* Makefile.in: Add rule for gnu-nat.o and i386gnu-nat.o (Gnu Hurd)
* config/i386/i386gnu.mh: remove rules for [i386]gnu-nat.o, now
in Makefile.in (as for other targets); add NATDEPFILE corelow.o to
satisfy symbol dependancy in solib.c (core_ops).
* target.[ch] conditionalize Mach-specific signals so that they
won't show up in non-Mach gdb's!
* thread.c: add function thread_id_to_pid; change name of static
function "thread_switch" to "switch_to_thread", to avoid conflict
with Mach global symbol; move thread_cmd_list to global scope
so targets can add their own thread commands.
* infrun.c: sig_print_info: allow for long signal names.
* gnu-nat.[ch]: tidying up comments.
* gnu-nat.c: move several functions to static scope, just for tidy;
remove calls to prune_threads and renumber_threads;
gnu_wait must not return -1 when inferior exits;
attach_to_child will modify inferior_pid in a way that allows
fork_inferior to remain unchanged;
remove extra arg from startup_inferior;
move mach-specific thread commands here from thread.c.
* procfs.c: Minor changes to make pre-ANSI compilers happy.
(procfs_notice_signals): Copy traced signal set back to
pi->prrun.pr_trace.
(unconditionally_kill_inferior): If PR_DEAD is defined,
rerun inferior after killing it.
floatformat_ieee_double_littlebyte_bigword for little endian
target byte order.
* utils.c (floatformat_to_doublest): Create local preswapped
copy of input for floatformat_littlebyte_bigword formats.
(get_field, put_field): Treat floatformat_littlebyte_bigword
the same as floatformat_little.
(floatformat_from_doublest): Postswap output words for
the floatformat_littlebyte_bigwords format.
does not require a command line arg, since most simulators
don't currently support passing such an arg into the simulated
program.
* gdb.base/commands.exp: Change tests to insert the proper
value as the arg to the first recursive factorial call. Change
compilation line to define FAKEARGV at compile time.
(DONT_RELOCATE_SYMFILE_OBJFILE): Removed.
* xcoffsolib.h (struct vmap): Add new members tvma, toffs and dvma,
remove tadj.
* exec.c (bfdsec_to_vmap): Initialize new vmap members, initialize
tstart and dstart with section VMA.
* rs6000-nat.c (vmap_symtab): Relocate relative to the VMA in the
object file.
(vmap_ldinfo, xcoff_relocate_core): Adjust tstart by section offset
of the text section, remove DONT_RELOCATE_SYMFILE_OBJFILE hack.
(vmap_exec): Relocate relative to the VMA in the object file,
relocate .bss section as well.
(xcoff_relocate_core): No longer adjust section addresses by VMA.
* rs6000-tdep.c (find_toc_address): Change type of tocbase
to CORE_ADDR.
* xcoffread.c (secnum_to_bfd_section): New routine to get
BFD section from CS section number.
(scan_xcoff_symtab): Make toc_offset section relative.
* symtab.c (total_number_of_methods): Avoid core dump if
baseclass type is still undefined.
that directory.
* defs.h utils.c (error warning): Make message be const.
* main.c (fputs_unfiltered): Only send gdb_stdout and gdb_stderr
to hook. Otherwise send it to fputs.
* monitor.c monitor.h (monitor_get_dev_name): New function. Does
the obvious.
* remote-e7000.c: Remove debugify stuff. Change printf, fprintf
to _filtered forms to make output appear in GUIs. Replace all
uses of SERIAL_READCHAR with readchar, which has better error
checking.
* (e7000_parse_device): Add prototype.
* (readchar): Improve doc. Handle random serial errors.
* (expect): Disable notice_quit code. It's busted. Remove
serial error handling (it's now handled in readchar). Remove
remote_debug echoing. That's handled in readchar as well.
* (e7000_parse_device): Remove serial_flag arg. It's not
necessary.
* (e7000_open): Split into two pieces. Second part is
e7000_start_remote, and is error protected. Now, when we connect
to the target, we setup the initial frame and registers so that
the user gets an immediate indication of where the target is.
* (gch): Remove debug output. That's handled by readchar.
* (e7000_read_inferior_memory): Handle errors better.
* (_initialize_remote_e7000): Get rid of `<xxx>' things from
command names. They show up when doing completion and confuse
things horribly.
* ser-e7kpc.c: Remove the last seven months of brain damage. Get
rid of the DLL's since we can access the device directly from Win32s
and Win95. Get rid of debugify crud.
* serial.c: Remove debugify cruft.
* (serial_logchar serial_log_command serial_write serial_readchar
serial_send_break serial_close): Merge common functionality into
serial_logchar. Clean up rest of routines.
* sparclet-rom.c: Disembowel. Leave only download routine.
Download routine now switches to remote target automatically.
* top.c (disconnect): Only define if SIGHUP is defined. Cleans
up MSVC/Win32 problem.
* utils.c (gdb_flush): Don't call hook unless it's for gdb_stdout
or gdb_stderr.
* config/sh/tm-sh.h: Define TARGET_SH for WinGDB.
* config/sparc/tm-sparclet.h: Remove override for prompt.
that directory.
* defs.h utils.c (error warning): Make message be const.
* main.c (fputs_unfiltered): Only send gdb_stdout and gdb_stderr
to hook. Otherwise send it to fputs.
* monitor.c monitor.h (monitor_get_dev_name): New function. Does
the obvious.
* remote-e7000.c: Remove debugify stuff. Change printf, fprintf
to _filtered forms to make output appear in GUIs. Replace all
uses of SERIAL_READCHAR with readchar, which has better error
checking.
