[Openswan Users] BUG

[Giovani Moda - MR Informática]

I've builded a new kernel for FC4, based on 2.6.14-1.1644_FC4 with 
openswan-2.4.5dr3 KLIPS and NAT-T patches. Everything went well, but when I 
installed the openswan userland and started the service, I got:

Dec  9 11:49:12 inet kernel: BUG: rwlock recursion on CPU#0, 
klipsdebug/9717, c03d95a0 (Not tainted)
Dec  9 11:49:12 inet kernel:  [<c01d231c>] _raw_write_lock+0x56/0x58
Dec  9 11:49:12 inet kernel:  [<c03171ff>] pfkey_sock_list_grab+0xf/0xc4
Dec  9 11:49:12 inet kernel:  [<c014a7c7>] do_no_page+0x16e/0x301
Dec  9 11:49:12 inet kernel:  [<c031762e>] pfkey_remove_socket+0x14/0xaf
Dec  9 11:49:12 inet kernel:  [<c03176e6>] pfkey_destroy_socket+0x1d/0x40d
Dec  9 11:49:12 inet kernel:  [<c03180e3>] pfkey_release+0x49/0x107
Dec  9 11:49:12 inet kernel:  [<c02a465d>] sock_fasync+0x9a/0x149
Dec  9 11:49:12 inet kernel:  [<c02a3b7d>] sock_release+0x14/0x72
Dec  9 11:49:12 inet kernel:  [<c02a45a9>] sock_close+0x1e/0x38
Dec  9 11:49:12 inet kernel:  [<c0159745>] __fput+0xb1/0x178
Dec  9 11:49:12 inet kernel:  [<c0158052>] filp_close+0x3e/0x62
Dec  9 11:49:12 inet kernel:  [<c0102eb1>] syscall_call+0x7/0xb

But this has happen only once. I rebooted the system and restarted the 
service several times after that, but I could not reproduce the error. The 
service seems to be running. Any ideas?

I'll test it later to see if the tunnel works as expected and I'll return 
any results I get.

BTW, I came to realize only now that this kernel is based on 
kernel-2.6.14.3. Reading openswan's documentation, I noticed that it 
mentioned to work with kernel-2.6.14.2. Could that be the problem?

I also haven't checked if this kernel compiles SETKEY as a module or static. 
If static, that could be the problem, right? I'll check it out.

Giovani