linux-stable.git/net/core/rtnetlink.c, branch linux-2.6.33.y

net: Fix ip link add netns oops

2011-03-21T19:44:55+00:00

commit 13ad17745c2cbd437d9e24b2d97393e0be11c439 upstream.

Ed Swierk  writes:
> On 2.6.35.7
>  ip link add link eth0 netns 9999 type macvlan
> where 9999 is a nonexistent PID triggers an oops and causes all network functions to hang:
> [10663.821898] BUG: unable to handle kernel NULL pointer dereference at 000000000000006d
>  [10663.821917] IP: [] __dev_alloc_name+0x9a/0x170
>  [10663.821933] PGD 1d3927067 PUD 22f5c5067 PMD 0
>  [10663.821944] Oops: 0000 [#1] SMP
>  [10663.821953] last sysfs file: /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
>  [10663.821959] CPU 3
>  [10663.821963] Modules linked in: macvlan ip6table_filter ip6_tables rfcomm ipt_MASQUERADE binfmt_misc iptable_nat nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack sco ipt_REJECT bnep l2cap xt_tcpudp iptable_filter ip_tables x_tables bridge stp vboxnetadp vboxnetflt vboxdrv kvm_intel kvm parport_pc ppdev snd_hda_codec_intelhdmi snd_hda_codec_conexant arc4 iwlagn iwlcore mac80211 snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_seq_midi snd_rawmidi i915 snd_seq_midi_event snd_seq thinkpad_acpi drm_kms_helper btusb tpm_tis nvram uvcvideo snd_timer snd_seq_device bluetooth videodev v4l1_compat v4l2_compat_ioctl32 tpm drm tpm_bios snd cfg80211 psmouse serio_raw intel_ips soundcore snd_page_alloc intel_agp i2c_algo_bit video output netconsole configfs lp parport usbhid hid e1000e sdhci_pci ahci libahci sdhci led_class
>  [10663.822155]
>  [10663.822161] Pid: 6000, comm: ip Not tainted 2.6.35-23-generic #41-Ubuntu 2901CTO/2901CTO
>  [10663.822167] RIP: 0010:[] [] __dev_alloc_name+0x9a/0x170
>  [10663.822177] RSP: 0018:ffff88014aebf7b8 EFLAGS: 00010286
>  [10663.822182] RAX: 00000000fffffff4 RBX: ffff8801ad900800 RCX: 0000000000000000
>  [10663.822187] RDX: ffff880000000000 RSI: 0000000000000000 RDI: ffff88014ad63000
>  [10663.822191] RBP: ffff88014aebf808 R08: 0000000000000041 R09: 0000000000000041
>  [10663.822196] R10: 0000000000000000 R11: dead000000200200 R12: ffff88014aebf818
>  [10663.822201] R13: fffffffffffffffd R14: ffff88014aebf918 R15: ffff88014ad62000
>  [10663.822207] FS: 00007f00c487f700(0000) GS:ffff880001f80000(0000) knlGS:0000000000000000
>  [10663.822212] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>  [10663.822216] CR2: 000000000000006d CR3: 0000000231f19000 CR4: 00000000000026e0
>  [10663.822221] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>  [10663.822226] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
>  [10663.822231] Process ip (pid: 6000, threadinfo ffff88014aebe000, task ffff88014afb16e0)
>  [10663.822236] Stack:
>  [10663.822240] ffff88014aebf808 ffffffff814a2bb5 ffff88014aebf7e8 00000000a00ee8d6
>  [10663.822251] <0> 0000000000000000 ffffffffa00ef940 ffff8801ad900800 ffff88014aebf818
>  [10663.822265] <0> ffff88014aebf918 ffff8801ad900800 ffff88014aebf858 ffffffff8149c413
>  [10663.822281] Call Trace:
>  [10663.822290] [] ? dev_addr_init+0x75/0xb0
>  [10663.822298] [] dev_alloc_name+0x43/0x90
>  [10663.822307] [] rtnl_create_link+0xbe/0x1b0
>  [10663.822314] [] rtnl_newlink+0x48a/0x570
>  [10663.822321] [] ? rtnl_newlink+0x1ac/0x570
>  [10663.822332] [] ? native_x2apic_icr_read+0x4/0x20
>  [10663.822339] [] rtnetlink_rcv_msg+0x177/0x290
>  [10663.822346] [] ? rtnetlink_rcv_msg+0x0/0x290
>  [10663.822354] [] netlink_rcv_skb+0xa9/0xd0
>  [10663.822360] [] rtnetlink_rcv+0x25/0x40
>  [10663.822367] [] netlink_unicast+0x2de/0x2f0
>  [10663.822374] [] netlink_sendmsg+0x1fe/0x2e0
>  [10663.822383] [] sock_sendmsg+0xf3/0x120
>  [10663.822391] [] ? _raw_spin_lock+0xe/0x20
>  [10663.822400] [] ? __d_lookup+0x136/0x150
>  [10663.822406] [] ? _raw_spin_lock+0xe/0x20
>  [10663.822414] [] ? _atomic_dec_and_lock+0x4d/0x80
>  [10663.822422] [] ? mntput_no_expire+0x30/0x110
>  [10663.822429] [] ? move_addr_to_kernel+0x65/0x70
>  [10663.822435] [] ? verify_iovec+0x88/0xe0
>  [10663.822442] [] sys_sendmsg+0x240/0x3a0
> [10663.822450] [] ? __do_fault+0x479/0x560
>  [10663.822457] [] ? _raw_spin_lock+0xe/0x20
>  [10663.822465] [] ? alloc_fd+0x10a/0x150
>  [10663.822473] [] ? do_page_fault+0x15e/0x350
>  [10663.822482] [] system_call_fastpath+0x16/0x1b
>  [10663.822487] Code: 90 48 8d 78 02 be 25 00 00 00 e8 92 1d e2 ff 48 85 c0 75 cf bf 20 00 00 00 e8 c3 b1 c6 ff 49 89 c7 b8 f4 ff ff ff 4d 85 ff 74 bd <4d> 8b 75 70 49 8d 45 70 48 89 45 b8 49 83 ee 58 eb 28 48 8d 55
>  [10663.822618] RIP [] __dev_alloc_name+0x9a/0x170
>  [10663.822627] RSP 
>  [10663.822631] CR2: 000000000000006d
>  [10663.822636] ---[ end trace 3dfd6c3ad5327ca7 ]---

