From mboxrd@z Thu Jan 1 00:00:00 1970 From: Michael Tokarev Subject: Re: [PATCH 1/2] Modernize udev rules Date: Mon, 11 Feb 2013 14:16:17 +0400 Message-ID: <5118C4F1.90203@msgid.tls.msk.ru> References: <1360432119-15910-1-git-send-email-thomas@archlinux.org> <20130211112016.3059ef95@notabene.brown> <5118ACE1.5020503@msgid.tls.msk.ru> <5118C3C0.2030708@archlinux.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <5118C3C0.2030708@archlinux.org> Sender: linux-raid-owner@vger.kernel.org To: =?ISO-8859-15?Q?Thomas_B=E4chler?= Cc: NeilBrown , linux-raid@vger.kernel.org List-Id: linux-raid.ids 11.02.2013 14:11, Thomas B=E4chler wrote: > Am 11.02.2013 09:33, schrieb Michael Tokarev: >> 11.02.2013 04:20, NeilBrown wrote: >>> On Sat, 9 Feb 2013 18:48:38 +0100 Thomas B=E4chler >>> wrote: >> >>>> -IMPORT{program}=3D"/sbin/blkid -o udev -p $tempnode" >>>> +IMPORT{builtin}=3D"blkid" >> >> Does this has additional udev version constrain? >> I mean, which is the minimum udev version to support this properly, >> do we care? > > Udev relies on devtmpfs and thus does not create devices since versio= n > 176, therefore $tempnode is deprecated and only $devnode should be us= ed > ($tempnode seems to be around for compatibility, no idea if/when it w= ill > be dropped). > > The blkid builtin is also available since udev 176. In latest > util-linux, the 'blkid -o udev' output is marked as deprecated, altho= ugh > Karel says it's not entirely going away for now. Thank you for the explanation. > Why do we care? We want correct udev rules ($tempnode vs. $devnode), = and > the using builtin is faster and less error-prone than calling an > external tool (plus, it is used everywhere else in udev). > > As for compatibility: I see no reason why anyone would use a version = of > udev that is more than one year old (175) together with a mdadm versi= on > that is to be released in the future (3.2.7). If someone really wants > that, they should be competent enough to adjust the udev rules for th= eir > local build. Well, it was just a question. Think about a distribution which is bein= g upgraded from one release to another, -- inter-package versioned depend= encies should be set correctly, so that the system will not be broken by upgra= ding mdadm which relies on udev, but not upgrading udev. It's a good idea t= o note somewhere that the next release of mdadm will require at least ude= v 175. That's about it. Thank you! /mjt > -- To unsubscribe from this list: send the line "unsubscribe linux-raid" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html