As the sub queue id in the broadcast forwarding table is always
set to absolute queue 0 rather than the interface's relative queue 0,
this will cause the received broadcast packets loopback to rcb.
This patch sets the sub queue id to relative queue 0 of each port.
Kejian Yan [Thu, 29 Sep 2016 17:09:13 +0000 (18:09 +0100)]
net: hns: fix port not available after testing loopback
After running command "ethtool -t eth0", eth0 can not be connected to
network. It is caused by the changing the inner loopback register and
this register cannot be changed when hns connected to network. The
routine of setting this register needs to be removed and using promisc
mode to let the packet looped back pass by dsaf mode.
Reported-by: Jun He <hjat2005@huawei.com> Signed-off-by: Kejian Yan <yankejian@huawei.com> Reviewed-by: Yisen Zhaung <yisen.zhuang@huawei.com> Signed-off-by: Salil Mehta <salil.mehta@huawei.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
Because we handle the received packets after napi, so delete the checking
before submitting. It delete the code of read the fbd number register,
which reduces the cpu usages while receiving packets
This patch adds fini_process for v2, it handles the packets recevied
by the hardware in the napi porcess. With this patch, the hardware irq
numbers will drop 50% per sec.
net: hns: bug fix about setting coalsecs-usecs to 0
When set rx/tx coalesce usecs to 0, the interrupt coalesce will be
disabled, but there is a interrupt rate limit which set to 1us, it
will cause no interrupt occurs. This patch disable interrupt limit
when sets coalsecs usecs to 0, and restores it to 1 in other case.
Salil [Tue, 23 Aug 2016 20:44:50 +0000 (04:44 +0800)]
IB/hns: Add support of ACPI to the Hisilicon RoCE driver
This patch is meant to add support of ACPI to the Hisilicon RoCE
driver.
Changes done are primarily meant to detect the type and then either
use DT specific or ACPI spcific functions. Where ever possible,
this patch tries to make use of Unified Device Property Interface
APIs to support both DT and ACPI through single interface.
This patch depends upon HNS ethernet driver to Reset RoCE. This
function within HNS ethernet driver has also been enhanced to
support ACPI and is part of other accompanying patch with this
patch-set.
NOTE: The changes in this patch are done over below branch,
https://github.com/dledford/linux/tree/hns-roce
Salil [Tue, 23 Aug 2016 20:44:49 +0000 (04:44 +0800)]
net: hns: Add support of ACPI to HNS driver RoCE Reset function
In the Hip06 SoC, the RoCE Engine is part of the Hisilicon Network
Subsystem and is dependent upon DSAF module. Therefore, certain
functions like RESET are exposed through the common registers of
HNS DSAF module which are memory-mapped by the HNS driver and
currently can only be accessed through DT/syscon interface.
This patch adds the support of ACPI to the existing RoCE reset
function in the HNS driver(please refer NOTE 2). Hisilicon RoCE
driver (please refer NOTE 1) shall call this reset function during
probe time to reset the RoCE Engine.
The HNS Reset function indirectly ends up in calling the _DSM()
function part of the DSDT ACPI Table. Actual reset functionality
for ACPI is implemented within the ACPI DSDT Table which also has
been enhanced to support this change.
Support of ACPI in the HNS RoCE driver shall be pushed through a
different accompanying below patch:
"IB/hns: Add support of ACPI to the Hisilicon RoCE Driver"
NOTE 1: HNS RoCE driver has already been accepted by its maintainer
Doug Ledford<dledford@redhat.com>. Please refer below link:
https://www.spinics.net/lists/linux-rdma/msg38850.html
NOTE 2: RoCE reset function patch has been accepted and now is
part of the net-next:
https://www.mail-archive.com/netdev@vger.kernel.org/msg123867.html
Wei Yongjun [Sat, 6 Aug 2016 01:01:24 +0000 (01:01 +0000)]
IB/hns: Fix return value check in hns_roce_get_cfg()
In case of error, the function devm_ioremap_resource() returns ERR_PTR()
and never returns NULL. The NULL test in the return value check should
be replaced with IS_ERR().
Philippe Reynes [Fri, 19 Aug 2016 22:52:19 +0000 (00:52 +0200)]
net: ethernet: renesas: ravb: use new api ethtool_{get|set}_link_ksettings
The ethtool api {get|set}_settings is deprecated.
We move this driver to new api {get|set}_link_ksettings.