* (e7000_parse_device): Add prototype.
* (readchar): Improve doc. Handle random serial errors.
* (expect): Disable notice_quit code. It's busted. Remove
serial error handling (it's now handled in readchar). Remove
remote_debug echoing. That's handled in readchar as well.
* (e7000_parse_device): Remove serial_flag arg. It's not
necessary.
* (e7000_open): Split into two pieces. Second part is
e7000_start_remote, and is error protected. Now, when we connect
to the target, we setup the initial frame and registers so that
the user gets an immediate indication of where the target is.
* (gch): Remove debug output. That's handled by readchar.
* (e7000_read_inferior_memory): Handle errors better.
* (_initialize_remote_e7000): Get rid of `<xxx>' things from
command names. They show up when doing completion and confuse
things horribly.
* ser-e7kpc.c: Remove the last seven months of brain damage. Get
rid of the DLL's since we can access the device directly from Win32s
and Win95. Get rid of debugify crud.
* serial.c: Remove debugify cruft.
* (serial_logchar serial_log_command serial_write serial_readchar
serial_send_break serial_close): Merge common functionality into
serial_logchar. Clean up rest of routines.
* sparclet-rom.c: Disembowel. Leave only download routine.
Download routine now switches to remote target automatically.
* top.c (disconnect): Only define if SIGHUP is defined. Cleans
up MSVC/Win32 problem.
* utils.c (gdb_flush): Don't call hook unless it's for gdb_stdout
or gdb_stderr.
* config/sh/tm-sh.h: Define TARGET_SH for WinGDB.
* config/sparc/tm-sparclet.h: Remove override for prompt.
o Provide poll_quit callback to simulators
so that they can poll for SIGINT on
clueless OS's.
o Add sim_stop to simulators so that clients
can request a halt (eg gdbtk's STOP button)
Works for PPC!
o Re-arange remote-sim.c so that the
hard work is moved from gdbsim_resume()
to gdbsim_wait() (where it should be).
(INSTALL_XFORM1): Remove.
(install-only): Use $(program_transform_name) directly, rather
than using $(INSTALL_XFORM) and $(INSTALL_XFORM1).
(uninstall): Transform name.
* procfs.c: Substantial (but incomplete) changes to support
sysv4.2mp procfs as implemented in UnixWare 2.1. The procinfo
struct now has substructs like struct flt_ctl instead of
just a fltset_t and has a ctl_fd, status_fd, as_fd, and
map_fd instead of a single fd. Non-sysv4.2mp procfs models
still have the structs and multiple fds, but don't use the
entire struct and the four fds all point to the same thing.
We use PROCFS_USE_READ_WRITE to decide whether to talk to
procfs with reads/writes or use ioctl instead. We use
HAVE_MULTIPLE_PROC_FDS to determine whether procfs really has
multiple fds or not. PROC_NAME_FMT is split out into
CTL_PROC_NAME_FMT, AS_PROC_NAME_FMT, MAP_PROC_NAME_FMT,
STATUS_PROC_NAME_FMT.
(procfs_notice_signals): now a necessary wrapper around
(notice_signals): which are the new guts for noticing signals
(open_proc_file): gets a new flag arg used in sysv4.2mp to
determine whether or not to attempt to open the ctl_fd.
(procfs_read_status): new local function, reads procfs status
(procfs_write_pcwstop): new local function, writes a PCWSTOP
(procfs_write_pckill): new local function, writes a PCKILL
(unconditionally_kill_inferior): remove signo since we now
just call procfs_write_pckill().
(procfs_xfer_memory): call lseek with SEEK_SET rather than 0
(proc_iterate_over_mappings): the whole function is ifdefed
on UNIXWARE to keep things readable.
Expanded the syscall_table to include new potential sysv4.2mp
members. Note that all ifdefs of UNIXWARE should be eliminated
if possible or renamed to describe what's being selected for a
bit better. Sysv4.2mp and IRIX both have SYS_sproc so the
IRIX specific code now also checks it's not UNIXWARE.
* config/i386/tm-i386v42mp.h: also define HAVE_PSTATUS_T,
HAVE_NO_PRRUN_T, PROCFS_USE_READ_WRITE, and UNIXWARE
* config/mips/nm-irix4.h: set CTL_PROC_NAME_FMT et al to
"/debug/%d" as PROC_NAME_FMT used to be
* (hexnumlen): Add prototype. Use max, not min.
* (remote_write_bytes remote_read_bytes): Fix max packet size
calculations to properly account for packet overhead. Also handle
(probably rare) case where remote_register_buf_size isn't set.
address.
* (remote_ops extended_remote_ops remote_desc remote_write_size):
Make static.
* (remote_fetch_registers remote_write_bytes remote_read_bytes):
Record size of response to fetch registers command, use this to
limit size of memory read and write commands.
* (push_remote_target): New function to make it possible to have
another target switch to the remote target.
* target.h: Add prototype for push_remote_target.
* sh-tdep.c (sh_frame_find_saved_regs): Fix sign extension bugs
for hosts which default to unsigned chars (such as SGI's).
* (_initialize_sh_tdep): Don't set remote_write_size. It's now
handled automatically in remote.c.
* gdb.base/help.exp: update help msg for INFO ADDR; add help test
for INFO SYMBOL; add help tests for OVERLAY commands.
* gdb.base/default.exp: add tests for INFO SYMBOL command;
add tests for OVERLAY commands
* gdb.base/overlays.c overlays.exp foo.c bar.c baz.c grbx.c ovlymgr.h
d10v.ld m32r.ld d10vovly.c m32rovly.c: add test case for overlays.
