Re: [PATCH v3 13/15] can: grcan: Add CANFD TX support alongside legacy CAN

[Arun Muthusamy <arun.muthusamy@gaisler.com>]

On 1/27/26 15:12, Marc Kleine-Budde wrote:
> On 27.01.2026 15:06:55, Arun Muthusamy wrote:
>> On 1/23/26 15:25, Marc Kleine-Budde wrote:
>>> On 22.01.2026 13:10:36, Arun Muthusamy wrote:
>>>> Include CANFD TX support with the legacy CAN support, enabling
>>>> support for extended data payloads to provide higher bit rates.
>>>>
>>>> Signed-off-by: Arun Muthusamy <arun.muthusamy@gaisler.com>
>>>> ---
>>>>    drivers/net/can/grcan.c | 102 +++++++++++++++++++++++++++++-----------
>>>>    1 file changed, 74 insertions(+), 28 deletions(-)
>>>>
>>>> diff --git a/drivers/net/can/grcan.c b/drivers/net/can/grcan.c
>>>> index 9fc18064fab1..508ad5320cff 100644
>>>> --- a/drivers/net/can/grcan.c
>>>> +++ b/drivers/net/can/grcan.c
>>>> @@ -174,6 +174,7 @@ struct grcan_registers {
>>>>    #define GRCAN_IRQ_DEFAULT (GRCAN_IRQ_RX | GRCAN_IRQ_TX | GRCAN_IRQ_ERRORS)
>>>>
>>>>    #define GRCAN_MSG_SIZE		16
>>>> +#define GRCAN_CLASSIC_DATA_SIZE 8
>>>>
>>>>    #define GRCAN_MSG_IDE		0x80000000
>>>>    #define GRCAN_MSG_RTR		0x40000000
>>>> @@ -195,6 +196,10 @@ struct grcan_registers {
>>>>    #define GRCAN_MSG_OFF		0x00000002
>>>>    #define GRCAN_MSG_PASS		0x00000001
>>>>
>>>> +#define GRCAN_MSG_EID_MASK      GENMASK(28, 0)
>>>> +#define GRCAN_MSG_BID_MASK      GENMASK(28, 18)
>>>> +#define GRCAN_MSG_DLC_MASK      GENMASK(31, 28)
>>>> +
>>>>    #define GRCAN_BUFFER_ALIGNMENT		1024
>>>>    #define GRCAN_DEFAULT_BUFFER_SIZE	1024
>>>>    #define GRCAN_VALID_TR_SIZE_MASK	0x001fffc0
>>>> @@ -227,6 +232,9 @@ struct grcan_registers {
>>>>    #define GRCANFD_FDBTR_PS2_BIT 5
>>>>    #define GRCANFD_FDBTR_SJW_BIT 0
>>>>
>>>> +#define GRCAN_TX_BRS  BIT(25)
>>>> +#define GRCAN_TX_FDF  BIT(26)
>>>> +
>>>>    /* Hardware capabilities */
>>>>    struct grcan_hwcap {
>>>>    	/* CAN-FD capable, indicates GRCANFD IP.
>>>> @@ -1218,6 +1226,13 @@ static void grcan_transmit_catch_up(struct net_device *dev)
>>>>    	spin_unlock_irqrestore(&priv->lock, flags);
>>>>    }
>>>>
>>>> +static int grcan_numbds(int len)
>>>> +{
>>>> +	if (len <= GRCAN_CLASSIC_DATA_SIZE)
>>>> +		return 1;
>>>> +	return 1 + ((len - GRCAN_CLASSIC_DATA_SIZE + GRCAN_MSG_SIZE) / GRCAN_MSG_SIZE);
>>> This looks strange, what is calculated here? Why is "<=
>>> GRCAN_CLASSIC_DATA_SIZE" a special case?
>> grcan_numbds() calculates the number of buffer descriptors needed based on the data length.
>> The condition "len <= GRCAN_CLASSIC_DATA_SIZE" addresses cases where the data length fits within a one buffer descriptor.
>> For data lengths greater than "GRCAN_CLASSIC_DATA_SIZE", it computes additional descriptors needed.
> What happens if "len = GRCAN_CLASSIC_DATA_SIZE + GRCAN_MSG_SIZE"?
>
> | return 1 + ((len - GRCAN_CLASSIC_DATA_SIZE + GRCAN_MSG_SIZE) / GRCAN_MSG_SIZE);
>
> = 1 + ((GRCAN_CLASSIC_DATA_SIZE + GRCAN_MSG_SIZE - GRCAN_CLASSIC_DATA_SIZE + GRCAN_MSG_SIZE) / GRCAN_MSG_SIZE);
> = 1 + ((                          GRCAN_MSG_SIZE -                         + GRCAN_MSG_SIZE) / GRCAN_MSG_SIZE);
> = 1 + 2;
> = 3
>
> Is this correct?
>
Thanks for pointing out.
That's not correct. It will be addressed in the next patch series.

-- 
-- 
BR,

Arun Muthusamy
Software Engineer
Frontgrade Gaisler
T : +46 (0) 700 558 528
arun.muthusamy@gaisler.com

Frontgrade Gaisler AB, Kungsgatan 12, SE-411 19 GÖTEBORG, Sweden.
+46 (0) 31 775 8650, www.gaisler.com