GDB using Angel error on an OKI board

GDB using Angel error on an OKI board

[Charles-Henri Balet]

Hello,

I'm using an OKI developpement board with the OKI ML671000 processor, this
is provide with an
ARM SDT 2.50 software and an ARM Angel 1.20 in the flash memory on the
board, when I use
the ARM Debugger with an angel link, all work fine in 9600 or 19200, the
message in the console
box at the start is the next :

Angel Debug Monitor V1.20 (ARM Ltd unreleased) for ML671000 CPU Board(OKI)
Built for ARM7TDMI Serial(x1), IRQ
Unreleased build Mar 14 2000
Serial Rate:   9600

all operation with memory, registers ... work fine with this debugger.

My problem is with arm-elf-gdb 5.1.1, when I start the software, the board
don't response any time
the next lines show how I start the link with the Angel on the OKI board :

$ arm-elf-gdb
GNU gdb 5.1.1
Copyright 2002 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain
conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for details.
This GDB was configured as "--host=i686-pc-cygwin --target=arm-elf".
(gdb) set remotebaud 9600
(gdb) target rdi com1
RDI_open: undefined error message, should reset target
RDI_open failed

(gdb) target rdi com1
Unexpected read on channel 4, length 192                                //
for each reset, I obtain this message
Unexpected read on channel 4, length 192
Unexpected read on channel 4, length 192

the only message back are this line that appear on each time I press on the
reset button.

more info, when I start other command like this :

I use cygwin under windows 2000 pro, the answer of the command "uname -a" is
:

CYGWIN_NT-5.0 CHARLES-HENRI 1.3.9(0.51/3/2) 2002-01-21 12:48 i686 unknow

the answer of the two command "arm-elf-gdb -v" and "arm-elf-gdb -nw" is the
same is:
GNU gdb 5.1.1
Copyright 2002 Free ....
This GDB was configured as "--host=i686-pc-cygwin --target=arm-elf"

many thanks for your help

have a nice day

balet charles-henri

Re: GDB using Angel error on an OKI board

[Grant Edwards]

On Sun, Feb 24, 2002 at 08:37:36PM +0100, Charles-Henri Balet wrote:
> Hello,
> 
> I'm using an OKI developpement board with the OKI ML671000
> processor, this is provide with an ARM SDT 2.50 software and an
> ARM Angel 1.20 in the flash memory on the board, when I use the
> ARM Debugger with an angel link, all work fine in 9600 or
> 19200, the message in the console box at the start is the next:

> (gdb) target rdi com1
> RDI_open: undefined error message, should reset target
> RDI_open failed

At this point things are probably already broken, so I don't
know if the messages below mean much.  ADP (the protocol used
by Angel) has a _lot_ of state to it, and restting the target
without restarting GDB never works (at least not for me).
Restarting GDB without resetting the target usually works, but
not always.

  1) Reset the board.
  2) Start GDB.
  3) Do the command "maint rdilogen on"
  4) Do the command "target rdi com1"
  5) Exit GDB

That should create a file named "rdi.log" with the initial ADP
packets logged to it.  Post that file and we'll see what we can
see.

-- 
Grant Edwards
grante@visi.com

Re: GDB using Angel error on an OKI board

[Grant Edwards]

On Wed, Apr 24, 2002 at 07:50:23PM +0200, Charles-Henri Balet wrote:

>>> I'm using an OKI developpement board with the OKI ML671000
>>> processor, this is provide with an ARM SDT 2.50 software and an
>>> ARM Angel 1.20 in the flash memory on the board
[...]
>>> (gdb) target rdi com1
>>> RDI_open: undefined error message, should reset target
>>> RDI_open failed
[...]
>>   1) Reset the board.
>>   2) Start GDB.
>>   3) Do the command "maint rdilogen on"
>>   4) Do the command "target rdi com1"
>>   5) Exit GDB

> tx: [T=0 L=36] 03 01 00 01 05 00 03 00 00 00 00 00 ff ff ff ff ff ff ff ff
> 01 00 00 00 00 c0 00 00 01 00 00 00 80 25 00 00
> R=00030005 H->T CI_HBOOT:  ADP_ParamNegotiate 00000001 0000c000 00000001 00002580
> rx: [T=0 L=36] 03 02 01 01 05 00 03 80 00 00 00 00 ff ff ff ff ff ff ff ff
> 00 00 00 00 01 00 00 00 00 c0 00 00 80 25 00 00
> R=80030005 H<-T CI_HBOOT:  ADP_ParamNegotiate 00000000 00000001 0000c000 00002580
> tx: [T=0 L=20] 03 02 00 01 06 00 03 00 00 00 00 00 ff ff ff ff ff ff ff ff
> R=00030006 H->T CI_HBOOT:  ADP_LinkCheck

The link check stuff has caused problems for me in the past
with other ADP devices.  Try doing a "set rdiheartbeat off"
before you do the "target" command.

