Hi Ramon,

On 13/05/24 7:20 pm, Ramón Nordin Rodriguez wrote:

EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe

On Mon, May 13, 2024 at 03:00:48PM +0200, Andrew Lunn wrote:

quoted

quoted

I've enabled some debugging options but so far nothing seems to hit.
What I've been able to conclude is that there still is SPI
communication, the macphy interrupt is still pulled low, and the cpu
does the ack so that it's reset to inactive.

Is it doing this in an endless cycle?

Exactly, so what I'm seeing is when the driver livelocks the macphy is
periodically pulling the irq pin low, the driver clears the interrupt
and repeat.

If I understand correctly, you are keep on getting interrupt without 
indicating anything in the footer?. Are you using LAN8650 Rev.B0 or B1?. 
If it is B0 then can you try with Rev.B1 once?

Best regards,
Parthiban V

quoted

Probably the debug tools are not showing anything because it is not
looping in just one location. It is a complex loop, interrupts
triggering a thread which runs to completion etc. So it looks like
normal behaviour.

Gotcha. The 'do work' func called in the worker threads loop does run
and return, so I guess there is not much to trigger on.

quoted

If it is an endless cycle, it sounds like an interrupt storm. Some
interrupt bit is not getting cleared, so it immediately fires again as
soon as interrupts are enabled.

Good input. I'll add some instrumentation/stats for how many jiffies
have elapsed between releases of the worker thread and for the irq
handler. I can probably find a gpio to toggle as well if it's really
tight timings.

The irq pin is inactive/high for 100s of us to ms in the measurments
I've done. But I've been using multiple channels and not the fanciest
equipment so samplerates might be playing tricks, I'll rerun some tests
while only measuring the irq pin.

quoted

Is this your dual device board? Do you have both devices on the same
SPI bus? Do they share interrupt lines?

It's on the dual device board, the macphys are using separate spi buses,
one chip shares the bus with another spi device, but the other is the
only tenant on the bus.

No device shares an irq line.

Pretty sure I can replicate the result for both devices, but need to
double check, been to much testing of random things for me to keep track.

I'll do some more digging, I think we're getting pretty close to
understading the behaviour now.

R