hi ：
当温度升高到51摄氏度时eth0 网速降低为0,当超过52或53摄氏度时网速恢复正常，可以达到接近千兆。
当温度降低到51摄氏度时网速在次降低到0,当降低到50摄氏度以下时，网速又恢复正常。
下面是抓到的一段日志，请帮忙分析一下。这个日子是代表什么问题。
Jul 5 21:49:51 compute kernel: [ 2843.588040] INFO: rcu_preempt self-detected stall on CPU
Jul 5 21:49:51 compute kernel: [ 2843.588175] 0-…: (2941 ticks this GP) idle=e51/140000000000002/0 softirq=317018/317018 fqs=2133
Jul 5 21:49:51 compute kernel: [ 2843.588355] (t=5250 jiffies g=190585 c=190584 q=46262)
Jul 5 21:49:51 compute kernel: [ 2843.588455] Task dump for CPU 0:
Jul 5 21:49:51 compute kernel: [ 2843.588459] ksoftirqd/0 R running task 0 3 2 0x00000002
Jul 5 21:49:51 compute kernel: [ 2843.588465] Call trace:
Jul 5 21:49:51 compute kernel: [ 2843.588477] [] dump_backtrace+0x0/0x198
Jul 5 21:49:51 compute kernel: [ 2843.588482] [] show_stack+0x24/0x30
Jul 5 21:49:51 compute kernel: [ 2843.588491] [] sched_show_task+0xf8/0x148
Jul 5 21:49:51 compute kernel: [ 2843.588495] [] dump_cpu_task+0x48/0x58
Jul 5 21:49:51 compute kernel: [ 2843.588500] [] rcu_dump_cpu_stacks+0xb8/0xec
Jul 5 21:49:51 compute kernel: [ 2843.588505] [] rcu_check_callbacks+0x728/0xa48
Jul 5 21:49:51 compute kernel: [ 2843.588509] [] update_process_times+0x34/0x60
Jul 5 21:49:51 compute kernel: [ 2843.588514] [] tick_sched_handle.isra.5+0x38/0x70
Jul 5 21:49:51 compute kernel: [ 2843.588517] [] tick_sched_timer+0x4c/0x90
Jul 5 21:49:51 compute kernel: [ 2843.588520] [] __hrtimer_run_queues+0xd8/0x360
Jul 5 21:49:51 compute kernel: [ 2843.588523] [] hrtimer_interrupt+0xa8/0x1e0
Jul 5 21:49:51 compute kernel: [ 2843.588529] [] arch_timer_handler_phys+0x38/0x58
Jul 5 21:49:51 compute kernel: [ 2843.588534] [] handle_percpu_devid_irq+0x90/0x2b0
Jul 5 21:49:51 compute kernel: [ 2843.588537] [] generic_handle_irq+0x34/0x50
Jul 5 21:49:51 compute kernel: [ 2843.588540] [] __handle_domain_irq+0x68/0xc0
Jul 5 21:49:51 compute kernel: [ 2843.588543] [] gic_handle_irq+0x5c/0xb0
Jul 5 21:49:51 compute kernel: [ 2843.588545] [] el1_irq+0xe8/0x18c
Jul 5 21:49:51 compute kernel: [ 2843.588551] [] iova_rcache_insert+0xc4/0x290
Jul 5 21:49:51 compute kernel: [ 2843.588554] [] free_iova_fast+0x34/0x58
Jul 5 21:49:51 compute kernel: [ 2843.588557] [] __iommu_dma_free_iova+0x78/0x90
Jul 5 21:49:51 compute kernel: [ 2843.588560] [] __iommu_dma_unmap+0x74/0xa8
Jul 5 21:49:51 compute kernel: [ 2843.588563] [] iommu_dma_unmap_page+0x38/0x48
Jul 5 21:49:51 compute kernel: [ 2843.588568] [] __iommu_unmap_page+0x60/0x160
Jul 5 21:49:51 compute kernel: [ 2843.588573] [] tx_swcx_free+0x54/0xe0
Jul 5 21:49:51 compute kernel: [ 2843.588576] [] eqos_napi_poll_tx+0x2b0/0x4f8
Jul 5 21:49:51 compute kernel: [ 2843.588581] [] net_rx_action+0xf4/0x358
Jul 5 21:49:51 compute kernel: [ 2843.588584] [] __do_softirq+0x13c/0x3b0
Jul 5 21:49:51 compute kernel: [ 2843.588589] [] irq_exit+0xd0/0x118
Jul 5 21:49:51 compute kernel: [ 2843.588592] [] __handle_domain_irq+0x6c/0xc0
Jul 5 21:49:51 compute kernel: [ 2843.588595] [] gic_handle_irq+0x5c/0xb0
Jul 5 21:49:51 compute kernel: [ 2843.588597] [] el1_irq+0xe8/0x18c
Jul 5 21:49:51 compute kernel: [ 2843.588609] [] smpboot_thread_fn+0xec/0x248
Jul 5 21:49:51 compute kernel: [ 2843.588613] [] kthread+0xec/0xf0
Jul 5 21:49:51 compute kernel: [ 2843.588616] [] ret_from_fork+0x10/0x40
Jul 5 21:49:51 compute kernel: [ 2843.600027] INFO: rcu_sched detected stalls on CPUs/tasks:
Jul 5 21:49:51 compute kernel: [ 2843.600156] 0-…: (2944 ticks this GP) idle=e51/140000000000002/0 softirq=317018/317018 fqs=1943
Jul 5 21:49:51 compute kernel: [ 2843.600299] (detected by 1, t=5253 jiffies, g=114818, c=114817, q=4289)
Jul 5 21:49:51 compute kernel: [ 2843.600435] Task dump for CPU 0:
Jul 5 21:49:51 compute kernel: [ 2843.600438] ksoftirqd/0 R running task 0 3 2 0x00000002
Please at least share detail about your board info.
https://forums.developer.nvidia.com/t/debug-tips-when-you-hit-any-kind-of-system-issue/177457/2
Is it possible for you to share your comment in English? So that other non-Chinese users can read your comment.
Please try with the latest JetPack to see if improved.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.