From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail172.messagelabs.com (mail172.messagelabs.com [216.82.254.3]) by kanga.kvack.org (Postfix) with ESMTP id 3D4016B004F for ; Fri, 17 Jul 2009 11:09:22 -0400 (EDT) Date: Fri, 17 Jul 2009 11:09:22 -0400 From: Christoph Hellwig Subject: Re: xfs mr_lock vs mmap_sem lock inversion? Message-ID: <20090717150922.GA434@infradead.org> References: <1247580955.7500.97.camel@twins> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1247580955.7500.97.camel@twins> Sender: owner-linux-mm@kvack.org To: Peter Zijlstra Cc: Christoph Hellwig , linux-kernel , linux-mm@kvack.org List-ID: It's a problem in the VM code, which we already discussed a while ago. The problem is that the VMA manipultation code calls fput under the mmap_sem, while we can take mmap_sem ue to a page fault from inside generic_file_aio_read/write. So any filesystem that nees the same lock held over read/write also in release is crewed. Now on the positive side I think we can actually get rid of taking the iolock in ->release in XFS, but I'm sure other filesystems might continue hitting similar issues. -- 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