Re: CM-ITC, pch_can/c_can_pci, sendto() returning ENOBUFS

[Jacob Kroon <jacob.kroon@gmail.com>]

On 9/2/22 17:13, Jacob Kroon wrote:
> Hi Oliver,
> 
> On 9/1/22 18:35, Oliver Hartkopp wrote:
>> Hi Jacob,
>>
>> On 01.09.22 11:38, Jacob Kroon wrote:
>>> On 8/30/22 21:15, Oliver Hartkopp wrote:
>>
>>>> I assume you have a transceiver, right? ;-)
>>>>
>>>
>>> Yes,, all nodes are using a TJA1050 transceiver
>>> (https://www.nxp.com/docs/en/data-sheet/TJA1050.pdf)
>>
>> Good!
>>
>>>> What is the other endpoint? The EG20T and another (automotive) ECU?
>>>
>>> Currently I have 4 nodes in the network, EG20T is in one end.
>>>
>>
>> Ok, that's a good base for testing.
>>
>>>>>> Do you have another CAN node which can be attached to the EG20T 
>>>>>> setup (e.g. some ECU or an USB CAN adapter)?
>>>>>
>>>>> Yes I do have a CAN analyser from Microchip. I guess I can record 
>>>>> all traffic with the analyzer, and compare it to what I see with 
>>>>> "candump can0" on the host. Or do you have some other suggestion ?
>>>>
>>>> Yes, please add the CAN analyzer from Microchip too!
>>>>
>>>> The problem with only two nodes is that you have to be very precise 
>>>> with bitrate settings and sampling points so that the receiving node 
>>>> needs to properly set the ACK to acknowlege the CAN frame.
>>>>
>>>> I had been working with a MSCAN system some time ago and that wasn't 
>>>> able to talk to a commercial CAN tool until I added another node 
>>>> (from another CAN tool provider).
>>>>
>>>> Maybe you can make the other node talk to the Microchip CAN analyzer 
>>>> and let the EG20T receive that traffic first.
>>>>
>>>
>>> I used "candump can0 -l" on the EG20T host to capture the traffic, 
>>> and then connected an CAN USB analyzer to the network and used that 
>>> to capture the traffic. One thing sticks out. This is the log from 
>>> the CAN USB analyzer:
>>>
>>>> ...
>>>> 505.7052;RX;0x464;3;0x01;0x01;0x00;;;;;;
>>>> 505.7052;RX;0x464;3;0x00;0x00;0x00;;;;;;
>>>> 505.7063;RX;0x65;64;;;;;;;;;
>>
>> What should this be?
>>
>> A length of 64 and no data ??
>>
>> This is no valid CAN frame.
>>
>>>> 505.7662;RX;0x440;3;0x32;0x20;0xFA;;;;;;
>> (..)
>>
>>>
>>> Note the third message from the top. This is what "candump" on the 
>>> host logs:
>>>
>>>> ...
>>>> (1662022485.638794) can0 464#010100
>>>> (1662022485.638940) can0 464#000000
>>>> (1662022485.699405) can0 440#3220FA
>>
>> The correct CAN frames are displayed correctly.
>>
>>>> ...
>>>
>>> It fails to see the 3rd message from the previous log. What would 
>>> that indicate ? The CAN analyzer sees the message, but the EG20T 
>>> doesn't.
>>
>> Don't know if this is an error on the CAN bus. You can also print 
>> error messages of detected CAN bus problems with adding an error 
>> message filter.
>>
>> See 'candump -h' :
>>
>> candump -l any,0:0,#FFFFFFFF
>>           (log error frames and also all data frames)
>>
>>
> 
> Thank you Oliver for all the good hints.
> 
> I've done some more logging, but there are no error frames being logged.
> 
> I can see that both pch_can and c_can_pci drivers call 
> can_put_echo_skb() in their ndo_start_xmit functions, but neither checks 
> the return value whether it succeeded or not. Shouldn't both these 
> return NETDEV_TX_BUSY if there are no echo slots available ?
> 
> One reason I ask is because whenever I strace the application, it would 
> seem the problem goes away, and I'm guessing strace:ing will slow down 
> my application.
> 

I did try the patch below, but then I just get the lockups without the 
warning messages:

> diff --git a/drivers/net/can/pch_can.c b/drivers/net/can/pch_can.c
> index 964c8a09226a..0a230368c443 100644
> --- a/drivers/net/can/pch_can.c
> +++ b/drivers/net/can/pch_can.c
> @@ -889,6 +889,10 @@ static netdev_tx_t pch_xmit(struct sk_buff *skb, struct net_device *ndev)
>                 return NETDEV_TX_OK;
>  
>         tx_obj_no = priv->tx_obj;
> +
> +       if (priv->can.echo_skb[tx_obj_no - PCH_RX_OBJ_END - 1])
> +               return NETDEV_TX_BUSY;
> +
>         if (priv->tx_obj == PCH_TX_OBJ_END) {
>                 if (ioread32(&priv->regs->treq2) & PCH_TREQ2_TX_MASK)
>                         netif_stop_queue(ndev);

Jacob