This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM5728: Linux reboot

Part Number: AM5728

Hi,

linux sdk5.0.2

The customer executed a program for camera capture + image recognition algorithm, CPU utilization is close to 80%. Process A cpu accounted for about 40%, and the algorithm process B cpu accounted for about 40%. The board reboot after print the below log.

[17:41:12] [ 57.232432] INFO: rcu_preempt detected stalls on CPUs/tasks:
[17:41:12] [ 57.238373] 1-...: (0 ticks this GP) idle=14a/140000000000000/0 softirq=15407/15407 fqs=0 
[17:41:12] [ 57.247117] (detected by 0, t=2102 jiffies, g=1197, c=1196, q=41)
[17:41:12] [ 57.253594] Sending NMI from CPU 0 to CPUs 1:
[17:41:22] [ 67.257281] rcu_preempt kthread starved for 2102 jiffies! g1197 c1196 f0x0 RCU_GP_WAIT_FQS(3) ->state=0x402 ->cpu=1
[17:41:22] [ 67.268211] rcu_preempt I 0 8 2 0x00000000
[17:41:22] [ 67.273961] Backtrace: 
[17:41:22] [ 67.276531] [<c0976768>] (__schedule) from [<c0976f60>] (schedule+0x58/0xc4)
[17:41:22] [ 67.283914] r10:c10181de r9:c10181d0 r8:c1002d00 r7:ee89bed8 r6:eed43580 r5:eed43580
[17:41:22] [ 67.292115] r4:ffffe000
[17:41:22] [ 67.294767] [<c0976f08>] (schedule) from [<c097a7b4>] (schedule_timeout+0x158/0x258)
[17:41:22] [ 67.302877] r5:eed43580 r4:ffff98f7
[17:41:22] [ 67.306625] [<c097a65c>] (schedule_timeout) from [<c028f988>] (rcu_gp_kthread+0x5a0/0x980)
[17:41:22] [ 67.315283] r8:00000001 r7:c10181dc r6:c1002d00 r5:00000001 r4:c1018080
[17:41:22] [ 67.322304] [<c028f3e8>] (rcu_gp_kthread) from [<c0248e28>] (kthread+0x164/0x16c)
[17:41:22] [ 67.330137] r7:ee89a000
[17:41:22] [ 67.332791] [<c0248cc4>] (kthread) from [<c0207d18>] (ret_from_fork+0x14/0x3c)
[17:41:22] [ 67.340358] r10:00000000 r9:00000000 r8:00000000 r7:00000000 r6:00000000 r5:c0248cc4
[17:41:22] [ 67.348556] r4:ee84ae40
[17:41:22] [ 67.351207] INFO: rcu_sched detected stalls on CPUs/tasks:
[17:41:22] [ 67.356952] 1-...: (1 GPs behind) idle=14a/140000000000000/0 softirq=15406/15407 fqs=34 
[17:41:22] [ 67.365515] (detected by 0, t=2102 jiffies, g=911, c=910, q=7)
[17:41:22] [ 67.371720] Sending NMI from CPU 0 to CPUs 1:
[17:41:32] [ 77.375403] rcu_sched kthread starved for 2034 jiffies! g911 c910 f0x0 RCU_GP_WAIT_FQS(3) ->state=0x0 ->cpu=0
[17:41:32] [ 77.385787] rcu_sched R running task 0 9 2 0x00000000
[17:41:32] [ 77.393172] Backtrace: 
[17:41:32] [ 77.395736] [<c0976768>] (__schedule) from [<c0976f60>] (schedule+0x58/0xc4)
[17:41:32] [ 77.403120] r10:c101841e r9:c1018410 r8:c1002d00 r7:ee89ded8 r6:eed34580 r5:eed34580
[17:41:32] [ 77.411319] r4:ffffe000
[17:41:32] [ 77.413970] [<c0976f08>] (schedule) from [<c097a7b4>] (schedule_timeout+0x158/0x258)
[17:41:32] [ 77.422078] r5:eed34580 r4:ffff993b
[17:41:32] [ 77.425822] [<c097a65c>] (schedule_timeout) from [<c028f988>] (rcu_gp_kthread+0x5a0/0x980)
[17:41:32] [ 77.434482] r8:00000000 r7:c101841c r6:c1002d00 r5:00000001 r4:c10182c0
[17:41:32] [ 77.441502] [<c028f3e8>] (rcu_gp_kthread) from [<c0248e28>] (kthread+0x164/0x16c)
[17:41:32] [ 77.449340] r7:ee89c000
[17:41:32] [ 77.451991] [<c0248cc4>] (kthread) from [<c0207d18>] (ret_from_fork+0x14/0x3c)
[17:41:32] [ 77.459555] r10:00000000 r9:00000000 r8:00000000 r7:00000000 r6:00000000 r5:c0248cc4
[17:41:32] [ 77.467757] r4:ee84ae40

  • Hi Nancy,

    Do you use AM572x TI board (EVM, IDK, X15) or custom board? If custom board, I can suggest to try this application on TI board.

    Also it seems to me that your are using RT Linux, is that correct? Can you also try with non-RT Linux? What is the result than?

    RT-Linux SDK 5.02:

    non-RT Linux SDK 5.02:

    Regards,
    Pavel

  • Hi,

    His core board is TI's EVM board, and he used non-RT linux.

    In addition, the programme crashed only when used TBB mechanism.

  • Nancy,

    According to the kernel log, an application loads the CPU for long time and causing stall then the kernel CPU stall detector activates.
    I suggest you to read instruction in ../<Linux_dir>/Documentation/RCU/stallwarn.txt which can help to make optimal kernel configuration from minimizing CPU stall time.

    Check also below e2e threads:

    Regards,
    Pavel