* gcc alpha/NetBSD 1.3 target files
@ 1998-04-30 18:55 Paul H. Anderson
1998-05-01 9:55 ` Richard Henderson
0 siblings, 1 reply; 4+ messages in thread
From: Paul H. Anderson @ 1998-04-30 18:55 UTC (permalink / raw)
To: egcs; +Cc: pha
egcs support folks,
The following patches are a initial effort to get egcs working on the
NetBSD/Alpha platform. The code comes entirely from the Linux work,
usually with simple substitutions and a few slightly larger modifications.
I've mailed a copyright assignment form and an employer release form to
Richard Stallman today. It should be at his office early next week
(certainly by Tuesday).
The changes in this message allow a user to build egcs-1.0.2 and create a
working compiler, with the following evident problems:
/usr/local/lib/gcc-lib/alpha-unknown-netbsd1.3/egcs-2.90.27/include/limits.h:31:
SCHAR_{MAX,MIN}, UCHAR_{}, CHAR_{}, etc, are redefined - I don't know why
egcs put these limits back in... must be a config problem here.
gdb has problems single stepping in c++ programs. The toolchain compiler
did this also, though, so I'm not sure that I made this any worse.
c++ static initializer seems to have problems (w/String object in libg++)
I have not tested C++ exceptions at all
egcs-1.0.2 won't build a NetBSD kernel, because of an unresolved
global, _memcpy. This is probably simple to fix.
Please let me know of any problems, questions, requests for changes, etc,
that I may be able to help with. I have some modest interest in further
work on the compiler, but it is more likely that some folks on the
netbsd-alpha mail list will be doing additional bug fixes, etc.
Paul Anderson
pha@pdq.com
ChangeLog follows:
-----------------------
Wed Apr 29 10:56:12 1998 Paul Anderson <pha@pdq.com>
* create config files needed for NetBSD/Alpha target
* config/alpha/alpha.c: don't include <stamp.h> if __NetBSD__
* config/alpha/elf-netbsd.h: copy and modify the Linux version
* config/alpha/netbsd.h: copy and modify the Linux version
* config/alpha/t-netbsd: copy and modify the Linux version
* config/alpha/xm-netbsd.h: copy and modify the Linux version
* configure: properly decode NetBSD/Alpha target setup
diffs relative to egcs-1.0.2 follow:
-----------------------
diff -Nrc3p egcs-1.0.2.cygnus/gcc/config/alpha/alpha.c egcs-1.0.2/gcc/config/alpha/alpha.c
*** egcs-1.0.2.cygnus/gcc/config/alpha/alpha.c Thu Nov 20 18:07:26 1997
--- egcs-1.0.2/gcc/config/alpha/alpha.c Fri Apr 17 10:21:58 1998
*************** direct_return ()
*** 1927,1933 ****
/* Write a version stamp. Don't write anything if we are running as a
cross-compiler. Otherwise, use the versions in /usr/include/stamp.h. */
! #if !defined(CROSS_COMPILE) && !defined(_WIN32) && !defined(__linux__) && !defined(VMS)
#include <stamp.h>
#endif
--- 1927,1933 ----
/* Write a version stamp. Don't write anything if we are running as a
cross-compiler. Otherwise, use the versions in /usr/include/stamp.h. */
! #if !defined(CROSS_COMPILE) && !defined(_WIN32) && !defined(__linux__) && !defined(VMS) && !defined(__NetBSD__)
#include <stamp.h>
#endif
diff -Nrc3p egcs-1.0.2.cygnus/gcc/config/alpha/elf-netbsd.h egcs-1.0.2/gcc/config/alpha/elf-netbsd.h
*** egcs-1.0.2.cygnus/gcc/config/alpha/elf-netbsd.h Wed Dec 31 16:00:00 1969
--- egcs-1.0.2/gcc/config/alpha/elf-netbsd.h Fri Apr 17 10:27:04 1998
***************
*** 0 ****
--- 1,521 ----
+ /* Definitions of target machine for GNU compiler, for DEC Alpha w/ELF.
+ Copyright (C) 1996, 1997 Free Software Foundation, Inc.
+ Contributed by Richard Henderson (rth@tamu.edu).
+
+ This file is part of GNU CC.
+
+ GNU CC 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 2, or (at your option)
+ any later version.
+
+ GNU CC 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 GNU CC; see the file COPYING. If not, write to
+ the Free Software Foundation, 59 Temple Place - Suite 330,
+ Boston, MA 02111-1307, USA. */
+
+ /* This is used on Alpha platforms that use the ELF format.
