On Mon, Sep 21, 2015 at 02:20:22PM -0400, Tejun Heo wrote:
>
> store_release and load_acquire are different from the usual memory
> barriers and can't be paired this way.  You have to pair store_release
> and load_acquire.  Besides, it isn't a particularly good idea to

OK I've decided to drop the acquire/release helpers as they don't
help us at all and simply pessimises the code by using full memory
barriers (on some architectures) where only a write or read barrier
is needed.

> depend on memory barriers embedded in other data structures like the
> above.  Here, especially, rhashtable_insert() would have write barrier
> *before* the entry is hashed not necessarily *after*, which means that
> in the above case, a socket which appears to have set bound to a
> reader might not visible when the reader tries to look up the socket
> on the hashtable.

But you are right we do need an explicit write barrier here to
ensure that the hashing is visible.

> There's no reason to be overly smart here.  This isn't a crazy hot
> path, write barriers tend to be very cheap, store_release more so.
> Please just do smp_store_release() and note what it's paired with.

It's not about being overly smart.  It's about actually understanding
what's going on with the code.  I've seen too many instances of
people simply sprinkling synchronisation primitives around without
any knowledge of what is happening underneath, which is just a recipe
for creating hard-to-debug races.

> > @@ -1539,7 +1546,7 @@ static int netlink_bind(struct socket *sock, struct sockaddr *addr,
> >  		}
> >  	}
> >
> > -	if (!nlk->portid) {
> > +	if (!nlk->bound) {
>
> I don't think you can skip load_acquire here just because this is the
> second deref of the variable.  That doesn't change anything.  Race
> condition could still happen between the first and second tests and
> skipping the second would lead to the same kind of bug.

The reason this one is OK is because we do not use nlk->portid or
try to get nlk from the hash table before we return to user-space.

However, there is a real bug here that none of these acquire/release
helpers discovered.  The two bound tests here used to be a single
one.  Now that they are separate it is entirely possible for another
thread to come in the middle and bind the socket.  So we need to
repeat the portid check in order to maintain consistency.

> > @@ -1587,7 +1594,7 @@ static int netlink_connect(struct socket *sock, struct sockaddr *addr,
> >  	    !netlink_allowed(sock, NL_CFG_F_NONROOT_SEND))
> >  		return -EPERM;
> >
> > -	if (!nlk->portid)
> > +	if (!nlk->bound)
>
> Don't we need load_acquire here too?  Is this path holding a lock
> which makes that unnecessary?

Ditto.

---8<---
The commit 1f770c0a ("netlink:
Fix autobind race condition that leads to zero port ID") created
some new races that can occur due to inconcsistencies between the
two port IDs.

Tejun is right that a barrier is unavoidable.  Therefore I am
reverting to the original patch that used a boolean to indicate
that a user netlink socket has been bound.

Barriers have been added where necessary to ensure that a valid
portid and the hashed socket is visible.

I have also changed netlink_insert to only return EBUSY if the
socket is bound to a portid different to the requested one.  This
combined with only reading nlk->bound once in netlink_bind fixes
a race where two threads that bind the socket at the same time
with different port IDs may both succeed.

Fixes: 1f770c0a ("netlink: Fix autobind race condition that leads to zero port ID")
Reported-by: Tejun Heo <tj@kernel.org>
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Nacked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

This is primarily for consistancy with vxlan and other tunnels which
use network byte order for similar parameters.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We are seeing unexplained TX timeouts under heavy load. Let's try to get
a better idea of what's going on.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The low 16 bits of the 'opts1' field in the TX descriptor are supposed
to still contain the buffer length when the descriptor is handed back to
us. In practice, at least on my hardware, they don't. So stash the
original value of the opts1 field and get the length to unmap from
there.

There are other ways we could have worked out the length, but I actually
want a stash of the opts1 field anyway so that I can dump it alongside
the contents of the descriptor ring when we suffer a TX timeout.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We calculate the value of the opts1 descriptor field in three different
places. With two different behaviours when given an invalid packet to
be checksummed — none of them correct. Sort that out.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When sending a TSO frame in multiple buffers, we were neglecting to set
the first descriptor up in TSO mode.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

After a certain amount of staring at the debug output of this driver, I
realised it was lying to me.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If an RX interrupt was already received but NAPI has not yet run when
the RX timeout happens, we end up in cp_tx_timeout() with RX interrupts
already disabled. Blindly re-enabling them will cause an IRQ storm.

