]> git.proxmox.com Git - mirror_ubuntu-kernels.git/log
mirror_ubuntu-kernels.git
10 years agobonding: permit using arp_validate with non-ab modes
Veaceslav Falico [Tue, 18 Feb 2014 06:48:37 +0000 (07:48 +0100)]
bonding: permit using arp_validate with non-ab modes

Currently it's disabled because it's sometimes hard, in typical configs, to
make it work - because of the nature how the loadbalance modes work - as
it's hard to deliver valid arp replies to correct slaves by the switch.

However we still can use arp_validation in loadbalance with several other
configs, per example with arp_validate == 2 for backup with one broadcast
domain, without the switch(es) doing any balancing - this way we'd be (a
bit more) sure that the slave is up.

So, enable it to let users decide which one works/suits them best. Also
correct the mode limitation from BOND_OPT_ARP_VALIDATE.

CC: Nikolay Aleksandrov <nikolay@redhat.com>
CC: Jay Vosburgh <fubar@us.ibm.com>
CC: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Acked-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobonding: remove bond->lock from bond_arp_rcv
Veaceslav Falico [Tue, 18 Feb 2014 06:48:36 +0000 (07:48 +0100)]
bonding: remove bond->lock from bond_arp_rcv

We're always called with rcu_read_lock() held (bond_arp_rcv() is only
called from bond_handle_frame(), which is rx_handler and always called
under rcu from __netif_receive_skb_core() ).

The slave active/passive and/or bonding params can change in-flight, however
we don't really care about that - we only modify the last time packet was
received, which is harmless.

CC: Jay Vosburgh <fubar@us.ibm.com>
CC: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Acked-by: Ding Tianhong <dingtianhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'r8152'
David S. Miller [Tue, 18 Feb 2014 21:40:17 +0000 (16:40 -0500)]
Merge branch 'r8152'

Hayes Wang says:

====================
r8152: improvement and new features

Change some flows or behavior to improve the efficiency or make the
code readable. Besides, support WOL and runtime suspend.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: support get_msglevel and set_msglevel
hayeswang [Tue, 18 Feb 2014 13:49:11 +0000 (21:49 +0800)]
r8152: support get_msglevel and set_msglevel

Support get_msglevel and set_msglevel.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: set disable_hub_initiated_lpm
hayeswang [Tue, 18 Feb 2014 13:49:10 +0000 (21:49 +0800)]
r8152: set disable_hub_initiated_lpm

Set disable_hub_initiated_lpm = 1.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: replace netif_rx with netif_receive_skb
hayeswang [Tue, 18 Feb 2014 13:49:09 +0000 (21:49 +0800)]
r8152: replace netif_rx with netif_receive_skb

Replace netif_rx with netif_receive_skb to avoid disabling irq frequently
for increasing the efficiency.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: disable teredo for RTL8152
hayeswang [Tue, 18 Feb 2014 13:49:08 +0000 (21:49 +0800)]
r8152: disable teredo for RTL8152

Disable teredo for RTL8152 by default.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: support runtime suspend
hayeswang [Tue, 18 Feb 2014 13:49:07 +0000 (21:49 +0800)]
r8152: support runtime suspend

Support runtime suspend for RTL8152 and RTL8153.

Move tx_bottom() from tasklet to delayed_work. That avoids to
transmit tx packets after calling autosuspend.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: support WOL
hayeswang [Tue, 18 Feb 2014 13:49:06 +0000 (21:49 +0800)]
r8152: support WOL

Support WOL for RTL8152 and RTL8153.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: move some functions from probe to open
hayeswang [Tue, 18 Feb 2014 13:49:05 +0000 (21:49 +0800)]
r8152: move some functions from probe to open

Add up method for rtl_ops and asign relative functions. Move
clear_bp() and hw_phy_cfg() from init method to up method of rtl_ops.
Call rtl_ops.up() for ndo_open() and rtl_ops.down for ndo_stop().

Replace allocating the memory in probe() with in ndo_open().

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: combine PHY reset with set_speed
hayeswang [Tue, 18 Feb 2014 13:49:04 +0000 (21:49 +0800)]
r8152: combine PHY reset with set_speed

PHY reset is necessary after some hw settings. However, it would
cause the linking down, and so does the set_speed function. Combine
the PHY reset with set_speed function. That could reduce the frequency
of linking down and accessing the PHY register.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: clear BMCR_PDOWN
hayeswang [Tue, 18 Feb 2014 13:49:03 +0000 (21:49 +0800)]
r8152: clear BMCR_PDOWN

Modify the method of enabling the PHY to clear BMCR_PDOWN only.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: reduce the frequency of spin_lock
hayeswang [Tue, 18 Feb 2014 13:49:02 +0000 (21:49 +0800)]
r8152: reduce the frequency of spin_lock

Replace getting one item from a list with getting the whole list one time.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: load the default MAC address
hayeswang [Tue, 18 Feb 2014 13:49:01 +0000 (21:49 +0800)]
r8152: load the default MAC address

Except for RTL_VER_01, replace loading the MAC address from PLA_IDR
with from PLA_BACKUP. The default MAC address may be modified by
the other OS, so the PLA_IDR may be not the default MAC address.