* gdb.base/sigall.c: add usestubs code frag
* gdb.base/watchpoint.c: turn on complex watchpoint test for M32R.
Except for the two m32r modules, this checkin is for
overlay support.
* blockframe.c: blockvector_for_pc_sect(), block_for_pc_sect(),
find_pc_sect_function(), find_pc_sect_partial_function(): new
functions for debugging overlays; pc without section is ambiguous.
* breakpoint.[ch]: add section pointer to breakpoint struct;
add section argument to check_duplicates(); check section as well
as pc in [breakpoint_here_p(), breakpoint_inserted_here_p(),
breakpoint_thread_match(), bpstat_stop_status()];
add section argument to describe_other_breakpoints();
use INIT_SAL() macro to zero-out new sal structures;
make resolve_sal_pc() fix up the sal's section as well as its pc;
match on section + pc in clear_command() and delete_breakpoint();
account for overlay sections in insert_breakpoints(),
remove_breakpoint() and breakpoint_re_set_one();
all this to support overlays where a PC is not unique.
* exec.c: change xfer_memory() to handle overlay sections.
* findvar.c: change read_var_value() to handle overlay sections.
* frame.h: declaration for block_for_pc_sect() [blockframe.c].
* infcmd.c: jump_command() warns against jumping into an overlay
that's not in memory. Also use INIT_SAL() to initialize sals.
* infrun.c: wait_for_inferior() sets a flag to invalidate cached
overlay state information; Also use INIT_SAL() to init sals.
* m32r-rom.c: modify load routines to use LMA instead of VMA.
* m32r-stub.c: mask exit value down to 8 bits; screen out any
memory read/writes in the range 600000 to a00000, and ff680000
to ff800000 (hangs because nothing is mapped there); fix strcpy().
* maint.c: maintenance command "translate-address" supports overlays.
* minsyms.c: lookup_minimal_symbol_by_pc_sect() supports overlays.
* objfiles.[ch]: add ovly_mapped field to the obj_section struct;
this constitutes gdb's internal overlay mapping table. Add macro
ALL_OBJSECTIONS() to loop thru the obj_structs and look at overlays.
Add function find_pc_sect_section().
* printcmd.c: modify print_address_symbolic() with overlay smarts;
modify address_info() with overlay smarts; add function sym_info()
to support the INFO SYMBOL command (translate address to symbol(s));
modify disassemble_command() to work on unmapped overlays.
* source.c: use INIT_SAL() to initialize sals.
* symfile.[ch]: change generic_load() to use section's LMA address
instead of VMA address, for overlay sections.
Add numerous functions for finding a PC's section / overlay,
translating between VMA and LMA address ranges, determining if an
overlay section is mapped, etc. Add several user commands for
overlay debugging. Add support for a "generic" form of automatically
reading overlay mapping info from the inferior (based on the default
(simple) overlay manager which Cygnus provides as an example).
* symtab.[ch]: add functions find_pc_sect_symtab(),
find_pc_sect_psymtab(), find_pc_sect_psymbol(), find_pc_sect_line()
for lookup; modify lookup_symbol and decode_line_1() to use them;
modify find_function_start_sal() to account for overlay sections;
add macro INIT_SAL() for initializing struct symtab_and_line.
* target.c: fix a comment in the declaration of target_ops.
mips_skip_stub, mips_ignore_helper): New functions for dealing
with MIPS16 call/return thunks.
(mips_init_frame_pc_first): New function to implement
INIT_FRAME_PC_FIRST macro; includes code from old macro plus
new code to skip over MIPS16 thunks.
(mips_frame_chain): Skip over MIPS16 thunks.
* config/mips/tm-mips.h (mips_in_call_stub, mips_in_return_stub,
mips_skip_stub, mips_ignore_helper): Declare.
(IN_SOLIB_CALL_TRAMPOLINE, IN_SOLIB_RETURN_TRAMPOLINE,
SKIP_TRAMPOLINE_CODE, IGNORE_HELPER_CALL): New macros that invoke
the above functions.
(INIT_FRAME_PC_FIRST): Change to invoke mips_init_frame_pc.
(mips_init_frame_pc): Declare.
* infrun.c (wait_for_inferior): Use new IGNORE_HELPER_CALL macro
to decide if certain library function calls should be ignored.
or netport.
* lib/gdb.exp: Set GDB to [transform gdb] if we're using a remote host
and it's not already set.
* gdb.base/setshow.exp: Only test the run command if the target
isn't using a stub and if it supports argument passing.
* symtab.c (find_pc_symtab): change to support the case
where the objfile is reordered and contains both coff and
stabs debugging info (continue on if a psymtab isn't found).
* gdb.disasm/mn10300.exp: Run mn10300 disassembler tests.
* gdb.disasm/Makefile.in: Remove "mn10300" when cleaning.
disassembler testsuite for the mn10300.
* config/mips/tm-mips.h (REGISTER_CONVERT_TO_TYPE,
REGISTER_CONVERT_FROM_TYPE): Swap words if target, not host,
is big-endian and if registers are 32 bits.
* mips-tdep.c (mips_print_register, mips_extract_return_value,
mips_store_return_value): Fix floating-point word-order problems on
little-endian targets introduced by changes of Mar 21.
globals.
(remote_resume, remote_wait): Use them.
* d10v-tdep.c: Set the above hooks.
(tracesource): New GDB variable, controls source display in
traces.
(display_trace): Find and display source line if requested.
(trace_info): Mention empty trace buffer if appropriate.
(tdisassemble_command): Robustify argument handling.
copy first two floating point arguments to general registers, so that
MIPS16 functions will receive the arguments correctly.