(This is made particularly horrid by the fact that cp_interrupt() always
returns that it's handled the interrupt, even when it hasn't actually
done anything. If it didn't do that, the core IRQ code would have
detected the storm and handled it, I'd have had a clear smoking gun
backtrace instead of just a spontaneously resetting router, and I'd have
at *least* two days of my life back. Changing the return value of
cp_interrupt() will be argued about under separate cover.)

Unconditionally leave RX interrupts disabled after the reset, and
schedule NAPI to check the receive ring and re-enable them.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Murali Karicheri says:

====================
net: netcp: a set of bug fixes

This patch series fixes a set of issues in netcp driver seen during internal
testing of the driver. While at it, do some clean up as well.

The fixes are tested on K2HK, K2L and K2E EVMs and the boot up logs can be
seen at

 http://pastebin.ubuntu.com/12533100/
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

A deadlock trace is seen in netcp driver with lockup detector enabled.
The trace log is provided below for reference. This patch fixes the
bug by removing the usage of netcp_modules_lock within ndo_ops functions.
ndo_{open/close/ioctl)() is already called with rtnl_lock held. So there
is no need to hold another mutex for serialization across processes on
multiple cores.  So remove use of netcp_modules_lock mutex from these
ndo ops functions.

ndo_set_rx_mode() shouldn't be using a mutex as it is called from atomic
context. In the case of ndo_set_rx_mode(), there can be call to this API
without rtnl_lock held from an atomic context. As the underlying modules
are expected to add address to a hardware table, it is to be protected
across concurrent updates and hence a spin lock is used to synchronize
the access. Same with ndo_vlan_rx_add_vid() & ndo_vlan_rx_kill_vid().

Probably the netcp_modules_lock is used to protect the module not being
removed as part of rmmod. Currently this is not fully implemented and
assumes the interface is brought down before doing rmmod of modules.
The support for rmmmod while interface is up is expected in a future
patch set when additional modules such as pa, qos are added. For now
all of the tests such as if up/down, reboot, iperf works fine with this
patch applied.

Deadlock trace seen with lockup detector enabled is shown below for
reference.

