[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <c550c690-7555-4ccd-bf8a-8c54657aea3c@paulmck-laptop>
Date: Mon, 1 Jul 2024 07:18:03 -0700
From: "Paul E. McKenney" <paulmck@...nel.org>
To: Markus Elfring <Markus.Elfring@....de>
Cc: Matthew Wilcox <willy@...radead.org>, kernel-janitors@...r.kernel.org,
Boqun Feng <boqun.feng@...il.com>,
Johannes Berg <johannes.berg@...el.com>,
Peter Zijlstra <peterz@...radead.org>,
Uladzislau Rezki <urezki@...il.com>,
LKML <linux-kernel@...r.kernel.org>, linux-fsdevel@...r.kernel.org,
maple-tree@...ts.infradead.org
Subject: Re: [v2 2/5] rosebush: Add new data structure
On Mon, Jul 01, 2024 at 07:21:18AM +0200, Markus Elfring wrote:
> >> Under which circumstances would you become interested to apply a statement
> >> like “guard(rcu)();”?
> >
> > Under no circumstances.
>
> I imagine that further contributors would like to discuss collateral evolution
> also according to the support for applications of scope-based resource management.
> https://elixir.bootlin.com/linux/v6.10-rc6/source/include/linux/rcupdate.h#L1093
>
> See also the commit 80cd613a9ae091dbf52e27a409d58da988ffc8f3 ("rcu:
> Mollify sparse with RCU guard") from 2024-04-15.
Although the guard(rcu)() statement is very helpful in some circumstances
and is seeing increasing use, it is not free of downsides in a number
of situations. For but one example, Matthew might expect that partially
overlapping critical sections will be needed, which would rule out use of
guards on one or the other of those two critical sections.
Thanx, Paul
Powered by blists - more mailing lists