Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Message-ID: <1148ee1105050215253ad27030@mail.gmail.com>
Date: Tue, 3 May 2005 00:25:52 +0200
From: Utku Ozcan <utku DOT ozcan AT gmail DOT com>
Reply-To: Utku Ozcan <utku DOT ozcan AT gmail DOT com>
To: cygwin AT cygwin DOT com
Subject: side effects of Cygwin's maximum memory test
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit

I *think* that the test below, which tests memory allocation limit of
Cygwin *might* produce problems in Windows XP:

http://sources.redhat.com/cygwin/cygwin-ug-net/setup-maxmem.html

In this page, I have compiled the C code, and after having run the
compiled executable, Windows XP gave suddenly a warning that virtual
memory setting has been changed (I think, that repeated malloc() calls
in this code somehow change the virtual memory settings in Windows
XP).

After that, I have observed harddisk related problems which I cannot
produce them repeated. Sometimes during power-on, harddisk locked, PC
never started (harddisk lamp lighted continuously). It helped to boot
the computer several times so that it can run. The last symptom was,
that PC could not boot itself and operating system dumped "ntfs
corrupt or damaged" message.

I have looked at internet, but could not find any information that I
could correlate the problems with it. The only thing I could find is
that Virtual Memory setting was user defined min: 768 MB, max: 1536 MB
(was not in automatical control of OS any more, compiled C code above
in the link must have forced the operating system to change it).

Then I have run defragmentation. The results can be viewed at:
http://www.geocities.com/utku_ozcan/vol_c_defrag.JPG

The only thing I can figure out is, that before and after the
defragmentation, harddisk has big holes in between. This cannot be
pagefile.sys, because it can be as much as 1,5 GB (max: 1536 MB). Big
holes between defragmented areas seems to be in 5-10 GB regions.

I know, I need proof, but sorry, not capable of producing it. Maybe my
harddisk had big fragment distances in Gigabytes before, OS cannot
boot because HD consume in that case excessive amount of time to read
fragmented sectors.

This problem occurred maybe when I run the C code above (fragmented
harddisk <-> manual change of virtual memory because of Cygwin test C
code).

Just a thought.

Utku

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/