linux-toradex.git/net/tipc/eth_media.c, branch v3.4.108

tipc: fix lockdep warning during bearer initialization

2013-09-14T13:02:11+00:00

[ Upstream commit 4225a398c1352a7a5c14dc07277cb5cc4473983b ]

When the lockdep validator is enabled, it will report the below
warning when we enable a TIPC bearer:

[ INFO: possible irq lock inversion dependency detected ]
---------------------------------------------------------
Possible interrupt unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock(ptype_lock);
                                local_irq_disable();
                                lock(tipc_net_lock);
                                lock(ptype_lock);
   
   lock(tipc_net_lock);

  *** DEADLOCK ***

the shortest dependencies between 2nd lock and 1st lock:
  -> (ptype_lock){+.+...} ops: 10 {
[...]
SOFTIRQ-ON-W at:
                      [] __lock_acquire+0x528/0x13e0
                      [] lock_acquire+0x90/0x100
                      [] _raw_spin_lock+0x38/0x50
                      [] dev_add_pack+0x3a/0x60
                      [] arp_init+0x1a/0x48
                      [] inet_init+0x181/0x27e
                      [] do_one_initcall+0x34/0x170
                      [] kernel_init+0x110/0x1b2
                      [] kernel_thread_helper+0x6/0x10
[...]
   ... key      at: [] ptype_lock+0x10/0x20
   ... acquired at:
    [] lock_acquire+0x90/0x100
    [] _raw_spin_lock+0x38/0x50
    [] dev_add_pack+0x3a/0x60
    [] enable_bearer+0xf2/0x140 [tipc]
    [] tipc_enable_bearer+0x1ba/0x450 [tipc]
    [] tipc_cfg_do_cmd+0x5c4/0x830 [tipc]
    [] handle_cmd+0x42/0xd0 [tipc]
    [] genl_rcv_msg+0x232/0x280
    [] netlink_rcv_skb+0x86/0xb0
    [] genl_rcv+0x1c/0x30
    [] netlink_unicast+0x174/0x1f0
    [] netlink_sendmsg+0x1eb/0x2d0
    [] sock_aio_write+0x161/0x170
    [] do_sync_write+0xac/0xf0
    [] vfs_write+0x156/0x170
    [] sys_write+0x42/0x70
    [] sysenter_do_call+0x12/0x38
[...]
}
  -> (tipc_net_lock){+..-..} ops: 4 {
[...]
    IN-SOFTIRQ-R at:
                     [] __lock_acquire+0x64a/0x13e0
                     [] lock_acquire+0x90/0x100
                     [] _raw_read_lock_bh+0x3d/0x50
                     [] tipc_recv_msg+0x1d/0x830 [tipc]
                     [] recv_msg+0x3f/0x50 [tipc]
                     [] __netif_receive_skb+0x22a/0x590
                     [] netif_receive_skb+0x2b/0xf0
                     [] pcnet32_poll+0x292/0x780
                     [] net_rx_action+0xfa/0x1e0
                     [] __do_softirq+0xae/0x1e0
[...]
}

>From the log, we can see three different call chains between
CPU0 and CPU1:

Time 0 on CPU0:

  kernel_init()->inet_init()->dev_add_pack()

At time 0, the ptype_lock is held by CPU0 in dev_add_pack();

Time 1 on CPU1:

  tipc_enable_bearer()->enable_bearer()->dev_add_pack()

At time 1, tipc_enable_bearer() first holds tipc_net_lock, and then
wants to take ptype_lock to register TIPC protocol handler into the
networking stack.  But the ptype_lock has been taken by dev_add_pack()
on CPU0, so at this time the dev_add_pack() running on CPU1 has to be
busy looping.

Time 2 on CPU0:

  netif_receive_skb()->recv_msg()->tipc_recv_msg()

At time 2, an incoming TIPC packet arrives at CPU0, hence
tipc_recv_msg() will be invoked. In tipc_recv_msg(), it first wants
to hold tipc_net_lock.  At the moment, below scenario happens:

On CPU0, below is our sequence of taking locks:

  lock(ptype_lock)->lock(tipc_net_lock)

On CPU1, our sequence of taking locks looks like:

  lock(tipc_net_lock)->lock(ptype_lock)

Obviously deadlock may happen in this case.

