From mboxrd@z Thu Jan 1 00:00:00 1970 From: Theodore Ts'o Subject: Re: [PATCH 1/3] ext4: pass allocation_request struct to ext4_(alloc,splice)_branch Date: Wed, 3 Sep 2014 13:17:21 -0400 Message-ID: <20140903171721.GG2504@thunk.org> References: <1409695549-18605-1-git-send-email-tytso@mit.edu> <1409695549-18605-2-git-send-email-tytso@mit.edu> <20140903162552.GD17066@quack.suse.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Ext4 Developers List To: Jan Kara Return-path: Received: from imap.thunk.org ([74.207.234.97]:35500 "EHLO imap.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932670AbaICRRZ (ORCPT ); Wed, 3 Sep 2014 13:17:25 -0400 Content-Disposition: inline In-Reply-To: <20140903162552.GD17066@quack.suse.cz> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Wed, Sep 03, 2014 at 06:25:52PM +0200, Jan Kara wrote: > This seems to suggest ext4_new_meta_blocks() would be better off by > taking allocation_request argument as well? I thought about it, but the problem is that ext4_new_meta_blocks() is called in many more places than ext4_alloc_branch(). So in this patch, it was just a metter of moving some code from one function to its (single) caller. In the case of ext4_new_meta_blocks(), we would needing to replicate that that code in four or five places, and we were passing in the mb_flags field anyway, so it wasn't a case of needing to add yet another argument to a function that had many more arguments to start with. So I decided it wasn't worth the effort. Cheers, - Ted