<br><br><div class="gmail_quote">On Fri, Aug 19, 2011 at 19:21, Greg KH <span dir="ltr"><<a href="mailto:greg@kroah.com">greg@kroah.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On Fri, Aug 19, 2011 at 11:20:44AM +0300, Giorgis Georgakoudis wrote:<br>
> Hello all,<br>
><br>
> I would like to ask for some pointers regarding <br>
> porting Linux kernel multiprocessor support in <br>
> new platforms. More specifically, what are the<br>
> HW requirements that a platform must have to<br>
> enable Linux multiprocessing<br>
<br>
</div>It must have more than one processor and both must be able to see the<br>
same system "image" at the same time.</blockquote><div><br></div><div>In fact, I'm trying to support multiprocessing in an experimental architecture</div><div>of Microblazes on an FPGA. There isn't (HW) cache coherency and RAM is</div>
<div>accessed through per processor dedicated buses, thus shared access locking </div><div>must be realized either in a software manner or through a dedicated hardware mutex.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
</blockquote><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im"><br>
> and some info<br>
> on kernel internals with respect to the booting<br>
> process, scheduling/migration etc.<br>
><br>
> I've been looking around for relevant documenta-<br>
> tion but found little to none.<br>
<br>
</div>Is the code not enough documentation for you for doing your porting<br>
work? What specifically do you find lacking?<br>
<br>
greg k-h<br></blockquote><div><br></div><div>No doubt that the ultimate documentation is the code but still allow me</div><div>to consider that a higher level overview, abstracting code intricacies is</div><div>helpful.</div>
<div><br></div><div>Best regards,</div><div><br></div><div>Giorgis</div></div><br>