If one doesn't subtract the stack bias back off %fp before attempting
to access variables stored relative to %fp the wrong address is picked up for
all locals and args. So, calls to the FRAME_LOCALS_ADDRESS or
FRAME_ARGS_ADDRESS macros need a function for sparc64 specific cases.
Provide a function to check for bias's addresses and fix them up.
support. Specifically, got the changes to md_prepare_relax_scan
from binutils' revision 1.11 of this file. As a bonus also
now define DWARF2_LINE_MIN_INSN_LENGTH.
first. This is necessary to avoid warnings with -fshort-enums. Casting
to an int really should be enough, but turns out not to be.
This change will be documented in doc/HACKS.
script files (and now that they are in cvs, they can't be fixed).
Future imports of any package should use "-ko" and do away with *2netbsd
altogether, as was the intent of gnu/dist in the first place.
set the C(arry) bit appropriately. All other leave it in an indeterminate
(to GCC) state. Mark that by setting CC_NO_OVERFLOW. Change vax.md emission
of branches that use the C bit to use OUTPUT_JUMP so that if CC_NO_OVERFLOW is
set, branches that would normally use the C bit use opcodes that don't use the
C bit (jgtru -> jneq && jlssu -> jeql). Delete the cmpdi pattern.
Define IN_SIGTRAMP() as nbsd_in_sigtramp(), a new function which knows
how to find the address of the signal trampoline at runtime, thus allowing
one gdb binary to work on all NetBSD/m68k machines.
(should make it into stock binutils as soon as Matt's copyright
assignment paperwork is finalized):
bfd:
* elf32-m68k.c (elf32_m68k_print_private_bfd_data): Recognize
EF_M68000.
binutils:
* readelf.c (get_machine_flags): Recognize EF_M68000.
gas:
* config/tc-m68k.c (md_show_usage): No longer display a
hard-coded "68020" for the default CPU, instead display the
canonical name of the true, configured default CPU.
(m68k_elf_final_processing): Mark objects for sub-68020
CPUs with the new EF_M68000 flag.
include/elf:
* m68k.h (EF_M68000): Define.
master binutils sources that fix gprof for LP64 platforms and
also fix some problems with cross-gprof:
* TODO: Remove "host architecture pointer size" item.
* gmon.h (GMON_HDRSIZE_BSD44_32): Define.
(GMON_HDRSIZE_BSD44_64): Ditto.
(GMON_HDRSIZE_OLDBSD_32): Ditto.
(GMON_HDRSIZE_OLDBSD_64): Ditto.
(struct raw_phdr): Wrap in #if 0, keeping it for
documentation purposes only.
(struct old_raw_phdr): Likewise.
(struct raw_arc): Likewise. Change type/size of
"count" member to long match 4.4BSD.
* gmon_io: Update copyright years.
(gmon_io_read_64): New function.
(gmon_io_read_vma): Use bfd_arch_bits_per_address to
determine target pointer size. Use gmon_io_read_32
and gmon_io_read_64.
(gmon_io_write_64): New function.
(gmon_io_write_vma): Use bfd_arch_bits_per_address to
determine target pointer size. Use gmon_io_write_32
and gmon_io_write_64.
(get_vma): Remove.
(put_vma): Ditto.
(gmon_read_raw_arc): New function.
(gmon_write_raw_arc): New function.
(gmon_out_read): Do not use struct raw_phdr or
struct old_raw_phdr to read the gmon header. Use
gmon_read_raw_arc to read call graph records.
(gmon_out_write): Do not use struct raw_phdr or
struct old_raw_phdr to write the gmon header. Use
gmon_write_raw_arc to write call graph records.
This makes the GMON_PTR_SIZE constant obsolete, however I have not
included the changes to the autoconf machinery that eliminate that
constant completely in order to minimize the changes here (they
will be picked up when we import Binutils 2.12 when it is released).
2001-03-04 Andrew Cagney <ac131313@redhat.com>
* emul_netbsd.c [WITH_NetBSD_HOST]: Include <sys/mount.h> and
<errno.h>.
(do_stat): Only do SYS test when SYS_stat defined.
(do_sigprocmask): Ditto for SYS_sigprocmask.
(do_fstat): Ditto for SYS_fstat.
(do_getdirentries): Ditto for SYS_getdirentries.
(do_lstat): Ditto for SYS_lstat.
2001-01-15 Geoffrey Keating <geoffk@redhat.com>
* emul_netbsd.c (do_open): Translate the flag parameter to the
open syscall to the numbers supported by the host.