Signed-off-by: Philippe Reynes <tremyfr@gmail.com> Acked-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Philippe Reynes [Fri, 19 Aug 2016 22:52:18 +0000 (00:52 +0200)]
net: ethernet: renesas: ravb: use phydev from struct net_device
The private structure contain a pointer to phydev, but the structure
net_device already contain such pointer. So we can remove the pointer
phy_dev in the private structure, and update the driver to use the
one contained in struct net_device.
Signed-off-by: Philippe Reynes <tremyfr@gmail.com> Acked-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 21 Aug 2016 04:58:49 +0000 (21:58 -0700)]
Merge branch '10GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue
Jeff Kirsher says:
====================
10GbE Intel Wired LAN Driver Updates 2016-08-20
This series contains updates to ixgbe and ixgbevf.
Veola fixes how the backplane reports the media in ethtool, as KR, KX or
KX4 based on the backplane interface present.
Emil fixes ixgbevf since an incorrect size parameter for
ixgbevf_write_msg_read_ack() ended up only giving the PF the first 4
bytes of the MAC address, so correct the size by calculating it on the
fly for all instances where we call ixgbevf_write_msg_read_ack(). Added
geneve receive offload support for x550em_a.
Don fixes the LED interface for x557 since it uses a different interface.
Added support for the new x557 copper device.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Don Skidmore [Thu, 18 Aug 2016 00:34:40 +0000 (20:34 -0400)]
ixgbe: Add support for new X557 device
This patch adds support for the new copper device X557.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Don Skidmore [Wed, 17 Aug 2016 21:34:07 +0000 (17:34 -0400)]
ixgbe: add device to MDIO speed setting
This shouldn't matter as nothing should be attached still to be
consisted control MDIO speed for these devices as well.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Don Skidmore [Wed, 17 Aug 2016 18:11:57 +0000 (14:11 -0400)]
ixgbe: Fix led interface for X557 devices
The X557 devices use a different interface to the LED for the port.
This patch reflect that change.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Emil Tantilov [Wed, 10 Aug 2016 18:19:23 +0000 (11:19 -0700)]
ixgbe: add support for geneve Rx offload
Add geneve Rx offload support for x550em_a.
The implementation follows the vxlan code with the lower 16 bits of
the VXLANCTRL register holding the UDP port for VXLAN and the upper
for Geneve.
Disabled NFS filters in the RFCTL register which allows us to simplify
the check for VXLAN and Geneve packets in ixgbe_rx_checksum().
Removed vxlan from the name of the callback functions and replaced it
with udp_tunnel which is more in line with the new API.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Emil Tantilov [Wed, 27 Jul 2016 17:55:08 +0000 (10:55 -0700)]
ixgbevf: fix incorrect MAC address on load
The PF driver was only receiving the first 4 bytes of the MAC due
to an incorrect size parameter for ixgbevf_write_msg_read_ack()
in ixgbevf_set_rar_vf().
Correct the size by calculating it on a fly for all instances where
we call ixgbevf_write_msg_read_ack()
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Veola Nazareth [Sun, 21 Aug 2016 02:35:37 +0000 (19:35 -0700)]
ixgbe: report correct media type for KR, KX and KX4 interfaces
ethtool reports backplane type interfaces as 1000/10000baseT link modes.
This has been corrected to report the media as KR, KX or KX4 based on the backplane interface present.
Signed-off-by: Veola Nazareth <veola.nazareth@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Markus Elfring [Sat, 20 Aug 2016 06:54:15 +0000 (08:54 +0200)]
tun: Use memdup_user() rather than duplicating its implementation
Reuse existing functionality from memdup_user() instead of keeping
duplicate source code.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net> Reviewed-by: Shmulik Ladkani <shmulik.ladkani@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 21 Aug 2016 01:56:30 +0000 (18:56 -0700)]
Merge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue
Jeff Kirsher says:
====================
40GbE Intel Wired LAN Driver Updates 2016-08-19
This series contains updates to i40e and i40evf only.
Jake provides several patches, first just moves a function to co-locate
the two functions related to configuring RSS via the admin queue, which
should help in spotting bugs when comparing the two functions. Fixed
an issue where commit e69ff813af35 ("i40e: rework the functions to
configure RSS with similar parameters") missed checking whether the seed
is NULL before using it and did not use the passed in *lut parameter.
Fixed an issue where a previous refactor missed i40e_vsi_config_rss()
and the values were being ignored, so checked for the fields and used
them instead of default values. Lastly replaced calls to
create_singlethread_workqueue() with alloc_workqueue() to provide more
control over workqueue creation and allows explicit setting of the
desired mode of operation.
Mitch adds link speed to log messages and reports speed through ethtool.
