netdev - Re: [RFC 1/1] eventpoll: support busy poll per epoll instance

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] [day] [month] [year] [list]

Message-ID: <20240122211616.GA1244@fastly.com>
Date: Mon, 22 Jan 2024 13:16:16 -0800
From: Joe Damato <jdamato@...tly.com>
To: Christian Brauner <brauner@...nel.org>
Cc: netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	chuck.lever@...cle.com, jlayton@...nel.org,
	linux-api@...r.kernel.org, edumazet@...gle.com, davem@...emloft.net,
	alexander.duyck@...il.com, sridhar.samudrala@...el.com,
	kuba@...nel.org
Subject: Re: [RFC 1/1] eventpoll: support busy poll per epoll instance

On Mon, Jan 22, 2024 at 04:25:01PM +0100, Christian Brauner wrote:
> On Sat, Jan 20, 2024 at 12:42:47AM +0000, Joe Damato wrote:
> > Add F_EPOLL_{S,G}ET_BUSY_POLL_USECS to allow setting a busy poll timeout
> > per epoll instance so that individual applications can enable (or
> > disable) epoll based busy poll as needed.
> > 
> > Prior to this change, epoll-based busy poll could only be enabled
> > system-wide, which limits the usefulness of busy poll.
> > 
> > Signed-off-by: Joe Damato <jdamato@...tly.com>
> > ---
> 
> This should be an ioctl on the epoll fd, not a fcntl(). fcntl()s
> should aim to be generic which this isn't. We've recently rejected a
> memfd addition as a fcntl() as well for the same reason.

OK, thanks for the review. An ioctl makes more sense, I agree.

I'll rewrite it as you've suggested and send another RFC.