* part of port-powerpc/14307
2001-10-19 Andrew Cagney <ac131313@redhat.com>
* configure.in: When Linux or NetBSD, enable PowerPC simulator.
* configure: Re-generate.
binutils-current respository.
2001-06-24 H.J. Lu <hjl@gnu.org>
* objcopy.c (strip_main): Revert the change made on 2001-05-30
by accident.
(copy_main): Apply the the change made to strip_main on
2001-05-30 by accident.
strip now passes the argument list as documented.
http://gcc.gnu.org/ml/gcc-patches/2000-06/msg00389.html
ChangeLog for this change (which was never committed to the master
GCC sources; the bug still exists in GCC 3.x!):
2000-05-09 Ethan Solomita ethan@cs.columbia.edu
* config/mips/mips.h (FUNCTION_PROFILER): Deal with
the case where jal is expanded as a macro.
This is done by removing the .reorder/.noreorder, and moving the
call to _mcount to after the stack adjust, thereby letting the
assembler deal with the delay slot, etc.
This fixes profiling with libraries built -mabicalls (default on
NetBSD).
I can't believe how long this problem has been discussed-but-not-fixed
on the GCC mailing list...
1) There is already a destructor that does this. The atexit() is superfluous.
2) If libstdc++ is loaded dynamically by dlopen() and then is removed by
dlclose(), the pointer is no longer valid and the program would core dump
at exit() time. (This caused xmms to core dump at exit when xmms-sid was
installed, for example.)
NOTE: THIS NEEDS TO BE APPLIED TO THE `TOOLCHAIN' VERSION.
* jump-format jumps: the target is in RB, so use the RB bits in
the insn, not the RA bits.
* branch-format displacements are 21 bits; extract all of them, and
sign-extend correctly.
* For conditional branches, we need to determine whether or not
the branch is actually taken.
and ASM_FINISH_DECLARE_OBJECT macros in <alpha/netbsd.h>,
and remove ASM_DECLARE_FUNCTION_NAME, ASM_DECLARE_OBJECT_NAME,
ASM_FINISH_DECLARE_OBJECT, and ASM_DECLARE_FUNCTION_SIZE from
<netbsd.h>.
apply gcc original rev. 1.59:
> * config/sh/sh.c (prepare_scc_operands): Apply force_reg to
> sh_compare_op1 when the mode is DImode and sh_compare_op1 isn't
> const0_rtx.
the patch I posted to tech-toolchain@netbsd.org as a workaround for the
GOT bug whereby the callee-saved register %a5 can be trashed in certain
optimised -fPIC functions.
This may yet need to be fixed a different way, pending input from
Andreas Schwab on his change in revison 1.33 of this file in the
FSF source...
In the meantime, I've built and installed several NetBSD shared libraries
(including libc) with my change and found no problems.
of this file in the master FSF sources; I obviously disagree
with the change, since it was apparently made to appease one
particular OS, even though previous binutils releases had been
made with ENTRY=__start, and have brought it up on
binutils@sources.redhat.com).
aligned. Corresponds to part of rev 1.21 of this file in the
master FSF sources.
Thanks to Michael Rauch <mrauch@netbsd.org> for fishing out the patch.
is in their revision 1.186, from schwab@suse.de, and said:
Don't call reg_used_between_p if the insn from BL2 is after
the insn from BL.
This fixes a cc1 SEGV on sys/dev/scsipi/ch.c rev 1.48 under -O2 or -Os,
seen on m68k and vax.
* move the code that supplies the integer and FP registers into
separate functions, rather than duplicating the code for the
ptrace and core file cases.
* Use supply_register() rather than just copying directly into
the register array and calling registers_fetched(). This way,
only the registers actually supplied are marked as valid within
the debugger.
* Add support for SSE/SSE2 registers via the PT_{GET,SET}XMMREGS
ptrace(2) request.
(Blocked on the FSF assignment clerk for feeding this back to
the master GDB sources.)
the 4.17 NetBSD support from our tree. Original port by Chris Sekiya, with
minimal tweaks and cleanups by me (mainly: build bi-endian by default, add
NetBSD kcore support back).
and place it in i386nbsd-tdep.c (new file).
- Move solib.o from TDEPFILES to NATDEPFILES (solib.c requires target
headers, namely <link_elf.h>).
Fixes build of cross-gdb pointed at i386--netbsd target.
jmc