Carolyn refactors tail bump check and fixes byte ordering problems found
when enabling this feature support. Adds support for HMC resources and
profile commands for x722 firmware.
Heinrich Schuchardt fixes format identifiers from %u to %d since the
variable is defined as an integer.
Catherine fixes an issue where there was a race condition between the
completion of the client open and calls to the client ops, so ensured
that client ops are not called until we are sure client is open.
Harshitha makes sure that i40e_client_release() does not try to use
an adapter pointer which may not be initialized, so make sure it is.
Joe Perches fixes the use of the local macro XSTRINGIFY() to use
__stringify() instead.
Avinash corrects the mutex usage in client_subtask(). Fixed the RDMA
client to open again after reset since it is closed during a PF reset.
Jeff (me) clean up whitespace issues, where indentation was done
inconsistently and with spaces versus tabs.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
William Tu [Fri, 19 Aug 2016 18:55:44 +0000 (11:55 -0700)]
samples/bpf: Add tunnel set/get tests.
The patch creates sample code exercising bpf_skb_{set,get}_tunnel_key,
and bpf_skb_{set,get}_tunnel_opt for GRE, VXLAN, and GENEVE. A native
tunnel device is created in a namespace to interact with a lwtunnel
device out of the namespace, with metadata enabled. The bpf_skb_set_*
program is attached to tc egress and bpf_skb_get_* is attached to egress
qdisc. A ping between two tunnels is used to verify correctness and
the result of bpf_skb_get_* printed by bpf_trace_printk.
Signed-off-by: William Tu <u9012063@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Haiyang Zhang [Fri, 19 Aug 2016 21:47:09 +0000 (14:47 -0700)]
hv_netvsc: Implement batching of receive completions
The existing code uses busy retry when unable to send out receive
completions due to full ring buffer. It also gives up retrying after limit
is reached, and causes receive buffer slots not being recycled.
This patch implements batching of receive completions. It also prevents
dropping receive completions due to full ring buffer.
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com> Reviewed-by: Stephen Hemminger <sthemmin@microsoft.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jeff Kirsher [Sat, 20 Aug 2016 04:47:41 +0000 (21:47 -0700)]
i40e/i40evf: Fix indentation
Several defines and code comments were indented with spaces instead
of tabs, correct the issue to make indentation consistent.
Change-ID: I0dc6bbb990ec4a9e856acc9ec526d876181f092c Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Avinash Dayanand [Wed, 17 Aug 2016 23:04:06 +0000 (16:04 -0700)]
i40e: Correcting mutex usage in client code
Correcting the mutex usage, in client_subtask(), mutex_unlock has
to be called just before client_del_instance() since this function opens
and later closes the same mutex again.
Similarly in client_is_registered removing the mutex since it closes
the mutex twice.
This is a patch suggested by RDMA team.
Change-ID: Icce519c266e4221b8a2a72a15ba5bf01750e5852 Signed-off-by: Avinash Dayanand <avinash.dayanand@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Joe Perches [Wed, 17 Aug 2016 10:37:31 +0000 (03:37 -0700)]
i40e: Remove XSTRINGIFY macro definitions and uses
Use __stringify instead.
Signed-off-by: Joe Perches <joe@perches.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
i40e: Initialize pointer in client_release function
The function i40e_client_release has a print statement that uses an
adapter pointer which is not initialized if a previous if statement
is not true. Hence, intialize it in the right place.
Change-ID: I1afdaa2c46771ac42be56edcc41bb56b455b06c8 Signed-off-by: Harshitha Ramamurthy <harshitha.ramamurthy@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
i40e: Check client is open before calling client ops
We were having a race between the completion of the client open and
calls to the client ops so don't call a client op unless we are sure the
client is open.
Testing Hints: Load IWARP driver and make sure it works as expected.
Change-Id: I741f4f2aa4fcbfdad3e40dabbbb1b005856c396b Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
i is defined as int but output as %u several times.
Adjust the format identifiers.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Carolyn Wyborny [Thu, 4 Aug 2016 18:37:04 +0000 (11:37 -0700)]
i40e: Fix byte ordering in ARP NS code for X722
This patch fixes byte ordering problems found when enabling this feature
support. Without this patch, the feature will not work correctly. This
patch fixes the definitions to have the correct byte order.
Change-ID: Ic7489fbcbe2195df7be62ff5e359201b827cefe6 Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Mitch Williams [Thu, 4 Aug 2016 18:37:02 +0000 (11:37 -0700)]
i40evf: report link speed
The PF driver tells us the link speed, so do something with that
information. Add link speed to log messages, and report speed through
ethtool.
