From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Mason Subject: Re: [PATCH] Btrfs-progs: update btrfs-debug-tree Date: Fri, 20 Feb 2009 09:28:28 -0500 Message-ID: <1235140108.13249.0.camel@think.oraclecorp.com> References: <499D04FF.5020600@cn.fujitsu.com> <499DFBAC.1020602@suse.com> Mime-Version: 1.0 Content-Type: text/plain Cc: Shen Feng , linux-btrfs@vger.kernel.org To: Jeff Mahoney Return-path: In-Reply-To: <499DFBAC.1020602@suse.com> List-ID: On Thu, 2009-02-19 at 19:39 -0500, Jeff Mahoney wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Shen Feng wrote: > > output objectid in btrfs_disk_key with human readable strings. > > Other updates are included for more readable output. > > This gets messy fast. I'd like to see some sort of generic extensible > printk mechanism in the kernel. This is a btrfs-progs patch ;) > Maybe I've been spoiled by the reiserfs > type printing but it seems a lot nicer to do eprintk(&ops, "key: %k\n"); > than to split them out into different printks. Since the core behavior > ends up in vsnprintf, you get it for things like panic for free too. > Ideally, the tprintk(&ops, ...) would be hidden behind a btrfs_warning() > or something. > > I have a patch set that implements an extensible printk so that every > single pointer type doesn't need to be added to lib/vsprintf.c's pointer(). > That would be interesting, I do miss the old reiserfs %k -chris