The slow path ulp_start and ulp_stop calls to the bnx2i driver
are sleepable calls and therefore should not be protected using
rcu_read_lock.  Fix it by using mutex and setting a bit during
these calls.  cnic_unregister_device() will now wait for the bit
to clear before completing the call.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Reviewed-by: NBenjamin Li <benli@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The slow path calls to the cnic driver are sleepable calls so we
cannot use rcu_read_lock().  Use mutex for these slow path calls
instead.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Reviewed-by: NBenjamin Li <benli@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Register and unregister with bnx2 during NETDEV_UP and NETDEV_DOWN
events.  This simplifies the sequence of events and allows locking
fixes in the next patch.
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Reviewed-by: NBenjamin Li <benli@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When the cnic driver tries to grab a symbol from bnx2 when bnx2 is
running init code, symbol_get() will succeed but symbol_put_addr()
will hit BUG() a moment later.  module_text_address() fails because
bnx2 is still in init code.

This is fixed by using symbol_put() instead which does the exact
opposite of symbol_get().
Signed-off-by: NMichael Chan <mchan@broadcom.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add new definition to 'pegasus.h' for support Japanese IO DATA
"ETX-US2" USB Ethernet Adapter.

PEGASUS_DEV( $B!H(BIO DATA USB ETX-US2$B!I(B, VENDOR_IODATA, 0x092a,
DEFAULT_GPIO_RESET | PEGASUS_II )
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The NETPOLL API requires that interrupts remain disabled in
netpoll_send_skb(). The use of spin_lock_irq() and spin_unlock_irq()
in the NETPOLL API callbacks causes the interrupts to get enabled and
can lead to kernel instability.

The solution is to use spin_lock_irqsave() and spin_unlock_restore()
to prevent the irqs from getting enabled while in netpoll_send_skb().

Call trace:
netpoll_send_skb()
{
   -> local_irq_save(flags)
     ---> dev->ndo_start_xmit(skb, dev)
         ---> spin_lock_irq()
         ---> spin_unlock_irq() *******here would enable the interrupt.
              ...
   -> local_irq_restore(flags)
}
Signed-off-by: NDongdong Deng <dongdong.deng@windriver.com>
Signed-off-by: NJason Wessel <jason.wessel@windriver.com>
Acked-by: NBruce Ashfield <bruce.ashfield@windriver.com>
Acked-by: NMatt Mackall <mpm@selenic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Test whether VELOCITY_DUPLEX_FULL bit is set in mii_status.
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Change rt2x00_rf_read() and rt2x00_rf_write() to subtract 1 from the rf
register number.  This is needed because the rf registers are enumerated
starting with one.  The size of the rf register cache is just enough to
hold all registers, so writing to the highest register was corrupting
memory.  Add a check to make sure that the rf register number is valid.
Signed-off-by: NPavel Roskin <proski@gnu.org>
Cc: stable@kernel.org
Acked-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Traffic received with a priority tag (VID = 0) and non-zero priority value was
incorrectly handled by the VLAN packet code path due to a check on zero for
the whole VLAN tag instead of just the VID.

This patch masked out the priority field when checking the vlan tag for
received VLAN packets.
Signed-off-by: NLucy Liu <lucy.liu@intel.com>
Acked-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We return the ddp->len in ixgbe_fcoe_ddp() to indicate the length of data that
have been DDPed. However, it is possible that the length is 0, e.g., for SCSI
READ, the FCP_RSP may come back w/ SCSI status 0x28 as Task Set Full with no FCP
data for DDP. In ixgbe_fcoe_ddp(), we return 0 to indicate not passing DDPed
packets to upper layer. Therefore in the case of ddp->len being 0 upon FCP_RSP,
we do not want to return the 0 ddp->len as we want FCP_RSP to be always
delivered to the upper layer. This patch fixes this bug by setting rc only if
ddp->len is non-zero.
Signed-off-by: NYi Zou <yi.zou@intel.com>
Acked-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

o Defer napi resouce allocation to device attach.

o Free napi resources and delete napi during detach.

