From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hans de Goede Subject: Re: [PATCH] Input :Added Check for EV_ABS event params Date: Wed, 22 Apr 2015 08:35:13 +0200 Message-ID: <55374121.8000308@redhat.com> References: <1429640392-30471-1-git-send-email-aksgarg1989@gmail.com> <20150421182916.GB16790@dtor-ws> <20150421215600.GA22958@jelly.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mx1.redhat.com ([209.132.183.28]:54097 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964847AbbDVGfR (ORCPT ); Wed, 22 Apr 2015 02:35:17 -0400 In-Reply-To: <20150421215600.GA22958@jelly.redhat.com> Sender: linux-input-owner@vger.kernel.org List-Id: linux-input@vger.kernel.org To: Peter Hutterer , Dmitry Torokhov Cc: Anshul Garg , linux-input@vger.kernel.org, anshul.g@samsung.com, Benjamin Tissoires , Jiri Kosina Hi, On 21-04-15 23:56, Peter Hutterer wrote: > On Tue, Apr 21, 2015 at 11:29:16AM -0700, Dmitry Torokhov wrote: >> Hi Anshul, >> >> On Tue, Apr 21, 2015 at 11:19:52AM -0700, Anshul Garg wrote: >>> From: Anshul Garg >>> >>> while handling EV_ABS event in input_handle_abs_event >>> function added check for out of range event value from >>> input driver. As input driver sets the ABS params at >>> registration time so input core should ignore events out >>> of the range set by the input driver. >> >> No, I do not think we want to do that, at least not unconditionally, >> especially since it is perfectly allowed to use 0 as min/max, which >> means that exact min and max are not defined. Historically min and max >> were provided to the userspace as a guidance and it was up to userspace >> to decide what to do with values outside of the limits. > > for the archives: > X requires us to drop coordinates outside the given range, at least the > ones we forward to the clients from absolute devices. We have used > out-of-bounds variables within the synaptics driver in the past though. Also for the record, for touchpad devices min-max typically give the edges not the absolute min / max values of the sensor. In some cases there may also be margins given in the min/max values coming from the touchpad-firmware which already take an edge-scrolling area into account. So with touchpads getting values outside of the min/max range is sometimes something which not only may happen, but actually is expected to happen. Regards, Hans