arm-*-freebsd* and who becomes armv6*-*-freebsd*?
>>>>
>>>> As tried above, we only know two triplets, so __ARM_ARCH >= 6 becomes
>>>> armv6*-*-freebsd* and __ARM_ARCH <=5 becomes arm-*-freebsd*.
>>>>
>>>> armv8 is not yet in the portfolio and it will become something
>>>> different, either arm64 or aarch64, I do not know.
>>>>
>>>> I'd like to keep this since our system compilers, clang and gcc-4.2.1
>>>> behave the same.
>>>> If we have to change here, we would confuse people quite a lot.
>>>>
>>>> The whole thing is FreeBSD specific and does not touch others.
>>>>
>>>> As usual, bootstrapped, cross compiled, tested.
>>>>
>>>> Ok for trunk?
>>>>
>>>> TIA,
>>>> Andreas
>>>>
>>>> toplevel:
>>>>
>>>> 	* configure.ac: Don't add ${libgcj} for arm*-*-freebsd*.
>>>> 	* configure: Regenerate.
>>>> gcc:
>>>> 	* config.gcc (arm*-*-freebsd*): New configuration.
>>>> 	* config/arm/freebsd.h: New file.
>>>> 	* config.host: Add extra components for arm*-*-freebsd*.
>>>> 	* config/arm/arm.h: Introduce MAX_SYNC_LIBFUNC_SIZE.
>>>> 	* config/arm/arm.c (arm_init_libfuncs): Use MAX_SYNC_LIBFUNC_SIZE.
>>>>
>>>> libgcc:
>>>>
>>>> 	* config.host (arm*-*-freebsd*): Add new configuration for
>>>> 	arm*-*-freebsd*.
>>>> 	* config/arm/freebsd-atomic.c: New file.
>>>> 	* config/arm/t-freebsd: Likewise.
>>>> 	* config/arm/unwind-arm.h: Add __FreeBSD__ to the list of
>>>> 	'PC-relative indirect' OS's.
>>>>
>>>> libatomic:
>>>>
>>>> 	* configure.tgt: Exclude arm*-*-freebsd* from try_ifunc.
>>>>
>>>> libstdc++-v3:
>>>>
>>>> 	* configure.host: Add arm*-*-freebsd* port_specific_symbol_files.
>>>>
>>>>
>>>
>>
>>
>>> Sorry for the delay in responding, I've been OoO quite a bit over the
>>> last month or so and not had much time for patch reviewing.
>>
>> Thank you for the review. Also being very busy, I know what you're
>> talking about.
>>
>>> This is OK to go in.  Most of the patch is now in OS specific files and
>>> thus not really relevant to the CPU maintenance.
>>>
>>> One item I do think you may need to think about is the definition of
>>> SUBTARGET_EXTRA_LINK_SPEC.  This picks the linker emulation setting.
>>> But it's a build-time choice and seems to ignore a run-time -mbig-endian
>>> or -mlittle-endian choice on the command line.  Is this selection even
>>> necessary?  I don't remember ever having to do it on other OS ports -
>>> the linker just picks the correct emulation based on the first object
>>> file in the link list.
>>
>> I don't think this is necessary. Certainly not yet. Might become a
>> future issue.
>>
>> I also got your second mail.
>>
>> If you're ok with, I'll commit the first hunk and send the undwind
>> update as separate patch again for review?
>>
>
> Yes, that's fine.
>

Committed. (r219388) Thanks.

Unfortunately I commited too much but I reverted the two files (r219391) 
which are not review yet.

Sorry for the troubles.

Andreas

Re: [PATCH][ARM] FreeBSD arm support, EABI, v2

[Gerald Pfeifer]

On Friday 2015-01-09 15:24, Andreas Tobler wrote:
> Committed. (r219388) Thanks.

Can this also go into the GCC 4.9 branch?  That is (and will be
for a while) the latest release stream and the patch is pretty
much isolated to FreeBSD.

Gerald

Re: [PATCH][ARM] FreeBSD arm support, EABI, v2

[Kyrill Tkachov]

On 17/01/15 15:18, Gerald Pfeifer wrote:
> On Friday 2015-01-09 15:24, Andreas Tobler wrote:
>> Committed. (r219388) Thanks.
> Can this also go into the GCC 4.9 branch?  That is (and will be
> for a while) the latest release stream and the patch is pretty
> much isolated to FreeBSD.

Also a note for the release note perhaps?

Kyrill

>
> Gerald
>