This bug was introduced in:
commit 81adee47dfb608df3ad0b91d230fb3cef75f0060
Author: Eric W. Biederman 
Date:   Sun Nov 8 00:53:51 2009 -0800

    net: Support specifying the network namespace upon device creation.

    There is no good reason to not support userspace specifying the
    network namespace during device creation, and it makes it easier
    to create a network device and pass it to a child network namespace
    with a well known name.

    We have to be careful to ensure that the target network namespace
    for the new device exists through the life of the call.  To keep
    that logic clear I have factored out the network namespace grabbing
    logic into rtnl_link_get_net.

    In addtion we need to continue to pass the source network namespace
    to the rtnl_link_ops.newlink method so that we can find the base
    device source network namespace.

    Signed-off-by: Eric W. Biederman 
    Acked-by: Eric Dumazet 

Where apparently I forgot to add error handling to the path where we create
a new network device in a new network namespace, and pass in an invalid pid.

Reported-by: Ed Swierk 
Signed-off-by: "Eric W. Biederman" 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

net: Fix userspace RTM_NEWLINK notifications.

2009-12-14T03:45:22+00:00

I received some bug reports about userspace programs having problems
because after RTM_NEWLINK was received they could not immediate access
files under /proc/sys/net/ because they had not been registered yet.

The original problem was trivially fixed by moving the userspace
notification from rtnetlink_event() to the end of
register_netdevice().

When testing that change I discovered I was still getting RTM_NEWLINK
events before I could access proc and I was also getting RTM_NEWLINK
events after I was seeing RTM_DELLINK.  Things practically guaranteed
to confuse userspace.