The data in the PLA_BACKUP address of the RTL_VER_01 may be destoryed,
so load MAC address from PLA_IDR for RTL_VER_01.

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: replace some types from int to bool
hayeswang [Tue, 18 Feb 2014 13:49:00 +0000 (21:49 +0800)]
r8152: replace some types from int to bool

Modify the following functions.
 - r8153_u1u2en
 - r8153_u2p3en
 - r8153_power_cut_en

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: add three functions
hayeswang [Tue, 18 Feb 2014 13:48:59 +0000 (21:48 +0800)]
r8152: add three functions

Replace some codes with the following three functions.
 - rtl_drop_queued_tx
 - rxdy_gated_en
 - r8152_power_cut_en

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agor8152: move some functions
hayeswang [Tue, 18 Feb 2014 13:48:58 +0000 (21:48 +0800)]
r8152: move some functions

Move the following functions which is for the further coding.
 - rtl_clear_bp
 - r8153_clear_bp
 - r8153_teredo_off
 - r8152b_disable_aldps
 - r8152b_enable_aldps
 - r8152b_hw_phy_cfg

Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'cxgb4'
David S. Miller [Tue, 18 Feb 2014 21:23:58 +0000 (16:23 -0500)]
Merge branch 'cxgb4'

Hariprasad Shenai says:

====================
Adds support for Chelsio T5 40G adapter and Misc. fixes

This patch series adds support for Chelsio T5 40G adapters and provides
miscelleneous fixes for cxgb4 driver.

It also adds device ids of two new T5 adapters.

We would like to request this patch series to get merged via David Miller's
'net-next' tree.

We have included all the maintainers of respective drivers. Kindly review the
change and let us know in case of any review comments.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Add more PCI device ids.
Hariprasad Shenai [Tue, 18 Feb 2014 12:26:15 +0000 (17:56 +0530)]
cxgb4: Add more PCI device ids.

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Don't assume LSO only uses SGL path in t4_eth_xmit()
Kumar Sanghvi [Tue, 18 Feb 2014 12:26:14 +0000 (17:56 +0530)]
cxgb4: Don't assume LSO only uses SGL path in t4_eth_xmit()

Also, modify is_eth_imm() to return header length so it doesn't
have to be recomputed in calc_tx_flits().

Based on original work by Mike Werner <werner@chelsio.com>

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Remove unused registers and add missing ones
Kumar Sanghvi [Tue, 18 Feb 2014 12:26:13 +0000 (17:56 +0530)]
cxgb4: Remove unused registers and add missing ones

Remove unused registers for registers list, and add missing ones
Based on original work by Santosh Rastapur <santosh@chelsio.com>

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Query firmware for T5 ULPTX MEMWRITE DSGL capabilities
Kumar Sanghvi [Tue, 18 Feb 2014 12:26:12 +0000 (17:56 +0530)]
cxgb4: Query firmware for T5 ULPTX MEMWRITE DSGL capabilities

Query firmware to see whether we're allowed to use T5 ULPTX MEMWRITE DSGL
capabilities.  Also pass that information to Upper Layer Drivers via the
new (struct cxgb4_lld_info).ulptx_memwrite_dsgl boolean.

Based on original work by Casey Leedom <leedom@chelsio.com>

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: LE-Workaround is not atomic in firmware
Kumar Sanghvi [Tue, 18 Feb 2014 12:26:11 +0000 (17:56 +0530)]
cxgb4: LE-Workaround is not atomic in firmware

The LE workaround in firmware is not atomic and fw_ofld_connection_wrs must not interleave.
Therefore, when the workaround is enabled, we need to send all ctrlq WRs on a single ctrl queue.

Based on original work by Santosh Rastapur <santosh@chelsio.com>

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Allow >10G ports to have multiple queues
Kumar Sanghvi [Tue, 18 Feb 2014 12:26:10 +0000 (17:56 +0530)]
cxgb4: Allow >10G ports to have multiple queues

Based on original work by Divy Le Ray.

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Print adapter VPD Part Number instead of Engineering Change field
Kumar Sanghvi [Tue, 18 Feb 2014 12:26:09 +0000 (17:56 +0530)]
cxgb4: Print adapter VPD Part Number instead of Engineering Change field

When we attach to adapter, print VPD Part Number instead of Engineering Change field.
Based on original work by Casey Leedom <leedom@chelsio.com>

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Add support to recognize 40G links
Kumar Sanghvi [Tue, 18 Feb 2014 12:26:08 +0000 (17:56 +0530)]
cxgb4: Add support to recognize 40G links

Also, create a new Common Code interface to translate Firmware Port Technology
Type values (enum fw_port_type) to string descriptions.  This will allow us
to maintain the description translation table in one place rather than in
every driver.

