From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Sun, 25 Feb 2007 04:06:57 -0800 From: Andrew Morton Subject: Re: [patch 3/3] mm: fix PageUptodate memorder Message-Id: <20070225040657.eb4fc159.akpm@linux-foundation.org> In-Reply-To: <20070215051851.7443.65811.sendpatchset@linux.site> References: <20070215051822.7443.30110.sendpatchset@linux.site> <20070215051851.7443.65811.sendpatchset@linux.site> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org Return-Path: To: Nick Piggin Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org List-ID: What an unpleasing patchset. I really really hope we really have a bug in there, and that all this crap isn't pointless uglification. We _do_ need a flush_dcaceh_page() in all cases which you're concerned about. Perhaps we should stick the appropriate barriers in there. > On Thu, 15 Feb 2007 08:31:31 +0100 (CET) Nick Piggin wrote: > +static inline void SetNewPageUptodate(struct page *page) > +{ > + /* > + * S390 sets page dirty bit on IO operations, which is why it is > + * cleared in SetPageUptodate. This is not an issue for newly > + * allocated pages that are brought uptodate by zeroing memory. > + */ > + smp_wmb(); > + __set_bit(PG_uptodate, &(page)->flags); > +} __SetPageUptodate() might be more conventional. Boy we'd better get the callers of this little handgrenade right. -- 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: email@kvack.org