Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 20 Sep 2000 23:26:18 +0100 (BST)
From:      Doug Rabson <dfr@nlsystems.com>
To:        Matthew Jacob <mjacob@feral.com>
Cc:        Mike Smith <msmith@freebsd.org>, cvs-committers@freebsd.org, cvs-all@freebsd.org
Subject:   Re: cvs commit: src/sys/boot/alpha/common main.c 
Message-ID:  <Pine.BSF.4.21.0009202324580.86297-100000@salmon.nlsystems.com>
In-Reply-To: <Pine.BSF.4.21.0009192254410.2289-100000@beppo.feral.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 19 Sep 2000, Matthew Jacob wrote:

> On Tue, 19 Sep 2000, Mike Smith wrote:
> 
> > > > >   Steal 512KB more from system memory for heap instead of 256KB. We died
> > > > >   without the extra space. What a pity.
> > > > 
> > > > Is there some way we can do this in boot1?  The loader isn't going to get 
> > > > a lot smaller anytime soon. 8(
> > > 
> > > Uh, no. boot1 is only just a smidgen less than 40KB. John Baldwin didn't have
> > > this fix things for him either.
> > 
> > Sorry, I meant "can't we have boot1 steal the memory rather than having 
> > the loader do it"?
> 
> This makes no sense. It's the loader that needs it for heap storage, not
> boot1.
> 
> > 
> > > The whole dance about memory allocation for the loader and the kernel isn't
> > > done by any stretch of the imagination. At any rate, up to 1MB for a loader
> > > that does prelinking and loading for a kernel seems not too unreasonable. If
> > > you could reclaim that later, that would be nice, but that would actually
> > > involve some real architecting instead of the current evolutionary methods.
> > 
> > The only real challenge would seem to be the trampoline necessary in
> > between leaving the loader and calling the kernel proper.  I'm sure one 
> > of our burgeoning Alpha hackers could cough something like that up in a 
> > day or so. 8)
> 
> No. The current method for stealing memory leaves no trace of memory as it's
> subtracted from the hw page list. What has to happen instead is that when one
> releases all boot resources you have to be able to call back into the loader
> to reclaim memory which was otherwise not on anyone else's list.
> 
> If you're doing this for alpha, you'll want to understand the implications
> for other architectures.

The pages which loader borrows for its nefarious mallocing purposes are
not subtracted from the system memory map. They are reused by the kernel
after it starts. The original 256k which the firmware allocated is not
reused though.

-- 
Doug Rabson				Mail:  dfr@nlsystems.com
Nonlinear Systems Ltd.			Phone: +44 20 8348 3944




To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0009202324580.86297-100000>