From: "H. Peter Anvin" <hpa@zytor.com>
To: Ferenc Wagner <wferi@niif.hu>
Cc: Alain Knaff <alain@knaff.lu>,
linux-kernel@vger.kernel.org,
Phillip Lougher <phillip@lougher.demon.co.uk>
Subject: Re: kernel decompressor interface
Date: Tue, 30 Mar 2010 10:45:55 -0700 [thread overview]
Message-ID: <4BB238D3.7050808@zytor.com> (raw)
In-Reply-To: <877hotu6x9.fsf@tac.ki.iif.hu>
On 03/30/2010 09:10 AM, Ferenc Wagner wrote:
> Hi,
>
> While working with SquashFS code recently, I got the impression that the
> current decompress_fn interface isn't best suited for general use: it
> rules out real scatter/gather operation, which -- one hopes -- is a
> general feature of stream decompressors. For example, if one has to
> decompress data from a series of buffer_heads into a bunch of (cache)
> pages (typical operation in compressed file systems), the inflate
> interface in zlib.h provides the possibility of changing input and
> output buffer addresses, but decompress_fn does not, necessitating extra
> memory copying. On the other hand, the latter is admittedly simpler.
>
> However, I feel like their advantages could be combined by making the
> fill() and flush() callbacks able to change the buffer addresses, like
> int fill(void **, unsigned int) for example. Even better, some context
> could also be passed around, so that they can easily carry private state.
>
> A cursory look at the LZMA decompressor implementation suggests that
> this would be straightforward change (although I haven't done it). Do
> you think that
> 1. it could be done for other decompressors (lzo, bunzip) as well, and
> 2. it should be done in any form (interface replacement or extension)?
I think the event-based interfaces in zlib.h are pretty good, myself,
and since they are so widely deployed they are also widely mimiced,
which makes it a good interface to rely on. A simpler "one shot"
interface can then be implemented as a wrapper (since it just means:
create the context, point it at the two buffers, and then tell it to to
run to the end.)
LZO has some super-compact variants that don't support start and stop
operation, but even the "full featured" versions are small enough, that
I don't think it matters for something as large as the Linux kernel. It
mattered to me when I was dealing with self-decompressing objects on the
order of a few tens of kilobytes.
-hpa
next parent reply other threads:[~2010-03-30 17:49 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <877hotu6x9.fsf@tac.ki.iif.hu>
2010-03-30 17:45 ` H. Peter Anvin [this message]
2010-03-31 0:25 ` kernel decompressor interface Phillip Lougher
2010-03-31 17:45 ` H. Peter Anvin
2010-04-01 12:11 ` Ferenc Wagner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4BB238D3.7050808@zytor.com \
--to=hpa@zytor.com \
--cc=alain@knaff.lu \
--cc=linux-kernel@vger.kernel.org \
--cc=phillip@lougher.demon.co.uk \
--cc=wferi@niif.hu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.