Skip site navigation (1)Skip section navigation (2)
Date:      13 Mar 2002 11:16:37 +0100
From:      Dag-Erling Smorgrav <des@ofug.org>
To:        Josef Karthauser <joe@tao.org.uk>
Cc:        arch@freebsd.org, peter@freebsd.org, jake@freebsd.org
Subject:   Re: dumpsys() rewrite
Message-ID:  <xzp7kogbwmy.fsf@flood.ping.uio.no>
In-Reply-To: <20020313093805.GA29679@genius.tao.org.uk>
References:  <xzp3cz5v7rp.fsf@flood.ping.uio.no> <20020313093805.GA29679@genius.tao.org.uk>

next in thread | previous in thread | raw e-mail | index | archive | help
Josef Karthauser <joe@tao.org.uk> writes:
> It would be fantastic to be able to compress the memory image as it is
> dumped to disk.  That would speed it up a bit in most cases.  If each
> dump_write reads memory directly rather than going through a common
> function this is never going to be easy to do.

dumpsys() controls what is written and when.  It is perfetly
conceivable for dumpsys() to do the same kind of hole compression that
savecore(1) does, but it would be *much* slower and it would have to
record information about every hole to disk, which would make
savecore(1) slower as well.  Currently, most drivers write 64 kB at a
time when dumping, and savecore(1) reads 1 MB at a time.  To perform
hole compression, dumpsys() would have to scan through memory word by
word *and* potentially write in far smaller increments than 64 kB.
Remember the bad old days when it wrote only a sector at a time?

DES
-- 
Dag-Erling Smorgrav - des@ofug.org

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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?xzp7kogbwmy.fsf>