[   16.863014] ======================================================
[   16.869183] [ INFO: possible circular locking dependency detected ]
[   16.875441] 4.1.6-01265-gfb1e101 #1 Tainted: G        W
[   16.881176] -------------------------------------------------------
[   16.887432] ifconfig/1662 is trying to acquire lock:
[   16.892386]  (netcp_modules_lock){+.+.+.}, at: [<c03e8110>]
netcp_ndo_open+0x168/0x518
[   16.900321]
[   16.900321] but task is already holding lock:
[   16.906144]  (rtnl_mutex){+.+.+.}, at: [<c053a418>] devinet_ioctl+0xf8/0x7e4
[   16.913206]
[   16.913206] which lock already depends on the new lock.
[   16.913206]
[   16.921372]
[   16.921372] the existing dependency chain (in reverse order) is:
[   16.928844]
-> #1 (rtnl_mutex){+.+.+.}:
[   16.932865]        [<c06023f0>] mutex_lock_nested+0x68/0x4a8
[   16.938521]        [<c04c5758>] register_netdev+0xc/0x24
[   16.943831]        [<c03e65c0>] netcp_module_probe+0x214/0x2ec
[   16.949660]        [<c03e8a54>] netcp_register_module+0xd4/0x140
[   16.955663]        [<c089654c>] keystone_gbe_init+0x10/0x28
[   16.961233]        [<c000977c>] do_one_initcall+0xb8/0x1f8
[   16.966714]        [<c0867e04>] kernel_init_freeable+0x148/0x1e8
[   16.972720]        [<c05f9994>] kernel_init+0xc/0xe8
[   16.977682]        [<c0010038>] ret_from_fork+0x14/0x3c
[   16.982905]
-> #0 (netcp_modules_lock){+.+.+.}:
[   16.987619]        [<c006eab0>] lock_acquire+0x118/0x320
[   16.992928]        [<c06023f0>] mutex_lock_nested+0x68/0x4a8
[   16.998582]        [<c03e8110>] netcp_ndo_open+0x168/0x518
[   17.004064]        [<c04c48f0>] __dev_open+0xa8/0x10c
[   17.009112]        [<c04c4b74>] __dev_change_flags+0x94/0x144
[   17.014853]        [<c04c4c3c>] dev_change_flags+0x18/0x48
[   17.020334]        [<c053a9fc>] devinet_ioctl+0x6dc/0x7e4
[   17.025729]        [<c04a59ec>] sock_ioctl+0x1d0/0x2a8
[   17.030865]        [<c0142844>] do_vfs_ioctl+0x41c/0x688
[   17.036173]        [<c0142ae4>] SyS_ioctl+0x34/0x5c
[   17.041046]        [<c000ff60>] ret_fast_syscall+0x0/0x54
[   17.046441]
[   17.046441] other info that might help us debug this:
[   17.046441]
[   17.054434]  Possible unsafe locking scenario:
[   17.054434]
[   17.060343]        CPU0                    CPU1
[   17.064862]        ----                    ----
[   17.069381]   lock(rtnl_mutex);
[   17.072522]                                lock(netcp_modules_lock);
[   17.078875]                                lock(rtnl_mutex);
[   17.084532]   lock(netcp_modules_lock);
[   17.088366]
[   17.088366]  *** DEADLOCK ***
[   17.088366]
[   17.094279] 1 lock held by ifconfig/1662:
[   17.098278]  #0:  (rtnl_mutex){+.+.+.}, at: [<c053a418>]
devinet_ioctl+0xf8/0x7e4
[   17.105774]
[   17.105774] stack backtrace:
[   17.110124] CPU: 1 PID: 1662 Comm: ifconfig Tainted: G        W
4.1.6-01265-gfb1e101 #1
[   17.118637] Hardware name: Keystone
[   17.122123] [<c00178e4>] (unwind_backtrace) from [<c0013cbc>]
(show_stack+0x10/0x14)
[   17.129862] [<c0013cbc>] (show_stack) from [<c05ff450>]
(dump_stack+0x84/0xc4)
[   17.137079] [<c05ff450>] (dump_stack) from [<c0068e34>]
(print_circular_bug+0x210/0x330)
[   17.145161] [<c0068e34>] (print_circular_bug) from [<c006ab7c>]
(validate_chain.isra.35+0xf98/0x13ac)
[   17.154372] [<c006ab7c>] (validate_chain.isra.35) from [<c006da60>]
(__lock_acquire+0x52c/0xcc0)
[   17.163149] [<c006da60>] (__lock_acquire) from [<c006eab0>]
(lock_acquire+0x118/0x320)
[   17.171058] [<c006eab0>] (lock_acquire) from [<c06023f0>]
(mutex_lock_nested+0x68/0x4a8)
[   17.179140] [<c06023f0>] (mutex_lock_nested) from [<c03e8110>]
(netcp_ndo_open+0x168/0x518)
[   17.187484] [<c03e8110>] (netcp_ndo_open) from [<c04c48f0>]
(__dev_open+0xa8/0x10c)
[   17.195133] [<c04c48f0>] (__dev_open) from [<c04c4b74>]
(__dev_change_flags+0x94/0x144)
[   17.203129] [<c04c4b74>] (__dev_change_flags) from [<c04c4c3c>]
(dev_change_flags+0x18/0x48)
[   17.211560] [<c04c4c3c>] (dev_change_flags) from [<c053a9fc>]
