bdm and multiprocessor machines

[Ulrich.Topp at de.abb.com]

Hello all,

let me give my question to Pavel to you all, perhaps there is someone out,
who solved already my problem.

Would be fine, if we could get help ...

Sincerly,
Ulrich

---------------------------------------------------------------------------------------------------------
[my first mail to pavel pisa (pisa at cmp.felk.cvut.cz) ]

Hello Pavel,

we are using your bdm-driver for linux here very successful and indeed, it
makes it possible to develop SW for our m68332-board.

But now we got a wonderful new PC with two CPUs.
We have SuSE-Linux 7.2 installed with kernel 2.4, BUT now your driver
doesn't work any longer.

We compiled it, and it is loadable ('insmod-able') but if we try to
download our application, it crashes.

Is it a known problem and do you have some hints/patches/workarounds
already?

It would be great to hear from you,
Ulrich Topp

PS: This is what the log-file says:

bdmlib_load: program name o-optimize/pecosdemo.nxe
bdmlib_set_sys_reg error: reg 0xf cont 0x5 error 617
bdmlib_set_sys_reg error: reg 0xe cont 0x5 error 617

PPS: And this is what the system-messages are:

Oct 18 16:49:13 linux kernel: BDM init_module
Oct 18 16:49:13 linux kernel:    $RCSfile: bdm.c,v $
Oct 18 16:49:13 linux kernel:    $Revision: 2.13 $
Oct 18 16:49:13 linux kernel:    $Date: 1998/11/19 13:15:56 $
Oct 18 16:49:13 linux kernel:    Version PD ICD +ISP
Oct 18 16:49:13 linux kernel:    Compiled at Oct  1 2001 14:55:49
Oct 18 16:49:13 linux kernel: BDM driver succesfully registered !!
Oct 18 16:58:23 linux kernel: bdm_ioctl minor=0 cmd=11 (BDM_SENSECABLE)
arg=0x0
Oct 18 16:58:23 linux kernel: bdm_ioctl: setting bdm_sense to 0
Oct 18 16:58:36 linux kernel: bdm_ioctl minor=0 cmd=6 (BDM_GET_STATUS)
arg=0x0
Oct 18 16:58:36 linux kernel: bdm_pd_getstatus minor=0 res=0x18
Oct 18 16:58:36 linux kernel: bdm_write minor 0 len 6
Oct 18 16:58:36 linux kernel: bdm_pd_getstatus minor=0 res=0x18
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: writing  248f
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: read 1ffff bits 0
Oct 18 16:58:36 linux kernel: bdm_write sent:  248f received: 1ffff
Oct 18 16:58:36 linux kernel: bdm_pd_getstatus minor=0 res=0x18
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: writing     0
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: read 1ffff bits 0
Oct 18 16:58:36 linux kernel: bdm_write sent:  0 received: 1ffff
Oct 18 16:58:36 linux kernel: bdm_write minor 0 len 6
Oct 18 16:58:36 linux kernel: bdm_pd_getstatus minor=0 res=0x18
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: writing  248e
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: read 1ffff bits 0
Oct 18 16:58:36 linux kernel: bdm_write sent:  248e received: 1ffff
Oct 18 16:58:36 linux kernel: bdm_pd_getstatus minor=0 res=0x18
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: writing     0
Oct 18 16:58:36 linux kernel: bdm_pd_ser_clk: read 1ffff bits 0
Oct 18 16:58:36 linux kernel: bdm_write sent:  0 received: 1ffff


------------------------------------------------------------------------
[pavels answer]

Hello Ulrich,

I would like to have your problems with SMP machine ;-)
I cannot test driver on the SMP box. (We have some SMP Linuxes
at University, but these machines are used as servers and vision
workstation and I have no right to test my kernel stuff on them).

I have some experience with multi-open and parallel access
drivers (uLan, DMSDOS), but I left BDM driver as simple as possible
=> only single open (there is no need for multiple GDB to one CPU)
and only blocking calls => I think, that SMP is no problem for BDM.
Speed could be bigger one, but ISA in/out frequency is constant
for long time now.

I see other problem. You have no response from target (1ffff).
There could be two reasons.
- Weak PC parallel port hardware
  or target wiring problem.
- BPP parallel port was not handled in my older patches

Change was so small that I have not renamed archive,
but I have updated it about three weeks ago.
I hope, that it solves your problem. I am sending last driver
source as well.

Best wishes

                Pavel Pisa
        e-mail: pisa at cmp.felk.cvut.cz
        www:    http://cmp.felk.cvut.cz/~pisa
        work:   http://www.pikron.com

-----------------------------------------------------------------------------
[my reply]

Hello Pavel,

thanks for your answer.
We checked the cable as well as the parallel port of the PC. Under Linux a
HP printer was perfectly running and the cable is in perfect use from
another PC.

What do you mean with 'BPP' (Is it  a typo for 'EPP' ?)
>From when are the newest sources of the driver? (I found on your server a
version from 09-Okt-01)

Sincerly,
Ulrich