[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZD0B3wGasaWT0rsr@debian.me>
Date: Mon, 17 Apr 2023 15:22:55 +0700
From: Bagas Sanjaya <bagasdotme@...il.com>
To: wenyang.linux@...mail.com,
Alexander Viro <viro@...iv.linux.org.uk>,
Jens Axboe <axboe@...nel.dk>,
Christian Brauner <brauner@...nel.org>
Cc: Christoph Hellwig <hch@....de>, Dylan Yudaken <dylany@...com>,
David Woodhouse <dwmw@...zon.co.uk>,
Paolo Bonzini <pbonzini@...hat.com>, Fu Wei <wefu@...hat.com>,
linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] eventfd: support delayed wakeup for non-semaphore
eventfd to reduce cpu utilization
On Sun, Apr 16, 2023 at 07:31:55PM +0800, wenyang.linux@...mail.com wrote:
> +eventfd_wakeup_delay_msec
> +------------------
Please match the section underline length as the section text above.
> +Frequent writing of an eventfd can also lead to frequent wakeup of the peer
> +read process, resulting in significant cpu overhead.
> +How ever for the NON SEMAPHORE eventfd, if it's counter has a nonzero value,
> +then a read(2) returns 8 bytes containing that value, and the counter's value
reading eventfd?
> +is reset to zero.
> +So it coule be optimized as follows: N event_writes vs ONE event_read.
> +By adding a configurable delay after eventfd_write, these unnecessary wakeup
> +operations are avoided.
What is the connection from optimization you described to eventfd_write
delay?
Thanks.
--
An old man doll... just what I always wanted! - Clara
Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)
Powered by blists - more mailing lists