Board Reboot while performing stress Test

cancel
Showing results for 
Search instead for 
Did you mean: 

Board Reboot while performing stress Test

Jump to solution
565 Views
himanshugusain
Contributor III

Hi

 

We have custom board based on mx6qsabresd board design

We use Temple run apk to test the board performance

Few of our boards are rebooting when the application runs

 

We suspected the DDR can be an issue,So we did a DDR test

that also passed

What can be the reason

Attaching the DDR configuration file

Uboot version : 2014 :  04

 

 

Regards

Himanshu

Original Attachment has been moved to: mx6q_4x_mt41j128.cfg.zip

Labels (3)
Tags (4)
0 Kudos
1 Solution
143 Views
himanshugusain
Contributor III

Hello Igor

This issue is resolved ,the issue was with our watchdog configuration

it was not configured properly in the devicetree imx6q-sabresd-ldo.dts

Thanks for you support

Regards

Himanshu

View solution in original post

0 Kudos
15 Replies
143 Views
igorpadykov
NXP TechSupport
NXP TechSupport

Hi himanshu

this may be caused by hardware and one can recheck

recommendations in

HW Design Checking List for i.MX6DQSDL

one can try following experiments to determine if it is a hardware PCB issue

(power voltages ripples):

hand solder 22uF capacitors directly under center of chip.  Solder on top of the 0402 chips.

have software Increase voltages of ARM_CAP, SOC_CAP, PU_CAP

Best regards

igor

-----------------------------------------------------------------------------------------------------------------------

Note: If this post answers your question, please click the Correct Answer button. Thank you!

-----------------------------------------------------------------------------------------------------------------------

0 Kudos
143 Views
himanshugusain
Contributor III

Hello Igor

Below is the scenario

The issue is not observed in all the hardwares

in 40% of the hardware issue is occurring

Also in those 40% of hardware in which issue occurs  if we put previous version of software

Android JB, issue is not reproduced

After further analysis In Issue Hardwares we found

1.40% of hardware in which issue occurs issue observed when ever temperature reaches above 60-62 Celsius

2.In the rest 60% of hardwares even if temperature reaches above 60-62 Celsius device works properly it doesn't reboots

Regards

Himanshu

0 Kudos
143 Views
igorpadykov
NXP TechSupport
NXP TechSupport

Hi himanshu

I would suggest to narrow down issue using suggested in previous

answer method.

Best regards

igor

0 Kudos
143 Views
himanshugusain
Contributor III

Hello Igor

Please find the attached schematic snapshot below

The changes you have mentioned are already taken care

pastedImage_0.png

Still the issue persist

Regards

Himanshu

0 Kudos
143 Views
igorpadykov
NXP TechSupport
NXP TechSupport

Hi Himanshu

one can try to decrease working frequencies

cpu and memory.

~igor

0 Kudos
143 Views
himanshugusain
Contributor III

Hello Igor

Below are the set working frequencies

cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq/scaling_available_frequencies

396000 792000 996000

cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq

396000

the cpu frequency is set as minimum only

With memory are we referring to DDR frequency ?

Regards

Himanshu

0 Kudos
143 Views
igorpadykov
NXP TechSupport
NXP TechSupport

Hi Himanshu

yes DDR frequency. Run Temple apk

with lower frequencies. If it will not fail,

then this failures may be from poor power

filtering.

~igor

0 Kudos
143 Views
himanshugusain
Contributor III

Hi Igor

Sure,will check that

Just in addition to the observation in the issue boards

The board reboots not just on running Temple Run apk

Infact we tried heating the processor(using heat gun) without running any application

and observed the board reset (watchdog) at the temperature range (62-66)

Lower than this temperature range board doesn't reset even if we run the Temple run apk

and keep on cooling the processor using the cooling gun .

But in similar hardware setup when we tried the older version JB,the board didn't reset

at all(neither by running Temple run apk nor by the heating processor by external means)

So this eliminates the issue on the hardware side.