Change-Id: I279dc9540cc5203376406050a3e8d67e128d5882 Signed-off-by: Mitch Williams <mitch.a.williams@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Jacob Keller [Thu, 4 Aug 2016 18:37:01 +0000 (11:37 -0700)]
i40e: use alloc_workqueue instead of create_singlethread_workqueue
Replace calls to create_singlethread_workqueue instead with alloc_workqueue
as is style with other Intel drivers. This provides more control over
workqueue creation, and allows explicit setting of the desired mode of
operation. It also makes it more obvious that driver name constant is
passed to a format "%s".
Change-ID: I6192b44caf5140336cd54c5b350d51c73b541fdb Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Jacob Keller [Tue, 19 Jul 2016 23:23:31 +0000 (16:23 -0700)]
i40e: use configured RSS key and lookup table in i40e_vsi_config_rss
A previous refactor added support to store user configuration for VSIs,
so that extra VSIs such as for VMDq can use this information when
configuring. Unfortunately the i40e_vsi_config_rss function was missed
in this refactor, and the values were being ignored. Fix this by
checking for the fields and using those instead of always using the
default values.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Jacob Keller [Tue, 19 Jul 2016 23:23:30 +0000 (16:23 -0700)]
i40e: fix broken i40e_config_rss_aq function
X722 hardware requires using the admin queue to configure RSS. This
function was previously re-written in commit e69ff813af35 ("i40e: rework
the functions to configure RSS with similar parameters").
However, the previous refactor did not work correctly for a few reasons
(a) it does not check whether seed is NULL before using it, resulting in
a NULL pointer dereference
(b) it does not even bother to use the passed in *lut parameter which
defines the requested lookup table. Instead it uses its own round robin
table.
Fix these issues by re-writing it to be similar to i40e_config_rss_reg
and i40e_get_rss_aq.
Fixes: e69ff813af35 ("i40e: rework the functions to configure RSS with similar parameters", 2015-10-21) Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Move this function below the two functions related to configuring RSS
via the admin queue. This helps co-locate the two functions, and made it
easier to spot a bug in the first i40e_config_rss_aq function as
compared to the i40e_get_rss_aq function.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Vegard Nossum [Fri, 19 Aug 2016 16:08:57 +0000 (18:08 +0200)]
net/irda: remove pointless assignment/check
We've already set sk to sock->sk and dereferenced it, so if it's NULL
we would have crashed already. Moreover, if it was NULL we would have
crashed anyway when jumping to 'out' and trying to unlock the sock.
Furthermore, if we had assigned a different value to 'sk' we would
have been calling lock_sock() and release_sock() on different sockets.
My conclusion is that these two lines are complete nonsense and only
serve to confuse the reader.
Signed-off-by: Vegard Nossum <vegard.nossum@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Yuval Mintz [Fri, 19 Aug 2016 06:33:31 +0000 (09:33 +0300)]
qed: utilize FW 8.10.10.0
This new firmware for the qed* adpaters fixes several issues:
- Better blocking of malicious VFs.
- After FLR, Tx-switching [internal routing] of packets might
be incorrect.
- Deletion of unicast MAC filters would sometime have side-effect
of corrupting the MAC filters configred for a device.
It also contains fixes for future qed* drivers that *hopefully* would be
sent for review in the near future.
In addition, it would allow driver some new functionality, including:
- Allowing PF/VF driver compaitibility with old drivers [running
pre-8.10.5.0 firmware].
- Better debug facilities.
This would also bump the qed* driver versions to 8.10.9.20.
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 20 Aug 2016 00:19:20 +0000 (17:19 -0700)]
Merge branch '10GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queue
Jeff Kirsher says:
====================
10GbE Intel Wired LAN Driver Updates 2016-08-18
This series contains updates to ixgbe and ixgbevf.
Emil cleans up confusing amongst the users by making an error message
into a debug message, since the TXDCTL.ENABLE (and comparable
VFTXDCTL.ENABLE for ixgbevf) bit is set only when the
transmit queue is actually enabled, which may not happen during the
configure phase eve if we waited for it. Converts to using netdev_dbg()
macro instead of our home brewed macro for ixgbevf. Converted the
service task to use atomic bitwise operations when setting and checking
reset requests to reduce the possibility of race conditions.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch series converts the bcm_sf2 driver from a traditional DSA driver
into a platform_device driver and makes it use the new DSA binding that Andrew
introduced in the latest merge window.