(mips_print_register): Print double registers correctly on
little-endian hosts.
(mips_extract_return_value): Return double values correctly
on little-endian hosts.
* mdebugread.c (parse_procedure): Adjust address of procedure relative
to address in file descriptor record; this accounts for constant
strings that may precede functions in the text section. Remove
now-useless lowest_pdr_addr from argument list and all calls.
New UnixWare 2.1 configuration
* config/i386/i386v42mp.mt: new
* config/i386/i386v42mp.mh: new
* config/i386/tm-i386v42mp.h: new
* config/i386/nm-i386v42mp.h: new
* configure.tgt: added new entries
* configure.host: added new entries
* Makefile.in: Install gdbtcl dir instead of gdbtk.tcl.
* gdbtk.c: Added some ifdefs for Windows. Changed GDBTK_FILENAME
to GDBTK_LIBRARY, which is now a path to search.
(gdb_path_conv): New function. Convert Cygwin32 pathname to
DOS-style pathname.
* aclocal.m4, configure.in: Changes for Windows builds.
* configure: Rebuilt.
bpstat_stop_status): Do not disable watchpoints going out of scope.
(insert_breakpoints): Make sure that the current frame is valid
before calling find_frame_addr_in_frame_chain.
* top.c (setup_user_args): Handle quotes and backslashes.
(print_gdb_version): Update copyright year.
(sim_open): Return a `descriptor' as result.
(*): New argument of descriptor result from sim_open.
* remote-sim.c (gdbsim_desc): Renamed from gdbsim_open_p.
(gdbsim_open): Record result of sim_open in gdbsim_desc.
Pass argv list to sim_open, argv[0] = pseudo program name.
(*): Pass gdbsim_desc to sim_foo fns.
* configure, configure.in: Allow gdbserver to be configured
for cross-target environments.
* gdbserver/Makefile.in: Add simulator support.
* gdbserver/configure.in: Eliminate assumption that host == target.
Simplify using gdb/configure.tgt and gdb/configure.host.
Fix other minor configuration errors.
* gdbserver/low-sparc.c: Fix compile error.
* gdbserver/remote-utils.c: Eliminate assumption that registers
and addresses are four bytes. Fix minor compile errors and warnings.
* gdbserver/server.c: Rewrite numerous instances of identical code
for starting inferior processes to call new function start_inferior.
Eliminate assumption that registers and addresses are four bytes.
* gdbserver/server.h: Add missing prototypes to eliminate compiler
warnings.
* gdbserver/low-sim.c: New file to mate gdbserver with simulators.
* config/mips/vr5000.mt: Add Vr5000 simulator support to gdbserver.
* config/i386/linux.mh: Eliminate gdbserver support as a first step
in moving such support from host to target makefile fragments.
* config/i386/linux.mt: Move gdbserver support here from linux.mh.
* symtab.h (INIT_SAL): New macro to initialize symtab_and_line,
to insure consistant initialization of unused fields to zero.
* symtab.c: replace initializations of sals with new macro INIT_SAL.
* breakpoint.c: ditto.
* infrun.c: ditto.
* infcmd.c: ditto.
* source.c: add call to INIT_SAL macro.
* gdb.base/help.exp: Ditto.
* gdb.base/list.exp: Ditto.
* gdb.base/signals.exp: Ditto.
* config/monitor.exp(gdb_load): If gdb,use_standard_load is
set, use remote_ld to download the testcase instead of
the GDB loader.
instead of in the Sun specific section.
(Include Files): The GNU linker supports the N_BINCL
optimization. Clarify the N_BINCL value, and what it is used
for.
(Procedures): Document N_FUN with an empty string to mark the end
of a function.
(Typedefs): Mention that Sun compilers may use N_GSYM for a type.
(Sun Differences): Remove this node, as the information is now
elsewhere in the main document.
(Stab Section Basics): Mention that the GNU linker may optimize
stabs and remove the leading N_UNDF symbol.
(mips_initialize): LSI PMON doesn't support 'set regsize' command.
(pmon_wait): Don't need to exit and re-enter debug mode on LSI
PMON after a continue; it causes target program misbehavior.
(mips_fetch_register): Don't fetch unsupported registers; this
cuts down on wasted serial traffic.
support for WinGDB. Add mswin to configdirs.
* configure.in configure: Move calls to configure.host and
configure.tgt to the top of configure.in to allow them to set
config variables before they are referenced.
with template parameters containing `::'.
* valops.c (search_struct_field, search_struct_method):
Pass correct valaddr parameter to baseclass_offset.
Prevent gdb crashes by making sure that the virtual base pointer
from an user object still points to accessible memory.
* gdb.base/ptype.exp: Don't call "gdb_exit; gdb_start" if we're
aborting; the testsuite driver will do that for us (see
gdb_finish). Also, use gdb_test in a few more places.
* gdb.disasm/configure.in: Remove references to target makefile
frags. Use "sh3.s" as the unique filename for this directory.
* gdb.disasm/configure: Rebuilt.
Cleaning up unused crud.
* gdb.disasm/mn10200.s: New test file for mn10200 disassembler.
* gdb.disasm/mn10200.exp: Run mn10200 disassembler tests.
* gdb.disasm/Makefile.in: Remove "mn10200" when cleaning.
mn10200 disassembler testsuite.
with 0 length. A sticky tag caused this fix from
revision 2.32 date: 1996/04/29 23:25:27; author: dawn;
state: Exp; lines: +1 -0 branches: 2.32.34;
to not make it into devo.
Add prototypes for make_final_cleanup (and the other cleanup
routines.
* remote-e7000.c: Fix sync code to timeout if unable to sync.