+ Currently only Linux and NetBSD uses this. */
+
+ #undef TARGET_VERSION
+ #define TARGET_VERSION fprintf (stderr, " (Alpha NetBSD/ELF)");
+
+ #undef OBJECT_FORMAT_COFF
+ #undef EXTENDED_COFF
+ #define OBJECT_FORMAT_ELF
+
+ #define SDB_DEBUGGING_INFO
+
+ #undef ASM_FINAL_SPEC
+
+ #undef CPP_PREDEFINES
+ #define CPP_PREDEFINES "\
+ -D__alpha -D__alpha__ -D__NetBSD__ -D_LONGLONG -Dunix \
+ -Asystem(unix) -Asystem(NetBSD) -Acpu(alpha) -Amachine(alpha) -D__ELF__"
+
+ #undef LINK_SPEC
+ #ifdef USE_GNULIBC_1
+ #define LINK_SPEC "-m elf64alpha -G 8 %{O*:-O3} %{!O*:-O1} \
+ %{shared:-shared} \
+ %{!shared: \
+ %{!static: \
+ %{rdynamic:-export-dynamic} \
+ %{!dynamic-linker:-dynamic-linker /usr/libexec/ld.elf_so}} \
+ %{static:-static}}"
+ #else
+ #define LINK_SPEC "-m elf64alpha -G 8 %{O*:-O3} %{!O*:-O1} \
+ %{shared:-shared} \
+ %{!shared: \
+ %{!static: \
+ %{rdynamic:-export-dynamic} \
+ %{!dynamic-linker:-dynamic-linker /usr/libexec/ld.elf_so}} \
+ %{static:-static}}"
+ #endif
+
+ #ifndef USE_GNULIBC_1
+ #undef DEFAULT_VTABLE_THUNKS
+ #define DEFAULT_VTABLE_THUNKS 1
+ #endif
+
+ #ifndef USE_GNULIBC_1
+ #undef LIB_SPEC
+ #define LIB_SPEC \
+ "%{shared: -lc} \
+ %{!shared: %{mieee-fp:-lieee} %{pthread:-lpthread} \
+ %{profile:-lc_p} %{!profile: -lc}}"
+ #endif
+
+ /* Output at beginning of assembler file. */
+
+ #undef ASM_FILE_START
+ #define ASM_FILE_START(FILE) \
+ { \
+ alpha_write_verstamp (FILE); \
+ output_file_directive (FILE, main_input_filename); \
+ fprintf (FILE, "\t.version\t\"01.01\"\n"); \
+ fprintf (FILE, "\t.set noat\n"); \
+ }
+
+ #define ASM_OUTPUT_SOURCE_LINE(STREAM, LINE) \
+ alpha_output_lineno (STREAM, LINE)
+ extern void alpha_output_lineno ();
+
+ extern void output_file_directive ();
+
+ /* Attach a special .ident directive to the end of the file to identify
+ the version of GCC which compiled this code. The format of the
+ .ident string is patterned after the ones produced by native svr4
+ C compilers. */
+
+ #define IDENT_ASM_OP ".ident"
+
+ #ifdef IDENTIFY_WITH_IDENT
+ #define ASM_IDENTIFY_GCC(FILE) /* nothing */
+ #define ASM_IDENTIFY_LANGUAGE(FILE) \
+ fprintf(FILE, "\t%s \"GCC (%s) %s\"\n", IDENT_ASM_OP, \
+ lang_identify(), version_string)
+ #else
+ #define ASM_FILE_END(FILE) \
+ do { \
+ fprintf ((FILE), "\t%s\t\"GCC: (GNU) %s\"\n", \
+ IDENT_ASM_OP, version_string); \
+ } while (0)
+ #endif
+
+ /* Allow #sccs in preprocessor. */
+
+ #define SCCS_DIRECTIVE
+
+ /* Output #ident as a .ident. */
+
+ #define ASM_OUTPUT_IDENT(FILE, NAME) \
+ fprintf (FILE, "\t%s\t\"%s\"\n", IDENT_ASM_OP, NAME);
+
+ /* This is how to allocate empty space in some section. The .zero
+ pseudo-op is used for this on most svr4 assemblers. */
+
+ #define SKIP_ASM_OP ".zero"
+
+ #undef ASM_OUTPUT_SKIP
+ #define ASM_OUTPUT_SKIP(FILE,SIZE) \
+ fprintf (FILE, "\t%s\t%u\n", SKIP_ASM_OP, (SIZE))
+
+ /* Output the label which precedes a jumptable. Note that for all svr4
+ systems where we actually generate jumptables (which is to say every
+ svr4 target except i386, where we use casesi instead) we put the jump-
