linux-kernel - Re: [PATCH v2] can: m_can: don't enable transceiver when probing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Wed, 8 May 2024 13:44:00 +0100
From: Simon Horman <horms@...nel.org>
To: Martin Hundebøll <martin@...nix.com>
Cc: Chandrasekar Ramakrishnan <rcsekar@...sung.com>,
	Marc Kleine-Budde <mkl@...gutronix.de>,
	Vincent Mailhol <mailhol.vincent@...adoo.fr>,
	"David S. Miller" <davem@...emloft.net>,
	Eric Dumazet <edumazet@...gle.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Markus Schneider-Pargmann <msp@...libre.com>,
	linux-can@...r.kernel.org, netdev@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] can: m_can: don't enable transceiver when probing

On Wed, May 01, 2024 at 02:42:03PM +0200, Martin Hundebøll wrote:
> The m_can driver sets and clears the CCCR.INIT bit during probe (both
> when testing the NON-ISO bit, and when configuring the chip). After
> clearing the CCCR.INIT bit, the transceiver enters normal mode, where it
> affects the CAN bus (i.e. it ACKs frames). This can cause troubles when
> the m_can node is only used for monitoring the bus, as one cannot setup
> listen-only mode before the device is probed.
> 
> Rework the probe flow, so that the CCCR.INIT bit is only cleared when
> upping the device. First, the tcan4x5x driver is changed to stay in
> standby mode during/after probe. This in turn requires changes when
> setting bits in the CCCR register, as its CSR and CSA bits are always
> high in standby mode.
> 
> Signed-off-by: Martin Hundebøll <martin@...nix.com>
> ---
> 
> Changes since v1:
>  * Implement Markus review comments:
>    - Rename m_can_cccr_wait_bits() to m_can_cccr_update_bits()
>    - Explicitly set CCCR_INIT bit in m_can_dev_setup()
>    - Revert to 5 timeouts/tries to 10
>    - Use m_can_config_{en|dis}able() in m_can_niso_supported()
>    - Revert move of call to m_can_enable_all_interrupts()
>    - Return -EBUSY on failure to enter normal mode
>    - Use tcan4x5x_clear_interrupts() in tcan4x5x_can_probe()
> 
>  drivers/net/can/m_can/m_can.c         | 131 +++++++++++++++-----------
>  drivers/net/can/m_can/tcan4x5x-core.c |  13 ++-
>  2 files changed, 85 insertions(+), 59 deletions(-)
> 
> diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c

..

> @@ -1694,8 +1708,12 @@ static int m_can_dev_setup(struct m_can_classdev *cdev)
>  		return -EINVAL;
>  	}
>  
> -	if (cdev->ops->init)
> -		cdev->ops->init(cdev);
> +	/* Forcing standby mode should be redunant, as the chip should be in

Hi Martin,

A minor nit, to consider fixing if you send a v3 for some other reason:
redunant -> redundant.

> +	 * standby after a reset. Write the INIT bit anyways, should the chip
> +	 * be configured by previous stage.
> +	 */
> +	if (!m_can_cccr_update_bits(cdev, CCCR_INIT, CCCR_INIT))
> +		return -EBUSY;
>  
>  	return 0;
>  }

..