Change sync code to report status while trying to sync-up
with hardware. Add debugging output and document.
* ser-e7kpc.c: Swap order of len & offset to match implementation.
Add debugging output and document.
* serial.c: Add debugging output.
* top.c: Add call to do_final_cleanups.
Remove conditionals preventing Win32 from getting SIGQUIT.
* utils.c: (*_cleanup): Modify cleanup routines to accept a cleanup
chain as a parameter. Extract this generic code from the cleanup
routines into separate funtions (*_my_cleanup). Keep old
functionality by passing "cleanup_chain" to the new funtions.
Define the cleanup chain "final_cleanup_chain" to be a cleanup
chain which will be executed only when gdb exits. Add functions
(*_final_cleanup) to match the original (*_cleanup) functions.
(pollquit, quit, notice_quit): Fix to read cntl-C from the
Windows message queue.
be two byte aligned. Round argument sizes up to two byte boundary.
Write out args in two byte hunks.
(mn10200_push_return_address): Implement.
* config/mn10200/tm-mn10200.h (EXTRACT_RETURN_VALUE): Abort for
structures > 8 bytes (temporary).
(STORE_RETURN_VALUE): Likewise.
(CALL_DUMMY): No longer undefine.
(USE_STRUCT_CONVENTION): Use for args > 8 bytes.
(REG_STRUCT_HAS_ADDR): Define.
Inferior function calls for the mn10200.
* config/vxworks.exp: New file.
* gdb.base/a2-run.exp: Change vxworks cases to use gdb_spawn_id.
* config/monitor.exp(gdb_target_monitor): Look for a prompt from
gdb before assuming everything worked. Send a ^C if a timeout
occurs.
Random fixes for vxworks and d10v.
(find_proc_desc): Move non-heuristic proc search code into separate
function.
(gdb_print_insn_mips): Use non-heuristic method to find procedure
descriptor, to avoid prologue examination when disassembling.
* remote-mips.c: Add support for new "lsi" target (LSI MiniRISC
aka MicroMeteor board).
(mips_exit_debug): Prevent protocol reinitialization if an error
occurs while exiting debug mode.
comments, formatting, etc. Plus other minor fixes for problems
I found during my first pass over the mn10200 port.
(mn10200_analyze_prologue): New function.
(mn10200_frame_chain, mn10200_init_extra_frame_info): Use it.
* config/mn10200/tm-mn10200.h: Lots of updates/improvements to
comments, formatting, etc. Minor fixes for problems I found during
my first pass over the mn10200 port.
(TARGET_*_BIT): Define appropriately for ints, long longs, doubles and
pointers.
(REGISTER_VIRTUAL_TYPE): Define as a long.
(EXTRACT_RETURN_VALUE): Rework to deal with long ints living
in register pairs.
(STORE_RETURN_VALUE): Similarly.
Checking in my initial changes, prologue scanning, etc.
Current gdb testsuite results:
=== gdb Summary ===
# of expected passes 3684
# of expected failures 40
# of unexpected failures 6
from being sent to 32-bit targets by masking off upper bits.
* mips-tdep.c (heuristic_proc_start): Mask off upper 32 bits
of PC on 32-bit targets.
(mips16_heuristic_proc_desc): Recognize 'addiu s1,sp,n' as a
frame setup instruction.
(mips32_heuristic_proc_desc): Fix warning found by gcc -Wall.
(mips16_skip_prologue): Recognize 'addiu s1,sp,n' as a valid
prologue instruction. Fix warnings and bugs found by gcc -Wall.
* buildsym.c (finish_block): Improve handling of overlapping blocks;
fixes problem on MIPS16 printing function arguments.
the linkage name of a die from DW_AT_MIPS_linkage_name or
DW_AT_name.
(read_func_scope, dwarf2_add_field, dwarf2_add_member_fn,
new_symbol): Use it instead of accessing DW_AT_name.
(read_partial_die): Use DW_AT_MIPS_linkage name as name of the
partial die if present.
(dwarf2_add_member_fn): Make a copy of physname on the type obstack.
exit status and/or results from the target.
* gdb.base/watchpoint.exp, gdb.base/langs.exp: Likewise.
* lib/gdb.exp: Remove old (now bogus) initialization of
noinferior, noargs, noresults and nosignals.
More cleanups for new testing framework.
* config/slite.exp: Try to connect multiple times to the board
before rebooting. Only send a "monitor run" if need_monitor_run
is set.
* gdb.base/break.exp: Don't do the "stub continue" test if
the target has gdb_stub set.
* gdb.base/callfuncs.exp: Increase the timeout.
* gdb.base/interrupt.exp: Don't even try to compile the testcase
if the target has gdb,noinferiorio set.
* gdb.base/list.exp: Increase match_max to 10000 characters.
* gdb.base/sigall.exp: Check for gdb,nosignals on the target.
* gdb.base/watchpoint.exp: Check for gdb,noinferiorio on the
target.
* lib/gdb.exp(gdb_run_cmd): Fix for targets that use stubs.
Misc fixes for sparclite and SH.
* config/slite.exp: Try to connect multiple times to the board
before rebooting. Only send a "monitor run" if need_monitor_run
is set.
* gdb.base/break.exp: Don't do the "stub continue" test if
the target has gdb_stub set.
* gdb.base/callfuncs.exp: Increase the timeout.
* gdb.base/interrupt.exp: Don't even try to compile the testcase
if the target has gdb,noinferiorio set.
* gdb.base/list.exp: Increase match_max to 10000 characters.
* gdb.base/sigall.exp: Check for gdb,nosignals on the target.