+ tables into the .rodata section and since other stuff could have been
+ put into the .rodata section prior to any given jumptable, we have to
+ make sure that the location counter for the .rodata section gets pro-
+ perly re-aligned prior to the actual beginning of the jump table. */
+
+ #define ALIGN_ASM_OP ".align"
+
+ #ifndef ASM_OUTPUT_BEFORE_CASE_LABEL
+ #define ASM_OUTPUT_BEFORE_CASE_LABEL(FILE,PREFIX,NUM,TABLE) \
+ ASM_OUTPUT_ALIGN ((FILE), 2);
+ #endif
+
+ #undef ASM_OUTPUT_CASE_LABEL
+ #define ASM_OUTPUT_CASE_LABEL(FILE,PREFIX,NUM,JUMPTABLE) \
+ do { \
+ ASM_OUTPUT_BEFORE_CASE_LABEL (FILE, PREFIX, NUM, JUMPTABLE) \
+ ASM_OUTPUT_INTERNAL_LABEL (FILE, PREFIX, NUM); \
+ } while (0)
+
+ /* The standard SVR4 assembler seems to require that certain builtin
+ library routines (e.g. .udiv) be explicitly declared as .globl
+ in each assembly file where they are referenced. */
+
+ #define ASM_OUTPUT_EXTERNAL_LIBCALL(FILE, FUN) \
+ ASM_GLOBALIZE_LABEL (FILE, XSTR (FUN, 0))
+
+ /* This says how to output assembler code to declare an
+ uninitialized external linkage data object. Under SVR4,
+ the linker seems to want the alignment of data objects
+ to depend on their types. We do exactly that here. */
+
+ #define COMMON_ASM_OP ".comm"
+
+ #undef ASM_OUTPUT_ALIGNED_COMMON
+ #define ASM_OUTPUT_ALIGNED_COMMON(FILE, NAME, SIZE, ALIGN) \
+ do { \
+ fprintf ((FILE), "\t%s\t", COMMON_ASM_OP); \
+ assemble_name ((FILE), (NAME)); \
+ fprintf ((FILE), ",%u,%u\n", (SIZE), (ALIGN) / BITS_PER_UNIT); \
+ } while (0)
+
+ /* This says how to output assembler code to declare an
+ uninitialized internal linkage data object. Under SVR4,
+ the linker seems to want the alignment of data objects
+ to depend on their types. We do exactly that here. */
+
+ #define LOCAL_ASM_OP ".local"
+
+ #undef ASM_OUTPUT_ALIGNED_LOCAL
+ #define ASM_OUTPUT_ALIGNED_LOCAL(FILE, NAME, SIZE, ALIGN) \
+ do { \
+ fprintf ((FILE), "\t%s\t", LOCAL_ASM_OP); \
+ assemble_name ((FILE), (NAME)); \
+ fprintf ((FILE), "\n"); \
+ ASM_OUTPUT_ALIGNED_COMMON (FILE, NAME, SIZE, ALIGN); \
+ } while (0)
+
+ /* This is the pseudo-op used to generate a 64-bit word of data with a
+ specific value in some section. */
+
+ #define INT_ASM_OP ".quad"
+
+ /* This is the pseudo-op used to generate a contiguous sequence of byte
+ values from a double-quoted string WITHOUT HAVING A TERMINATING NUL
+ AUTOMATICALLY APPENDED. This is the same for most svr4 assemblers. */
+
+ #undef ASCII_DATA_ASM_OP
+ #define ASCII_DATA_ASM_OP ".ascii"
+
+ /* Support const sections and the ctors and dtors sections for g++.
+ Note that there appears to be two different ways to support const
+ sections at the moment. You can either #define the symbol
+ READONLY_DATA_SECTION (giving it some code which switches to the
+ readonly data section) or else you can #define the symbols
+ EXTRA_SECTIONS, EXTRA_SECTION_FUNCTIONS, SELECT_SECTION, and
+ SELECT_RTX_SECTION. We do both here just to be on the safe side. */
+
+ #define USE_CONST_SECTION 1
+
+ #define CONST_SECTION_ASM_OP ".section\t.rodata"
+
+ /* Define the pseudo-ops used to switch to the .ctors and .dtors sections.
