This is the mail archive of the gdb-patches@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFA/m32r] m32r-linux support


Hello,

I revised the patch of m32r-linux support.

Thanks to Mark Kettenis' advice, I removed many unnecessary codes
from it and simplified it. But it still contains some stuff
to support older versions of Linux because Linux/M32R does not
support PTRACE_GETREGS.

I also removed the gdbserver support from it. I'll post it
as another patch after this patch is accepted.

OK to commit?

2004-08-20    Kei Sakamoto  <sakamoto.kei@renesas.com>

    Add m32r-linux support.
    * configure.tgt: Add m32r*-*-linux*.
    * Makefile.in (ALLDEPFILES): Add m32r-tdep.c, m32r-linux-nat.c and
    m32r-linux-tdep.c.
    (m32r-linux-nat.o, m32r-linux-tdep.o): New dependencies.
    * m32r-linux-nat.c, m32r-linux-tdep.c, config/m32r/linux.mh,
    config/m32r/linux.mt, config/m32r/m32r.mh, config/m32r/nm-linux.h,
    config/m32r/tm-linux.h: New files.

Kei Sakamoto said:

> Hello, Mark,
> 
> Thank you for your reply. It is very useful.
> 
> There is the summer holidays in Japan next week. So I'll modify my
> patch and post it after that.
> 
> Kei Sakamoto
> 
> Mark Kettenis said:
> >    From: "Kei Sakamoto" <sakamoto.kei@renesas.com>
> >    Date: Thu, 5 Aug 2004 14:11:07 +0900
> >
> >    Hello,
> >
> > Hi Kei,
> >
> > First can you seperate out the gdbserver support, and submit that as a
> > seperate patch?  Daniel will have to review that part of the code.
> > Making it seperate will make it more likely that the code can be
> > checked in soon.
> >
> >    The attached patch adds m32r-linux support.
> >    The M32R architecture is not added to the official Linux kernel yet.
> >    But Linux on M32R is already available.
> >    http://www.linux-m32r.org/
> >
> > So Linux/M32R is a brand new port?  Great, but in that case I have
> > some comments.  It looks like you based your GDB port on the
> > Linux/i386 stuff.  That's great, but you must realize that part of the
> > i386 code that you copied and modified is only there to support older
> > versions of Linux.  I hope it isn't necessary for Linux/M32R and would
> > like you to remove it before this gets committed.  I've got a few
> > other comments too.  Here we go...
> >
> >    diff -u -r1.144 configure.tgt=0A=
> >    --- configure.tgt 26 Jun 2004 10:06:34 -0000 1.144=0A=
> >    +++ configure.tgt 5 Aug 2004 04:48:33 -0000=0A=
> >    @@ -102,7 +103,10 @@
> >    ;;
> >     ia64*-*-*) gdb_target=ia64 ;;
> >
> >    -m32r-*-*) gdb_target=m32r ;;
> >    +m32r*-*-elf*) gdb_target=m32r ;;
> >    +m32r*-*-linux*) gdb_target=linux
> >    + build_gdbserver=3Dyes
> >    + ;;
> >
> >     m68hc11*-*-*|m6811*-*-*) gdb_target=m68hc11 ;;
> >
> > It might be a good idea to keep m32r-*-* as the generic M32R target.
> > Just put the m32r-*-* entry in front of it.
> >
> >    diff -N m32r-linux-nat.c
> >    --- /dev/null 1 Jan 1970 00:00:00 -0000
> >    +++ m32r-linux-nat.c 5 Aug 2004 04:48:33 -0000
> >    +/* The register sets used in GNU/Linux ELF core-dumps are identical to
> >    +   the register sets in `struct user' that is used for a.out
> >    +   core-dumps, and is also used by `ptrace'.  The corresponding types
> >    +   are `elf_gregset_t' for the general-purpose registers (with
> >    +   `elf_greg_t' the type of a single GP register) and `elf_fpregset_t'
> >    +   for the floating-point registers.
> >    +
> >    +   Those types used to be available under the names `gregset_t' and
> >    +   `fpregset_t' too, and this file used those names in the past.  But
> >    +   those names are now used for the register sets used in the
> >    +   `mcontext_t' type, and have a different size and layout.  */
> >    +
> >    +/* Mapping between the general-purpose registers in `struct user'
> >    +   format and GDB's register array layout.  */
> >    +static int regmap[] = {
> >    +  4, 5, 6, 7, 0, 1, 2, 8,
> >    +  9, 10, 11, 12, 13, 24, 25, 23,
> >    +  19, 31, 26, 23, 20, 30, 16, 15,
> >    +  32
> >    +};
> >
> > Does Linux/M32R support a.out?  That would really surprise me.  So I
> > don't think you need this stuff.
> >
> >    +#define M32R_NUM_GREGS 25=0A=
> >
> > I think that one really belongs in m32r-tdep.h.
> >
> >    +/* Doee apply to the corresponding SET requests as well.  *
> >    +#define GETREGS_SUPPLIES(regno) (0 <=3D (regno) && (regno) <=
> M32R_NUM_GREGS)
> >    +#define GETFPREGS_SUPPLIES(regno) 0
> >
> > Does the M32R support floating point registers at all?  If not, you
> > could rip out all the FPREGS stuff.
> >
> >    +/* Does the current host support the GETREGS request?  */
> >    +int have_ptrace_getregs =
> >    +#ifdef HAVE_PTRACE_GETREGS
> >    +  1
> >    +#else
> >    +  0
> >    +#endif
> >    +  ;
> >
> > I really hope that all released M32R kernels support the
