linux-kernel - Re: [PATCH] btrfs: index buffer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20250513171039.GC9140@twin.jikos.cz>
Date: Tue, 13 May 2025 19:10:40 +0200
From: David Sterba <dsterba@...e.cz>
To: Daniel Vacek <neelx@...e.com>
Cc: dsterba@...e.cz, Chris Mason <clm@...com>,
	Josef Bacik <josef@...icpanda.com>, David Sterba <dsterba@...e.com>,
	linux-btrfs@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] btrfs: index buffer_tree using node size

On Tue, May 13, 2025 at 09:56:19AM +0200, Daniel Vacek wrote:
> On Mon, 12 May 2025 at 22:20, David Sterba <dsterba@...e.cz> wrote:
> >
> > On Mon, May 12, 2025 at 07:23:20PM +0200, Daniel Vacek wrote:
> > > So far we are deriving the buffer tree index using the sector size. But each
> > > extent buffer covers multiple sectors. This makes the buffer tree rather sparse.
> > >
> > > For example the typical and quite common configuration uses sector size of 4KiB
> > > and node size of 16KiB. In this case it means the buffer tree is using up to
> > > the maximum of 25% of it's slots. Or in other words at least 75% of the tree
> > > slots are wasted as never used.
> > >
> > > We can score significant memory savings on the required tree nodes by indexing
> > > the tree using the node size instead. As a result far less slots are wasted
> > > and the tree can now use up to all 100% of it's slots this way.
> >
> > This looks interesting. Is there a way to get xarray stats? I don't see
> > anything in the public API, e.g. depth, fanout, slack per level. For
> > debugging purposes we can put it to sysfs or as syslog message,
> > eventually as non-debugging output to commit_stats.
> 
> I'm using a python script in crash (even live on my laptop). I believe
> you could do the same in dragon. Though that's not the runtime stats
> you described. And I don't really think it's worth it.

How come you don't think it's worth it? You claim some numbers and we
don't have a way to verify that or gather on various setups or
workloads. I'd be interested in the numbers also to better understand
how xarray performs with the extent buffers but I don't now how to write
the analysis scripts in any of the tools, nor have time for that.