请注意，本文内容源自机器翻译，可能存在语法或其它翻译错误，仅供参考。如需获取准确内容，请参阅链接中的英语原文或自行翻译。
器件型号： J784S4XEVM
尊敬的专家：
我将 J784S4 EVM 与 SDK 10 配合使用。
我正在进行 2 个 EVM 之间的 PCIe 连接。
在 RC 侧、我能够启用 BAR 1。
但当我启用 bar 0 时、Linux 将崩溃
root@j784s4-evm:~# devmem2 0xd100300 w 0x106
/dev/mem opened.
Memory mapped at address 0xffff94b27000.
Read at address 0x0D100300 (0xffff94b27300): 0x00012400
Write at address 0x0D100300 (0xffff94b27300): 0x00000106, readback 0x00000106
root@j784s4-evm:~# echo 1 > /sys/bus/pci/devices/0000\:00\:00.0/remove
-sh: /sys/bus/pci/devices/0000:00:00.0/remove: No such file or directory
root@j784s4-evm:~# echo 1 > /sys/bus/pci/rescan
[ 170.773002] pci 0000:00:00.0: [104c:b012] type 01 class 0x060400
[ 170.779042] pci 0000:00:00.0: reg 0x10: [mem 0x00000000-0x000000ff]
[ 170.785411] pci 0000:00:00.0: supports D1
[ 170.789417] pci 0000:00:00.0: PME# supported from D0 D1 D3hot
[ 170.797566] pci 0000:00:00.0: BAR 0: assigned [mem 0x10011000-0x100110ff]
[ 170.804364] pci 0000:00:00.0: PCI bridge to [bus 01]
[ 170.809446] pcieport 0000:00:00.0: of_irq_parse_pci: failed with rc=-22
[ 191.804815] rcu: INFO: rcu_preempt detected stalls on CPUs/tasks:
[ 191.810899] rcu: 4-...!: (5 ticks this GP) idle=d46c/1/0x4000000000000000 softirq=2662/2662 fqs=3
[ 191.819840] rcu: (detected by 0, t=5254 jiffies, g=4301, q=670 ncpus=8)
[ 191.826523] Task dump for CPU 4:
[ 191.829737] task:sh state:R running task stack:0 pid:1124 ppid:678 flags:0x00000205
[ 191.839632] Call trace:
[ 191.842067] __switch_to+0xcc/0x12c
[ 191.845557] msix_setup_msi_descs+0x40/0x134
[ 191.849817] __pci_enable_msix_range+0x368/0x610
[ 191.854420] pci_alloc_irq_vectors_affinity+0xc4/0x144
[ 191.859546] pci_alloc_irq_vectors+0x14/0x20
[ 191.863803] pcie_portdrv_probe+0x1d8/0x5f0
[ 191.867973] pci_device_probe+0xa4/0x168
[ 191.871882] really_probe+0x148/0x2b0
[ 191.875533] __driver_probe_device+0x78/0x12c
[ 191.879876] driver_probe_device+0xd8/0x15c
[ 191.884046] __device_attach_driver+0xb8/0x134
[ 191.888476] bus_for_each_drv+0x88/0xe8
[ 191.892298] __device_attach+0xa0/0x190
[ 191.896120] device_attach+0x14/0x20
[ 191.899684] pci_bus_add_device+0x64/0xd4
[ 191.903682] pci_bus_add_devices+0x3c/0x88
[ 191.907766] pci_rescan_bus+0x30/0x44
[ 191.911414] rescan_store+0x84/0xa8
[ 191.914889] bus_attr_store+0x28/0x3c
[ 191.918538] sysfs_kf_write+0x44/0x54
[ 191.922189] kernfs_fop_write_iter+0x120/0x1cc
[ 191.926618] vfs_write+0x1c0/0x308
[ 191.930010] ksys_write+0x74/0x10c
[ 191.933400] __arm64_sys_write+0x1c/0x28
[ 191.937310] invoke_syscall+0x48/0x114
[ 191.941048] el0_svc_common.constprop.0+0xc0/0xe0
[ 191.945738] do_el0_svc+0x1c/0x28
[ 191.949040] el0_svc+0x2c/0x84
[ 191.952083] el0t_64_sync_handler+0x120/0x12c
[ 191.956426] el0t_64_sync+0x190/0x194
[ 191.960076] rcu: rcu_preempt kthread timer wakeup didn't happen for 5282 jiffies! g4301 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402
[ 191.971352] rcu: Possible timer handling issue on cpu=5 timer-softirq=356
[ 191.978206] rcu: rcu_preempt kthread starved for 5288 jiffies! g4301 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402 ->cpu=5
[ 191.988529] rcu: Unless rcu_preempt kthread gets sufficient CPU time, OOM is now expected behavior.
[ 191.997637] rcu: RCU grace-period kthread stack dump:
[ 192.002670] task:rcu_preempt state:I stack:0 pid:16 ppid:2 flags:0x00000008
[ 192.011002] Call trace:
[ 192.013437] __switch_to+0xcc/0x12c
[ 192.016914] __schedule+0x264/0x9c4
[ 192.020389] schedule+0x5c/0xc4
[ 192.023517] schedule_timeout+0x90/0x108
[ 192.027427] rcu_gp_fqs_loop+0x114/0x4c8
[ 192.031339] rcu_gp_kthread+0x134/0x160
[ 192.035163] kthread+0x110/0x114
[ 192.038384] ret_from_fork+0x10/0x20
[ 192.041947] rcu: Stack dump where RCU GP kthread last ran:
[ 192.047413] Task dump for CPU 5:
[ 192.050627] task:swapper/5 state:R running task stack:0 pid:0 ppid:1 flags:0x00000008
[ 192.060519] Call trace:
[ 192.062952] __switch_to+0xcc/0x12c
[ 192.066429] 0x0
感谢你的帮助
Charles
