linux-kernel - Re: [PATCH 0/3] dm-verity: optionally use tasklets in dm-verity

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: <YtrTI/CJMoLihA/1@infradead.org>
Date:   Fri, 22 Jul 2022 09:41:07 -0700
From:   Christoph Hellwig <hch@...radead.org>
To:     Nathan Huckleberry <nhuck@...gle.com>
Cc:     linux-kernel@...r.kernel.org, dm-devel@...hat.com,
        Alasdair Kergon <agk@...hat.com>,
        Mike Snitzer <snitzer@...nel.org>,
        Eric Biggers <ebiggers@...nel.org>,
        Sami Tolvanen <samitolvanen@...gle.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Subject: Re: [PATCH 0/3] dm-verity: optionally use tasklets in dm-verity

We've been tying to kill off task lets for about 15 years.  I don't
think adding new users will make you a whole lot of friends..

On Fri, Jul 22, 2022 at 09:38:20AM +0000, Nathan Huckleberry wrote:
> Using tasklets for disk verification can reduce IO latency.  When there are
> accelerated hash instructions it is often better to compute the hash immediately
> using a tasklet rather than deferring verification to a work-queue.  This
> reduces time spent waiting to schedule work-queue jobs, but requires spending
> slightly more time in interrupt context.
> 
> A tasklet can only be used for verification if all the required hashes are
> already in the dm-bufio cache.  If verification cannot be done in a tasklet, we
> fallback the existing work-queue implementation.
> 
> To allow tasklets to query the dm-bufio cache, the dm-bufio code must not sleep.
> This patchset adds a flag to dm-bufio that disallows sleeping.
> 
> The following shows a speed comparison of random reads on a dm-verity device.
> The dm-verity device uses a 1G ramdisk for data and a 1G ramdisk for hashes.
> One test was run using tasklets and one test was run using the existing
> work-queue solution.  Both tests were run when the dm-bufio cache was hot.  The
> tasklet implementation performs significantly better since there is no time
> waiting for work-queue jobs to be scheduled.
> 
> # /data/fio /data/tasklet.fio | grep READ
>    READ: bw=181MiB/s (190MB/s), 181MiB/s-181MiB/s (190MB/s-190MB/s), io=512MiB
>    (537MB), run=2827-2827msec
> # /data/fio /data/workqueue.fio | grep READ
>    READ: bw=23.6MiB/s (24.8MB/s), 23.6MiB/s-23.6MiB/s (24.8MB/s-24.8MB/s),
>    io=512MiB (537MB), run=21688-21688msec
> 
> Nathan Huckleberry (3):
>   dm-bufio: Add flags for dm_bufio_client_create
>   dm-bufio: Add DM_BUFIO_GET_CANT_SLEEP
>   dm-verity: Add try_verify_in_tasklet
> 
>  drivers/md/dm-bufio.c                         | 29 +++++--
>  drivers/md/dm-ebs-target.c                    |  3 +-
>  drivers/md/dm-integrity.c                     |  2 +-
>  drivers/md/dm-snap-persistent.c               |  2 +-
>  drivers/md/dm-verity-fec.c                    |  4 +-
>  drivers/md/dm-verity-target.c                 | 87 ++++++++++++++++---
>  drivers/md/dm-verity.h                        |  5 ++
>  drivers/md/persistent-data/dm-block-manager.c |  3 +-
>  include/linux/dm-bufio.h                      |  8 +-
>  9 files changed, 119 insertions(+), 24 deletions(-)
> 
> -- 
> 2.37.1.359.gd136c6c3e2-goog
> 
---end quoted text---