linux.git/drivers/net, branch v4.1-rc2

net: fec: Fix RGMII-ID mode

2015-04-30T20:48:53+00:00

RGMII-ID uses an internal delay within the transmitter or receiver. This
feature is phy specific. The rest of the communication is normal RGMII.

So the fec driver has to check for all RGMII modes, not only
'PHY_INTERFACE_MODE_RGMII'.

Signed-off-by: Markus Pargmann 
Signed-off-by: David S. Miller

net/mlx4_en: Schedule napi when RX buffers allocation fails

2015-04-30T20:47:50+00:00

When system is out of memory, refilling of RX buffers fails while
the driver continue to pass the received packets to the kernel stack.
At some point, when all RX buffers deplete, driver may fall into a
sleep, and not recover when memory for new RX buffers is once again
availible. This is because hardware does not have valid descriptors,
so no interrupt will be generated for the driver to return to work
in napi context. Fix it by schedule the napi poll function from
stats_task delayed workqueue, as long as the allocations fail.

Signed-off-by: Ido Shamay 
Signed-off-by: Amir Vadai 
Signed-off-by: David S. Miller

netxen_nic: use spin_[un]lock_bh around tx_clean_lock

2015-04-30T20:37:29+00:00

While testing this driver with DEBUG_LOCKDEP and DEBUG_SPINLOCK
enabled did not produce any traces, it would be more prudent in the
case of tx_clean_lock to use spin_[un]lock_bh, since this lock is
manipulated in both the process and softirq contexts.

This patch was tested for functionality and regressions with netperf
and DEBUG_LOCKDEP and DEBUG_SPINLOCK enabled.

Signed-off-by: Tony Camuso 
Acked-by: Neil Horman 
Signed-off-by: David S. Miller

net/mlx4_core: Fix unaligned accesses

2015-04-30T20:26:30+00:00

Addresses the following kernel logs seen during boot:

Kernel unaligned access at TPC[100ee150] mlx4_QUERY_HCA+0x80/0x248 [mlx4_core]
Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core]
Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core]
Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core]
Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core]

Signed-off-by: David Ahern 
Acked-by: Or Gerlitz 
Signed-off-by: David S. Miller

mlx4_en: Use correct loop cursor in error path.

2015-04-30T20:25:14+00:00

Signed-off-by: Benjamin Poirier 
Fixes: 9e311e7 ("net/mlx4_en: Use affinity hint")
Acked-by: Amir Vadai 
Signed-off-by: David S. Miller

cxgb4: Fix MC1 memory offset calculation

2015-04-29T19:50:16+00:00

Commit 6559a7e8296002b4 ("cxgb4: Cleanup macros so they follow the same
style and look consistent") introduced a regression where reading MC1
memory in adapters where MC0 isn't present or MC0 size is not equal to MC1
size caused the adapter to crash due to incorrect computation of memoffset.
Fix is to read the size of MC0 instead of MC1 for offset calculation

Signed-off-by: Steve Wise 
Signed-off-by: Hariprasad Shenai 
Signed-off-by: David S. Miller

bnx2x: Delay during kdump load

2015-04-29T19:49:21+00:00

In a kdump environment interfaces might be re-loaded without a proper
unload sequence in the previous running kernel.
bnx2x management FW and driver maintains a `pulse' that notifies the FW
that the driver is still up and running.

Driver load on the kdump kernel should be performed only after the pulse
has been out-of-sync long enough for the management FW to identify that
the driver has crashed, on which point it will perform some necessary
cleanup of the HW.

In today's distros kdump loading is quite fast, sometimes too fast for our
FW to get out-of-sync. This patch delays the bnx2x's probe during kdump
to allow a proper re-load on the kdump kernel.

Signed-off-by: Yuval Mintz 
Signed-off-by: David S. Miller

net: Fix Kernel Panic in bonding driver debugfs file: rlb_hash_table

2015-04-29T19:37:19+00:00

This patch fixes a Kernel Panic in bonding driver debugfs file: rlb_hash_table.

$> modprobe bonding mode=6
$> cat /sys/kernel/debug/bonding/bond0/rlb_hash_table

This will crash the kernel. The struct alb_bond_info is initialized only when
the bonding interface is initialized (ip link set bond0 up) and not at the time
it is allocated. If we try to read the table before that, it'll result in a
kernel panic.

The patch applies against both net and net-next

Signed-off-by: Vishwanath Pai 
Signed-off-by: Andy Gospodarek 
Signed-off-by: David S. Miller

net: macb: Fix race condition in driver when Rx frame is dropped

2015-04-29T19:24:46+00:00

Under heavy Rx load, observed that the Hw is updating the USED bit
and it is not updating the received frame status to the BD control
field. This could be lack of resources for processing the BDs at high
data rates. Driver drops the frame associated with this BD but not
clearing the USED bit. So, this is causing hang condition as Hw
expects USED bit to be cleared for this BD.

Signed-off-by: Punnaiah Choudary Kalluri 
Signed-off-by: David S. Miller

hv_netvsc: Fix a bug in netvsc_start_xmit()

2015-04-29T19:20:39+00:00

Commit b08cc79155fc26d0d112b1470d1ece5034651a4b eliminated memory
allocation in the packet send path:

    "hv_netvsc: Eliminate memory allocation in the packet send path

    The network protocol used to communicate with the host is the remote ndis (rndis)
    protocol. We need to decorate each outgoing packet with a rndis header and
    additional rndis state (rndis per-packet state). To manage this state, we
    currently allocate memory in the transmit path. Eliminate this allocation by
    requesting additional head room in the skb."

This commit introduced a bug since it did not account for the case if the skb
was cloned. Fix this bug.

Signed-off-by: K. Y. Srinivasan 
Tested-by: Dexuan Cui 
Signed-off-by: David S. Miller