Based on original work by Scott Bardone and Casey Leedom <leedom@chelsio.com>

Signed-off-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'msix'
David S. Miller [Tue, 18 Feb 2014 20:34:14 +0000 (15:34 -0500)]
Merge branch 'msix'

Alexander Gordeev says:

====================
net: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Cc: e1000-devel@lists.sourceforge.net
Cc: linux-driver@qlogic.com
Cc: linux-net-drivers@solarflare.com
Cc: linux-pci@vger.kernel.org
Cc: linux-rdma@vger.kernel.org
Cc: netdev@vger.kernel.org
Cc: pv-drivers@vmware.com
Cc: wil6210@qca.qualcomm.com
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agowil6210: Use pci_enable_msi_range() instead of pci_enable_msi_block()
Alexander Gordeev [Tue, 18 Feb 2014 10:12:04 +0000 (11:12 +0100)]
wil6210: Use pci_enable_msi_range() instead of pci_enable_msi_block()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Cc: "John W. Linville" <linville@tuxdriver.com>
Cc: wil6210@qca.qualcomm.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Acked-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agovmxnet3: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:12:03 +0000 (11:12 +0100)]
vmxnet3: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Shreyas Bhatewara <sbhatewara@vmware.com>
Cc: pv-drivers@vmware.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agovmxnet3: Fix MSI-X/MSI enablement code
Alexander Gordeev [Tue, 18 Feb 2014 10:12:02 +0000 (11:12 +0100)]
vmxnet3: Fix MSI-X/MSI enablement code

This update cleans up the MSI-X/MSI enablement code, fixes
vmxnet3_acquire_msix_vectors() invalid return values and
enables a dead code in case VMXNET3_LINUX_MIN_MSIX_VECT
MSI-X vectors were allocated.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Shreyas Bhatewara <sbhatewara@vmware.com>
Cc: pv-drivers@vmware.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoniu: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:12:01 +0000 (11:12 +0100)]
niu: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jingoo Han <jg1.han@samsung.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agosfc: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:12:00 +0000 (11:12 +0100)]
sfc: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Shradha Shah <sshah@solarflare.com>
Cc: linux-net-drivers@solarflare.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Acked-by: Shradha Shah <sshah@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoqlge: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:59 +0000 (11:11 +0100)]
qlge: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Shahed Shaikh <shahed.shaikh@qlogic.com>
Cc: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Cc: Ron Mercer <ron.mercer@qlogic.com>
Cc: linux-driver@qlogic.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Acked-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoqlge: Get rid of an redundant assignment
Alexander Gordeev [Tue, 18 Feb 2014 10:11:58 +0000 (11:11 +0100)]
qlge: Get rid of an redundant assignment

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Shahed Shaikh <shahed.shaikh@qlogic.com>
Cc: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Cc: Ron Mercer <ron.mercer@qlogic.com>
Cc: linux-driver@qlogic.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Acked-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoqlcnic: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:57 +0000 (11:11 +0100)]
qlcnic: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Himanshu Madhani <himanshu.madhani@qlogic.com>
Cc: Rajesh Borundia <rajesh.borundia@qlogic.com>
Cc: Shahed Shaikh <shahed.shaikh@qlogic.com>
Cc: linux-driver@qlogic.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoqlcnic: Cleanup qlcnic_enable_msix() return values
Alexander Gordeev [Tue, 18 Feb 2014 10:11:56 +0000 (11:11 +0100)]
qlcnic: Cleanup qlcnic_enable_msix() return values

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Himanshu Madhani <himanshu.madhani@qlogic.com>
Cc: Rajesh Borundia <rajesh.borundia@qlogic.com>
Cc: Shahed Shaikh <shahed.shaikh@qlogic.com>
Cc: linux-driver@qlogic.com
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonetxen: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:55 +0000 (11:11 +0100)]
netxen: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Manish Chopra <manish.chopra@qlogic.com>
Cc: Sony Chacko <sony.chacko@qlogic.com>
Cc: Rajesh Borundia <rajesh.borundia@qlogic.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoforcedeth: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:54 +0000 (11:11 +0100)]
forcedeth: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Patrick McHardy <kaber@trash.net>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoforcedeth: Cleanup MSI-X to MSI to INTx fallback code
Alexander Gordeev [Tue, 18 Feb 2014 10:11:53 +0000 (11:11 +0100)]
forcedeth: Cleanup MSI-X to MSI to INTx fallback code

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Patrick McHardy <kaber@trash.net>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoforcedeth: Fix invalid errno reporting in nv_request_irq()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:52 +0000 (11:11 +0100)]
forcedeth: Fix invalid errno reporting in nv_request_irq()

