i.MX8M Mini unstable issues on BSP 4.19.35_1.1.0/4.14.98_2.3.0

[kunyichen]

Hi all,
   we meeting a very serious problem on i.mx8mm,
   tested with BSP：4.14.98_2.3.0 & 4.19.35_1.1.0）。
   when pilot run over 250pcs customization board,
   but almost have 4% board meet the issues
   system maybe freeze sometime when run memtester for burn in
   or complete memtest cycle
   the issues like cpufreq scaling driver not stable?
   have anyone give me some advise for debug?

the board design to reference i.mx8mm EVK
run DDR stress tool passed over night
[ 6764.162531] rcu: rcu_sched kthread starved for 5252 jiffies! g100017 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402 ->cpu=2
[ 6764.172875] rcu: RCU grace-period kthread stack dump:
[ 6764.177924] rcu_sched       I    0    11      2 0x00000028
[ 6764.183410] Call trace:
[ 6764.185856]  __switch_to+0xdc/0x148
[ 6764.189345]  __schedule+0x1d8/0x5c8
[ 6764.192833]  schedule+0x38/0xa0
[ 6764.195975]  schedule_timeout+0x90/0x368
[ 6764.199900]  rcu_gp_kthread+0x494/0x860
[ 6764.203737]  kthread+0x128/0x130
[ 6764.206965]  ret_from_fork+0x10/0x18
[ 6827.030347] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
[ 6827.036286] rcu:     2-...!: (0 ticks this GP) idle=c50/0/0x0 softirq=109760/109760 fqs=1
[ 6827.044287] rcu:     (detected by 3, t=21007 jiffies, g=100017, q=7)
[ 6827.050383] Task dump for CPU 2:
[ 6827.053610] swapper/2       R  running task        0     0      1 0x00000028
[ 6827.060660] Call trace:
[ 6827.063114]  __switch_to+0xdc/0x148
[ 6827.066603]  0xffff800035158000
[ 6827.069747] rcu: rcu_sched kthread starved for 15755 jiffies! g100017 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402 ->cpu=2
[ 6827.080178] rcu: RCU grace-period kthread stack dump:
[ 6827.085227] rcu_sched       I    0    11      2 0x00000028
[ 6827.090713] Call trace:
[ 6827.093159]  __switch_to+0xdc/0x148
[ 6827.096650]  __schedule+0x1d8/0x5c8
[ 6827.100138]  schedule+0x38/0xa0
[ 6827.103281]  schedule_timeout+0x90/0x368
[ 6827.107206]  rcu_gp_kthread+0x494/0x860
[ 6827.111042]  kthread+0x128/0x130
[ 6827.114271]  ret_from_fork+0x10/0x18
[ 7137.266393] cpu cpu0: failed to scale arm_reg down: -110
[ 7137.271773] cpufreq: __target_index: Failed to change cpu frequency: -110
[ 7137.694380] cpu cpu0: failed to scale arm_reg down: -110
[ 7137.699751] cpufreq: __target_index: Failed to change cpu frequency: -110
[ 7138.122540] cpu cpu0: failed to scale arm_reg down: -110
[ 7138.128097] cpufreq: __target_index: Failed to change cpu frequency: -110
[ 7138.550539] cpu cpu0: failed to scale arm_reg down: -110
[ 7138.556091] cpufreq: __target_index: Failed to change cpu frequency: -110

[igorpadykov]

Hi kunyichen

one can try attached patch, however it is highly recommended to use

latest bsps, described on linux documentation link

https://www.nxp.com/design/software/embedded-software/i-mx-software/embedded-linux-for-i-mx-applicat...

additionally one can try to change signals drive strength (search in RPA tool for "stregth").

Best regards
igor

[kunyichen]

Hi @igorpadykov,

       Thank you for your quick reply,

       okay, we will try apply this patch on u-boot

       or migration our u-boot to latest version