(CGEN_ARCH_SCM): New variable.
* cgen-engine.h (EXTRACT_[ML]SB0_{INT,UINT}): New macros.
(EXTRACT_INT,EXTRACT_UINT): New macros.
(SEM_SEM_ARG): New macro.
(SEM_NEXT_VPC): New arg `pc'.
* cgen-sim.h (EXTRACT_SIGNED,EXTRACT_UNSIGNED): Delete.
(sim_disassemble_insn): Update prototype.
* cgen-trace.c (current_insn,insn_fields): New static locals.
(trace_insn): Set them.
* cgen-utils.scm: #include cgen-engine.h.
(sim_disassemble_insn): New arg insn_fields.
Handle variable length insns.
* genmloop.sh: Only emit pbb decls if -pbb.
(${cpu}_scache_lookup): New arg `vpc'.
(scache support): Fetch pc before entering loop.
* elf.c (_bfd_elf_link_hash_newfunc): Clear it.
* elflink.h (elf_gc_propagate_vtable_entries_used): Copy it, and
respect it as an upper bound on what memory to reference.
(elf_gc_smash_unused_vtentry_relocs): Likewise.
(elf_gc_record_vtentry): Handle as-yet undefined vtables. Set and
update vtable_entries_size appropriately.
bfds; they have no "file" so we fake it.
* bfd-in2.h: add bfd_make_writable and bfd_make_readable (regen)
* libbfd.c (bfd_write): Allow writing to BFD_IN_MEMORY bfds by
resizing the memory buffer.
* opncls.c (bfd_make_writable): New function; lets you write a
bfd_create'd bfd to an in-memory buffer.
(bfd_make_readable): New function; lets you read back that buffer
as if it were from bfd_openr. Used to build generated archive
members (winsup's ld's dlltool)
* deffile.h: add hint member.
* pe-dll.c (pe_dll_generate_implib): New function with helpers;
generates the import library directly from the export table.
(fill_edata): remember the actual hint for the import library.
element of the gotentry list for local symbols too.
(elf64_alpha_relocate_section): Unify local and global handling.
Assert that gotent->use_count > 0.
Changes to account for name change from cygwin32 to cygwin and
clean up Win32-related ifdefs.
* configure.tgt: check for cygwin* instead of cygwin32.
New cygwin gdb_target variable loses the "32".
* configure.host: check for cygwin* instead of cygwin32.
New cygwin gdb_host variable loses the "32".
* configure.in: test __CYGWIN__ instead of __CYGWIN32__,
rename gdb_cv_os_cygwin32 variable to drop the "32". Call
AM_EXEEXT instead of AC_EXEEXT since that isn't in a released
autoconf yet.
* configure: regenerate.
* main.c: drop "32" from cygwin_ funcs, include sys/cygwin.h
where
cygwin path conv protos live, instead of adding a proto here
for
them here.
* {main.c, ser-tcp.c, ser-unix.c, top.c}: check __CYGWIN__
instead of __CYGWIN32__.
* source.c: thoughout, check _WIN32 instead of WIN32.
* config/i386/cygwin32.mh: delete.
* config/i386/cygwin.mh: new file, was cygwin32.mh.
* config/i386/cygwin32.mt: delete.
* config/i386/cygwin.mt: new file, was cygwin32.mt.
* config/i386/tm-cygwin32.h: delete.
* config/i386/tm-cygwin.h: new file, was tm-cygwin32.h.
* config/i386/xm-cygwin32.h: delete.
* config/i386/xm-cygwin.h: new file, was xm-cygwin32.h.
* config/i386/xm-windows.h: #include xm-cygwin.h now.
* config/powerpc/cygwin32.mh: delete.
* config/powerpc/cygwin.mh: new file, was cygwin32.mh.
* config/powerpc/cygwin32.mt: delete.
* config/powerpc/cygwin.mt: new file, was cygwin32.mt.
* config/powerpc/tm-cygwin32.h: delete.
* config/powerpc/tm-cygwin.h: new file, was tm-cygwin32.h.
* config/powerpc/xm-cygwin32.h: delete.
* config/powerpc/xm-cygwin.h: new file, was xm-cygwin32.h.
* rdi-share/aclocal.m4: regenerate with aclocal.
* rdi-share/configure: regenerate with autoconf.
* rdi-share/{host.h, hostchan.c, hostchan.h, serdrv.c,
* serpardr.c,
unixcomm.c}: check __CYGWIN__ instead of __CYGWIN32__.
multithread debugging.
(remote_get_threadlist) : get a partial list of threads
(remote_threadlist_iterator) : Step through all the threads
(init_remote_threadtests) : Optional builtin unit test commands.
* thread.c (bind_target_thread_vector) : Implementa a more dynamic
way of accessing target specific thread info functions than
FIND_NEW_THREADS.
(target_thread_info) : Function to get extended thread information.
* gdbthread.h : Export internal data structures corresponding to
external detailed thread info response. This is more like a 'ps'
command than what might be expected of host based threads. This
is for embedded systems.