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Patrick McHardy <kaber@trash.net>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agovxge: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:51 +0000 (11:11 +0100)]
vxge: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jon Mason <jdmason@kudzu.us>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agos2io: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:50 +0000 (11:11 +0100)]
s2io: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jon Mason <jdmason@kudzu.us>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agomyri10ge: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:49 +0000 (11:11 +0100)]
myri10ge: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Hyong-Youb Kim <hykim@myri.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agomlx5: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:48 +0000 (11:11 +0100)]
mlx5: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Eli Cohen <eli@mellanox.com>
Cc: linux-rdma@vger.kernel.org
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agomlx4: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:47 +0000 (11:11 +0100)]
mlx4: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Amir Vadai <amirv@mellanox.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Acked-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoixgbevf: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:46 +0000 (11:11 +0100)]
ixgbevf: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: Bruce Allan <bruce.w.allan@intel.com>
Cc: e1000-devel@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoixgbe: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:45 +0000 (11:11 +0100)]
ixgbe: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: Bruce Allan <bruce.w.allan@intel.com>
Cc: e1000-devel@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoigbvf: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:44 +0000 (11:11 +0100)]
igbvf: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: Bruce Allan <bruce.w.allan@intel.com>
Cc: e1000-devel@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoigb: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:43 +0000 (11:11 +0100)]
igb: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: e1000-devel@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoi40e: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:42 +0000 (11:11 +0100)]
i40e: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: e1000-devel@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoe1000e: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:41 +0000 (11:11 +0100)]
e1000e: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: Bruce Allan <bruce.w.allan@intel.com>
Cc: e1000-devel@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobenet: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:11:40 +0000 (11:11 +0100)]
benet: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Sathya Perla <sathya.perla@emulex.com>
Cc: Subbu Seetharaman <subbu.seetharaman@emulex.com>
Cc: Ajit Khaparde <ajit.khaparde@emulex.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Acked-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoenic: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:08:02 +0000 (11:08 +0100)]
enic: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Christian Benvenuti <benve@cisco.com>
Cc: Sujith Sankar <ssujith@cisco.com>
Cc: Govindarajulu Varadarajan <govindarajulu90@gmail.com>
Cc: Neel Patel <neepatel@cisco.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Acked-by: Govindarajulu Varadarajan <govindarajulu90@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4vf: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:08:01 +0000 (11:08 +0100)]
cxgb4vf: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Casey Leedom <leedom@chelsio.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4vf: Remove superfluous call to pci_disable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:08:00 +0000 (11:08 +0100)]
cxgb4vf: Remove superfluous call to pci_disable_msix()

There is no need to call pci_disable_msix() in case
the previous call to pci_enable_msix() failed

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Casey Leedom <leedom@chelsio.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb4: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:07:59 +0000 (11:07 +0100)]
cxgb4: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Dimitris Michailidis <dm@chelsio.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb3: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:07:58 +0000 (11:07 +0100)]
cxgb3: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Santosh Raspatur <santosh@chelsio.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocxgb3: Remove superfluous call to pci_disable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:07:57 +0000 (11:07 +0100)]
cxgb3: Remove superfluous call to pci_disable_msix()

There is no need to call pci_disable_msix() in case
the previous call to pci_enable_msix() failed

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Santosh Raspatur <santosh@chelsio.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobna: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:07:56 +0000 (11:07 +0100)]
bna: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Rasesh Mody <rmody@brocade.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agotg3: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:07:55 +0000 (11:07 +0100)]
tg3: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Nithin Nayak Sujir <nsujir@broadcom.com>
Cc: Michael Chan <mchan@broadcom.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobnx2x: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:07:54 +0000 (11:07 +0100)]
bnx2x: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Ariel Elior <ariele@broadcom.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobnx2: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Tue, 18 Feb 2014 10:07:53 +0000 (11:07 +0100)]
bnx2: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Michael Chan <mchan@broadcom.com>
Cc: netdev@vger.kernel.org
Cc: linux-pci@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'gianfar'
David S. Miller [Tue, 18 Feb 2014 20:32:51 +0000 (15:32 -0500)]
Merge branch 'gianfar'

Claudiu Manoil says:

====================
gianfar: Device configuration fixes

This patchset represents the first part of an effort
to solve some old device configuration issues in gianfar,
especially run-time reset and re-configuration problems.
I'm referring to "on-the-fly" configuration of registers
against HW specification, concurrency issues during device
reset / re-configuration operations, and implementing HW
advisories for these operations.

There's also a good deal of code cleanup and refactoring,
and some other (minor) fixes as well.

v2: Remove sysfs stubs w/o replacing them with module
    params (patch 2).
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogianfar: Remove clean_rx_ring race from gfar_ethtool
Claudiu Manoil [Mon, 17 Feb 2014 10:53:19 +0000 (12:53 +0200)]
gianfar: Remove clean_rx_ring race from gfar_ethtool

gfar_clean_rx_ring() was designed to be called from napi
(rx softirq) context to do the Rx processing. Calling it
from a process context like this is a bug as it will
clearly race with the napi Rx processing.

