From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pf0-f197.google.com (mail-pf0-f197.google.com [209.85.192.197]) by kanga.kvack.org (Postfix) with ESMTP id 399B36B0038 for ; Fri, 3 Feb 2017 10:14:04 -0500 (EST) Received: by mail-pf0-f197.google.com with SMTP id d123so26096071pfd.0 for ; Fri, 03 Feb 2017 07:14:04 -0800 (PST) Received: from bombadil.infradead.org (bombadil.infradead.org. [65.50.211.133]) by mx.google.com with ESMTPS id l33si25751008pld.26.2017.02.03.07.14.03 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Fri, 03 Feb 2017 07:14:03 -0800 (PST) Date: Fri, 3 Feb 2017 07:13:59 -0800 From: Matthew Wilcox Subject: Re: [PATCH] mm: Avoid returning VM_FAULT_RETRY from ->page_mkwrite handlers Message-ID: <20170203151356.GB2267@bombadil.infradead.org> References: <20170203150729.15863-1-jack@suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20170203150729.15863-1-jack@suse.cz> Sender: owner-linux-mm@kvack.org List-ID: To: Jan Kara Cc: Al Viro , linux-fsdevel@vger.kernel.org, linux-mm@kvack.org, Andrew Morton , lustre-devel@lists.lustre.org, cluster-devel@redhat.com On Fri, Feb 03, 2017 at 04:07:29PM +0100, Jan Kara wrote: > Some ->page_mkwrite handlers may return VM_FAULT_RETRY as its return > code (GFS2 or Lustre can definitely do this). However VM_FAULT_RETRY > from ->page_mkwrite is completely unhandled by the mm code and results > in locking and writeably mapping the page which definitely is not what > the caller wanted. Fix Lustre and block_page_mkwrite_ret() used by other > filesystems (notably GFS2) to return VM_FAULT_NOPAGE instead which > results in bailing out from the fault code, the CPU then retries the > access, and we fault again effectively doing what the handler wanted. Reading this commit message makes me wonder if this is the best fix. It would seem logical that if I want the fault to be retried that I should return VM_FAULT_RETRY, not VM_FAULT_NOPAGE. Why don't we have the MM treat VM_FAULT_RETRY the same way that it treats VM_FAULT_NOPAGE and give driver / filesystem writers one fewer way to shoot themselves in the foot? -- 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