+
+ Note that we want to give these sections the SHF_WRITE attribute
+ because these sections will actually contain data (i.e. tables of
+ addresses of functions in the current root executable or shared library
+ file) and, in the case of a shared library, the relocatable addresses
+ will have to be properly resolved/relocated (and then written into) by
+ the dynamic linker when it actually attaches the given shared library
+ to the executing process. (Note that on SVR4, you may wish to use the
+ `-z text' option to the ELF linker, when building a shared library, as
+ an additional check that you are doing everything right. But if you do
+ use the `-z text' option when building a shared library, you will get
+ errors unless the .ctors and .dtors sections are marked as writable
+ via the SHF_WRITE attribute.) */
+
+ #define CTORS_SECTION_ASM_OP ".section\t.ctors,\"aw\""
+ #define DTORS_SECTION_ASM_OP ".section\t.dtors,\"aw\""
+
+ /* On svr4, we *do* have support for the .init and .fini sections, and we
+ can put stuff in there to be executed before and after `main'. We let
+ crtstuff.c and other files know this by defining the following symbols.
+ The definitions say how to change sections to the .init and .fini
+ sections. This is the same for all known svr4 assemblers. */
+
+ #define INIT_SECTION_ASM_OP ".section\t.init"
+ #define FINI_SECTION_ASM_OP ".section\t.fini"
+
+ /* A default list of other sections which we might be "in" at any given
+ time. For targets that use additional sections (e.g. .tdesc) you
+ should override this definition in the target-specific file which
+ includes this file. */
+
+ #undef EXTRA_SECTIONS
+ #define EXTRA_SECTIONS in_const, in_ctors, in_dtors
+
+ /* A default list of extra section function definitions. For targets
+ that use additional sections (e.g. .tdesc) you should override this
+ definition in the target-specific file which includes this file. */
+
+ #undef EXTRA_SECTION_FUNCTIONS
+ #define EXTRA_SECTION_FUNCTIONS \
+ CONST_SECTION_FUNCTION \
+ CTORS_SECTION_FUNCTION \
+ DTORS_SECTION_FUNCTION
+
+ #undef READONLY_DATA_SECTION
+ #define READONLY_DATA_SECTION() const_section ()
+
+ extern void text_section ();
+
+ #define CONST_SECTION_FUNCTION \
+ void \
+ const_section () \
+ { \
+ if (!USE_CONST_SECTION) \
+ text_section(); \
+ else if (in_section != in_const) \
+ { \
+ fprintf (asm_out_file, "%s\n", CONST_SECTION_ASM_OP); \
+ in_section = in_const; \
+ } \
+ }
+
+ #define CTORS_SECTION_FUNCTION \
+ void \
+ ctors_section () \
+ { \
+ if (in_section != in_ctors) \
+ { \
+ fprintf (asm_out_file, "%s\n", CTORS_SECTION_ASM_OP); \
+ in_section = in_ctors; \
+ } \
+ }
+
+ #define DTORS_SECTION_FUNCTION \
+ void \
+ dtors_section () \
+ { \
+ if (in_section != in_dtors) \
+ { \
+ fprintf (asm_out_file, "%s\n", DTORS_SECTION_ASM_OP); \
+ in_section = in_dtors; \
+ } \
+ }
+
+ /* Switch into a generic section.
+ This is currently only used to support section attributes.
+
+ We make the section read-only and executable for a function decl,
+ read-only for a const data decl, and writable for a non-const data decl. */
+ #define ASM_OUTPUT_SECTION_NAME(FILE, DECL, NAME, RELOC) \
+ fprintf (FILE, ".section\t%s,\"%s\",@progbits\n", NAME, \
+ (DECL) && TREE_CODE (DECL) == FUNCTION_DECL ? "ax" : \
+ (DECL) && DECL_READONLY_SECTION (DECL, RELOC) ? "a" : "aw")
+
+
+ /* A C statement (sans semicolon) to output an element in the table of
+ global constructors. */
+ #define ASM_OUTPUT_CONSTRUCTOR(FILE,NAME) \
+ do { \
+ ctors_section (); \
+ fprintf (FILE, "\t%s\t ", INT_ASM_OP); \
+ assemble_name (FILE, NAME); \
+ fprintf (FILE, "\n"); \
+ } while (0)
+
+ /* A C statement (sans semicolon) to output an element in the table of
+ global destructors. */
+ #define ASM_OUTPUT_DESTRUCTOR(FILE,NAME) \
+ do { \
+ dtors_section (); \
+ fprintf (FILE, "\t%s\t ", INT_ASM_OP); \
+ assemble_name (FILE, NAME); \
+ fprintf (FILE, "\n"); \
+ } while (0)
+
+ /* A C statement or statements to switch to the appropriate
+ section for output of DECL. DECL is either a `VAR_DECL' node
+ or a constant of some sort. RELOC indicates whether forming
+ the initial value of DECL requires link-time relocations. */
+
+ #define SELECT_SECTION(DECL,RELOC) \
+ { \
+ if (TREE_CODE (DECL) == STRING_CST) \
+ { \
+ if (! flag_writable_strings) \
+ const_section (); \
+ else \
+ data_section (); \
+ } \
+ else if (TREE_CODE (DECL) == VAR_DECL) \
+ { \
+ if ((flag_pic && RELOC) \
+ || !TREE_READONLY (DECL) || TREE_SIDE_EFFECTS (DECL) \
+ || !DECL_INITIAL (DECL) \
+ || (DECL_INITIAL (DECL) != error_mark_node \
+ && !TREE_CONSTANT (DECL_INITIAL (DECL)))) \
+ data_section (); \
+ else \
+ const_section (); \
+ } \
+ else \
+ const_section (); \
+ }
+
+ /* A C statement or statements to switch to the appropriate
+ section for output of RTX in mode MODE. RTX is some kind
+ of constant in RTL. The argument MODE is redundant except
+ in the case of a `const_int' rtx. Currently, these always
+ go into the const section. */
+
+ #undef SELECT_RTX_SECTION
+ #define SELECT_RTX_SECTION(MODE,RTX) const_section()
+
+ /* Define the strings used for the special svr4 .type and .size directives.