There's also no point in initializing num_txbdfree since
startup_gfar() already does that, when bringing the device
up again (after reset). Changing num_txbdfree "on-the-fly"
like this is also subject to race conditions.  num_txbdfree
is handled by the Tx processing path and the device reset
procedure.  Also, don't assume that num_rx_queues is always
equal to num_tx_queues.

Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogianfar: Add missing graceful reset steps and fixes
Claudiu Manoil [Mon, 17 Feb 2014 10:53:18 +0000 (12:53 +0200)]
gianfar: Add missing graceful reset steps and fixes

gfar_halt() and gfar_start() are responsible for stopping
and starting the DMA and the Rx/Tx hw rings. They implement
the support for the "graceful Rx/Tx stop/start" hw procedure,
and also disable/enable eTSEC's hw interrupts in the process.

The GRS/GTS procedure requires however to have the RQUEUE/TQUEUE
registers cleared first and to wait for a period of time for the
current frame to pass through the interface (around ~10ms for a
jumbo frame). Only then may the GTS and GRS bits from DMACTRL be
set to shut down the DMA, and finally the Tx_EN and Rx_EN bits in
MACCFG1 may be cleared to disable the Tx/Rx blocks.

The same register programming order applies to start the Rx/Tx:
enabling the RQUEUE/TQUEUE *before* clearing the GRS/GTS bits.

This is a HW recommendation in order to avoid a possible
controller "lock up" during graceful reset.

Cleanup the gfar_halt()/start() prototypes, to take priv instead
of ndev as their purpose is to operate on HW. Enabling the
RQUEUE/TQUEUE in the hw_init() is not needed anymore since
that's the job of gfar_start().

Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogianfar: Factor out enabling/disabling of hw interrupts
Claudiu Manoil [Mon, 17 Feb 2014 10:53:17 +0000 (12:53 +0200)]
gianfar: Factor out enabling/disabling of hw interrupts

Throughout the code there are places where the controller's
hw interrupt sources need to get disabled/enabled (masked/
un-masked) all at once.  The recommendation for disabling
the interrupts is to clear the ievent first then the imask
register (not the other way around).
Use the gfar_ints_enable/disable() helpers to make these
operations consistent.

Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogianfar: Remove useless HAS_PADDING device flag
Claudiu Manoil [Mon, 17 Feb 2014 10:53:16 +0000 (12:53 +0200)]
gianfar: Remove useless HAS_PADDING device flag

The RCTRL updates of the FSL_GIANFAR_DEV_HAS_PADDING device
flag get overriden by the FSL_GIANFAR_DEV_HAS_TIMER flag
settings, which impose a Rx padding alignment of 8 bytes.
As all the eTSEC devices that set HAS_PADDING also set the
HAS_TIMER flag, the HAS_PADDING flag is now obsolete.

Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogianfar: Remove sysfs stubs for FIFOCFG and stashing
Claudiu Manoil [Mon, 17 Feb 2014 10:53:15 +0000 (12:53 +0200)]
gianfar: Remove sysfs stubs for FIFOCFG and stashing

Removing the sysfs stubs for the Tx FIFOCFG and ATTRELI
(stashing) config registers, as these registers may only
be configured after a MAC reset, with the controller stopped
(i.e. during hw init, at probe() time).  The current sysfs
stubs allow on-the-fly updates of these registers (the locking
measures are useless and only add unecessary code).

Changing these registers is discouraged. Only the default values
will be used instead.

Moreover, the stashing (ATTRELI) configuration options were
effectively disabled (didn't get to the hw anyway if changed)
because the stashing device_flags (HAS_BD_STASHING|HAS_BUF_STASHING)
were "accidentally" cleared during probe().

Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agogianfar: Cleanup/Fix gfar_probe and the hw init code
Claudiu Manoil [Mon, 17 Feb 2014 10:53:14 +0000 (12:53 +0200)]
gianfar: Cleanup/Fix gfar_probe and the hw init code

Factor out gfar_hw_init() to contain all the controller hw
initialization steps for a better control of register writes,
and to significantly simplify the tangled code from gfar_probe().
This results in code size and stack usage reduction (besides
code readability).

Fix memory leak on device removal, by freeing the rx_/tx_queue
structures.

Replace custom bit swapping function with a library one (bitrev8).

Move allocation of rx_/tx_queue struct arrays before the group
structure init, because in order to assign Rx/Tx queues
to groups we need to have the queues first.  This also allows
earlier bail out of gfar_probe(), in case the memory allocation
fails.

The flow control checks for maccfg1 were removed from gfar_probe(),
since flow control is disabled at probe time (priv->rx_/tx_pause_en
are 0). Redundant initializations (by 0) also removed.

Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: i40evf: Remove duplicate include
Sachin Kamat [Mon, 17 Feb 2014 09:59:21 +0000 (15:29 +0530)]
net: i40evf: Remove duplicate include

linux/sctp.h was included twice.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoipv6:fix checkpatch errors with assignment in if condition
Wang Yufen [Mon, 17 Feb 2014 08:53:35 +0000 (16:53 +0800)]
ipv6:fix checkpatch errors with assignment in if condition

Signed-off-by: Wang Yufen <wangyufen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoipv6: fix checkpatch errors with space required or prohibited
Wang Yufen [Mon, 17 Feb 2014 08:53:34 +0000 (16:53 +0800)]
ipv6: fix checkpatch errors with space required or prohibited

Signed-off-by: Wang Yufen <wangyufen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoipv6: fix checkpatch errors with brace and "foo *bar"
Wang Yufen [Mon, 17 Feb 2014 08:53:33 +0000 (16:53 +0800)]
ipv6: fix checkpatch errors with brace and "foo *bar"

Signed-off-by: Wang Yufen <wangyufen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonetlink: fix checkpatch errors space and "foo *bar"
Wang Yufen [Mon, 17 Feb 2014 08:53:32 +0000 (16:53 +0800)]
netlink: fix checkpatch errors space and "foo *bar"

ERROR: spaces required and "(foo*)" should be "(foo *)"

Signed-off-by: Wang Yufen <wangyufen@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agosch_netem: replace magic numbers with enumerate in get_loss_clg
Yang Yingliang [Mon, 17 Feb 2014 08:48:21 +0000 (16:48 +0800)]
sch_netem: replace magic numbers with enumerate in get_loss_clg

Replace two magic numbers which intialize clgstate::state.

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'ieee802154'
David S. Miller [Mon, 17 Feb 2014 21:43:17 +0000 (16:43 -0500)]
Merge branch 'ieee802154'

Phoebe Buckheister says:

====================
ieee802154: support rf212 and extended mac features

this patch set adds support for the RF212 radio chip to the existing
at86rf230 driver and adds support for numerous features of the RF212
chips to the ieee802154 stack. These features include CSMA parameter
configuration, transmit power control, CCA parameter configuration, and
automatic retransmission of frames. Netlink APIs are provided for all
new options introduced in this set.

Many features might also work for RF230, but since I have no such chips
at my disposal, most new features are implemented only for RF212.

Changes since v2:
 * Indentation
Changes since v1:
 * CodingStyle compliance. Thanks Sergei Shtylyov
 * Add CSMA parameters to netlink phy list that were forgotten in v1
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: enable smart transmitter features of RF212
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:15 +0000 (11:34 +0100)]
ieee802154: enable smart transmitter features of RF212

This patch does multiple things:

 * add .set_csma_params and .set_frame_retries for the RF212 radio.
   This should work fine with RF230, but since I have no RF230 radios to
   test with, RF230 does not implement these right now
 * enable TX_ARET for frame retransmission limits greater than -1

Since RF230 has no operations to change CSMA parameters or frame retry
limits, RF230 will not be able to enter TX_ARET with this patch.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: add netlink APIs for smartMAC configuration
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:14 +0000 (11:34 +0100)]
ieee802154: add netlink APIs for smartMAC configuration

Introduce new netlink attributes for SET_PHY_ATTRS:
 * CSMA minimal backoff exponent
 * CSMA maximal backoff exponent
 * CSMA retry limit
 * frame retransmission limit

The CSMA attributes shall correspond to minBE, maxBE and maxCSMABackoffs of
802.15.4, respectively. The frame retransmission shall correspond to
maxFrameRetries of 802.15.4, unless given as -1: then the old behaviour
of the stack shall apply. For RF2xy, the old behaviour is to not do
channel sensing at all and simply send *right now*, which is not
intended behaviour for most applications and actually prohibited for
some channel/page combinations.

For all values except frame retransmission limit, the defaults of
802.15.4 apply. Frame retransmission limits are set to -1 to indicate
backward-compatible behaviour.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: ensure that first RF212 state comes from TRX_OFF
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:13 +0000 (11:34 +0100)]
ieee802154: ensure that first RF212 state comes from TRX_OFF

A documented erratum of the RF212 chip describes that some versions of
RF212 may not be properly reset by the standard reset procedure. The
described workaround seems to not fully work either; my RF212 chips will
not correctly receive any frames unless they are taken from RX mode to
TRX_OFF and back to RX mode, effectively forcing a up-down-up-cycle.
Going to TRX_OFF and to RX without intermediate state changes at least
once fixes this.

The same transition is allowed for RF230, so simply do it all the time.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: add support for setting CCA energy detection levels
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:12 +0000 (11:34 +0100)]
ieee802154: add support for setting CCA energy detection levels

Since three of the four clear channel assesment modes make use of energy
detection, provide an API to set the energy detection threshold.
Driver support for this is available in at86rf230 for the RF212 chips.
Since for these chips the minimal energy detection threshold depends on
page and channel used, add a field to struct at86rf230_local that stores
the minimal threshold. Actual ED thresholds are configured as offsets
from this value.