(devinet_ioctl+0x6dc/0x7e4)
[   17.219729] [<c053a9fc>] (devinet_ioctl) from [<c04a59ec>]
(sock_ioctl+0x1d0/0x2a8)
[   17.227378] [<c04a59ec>] (sock_ioctl) from [<c0142844>]
(do_vfs_ioctl+0x41c/0x688)
[   17.234939] [<c0142844>] (do_vfs_ioctl) from [<c0142ae4>]
(SyS_ioctl+0x34/0x5c)
[   17.242242] [<c0142ae4>] (SyS_ioctl) from [<c000ff60>]
(ret_fast_syscall+0x0/0x54)
[   17.258855] netcp-1.0 2620110.netcp eth0: Link is Up - 1Gbps/Full - flow
control off
[   17.271282] BUG: sleeping function called from invalid context at
kernel/locking/mutex.c:616
[   17.279712] in_atomic(): 1, irqs_disabled(): 0, pid: 1662, name: ifconfig
[   17.286500] INFO: lockdep is turned off.
[   17.290413] Preemption disabled at:[<  (null)>]   (null)
[   17.295728]
[   17.297214] CPU: 1 PID: 1662 Comm: ifconfig Tainted: G        W
4.1.6-01265-gfb1e101 #1
[   17.305735] Hardware name: Keystone
[   17.309223] [<c00178e4>] (unwind_backtrace) from [<c0013cbc>]
(show_stack+0x10/0x14)
[   17.316970] [<c0013cbc>] (show_stack) from [<c05ff450>]
(dump_stack+0x84/0xc4)
[   17.324194] [<c05ff450>] (dump_stack) from [<c06023b0>]
(mutex_lock_nested+0x28/0x4a8)
[   17.332112] [<c06023b0>] (mutex_lock_nested) from [<c03e9840>]
(netcp_set_rx_mode+0x160/0x210)
[   17.340724] [<c03e9840>] (netcp_set_rx_mode) from [<c04c483c>]
(dev_set_rx_mode+0x1c/0x28)
[   17.348982] [<c04c483c>] (dev_set_rx_mode) from [<c04c490c>]
(__dev_open+0xc4/0x10c)
[   17.356724] [<c04c490c>] (__dev_open) from [<c04c4b74>]
(__dev_change_flags+0x94/0x144)
[   17.364729] [<c04c4b74>] (__dev_change_flags) from [<c04c4c3c>]
(dev_change_flags+0x18/0x48)
[   17.373166] [<c04c4c3c>] (dev_change_flags) from [<c053a9fc>]
(devinet_ioctl+0x6dc/0x7e4)
[   17.381344] [<c053a9fc>] (devinet_ioctl) from [<c04a59ec>]
(sock_ioctl+0x1d0/0x2a8)
[   17.388994] [<c04a59ec>] (sock_ioctl) from [<c0142844>]
(do_vfs_ioctl+0x41c/0x688)
[   17.396563] [<c0142844>] (do_vfs_ioctl) from [<c0142ae4>]
(SyS_ioctl+0x34/0x5c)
[   17.403873] [<c0142ae4>] (SyS_ioctl) from [<c000ff60>]
(ret_fast_syscall+0x0/0x54)
[   17.413772] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
udhcpc (v1.20.2) started
Sending discover...
[   18.690666] netcp-1.0 2620110.netcp eth0: Link is Up - 1Gbps/Full - flow
control off
Sending discover...
[   22.250972] netcp-1.0 2620110.netcp eth0: Link is Up - 1Gbps/Full - flow
control off
[   22.258721] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[   22.265458] BUG: sleeping function called from invalid context at
kernel/locking/mutex.c:616
[   22.273896] in_atomic(): 1, irqs_disabled(): 0, pid: 342, name: kworker/1:1
[   22.280854] INFO: lockdep is turned off.
[   22.284767] Preemption disabled at:[<  (null)>]   (null)
[   22.290074]
[   22.291568] CPU: 1 PID: 342 Comm: kworker/1:1 Tainted: G        W
4.1.6-01265-gfb1e101 #1
[   22.300255] Hardware name: Keystone
[   22.303750] Workqueue: ipv6_addrconf addrconf_dad_work
[   22.308895] [<c00178e4>] (unwind_backtrace) from [<c0013cbc>]
(show_stack+0x10/0x14)
[   22.316643] [<c0013cbc>] (show_stack) from [<c05ff450>]
(dump_stack+0x84/0xc4)
[   22.323867] [<c05ff450>] (dump_stack) from [<c06023b0>]
(mutex_lock_nested+0x28/0x4a8)
[   22.331786] [<c06023b0>] (mutex_lock_nested) from [<c03e9840>]
(netcp_set_rx_mode+0x160/0x210)
[   22.340394] [<c03e9840>] (netcp_set_rx_mode) from [<c04c9d18>]
(__dev_mc_add+0x54/0x68)
[   22.348401] [<c04c9d18>] (__dev_mc_add) from [<c05ab358>]
(igmp6_group_added+0x168/0x1b4)
[   22.356580] [<c05ab358>] (igmp6_group_added) from [<c05ad2cc>]
(ipv6_dev_mc_inc+0x4f0/0x5a8)
[   22.365019] [<c05ad2cc>] (ipv6_dev_mc_inc) from [<c058f0d0>]
(addrconf_dad_work+0x21c/0x33c)
[   22.373460] [<c058f0d0>] (addrconf_dad_work) from [<c0042850>]
(process_one_work+0x214/0x8d0)
[   22.381986] [<c0042850>] (process_one_work) from [<c0042f54>]
(worker_thread+0x48/0x4bc)
[   22.390071] [<c0042f54>] (worker_thread) from [<c004868c>]
(kthread+0xf0/0x108)
[   22.397381] [<c004868c>] (kthread) from [<c0010038>]

