From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Dave Young" Subject: Re: [PATCH -mm] __ratelimit rewrite Date: Mon, 5 May 2008 13:18:30 +0800 Message-ID: References: <20080504070738.GA11832@darkstar.te-china.tietoenator.com> <1209912447.16966.42.camel@localhost> <1209954305.13011.1.camel@localhost> <1209955974.13011.10.camel@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: akpm@linux-foundation.org, paulmck@us.ibm.com, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, "David Miller" To: "Joe Perches" Return-path: Received: from yw-out-2324.google.com ([74.125.46.30]:43714 "EHLO yw-out-2324.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752247AbYEEFSb (ORCPT ); Mon, 5 May 2008 01:18:31 -0400 Received: by yw-out-2324.google.com with SMTP id 9so375717ywe.1 for ; Sun, 04 May 2008 22:18:30 -0700 (PDT) In-Reply-To: <1209955974.13011.10.camel@localhost> Content-Disposition: inline Sender: netdev-owner@vger.kernel.org List-ID: On Mon, May 5, 2008 at 10:52 AM, Joe Perches wrote: > > On Mon, 2008-05-05 at 10:58 +0800, Dave Young wrote: > > On Mon, May 5, 2008 at 10:25 AM, Joe Perches wrote: > > > On Mon, 2008-05-05 at 09:37 +0800, Dave Young wrote: > > > > > I think your changes should then be done in > > > > > printk.h not creating a new ratelimit.h. > > > > IMO ratelimit is not just for printk use now, so a standalone head > > > > file is necessary. > > > What other uses would ratelimit have? > > ratelimit is a general function, another user of it is > > WARN_ON_RATELIMIT in this patch. It could have other usage in future. > > You put WARN_ON_RATELIMIT in bug.h right? > WARN_ON in bug.h is a printk Yes, warn is indeed printk, but it's not equal to printk. > > printk.h should #include > kernel.h should not include that file. > > What non print related function could be served by ratelimit? You can do some other handling except printk in ratelimit logic. After searching the code, I guess DCCP-Sync ratelimit can use this feature. David, am I right? BTW, most of printk_ratelimit users should use their own ratelimit_state, maybe I should do the replace work as a patch series.