From mboxrd@z Thu Jan 1 00:00:00 1970 From: Oliver Hartkopp Subject: Re: Associate physical device with device name Date: Fri, 9 Feb 2018 18:09:36 +0100 Message-ID: <16bc3c4c-4013-9d9c-4afe-4cbffc2d8ad7@hartkopp.net> References: <87ddf56f-c829-6aad-5b86-474359d282d4@pengutronix.de> <464f89e6-82c8-a2fd-25e2-f847508b41a3@pengutronix.de> <35989d02-3134-1a84-8bec-61a863fa0bf3@kvaser.com> <771a0c6e-c94c-7dc2-ffa0-06507aab077d@pengutronix.de> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from mo4-p00-ob.smtp.rzone.de ([81.169.146.220]:24473 "EHLO mo4-p00-ob.smtp.rzone.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751145AbeBIRJq (ORCPT ); Fri, 9 Feb 2018 12:09:46 -0500 In-Reply-To: Content-Language: en-US Sender: linux-can-owner@vger.kernel.org List-ID: To: Jimmy Assarsson , Marc Kleine-Budde , "linux-can@vger.kernel.org" On 02/09/2018 04:35 PM, Jimmy Assarsson wrote: > On 2018-02-09 14:24, Marc Kleine-Budde wrote: >> On 02/09/2018 12:21 PM, Jimmy Assarsson wrote: >>> All devices with the same EAN look identical, since we do not set serial >>> number in the USB descriptor. >> >> That should be done on the next FW revision. > > It is left out on purpose. I don't remember the details, but it had > something to do with Windows reinstalling the driver whenever a device > with different serial number is connected to the same USB port. > >> Ok - I see two possibilities. >> >> 1) sysfs file >> Add a file to sysfs that contains the serial number, see: >> http://lxr.bootlin.com/linux/latest/source/drivers/net/can/softing/softing_main.c#L718 >> >> >> 2) ethtool eeprom >> You can export the serial number via ethtool's dump eeprom functionality. >> >> Better try 1). > > Great :) Will try this next week. Would it make sense to provide a common sysfs infrastructure in linux/drivers/net/can/dev.c to populate /sys/class/net// instead of maintaining this stuff in each single driver? We have this requirement in the softing/kvaser/peak adapters so far. We could provide some pointer in struct can_priv which points to a struct which contains the sysfs content - or something like this. Any thoughts? Best, Oliver