RE: [PATCH] ARM: OMAP: AM33XX: CAN: d_can: Add support for BoschD_CAN controller

From: AnilKumar, Chimata
Date: Tue Apr 03 2012 - 10:30:13 EST


On Tue, Apr 03, 2012 at 19:19:32, Marc Kleine-Budde wrote:
> On 04/03/2012 03:41 PM, AnilKumar, Chimata wrote:
> > Hi Wolfgang,
> >
> > Thanks for reviewing the patch
> >
> > On Tue, Apr 03, 2012 at 18:14:55, Wolfgang Grandegger wrote:
> >> On 04/03/2012 02:32 PM, AnilKumar Ch wrote:
> >>> This patch adds the support for Bosch D_CAN controller.
> >>>
> >>> Bosch D_CAN controller is a full-CAN implementation compliant to
> >>> CAN protocol version 2.0 part A and B. Bosch D_CAN user manual
> >>> can be obtained from: http://www.semiconductors.bosch.de/media/
> >>> en/pdf/ipmodules_1/can/d_can_users_manual_111.pdf
> >>>
> >>> D_CAN device is used on many SoCs like AM335x, DM8148 and DM813x
> >>> EVMs from TI, D_CAN details on AM335x can be accessed from:
> >>> http://www.ti.com/lit/ug/spruh73c/spruh73c.pdf
> >>>
> >>> D_CAN can be configurable for 16, 32, 64 and 128 message objects.
> >>> The driver implementation is based on 64 message objects.
> >>>
> >>> Following are the design choices made while writing the controller
> >>> driver:
> >>> 1. Interface Register set IF0 has be used for receive and IF1 is
> >>> used for transmit message objects.
> >>> 2. Out of the total Message objects available, half of it are kept
> >>> aside for RX purposes and the rest for TX purposes.
> >>> 3. NAPI implementation is such that both the TX and RX paths
> >>> functions in polling mode.
> >>>
> >>> Signed-off-by: AnilKumar Ch <anilkumar@xxxxxx>
> >>
> >> Please explain why this CAN controller cannot be handled by the existing
> >> C_CAN driver, eventually with some extensions. The register layout seems
> >> almost identical, at least.
> >>
> >> Wolfgang.
> >>
> >
> > These are the some of the pointers I can say, why I had gone for separate
> > file instead of existing driver:
> > * In case of D_CAN driver we can see all the registers are 32bit length
> > but in case of C_CAN registers are in 16bit length.
>
> How many bits in these 32 bit registers are used?

In some cases (D_CAN_TXRQ, D_CAN_INTPND, D_CAN_MSGVAL) I have used all the
bits, in some cases used few bits.

Roughly I can say that its (higher 16bits) % of usages is similar as compare
to 16bits

While checking the status of TXRequest registers and INT pending register,
which is a hot code path, we have to put if checks for register access.

>
> > * Some of the registers, bit masks are different, so we have to add
> > checks on every API for differentiating the kind of device
>
> Which registers are this? Can you give us an example?

I am pointing out some of the resisters
* Single registers in case of D_CAN but multiple register in case of C_CAN
So masks will change MASK, ARB, INTPND
* D_CAN_IFCMD is the combination of COMM request and COMM mask registers

>
> > * In case of D_CAN we have some extra features like direct message RAM
> > access, DMA support, TX/RX pins can be used as GPIO lines (if applicable),
> > more interrupt lines and three sets of interface registers.
>
> Which of these features are used in you driver?

Right now I added only CAN functionality, in that I am using two set
of interface registers (0 and 1).

Based on other features status register are extended, we have to handle those
as well.

>
> > * Wait timings while init bit set/reset during bit-timing initialization
> > are different in both the cases
>
> That's not the hot code path, so some ifs shouldn't hurt.
>
> > * bittiming configurations are different.
>
> see above.
>
> Marc
> --
> Pengutronix e.K. | Marc Kleine-Budde |
> Industrial Linux Solutions | Phone: +49-231-2826-924 |
> Vertretung West/Dortmund | Fax: +49-5121-206917-5555 |
> Amtsgericht Hildesheim, HRA 2686 | http://www.pengutronix.de |
>
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/