Date:   Fri, 11 Jan 2019 06:27:02 -0600
From:   Dan Murphy <dmurphy@...com>
To:     Wolfgang Grandegger <wg@...ndegger.com>, <mkl@...gutronix.de>,
        <davem@...emloft.net>
CC:     <linux-can@...r.kernel.org>, <netdev@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [RFC PATCH 1/3] can: m_can: Create m_can core to leverage common
 code

Wolfgang

On 1/11/19 2:27 AM, Wolfgang Grandegger wrote:
> Hello Dan,
> 
> Am 10.01.19 um 13:53 schrieb Dan Murphy:
>> Wolfgang
>>
>> On 1/10/19 1:44 AM, Wolfgang Grandegger wrote:
>>> Hello Dan,
>>>
>>> sorry for my late response on that topic...
>>>
>>> Am 09.01.19 um 21:58 schrieb Dan Murphy:
>>>> Wolfgang
>>>>
>>>> On 11/3/18 5:45 AM, Wolfgang Grandegger wrote:
>>>>> Hello Dan,
>>>>>
>>>>> Am 31.10.2018 um 21:15 schrieb Dan Murphy:
>>>>>> Wolfgang
>>>>>>
>>>>>> Thanks for the review
>>>>>>
>>>>>> On 10/27/2018 09:19 AM, Wolfgang Grandegger wrote:
>>>>>>> Hello Dan,
>>>>>>>
>>>>>>> for the RFC, could you please just do the necessary changes to the
>>>>>>> existing code. We can discuss about better names, etc. later. For
>>>>>>> the review if the common code I quickly did:
>>>>>>>
>>>>>>>   mv m_can.c m_can_platform.c
>>>>>>>   mv m_can_core.c m_can.c
>>>>>>>
>>>>>>> The file names are similar to what we have for the C_CAN driver.
>>>>>>>
>>>>>>>   s/classdev/priv/
>>>>>>>   variable name s/m_can_dev/priv/
>>>>>>>
>>>>>>> Then your patch 1/3 looks as shown below. I'm going to comment on that
>>>>>>> one. The comments start with "***"....
>>>>>>>
>>>>>>
>>>>>> So you would like me to align the names with the c_can driver?
>>>>>
>>>>> That would be the obvious choice.
>>>>>> <snip>
>>>>>>>
>>>>>>> *** I didn't review the rest of the patch for now.
>>>>>>>
>>>>>>
>>>>>> snipped the code to reply to the comment.
>>>>>>
>>>>>>> Looking to the generic code, you didn't really change the way
>>>>>>> the driver is accessing the registers. Also the interrupt handling
>>>>>>> and rx polling is as it was before. Does that work properly using
>>>>>>> the SPI interface of the TCAN4x5x?
>>>>>>
>>>>>> I don't want to change any of that yet.  Maybe my cover letter was not clear
>>>>>> or did not go through.
>>>>>>
>>>>>> But the intention was just to break out the functionality to create a MCAN framework
>>>>>> that can be used by devices that contain the Bosch MCAN core and provider their own protocal to access
>>>>>> the registers in the device.
>>>>>>
>>>>>> I don't want to do any functional changes at this time on the IP code itself until we have a framework.
>>>>>> There should be no regression in the io mapped code.
>>>>>>
>>>>>> I did comment on the interrupt handling and asked if a threaded work queue would affect CAN timing.
>>>>>> For the original TCAN driver this was the way it was implemented.
>>>>>
>>>>> Do threaded interrupts with RX polling make sense? I think we need a
>>>>> common interface allowing to select hard-irqs+napi or threaded-irqs.
>>>>>
>>>>
>>>> I have been working on this code for about a month now and I am *not happy* with the amount of change that needs
>>>> to be done to make the m_can a framework.
>>>>
>>>> I can tx/rx frames from another CAN device to the TCAN part but I have not even touched the iomapped code.
>>>>
>>>> The challenging part is that the m_can code that is currently available does not have to worry about atomic context because
>>>> there is no peripheral waiting.  Since the TCAN is a peripheral device we need to take into about the hard waits in IRQ context
>>>> as well as the atomic context.  Doing this creates many deltas in the base code that may break iomapped devices.  I have had to 
>>>> add the thread_irqs and now I am in the midst of the issue you brought up with napi.  I would have to schedule a queue for perp devices
>>>> and leave the non-threaded iomapped irq.
>>>>
>>>> At this point I think it may be wise to leave the m_can code alone as it is working and stable and just work on the TCAN driver as
>>>> a standalone driver.  A framework would be nice but I think it would destablize the m_can driver which is embedded in many SoC's and
>>>> we cannot possibly test everyone of them.
>>>
>>> Unfortunately, I do not have m_can hardware at hand.
>>>
>>>> What are your thoughts?
>>>
>>> What we need is a common set of functions doing tx, rx, error and state
>>> handling. This will requires substantial changes to the existing
>>> io-mapped m_can driver, of course. I still believe it's worth the
>>> effort, but I agree that it's difficult for you to re-write and test the
>>> existing m_can driver.
>>
>> OK I will keep working on it.  What you are describing is what I have done.
>> I have abstracted the register reads and writes away and I am in the process
>> of abstracting away the device specific initialization.
> 
> Would be nice if you could show your current implementation...
> 

I did submit v2 with the current implementation but the code has changed a bit since then
I clean up the code and post v3 today after I get the Rx working properly.

>>>
>>> What about implementing such a set of common functions plus the SPI
>>> specific part for your TCAN device. What do you/others think?
>>
>> As stated above this is what I have.  But the m_can driver was written for io-mapped that has no delays
>> so we need to take into about peripheral wait time in IRQ and atomic context.
>>
>> This is where the issues are stemming from mainly in the atomic context.
> 
> ... to understand a bit better what you exactly mean. Or does the last
> patch you sent already highlight them.

v3 will show the deltas

Dan

> 
> Wolfgang.
> 


-- 
------------------
Dan Murphy

Powered by blists - more mailing lists