This is the mail archive of the
gdb@sourceware.org
mailing list for the GDB project.
Re: GDB default breakpoint instruction for SuperH
- From: Mark Kettenis <mark dot kettenis at xs4all dot nl>
- To: zou dot min at renesas dot com
- Cc: gdb at sources dot redhat dot com
- Date: Sat, 17 Jun 2006 21:09:07 +0200 (CEST)
- Subject: Re: GDB default breakpoint instruction for SuperH
- References: <447261B2.6090805@renesas.com>
> Date: Tue, 23 May 2006 09:13:22 +0800
> From: Min ZOU <zou.min@renesas.com>
>
> Hi,
>
> I looked at this GDB (v6.4) source file: gdb/config/sh/tm-sh.h . In
> side, it specifies the default breakpoint instruction for SH, which
> causes GDB to insert 0xc320 (trapa #32) into target when breakpoint is
> set. See below:
>
> #define DEPRECATED_BIG_REMOTE_BREAKPOINT { 0xc3, 0x20 } /* Used in
> remote.c *
> /
> #define DEPRECATED_LITTLE_REMOTE_BREAKPOINT { 0x20, 0xc3 } /* Used in
> remote.c *
> /
>
> I wonder why by default 0x20 is used instead of other number without
> knowing the target SuperH processor type ?
>
> Thanks!
I don't really know the SuperH instruction set, so I can't answer your
question. But these defines are for code that should be removed from
gdb altogether. If we do that, gdb will use 0xc3, 0xc3 which is the
same for both little and big endian (see sh-tdep.c).
Can you verify whether that works correcltly?
Mark