From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Sun, 29 Jul 2007 20:56:54 -0700 (PDT) From: Linus Torvalds Subject: Re: [patch][rfc] remove ZERO_PAGE? In-Reply-To: Message-ID: References: <20070727021943.GD13939@wotan.suse.de> <20070727055406.GA22581@wotan.suse.de> <20070730030806.GA17367@wotan.suse.de> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=us-ascii Sender: owner-linux-mm@kvack.org Return-Path: To: Nick Piggin Cc: Andrew Morton , Hugh Dickins , Andrea Arcangeli , Linux Memory Management List List-ID: On Sun, 29 Jul 2007, Linus Torvalds wrote: > > I'd love to hear "here's a real-life load, and yes, the ZERO_PAGE logic > really does hurt more than it helps, it's time to remove it". At that > point I'll happily apply the patch. Btw, in the absense of that, I'd at least like to hear an acknowledgement that the complexity isn't worth it, and that what used to work fine was broken because the reference counting overhead gets us on large-scale machines. IOW, I certainly like removing lines of code. In that sense I _love_ that patch. I really just react negatively because I really think you first set ZERO_PAGE up to fail. So even if you cannot find a load where this all matters, at least point to commit b5810039a54e5babf428e9a1e89fc1940fabff11 (or exactly whichever one it was that started ref-counting ZERO_PAGE) and blame *that* one, rather than blaming ZERO_PAGE for the problem. It has served us well for fifteen years, we shouldn't blame it for problems that came from elsewhere. Linus -- 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