+ These strings generally do not vary from one system running svr4 to
+ another, but if a given system (e.g. m88k running svr) needs to use
+ different pseudo-op names for these, they may be overridden in the
+ file which includes this one. */
+
+ #define TYPE_ASM_OP ".type"
+ #define SIZE_ASM_OP ".size"
+
+ /* This is how we tell the assembler that a symbol is weak. */
+
+ #define ASM_WEAKEN_LABEL(FILE,NAME) \
+ do { fputs ("\t.weak\t", FILE); assemble_name (FILE, NAME); \
+ fputc ('\n', FILE); } 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)
+
+ /* The following macro defines the format used to output the second
+ operand of the .type assembler directive. Different svr4 assemblers
+ expect various different forms for this operand. The one given here
+ is just a default. You may need to override it in your machine-
+ specific tm.h file (depending upon the particulars of your assembler). */
+
+ #define TYPE_OPERAND_FMT "@%s"
+
+ /* Write the extra assembler code needed to declare a function's result.
+ Most svr4 assemblers don't require any special declaration of the
+ result value, but there are exceptions. */
+
+ #ifndef ASM_DECLARE_RESULT
+ #define ASM_DECLARE_RESULT(FILE, RESULT)
+ #endif
+
+ /* These macros generate the special .type and .size directives which
+ are used to set the corresponding fields of the linker symbol table
+ entries in an ELF object file under SVR4. These macros also output
+ the starting labels for the relevant functions/objects. */
+
+ /* Write the extra assembler code needed to declare an object properly. */
+
+ #define ASM_DECLARE_OBJECT_NAME(FILE, NAME, DECL) \
+ do { \
+ fprintf (FILE, "\t%s\t ", TYPE_ASM_OP); \
+ assemble_name (FILE, NAME); \
+ putc (',', FILE); \
+ fprintf (FILE, TYPE_OPERAND_FMT, "object"); \
+ putc ('\n', FILE); \
+ size_directive_output = 0; \
+ if (!flag_inhibit_size_directive && DECL_SIZE (DECL)) \
+ { \
+ size_directive_output = 1; \
+ fprintf (FILE, "\t%s\t ", SIZE_ASM_OP); \
+ assemble_name (FILE, NAME); \
+ fprintf (FILE, ",%d\n", int_size_in_bytes (TREE_TYPE (DECL))); \
+ } \
+ ASM_OUTPUT_LABEL(FILE, NAME); \
+ } while (0)
+
+ /* Output the size directive for a decl in rest_of_decl_compilation
+ in the case where we did not do so before the initializer.
+ Once we find the error_mark_node, we know that the value of
+ size_directive_output was set
+ by ASM_DECLARE_OBJECT_NAME when it was run for the same decl. */
+
+ #define ASM_FINISH_DECLARE_OBJECT(FILE, DECL, TOP_LEVEL, AT_END) \
+ do { \
+ char *name = XSTR (XEXP (DECL_RTL (DECL), 0), 0); \
+ if (!flag_inhibit_size_directive && DECL_SIZE (DECL) \
+ && ! AT_END && TOP_LEVEL \
+ && DECL_INITIAL (DECL) == error_mark_node \
+ && !size_directive_output) \
+ { \
+ size_directive_output = 1; \
+ fprintf (FILE, "\t%s\t ", SIZE_ASM_OP); \
+ assemble_name (FILE, name); \
+ fprintf (FILE, ",%d\n", int_size_in_bytes (TREE_TYPE (DECL))); \
+ } \
+ } while (0)
+
+ /* A table of bytes codes used by the ASM_OUTPUT_ASCII and
+ ASM_OUTPUT_LIMITED_STRING macros. Each byte in the table
+ corresponds to a particular byte value [0..255]. For any
+ given byte value, if the value in the corresponding table
+ position is zero, the given character can be output directly.
