From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Biedl Subject: Re: Testing the lws_compare_and_swap_2 syscall Date: Tue, 7 Nov 2017 22:46:25 +0100 Message-ID: <1510090803@msgid.manchmal.in-ulm.de> References: <1508874207@msgid.manchmal.in-ulm.de> <1508973952@msgid.manchmal.in-ulm.de> <20171106212734.GA29237@ls3530.fritz.box> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii To: Helge Deller , Richard Henderson , John David Anglin , linux-parisc@vger.kernel.org Return-path: In-Reply-To: <20171106212734.GA29237@ls3530.fritz.box> List-ID: List-Id: linux-parisc.vger.kernel.org Helge Deller wrote... > * Christoph Biedl : > > Status: My system boots and all my tests pass. However, I find smartd > > stalling in 100% CPU, might be coincidence. Now I'm putting some more > > load onto the box to see whether it's less crashy then it used to be the > > previous weeks. > > Did you continued your tests? > How were the results. Basically I gave up on this. My change made things worse, and while I still didn't get the point (read: For my understanding there's a discrepancy between the syscall's description and what actually happens), it's certainly better to keep things the way they are. > > Aside, there is another ",ma" modifier in lws_compare_and_swap that I > > fail to understand. Haven't checked yet in detail yet, though. > > I'd suggest to keep the other ",ma" modifiers for now. Since, as I presume, this has no impact on performance, this boils down to aid a human reviewing the code in understanding what's happening here. Christoph