From mboxrd@z Thu Jan 1 00:00:00 1970 From: Miklos Szeredi Subject: Re: [GIT PULL for 4.8] consolidate vfs API for overlayfs Date: Thu, 30 Jun 2016 09:22:07 +0200 Message-ID: References: <20160628081810.GB2388@veci.piliscsaba.szeredi.hu> <20160629230952.GE14480@ZenIV.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Return-path: In-Reply-To: <20160629230952.GE14480@ZenIV.linux.org.uk> Sender: linux-kernel-owner@vger.kernel.org To: Al Viro Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, "linux-unionfs@vger.kernel.org" List-Id: linux-unionfs@vger.kernel.org On Thu, Jun 30, 2016 at 1:09 AM, Al Viro wrote: > On Tue, Jun 28, 2016 at 10:18:10AM +0200, Miklos Szeredi wrote: >> Hi Al, >> >> Please pull from: >> >> git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs.git d_real >> >> This consolidates the d_real/d_select_inode API as well as some documentation >> cleanups. > > Two notes: > > 1) you have 4 places with identical "if DCACHE_OP_REAL is there, > call ->d_real(dentry, ..., ...), otherwise return dentry". And you have > exactly one place where d_real() wrapper is called. Looks like we would > be better off with the calling conventions for wrapper updated as well. > > 2) while we can't turn dentry argument into const struct dentry * (due to > ovl_copy_up()), could we at least do that to the inode one? > > IOW, how about this, on top of that branch? Or fold it into the commit > changing the arguments of ->d_real(), for that matter... Makes sense. Folded, updated doc and force pushed to the same branch. Thanks, Miklos