Trace related to incorrect usage of mutex inside ndo_set_rx_mode

[   24.086066] BUG: sleeping function called from invalid context at
kernel/locking/mutex.c:616
[   24.094506] in_atomic(): 1, irqs_disabled(): 0, pid: 1682, name: ifconfig
[   24.101291] INFO: lockdep is turned off.
[   24.105203] Preemption disabled at:[<  (null)>]   (null)
[   24.110511]
[   24.112005] CPU: 2 PID: 1682 Comm: ifconfig Tainted: G        W
4.1.6-01265-gfb1e101 #1
[   24.120518] Hardware name: Keystone
[   24.124018] [<c00178e4>] (unwind_backtrace) from [<c0013cbc>]
(show_stack+0x10/0x14)
[   24.131772] [<c0013cbc>] (show_stack) from [<c05ff450>]
(dump_stack+0x84/0xc4)
[   24.138989] [<c05ff450>] (dump_stack) from [<c06023b0>]
(mutex_lock_nested+0x28/0x4a8)
[   24.146908] [<c06023b0>] (mutex_lock_nested) from [<c03e9840>]
(netcp_set_rx_mode+0x160/0x210)
[   24.155523] [<c03e9840>] (netcp_set_rx_mode) from [<c04c483c>]
(dev_set_rx_mode+0x1c/0x28)
[   24.163787] [<c04c483c>] (dev_set_rx_mode) from [<c04c490c>]
(__dev_open+0xc4/0x10c)
[   24.171531] [<c04c490c>] (__dev_open) from [<c04c4b74>]
(__dev_change_flags+0x94/0x144)
[   24.179528] [<c04c4b74>] (__dev_change_flags) from [<c04c4c3c>]
(dev_change_flags+0x18/0x48)
[   24.187966] [<c04c4c3c>] (dev_change_flags) from [<c053a9fc>]
(devinet_ioctl+0x6dc/0x7e4)
[   24.196145] [<c053a9fc>] (devinet_ioctl) from [<c04a59ec>]
(sock_ioctl+0x1d0/0x2a8)
[   24.203803] [<c04a59ec>] (sock_ioctl) from [<c0142844>]
(do_vfs_ioctl+0x41c/0x688)
[   24.211373] [<c0142844>] (do_vfs_ioctl) from [<c0142ae4>]
(SyS_ioctl+0x34/0x5c)
[   24.218676] [<c0142ae4>] (SyS_ioctl) from [<c000ff60>]
(ret_fast_syscall+0x0/0x54)
[   24.227156] IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently netcp_rxpool_refill() that refill descriptors and attached
buffers to fdq while interrupt is enabled as part of NAPI poll. Doing
it while interrupt is disabled could be beneficial as hardware will
not be starved when CPU is busy with processing interrupt.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently netcp_module_probe() doesn't check the return value of
of_parse_phandle() that points to the interface data for the
module and then pass the node ptr to the module which is incorrect.
Check for return value and free the intf_modpriv if there is error.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently, if netcp_allocate_rx_buf() fails due no descriptors
in the rx free descriptor queue, inside the netcp_rxpool_refill() function
the iterative loop to fill buffers doesn't terminate right away. So modify
the netcp_allocate_rx_buf() to return an error code and use it break the
loop when there is error.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The netcp interface is not fully initialized before attach the module
to the interface. For example, the tx pipe/rx pipe is initialized
in ethss module as part of attach(). So until this is complete, the
interface can't be registered.  So move registration of interface to
net device outside the current loop that attaches the modules to the
interface.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

netcp_core is the first driver that will get initialized and the modules
(ethss, pa etc) will then get initialized. So the code at the end of
netcp_probe() that iterate over the modules is a dead code as the module
list will be always be empty. So remove this code.
Signed-off-by: Murali Karicheri <m-karicheri2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

