Commit Graph

3283 Commits

Author SHA1 Message Date
bjh21
7bdb61e034 Fix from Richard Earnshaw to make MASK_RETURN_ADDR work on both 26-bit and
32-bit ARMs, switching at run time.
2001-11-08 20:47:56 +00:00
rafal
9f74cbc2da Oops, last round checked in make droppings here. Check in the actual
toolchain2netbsd generated files, though this is as of yet untested.
2001-11-08 08:20:31 +00:00
rafal
f317d28b1a New toolchain files for mipseb. Mainly toolchain2netbsd output with some
changes to configuration stuff to (a) recognize `mipseb', and (b) build a
BE-default GCC on mipseb.  gprof and gdb still not done.

WARNING: Binutils 2.11.2 (maybe earlier) changed the MIPS ABI, so any
shared libs built by this toolchain WILL NOT WORK without either a whack
to BFD to fix that or a patch to ld_elf.so to work around it.  I need to
chase the binutils folks on this issue still.

That said, the new toolchain seems to work quite well once the ABI change
is worked around/fixed -- I'm committing from a machine running a user-
land built with the new compiler.
2001-11-08 07:21:53 +00:00
simonb
1adc8dd435 Don't rebuild "fini" _every_ time make(1) is run; avoids a complete
rebuild of f771 all the time.
2001-11-08 00:13:53 +00:00
tv
f965ecc2a3 Give mips*-*-netbsd* a gdb_target=nbsd line. 2001-11-06 16:32:33 +00:00
aymeric
589ec775e4 Make xmalloc() and xrealloc()'s definitions match their prototypes in
scan.h
2001-11-05 20:49:47 +00:00
scw
e994d009cd Add target/rule for m68kelfnbsd.o 2001-11-05 19:50:35 +00:00
scw
faad780fa8 Change target_emul for m68*-*-netbsdelf* to m68kelfnbsd. 2001-11-05 19:49:45 +00:00
scw
ec53366faa Add m68k-specific `toolchain2netbsd' generated toolchain build/config files. 2001-11-05 19:10:55 +00:00
scw
dd0e9d09ce Need "gdbcore.h" for struct core_fns. 2001-11-04 09:22:50 +00:00
scw
8f7fc0baa2 Build kcore and gdbserver support into m68k gdb. 2001-11-04 09:22:07 +00:00
scw
116d3a49bf Define SVR4_SHARED_LIBS when __ELF__ is defined now that most of NetBSD's
m68k ports have made the switch.
2001-11-04 09:20:26 +00:00
tv
ec5a9a312e Add REQUIRETOOLS to make version-specific new toolchain compiler libs fail
to build if USETOOLS is set to "no" by the builder.  (Sanity check; as an
override, USETOOLS may be set to "never", but see warnings in BUILDING.)
2001-10-31 16:27:31 +00:00
bjh21
18f0b6b3bd Move the code for splitting and joining a 26-bit R15 into arm-tdep.c, since
it should be sharable with other 26-bit systems (Linux, RISC iX?).

Also use supply_register(), since it seems to be the modern thing to do.
2001-10-29 20:37:59 +00:00
mrg
d59ba42415 disable gdb on alpha host. XXX FIXME 2001-10-28 13:53:37 +00:00
matt
37a2ed9298 Updated ARM defs.mk 2001-10-27 05:46:02 +00:00
briggs
1abee6c6fd Add:
-F (default) h/w flow control
	-f        no h/w flow control
	-q           be quiet during downloads
2001-10-26 14:44:13 +00:00
mrg
857523fed5 fix a compile error when _not_:
#if HOST_BITS_PER_WIDE_INT == 32

found building cross powerpc compiler on sparc64 host.
2001-10-26 03:14:55 +00:00
bjh21
b685397e2d When fetching registers, if CPSR is zero (and hence we've just fetched from
a 26-bit target), split R15 into PC and CPSR.  Similarly, when storing
registers, copy CPSR back into R15 if CPSR indicates a 26-bit mode.
This makes single-stepping work correctly on arm26.

Also, when fetching registers, set arm_apcs_32 based on the mode the target's
in.  This removes the need to kludge it manually.  The variable should really
be lost entirely, and arm-tdep.c should look at the mode bits in CPSR, but
the ARM Linux target is horribly broken in this area, and will either need
fixing or working around.
2001-10-25 20:55:13 +00:00
thorpej
f2cc4f7bb2 Give the host tool build more control over the HOST_CPPFLAGS. 2001-10-25 02:37:18 +00:00
thorpej
a0e5722800 Descend into mdsetimage. 2001-10-24 23:58:58 +00:00
thorpej
e46fa74fd3 Add support for building as a target-prefixed cross-tool from src/tools. 2001-10-24 23:56:38 +00:00
thorpej
98f31d05a1 Apply patch from misc/13927. Also corresponds to master binutils
rev 1.24 of this file.  Binutils log message:

