Re: selinux networking: sleeping functin called from invalid context in 2.6.20-rc[12]

[Parag Warudkar <paragw@paragw.zapto.org>]

On Mon, 25 Dec 2006, Adam J. Richter wrote:

> 	Under 2.6.20-rc1 and 2.6.20-rc2, I get the following complaint
> for several network programs running on my system:
>
> [  156.381868] BUG: sleeping function called from invalid context at net/core/sock.c:1523
> [  156.381876] in_atomic():1, irqs_disabled():0
> [  156.381881] no locks held by kio_http/9693.
> [  156.381886]  [<c01057a2>] show_trace_log_lvl+0x1a/0x2f
> [  156.381900]  [<c0105dab>] show_trace+0x12/0x14
> [  156.381908]  [<c0105e48>] dump_stack+0x16/0x18
> [  156.381917]  [<c011e30f>] __might_sleep+0xe5/0xeb
> [  156.381926]  [<c025942a>] lock_sock_nested+0x1d/0xc4
> [  156.381937]  [<c01cc570>] selinux_netlbl_inode_permission+0x5a/0x8e
> [  156.381946]  [<c01c2505>] selinux_file_permission+0x96/0x9b
> [  156.381954]  [<c0175a0a>] vfs_write+0x8d/0x167
> [  156.381962]  [<c017605a>] sys_write+0x3f/0x63
> [  156.381971]  [<c01040c0>] syscall_call+0x7/0xb
> [  156.381980]  =======================

lock_sock_nested can sleep, its BH counterpart doesn't.
selinux_netlbl_inode_permission() probably needs to use the BH counterpart 
unconditionally. But I am not sure if that function is always called from an atomic context. Assuming it is, the 
attached patch should fix this.

Compile tested.

Signed-off-by: Parag Warudkar <paragw@paragw.zapto.org>

Parag

--- linux-2.6/security/selinux/ss/services.c.orig	2006-12-24 18:52:42.000000000 -0500
+++ linux-2.6/security/selinux/ss/services.c	2006-12-24 19:00:22.000000000 -0500
@@ -2660,9 +2660,9 @@
 		rcu_read_unlock();
 		return 0;
 	}
-	lock_sock(sock->sk);
+	bh_lock_sock_nested(sock->sk);
 	rc = selinux_netlbl_socket_setsid(sock, sksec->sid);
-	release_sock(sock->sk);
+	bh_unlock_sock(sock->sk);
 	rcu_read_unlock();
 
 	return rc;