From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Mosberger Date: Fri, 29 Apr 2005 16:01:14 +0000 Subject: Re: write_unlock: replace clear_bit with byte store Message-Id: <17010.23114.560223.6706@napali.hpl.hp.com> List-Id: References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-ia64@vger.kernel.org >>>>> On Fri, 29 Apr 2005 16:51:53 +0100, Christoph Hellwig said: Christoph> On Fri, Apr 29, 2005 at 08:48:43AM -0700, David Mosberger wrote: >> >> Yes, but _if_ it's a good idea to use .nta with GCC, there is no >> >> reason not to do the same with ICC. Don't introduce unnecessary >> >> divergence. Christoph> The same situation of .nta only for GCC already exists Christoph> for regular spinlocks as a result of my nta unlock patch Christoph> that I posted a week or so ago. >> And that's an argument to make the situation worse? How about >> cleaning up the previous patch instead? Christoph> I don't think it's fair to expect contributors to fix up Christoph> ifdef'ed bits for a propritary compiler. If HP and Intel Christoph> care about it they can add the features for icc later. I don't think this is about proprietary vs non-proprietary. It's about whether the code is clean. I'd _love_ to get rid of inline asm in the future when GCC supports intrinsics since that could lead to significantly improved code. (And no, I don't expect Chris to necessarily fix up the ICC bits, though a reasonable best-effort wouldn't take much time.) --david