Prior attempts used to coerce the code in net/dsa/dsa2.c to accept the old
binding, while really there is only one broken single user out there: bcm_sf2,
so instead, just assume the new DT binding is deployed and use it accordingly.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Fainelli [Thu, 18 Aug 2016 22:30:16 +0000 (15:30 -0700)]
net: dsa: bcm_sf2: Remove probing through old DSA binding
Remove our dsa_switch_driver::drv_probe callback to prevent probing
through the old DSA binding, not that this could happen anymore now that
we have moved the matching compatible string from net/dsa/dsa.c to
drivers/net/dsa/bcm_sf2.c, so this is essentially dead code.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Fainelli [Thu, 18 Aug 2016 22:30:15 +0000 (15:30 -0700)]
net: dsa: bcm_sf2: Use device managed helpers
Now that we have converted the drivers into a proper platform device
driver, we can use the device managed helper functions to simplify the
error paths a bit wrt. register resources and IRQs.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Fainelli [Thu, 18 Aug 2016 22:30:14 +0000 (15:30 -0700)]
net: dsa: bcm_sf2: Make it a real platform device driver
The Broadcom Starfighter 2 switch driver should be a proper platform
driver, now that the DSA code has been updated to allow that, register a
switch device, feed it with the proper configuration data coming from
Device Tree and register our switch device with DSA.
The bulk of the changes consist in moving what bcm_sf2_sw_setup() did
into the platform driver probe function.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Florian Fainelli [Thu, 18 Aug 2016 22:30:12 +0000 (15:30 -0700)]
net: dsa: Export suspend/resume functions
In preparation for allowing switch drivers to implement system-wide
suspend/resume functions, export dsa_switch_suspend and
dsa_switch_resume() such that these are callable from the appropriate
driver specific suspend/resume functions.
Reviewed-by: Andrew Lunn <andrew@lunn.ch> Tested-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 20 Aug 2016 00:14:08 +0000 (17:14 -0700)]
Merge branch 'mv88e6xxx-fix-wait'
Andrew Lunn says:
====================
Fix mv88e6xxx wait function
The mv88e6xxx wait function can be upset of the system has nots of
other things to do and a sleep takes a lot longer than expected. Fix
this be using a fixed number of iterations, rather than a fixed
walkclock time.
Witht that change made, it is possible to consoliate another
wait function.
A wait actually timing out should not happen and when it does, it
means something serious is wrong. Make sure an error is logged,
since not all callers will log an error.
====================
Reviewed-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Thu, 18 Aug 2016 22:01:56 +0000 (00:01 +0200)]
dsa: mv88e6xxx: Use mv88e6xx_wait in mv88e6xxx_update()
Now that mv88e6xx_wait() iterated on a counter than a fixed time
interval, it implements the same mechanism as mv88e6xxx_update() uses.
So use it in mv88e6xx_wait().
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Thu, 18 Aug 2016 22:01:55 +0000 (00:01 +0200)]
dsa: mv88e6xxx: Timeout based on iterations, not time
The mv88e6xxx driver times out operations on the switch based on
looping until an elapsed wall clock time is reached. However, if
usleep_range() sleeps much longer than expected, it could timeout with
an error without actually checking to see if the devices has completed
the operation. So replace the elapsed time with a fixed upper bound on
the number of loops.
Testing on various switches has shown that switches takes either 0 or
1 iteration, so a maximum of 16 iterations is a safe limit.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 20 Aug 2016 00:11:50 +0000 (17:11 -0700)]
Merge branch 'phy-next'
Andrew Lunn says:
====================
PHY Kconfig and Makefile cleanup
The Ethernet PHY directory has slowly been getting more entries.
Split the entries in the Makefile and Kconfig into MDIO bus drivers
and PHYs. Within these two groups, sort them. This should reduce merge
conflicts and aid finding what one searches for.
The Kconfig text contains redundant "Driver for" and "Support for"
which add little value, make the vendor less obvious, and defeat the
shortcut key in the menu. Remove such text.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Thu, 18 Aug 2016 21:56:06 +0000 (23:56 +0200)]
net: phy: Kconfig: Remove redundant "Support for"
Remove the redundant "Support for" and "Drivers for" from the Kconfig
short description. This makes the manufacture much more prominent in
the list and makes the shortcut keys useful.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Acked-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn [Thu, 18 Aug 2016 21:56:05 +0000 (23:56 +0200)]
net: phy: Sort Makefile and Kconfig
Sort the files to reduce merge conflicts and to make it easier to find
drivers by name. Also separate the MDIO bus drivers from the PHY
drivers, again to help find what you need.
