Return-Path: linux-nfs-owner@vger.kernel.org Received: from mail-qa0-f53.google.com ([209.85.216.53]:35121 "EHLO mail-qa0-f53.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754257AbaLHUtX (ORCPT ); Mon, 8 Dec 2014 15:49:23 -0500 Received: by mail-qa0-f53.google.com with SMTP id bm13so3968598qab.12 for ; Mon, 08 Dec 2014 12:49:22 -0800 (PST) From: Jeff Layton Date: Mon, 8 Dec 2014 15:49:21 -0500 To: "J. Bruce Fields" Cc: linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org, Tejun Heo , Al Viro Subject: Re: [RFC PATCH 09/14] sunrpc: add basic support for workqueue-based services Message-ID: <20141208154921.3e200c12@tlielax.poochiereds.net> In-Reply-To: <20141208204709.GA19437@fieldses.org> References: <1417544663-13299-1-git-send-email-jlayton@primarydata.com> <1417544663-13299-10-git-send-email-jlayton@primarydata.com> <20141208204709.GA19437@fieldses.org> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-nfs-owner@vger.kernel.org List-ID: On Mon, 8 Dec 2014 15:47:09 -0500 "J. Bruce Fields" wrote: > On Tue, Dec 02, 2014 at 01:24:18PM -0500, Jeff Layton wrote: > > diff --git a/net/sunrpc/Kconfig b/net/sunrpc/Kconfig > > index fb78117b896c..08e01949bdc5 100644 > > --- a/net/sunrpc/Kconfig > > +++ b/net/sunrpc/Kconfig > > @@ -71,3 +71,13 @@ config SUNRPC_XPRT_RDMA_SERVER > > choose M here: the module will be called svcrdma. > > > > If unsure, say N. > > + > > +config SUNRPC_SVC_WORKQUEUE > > + bool "Support for workqueue-based SUNRPC services" > > + depends on SUNRPC > > + default n > > + help > > + Traditional SUNRPC services have required a dedicated thread > > + to handle incoming requests. This option enables support for > > + queueing incoming reqests to a workqueue instead, eliminating > > + the need for a dedicated thread pool. > > Minor point, but: If people don't want this, they can turn it off at > runtime. It's annoying to test all the possible combination of build > options, and this doesn't seem likely to let people e.g. build a > significantly smaller kernel, so I wouldn't bother. > > --b. Ok, fair enough. I added the config option since this was experimental, but if we keep it runtime-switchable then I think you're probably right and there's no need. -- Jeff Layton