[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABGWkvq2nbfYRww0KWB1YxLQ92hjsWjfV9+nT-cKzc5FPF=DzA@mail.gmail.com>
Date: Thu, 9 Jun 2022 10:52:03 +0200
From: Dario Binacchi <dario.binacchi@...rulasolutions.com>
To: Marc Kleine-Budde <mkl@...gutronix.de>
Cc: linux-kernel@...r.kernel.org,
Amarula patchwork <linux-amarula@...rulasolutions.com>,
michael@...rulasolutions.com,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Jakub Kicinski <kuba@...nel.org>,
Jiri Slaby <jirislaby@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
Vincent Mailhol <mailhol.vincent@...adoo.fr>,
Wolfgang Grandegger <wg@...ndegger.com>,
linux-can@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: [RFC PATCH 11/13] can: slcan: add ethtool support to reset
adapter errors
Hi Marc,
On Thu, Jun 9, 2022 at 10:01 AM Marc Kleine-Budde <mkl@...gutronix.de> wrote:
>
> On 09.06.2022 09:24:14, Dario Binacchi wrote:
> > > > > I'm a big fan of bringing the device into a well known good state during
> > > > > ifup. What would be the reasons/use cases to not reset the device?
> > > >
> > > > Because by default either slcand and slcan_attach don't reset the
> > > > error states, but you must use the `-f' option to do so. So, I
> > > > followed this use case.
> > >
> > > Is this a CAN bus error state, like Bus Off or some controller (i.e. non
> > > CAN related) error?
> >
> > The help option of slcan_attach and slcand prints " -f (read status
> > flags with 'F\\r' to reset error states)\n" I looked at the sources of
> > the adapter I am using (USBtin, which uses the mcp2515 controller).
> > The 'F' command reads the EFLG register (0x2d) without resetting the
> > RX0OVR and RX1OVR overrun bits.
>
> The Lawicel doc [1] says 'F' is to read the status flags not to clear
> it. However commit 7ef581fec029 ("slcan_attach: added '-f' commandline
> option to read status flags") [2] suggests that there are some adapters
> that the reading of the status flag clears the errors. IMHO the 'F'
> command should be send unconditionally during open.
When in doubt I would follow the slcand / slcan_attach approach, that don't
send the 'F \ r' command by default. We would be more backward compatible
as regards the sequence of commands to be sent to the controller.
>
> [1] http://www.can232.com/docs/can232_v3.pdf
> [2] https://github.com/linux-can/can-utils/commit/7ef581fec0298a228baa71372ea5667874fb4a56
>
> > The error states reset is done by 'f <subcmd>' command, that is not
> > managed by slcan_attach/slcand.
>
> Is the 'f' command is non-standard?
It's possible.
Thanks and regards,
Dario
>
> > switch (subcmd) {
> > case 0x0: // Disable status reporting
> > mcp2515_write_register(MCP2515_REG_CANINTE, 0x00);
> > return CR;
> > case 0x1: // Enable status reporting
> > mcp2515_write_register(MCP2515_REG_CANINTE, 0x20); //
> > ERRIE interrupt to INT pin
> > return CR;
> > case 0x2: // Clear overrun errors
> > mcp2515_write_register(MCP2515_REG_EFLG, 0x00);
> > return CR;
> > case 0x3: // Reinit/reset MCP2515 to clear all errors
> > if (state == STATE_CONFIG) {
> > mcp2515_init();
> > return CR;
> > }
> > break;
> > }
>
> Marc
>
> --
> Pengutronix e.K. | Marc Kleine-Budde |
> Embedded Linux | https://www.pengutronix.de |
> Vertretung West/Dortmund | Phone: +49-231-2826-924 |
> Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
--
Dario Binacchi
Embedded Linux Developer
dario.binacchi@...rulasolutions.com
__________________________________
Amarula Solutions SRL
Via Le Canevare 30, 31100 Treviso, Veneto, IT
T. +39 042 243 5310
info@...rulasolutions.com
www.amarulasolutions.com
Powered by blists - more mailing lists