+ If the table value is 1, the byte must be output as a \ooo
+ octal escape. If the tables value is anything else, then the
+ byte value should be output as a \ followed by the value
+ in the table. Note that we can use standard UN*X escape
+ sequences for many control characters, but we don't use
+ \a to represent BEL because some svr4 assemblers (e.g. on
+ the i386) don't know about that. Also, we don't use \v
+ since some versions of gas, such as 2.2 did not accept it. */
+
+ #define ESCAPES \
+ "\1\1\1\1\1\1\1\1btn\1fr\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\
+ \0\0\"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\
+ \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\\\0\0\0\
+ \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\1\
+ \1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\
+ \1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\
+ \1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\
+ \1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1\1"
+
+ /* Some svr4 assemblers have a limit on the number of characters which
+ can appear in the operand of a .string directive. If your assembler
+ has such a limitation, you should define STRING_LIMIT to reflect that
+ limit. Note that at least some svr4 assemblers have a limit on the
+ actual number of bytes in the double-quoted string, and that they
+ count each character in an escape sequence as one byte. Thus, an
+ escape sequence like \377 would count as four bytes.
+
+ If your target assembler doesn't support the .string directive, you
+ should define this to zero.
+ */
+
+ #define STRING_LIMIT ((unsigned) 256)
+
+ #define STRING_ASM_OP ".string"
+
+ /*
+ * We always use gas here, so we don't worry about ECOFF assembler problems.
+ */
+ #undef TARGET_GAS
+ #define TARGET_GAS (1)
+
+ #undef PREFERRED_DEBUGGING_TYPE
+ #define PREFERRED_DEBUGGING_TYPE DBX_DEBUG
+
+ #ifdef NOTDEF
+ /* Provide a STARTFILE_SPEC appropriate for NetBSD. Here we add
+ the NetBSD magical crtbegin.o file (see crtstuff.c) which
+ provides part of the support for getting C++ file-scope static
+ object constructed before entering `main'. */
+
+ #undef STARTFILE_SPEC
+ #define STARTFILE_SPEC \
+ "%{!shared: \
+ %{pg:gcrt1.o%s} %{!pg:%{p:gcrt1.o%s} %{!p:crt1.o%s}}}\
+ crti.o%s crtbegin.o%s"
+
+ /* Provide a ENDFILE_SPEC appropriate for NetBSD. Here we tack on
+ the NetBSD magical crtend.o file (see crtstuff.c) which
+ provides part of the support for getting C++ file-scope static
+ object constructed before entering `main', followed by a normal
+ NetBSD "finalizer" file, `crtn.o'. */
+
+ #undef ENDFILE_SPEC
+ #define ENDFILE_SPEC \
+ "crtend.o%s crtn.o%s"
+ #endif
diff -Nrc3p egcs-1.0.2.cygnus/gcc/config/alpha/netbsd.h egcs-1.0.2/gcc/config/alpha/netbsd.h
*** egcs-1.0.2.cygnus/gcc/config/alpha/netbsd.h Wed Dec 31 16:00:00 1969
--- egcs-1.0.2/gcc/config/alpha/netbsd.h Fri Apr 17 10:09:57 1998
***************
*** 0 ****
--- 1,98 ----
+ /* Definitions of target machine for GNU compiler, for Alpha NetBSD,
+ using ECOFF.
+ Copyright (C) 1996 Free Software Foundation, Inc.
+ Contributed by Bob Manson.
+
+ This file is part of GNU CC.
+
+ GNU CC 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 2, or (at your option)
+ any later version.
