From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail105.syd.optusnet.com.au ([211.29.132.249]:39173 "EHLO mail105.syd.optusnet.com.au" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725857AbfEOGHw (ORCPT ); Wed, 15 May 2019 02:07:52 -0400 Date: Wed, 15 May 2019 16:07:50 +1000 From: Dave Chinner Subject: Re: [PATCH 06/11] libxfs: create new file trans_buf.c Message-ID: <20190515060750.GY29573@dread.disaster.area> References: <1557519510-10602-1-git-send-email-sandeen@redhat.com> <1557519510-10602-7-git-send-email-sandeen@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1557519510-10602-7-git-send-email-sandeen@redhat.com> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Eric Sandeen Cc: linux-xfs@vger.kernel.org On Fri, May 10, 2019 at 03:18:25PM -0500, Eric Sandeen wrote: > Pull functions out of libxfs/*.c into trans_buf.c, if they roughly match > the kernel's xfs_trans_buf.c file. > > Signed-off-by: Eric Sandeen So I have no problems with this, but I'm not sure what the eventual goal is? Just sharing code, or is there some functionality that requires a more complete transaction subsystem in userspace? I'm asking because if the goal is eventual unification with the kernel code, then we probably should name the files the same as the kernel code so we don't have to rename them again when we do the unification. That will make history searching a bit easier - less file names to follow across and git blame works a whole lot better... > +int > +xfs_trans_read_buf_map( > + xfs_mount_t *mp, > + xfs_trans_t *tp, > + struct xfs_buftarg *btp, > + struct xfs_buf_map *map, > + int nmaps, > + uint flags, > + xfs_buf_t **bpp, > + const struct xfs_buf_ops *ops) Hmmmm. Will there be any follow-up to de-typedef these new files? /me would love to just have a flag day that de-typedefs all of the userspace code. Cheers, Dave. -- Dave Chinner david@fromorbit.com