[openib-general] [rdma_ucm] enabling the rdma_ucm and restarting the driver several times causes kernel oops

[dotanb at dev.mellanox.co.il]

> Dotan Barak wrote:
>> here is the backtrace from the /var/log/messages:
>> Dec 27 15:36:25 sw086 kernel: Unable to handle kernel NULL pointer
>> dereference at 0000000000000001 RIP:
>> Dec 27 15:36:25 sw086 kernel:  [<0000000000000001>]
>> Dec 27 15:36:25 sw086 kernel: PGD 11f4c3067 PUD 11fed7067 PMD 0
>> Dec 27 15:36:25 sw086 kernel: Oops: 0000 [1] SMP
>> Dec 27 15:36:25 sw086 kernel: CPU 1
>> Dec 27 15:36:25 sw086 kernel: Modules linked in: rdma_ucm ib_sdp rdma_cm
>> iw_cm ib_addr ib_ipoib ib_mthca ib_umad ib_ucm ib_u
>> verbs ib_cm ib_sa ib_mad ib_core nfsd exportfs ipv6 parport_pc lp
>> parport autofs4 nfs lockd nfs_acl sunrpc dm_mirror dm_mod
>> button battery asus_acpi ac uhci_hcd ehci_hcd i2c_i801 i2c_core tg3 sg
>> ext3 jbd sd_mod
>
> Can you narrow down which module unload is causing the issue?  Is anything
> using
> the rdma_ucm or ib_uverbs?  Is ib_sdp the first module unloaded?
>
> - Sean
>
vlad: can you please tell us what the order of the modules
loading/unloading is?

As i wrote in the problem description, I only enabled the rdma_ucm
module and restarted the driver (without even using this module or any
other module in the driver) and i got this kernel oops.

Dotan