Re: 3.3G RAM limit?

[Mouse <mouse%Rodents-Montreal.ORG@localhost>]

>> I have a fuzzy memory that there's some kind of limit that leads to
>> i386 topping out at three-point-something gigs of RAM.  Could
>> someone remind me of details?
> It's got a single address space for user and kernel.

Okay, that's one limit, but that doesn't explain the BIOS reporting
only 3327M.  Of course, I don't expect people here to know the
intricacies of Dell's BIOS, but the fuzzy memory indicated there was
some hardware basis for this.  Perhaps that memory just got corrupted.

> I can't remember if that impacts our memory size reporting (likely
> not) but you will need to run in PAE mode if you want to address
> anywhere near the full 4GB from non-kernel code.

Well, cpuctl does report that the CPU is PAE-capable.  I'd probably
have to backport something, since the only PAE I see in this release is
in Xen config files; I'll have to investigate and decide whether that's
true, and if so whether it's worth the effort.

But that sounds like a per-process VA limit.  That's of less interest
to me than 18.75% of the RAM apparently being completely unavailable
regardless of how many different processes are involved.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B