Re: Julian's ps broken on 111h2 (fwd)

[roemer@informatik.uni-frankfurt.de]

i've notices similar problems with the 'top'-Port from the mintutl-package.

First of all the first line says: ........ (idle).
This is because top assumes tha idle time is counted on pid 0. But _MiNT_ has
pid 0 (is this beacuse of the hack that forces init to have pid 1 ?)

Trying to track down the other problems with the strange process-times I ran
into another problem:
gdb can't display variable contents any more.
I've inversigated this a bit and traced the following kernel-calls:

all on the open process handle for the debugged process
Fseek(0, 1)  (don't know what this one is worth for)
Fseek(<address>, 0) (the address doesn't corresopnd to the variable address in
							memory)
read(<variable-address>, length) (to copy the variable contents from the process
											 image to the variable address)
proc_read(same as above)

But proc_read only gives 0's.

If I try to access the variable directly via the gdx-x-command, I get something
like:
ptrace: can't access ....

As I haven't got the gdb-source, I can't investigate any further with those
stange seek/read things, but I'll try to find out a bit about the ptrace stuff.

	** Dominik **

-- 
Dominik Strasser                            | Phone:  +49 89 636-43691
SIEMENS NIXDORF Information Systems         | Fax:    +49 89 636-40140
SU BS2000 SD225 (C/C++ Compilers)           | E-Mail:Dominik.Strasser@mch.sni.de
Otto-Hahn-Ring 6, D-81739 Muenchen, Germany |