Received: by 2002:a05:6a10:c7d3:0:0:0:0 with SMTP id h19csp685987pxy; Sat, 14 Aug 2021 20:35:20 -0700 (PDT) X-Google-Smtp-Source: ABdhPJxGDSRRSUXHMLSdx+EJJ/fi6ezqnfF4lCtvr1YTMHoDrCuXC1PywXmtzKlh+ZuBccTnBJyl X-Received: by 2002:a05:6402:14d1:: with SMTP id f17mr11826237edx.167.1628998520559; Sat, 14 Aug 2021 20:35:20 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1628998520; cv=none; d=google.com; s=arc-20160816; b=YGkuiI3pcGFRv8CPD2cTrwUqKDxoyUMQ8a8z07MRq3RerENEUWegMHLa/IIMp4UzPd 6fmB7HfG3pjhMu0h2o0SSoVDimgT9+RwDilXORak1ED8j6bxIvaZ0ORw2W7pyDaJW1dq 9e5LW5zTCDYkSZY5XtGdByDACE5KS490oVEn3aW4QTSJwwwa086HZQ2a5B1iwpzNbGkS RZ/6ni59ojtTdod4B7AtU14ARrWbBREXPIZk0WYmyhU5J/CO9/CBGO9sOJw1sCTVQDVP OCbuxVpXwS6nNi2PiJos7YvgUoQnay2b8ZuMi5SR3WafwvfALqRXKmPpIbxeZ/M5f8iV bU2g== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:content-transfer-encoding:mime-version :message-id:date:subject:cc:to:from; bh=wnAfnh2WKaNS7EAgIUCMCrsl9z2JaAYl5E72LviGoAg=; b=a9JZyUyr8ylxjHRAx4OuQ9owljnVYRgI5aMIb9ri0fZXolMOiO3JklrZFMAUG1a9wd dzC7WmJYWRSKFPg2nCw2SNTUKs/wn48abgeGWUq+n9+x0TK6XRmKNZGbMyW3/vsEKo76 +z5BjA4MGQGhBerdn989Wrfwrd3y6T6elYMZ79p9uAgGlGxtIOGXiNRjC9BlZgtrh6+0 RRJEYumsEvmsHjtR39sN+RHK4ATV7HzouPA2MnY2MsFCr+9U4VvkG4jLW9NggA0+ZD5A PEyNj0+J0e+9LXdFQLvIW4x6buTbndk6jL/fwnLm68Ssuil4Y4aFzUvPyCSXDQ8Surmk knqg== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Return-Path: Received: from vger.kernel.org (vger.kernel.org. [23.128.96.18]) by mx.google.com with ESMTP id nc13si6958726ejc.291.2021.08.14.20.34.58; Sat, 14 Aug 2021 20:35:20 -0700 (PDT) Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18; Authentication-Results: mx.google.com; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235414AbhHODdv (ORCPT + 99 others); Sat, 14 Aug 2021 23:33:51 -0400 Received: from smtp11.smtpout.orange.fr ([80.12.242.133]:36235 "EHLO smtp.smtpout.orange.fr" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235413AbhHODdt (ORCPT ); Sat, 14 Aug 2021 23:33:49 -0400 Received: from tomoyo.flets-east.jp ([114.149.34.46]) by mwinf5d21 with ME id hfZ7250050zjR6y03fZEwn; Sun, 15 Aug 2021 05:33:17 +0200 X-ME-Helo: tomoyo.flets-east.jp X-ME-Auth: bWFpbGhvbC52aW5jZW50QHdhbmFkb28uZnI= X-ME-Date: Sun, 15 Aug 2021 05:33:17 +0200 X-ME-IP: 114.149.34.46 From: Vincent Mailhol To: Marc Kleine-Budde , linux-can@vger.kernel.org Cc: =?UTF-8?q?Stefan=20M=C3=A4tje?= , netdev@vger.kernel.org, linux-kernel@vger.kernel.org, Vincent Mailhol Subject: [PATCH v5 0/7] add the netlink interface for CAN-FD Transmitter Delay Compensation (TDC) Date: Sun, 15 Aug 2021 12:32:41 +0900 Message-Id: <20210815033248.98111-1-mailhol.vincent@wanadoo.fr> X-Mailer: git-send-email 2.31.1 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org The main goal of this series is to add a netlink interface for the TDC parameters using netlink nested attributes. The series also contains a few fix on the current TDC implementation. This series completes the implementation of the Transmitter Delay Compensation (TDC) in the kernel which started in March with those two patches: - commit 289ea9e4ae59 ("can: add new CAN FD bittiming parameters: Transmitter Delay Compensation (TDC)") - commit c25cc7993243 ("can: bittiming: add calculation for CAN FD Transmitter Delay Compensation (TDC)") The netlink interface was missing from this series because the initial patch needed rework in order to make it more flexible for future changes. At that time, Marc suggested to take inspiration from the recently released ethtool-netlink interface (Ref [1]). ethtool uses nested attributes (c.f. NLA_NESTED type in validation policy). A bit of trivia: the NLA_NESTED type was introduced in