* gdb.base/watchpoint.exp: Check for gdb,noinferiorio on the
target.
* lib/gdb.exp(gdb_run_cmd): Fix for targets that use stubs.
Misc fixes for sparclite and SH.
sd and sw instructions correctly.
(heuristic_proc_start): Add support for MIPS16.
(mips16_get_imm, mips16_heuristic_proc_desc,
mips32_heuristic_proc_desc): New helper functions for
heuristic_proc_desc.
(heuristic_proc_desc): Rewrite and reorganize to support MIPS16.
(mips_push_arguments): Don't align small arguments in EABI.
(mips32_skip_prologue): Attempt to shrink code size a little.
trying to delete the copy of gdb. Catch the file delete so we
don't die if the delete fails; also, the file should be copied to
the host, not to the build.
Fixes problems when gdb isn't built with debugging.
* m32r-stub.c: New -- remote protocol support for M32R cpu.
* m32r-rom.c: Several experiments with improved download time.
* .Sanitize: Add entry for m32r-stub.c
parenthesis in case it contains multiple regexps separated
with |.
Fixes last nights native failures.
* gdb.base/watchpoint.exp: Use gdb_test.
* gdb.base/default.exp: Ditto.
More improvements.
(mips_find_saved_regs): Replace hardcoded 2's with MIPS16_INSTLEN.
(heuristic_proc_start): Recognize 'entry' pseudo-op as a start
of function on MIPS16.
(mips32_skip_prologue, mips16_skip_prologue): New helper functions
for mips_skip_prologue.
(mips_skip_prologue): Recognize both 16- and 32-bit prologues.
* gdb.base/signals.exp: Check for a gdb,nosignals feature of
the target.
* gdb.base/watchpoint.exp: Fix regexp.
Random miscellaneous fixes. Gotta dig up all these global variables someday..
close and wait commands, as the descriptor may now be
invalid. Always call "remote_close host".
Makes sure things get cleaned up if gdb dies in the middle of a test.
* config/mips/tm-mips.h (MIPS16_BIG_BREAKPOINT,
MIPS16_LITTLE_BREAKPOINT, BREAKPOINT_FROM_PC): Define.
(ABOUT_TO_RETURN): Call new function mips_about_to_return.
(mips_breakpoint_from_pc, mips_about_to_return): Declare.
* mem-break.c (memory_breakpoint_from_pc): New function.
(memory_insert_breakpoint, memory_remove_breakpoint): Use
memory_breakpoint_from_pc to determine breakpoint contents and size.
* target.h (memory_breakpoint_from_pc): Declare.
* monitor.c (monitor_insert_breakpoint): Use memory_breakpoint_from_pc
to determine size of breakpoint instruction.
* mips-tdep.c (mips32_decode_reg_save, mips16_decode_reg_save):
New helper functions for mips_find_saved_regs.
(mips_find_saved_regs): Recognize mips16 prologues.
(mips_addr_bits_remove): Strip off upper 32 bits of address
when target CPU is 32 bits but CORE_ADDR is 64 bits.
(mips_step_skips_delay): No branch delay slot on mips16.
(gdb_print_insn_mips): Disassemble mips16 code.
(mips_breakpoint_from_pc, mips_about_to_return): New functions.
(REGISTER_NAMES): Elimination registers not found on the mn10200.
(PC_REGNUM, MDR_REGNUM, PSW_REGNUM): Corresponding changes.
(LIR_REGNUM, LAR_REGNUM): Delete. They don't exist on the mn10200.
internal structure definitions with the corresponding enumeration
types from dwarf2.h. Add default cases to switches on enumerations
where appropriate.
Make quoting of string arguments in complaint messages consistent.
Check for NULL returns from DW_STRING.
(struct partial_die_info): Add sibling and has_type fields, remove
unused value field.
(DW_*): Move access macro definitions near the definition of the
attribute structure.
(struct field_info): New structure to pass information about fields
and member functions between die processing routines.
(dwarf2_build_psymtabs_hard): Set cu_header_offset.
(scan_partial_symbols): Do not enter DW_TAG_subprogram dies into
the partial symbol table if the DW_AT_*_pc attributes are missing.
Add file scope base type definitions to the partial symbol table.
Skip over child dies if the die has a sibling attribute.
(add_partial_symbol): Enter global variables with type attributes
and without location descriptors into the partial symbol table.
Store value of DW_TAG_variable dies in the partial symbol table.
Do not enter global variables into the minimal symbol table.
Add base type definitions to the partial symbol table.
(psymtab_to_symtab_1): Use dwarf2_get_pc_bounds to determine highpc.
(process_die): Move check for DW_AT_low_pc to read_func_scope.
Add a typedef symbol for base type definitions to the symbol table.
Ignore DW_TAG_inlined_subroutine tags for now.
(read_file_scope): Use dwarf2_get_pc_bounds to determine pc bounds.
(read_func_scope, read_lexical_block_scope): Use dwarf2_get_pc_bounds
to determine pc bounds, ignore dies with invalid bounds.
(dwarf2_get_pc_bounds): New routine to extract and validate the
DW_AT_*_pc attributes of a die.
(dwarf2_add_field, dwarf2_attach_fields_to_type, skip_member_fn_name,
dwarf2_add_member_fn, dwarf2_attach_fn_fields_to_type):
New functions to handle fields and member functions.
(read_structure_scope): Rewritten to use them.
(read_array_type): Renamed from dwarf_read_array_type.
Default upper array bound to describe an array with unspecified
length.
Create array types in backwards order, as dwarf2 puts out the array
dimensions from left to right.
(read_subroutine_type): Handle DW_TAG_unspecified_parameters,
DW_AT_artificial and DW_AT_prototyped.
