Kernel oops during access broadcom switch with PCIe

[zwg132883]

Hi,

A broadcom switch device is connected to T1042 with PCIe on our board. switch is working on EP mode. linux kernel version is 4.14. Kernel oops is raised in field card with pcie error detected. but this issue cannot be reproduced on our lab card.

Please see kernel oops, interrupts and dmesg info in attachment.

What's the root cause of kernel oops? How can we debug the kernel oops? thanks.

[Bio_TICFSL]

Hello,

An oops means that something running on the system has violated the kernel’s rules about proper behavior. Maybe the code tried to take a code path that was not allowed or use an invalid pointer. Whatever it was, the kernel — always on the lookout for process misbehavior — most likely will have stopped the particular process in its tracks and written some messages about what it did to the console, to /var/log/dmesg or the /var/log/kern.log file.

You need a realview DS5 to debug the kernel.

Regards

[zwg132883]

Hello,

I have no realview DS5. Could you please explain why "mpc85xx_pci_err mpc85xx-pci-edac.1.auto" happened? And why so many "PCI err" interrrupt raised on my board? thanks.

[Bio_TICFSL]

Hello,

The problem is that a pointer to mpc85xx... is pointed out, you need to see where the fails is, or you need to program your board correctly.

Regards