version 2.6.15 of the kernel and thus actually predates Socket CAN. Ref: commit bfa83a9e03cf ("[NETLINK]: Type-safe netlink messages/attributes interface") Since then, Stephan shared additional remarks for improvement which are addressed in revision v4 of this series [2]. The first patch allow a user to turn off a feature even if not supported (e.g. allow "ip link set can0 type can bitrate 500000 fd off" even if fd is not supported). This feature will be used later by the fifth patch of the series. The second patch allows TDCV and TDCO to be zero (previously, those values were assigned a special meaning). The third patch fixes the unit of the TDC parameters. In fact, those should be measured in clock period (a.k.a. minimum time quantum) and not in time quantum as it is done in current implementation. The fourth patch addresses the concern of Marc and Stefan concerning some devices which use a TDCO value relative to the Sample Point (so far, the TDC implementation used the formula SSP = TDCV + TDCO). To do so, an helper function to convert TDCO from an absolute value to a value relative to the sample point is added. The fifth patch is the real thing: the TDC netlink interface. The sixth patch allows to retrieve TDCV from the device through a callback function. The seventh and last patch does a bit of cleanup in the ES58x driver to remove some redundant TDC information from the documentation. [1] https://lore.kernel.org/linux-can/20210407081557.m3sotnepbgasarri@pengutronix.de/ [2] https://lore.kernel.org/linux-can/75fa87a71a3f5fd7d7407a2c514b71690e56eb4e.camel@esd.eu/ ** Changelog ** v4 -> v5: - move to can_validate() all of the checks on TDC parameters that do not need access to priv. - in can_tdc_get_size(), field IFLA_CAN_TDCV was not taken in account for size calculation when in automatic mode. - if can_tdc_is_enabled(priv) returns true, we know that one of either CAN_CTRLMODE_TDC_AUTO or CAN_CTRLMODE_TDC_MANUAL is set. After confirming that CAN_CTRLMODE_TDC_MANUAL if off, we then implicitly know that CAN_CTRLMODE_TDC_AUTO is set and do not need to check it again. Remove such redundant check in can_tdc_fill_info(). v3 -> v4: - add a patch to the series to change the unit from time quantum to clock period (a.k.a. minimum time quantum). - add a callback function to retrieve TDCV from the device. v2 -> v3: - allows both TDCV and TDCO to be zero. - introduce the can_tdc_get_relative_tdco() helper function - other path of the series modified accordingly to above changes. RFC v1 -> v2: The v2 fixes several issue in can_tdc_get_size() and can_tdc_fill_info(). Namely: can_tdc_get_size() returned an incorrect size if TDC was not implemented and can_tdc_fill_info() did not include a fail path with nla_nest_cancel(). Vincent Mailhol (7): can: netlink: allow user to turn off unsupported features can: bittiming: allow TDC{V,O} to be zero and add can_tdc_const::tdc{v,o,f}_min can: bittiming: change unit of TDC parameters to clock periods can: dev: add can_tdc_get_relative_tdco() helper function can: netlink: add interface for CAN-FD Transmitter Delay Compensation (TDC) can: netlink: add can_priv::do_get_auto_tdcv() to retrieve tdcv from device can: etas_es58x: clean-up documentation of struct es58x_fd_tx_conf_msg drivers/net/can/dev/bittiming.c | 19 +- drivers/net/can/dev/netlink.c | 213 +++++++++++++++++++++- drivers/net/can/usb/etas_es58x/es58x_fd.c | 7 +- drivers/net/can/usb/etas_es58x/es58x_fd.h | 23 +-- include/linux/can/bittiming.h | 80 +++++--- include/linux/can/dev.h | 34 ++++ include/uapi/linux/can/netlink.h | 31 +++- 7 files changed, 353 insertions(+), 54 deletions(-) -- 2.31.1