+
+ GNU CC 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 GNU CC; see the file COPYING. If not, write to
+ the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */
+
+ #undef TARGET_DEFAULT
+ #define TARGET_DEFAULT (3 | MASK_GAS)
+
+ #undef TARGET_VERSION
+ #define TARGET_VERSION fprintf (stderr, " (NetBSD/Alpha)");
+
+ #undef CPP_PREDEFINES
+ #define CPP_PREDEFINES "\
+ -D__alpha -D__alpha__ -D__netbsd -D__netbsd -D_LONGLONG -Dnetbsd -Dunix \
+ -Asystem(netbsd) -Acpu(alpha) -Amachine(alpha)"
+
+ /* We don't actually need any of these; the MD_ vars are ignored
+ anyway for cross-compilers, and the other specs won't get picked up
+ because the user is supposed to do ld -r (hmm, perhaps that should be
+ the default). In any case, setting them thus will catch some
+ common user errors. */
+
+ #undef MD_EXEC_PREFIX
+ #undef MD_STARTFILE_PREFIX
+
+ #undef LIB_SPEC
+ #define LIB_SPEC "%{pg:-lgmon} %{pg:-lc_p} %{!pg:-lc}"
+
+ #undef LINK_SPEC
+ #define LINK_SPEC "-G 8 %{O*:-O3} %{!O*:-O1}"
+
+ #undef ASM_SPEC
+ #define ASM_SPEC "-nocpp"
+
+ /* Can't do stabs */
+ #undef SDB_DEBUGGING_INFO
+
+ /* Prefer dbx. */
+ #undef PREFERRED_DEBUGGING_TYPE
+ #define PREFERRED_DEBUGGING_TYPE DBX_DEBUG
+
+ #undef FUNCTION_PROFILER
+ #define FUNCTION_PROFILER(FILE, LABELNO) \
+ fputs ("\tjsr $28,_mcount\n", (FILE))
+
+ /* Generate calls to memcpy, etc., not bcopy, etc. */
+ #define TARGET_MEM_FUNCTIONS
+
+ /* Show that we need a GP when profiling. */
+ #define TARGET_PROFILING_NEEDS_GP
+
+ /* We need that too. */
+ #define HANDLE_SYSV_PRAGMA
+
+ #undef ASM_FINAL_SPEC
+
+ /* Emit RTL insns to initialize the variable parts of a trampoline.
+ FNADDR is an RTX for the address of the function's pure code.
+ CXT is an RTX for the static chain value for the function.
+
+ This differs from the standard version in that:
+
+ We do not initialize the "hint" field because it only has an 8k
+ range and so the target is in range of something on the stack.
+ Omitting the hint saves a bogus branch-prediction cache line load.
+
+ Linux always has an executable stack -- no need for a system call.
+ */
+
+ #undef INITIALIZE_TRAMPOLINE
+ #define INITIALIZE_TRAMPOLINE(TRAMP, FNADDR, CXT) \
+ { \
+ rtx _addr; \
+ \
+ _addr = memory_address (Pmode, plus_constant ((TRAMP), 16)); \
+ emit_move_insn (gen_rtx (MEM, Pmode, _addr), (FNADDR)); \
+ _addr = memory_address (Pmode, plus_constant ((TRAMP), 24)); \
+ emit_move_insn (gen_rtx (MEM, Pmode, _addr), (CXT)); \
+ \
+ emit_insn (gen_rtx (UNSPEC_VOLATILE, VOIDmode, \
+ gen_rtvec (1, const0_rtx), 0)); \
+ }
diff -Nrc3p egcs-1.0.2.cygnus/gcc/config/alpha/t-netbsd egcs-1.0.2/gcc/config/alpha/t-netbsd
*** egcs-1.0.2.cygnus/gcc/config/alpha/t-netbsd Wed Dec 31 16:00:00 1969
--- egcs-1.0.2/gcc/config/alpha/t-netbsd Fri Apr 17 10:09:57 1998
***************
*** 0 ****
--- 1 ----
+ USER_H = $(EXTRA_HEADERS) $(LANG_EXTRA_HEADERS)
diff -Nrc3p egcs-1.0.2.cygnus/gcc/config/alpha/xm-netbsd.h egcs-1.0.2/gcc/config/alpha/xm-netbsd.h
*** egcs-1.0.2.cygnus/gcc/config/alpha/xm-netbsd.h Wed Dec 31 16:00:00 1969
--- egcs-1.0.2/gcc/config/alpha/xm-netbsd.h Fri Apr 17 10:09:57 1998
***************
*** 0 ****
--- 1,2 ----
+ #undef USE_BFD
+ #define USE_BFD
diff -Nrc3p egcs-1.0.2.cygnus/gcc/configure egcs-1.0.2/gcc/configure
*** egcs-1.0.2.cygnus/gcc/configure Mon Mar 16 00:26:57 1998
--- egcs-1.0.2/gcc/configure Fri Apr 17 10:09:57 1998
*************** for machine in $build $host $target; do
*** 1827,1832 ****
--- 1827,1841 ----
thread_file='posix'
fi
;;
+ alpha*-*-netbsd*)
+ tm_file="${tm_file} alpha/netbsd.h alpha/elf-netbsd.h"
+ xm_file="${xm_file} alpha/xm-netbsd.h"
+ # On NetBSD, the headers are already okay, except for math.h.
