Re: [PATCH 0/6] Add Microchip MCP25XXFD CAN driver

[Marc Kleine-Budde]

On 6/11/20 10:30 PM, Marc Kleine-Budde wrote:
> On 6/11/20 6:26 PM, Marc Kleine-Budde wrote:
>> I initially started looking at Martin's driver and it was not using several
>> modern CAN driver infrastructures. I then posted some cleanup patches but Martin
>> was not working on the driver any more. Then I decided to rewrite the driver,
>> that is the one I'm hoping to mainline soon.
> 
> Seems the driver still suffers from the same problems the last time I looked at it:
> 
> I'm on a raspi4 with a mcp2518fd connected to spi0 cs0 and cs1.
> 
> Running a canfdtest -vg can0; canfdtest -v can1. It  runs into this problem on a
> unloaded system and scaling_governor=performance  within minutes:
> 
>> Jun 11 21:27:08 rpi4 kernel: mcp25xxfd spi0.0 can1: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
>> Jun 11 21:27:08 rpi4 kernel: mcp25xxfd spi0.0 can1: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
>> Jun 11 21:27:08 rpi4 kernel: mcp25xxfd spi0.0 can1: tefif: fifo 6 not pending - tef data: id: 00000078 flags: 00000c08, ts: 0addbed3 - this may be a problem with spi signal quality- try reducing spi-clock speed if this can get reproduced
>> Jun 11 21:27:08 rpi4 kernel: mcp25xxfd spi0.0 can1: tefif: fifo 7 not pending - tef data: id: 00000078 flags: 00000e08, ts: 0addc2e5 - this may be a problem with spi signal quality- try reducing spi-clock speed if this can get reproduced
>> Jun 11 21:27:08 rpi4 kernel: mcp25xxfd spi0.0 can1: tefif: fifo 1 not pending - tef data: id: 00000078 flags: 00000208, ts: 0addc701 - this may be a problem with spi signal quality- try reducing spi-clock speed if this can get reproduced
>> Jun 11 21:27:08 rpi4 kernel: mcp25xxfd spi0.0 can1: tefif: fifo 2 not pending - tef data: id: 00000078 flags: 00000408, ts: 0addcb1b - this may be a problem with spi signal quality- try reducing spi-clock speed if this can get reproduced

Further two mcp2517 on the rpi4, unloaded system, performance dovernor running
cansequence -v can0 -p; cansequence can1 -v; gives:

> Jun 12 12:17:53 rpi4 kernel: mcp25xxfd spi1.0: ECC single bit error at 800
> Jun 12 12:17:53 rpi4 kernel: mcp25xxfd spi1.0: ECC single bit error at 800
> Jun 12 12:17:54 rpi4 kernel: mcp25xxfd spi1.0: ECC double bit error at 800
> Jun 12 12:17:54 rpi4 kernel: mcp25xxfd spi1.0: unidentified system interrupt - intf =  b91a1118
> Jun 12 12:17:54 rpi4 kernel: mcp25xxfd spi1.0: ECC double bit error at 800

I don't see these problems with my driver.

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 |