This ensures right behavior across firmware reset.
Signed-off-by: NDhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

o Remove private workqueue in the driver, move all
  scheduled tasks to keventd workqueues. This makes
  ports (interfaces) of same / different NIC boards
  independent, in terms of their link watchdog and
  reset tasks.

o Move quick checks for link status and temperature
  in timer callback, schedule watchdog task only if
  link status changed or temperature reached critical
  threshold.

This also fixes deadlock when thermal panic occurs,
watchdog work was flushing workqueue that it was
sitting on.
Signed-off-by: NAmit Kumar Salecha <amit@qlogic.com>
Signed-off-by: NDhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently setting rx-usecs when the interface is in legacy interrupt
mode it is not immediate.  We were only setting EITR for each MSIx
vector and since this count would be zero for legacy mode it wasn't
set until after a reset.  This patch corrects that by checking what
mode we are in and then setting EITR accordingly.
Signed-off-by: NDon Skidmore <donald.skidmore@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For 'real' hardware CAN devices the netlink interface is used to set CAN
specific communication parameters. Real CAN hardware can not be created with
the ip tool ...

The invocation of 'ip link add type can' lead to an oops as the standard rtnl
newlink function was called:

http://bugzilla.kernel.org/show_bug.cgi?id=13954

This patch adds a private newlink function for the CAN device driver interface
that unconditionally returns -EOPNOTSUPP.
Signed-off-by: NOliver Hartkopp <oliver@hartkopp.net>
Reported-by: NDmitry Eremin-Solenikov <dbaryshkov@gmail.com>
CC: Patrick McHardy <kaber@trash.net>
CC: Wolfgang Grandegger <wg@grandegger.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The driver always:
1. allocate cp->rx_buf_sz + NET_IP_ALIGN
2. map cp->rx_buf_sz
Signed-off-by: NFrancois Romieu <romieu@fr.zoreil.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This reverts commit 57921c31.

On request from John Linville:

	It has been shown to create a new problem.  There is work
	towards a solution to that one, but it isn't a simple
	clean-up.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

queue == __AR9170_NUM_TXQ would cause a bug on the next line.

found by Smatch ( http://repo.or.cz/w/smatch.git ).

Cc: stable@kernel.org
Reported-by: NDan Carpenter <error27@gmail.com>
Signed-off-by: NDan Carpenter <error27@gmail.com>
Signed-off-by: NChristian Lamparter <chunkeey@web.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When ar9170-2.fw was missing, the driver erroneously complained
about missing the initialization values file ar9170-1.fw...
Signed-off-by: NChristian Lamparter <chunkeey@web.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

If the length is less or equal to frag_prefix_size in the first iteration
we write skb_frags_rx[-1] and read from priv->frag_info[-1]
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Prevent read from cards[-1] when no card was found.
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

An `options[cards_found]' that equals `sizeof(options_mapping)' is already beyond
the array.
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If vlan has been enabled. ifdown followed by ifup will lost hardware
related state.

Also remove duplicated operation in gfar_vlan_rx_register().
Signed-off-by: NYong Zhang <yong.zhang0@gmail.com>
Acked-by: NDai Haruki <dai.haruki@freescale.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The bank offset was being incorrectly calculated on ICH9 parts with a bank
size of 8K (instead of the more common 4K bank) which would cause any NVM
writes to be done on the wrong address after switching from bank 1 to bank
0.  Additionally, assume we are meant to use bank 0 if a valid bank is not
detected, and remove the unnecessary acquisition of the SW/FW/HW semaphore
when writing to the shadow ram version of the NVM image.
Signed-off-by: NBruce Allan <bruce.w.allan@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For ICHx parts, write the EXTCNF_CTRL.SWFLAG bit once when trying to
acquire the SW/FW/HW semaphore instead of multiple times to prevent the
hardware from having problems (especially for systems with manageability
enabled), and extend the timeout for the hardware to set the SWFLAG bit.
Signed-off-by: NBruce Allan <bruce.w.allan@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For 82599, packet split has to be disabled for FCoE direct data placement.
However, this is only required on received queues allocated for FCoE. This
patch adds a per ring flags to indicate if packet split is disabled on a
per queue basis, particularly for FCoE, as packet split must be disabled
for large receive using direct data placement (DDP).
Signed-off-by: NYi Zou <yi.zou@intel.com>
Acked-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Instead of passing the register index of the corresponding rx_ring and find
the way back to get to corresponding rx_ring in ixgbe_configure_srrctl(),
simplify the function ixgbe_configure_srrctl() by passing the rx_ring into
it. Then the register index for that rx_ring is already available from
rx_ring->reg_idx.
Signed-off-by: NYi Zou <yi.zou@intel.com>
Acked-by: NPeter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As it is, parts of the ioctl runs under the RTNL and parts of
it do not.  The unlocked section is still protected by the BKL,
but there can be subtle races.  For example, Eric Biederman and
Paul Moore observed that if two threads tried to create two tun
devices on the same file descriptor, then unexpected results
may occur.