On K2HK, sgmii module registers of slave 0 and 1 are mem
mapped to one contiguous block, while those of slave 2
and 3 are mapped to another contiguous block.  However,
on K2E and K2L, sgmii module registers of all slaves are
mem mapped to one contiguous block.  SGMII APIs expect
slave 0 sgmii base when API is invoked for slave 0 and 1,
and slave 2 sgmii base when invoked for other slaves.
Before this patch, slave 0 sgmii base is always passed to
sgmii API for K2E regardless which slave is the API invoked
for.  This patch fixes the problem.
Signed-off-by: WingMan Kwok <w-kwok2@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 7d824109 ("virtio: add explicit big-endian support to memory
accessors") accidentally changed the virtio_net header used by
AF_PACKET with PACKET_VNET_HDR from host-endian to big-endian.

Since virtio_legacy_is_little_endian() is a very long identifier,
define a vio_le macro and use that throughout the code instead of the
hard-coded 'false' for little-endian.

This restores the ABI to match 4.1 and earlier kernels, and makes my
test program work again.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Drivers might call napi_disable while not holding the napi instance poll_lock.
In those instances, its possible for a race condition to exist between
poll_one_napi and napi_disable.  That is to say, poll_one_napi only tests the
NAPI_STATE_SCHED bit to see if there is work to do during a poll, and as such
the following may happen:

CPU0				CPU1
ndo_tx_timeout			napi_poll_dev
 napi_disable			 poll_one_napi
  test_and_set_bit (ret 0)
				  test_bit (ret 1)
   reset adapter		   napi_poll_routine

If the adapter gets a tx timeout without a napi instance scheduled, its possible
for the adapter to think it has exclusive access to the hardware  (as the napi
instance is now scheduled via the napi_disable call), while the netpoll code
thinks there is simply work to do.  The result is parallel hardware access
leading to corrupt data structures in the driver, and a crash.

Additionaly, there is another, more critical race between netpoll and
napi_disable.  The disabled napi state is actually identical to the scheduled
state for a given napi instance.  The implication being that, if a napi instance
is disabled, a netconsole instance would see the napi state of the device as
having been scheduled, and poll it, likely while the driver was dong something
requiring exclusive access.  In the case above, its fairly clear that not having
the rings in a state ready to be polled will cause any number of crashes.

The fix should be pretty easy.  netpoll uses its own bit to indicate that that
the napi instance is in a state of being serviced by netpoll (NAPI_STATE_NPSVC).
We can just gate disabling on that bit as well as the sched bit.  That should
prevent netpoll from conducting a napi poll if we convert its set bit to a
test_and_set_bit operation to provide mutual exclusion

Change notes:
V2)
	Remove a trailing whtiespace
	Resubmit with proper subject prefix

V3)
	Clean up spacing nits
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: jmaxwell@redhat.com
Tested-by: jmaxwell@redhat.com
Signed-off-by: David S. Miller <davem@davemloft.net>

RST packets sent on behalf of TCP connections with TS option (RFC 7323
TCP timestamps) have incorrect TS val (set to 0), but correct TS ecr.

A > B: Flags [S], seq 0, win 65535, options [mss 1000,nop,nop,TS val 100
ecr 0], length 0
B > A: Flags [S.], seq 2444755794, ack 1, win 28960, options [mss
1460,nop,nop,TS val 7264344 ecr 100], length 0
A > B: Flags [.], ack 1, win 65535, options [nop,nop,TS val 110 ecr
7264344], length 0

B > A: Flags [R.], seq 1, ack 1, win 28960, options [nop,nop,TS val 0
ecr 110], length 0

We need to call skb_mstamp_get() to get proper TS val,
derived from skb->skb_mstamp

Note that RFC 1323 was advocating to not send TS option in RST segment,
but RFC 7323 recommends the opposite :

  Once TSopt has been successfully negotiated, that is both <SYN> and
  <SYN,ACK> contain TSopt, the TSopt MUST be sent in every non-<RST>
  segment for the duration of the connection, and SHOULD be sent in an
  <RST> segment (see Section 5.2 for details)

Note this RFC recommends to send TS val = 0, but we believe it is
premature : We do not know if all TCP stacks are properly
handling the receive side :

   When an <RST> segment is
   received, it MUST NOT be subjected to the PAWS check by verifying an
   acceptable value in SEG.TSval, and information from the Timestamps
   option MUST NOT be used to update connection state information.
   SEG.TSecr MAY be used to provide stricter <RST> acceptance checks.