Signed-off-by: Andrew Lunn <andrew@lunn.ch> Acked-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Thu, 18 Aug 2016 17:19:34 +0000 (10:19 -0700)]
net: ipv4: fix sparse error in fib_good_nh()
Fixes following sparse errors :
net/ipv4/fib_semantics.c:1579:61: warning: incorrect type in argument 2
(different base types)
net/ipv4/fib_semantics.c:1579:61: expected unsigned int [unsigned]
[usertype] key
net/ipv4/fib_semantics.c:1579:61: got restricted __be32 const
[usertype] nh_gw
Fixes: a6db4494d218c ("net: ipv4: Consider failed nexthops in multipath routes") Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Thomas Falcon [Thu, 18 Aug 2016 16:37:51 +0000 (11:37 -0500)]
ibmvnic: Handle backing device failover and reinitialization
An upcoming feature of IBM VNIC protocol is the ability to configure
redundant backing devices for a VNIC client. In case of a failure
on the current backing device, the driver will receive a signal
from the hypervisor indicating that a failover will occur. The driver
will then wait for a message from the backing device before
establishing a new connection.
Signed-off-by: Thomas Falcon <tlfalcon@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
oulijun [Thu, 18 Aug 2016 12:32:52 +0000 (20:32 +0800)]
net: hns: Add reset function support for RoCE driver
It added reset function for RoCE driver. RoCE is a feature of hns.
In hip06 SoC, in RoCE reset process, it's needed to configure dsaf
channel reset, port and sl map info. Reset function of RoCE is
located in dsaf module, we only call it in RoCE driver when needed.
This patch is used to fix the conflict, please refer to this link:
https://www.spinics.net/lists/linux-rdma/msg39114.html
Signed-off-by: Wei Hu <xavier.huwei@huawei.com> Signed-off-by: Nenglong Zhao <zhaonenglong@hisilicon.com> Signed-off-by: Lijun Ou <oulijun@huawei.com> Signed-off-by: Sheng Li <lisheng011@huawei.com> Reviewed-by: Yisen Zhuang <yisen.zhuang@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 19 Aug 2016 21:40:25 +0000 (14:40 -0700)]
Merge branch 'rhash-raw-walkers-remove-part-1'
Herbert Xu says:
====================
rhashtable: Get rid of raw table walkers part 1
This series starts the process of getting rid of all raw rhashtable
walkers (e.g., using any of the rht_for_each helpers) from the
kernel.
We need to do this before I can fix the resize kmalloc failure issue
by using multi-layered tables.
We should do this anyway because almost all raw table walkers are
already buggy in that they don't handle multiple rhashtables during
a resize.
====================
Dave/Tomas, please keep an eye out for any new patches that try
to introduce raw table walkers and nack them.
Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Fri, 19 Aug 2016 08:21:37 +0000 (16:21 +0800)]
netlink: Use rhashtable walk interface in diag dump
This patch converts the diag dumping code to use the rhashtable
walk code instead of going through rhashtable by hand. The lock
nl_table_lock is now only taken while we process the multicast
list as it's not needed for the rhashtable walk.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Thu, 18 Aug 2016 08:50:56 +0000 (16:50 +0800)]
rhashtable: Remove GFP flag from rhashtable_walk_init
The commit 8f6fd83c6c5ec66a4a70c728535ddcdfef4f3697 ("rhashtable:
accept GFP flags in rhashtable_walk_init") added a GFP flag argument
to rhashtable_walk_init because some users wish to use the walker
in an unsleepable context.
In fact we don't need to allocate memory in rhashtable_walk_init
at all. The walker is always paired with an iterator so we could
just stash ourselves there.
This patch does that by introducing a new enter function to replace
the existing init function. This way we don't have to churn all
the existing users again.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 19 Aug 2016 07:00:45 +0000 (00:00 -0700)]
Merge branch 'cxgb-crypto'
Hariprasad Shenai says:
====================
crypto/chcr: Add support for Chelsio Crypto Driver
This patch series adds support for Chelsio Crypto driver.
The patch series has been created against net-next tree and includes
patches for Chelsio Low Level Driver(cxgb4) and adds the new crypto
Upper Layer Driver(chcr) under a new directory drivers/crypto/chelsio.
Patch 1/4 ("cxgb4: Add support for dynamic allocation of resources for
ULD") adds support for dynamic allocation of resources for ULD. The
objective of this patch is to provide generic interface for upper layer
drivers to allocate and initialize hardware resources.
The present cxgb4 (network driver) apart from network functionality, also
initializes hardware and thus acts as lower layer driver for other drivers
to use hardware resources. Thus it acts as both a Low level driver for
Upper layer driver's like iw_cxgb4, cxgb4i and cxgb4it and a Network Driver.
