[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAJaqyWdSgTcH1sLcmMbySgW+=G3VvjHoy5mW1xz+QNmnXhEbgw@mail.gmail.com>
Date: Fri, 17 Oct 2025 08:32:26 +0200
From: Eugenio Perez Martin <eperezma@...hat.com>
To: "Michael S. Tsirkin" <mst@...hat.com>, Yongji Xie <xieyongji@...edance.com>
Cc: Stefano Garzarella <sgarzare@...hat.com>, Xuan Zhuo <xuanzhuo@...ux.alibaba.com>,
virtualization@...ts.linux.dev, linux-kernel@...r.kernel.org,
jasowang@...hat.com, Maxime Coquelin <mcoqueli@...hat.com>,
Laurent Vivier <lvivier@...hat.com>, Cindy Lu <lulu@...hat.com>
Subject: Re: [PATCH v7 1/7] vduse: make domain_lock an rwlock
On Fri, Oct 10, 2025 at 2:34 PM Michael S. Tsirkin <mst@...hat.com> wrote:
>
> On Fri, Oct 10, 2025 at 10:58:21AM +0200, Eugenio Pérez wrote:
> > It will be used in a few more scenarios read-only so make it more
> > scalable.
>
> Well a mutex is sleepable and rwlock is a spinlock.
>
> So this does much more than "make it more scalable".
>
> "A few more scenarios" and "scalable" is also vague.
> RW has its own issues such as fairness.
>
> So tell us please, which operations do you want to speed up and why?
> What kind of speedup was observed?
>
I think you're totally right here.
Yongji, would it work for you if we leave this conversion on top of
the series with a simple mutex, and we have bandwith to explore the
benefits of this? You were the one proposing it [1].
Thanks!
[1] https://lore.kernel.org/all/CACycT3v=_Nm6fefJGFEyoU+Xf5G=Kzi0sXhhaBHnJQZcG-4EqA@mail.gmail.com/
Powered by blists - more mailing lists