Complete system crash when sending many CAN Frames

[pd0wm]

I'm running into very reproducible system crashes when using the SocketCAN driver. This seems to happen when sending a burst of CAN messages to different Arbiration IDs while running candump. I cannot trigger the crash when only sending on a single Arbitration ID, or if candump is not running.

Kernel version: Linux workstation-arch 6.8.2-arch2-1
intrepid-socketcan-kernel-module: 3.0.4 (4841b2863c1ad0a70e9d8d9334f82997195f6b2e)
icsscand: 3.1.2 (52db71754dfcceef1e80c893b9d78d163467248c).
Hardware Setup: ValueCAN 4-2 connected to another CAN adapter that will ensure frames are ACKed, but bus is otherwise empty.

Steps to reproduce:

1. Ensure the adapter is sending messages on different Arbitration IDs. e.g. while true; do for i in {100..999}; do cansend can0 $i#1122334455667788; done; done
2. Start candump. e.g. candump can0

Unfortunately I don't have any logs or crashdumps, as my whole system freezes instantly.

[GG-UP1]

I can confirm I'm experiencing the same issue.

I've encountered this system lockup on multiple neovi-Pi devices. The CAN interface becomes unusable when there's a high volume of differing CAN IDs on a single channel, impacting both sending and receiving operations.

Like the OP, the system freezes completely, making it impossible to capture any logs. This is a significant problem as SocketCAN is often the most straightforward way to interface with CAN on Linux.

Hoping for a resolution or guidance on how to address this.