In 5 years, if/when all TCP stack are RFC 7323 ready, we might consider
to decide to send TS val = 0, if it buys something.

Fixes: 7faee5c0 ("tcp: remove TCP_SKB_CB(skb)->when")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The Marvell Egress rx trailer check must be fixed to
correctly detect bad bits in the third byte of the
Eggress trailer as described in the Table 28 of the
88E6060 datasheet.
The current code incorrectly omits to check the third
byte and checks the fourth byte twice.
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

rhashtable_rehash_one() uses complex logic to update entry->next field,
after INIT_RHT_NULLS_HEAD and NULLS_MARKER expansion:

entry->next = 1 | ((base + off) << 1)

This can be compiled along the lines of:

entry->next = base + off
entry->next <<= 1
entry->next |= 1

Which will break concurrent readers.

NULLS value recomputation is not needed here, so just remove
the complex logic.

The data race was found with KernelThreadSanitizer (KTSAN).
Signed-off-by: Dmitry Vyukov <dvyukov@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Converts the ch9200 driver to use the module_usb_driver() macro which
makes the code smaller and a bit simpler.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Acked-by: Matthew Garrett <mjg59@srcf.ucam.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

When support for megaflows was introduced, OVS needed to start
installing flows with a mask applied to them. Since masking is an
expensive operation, OVS also had an optimization that would only
take the parts of the flow keys that were covered by a non-zero
mask. The values stored in the remaining pieces should not matter
because they are masked out.

While this works fine for the purposes of matching (which must always
look at the mask), serialization to netlink can be problematic. Since
the flow and the mask are serialized separately, the uninitialized
portions of the flow can be encoded with whatever values happen to be
present.

In terms of functionality, this has little effect since these fields
will be masked out by definition. However, it leaks kernel memory to
userspace, which is a potential security vulnerability. It is also
possible that other code paths could look at the masked key and get
uninitialized data, although this does not currently appear to be an
issue in practice.

This removes the mask optimization for flows that are being installed.
This was always intended to be the case as the mask optimizations were
really targetting per-packet flow operations.

Fixes: 03f0d916 ("openvswitch: Mega flow implementation")
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 54d792f2 ("net: dsa: Centralise global and port setup
code into mv88e6xxx.") merged in the 4.2 merge window broke the link
speed forcing for the CPU port of Marvell DSA switches.  The original
code was:

        /* MAC Forcing register: don't force link, speed, duplex
         * or flow control state to any particular values on physical
         * ports, but force the CPU port and all DSA ports to 1000 Mb/s
         * full duplex.
         */
        if (dsa_is_cpu_port(ds, p) || ds->dsa_port_mask & (1 << p))
                REG_WRITE(addr, 0x01, 0x003e);
        else
                REG_WRITE(addr, 0x01, 0x0003);

but the new code does a read-modify-write:

                reg = _mv88e6xxx_reg_read(ds, REG_PORT(port), PORT_PCS_CTRL);
                if (dsa_is_cpu_port(ds, port) ||
                    ds->dsa_port_mask & (1 << port)) {
                        reg |= PORT_PCS_CTRL_FORCE_LINK |
                                PORT_PCS_CTRL_LINK_UP |
                                PORT_PCS_CTRL_DUPLEX_FULL |
                                PORT_PCS_CTRL_FORCE_DUPLEX;
                        if (mv88e6xxx_6065_family(ds))
                                reg |= PORT_PCS_CTRL_100;
                        else
                                reg |= PORT_PCS_CTRL_1000;

The link speed in the PCS control register is a two bit field.  Forcing
the link speed in this way doesn't ensure that the bit field is set to
the correct value - on the hardware I have here, the speed bitfield
remains set to 0x03, resulting in the speed not being forced to gigabit.

We must clear both bits before forcing the link speed.

Fixes: 54d792f2 ("net: dsa: Centralise global and port setup code into mv88e6xxx.")
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Acked-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Partially due to a pre-exising "thinko", the new metadata-based tx/rx
paths were handling ECN propagation differently than the traditional
tx/rx paths.  This patch removes the "thinko" (involving multiple
ip_hdr assignments) on the rx path and corrects the ECN handling on
both the rx and tx paths.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Before allowing lockless LISTEN processing, we need to make
sure to arm the SYN_RECV timer before the req socket is visible
in hash tables.