But please note the deadlock possibly doesn't occur at all when the
first TIPC bearer is enabled.  Before enable_bearer() -- running on
CPU1 does not hold ptype_lock, so the TIPC receive handler (i.e.
recv_msg()) is not registered successfully via dev_add_pack(), so
the tipc_recv_msg() cannot be called by recv_msg() even if a TIPC
message comes to CPU0. But when the second TIPC bearer is
registered, the deadlock can perhaps really happen.

To fix it, we will push the work of registering TIPC protocol
handler into workqueue context. After the change, both paths taking
ptype_lock are always in process contexts, thus, the deadlock should
never occur.

Signed-off-by: Ying Xue 
Signed-off-by: Jon Maloy 
Signed-off-by: Paul Gortmaker 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

tipc: rename struct media to struct tipc_media

2011-12-30T02:53:29+00:00

Give it a meaningful prefix, as suggested by DaveM, so that it
is consistent with things like struct tipc_bearer, and so it isn't
confused with anything else.  This has no impact on the actual
runtime code behaviour.

Signed-off-by: Paul Gortmaker

tipc: Eliminate useless memset operations in Ethernet media support

2011-12-27T16:33:44+00:00

Gets rid of two pointless operations that zero out the array used to
record information about TIPC's Ethernet bearers. There is no need to
initialize the array on start up since it is a global variable that is
already zero'd out, and there is no need to zero it out on exit because
the array is never referenced again.

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker

tipc: Do timely cleanup of disabled Ethernet bearer resources

2011-12-27T16:33:42+00:00

Modifies Ethernet bearer disable logic to break the association between
the bearer and its device driver at the time the bearer is disabled,
rather than when the TIPC module is unloaded. This allows the array
entry used by the disabled bearer to be re-used if the same bearer (or
a different one) is subsequently enabled.

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker

tipc: Minor optimization to deactivation of Ethernet media suppot

2011-12-27T16:33:41+00:00

Change TIPC's shutdown code to deactivate generic networking support
before terminating Ethernet media support. The deactivation of generic
networking support causes all existing bearers to be destroyed, meaning
the Ethernet media termination routine no longer has to bother marking
them as unavailable.

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker

tipc: Hide media-specific addressing details from generic bearer code

2011-12-27T16:33:36+00:00

Reworks TIPC's media address data structure and associated processing
routines to transfer all media-specific details of address conversion
to the associated TIPC media adaptation code. TIPC's generic bearer code
now only needs to know which media type an address is associated with
and whether or not it is a broadcast address, and totally ignores the
"value" field that contains the actual media-specific addressing info.

These changes eliminate the need for a number of endianness conversion
operations and will make it easier for TIPC to support new media types
in the future.

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker

tipc: Add new address conversion routines for Ethernet media

2011-12-27T16:33:27+00:00

Enhances TIPC's Ethernet media support to provide 3 new address conversion
routines, which allow TIPC to interpret an address that is in string form
and to convert an address to and from the 20 byte format used in TIPC's
neighbor discovery messages.

These routines are pre-requisites to a follow on commit that hides all
media-specific addressing details from TIPC's generic bearer code.

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker

tipc: Improve handling of media address printing errors

2011-12-27T16:13:06+00:00

Enhances conversion of a media address to printable form so that an
unconvertable address will be displayed as a string of hex digits,
rather than not being displayed at all. (Also removes a pointless check
for the existence of the media-specific address conversion routine,
since the routine is not optional.)

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker

tipc: Register new media using pre-compiled structure

2011-12-27T16:13:05+00:00

Speeds up the registration of TIPC media types by passing in a structure
containing the required information, rather than by passing in the various
fields describing the media type individually.

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker

tipc: Lower limits for number of bearers and media types

2011-09-18T02:55:07+00:00

Reduces the number of bearers a node can support to 2, which can use
identical or non-identical media. This change won't impact users,
since they are currently limited to a maximum of 2 Ethernet bearers,
and will save memory by eliminating a number of unused entries in
TIPC's media and bearer arrays.

Signed-off-by: Allan Stephens 
Signed-off-by: Paul Gortmaker