Right now the allocation of resources for Upper layer driver's is done
statically. Patch 1/4 adds a new infrastructure for dynamic allocation of
resources. cxgb4 will read the hardware capability through firmware and
allocate/free the queues for Upper layer drivers when the respective
driver's are loaded and freed when unloaded.
Patch 2/3, 3/4 and 4/4 adds support for Chelsio Crypto Driver. The Crypto
driver will act as another ULD on top of cxgb4.
In this patch series, the ULD API framework is used only by crypto and other
ULD's will make use of it in the next series.
This patch series is only for review, if this looks ok we will test it
thoroughly and send request for merge.
We have included all the maintainers of respective drivers. Kindly
review the changes and provide feedback on the same.
V3: - Removed crypto queues from cxgb4 and added support for dynamic
allocation of resources for Upper layer drivers
- Dependency fix in Kconfig.
V2: - Some residual code cleanup
- Adds pr_fmt with chcr (KBUILD_MODNAME) added
- Changes var name to accomodate them <80 columns in the chcr_register_alg
- Support for printing the crypto queue stats
- Fix compile warnings reported by kbuild bot for certain architectures
- Dependency fix in Kconfig.
- If the request has the MAY_BACKLOG bit set and hardware queue is
full the request is queued up else -EBUSY is returned to throttle
the user. The queue when executed and processed returns -EINPROGRESS
in completion.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Adds the config entry for the Chelsio Crypto Driver, Makefile changes
for the same.
Signed-off-by: Atul Gupta <atul.gupta@chelsio.com> Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The Chelsio's Crypto Hardware can perform the following operations:
SHA1, SHA224, SHA256, SHA384 and SHA512, HMAC(SHA1), HMAC(SHA224),
HMAC(SHA256), HMAC(SHA384), HAMC(SHA512), AES-128-CBC, AES-192-CBC,
AES-256-CBC, AES-128-XTS, AES-256-XTS
This patch implements the driver for above mentioned features. This
driver is an Upper Layer Driver which is attached to Chelsio's LLD
(cxgb4) and uses the queue allocated by the LLD for sending the crypto
requests to the Hardware and receiving the responses from it.
The crypto operations can be performed by Chelsio's hardware from the
userspace applications and/or from within the kernel space using the
kernel's crypto API.
The above mentioned crypto features have been tested using kernel's
tests mentioned in testmgr.h. They also have been tested from user
space using libkcapi and Openssl.
Signed-off-by: Atul Gupta <atul.gupta@chelsio.com> Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Atul Gupta <atul.gupta@chelsio.com> Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
cxgb4: Add support for dynamic allocation of resources for ULD
Add a new commmon infrastructure to allocate reosurces dynamically to
Upper layer driver's(ULD) when they register with cxgb4 driver and free
them during unregistering. All the queues and the interrupts for
them will be allocated during ULD probe only and freed during remove.
Signed-off-by: Atul Gupta <atul.gupta@chelsio.com> Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net>
LABBE Corentin [Wed, 17 Aug 2016 13:56:45 +0000 (15:56 +0200)]
atm: fore200e: Do not drop const qualifier
The data member of structure firmware is const and this constness is
dropped by some cast.
This patch add some const for keeping the const information.
Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 19 Aug 2016 06:38:17 +0000 (23:38 -0700)]
Merge branch 'bpf-next'
Daniel Borkmann says:
====================
BPF helper improvements and cleanups
This set adds various improvements to BPF helpers, a cleanup to use
skb_pkt_type_ok() helper, addition of bpf_skb_change_tail(), a follow
up for event output helper and removing ifdefs around the cgroupv2
helper bits. For details please see individual patches.
The set is based against net-next tree, but requires a merge of net
into net-next first.
Thanks a lot!
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Wed, 17 Aug 2016 23:00:41 +0000 (01:00 +0200)]
bpf: get rid of cgroup helper related ifdefs
As recently discussed during the task_under_cgroup_hierarchy() addition,
we should get rid of the ifdefs surrounding the bpf_skb_under_cgroup()
helper. If related functionality is not built-in, the helper cannot be
used anyway, which is also in line with what we do for all other helpers.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Wed, 17 Aug 2016 23:00:40 +0000 (01:00 +0200)]
bpf: enable event output helper also for xdp types
Follow-up to 555c8a8623a3 ("bpf: avoid stack copy and use skb ctx for
event output") for also adding the event output helper for XDP typed
programs. The event output helper has been very useful in particular for
debugging or event notification purposes, since it's much faster and
flexible than regular trace printk due to programmatically being able to
attach meta data. Same flags structure applies as with tc BPF programs.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Wed, 17 Aug 2016 23:00:39 +0000 (01:00 +0200)]
bpf: add bpf_skb_change_tail helper
This work adds a bpf_skb_change_tail() helper for tc BPF programs. The
basic idea is to expand or shrink the skb in a controlled manner. The
eBPF program can then rewrite the rest via helpers like bpf_skb_store_bytes(),
bpf_lX_csum_replace() and others rather than passing a raw buffer for
writing here.
