From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from imap.thunk.org ([74.207.234.97]:53636 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751882AbeDQXfN (ORCPT ); Tue, 17 Apr 2018 19:35:13 -0400 Date: Tue, 17 Apr 2018 19:34:59 -0400 From: "Theodore Y. Ts'o" Subject: Re: ioctl FIBMAP for dax gone in v4.17-rc1 Message-ID: <20180417233459.GB23194@thunk.org> References: <20180417144059.nwbbynhgq3k3i63q@XZHOUW.usersys.redhat.com> <20180417161056.GA24257@infradead.org> <20180417165756.GC24738@magnolia> <20180417175007.GB5210@magnolia> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180417175007.GB5210@magnolia> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: "Darrick J. Wong" Cc: Andreas Dilger , Dan Williams , Christoph Hellwig , Xiong Zhou , linux-fsdevel , linux-xfs , Eric Sandeen , fstests , linux-nvdimm On Tue, Apr 17, 2018 at 10:50:07AM -0700, Darrick J. Wong wrote: > syslinux does FIEMAP w/ FIBMAP fallback. > > lilo & elilo still use FIBMAP only, but lilo hasn't seen a release since > 2015 and elilo 2013. I prefer not to add FIBMAP support in XFS-DAX for > the singular case of booting off pmem via lilo. It shouldn't be that hard to write a userspace emulation layer which exports a FIBMAP-like functional interface but gets the information from the kernel using the FIMAP ioctl.... - Ted