As there isn't anything in the ioctl that is expected to sleep
indefinitely, we can prevent this from occurring by extending
the RTNL lock coverage.

This also allows to get rid of the BKL.

Finally, I changed tun_get_iff to take a tun device in order to
avoid calling tun_put which would dead-lock as it also tries to
take the RTNL lock.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit f0b3fbea ("FEC Buffer rework")
breaks transmission of packets where the skb data buffer is not memory
aligned according to FEC_ALIGNMENT. It incorrectly passes to
dma_sync_single() the buffer address directly from the skb, instead of
the address calculated for use (which may be the skb address or one of
the bounce buffers).

It seems there is no use converting the cpu address of the buffer to
a physical either, since dma_map_single() expects the cpu address and
will return the dma address to use in the descriptor. So remove the use
of __pa() on the buffer address as well.

This patch is against 2.6.30-rc5. This breakage is a regression over
2.6.30, which does not have this problem.
Signed-off-by: NGreg Ungerer <gerg@uclinux.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

strlcpy() will always null terminate the string.
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Cc: Jay Cliburn <jcliburn@gmail.com>
Cc: Chris Snook <csnook@redhat.com>
Cc: Jie Yang <jie.yang@atheros.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

io[i] is read before the bounds check on i, order should be reversed.
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Cc: Samuel Ortiz <samuel@sortiz.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Older Gcc compilers (3.4.5 tested) need additional hints in order to get
the packing of the rxpd structure (which contains a 16 bit union)
correct on the ARM processor.

struct txpd does not need these hints since it contains a 32 bit union
that packs naturally.
Signed-off-by: NR.J.Dunlop <rdunlop@guralp.com>
Acked-by: NDan Williams <dcbw@redhat.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We need to unregister our ieee80211_hw before resetting the chip, as
the former causes firmware commands to be issued which will time out
once the chip has been reset.
Signed-off-by: NLennert Buytenhek <buytenh@marvell.com>
Acked-by: NNicolas Pitre <nico@marvell.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NLennert Buytenhek <buytenh@marvell.com>
Acked-by: NNicolas Pitre <nico@marvell.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NLennert Buytenhek <buytenh@marvell.com>
Acked-by: NNicolas Pitre <nico@marvell.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NLennert Buytenhek <buytenh@marvell.com>
Acked-by: NNicolas Pitre <nico@marvell.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When we go into out-of-memory and fail to allocate skbuffs to
refill the receive ring with, rxq_process can end up running into
a receive ring entry that is marked as host-owned but doesn't have
an associated skbuff.  If this happens, we must break out of the
rx processing loop instead of trying to process the descriptor.
Signed-off-by: NLennert Buytenhek <buytenh@marvell.com>
Acked-by: NNicolas Pitre <nico@marvell.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

> channel_index loops up to IPW_SCAN_CHANNELS, but is used after being
> incremented. This might be able to access 1 past the end of the array
Reported-by: NRoel Kluin <roel.kluin@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We found this old card which was not supported, and physically
looks similar to the other 3C905B we have (9055).

After adding the IDs it seems to work fine (MII report, dhcp, scp, ...)
Acked-by: NSteffen Klassert <klassert@mathematik.tu-chemnitz.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix dma mask calculation that caps at 63-bit addressing even
when firmware advertises full 64-bit support.
Signed-off-by: NDhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch fixes the napi list handling when an ehea interface is shut
down to avoid corruption of the napi list.
Signed-off-by: NHannes Hering <hering2@de.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>