For RF212, setting the ED threshold will not work before a channel/page
has been set due to the dependency of energy detection in the chip and
the actual channel/page selected.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: add support for CCA mode in wpan phys
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:11 +0000 (11:34 +0100)]
ieee802154: add support for CCA mode in wpan phys

The standard describes four modes of clear channel assesment: "energy
above threshold", "carrier found", and the logical and/or of these two.
Support for CCA mode setting is included in the at86rf230 driver,
predicated for RF212 chips.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: add support for listen-before-talk in wpan_phy
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:10 +0000 (11:34 +0100)]
ieee802154: add support for listen-before-talk in wpan_phy

Listen-before-talk is an alternative to CSMA in uncoordinated networks
and prescribed by european regulations if one wants to have a device
with radio duty cycles above 10% (or less in some bands). Add a phy
property to enable/disable LBT in the phy, including support in the
at86rf230 driver for RF212 chips.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: support 100kbps QPSK/EU in at86rf230
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:09 +0000 (11:34 +0100)]
ieee802154: support 100kbps QPSK/EU in at86rf230

The standard assigns channel 0 on page 2 to be 100kbps QPSK in the
868.3MHz band. Add support to the at86rf230 driver for this channel and
page, at the moment predicated only for the RF212 chip.

Per the datasheet, configurations for page 0, channels 0 to 10 and page
2, channels 0 to 10 differ only in the BPSK_QPSK bit. Support for
channels 1 to 10 is untested.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: add TX power control to wpan_phy
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:08 +0000 (11:34 +0100)]
ieee802154: add TX power control to wpan_phy

Replace the current u8 transmit_power in wpan_phy with s8 transmit_power.
The u8 field contained the actual tx power and a tolerance field,
which no physical radio every used. Adjust sysfs entries to keep
compatibility with userspace, give tolerances of +-1dB statically there.

This patch only adds support for this in the at86rf230 driver and the
RF212 chip. Configuration calculation for RF212 is also somewhat basic,
but does the job - the RF212 datasheet gives a large table with
suggested values for combinations of TX power and page/channel, if this
does not work well, we might have to copy the whole table.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: default to AACK in at86rf230
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:07 +0000 (11:34 +0100)]
ieee802154: default to AACK in at86rf230

The current IEEE802.15.4 stack assumes that a radio will never deliver
packets with a bad CRC into the stack, as required by the standard.
at86rf230 driven radios violates this assumption because of another
incompatibility: devices are required to send ACKs if requested by a
sender, but RF2xx will only send ACKs from a special receive mode that
is currently not used by the driver.

Enable this receive mode to fix both bugs. Frames with bad CRCs will not
be received at all, and ACKs will be sent if so requested.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoieee802154: add basic support for RF212 to at86rf230 driver
Phoebe Buckheister [Mon, 17 Feb 2014 10:34:06 +0000 (11:34 +0100)]
ieee802154: add basic support for RF212 to at86rf230 driver

Since the AT86RF2xy chips are mostly compatible, this is only a small
change to the actual driver code. The at86rf230 driver already supports the
RF212 in most places, only three small adjustments are required:

 * force the initial state after P_ON to FORCE_TRX_OFF to work around a
   documented erratum
 * channels_supported depends on the frequency of the transceiver, and
   thus is_rf212
 * do early detection of chip version select an appropriate _ops struct
   based on the chip version.

Signed-off-by: Phoebe Buckheister <phoebe.buckheister@itwm.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoMerge branch 'phy'
David S. Miller [Mon, 17 Feb 2014 21:40:24 +0000 (16:40 -0500)]
Merge branch 'phy'

Florian Fainelli says:

====================
net: phy: soft reset rework for 10G PHYs

As reported by Shaohui, 10G PHYs may have a slightly more complex reset
sequence for which a BMCR_RESET software reset might not suffice. This
patchset offers a solution for those by allowing them to implement their
own soft_reset() callback. Finally there is an update to the PHY library
Documentation to cover for the newly added callbacks of the PHY driver
structure.

Changes in v2:
- fixed callback testing
- fixed typo in Documentation
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoDocumentation: networking: update phy.txt with recent changes
Florian Fainelli [Mon, 17 Feb 2014 21:34:04 +0000 (13:34 -0800)]
Documentation: networking: update phy.txt with recent changes

The PHY library was missing a bunch of newly added PHY driver callbacks
along with a smallish description of what they do, fix that.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: phy: allow PHY drivers to implement their own software reset
Florian Fainelli [Mon, 17 Feb 2014 21:34:03 +0000 (13:34 -0800)]
net: phy: allow PHY drivers to implement their own software reset

As pointed out by Shaohui, most 10G PHYs out there have a non-standard
compliant software reset sequence, eventually something much more
complex than just toggling the BMCR_RESET bit. Allow PHY driver to
implement their own soft_reset() callback to deal with that. If no
callback is provided, call into genphy_soft_reset() which makes sure the
existing behavior is kept intact.

