Re: Recent mm-system changes broke dosemu, was Re: dosemu & newer kernels

Steven N. Hirsch (hirsch@uvm-gen.emba.uvm.edu)
Fri, 19 Apr 1996 18:05:43 -0400

Uwe Bonnes wrote:
>
> >>>>> "Steven" == Steven N Hirsch <hirsch@uvm-gen.emba.uvm.edu> writes:
>
> Steven> As of 1.3.89 (or so) my old, trusty, non-emumodule dosemu-0.60.4
> Steven> stopped working :-(. Strace suggests that this is the culprit:
>
> Steven> ... <strace deleted>
> Steven>
>
> Steven> Question is, why does shmat fail? I'm far from an expert on the
> Steven> Linux mm system, but perhaps a problem now exists. Again,
> Steven> perhaps dosemu's behavior was taking advantage of something that
> Steven> had actually been broken until recently :-)..
>
> Steven> Any takers? Which is it?
>
> This strace is very instructional. But I think it will be usefull, if you
> too report where in dosemu ( with what program) the error occured. And log a
> debug output "dos -o /tmp/dos.debug -D +E" of your program. Scan that file
> for "unsucess", "fail" etc, to see where the error occured. This should
> point you to the sourcecode where the error occured. Perhaps you should post
> your results to dosemu-devel@ednet.ns.ca or at least to
> linux-msdos@rutgers.vger.edu?
>

Oh, sorry that I was not specific about this, but this happens _instantly_ on
attempting to run dosemu! I cannot get past the linux shell prompt :-(.

> Recently (1.3.79) there have been some not well documented changes to the
> area of mmap, which broke dosemu too. By debugging Hans Lermen found out,
> that pages have to be touched before they can be mmap(). Perhaps this
> shmat-thing is related. Can those who applied these changes to the mm-system
> enlighten us, what else changed?
>
> Uwe Bonnes bon@elektron.ikp.physik.th-darmstadt.de
>
> Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt
> --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

Yes, indeed, inquiring minds want to know. I did try building the latest
pre-release of dosemu (0.63.1.8), but it's hopeless. Programs using VESA SVGA
modes that used to work flawlessly are completely broken. I'm afraid that this
stuff is beyond my abilities (and time) at the moment for debugging.

- Steve

-- 
___________________________________________________________
|Steven N. Hirsch	"Anything worth doing is worth     |
|University of Vermont	 overdoing.." - Hunter S. Thompson |
|Computer Science / EE                                     |
------------------------------------------------------------