+ fixincludes=fixinc.math
+ tmake_file="t-netbsd alpha/t-netbsd"
+ xmake_file=none
+ gas=yes gnu_ld=yes
+ ;;
alpha*-dec-osf[456789]*)
if [ x$stabs = xyes ]
then
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: gcc alpha/NetBSD 1.3 target files
1998-04-30 18:55 gcc alpha/NetBSD 1.3 target files Paul H. Anderson
@ 1998-05-01 9:55 ` Richard Henderson
1998-05-04 10:58 ` Paul H. Anderson
0 siblings, 1 reply; 4+ messages in thread
From: Richard Henderson @ 1998-05-01 9:55 UTC (permalink / raw)
To: Paul H. Anderson; +Cc: egcs
On Thu, Apr 30, 1998 at 03:43:49PM -0400, Paul H. Anderson wrote:
> The following patches are a initial effort to get egcs working on the
> NetBSD/Alpha platform. The code comes entirely from the Linux work,
> usually with simple substitutions and a few slightly larger modifications.
This will not be added to the 1.0.2 branch, as such changes at this
time are reserved for drop-dead bug fixes.
However, I will be happy to help you integrate these into the
development sources, destined to become egcs 1.1. You will also find
that you have an easier time of it, since sometime after the 1.0
branch I reorganized things such that another Alpha/ELF system would
not have to duplicate all the files.
I'll take a look at integrating this patch this weekend, after which
it would need some verification.
r~
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: gcc alpha/NetBSD 1.3 target files
1998-05-01 9:55 ` Richard Henderson
@ 1998-05-04 10:58 ` Paul H. Anderson
1998-05-16 23:27 ` Richard Henderson
0 siblings, 1 reply; 4+ messages in thread
From: Paul H. Anderson @ 1998-05-04 10:58 UTC (permalink / raw)
To: Richard Henderson; +Cc: egcs
On Fri, 1 May 1998, Richard Henderson wrote:
> On Thu, Apr 30, 1998 at 03:43:49PM -0400, Paul H. Anderson wrote:
> > The following patches are a initial effort to get egcs working on the
> > NetBSD/Alpha platform. The code comes entirely from the Linux work,
> > usually with simple substitutions and a few slightly larger modifications.
>
> This will not be added to the 1.0.2 branch, as such changes at this
> time are reserved for drop-dead bug fixes.
Okie dokie.
> However, I will be happy to help you integrate these into the
> development sources, destined to become egcs 1.1. You will also find
> that you have an easier time of it, since sometime after the 1.0
> branch I reorganized things such that another Alpha/ELF system would
> not have to duplicate all the files.
>
> I'll take a look at integrating this patch this weekend, after which
> it would need some verification.
I'll be happy to either re-integrate the changes into the development
sources, or to do verification of whatever patches you make. Just let me
know what you want me to work on. My project strongly requires a working
g++ for *BSD on the Alpha, so I'm pretty motivated to make this go.
Just keep in mind that I'm new to the Cygnus/GNU way of doing software
development. I have done large scale development using GNU tools, just
not on them.
Paul
+------------------------------------------------------+
| Paul Anderson Public Data Queries, Inc. |
| pha@pdq.com 734-213-4964(W) 994-3734(H) |
+------------------------------------------------------+
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: gcc alpha/NetBSD 1.3 target files
1998-05-04 10:58 ` Paul H. Anderson
@ 1998-05-16 23:27 ` Richard Henderson
0 siblings, 0 replies; 4+ messages in thread
From: Richard Henderson @ 1998-05-16 23:27 UTC (permalink / raw)
To: Paul H. Anderson; +Cc: egcs
On Mon, May 04, 1998 at 10:23:19AM -0400, Paul H. Anderson wrote:
> I'll be happy to either re-integrate the changes into the development
> sources, or to do verification of whatever patches you make. Just let me
> know what you want me to work on.
I've checked in what I believe to be the appropriate changes to the
development sources. Let me know if I got it right.
r~
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~1998-05-16 23:27 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1998-04-30 18:55 gcc alpha/NetBSD 1.3 target files Paul H. Anderson
1998-05-01 9:55 ` Richard Henderson
1998-05-04 10:58 ` Paul H. Anderson
1998-05-16 23:27 ` Richard Henderson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).