* monitor/mon-command.c: Per PR192 the RTEMS monitor makes everything
lowercase. The capture engine need to set triggers or watches on task
with uppercase names.
Also stop the monitor repeating command when enter is pressed.
* untar/Makefile.am: Use 'PREINSTALL_FILES ='.
* monitor/Makefile.am: Use 'PREINSTALL_FILES ='.
* shell/Makefile.am: Use 'PREINSTALL_FILES ='.
* devnull/Makefile.am: Use 'PREINSTALL_FILES ='.
* dumpbuf/Makefile.am: Use 'PREINSTALL_FILES ='.
* mw-fb/Makefile.am: Use 'PREINSTALL_FILES ='.
* stackchk/Makefile.am: Use 'PREINSTALL_FILES ='.
* rtmonuse/Makefile.am: Use 'PREINSTALL_FILES ='.
* cpuuse/Makefile.am: Use 'PREINSTALL_FILES ='.
* rootfs/Makefile.am: Use 'PREINSTALL_FILES ='.
* configure.in, rootfs/Makefile.am: Added conditional to make sure
rootfs does not get built with networking is disabled. Also
added include of compile.am to rootfs/Makefile.am.
* wrapup/Makefile.am: Added rootfs conditionally if networking enabled.
* Added support for populating the initial "root" filesystem
with information obtained via the DHCP response.
* rootfs: New directory.
* rootfs/.cvsignore, rootfs/Makefile.am, rootfs/mkrootfs.c,
rootfs/mkrootfs.h: New files.
* configure.in, Makefile.am: Modified to reflect addition.
* monitor/mon-command.c: Add support for partial command matching.
The monitor used to have this functionality before it was overhauled
to support addition of user commands.
* shell/telnetd.c, shell/telnetd.h: Moved to
libnetworking/rtems_servers so the network stack to address network
depenendency.
* shell/Makefile.am: Modified to reflect above.
* monitor/mon-prmisc.c: Correct print line.
* shell/Makefile.am: Added new file telnetd.c.
* shell/telnetd.c, shell/telnetd.h, shell/pty.c: New files.
* shell/shell.c, shell/cmds.c, shell/shell.h: Numerous improvments:
- The shell_init has a new parameter 'forever' because in
/dev/console you need that this process runs forever but in
tcp/ip not. (respawn?)
- A new task for every session opened trought tcp/ip telnet client.
(the chargen,daytime and more are possible of implementation but
I ask me if they are necesary)
- Exit from the session delete the task and when the client fails too.
- More cmds have been implemented. (very reduced version of these)
umask, chmod, id, whoami, rm, cat, ...
- A reduced line edit has been implemented.
Ctrl-C abort the input,
Ctrl-d in the first position gives EOF (logout).
'\b' and DEL makes the rubout operation.
I think that readline() for every session spents a lot of resources.
* devnull/devnull.c (null_initialize): Added check to
ensure that the driver is only initialized once. Otherwise,
it will abort when the device /dev/null is present.
(null_write): Do not dereference a NULL pointer.
* devnull: New addition. Moved from libcpu/sh.
* devnull/devnull.c, devnull/devnull.h, devnull/.cvsignore,
devnull/Makefile.am: New files.
* Makefile.am, configure.in, wrapup/Makefile.am: Modified to reflect
addition of above.
* Added generic Micro FrameBuffer interface for MicroWindows.
This interface allows MicroWindows to under RTEMS. A sample
driver has been developed for the pc386 BSP. See
pc386/fb_vga.c as a sample.
* Added Uniform Input Device interface for MicroWindows.
See PC386 bsp for sample drivers for mouse and keyboard (console).
* mw-bf: New directory.
* Makefile.am, configure.in, wrapup/Makefile.am: Account for mw-fb.
* mw-fb/Makefile.am: New file.
* mw-fb/mw_fb.c: New file.
* mw-fb/mw_fb.h: New file.
* mw-fb/mw_uid.c: New file.
* mw-fb/mw_uid.h: New file.
a BSP (c4xsim) supporting the simulator included with gdb. This port
was done by Joel Sherrill and Jennifer Averett of OAR Corporation.
Also included with this port is a space/time optimization to eliminate
FP context switch management on CPUs without hardware or software FP.
An issue with this port was that sizeof(unsigned32) = sizeof(unsigned8)
on this CPU. This required addressing alignment checks and assumptions
as well as fixing code that assumed sizeof(unsigned32) == 4.
that breaks when the target has 16 bit address space. One of the H8
multilibs is a 16-bit address space CPU. When a real attempt is
made to support this CPU model, the code that assumes an address
is 32 bits will have to change. This constant is probably not
flagging all impacted code.
to:
I have also added the ability to register and unregister commands. This
allows me to create a set of monitor commands for the network stack plus
basic memory dump/patch commands (needs a working probe interface). I
will also look at a basic ls/cd/rm/mv/cp command set at some stage.
I have been thinking about changing the monitor in the future to more
like a light weight RTEMS shell, `eshell' for embedded shell. This is a
story for another day but is a process or getting the commands to map to
the filesystem (eg, major=commands, minor=command) and supporting an
environment. The filesystem provide a structure for the commands.
Joel Sherrill