linux-kernel - Re: [PATCH] xfs: make xfs_log_iovec independent from xfs_log

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: <20240625112438.1925184-1-alexjlzheng@tencent.com>
Date: Tue, 25 Jun 2024 19:24:38 +0800
From: alexjlzheng@...il.com
To: hch@...radead.org
Cc: alexjlzheng@...il.com,
	alexjlzheng@...cent.com,
	chandan.babu@...cle.com,
	david@...morbit.com,
	djwong@...nel.org,
	linux-kernel@...r.kernel.org,
	linux-xfs@...r.kernel.org
Subject: Re: [PATCH] xfs: make xfs_log_iovec independent from xfs_log_vec and release it early

On Tue, 25 Jun 2024 04:14:55 -0700, Christoph Hellwig wrote:
> On Tue, Jun 25, 2024 at 12:06:14AM +0800, Jinliang Zheng wrote:
> > xfs_log_iovec is where all log data is saved. Compared to xfs_log_vec itself,
> > xfs_log_iovec occupies a larger memory space.
> > 
> > When their memory spaces are allocated together, the memory occupied by
> > xfs_log_iovec can only be released after iclog is written to the disk log
> > space. But when xfs_log_iovec is written to iclog, its existence becomes
> > meaningless, because a copy of its content is already saved in iclog at this
> > time.
> > 
> > And if they are separated, we can release its memory when the data in
> > xfs_log_iovec is written to iclog. The interval between these two time points
> > is not too small.
> > 
> > Since xfs_log_iovec is the area that currently uses the most memory in
> > xfs_log_vec, this means that we have released quite a lot of memory. Freeing
> > memory that occupies a larger size earlier means smaller memory usage.
> 
> This all needs to go into the commit log.  Preferably including the
> actual quantity of memory saved for a useful workload.

I am sorry, but I didn't get your point. May I ask if you could clarify your
viewpoint more clearly?

Thank you very much. :)
Jinliang Zheng