(read_base_type): Make an unsigned type for DW_ATE_boolean.
Pass objfile to dwarf_base_type.
(read_partial_die): Use read_attribute to read in the attributes.
Handle DW_AT_sibling and DW_AT_type.
Follow references when determining DW_AT_name and DW_AT_external
attributes of the die.
Validate DW_AT_*_pc attributes.
(read_full_die): Use read_attribute to read in the attributes.
(read_attribute): New function to read an attribute described
by an abbreviated attribute.
(new_symbol): Relocate symbol value for DW_TAG_label with baseaddr.
Do not set SYMBOL_VALUE_ADDRESS for DW_TAG_subprogram,
SYMBOL_BLOCK_VALUE for the symbol will be set later by finish_block.
Change symbol class for global variables with a zero valued location
descriptor to LOC_UNRESOLVED.
Handle DW_AT_const_value attributes for DW_TAG_variable,
DW_TAG_formal_parameter and DW_TAG_enumerator.
Build a typedef symbol for DW_TAG_base_type.
(dwarf2_const_value): New routine to copy a constant value from an
attribute to a symbol.
(dwarf_base_type): Use passed in objfile, not current_objfile
when calling dwarf2_fundamental_type.
(dump_die): Use DW_* accessor macros to access values of attributes.
(decode_locdesc): Handle DW_OP_plus_uconst.
Start mn10200 gdb port by adding copies of mn10300
target-specific files with all instances of mn10300 changed
to mn10200 to start with.
* mn10200-tdep.c: new
* config/mn10200/tm-mn10200.h: new, REGISTER_SIZE is 24 bits not 32,
SP_REGNUM and FP_REGNUM are different, also no lar or lir.
* config/mn10200/mn10200.mt: new
* configure.tgt: add mn10200 entry
Start mn10200 gdb port by adding copies of mn10300
target-specific files with all instances of mn10300 changed
to mn10200 to start with.
* mn10200-tdep.c: new
* config/mn10200/tm-mn10200.h: new, REGISTER_SIZE is 24 bits not 32,
SP_REGNUM and FP_REGNUM are different, also no lar or lir.
* config/mn10200/mn10200.mt: new
* mn10300-tdep.c: wrote/fixed implementations of
mn10300_frame_chain, mn10300_init_extra_frame_info,
mn10300_frame_saved_pc
* config/mn10300/tm-mn10300.h: redefine INIT_EXTRA_FRAME_INFO
and INIT_FRAME_PC macros
Backtracing starting to work correctly.
* mn10300-tdep.c: made a lot more generic, ripping out code
from copied target (no more mn10300_scan_prologue,
init_extra_frame_info, and mn10300_fix_call_dummy calls)
* config/mn10300/tm-mn10300.h: undefine INIT_EXTRA_FRAME_INFO
and INIT_FRAME_PC macros
FP_REGISTER_BYTES to compute offsets into the saved frame,
since it fails for SPARC targets configured without any
FP regs. Instead, use DUMMY_STACK_REG_BUF_SIZE.
--------------------------------------------------------------------
ENABLE_OBS, and THREAD_DB_OBS. These are consolidated into LIBS
and CONFIG_OBS.
* configure configure.in: Clean up test cases around thread support.
start-sanitize-v850
* configure.tgt (v850-*-*): Include v850ice.o and v850.lib if
host is Windows.
end-sanitize-v850
* c-valprint.c ch-valprint.c cp-valprint.c eval.c expprint.c
printcmd.c valops.c value.h values.c: Add bfd_section arg to
value_at and value_at_lazy.
* coffread.c dbxread.c elfread.c mdebugread.c minsyms.c symtab.h:
Add bfd_section arg to prim_record_minimal_symbol_and_info.
* corefile.c gdbcore.h printcmd.c valops.c: Use read_memory_section
instead of read_memory. It takes a bfd_section arg.
* coffread.c dbxread.c elfread.c gdb-stabs.h objfiles.h: Remove
unnecessary cast for assignment of struct dbx_symfile_info.
Struct objfile now uses a real pointer instead of PTR for this
element.
* dbxread.c (dbx_symfile_init): Stash bfd section pointers for
text, data and bss into dbx_symfile_info.
* exec.c (xfer_memory): Handle transfers for user-specified
sections.
* findvar.c (read_var_value locate_var_value): Copy bfd section
from the symbol to the value.
* gdb-stabs.h: Add section pointers for text, data and bss
sections.
* maint.c (translate address command): Add test code for overlay
address translation.
* printcmd.c (do_examine do_one_display): Now takes a bfd section
arg.
* (print_formatted x_command): Record current section along with
current address for repeated commands.
* sparc-nat.c (fetch_inferior_registers): Change
target_xfer_memory to target_{read write}_memory to allow changes
to target_xfer_memory interface for section info.
* symmisc.c (dump_msymbols print_symbol): Print section
assocaited with symbol.
* symtab.c (fixup_symbol_section): New routine to
add section info to symbols returned by lookup_symbol.
* symtab.h (struct general_symbol_info): Add bfd section to
symbols.
* target.c target.h (target_xfer_memory): Add bfd section to
args.
* (target_read_memory_section): New routine to read data from a
specific section.
* (target_memory_bfd_section): New global variable to pass bfd
section in to targets.
* valarith.c (value_add value_addr value_array): Preserve bfd
section when computing new value.
* value.h (struct value): Add bfd section to values.
* values.c (allocate_value value_copy): Initialize/preserve bfd
section.
* (unpack_double): Clean up _MSC_VER conditionals to remove
duplicate code.
start-sanitize-v850
* v850ice.c: New module to support communication with NEC's
PC-based ICE.