-- 
Grant Edwards
grante@visi.com

Re: GDB using Angel error on an OKI board

[Grant Edwards]

> I have tried this command "set rdiheatbeat off" but the result
> is the same, with the "maint rdilogen on" as :

> tx: [T=0 L=36] 03 01 00 01 05 00 03 00 00 00 00 00 ff ff ff ff ff ff ff ff 01 00 00 00 00 c0 00 00 01 00 00 00 80 25 00 00
> R=00030005 H->T CI_HBOOT:  ADP_ParamNegotiate 00000001 0000c000 00000001 00002580
> rx: [T=0 L=36] 03 02 01 01 05 00 03 80 00 00 00 00 ff ff ff ff ff ff ff ff 00 00 00 00 01 00 00 00 00 c0 00 00 80 25 00 00
> R=80030005 H<-T CI_HBOOT:  ADP_ParamNegotiate 00000000 00000001 0000c000 00002580
> tx: [T=0 L=20] 03 02 00 01 06 00 03 00 00 00 00 00 ff ff ff ff ff ff ff ff 
> R=00030006 H->T CI_HBOOT:  ADP_LinkCheck

Ah yes.  That link check command isn't sent by the heartbeat
feature -- it's part of the serial-port negotiation sequence.
It's possible for both ends to switch baud rates after the
initial parmeter negotiation, so a link check is sent to make
sure both ends are still talking.

The board isn't responding to the link check command.  It is
supposed to reply with another link check frame.  [At least
that is what happens with the ADP devices I have.]

I suppose you could go into the rdi code and try to skip over
that step in the start-up process to see if the board handles
the rest of the ADP commands correctly.

Or, you could try to capture the sequence used by the debugger
that works.  

[Please cc: the mailing list, since it is the collective memory
of the gdb community.  We certainly don't want to depend upon
either my memory or my record-keeping skills.]

-- 
Grant Edwards
grante@visi.com

Re: GDB using Angel error on an OKI board

[Grant Edwards]

> I have obtain the serial TX and RX of the port for the ANGEL
> ARM SDT 2.50, the code of start the sending and receiving are
> the next :

> 000000  1c 14 01 c5 c4 01 15 01 ff ff ff ff ff ff ff ff 85 49 c7 47 1d 1c 08 c5
> 000018  c4 04 4b be c7 38 e1 10 53 74 1d 1c 24 03 c6 c5 01 05 03 ff ff ff ff ff
> 000030  ff ff ff 01 c0 01 80 25 1b 53 f9 16 d3 1d 1c 14 03 c7 c6 01 06 03 ff ff
[...]

That looks like ADP framing, 0x1c starts a frame, 0x1d ends a
frame, and 0x1b is the escape character (toggles bit 4 in the
next byte).  But, what's inside the frame doesn't match up with
the ADP frame format.  Parts of it look OK, but there are bytes
missing.  It looks like whatever you used to capture the data
has discarded all of the bytes with a value of 0x00.

-- 
Grant Edwards
grante@visi.com

Re: GDB using Angel error on an OKI board

[Grant Edwards]

On Mon, Feb 25, 2002 at 10:59:29PM +0100, Charles-Henri Balet wrote:

> I have modified the source code of the ardi.c and put the delay time to 0.5
> sec. and all work fine.

Interesting.  Should this be increased "permanently?" or should
it be made configurable?

-- 
Grant Edwards
grante@visi.com

Re: GDB using Angel error on an OKI board

[Charles-Henri Balet]

Hi,

I have modified the source code of the ardi.c and put the delay time to 0.5
sec. and all work fine.

thanks for your help

balet charles-henri

/*
* 960430 KWelton
*
* There is a race in the renegotiation protocol: the
* target has to have had time to load new config before
* we send the link check packet - insert a deliberate
* pause (100ms) to give the target some time
*/
Adp_delay(500000);