On Thu, Sep 30, 2021 at 12:21:55AM +0300, Nikolay Aleksandrov wrote: > To answer the question about the patch - it is correct and ok, the dereference needed > is done inside the function and after that the value is copied. I'd only check how the > DSA bridge pointer is handled, it seems it is dereferenced directly and I don't know > the DSA locking rules (i.e. if anything is held by that time to ensure the pointer is > valid). As for the patch: > Reviewed-by: Nikolay Aleksandrov <nikolay@xxxxxxxxxx> I think you're right in the fact that nobody ensures dp->bridge_dev cannot be reset to NULL by dsa_port_bridge_leave(). I'll try to simulate an issue by spin-looping for a while in the xmit function while I add and remove the port from a bridge and see what happens. Worst case, we'll have to add READ_ONCE and WRITE_ONCE annotations for all users, and work with a local, on-stack reference. Because the bridge net_device itself will not go away until the xmit itself finishes, but it's just that it may not be our upper anymore.
