Hi marc, > -----Original Message----- > From: Marc Kleine-Budde <mkl@xxxxxxxxxxxxxx> > Sent: Friday, March 18, 2022 2:01 PM > To: Srinivas Neeli <sneeli@xxxxxxxxxx> > Cc: Srinivas Goud <sgoud@xxxxxxxxxx>; Naveen Kumar Gaddipati > <naveenku@xxxxxxxxxx>; Prasad Beer Prasad <prasadbe@xxxxxxxxxx>; linux- > can <linux-can@xxxxxxxxxxxxxxx>; Stefan Mätje <stefan.maetje@xxxxxx>; > Vincent MAILHOL <mailhol.vincent@xxxxxxxxxx> > Subject: Re: Queries on CAN frame work > > On 18.03.2022 06:12:30, Srinivas Neeli wrote: > > > Are you using the mainline xilinx_can driver? The "ip" command says, > > > you're > > > using: > > > > Yes , mainline and internal drivers are same. > > No, you are not. As explained in my last mail, there's no struct > can_bittiming_const in the mainline driver that matches your "ip" > output. Just cross checked the mainline driver and our internal driver. I am not seeing difference in struct can_bittiming_const structures. We are using 5 structures CAN - 1 structure CANFD 1.0 - Arbitration 1, data phase 1 CANFD 2.0 - Arbitration 1, data phase 1 Please check below link once. https://elixir.bootlin.com/linux/latest/source/drivers/net/can/xilinx_can.c#L216 Could you please let me know if I miss anything? > > > Xilinx maintains common driver for multiple CAN and CANFD IP's. > > Xilinx supports below CAN IP's: > > xlnx,zynq-can-1.0 (zynq platform & zynqmp platform) > > xlnx,axi-can-1.00.a ( Soft IP version 1.00) > > xlnx,canfd-1.0 (Soft canfd ip version 1.0) xlnx,canfd-2.0(Soft canfd > > ip Version 2.0) > > xlnx,canfd2.0 equal to versal canfd ip, so using same compatible string for > both. > > > > For each ip sjw_max and brp values are different. > > Due to this reason created multiple bit timing structures. > > ACK. But your driver is obviously different to the mainline driver (see above). > > > > > sjw 1..128 brp 1..256 > > > > dsjw 1..16 dbrp 1..256 > > > > > > Looking at the mainline xilinx driver, we have: > > > > > > | /* CAN Bittiming constants as per Xilinx CAN specs */ static const > > > | struct can_bittiming_const xcan_bittiming_const = { > > > | .name = DRIVER_NAME, > > > | .tseg1_min = 1, > > > | .tseg1_max = 16, > > > | .tseg2_min = 1, > > > | .tseg2_max = 8, > > > | .sjw_max = 4, > > > ^ > > > Does not match due to "sjw 1..128". > > > > > > | .brp_min = 1, > > > | .brp_max = 256, > > > | .brp_inc = 1, > > > | }; > > > | > > > | /* AXI CANFD Arbitration Bittiming constants as per AXI CANFD 1.0 > > > | spec */ static const struct can_bittiming_const > xcan_bittiming_const_canfd = { > > > | .name = DRIVER_NAME, > > > | .tseg1_min = 1, > > > | .tseg1_max = 64, > > > | .tseg2_min = 1, > > > | .tseg2_max = 16, > > > | .sjw_max = 16, > > > ^^ > > > Does not match due to "sjw 1..128". > > > > > > | .brp_min = 1, > > > | .brp_max = 256, > > > | .brp_inc = 1, > > > | }; > > > | > > > | /* AXI CANFD Data Bittiming constants as per AXI CANFD 1.0 specs > > > | */ static struct can_bittiming_const xcan_data_bittiming_const_canfd = > { > > > | .name = DRIVER_NAME, > > > | .tseg1_min = 1, > > > | .tseg1_max = 16, > > > | .tseg2_min = 1, > > > | .tseg2_max = 8, > > > | .sjw_max = 8, > > > ^ > > > Does not match due to "dsjw 1..16" > > > > > > | .brp_min = 1, > > > | .brp_max = 256, > > > | .brp_inc = 1, > > > | }; > > > | > > > | /* AXI CANFD 2.0 Arbitration Bittiming constants as per AXI CANFD > > > | 2.0 spec */ static const struct can_bittiming_const > > > xcan_bittiming_const_canfd2 = { > > > | .name = DRIVER_NAME, > > > | .tseg1_min = 1, > > > | .tseg1_max = 256, > > > | .tseg2_min = 1, > > > | .tseg2_max = 128, > > > | .sjw_max = 128, > > > | .brp_min = 2, > > > ^ > > > Does not match due to "brp 1..256". > > > > > > | .brp_max = 256, > > > | .brp_inc = 1, > > > | }; > > > | > > > | /* AXI CANFD 2.0 Data Bittiming constants as per AXI CANFD 2.0 > > > | spec */ static struct can_bittiming_const > xcan_data_bittiming_const_canfd2 = { > > > | .name = DRIVER_NAME, > > > | .tseg1_min = 1, > > > | .tseg1_max = 32, > > > | .tseg2_min = 1, > > > | .tseg2_max = 16, > > > | .sjw_max = 16, > > > | .brp_min = 2, > > > ^ > > > Does not match due to "dbrp 1..256". > > > > Versal early silicon we observed bit enlarging issue with brp =1 in > > side IP using simulaitons(Not observed on software side) , but it > > fixed in production silicon. > > > > Plan to revert this patch to support TDC feature(To support higher > > baud rates we need brp =1). > > Ok. > > regards, > 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 |
