"Cannot allocate memory" or kernel backtrace in "qdisc_create" when running "tc"

Issue

• I have written a script which runs several thousand invocations of the linux tc application. When I run the script, I get a kernel oops. It is quite reproducible.
• Running tc traffic control returns RTNETLINK answers: Cannot allocate memory
• Kernel BUG or OOPS with backtrace similar to the following when running tc:

kernel: Call Trace:
kernel: [<ffffffff8143d6ac>] qdisc_create+0x1ac/0x370
kernel: [<ffffffff811f489a>] ? avc_has_perm_flags+0x7a/0xa0
kernel: [<ffffffff81258f64>] ? nla_parse+0x34/0x110
kernel: [<ffffffff8143ee5d>] tc_modify_qdisc+0x18d/0x490
kernel: [<ffffffff8142c617>] rtnetlink_rcv_msg+0x2b7/0x320
kernel: [<ffffffff8142c360>] ? rtnetlink_rcv+0x40/0x40
kernel: [<ffffffff814462e9>] netlink_rcv_skb+0xa9/0xd0
kernel: [<ffffffff8142c345>] rtnetlink_rcv+0x25/0x40
kernel: [<ffffffff81445f8e>] netlink_unicast+0x2de/0x2f0
kernel: [<ffffffff81446b15>] netlink_sendmsg+0x225/0x340
kernel: [<ffffffff81407eaf>] sock_sendmsg+0xef/0x120
kernel: [<ffffffff81107a72>] ? filemap_fault+0xd2/0x4b0
kernel: [<ffffffff8110685a>] ? unlock_page+0x2a/0x40
kernel: [<ffffffff8112e4e1>] ? __do_fault+0x481/0x5a0
kernel: [<ffffffff81407095>] ? move_addr_to_kernel+0x65/0x70
kernel: [<ffffffff81414415>] ? verify_iovec+0x85/0xf0
kernel: [<ffffffff81409baa>] __sys_sendmsg+0x41a/0x440
kernel: [<ffffffff8112e9e5>] ? handle_mm_fault+0x1d5/0x350
kernel: [<ffffffff814ea020>] ? do_page_fault+0x1e0/0x460
kernel: [<ffffffff81409dd9>] sys_sendmsg+0x49/0x90
kernel: [<ffffffff814ee802>] system_call_fastpath+0x16/0x1b