Also, req->rsk_hash should be written before we set rsk_refcnt
to a non zero value.

Fixes: fa76ce73 ("inet: get rid of central tcp/dccp listener timer")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ying Cai <ycai@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When creating a timewait socket, we need to arm the timer before
allowing other cpus to find it. The signal allowing cpus to find
the socket is setting tw_refcnt to non zero value.

As we set tw_refcnt in __inet_twsk_hashdance(), we therefore need to
call inet_twsk_schedule() first.

This also means we need to remove tw_refcnt changes from
inet_twsk_schedule() and let the caller handle it.

Note that because we use mod_timer_pinned(), we have the guarantee
the timer wont expire before we set tw_refcnt as we run in BH context.

To make things more readable I introduced inet_twsk_reschedule() helper.

When rearming the timer, we can use mod_timer_pending() to make sure
we do not rearm a canceled timer.

Note: This bug can possibly trigger if packets of a flow can hit
multiple cpus. This does not normally happen, unless flow steering
is broken somehow. This explains this bug was spotted ~5 months after
its introduction.

A similar fix is needed for SYN_RECV sockets in reqsk_queue_hash_req(),
but will be provided in a separate patch for proper tracking.

Fixes: 789f558c ("tcp/dccp: get rid of central timewait timer")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Ying Cai <ycai@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

`ls /sys/devices/channel-devices/vnet-port-0-0/net' is missing without
this change, and applications like NetworkManager are looking in
sysfs for the information.
Signed-off-by: Sowmini Varadhan <sowmini.varadhan@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The code handling vlan tag insertion was dropped in commit 371bd106
("geneve: Consolidate Geneve functionality in single module.").  Now we
need to drop the related vlan feature bits in the netdev structure.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reviewed-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There's a bunch of cheap USB 10/100 devices based on QinHeng chipsets. The
vendor driver supports the CH9100 and CH9200 devices, but the majority of
the code is of the if (ch9100) {} else {} form, with the most significant
difference being that CH9200 provides a real MII interface but CH9100 fakes
one with a bunch of global variables and magic commands. I don't have a
CH9100, so it's probably better if someone who does provides an independent
driver for it. In any case, this is a lightly cleaned up version of the
vendor driver with all the CH9100 code dropped.
Signed-off-by: Matthew Garrett <mjg59@srcf.ucam.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Luis de Bethencourt says:

====================
net: phy: Fix module autoload for OF platform drivers

These patches add the missing MODULE_DEVICE_TABLE() for OF to export
the information so modules have the correct aliases built-in and
autoloading works correctly.

A longer explanation by Javier Canillas can be found here:
https://lkml.org/lkml/2015/7/30/519
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This platform driver has a OF device ID table but the OF module
alias information is not created so module autoloading won't work.
Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This platform driver has a OF device ID table but the OF module
alias information is not created so module autoloading won't work.
Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Luis de Bethencourt says:

====================
net: Fix module autoload for OF platform drivers

These patches add the missing MODULE_DEVICE_TABLE() for OF to export
the information so modules have the correct aliases built-in and
autoloading works correctly.

A longer explanation by Javier Canillas can be found here:
https://lkml.org/lkml/2015/7/30/519
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This platform driver has a OF device ID table but the OF module
alias information is not created so module autoloading won't work.
Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This platform driver has a OF device ID table but the OF module
alias information is not created so module autoloading won't work.
Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This platform driver has a OF device ID table but the OF module
alias information is not created so module autoloading won't work.
Signed-off-by: Luis de Bethencourt <luis@osg.samsung.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This platform driver has a OF device ID table but the OF module
alias information is not created so module autoloading won't work.
Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This platform driver has a OF device ID table but the OF module
alias information is not created so module autoloading won't work.
Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The commit c0bb07df ("netlink:
Reset portid after netlink_insert failure") introduced a race
condition where if two threads try to autobind the same socket
one of them may end up with a zero port ID.  This led to kernel
deadlocks that were observed by multiple people.

This patch reverts that commit and instead fixes it by introducing
a separte rhash_portid variable so that the real portid is only set
after the socket has been successfully hashed.

Fixes: c0bb07df ("netlink: Reset portid after netlink_insert failure")
Reported-by: Tejun Heo <tj@kernel.org>
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>