From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Wed, 26 May 2004 09:06:17 +0200 From: Andrea Arcangeli Subject: Re: [PATCH] ppc64: Fix possible race with set_pte on a present PTE Message-ID: <20040526070617.GN29378@dualathlon.random> References: <20040525114437.GC29154@parcelfarce.linux.theplanet.co.uk> <20040525212720.GG29378@dualathlon.random> <20040525215500.GI29378@dualathlon.random> <20040526021845.A1302@den.park.msu.ru> <20040525224258.GK29378@dualathlon.random> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: owner-linux-mm@kvack.org Return-Path: To: Linus Torvalds Cc: Ivan Kokshaysky , Matthew Wilcox , Benjamin Herrenschmidt , Andrew Morton , Linux Kernel list , Ingo Molnar , Ben LaHaise , linux-mm@kvack.org, Architectures Group List-ID: On Tue, May 25, 2004 at 07:26:21PM -0700, Linus Torvalds wrote: > You're reading it wrong. > > The "including when the present flag is set to zero" part does not mean > that the present flag was zero _before_, it means "is being set to zero" > as in "having been non-zero before that". "having been non-zero before that" makes a lot more sense indeed, the wording in the specs wasn't the best IMHO. Interestingly the ptep_establish at the end of handle_pte_fault would have hidden any double fault completely, nobody but a tracer would have noticed that, but it made very little sense that non-present entries can be cached. It's all clear now thanks. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: aart@kvack.org