From mboxrd@z Thu Jan 1 00:00:00 1970 From: Robin Holt Date: Mon, 07 Aug 2006 17:27:00 +0000 Subject: Re: Silent data corruption caused by XPC. Message-Id: <20060807172700.GA24663@lnx-holt.americas.sgi.com> List-Id: References: <20060807154608.GC15674@lnx-holt.americas.sgi.com> In-Reply-To: <20060807154608.GC15674@lnx-holt.americas.sgi.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-ia64@vger.kernel.org On Mon, Aug 07, 2006 at 08:57:51AM -0700, Luck, Tony wrote: > > and verified it works. I also added DBUG_ON statements to help catch > > this if a similar situation is encountered. > > Can you just do this once at alloc time rather than at every use? Dean and I liked the idea at first, but hashed it over more. My intent with putting them right before the bte_copy was to trap the event where a copy was going to be initiated that would corrupt memory. If I put them at allocation time only, any changes in the allocation would not necessarily be caught. I think we would like to leave them where they are at unless you have a strong objection. They are DBUG_ON() instead of BUG_ON() so they should not have a performance impact normally. If you would like it moved to allocation time, it will make the patch smaller and more clean as we will move it into xpc_kmalloc_cacheline_aligned. Dean said he will handle that if you desire. Thanks, Robin