After a little more investigation these extra notifications proved to
be from the new notifiers NETDEV_POST_INIT and NETDEV_UNREGISTER_BATCH
hitting the default case in rtnetlink_event, and triggering
unnecessary RTM_NEWLINK messages.

rtnetlink_event now explicitly handles NETDEV_UNREGISTER_BATCH and
NETDEV_POST_INIT to avoid sending the incorrect userspace
notifications.

Signed-off-by: Eric W. Biederman 
Signed-off-by: David S. Miller

net: Support specifying the network namespace upon device creation.

2009-11-08T08:53:51+00:00

There is no good reason to not support userspace specifying the
network namespace during device creation, and it makes it easier
to create a network device and pass it to a child network namespace
with a well known name.

We have to be careful to ensure that the target network namespace
for the new device exists through the life of the call.  To keep
that logic clear I have factored out the network namespace grabbing
logic into rtnl_link_get_net.

In addtion we need to continue to pass the source network namespace
to the rtnl_link_ops.newlink method so that we can find the base
device source network namespace.

Signed-off-by: Eric W. Biederman 
Acked-by: Eric Dumazet

rtnetlink: Cleanups

2009-11-07T09:26:17+00:00

Pure cleanups patch

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: add a list_head parameter to dellink() method

2009-10-28T09:22:07+00:00

Adding a list_head parameter to rtnl_link_ops->dellink() methods
allow us to queue devices on a list, in order to dismantle
them all at once.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

rtnetlink: speedup rtnl_dump_ifinfo()

2009-10-24T13:13:17+00:00

When handling large number of netdevice, rtnl_dump_ifinfo()
is very slow because it has O(N^2) complexity.

Instead of scanning one single list, we can use the 256 sub lists
of the dev_index hash table.

This considerably speedups "ip link" operations

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

rtnetlink: rtnl_setlink() and rtnl_getlink() changes

2009-10-22T11:33:48+00:00

rtnl_getlink() & rtnl_setlink() run with RTNL held, we can use
__dev_get_by_index() and __dev_get_by_name() variants and avoid
dev_hold()/dev_put()

Adds to rtnl_getlink() the capability to find a device by its name,
not only by its index.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net_sched: reintroduce dev->qdisc for use by sch_api

2009-09-06T09:07:03+00:00

Currently the multiqueue integration with the qdisc API suffers from
a few problems:

- with multiple queues, all root qdiscs use the same handle. This means
  they can't be exposed to userspace in a backwards compatible fashion.

- all API operations always refer to queue number 0. Newly created
  qdiscs are automatically shared between all queues, its not possible
  to address individual queues or restore multiqueue behaviour once a
  shared qdisc has been attached.

- Dumps only contain the root qdisc of queue 0, in case of non-shared
  qdiscs this means the statistics are incomplete.

This patch reintroduces dev->qdisc, which points to the (single) root qdisc
from userspace's point of view. Currently it either points to the first
(non-shared) default qdisc, or a qdisc shared between all queues. The
following patches will introduce a classful dummy qdisc, which will be used
as root qdisc and contain the per-queue qdiscs as children.

Signed-off-by: Patrick McHardy 
Signed-off-by: David S. Miller

vlan: multiqueue vlan device

2009-09-03T01:03:00+00:00

vlan devices are currently not multi-queue capable.

We can do that with a new rtnl_link_ops method,
get_tx_queues(), called from rtnl_create_link()

This new method gets num_tx_queues/real_num_tx_queues
from real device.

register_vlan_device() is also handled.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: move and export get_net_ns_by_pid

2009-07-12T21:03:28+00:00

The function get_net_ns_by_pid(), to get a network
namespace from a pid_t, will be required in cfg80211
as well. Therefore, let's move it to net_namespace.c
and export it. We can't make it a static inline in
the !NETNS case because it needs to verify that the
given pid even exists (and return -ESRCH).

Signed-off-by: Johannes Berg 
Signed-off-by: David S. Miller