* config/v850/tm-v850.h (REGISTER_NAMES): Replace sp, gp, fp, and
ep names with rxx names. sp and fp are renamed via a different
mechanism.
end-sanitize-v850
ENABLE_OBS, and THREAD_DB_OBS. These are consolidated into LIBS
and CONFIG_OBS.
* configure configure.in: Clean up test cases around thread support.
start-sanitize-v850
* configure.tgt (v850-*-*): Include v850ice.o and v850.lib if
host is Windows.
end-sanitize-v850
* c-valprint.c ch-valprint.c cp-valprint.c eval.c expprint.c
printcmd.c valops.c value.h values.c: Add bfd_section arg to
value_at and value_at_lazy.
* coffread.c dbxread.c elfread.c mdebugread.c minsyms.c symtab.h:
Add bfd_section arg to prim_record_minimal_symbol_and_info.
* corefile.c gdbcore.h printcmd.c valops.c: Use read_memory_section
instead of read_memory. It takes a bfd_section arg.
* coffread.c dbxread.c elfread.c gdb-stabs.h objfiles.h: Remove
unnecessary cast for assignment of struct dbx_symfile_info.
Struct objfile now uses a real pointer instead of PTR for this
element.
* dbxread.c (dbx_symfile_init): Stash bfd section pointers for
text, data and bss into dbx_symfile_info.
* exec.c (xfer_memory): Handle transfers for user-specified
sections.
* findvar.c (read_var_value locate_var_value): Copy bfd section
from the symbol to the value.
* gdb-stabs.h: Add section pointers for text, data and bss
sections.
* maint.c (translate address command): Add test code for overlay
address translation.
* printcmd.c (do_examine do_one_display): Now takes a bfd section
arg.
* (print_formatted x_command): Record current section along with
current address for repeated commands.
* sparc-nat.c (fetch_inferior_registers): Change
target_xfer_memory to target_{read write}_memory to allow changes
to target_xfer_memory interface for section info.
* symmisc.c (dump_msymbols print_symbol): Print section
assocaited with symbol.
* symtab.c (fixup_symbol_section): New routine to
add section info to symbols returned by lookup_symbol.
* symtab.h (struct general_symbol_info): Add bfd section to
symbols.
* target.c target.h (target_xfer_memory): Add bfd section to
args.
* (target_read_memory_section): New routine to read data from a
specific section.
* (target_memory_bfd_section): New global variable to pass bfd
section in to targets.
* valarith.c (value_add value_addr value_array): Preserve bfd
section when computing new value.
* value.h (struct value): Add bfd section to values.
* values.c (allocate_value value_copy): Initialize/preserve bfd
section.
* (unpack_double): Clean up _MSC_VER conditionals to remove
duplicate code.
start-sanitize-v850
* v850ice.c: New module to support communication with NEC's
PC-based ICE.
* config/v850/tm-v850.h (REGISTER_NAMES): Replace sp, gp, fp, and
ep names with rxx names. sp and fp are renamed via a different
mechanism.
end-sanitize-v850
* mn10300-tdep.c (mn10300_push_arguments): rewrote
also removed code elsewhere that made use of RP_REGNUM
* config/mn10300/tm-mn10300.h: ripped out RP_REGNUM, V0_REGNUM,
ARG0_REGNUM, ARGLAST_REGNUM (all not appropriate for mn10300
arch.), redefined SAVED_PC_AFTER_CALL, EXTRACT_RETURN_VALUE,
EXTRACT_STRUCT_VALUE_ADDRESS, STORE_RETURN_VALUE
with separate LITTLE_BREAKPOINT and BIG_BREAKPOINT definitions;
this fixes problem with setting breakpoints in little-endian
programs in the simulator.
and SIGQUIT values; it messed up non-MIPS targets.
* config/mips/tm-mips.h: Undefine BREAKPOINT, replace
with separate LITTLE_BREAKPOINT and BIG_BREAKPOINT definitions;
this fixes problem with setting breakpoints in little-endian
programs in the simulator.
of 64-bit values on Windows.
* config/i386/xm-windows.h: Leave CC_HAS_LONG_LONG defined,
undefine PRINTF_HAS_LONG_LONG, so that 64-bit values will
be printed without loss of upper bits.
* sparclet-stub.c: a few clean-ups, removing dead code
* config/sparc/tm-sparclet.h: make registers ASR15, ASR19 invisible
(they're not useful, you can't change, write or even read them)
* sparclet-rom.c: Remove includes of Unix system files.
Add function "sparclet_supply_register" so that parse_register_dump
will not seg-fault by calling a null function pointer.
Remove XMODEM support (unfinished work?).
Remove flag "MO_HANDLE_NL", so monitor's output can be read by humans.
Add fill command.
Remove colon from getreg.resp_delim so PSR register will work.
Remove pointer to sparclet_load (downloading SREC's doesn't work).
Null out local register names for %g0, all %cc and all %asr regs,
since the monitor can't report them. Will return zero instead.
* sparclet-stub.c: New -- remote protocol support for sparclet CPU.
* config/sparc/tm-sparclet.h: Re-arrange REGISTER_NAMES:
Add back %g0 and %psr, add %cc coprocessor regs, add %asr regs.
Adjust NUM_REGS and REGISTER_BYTES accordingly
from want and want_nopc.
(want_h8300s, want_nopc_h8300s): New variables for H8/S register
lists.
(e7000_fetch_registers): Use H8/300H or H8/S register list string
as needed.
(e7000_wait): Likewise.
Bringing over a change from hmse-h8s-beta branch.