Fri Sep 14 13:56:45 2001 UTC by nickc:
Fix bug emitting R_SPARC_UA32 relocs into shared objects
2001-10-24 21:01:31 +00:00
bjh21
61d38b1990 Fix up the Makefile fragments for GDB on NetBSD/arm so they actually
correspond to the files we've got.
2001-10-24 13:47:52 +00:00
matt
e1c1698075 Regen from toolchain2netbsd 2001-10-23 16:53:45 +00:00
fredb
8dd298e978 Get the information for the "Release" field via running "uname" on the
host that's doing the filing (with a suitable comment for non-usual
cases), as suggested by Don Yuniskis in PR 14217 and lukem on tech-pkg.
Also closes PR's 13938, 14104.
2001-10-23 14:29:07 +00:00
bjh21
8523ec52e8 Replace Mark Brinicombe's NetBSD/arm glue with Richard Earnshaw's, since the
latter is likely to be easier to get into the GNU sources.  I believe the only
functional change is the loss of a working get_longjmp_target().
2001-10-22 21:09:47 +00:00
bjh21
44b78e0623 Correct a comment: NetBSD uses an undefined instruction as a breakpoint,
not a SWI.
2001-10-22 19:07:10 +00:00
bjh21
27c601625f Apply Richard Earnshaw's gdb.step.patch, which adds software single-step
support to GDB ARM targets in general, and make corresponding changes to
NetBSD-specific code.

The first half of this has already been send to gdb-patches by Richard.
The second half is irrelevant to them since they don't yet have NetBSD/arm
support in their tree yet.
2001-10-22 18:58:50 +00:00
simonb
ced4db818c Bah, move ".include <bsd.own.mk>" to before USE_NEW_TOOLCHAIN check. 2001-10-22 03:31:41 +00:00
simonb
83a2424500 Add a Makefile for mdsetimage that tries to do the new/old toolchain
dance to get to libbfd.  Tested on new and old toolchain, makefile
messiness to be verified before being enabled.
2001-10-22 03:23:00 +00:00
simonb
ad8890a6fa Add the bfd-mdsetimage pkgsrc version of mdsetimage - this is Chris
Demetriou's origianal mdsetimage with modifications by Kazuki Sakamoto
to use libbfd.  It lives in gnusrc because libbfd is GPL, not LGPL.
2001-10-22 03:19:29 +00:00
bjh21
318cfb14be Add a hack to automatically switch to 26-bit mode when starting up on an
arm26 system.  This could be done better.
2001-10-20 17:44:58 +00:00
bjh21
408ac67774 Oops: Get the test for 32-bitness the right way around. 2001-10-20 17:14:36 +00:00
bjh21
f3da28335b arm_addr_bits_remove: Stop believing in the existence of 26-bit Thumb.
arm_frame_saved_pc: Use arm_addr_bits_remove to strip the retrieved PC.

These seem to make stack backtraces work on arm26.
2001-10-20 16:57:42 +00:00
tv
32bac67e87 Exclude Y2K bug flag files from the build; fixes undefined symbol errors
reported by jbernard@mines.edu in PR bin/14222.
2001-10-19 03:33:35 +00:00
tv
d20ff568a2 Remove all traces of "aux.m4" so as not to conflict with Microsoft host OS's.
(Since we aren't A/UX, this is not much of a loss for a NetBSD host.)
2001-10-19 02:53:02 +00:00
bjh21
e426602284 Update ARM software single-stepping support to work with GDB 5.0. 2001-10-19 00:53:21 +00:00
bjh21
02e1edfb30 Small changes to make GDB for arm--netbsdelf compile:
1: Don't include arm-convert.o.  arm-convert.s contains only comments, and our
   build procedure can't cope with .s files.
2: Add "check_format" and "core_sniffer" entries to netbsd_core_fns, the same
   as in i386nbsd_nat.c.

The first of these will need toolchain2netbsd to be re-run.  While these
changes let GDB compile, I haven't tried running it on an ARM ELF system yet.
2001-10-18 15:52:57 +00:00
matt
9744d3aebc If MACHINE_ARCH == arm, install into /usr/include/${MACHINE_ARCH} 2001-10-18 06:12:48 +00:00
matt
1532b67157 Update from toolchain2netbsd run. (though some only have the updated
toolchain2netbsd version run)
2001-10-18 05:45:26 +00:00
matt
e42e6e66e6 Updated files from new toolchain2netbsd 2001-10-18 05:42:26 +00:00
jmc
e5447065e5 Add cpp.1 as an MLINK 2001-10-17 05:39:04 +00:00
matt
73b0a5ab88 disable minmax_arithsi pattern (from Richard Earnshaw). 2001-10-15 20:54:06 +00:00
bjh21
f5e1b9570e Update version numbers to match the current state of the world. 2001-10-15 18:28:28 +00:00
nathanw
7f59122d03 Use @true instead of @${TRUE} in includes-foo targets, since there is no
variable TRUE defined in our makefile system.

This prevents "make includes" from breaking with older bsd.subdir.mk, and is
more consistent with the uses of "true" in the rest of the tree.
2001-10-12 21:05:08 +00:00
tv
e2fa0e8e96 CLEANFILES has to be set before including <bsd.prog.mk> to make sure that
clean/cleandir does its job (this should be fixed).  Work around it, and
fix some parallelism problems.
2001-10-08 23:18:11 +00:00
tv
c88b6a3e03 Deal brute-force with parallelism. In order to accomodate breaking the
build in the middle and restarting on another platform (requiring atomic
host tool builds), and keep parallelism, the ".lo" rules can't be used
at all.  Instead, compile all host .c files directly into executables.
2001-10-08 22:56:31 +00:00
bjh21
59ba87fb54 Add an entry for NetBSD/arm. 2001-10-08 22:47:31 +00:00
tv
68c509655a Single-target subdir builds no longer work at all since the introduction
of "dependall" and .WAIT; add a .WAIT here and nuke the dependencies for now.
2001-10-08 22:27:34 +00:00