From mboxrd@z Thu Jan 1 00:00:00 1970 From: Martin Pitt Date: Mon, 14 Dec 2009 13:32:45 +0000 Subject: Re: [PATCH] keymap: support for force_release quirk Message-Id: <20091214133245.GB2039@piware.de> List-Id: References: <20091214001219.GA17792@sig21.net> In-Reply-To: <20091214001219.GA17792@sig21.net> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-hotplug@vger.kernel.org Johannes Stezenbach [2009-12-14 13:55 +0100]: > Yes, that was my question: Can I just save the original at boot > (first "add" event) into a file (like /dev/.udev/force-release/serioX)? No, please don't; such external state keeping is very ugly and error prone IMHO, and not what udev is designed to do. I rather think we should do one of: (1) Demand to have complete force-release tables, and completely reset the value with each invocation of the udev rule. Pro: Can remove values Con: Hard to just fix a single key (2) Only ever add new values by reading the current value, appending our tables, and writing it back. Pro: Can fix single keys and potentially simplifies rules Con: Needs reboot for bad tables which set quirk erroneously I suppose that on computers which need these quirks, all Fn keys are affected equally, so the "fix single key" case is probably irrelevant. OTOH, I don't think we'll get too many errors like "erroneous quirk", so both approaches should work well in practice. Personally I tend to prefer (2). > BTW, 95-keymap.rules also doesn't handle "change". Does it need > updating or is there a difference in handling "change" for > force_release? Fixed in git, thanks for pointing out. Martin -- Martin Pitt | http://www.piware.de Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org)