Boot hangs on serial8250: too much work for irq36

[EliteHawk]

I am currently working on Poky on an embedded PPC system.

I am experiencing a "serial8250: too much work for irq36" problem around 50-75% times I try to boot.

What I was thinking to do is

1. Reduce printk lines
2. Check if there is any fix about this error on Linux source files update

The error I am observing is this one:

serial8250_interrupt: 8544 callbacks suppressed
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
(several lines same of the ones before)
INFO: rcu_sched self-detected stall on CPU
3: (1 GPs behind) idle=f9f/140000000000001/0 softirq=0/0 fqs=1699
INFO: rcu_sched detected stalls on CPUs/tasks:
0: (1 GPs behind) idle=965/140000000000001/0 softirq=0/0 fqs=1699
3: (1 GPs behind) idle=f9f/140000000000001/0 softirq=0/0 fqs=1699
(detected by 1, t=52517 jiffies, g=91, c=90, q=161)
Task dump for CPU 0:
(agetty) R running 0 1915 1 0x00000084
Call Trace:
[ed0079a0] [c005b408] update_cfs_rq_blocked_load+0xb8/0x1c0 (unreliable)
[ed007a60] [00000211] 0x211
Task dump for CPU 3:
systemd-journal R running 0 957 1 0x00000004
Call Trace:
rcu_sched kthread starved for 50818 jiffies!
(t=52529 jiffies g=91 c=90 q=161)
rcu_sched kthread starved for 50831 jiffies!
Task dump for CPU 0:
(agetty) R running 0 1915 1 0x00000084
Call Trace:
[ed0079a0] [c005b408] update_cfs_rq_blocked_load+0xb8/0x1c0 (unreliable)
[ed007a60] [00000211] 0x211
Task dump for CPU 3:
systemd-journal R running 0 957 1 0x00000004
Call Trace:
[e9ee9b80] [c007f948] rcu_dump_cpu_stacks+0xa8/0x100 (unreliable)
[e9ee9ba0] [c008386c] rcu_check_callbacks+0x4fc/0x7b0
[e9ee9c10] [c008725c] update_process_times+0x3c/0x70
[e9ee9c20] [c009afb8] tick_sched_timer+0x68/0xe0
[e9ee9c50] [c00880b4] __run_hrtimer.isra.34+0x54/0xf0
[e9ee9c70] [c00889b8] hrtimer_interrupt+0x118/0x330
[e9ee9ce0] [c0009658] __timer_interrupt+0xa8/0x1a0
[e9ee9d00] [c0009980] timer_interrupt+0xb0/0xe0
[e9ee9d20] [c000f7e0] ret_from_except+0x0/0x18
--- interrupt: 901 at smp_call_function_many+0x2a0/0x300
LR = smp_call_function_many+0x270/0x300
[e9ee9de0] [c00a033c] smp_call_function_many+0x24c/0x300 (unreliable)
[e9ee9e20] [c00177cc] flush_tlb_mm+0x9c/0xa0
[e9ee9e40] [c00f8dd8] tlb_flush_mmu_tlbonly.part.86+0x18/0x90
[e9ee9e50] [c00f8f54] tlb_flush_mmu+0x24/0x40
[e9ee9e60] [c00f8f88] tlb_finish_mmu+0x18/0x70
[e9ee9e70] [c00fee00] unmap_region+0xc0/0x140
[e9ee9ef0] [c01013ec] do_munmap+0x26c/0x430
[e9ee9f20] [c01015e8] vm_munmap+0x38/0x60
[e9ee9f40] [c000f130] ret_from_syscall+0x0/0x3c
--- interrupt: c01 at 0x201842fc
LR = 0x20117090
serial8250_interrupt: 8544 callbacks suppressed
serial8250: too much work for irq36
serial8250: too much work for irq36
serial8250: too much work for irq36
(several lines same of the ones before)

[Pavel]

Is there problem for using NXP SDK 2.0 on your board?

Is there problem if default configuration of SDK 2.0 is used on your board?

[jobs]

I had a similar problem,MPC85XX T2080 E6500 RCU检测CPU X卡顿 - 恩智浦社区 (nxp.com)