> > PTRACE_GETREGS request.  In that case this should not be necessary.
> >
> >
> >    +/* Support for the user struct.  */
> >    +
> >    +/* Return the address of register REGNUM.  BLOCKEND is the value of
> >    +   u.u_ar0, which should point to the registers.  */
> >    +
> >    +CORE_ADDR
> >    +register_u_addr (CORE_ADDR blockend, int regnum)
> >    +{
> >    +  return (blockend + 4 * regmap[regnum]);
> >    +}
> >    +
> >    +/* Return the size of the user struct.  */
> >    +
> >    +int
> >    +kernel_u_size (void)
> >    +{
> >    +  return (sizeof (struct user));
> >    +}
> >
> > The `user struct' is a wart from ancient UNIX.  There really is no
> > point in supporting it on modern versions of Linux.  If Linux/M32R is
> > ELF-only and fully supports PTRACE_GETREGS, this can go too.
> >
> >    +#ifdef HAVE_PTRACE_GETREGS
> >    +
> >    +/* Fetch all general-purpose registers from process/thread TID and
> >    +   store their values in GDB's register array.  */
> >    +
> >    +static void
> >    +fetch_regs (int tid)
> >    +{
> >    +  elf_gregset_t regs;
> >    +
> >
> > This code in #ifdef HAVE_PTRACE_GETREGS can be simplified considerably
> > if you assume that PTRACE_GETREGS works.  Move the code to
> > fetch_inferior_registers() and store_inferior_registers() below.
> >
> >    +/* Transferring arbitrary registers between GDB and inferior.  */
> >    +
> >    +/* Check if register REGNO in the child process is accessible.
> >    +   If we are accessing registers directly via the U area, only the
> >    +   general-purpose registers are available.
> >    +   All registers should be accessible if we have GETREGS support.  */
> >    +
> >    +int
> >    +cannot_fetch_register (int regno)
> >    +{
> >    +  gdb_assert (regno >= 0 && regno < NUM_REGS);
> >    +  return (!have_ptrace_getregs && regmap[regno] == -1);
> >    +}
> >    +
> >    +int
> >    +cannot_store_register (int regno)
> >    +{
> >    +  gdb_assert (regno >= 0 && regno < NUM_REGS);
> >    +  return (!have_ptrace_getregs && regmap[regno] == -1);
> >    +}
> >
> > These functions should no longer be necessary.
> >
> >    +/* Accessing registers through the U area, one at a time.  */
> >    +
> >    +/* Fetch one register.  */
> >    +
> >    +static void
> >    +fetch_register (int regno)
> >    +{
> >    +  int tid;
> >    +  int val;
> >    +
> >    +  gdb_assert (!have_ptrace_getregs);
> >    +  if (cannot_fetch_register (regno))
> >    +    {
> >    +      regcache_raw_supply (current_regcache, regno, NULL);
> >    +      return;
> >    +    }=
> >
> > This can go too; no U area anymore :-).
> >
> >    diff -N config/m32r/linux.mh
> >    --- /dev/null 1 Jan 1970 00:00:00 -0000
> >    +++ config/m32r/linux.mh 5 Aug 2004 04:48:34 -0000
> >    @@ -0,0 +1,9 @@
> >    +# Host: M32R based machine running GNU/Linux
> >    +
> >    +NAT_FILE= nm-linux.h
> >    +NATDEPFILES= infptrace.o inftarg.o fork-child.o corelow.o \
> >    + core-aout.o m32r-linux-nat.o linux-proc.o gcore.o \
> >    + proc-service.o thread-db.o lin-lwp.o linux-nat.o
> >    +
> >    +LOADLIBES=3D -ldl -rdynamic
> >
> > core-aout.o shouldn't be necessary.
> >
> >    diff -N config/m32r/m32r.mh
> >
> > Please remove this file.
> >
> >    diff -u -r1.7 m32r.mt
> >    --- config/m32r/m32r.mt 16 Oct 2003 02:36:39 -0000 1.7
> >    +++ config/m32r/m32r.mt 5 Aug 2004 04:48:34 -0000
> >    @@ -1,4 +1,5 @@
> >     # Target: Renesas m32r processor
> >     TDEPFILES= m32r-tdep.o monitor.o m32r-rom.o dsrec.o remote-m32r-sdi.o
> >    +TM_FILE= tm-m32r.h
> >     SIM_OBS = remote-sim.o
> >     SIM = ../sim/m32r/libsim.a
> >
> > This shouldn't be necessary.  See below.
> >
> >    diff -N config/m32r/nm-linux.h
> >    --- /dev/null 1 Jan 1970 00:00:00 -0000
> >    +++ config/m32r/nm-linux.h 5 Aug 2004 04:48:34 -0000
> >
> >    [snip]
> >
> >    +/* Support for the user area.  */
> >    +
> >    +/* Return the size of the user struct.  */
> >    +extern int kernel_u_size (void);
> >    +#define KERNEL_U_SIZE kernel_u_size()
> >    +
> >    +/* This is the amount to subtract from u.u_ar0
> >    +   to get the offset in the core file of the register values.  */
> >    +#define KERNEL_U_ADDR 0x0
> >    +
> >    +/* Offset of the registers within the user area.  */
> >    +#define U_REGS_OFFSET 0
> >    +
> >    +extern CORE_ADDR register_u_addr (CORE_ADDR blockend, int regnum);
> >    +#define REGISTER_U_ADDR(addr, blockend, regnum) \
> >    +  (addr) = register_u_addr (blockend, regnum)
> >
> > This user area stuff can go too.
> >
> >    diff -N config/m32r/tm-m32r.h
> >
> > This file contains nothing useful.  Please remove and adjust tm-linux.h.
> >
> > I hope you can make the changes.  The simplifications will make
> > maintaining the code much easier.

Attachment: m32r-linux.patch
Description: Binary data


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]