<HTML><BODY><span>I have many megabytes of RAM, and want to set aside a large portion for use as a RAM disk. I've looked about a bit on the message list, and couldn't find the answers I need, and some info I did found seemed contradictory - that, and I suspect things have changed a bit in the last 4 years anyway!<br>
<br>
>From what I've read, it seems I don't really want to use just the IMFS, as it uses the C heap; I'd like to keep the disk ram usage separate from my app's heap, out of concern for fragmentation. </span><span>Looking at http://www.rtems.org/wiki/index.php/File_Systems, I think I have figured out the basics, but then come my questions:<br>
<br>
1. It seems </span><span>I want to use RFS, is that correct? How do I configure it? (I could not find the configuration documentation).<br>
<br>
2. RFS then needs a block device, specifically the RAM Disk, right? How do I configure that?<br>
<br>
3. Is the Block Device Cache required in this setup, and if so, how is that configured?<br>
<br>
4. Is there a way to pre-populate the filesystem with a set of files? In other words, is there a reasonable way to bind a "disk image" into the final output of my build process, such that when my app is loaded and begins execution, it can mount this filesystem and use the files it finds there?<br>
<br>
What I'm looking for is a nice "how-to". Given that our project team will have several RTEMS-based systems in it, and I'm the first one out the gate, I'll become a local resource for the others. Therefore, I'll be putting together one for our use here, and will post it to the RTEMS wiki for others to use too.<br>
<br>
Thanks,<br>
-Bob<br>
<br>
<div class="mirapoint-signature">
<pre><br>
</pre>
</div>
</span></BODY></HTML>