Re: EMX problem

Path	csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!news.dougwise.org!gegeweb.org!eternal-september.org!feeder.eternal-september.org!.POSTED!not-for-mail
From	Ilya Zakharevich <nospam-abuse@ilyaz.org>
Newsgroups	comp.os.os2.programmer.misc
Subject	Re: EMX problem
Date	Tue, 5 Apr 2011 18:55:16 +0000 (UTC)
Organization	U.C. Berkeley Math. Department.
Lines	36
Message-ID	<slrnipmpck.p8d.nospam-abuse@powdermilk.math.berkeley.edu> (permalink)
References	<fV45K0OBJxbE-pn2-Po0DVV9WtXII@localhost> <fV45K0OBJxbE-pn2-UW9Gwdgwbd4f@localhost> <slrnipjvr6.kmq.nospam-abuse@powdermilk.math.berkeley.edu> <fV45K0OBJxbE-pn2-T1if6jK8e8Ni@localhost> <slrnipke37.ljg.nospam-abuse@powdermilk.math.berkeley.edu> <fV45K0OBJxbE-pn2-1bCYdMdR4XW2@localhost>
Injection-Date	Tue, 5 Apr 2011 18:55:16 +0000 (UTC)
Injection-Info	mx02.eternal-september.org; posting-host="LlLjF8LevGdIZWbTvv+SXA"; logging-data="4802"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+wK+Zfa7WHIzll7b6rksfo"
User-Agent	slrn/0.9.8.1pl1 (Linux)
Cancel-Lock	sha1:Xb3mmIDrolxtPmzUgJkLlk6WMA0=
Xref	x330-a1.tempe.blueboxinc.net comp.os.os2.programmer.misc:174

Show key headers only | View raw

On 2011-04-05, Dave Saville <dave@invalid.invalid> wrote:
> On Mon, 4 Apr 2011 21:30:15 UTC, Ilya Zakharevich 
><nospam-abuse@ilyaz.org> wrote:
>
>> On 2011-04-04, Dave Saville <dave@invalid.invalid> wrote:
>> >> a) Update the max command line to the correct value, 32K.
>> 
>> >> b) Read what the crash errorlog tells you.
>> 
>> > Not easy when there is none - Just a SYS* message.
>> 
>> It shoulda have been quite enough.  Do the corresponding
>> 
>>     help sys*
>> 
>> (actually, one can drop SYS as well ;-).
>> 
>> Ilya
>
> Which boiled down to "you have accessed RAM that does not belong to 
> you" Really helpful :-)

Stupid me; yes, stack overflows do not ALWAYS produce a meaningful
SYS* code if one allocates a large enough chunk on stack!

A similar topic: brain-damaged layout of "native" OS/2 executables:
while executables produced with EMX's bind.exe have a large no-read
no-write gap between stack and other segments, it is (at least)
non-obvious how to do it with link386 (or ilink).

Therefore, if one allocates 1MB on stack, one would just jump from one
segment to another; if the other one is writable, one may get into a
lot of trouble...

Ilya

Back to comp.os.os2.programmer.misc | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

Re: EMX problem "Dave Saville" <dave@invalid.invalid> - 2011-04-03 20:31 +0000
  Re: EMX problem Ilya Zakharevich <nospam-abuse@ilyaz.org> - 2011-04-04 17:27 +0000
    Re: EMX problem "Dave Saville" <dave@invalid.invalid> - 2011-04-04 19:01 +0000
      Re: EMX problem Ilya Zakharevich <nospam-abuse@ilyaz.org> - 2011-04-04 21:30 +0000
        Re: EMX problem "Dave Saville" <dave@invalid.invalid> - 2011-04-05 13:43 +0000
          Re: EMX problem Ilya Zakharevich <nospam-abuse@ilyaz.org> - 2011-04-05 18:55 +0000
      Re: EMX problem "A.D. Fundum" <what.ever@neverm.ind> - 2011-04-06 13:24 +0200

csiph-web