From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: martin.petersen@oracle.com To: =?utf-8?B?THVrw6HFoQ==?= Czerner Cc: "Martin K. Petersen" , Karel Zak , util-linux@vger.kernel.org, linux-fsdevel@vger.kernel.org Subject: Re: [PATCH] blkdiscard: add new command From: "Martin K. Petersen" References: <1347486555-24330-1-git-send-email-lczerner@redhat.com> <20120927094256.GA18644@x2.net.home> Date: Thu, 27 Sep 2012 13:33:30 -0400 In-Reply-To: (=?utf-8?B?Ikx1a8OhxaE=?= Czerner"'s message of "Thu, 27 Sep 2012 13:03:30 -0400 (EDT)") Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 List-ID: >>>>> "Luk=C3=A1=C5=A1" =3D=3D Luk=C3=A1=C5=A1 Czerner writes: Luk=C3=A1=C5=A1> Hmm, I do not know. blktool seems to be very similar to bl= ockdev Luk=C3=A1=C5=A1> in the sense that it is for querying and changing block de= vice Luk=C3=A1=C5=A1> setting which is not what we do with, discard, write same = or Luk=C3=A1=C5=A1> zeroout where we actually change the data on the device Luk=C3=A1=C5=A1> itself. In my mind it just somehow does not fit there. I was merely picking up on the fact that blockdev's argument handling didn't seem particularly friendly to the discard/zero/ws case. Whereas blktool looked more flexible. Luk=C3=A1=C5=A1> But you're right that it is not just about discard or secu= re Luk=C3=A1=C5=A1> discard, we have other quite similar functionalities and i= t might Luk=C3=A1=C5=A1> be worth having one separate tool for all those, thanks for Luk=C3=A1=C5=A1> pointing this out. I'm perfectly ok with a separate blkswissarmyknife command. I'm also ok with separate blkdiscard/blkwritesame/blkzeroout commands as long as they look and feel the same. --=20 Martin K. Petersen Oracle Linux Engineering