IMX6Q rev 1.6 hangs randomly custom board with 2GB DDR

[agnel_terry]

Hi ,

In one of our products which runs on imx6Q rev 1.6 with 2GB DDR, has random hangs. we have primary ethernet interface eth0 through PCIE +intel 120 and the FEC is configured as eth1 +micrel Ksz9031 PHY.

No stress testing done, just powered on and PING has been done through eth0. the unit is in idle state almost 95%.

We see random hangs and some units exhibit RCU preempt CPU Stall lockup. 

I learned from the forum that imx6Q rev 1.0 and rev1.1 had random hangs when entering and exiting from idle state. the workaround was to comment out CPU idle in arch/arm.

Can entering and exiting idle state cause power ripple?  we are powering through external PMIC PFUZE100.

can you provide any suggestion or debugging steps to try out. I am not able to reproduce the issue frequently. sometimes it fails in a week some time it takes more than 30 days.

I am running a test on the units by increasing the VDDARM from 1.375 to 1.4 and running at 792Mhz.

is there any other recommended steps I can try ?

Note: Same Product has a version with IMX6SOLO +1GB DDR +single ethernet port through PCIE + FEC is disabled in that products, I am not seeing this issue in those boards.

Thanks and regards

Terry

[Bio_TICFSL]

Hello,

Just for the top of my head which bsp version are you using? since it similar happens with previous version 5.15.71 and 4.14v. But is now suppose to be fixed.

Regards

[agnel_terry]

Hi 

Thank you for the response.

We are using  linux kernel version rel_imx_5.4.70_2.3.0  from NXP git hub.

GitHub - nxp-imx/linux-imx at rel_imx_5.4.70_2.3.0

When you mean fixed now, in which version it is fixed ?  could you please give more info about what is fixed . ?

Highly appreciate your help.

Regards

Terry

[agnel_terry]

Can you provide some details about the fix done for Kernel hang with IMX6Q.

[agnel_terry]

HI,

Can someone have some suggestion for RCU stalls ? with IMX6Q.

its mostly failing. on the RCU layer.

Thanks
Terry