Hello,
-----Messaggio originale-----
Da: Wolfgang Grandegger [mailto:[email protected]]
Inviato: giovedì 24 settembre 2009 14.49
A: SETTE AGOSTINO - technolabs
Cc: [email protected]
Oggetto: Re: CC770 and AN82527
Hello,
SETTE AGOSTINO - technolabs wrote:
> Hi All,
>
> I have a card with a Bosch CC770 CAN Controller but I have some
> problems:
>
> 1. I Applied the patch suggested by BOSCH, Ref. CC770 (Stand Alone CAN
> Controller) Data Sheet Revision 1.6 30.03.2009
Interesting, could you give me the link to that patch?
[Agostino] I do not have a link up to now, I have only a pdf file named
"13464_Datasheet.pdf"
titled "CC770 (Stand Alone CAN Controller) Data Sheet Revision 1.6
20.03.2009"(I made
a mistake before on the date). Concerning the Software patch I had
an old code it seems to work
but in my opinion in not exactly what suggested by Bosch. For this
reason I wrote this e-mail, to
double check with the mailing list if I am right or wrong. Here is
the management of the bus_off error:
if (status_reg&iSTAT_BOFF) {
flags = can_read_reg(chip, iCTL) &
(iCTL_IE|iCTL_SIE|iCTL_EIE|iCTL_CCE|iCTL_INI);
can_write_reg(chip, flags, iCTL);
CANMSG("CAN Detecting BusOff - Error count >=256 boff_count %d \n",
chip->canstat.boff_count);
chip->canstat.boff_count++;
flags = can_read_reg(chip, iCTL) &
(iCTL_IE|iCTL_SIE|iCTL_EIE|iCTL_CCE|~iCTL_INI);
can_write_reg(chip, flags, iCTL);
}
>
> 9.1 Description
>
> In chapter chapter 4.2 the description of the EIE bit suggests,
> that a change in the error status
> will cause an interrupt to be generated. That means, if bits BOff
> or Warn in Status register
> (01H) are changing from 0 to 1 or vice versa from 1 to 0 an
> interrupt will be generated. In
> such a case, usually an interrupt routine reads the Interrupt
> register (5FH) to determine the
> interrupt source and then reads the Status register (01H) to
> acknowledge the interrupt and
> this deactivates the interrupt line.
> The behaviour of the error status Warn is as described, but the
> behaviour of the error status
> BOff is not as described. The error status BOff does not trigger
> the Interrupt register by a
> change (edge triggered), but by level. Thus, in Bus Off state, the
> Interrupt register is forced
> to 01H and hence the interrupt line remains active. The only way to
> deactivate the interrupt
> line is to reset the IE bit in the Control register, a read to the
> Status register has no effect.
>
> 9.2 Work-around
>
> On occurrence of a Bus Off interrupt, the CPU shall reset the IE
> bit in the Control register to
> deactivate the interrupt line. To re-synchronize the CC770 to the
> CAN bus, the Bus Off
> Recovery Sequence shall be started by resetting the Init bit in the
> Control register. After
> successful resynchronisation, the BOff flag in the Control register
> will be reset by CC770.
> The software shall poll for this state and, if reached, it may
> re-enable the CC770 interrupt by
> setting IE bit in the Control register.
> Please note that leaving BusOff state (resetting BOff and Warn
> bits) will also cause an Error
> interrupt. In this case, reading the Status register clears the
> Interrupt register.
>
> And it works properly. Unfortunately now I have some "Stuff Error"
> and "Form Error".
> The causes of these errors, AFAIK, are:
>
> Stuff Error
> 1. a disturbance generates more consecutive bits of equal level
> than allowed by the rule of
> bit stuffing and therefore the error is detected by all nodes.
> 2. a disturbance falsifies one or more of the five bits preceeding
> the stuff bit. This error
> is not recognized by a receiver; however, if the error also
> appears at the transmitter,
> it will be detected as a bit error (transmitter monitors bus as
> it transmits).
>
> In any case, the error is detected by a receiver either by the bit
> stuffing mechanism (the
> stuff bit of the transmitter is not dropped but taken as an
> information bit) or by the CRC check.
>
> Form Error
> Form Errors result from the violation of the fixed form of the
> following bit fields:
> - end of frame
> - interframe space
> - ACK delimiter
> - CRC delimiter
> During the transmission of these bit fields, an error condition is
> recognized if a "dominant" bit
> level is detected.
>
> My question is: are these errors detected on the TX side or on the
> RX side? It seems to be in TX, as far
> as I can understand but I would like to have a confirmation.
I will first answer the following questions.
> 2. Where can I find the Linux drivers for the Bosch CC770?
There is a Socket-CAN driver for the CC770 or AN82527 in the SVN repository at
BerliOS:
http://developer.berlios.de/projects/socketcan/
http://svn.berlios.de/svnroot/repos/socketcan/trunk/kernel/2.6/drivers/net/can/cc770/
How is the chip connected to your system (what bus)?
[Agostino] the connection between Microprocessor MPC852 and the Bosch CC770 CAN
controller is
through a parallel non multiplexed interface (8 data and 8 address).
> 3. I would like to know how the Bus_off problem in the interrupt
> handler is managed in the kernel tree.
I will have a close look to what you are speaking about later on.
Wolfgang.
_______________________________________________
Socketcan-core mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/socketcan-core