Just wanted to let you know so that we can focus on the right direction

Regards

Himanshu

0 Kudos
143 Views
igorpadykov
NXP TechSupport
NXP TechSupport

Hi Himanshu

you can also check if all latest patches are applied,

on http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=RDIMX6SABREBRD&fpsp=1&tab=Design_Tool...

Also try to reproduce this issue on Sabre reference board, this will point direction:

if this relates to custom hardware board or this is just software issue.

~igor

0 Kudos
143 Views
himanshugusain
Contributor III

Hi Igor

Please find the register dump of DDR register for Lollipop uboot

and Jellybean uboot ,I found few differences w.r.t few register

Jellybean Uboot

HIM:DDR 0x021b08b8  value read   =  690000

HIM:DDR 0x021b48b8  value read   =  6c0000

HIM:REG 0x020e0010 value read   =  0

HIM:REG 0x020e0018 value read   =  22222222

HIM:REG 0x020e001C value read   =  22222222

HIM:REG 0x020c4060 value read   =  a0001

HIM:REG 0x020c4068 value read   =  ffffffff

HIM:REG 0x020c406c value read   =  ffffffff

HIM:REG 0x020c4070 value read   =  ffffcfff

HIM:REG 0x020c4074 value read   =  ffff3fcf

HIM:REG 0x020c4078 value read   =  ffffffff

HIM:REG 0x020c407c value read   =  ffffffff

HIM:REG 0x020c4080 value read   =  ffffffff

Lollipop uboot

HIM:DDR 0x021b08b8  value read   =  640000

HIM:DDR 0x021b48b8  value read   =  670000

HIM:REG 0x020e0010 value read   =  f00000cf

HIM:REG 0x020e0018 value read   =  7f007f

HIM:REG 0x020e001C value read   =  7f007f

HIM:REG 0x020c4060 value read   =  fb

HIM:REG 0x020c4068 value read   =  c03f3f

HIM:REG 0x020c406c value read   =  30fc03

HIM:REG 0x020c4070 value read   =  fffc033

HIM:REG 0x020c4074 value read   =  3ff0f003

HIM:REG 0x020c4078 value read   =  fff300

HIM:REG 0x020c407c value read   =  f0000f3

HIM:REG 0x020c4080 value read   =  3ff

Note: we dont see device reboot issue with JB

Can these register differences can cause the issue in Lollipop

Also Iam not sure where the issue is ,Is it at kernel Level or at Uboot level

Regards

Himanshu

0 Kudos
143 Views
igorpadykov
NXP TechSupport
NXP TechSupport

Hi Himanshu

suggest to run ddr test and set ddr settings according to

its results

i.MX6/7 DDR Stress Test Tool V2.40

~igor

0 Kudos
143 Views
himanshugusain
Contributor III

Hi Igor

tried this ,the issue still persist

seems like the issue is at the kernel level

As at uboot level i tried running DDR test and externalyy heated the processor

was unable to see the board rebooting

You have any idea on what in the kernel level can affect rebooting of the tool

at certain temperature

Regards

Himanshu

0 Kudos
143 Views
igorpadykov
NXP TechSupport
NXP TechSupport

Hi Himanshu

if you think that this is kernel bug, please

try to reproduce this on Sabre NXP/FSL reference board.

If this issue will not appear on reference board, then

problem in custom board.

~igor

0 Kudos
144 Views
himanshugusain
Contributor III

Hello Igor

This issue is resolved ,the issue was with our watchdog configuration

it was not configured properly in the devicetree imx6q-sabresd-ldo.dts

Thanks for you support

Regards

Himanshu

View solution in original post

0 Kudos
143 Views
himanshugusain
Contributor III

I Igor

I booted my device with buildroot (uramdisk) (using same rest of LP images)

and i am not able to see the issue of device rebooting at that particular temp.

which means issue might be related to graphics

Do you have any clue on this

Regards

Himanshu

0 Kudos