Reported-by: Shaohui Xie <Shaohui.Xie@freescale.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet: phy: move PHY software reset to genphy_soft_reset
Florian Fainelli [Mon, 17 Feb 2014 21:34:02 +0000 (13:34 -0800)]
net: phy: move PHY software reset to genphy_soft_reset

As pointed out by Shaohui, this function is generic for 10/100/1000
PHYs, but 10G PHYs might have a slightly different reset sequence which
prevents most of them from using this function.

Move the BMCR_RESET based software resent sequence to
genphy_soft_reset() in preparation for allowing PHY drivers to implement
a soft_reset() callback.

Reported-by: Shaohui Xie <Shaohui.Xie@freescale.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoDrivers: net: hyperv: Cleanup the netvsc receive callback functio
KY Srinivasan [Mon, 17 Feb 2014 00:38:45 +0000 (16:38 -0800)]
Drivers: net: hyperv: Cleanup the netvsc receive callback functio

Get rid of the buffer allocation in the receive path for normal packets.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoDrivers: net: hyperv: Cleanup the receive path
KY Srinivasan [Mon, 17 Feb 2014 00:38:44 +0000 (16:38 -0800)]
Drivers: net: hyperv: Cleanup the receive path

Make the receive path a little more efficient by parameterizing the
required state rather than re-establishing that state.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoDrivers: net: hyperv: Get rid of the rndis_filter_packet structure
KY Srinivasan [Mon, 17 Feb 2014 00:38:43 +0000 (16:38 -0800)]
Drivers: net: hyperv: Get rid of the rndis_filter_packet structure

This structure is redundant; get rid of it make the code little more efficient -
get rid of the unnecessary indirection.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agonet:cpsw: Pass unhandled ioctl's on to generic phy ioctl
Stefan Sørensen [Sun, 16 Feb 2014 13:54:25 +0000 (14:54 +0100)]
net:cpsw: Pass unhandled ioctl's on to generic phy ioctl

This patch allows the use of a generic timestamping phy connected
to the cpsw if CPTS support is not enabled. This also adds support
of the SIOCGMIIREG and SIOCSMIIREG, and moves handling of SIOCGMIIPHY
to the generic driver.

Signed-off-by: Stefan Sørensen <stefan.sorensen@spectralink.com>
Reviewed-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobonding: Convert memcpy(foo, bar, ETH_ALEN) to ether_addr_copy(foo, bar)
Joe Perches [Sun, 16 Feb 2014 00:02:17 +0000 (16:02 -0800)]
bonding: Convert memcpy(foo, bar, ETH_ALEN) to ether_addr_copy(foo, bar)

ether_addr_copy is smaller and faster for some architectures.

This relies on a stack frame being at least __aligned(2)
for one use of an Ethernet address on the stack.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobonding: Convert c99 comments
Joe Perches [Sun, 16 Feb 2014 00:02:03 +0000 (16:02 -0800)]
bonding: Convert c99 comments

Neatening only.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobonding: Neaten pr_<level>
Joe Perches [Sun, 16 Feb 2014 00:01:45 +0000 (16:01 -0800)]
bonding: Neaten pr_<level>

Add missing terminating newlines.
Convert uses of pr_info to pr_cont in bond_check_params.
Standardize upper/lower case styles.
Typo fixes, remove unnecessary parentheses and periods.
Alignment neatening.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agobonding: Convert pr_warning to pr_warn, neatening
Joe Perches [Sat, 15 Feb 2014 23:57:04 +0000 (15:57 -0800)]
bonding: Convert pr_warning to pr_warn, neatening

Use more current logging style.

Coalesce formats, realign arguments, drop unnecessary periods.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agocaif: delete unnecessary field initialization
Julia Lawall [Sat, 15 Feb 2014 07:36:13 +0000 (08:36 +0100)]
caif: delete unnecessary field initialization

On success, the function netdev_alloc_skb initializes the dev field of its
result to its first argument, so this doesn't have to be done in the
calling context.

The semantic patch that fixes this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression skb,privn,e;
@@

skb = netdev_alloc_skb(privn,...);
... when strict
(
-skb->dev = privn;
|
?skb = e
)
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
10 years agoDocumentation: broadcom-bcmgenet: add better clocks documentation
Florian Fainelli [Fri, 14 Feb 2014 23:21:42 +0000 (15:21 -0800)]
Documentation: broadcom-bcmgenet: add better clocks documentation

Document the required "clocks" phandles and their corresponding
"clock-names" properties for the two clocks used by the GENET hardware
block ("enet" and "enet-wol").

CC: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>