bpf_skb_change_tail() is really a slow path helper and intended for
replies with f.e. ICMP control messages. Concept is similar to other
helpers like bpf_skb_change_proto() helper to keep the helper without
protocol specifics and let the BPF program mangle the remaining parts.
A flags field has been added and is reserved for now should we extend
the helper in future.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann [Wed, 17 Aug 2016 23:00:38 +0000 (01:00 +0200)]
bpf: use skb_pkt_type_ok helper in bpf_skb_change_type
Since we have a skb_pkt_type_ok() helper for checking the type before
mangling, make use of it instead of open coding. Follow-up to commit 8b10cab64c13 ("net: simplify and make pkt_type_ok() available for other
users") that came in after d2485c4242a8 ("bpf: add bpf_skb_change_type
helper").
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Acked-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Richard Alpe [Thu, 18 Aug 2016 08:33:52 +0000 (10:33 +0200)]
tipc: add peer removal functionality
Add TIPC_NL_PEER_REMOVE netlink command. This command can remove
an offline peer node from the internal data structures.
This will be supported by the tipc user space tool in iproute2.
Signed-off-by: Richard Alpe <richard.alpe@ericsson.com> Reviewed-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 17 Aug 2016 21:17:09 +0000 (14:17 -0700)]
tcp: refine tcp_prune_ofo_queue() to not drop all packets
Over the years, TCP BDP has increased a lot, and is typically
in the order of ~10 Mbytes with help of clever Congestion Control
modules.
In presence of packet losses, TCP stores incoming packets into an out of
order queue, and number of skbs sitting there waiting for the missing
packets to be received can match the BDP (~10 Mbytes)
In some cases, TCP needs to make room for incoming skbs, and current
strategy can simply remove all skbs in the out of order queue as a last
resort, incurring a huge penalty, both for receiver and sender.
Unfortunately these 'last resort events' are quite frequent, forcing
sender to send all packets again, stalling the flow and wasting a lot of
resources.
This patch cleans only a part of the out of order queue in order
to meet the memory constraints.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Neal Cardwell <ncardwell@google.com> Cc: Yuchung Cheng <ycheng@google.com> Cc: Soheil Hassas Yeganeh <soheil@google.com> Cc: C. Stephen Gun <csg@google.com> Cc: Van Jacobson <vanj@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Acked-by: Yuchung Cheng <ycheng@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Rafał Miłecki [Wed, 17 Aug 2016 21:11:52 +0000 (23:11 +0200)]
net: bgmac: make it clear when setting interface type to RMII
It doesn't really change anything as BGMAC_CHIPCTL_1_IF_TYPE_RMII is
equal to 0. It make code a bit clener, so far when reading it one could
think we forgot to set a proper mode. It also keeps this mode code in
sync with other ones.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl> Signed-off-by: David S. Miller <davem@davemloft.net>
Rafał Miłecki [Wed, 17 Aug 2016 21:00:30 +0000 (23:00 +0200)]
net: bgmac: support Ethernet core on BCM53573 SoCs
BCM53573 is a new series of Broadcom's SoCs. It's based on ARM and can
be found in two packages (versions): BCM53573 and BCM47189. It shares
some code with the Northstar family, but also requires some new quirks.
First of all there can be up to 2 Ethernet cores on this SoC. If that is
the case, they are connected to two different switch ports allowing some
more complex/optimized setups. It seems the second unit doesn't come
fully configured and requires some IRQ quirk.
Other than that only the first core is connected to the PHY. For the
second one we have to register fixed PHY (similarly to the Northstar),
otherwise generic PHY driver would get some invalid info.
This has been successfully tested on Tenda AC9 (BCM47189B0).
Signed-off-by: Rafał Miłecki <rafal@milecki.pl> Signed-off-by: David S. Miller <davem@davemloft.net>
Paul Durrant [Wed, 17 Aug 2016 15:13:29 +0000 (16:13 +0100)]
xen-netback: create a debugfs node for hash information
It is useful to be able to see the hash configuration when running tests.
This patch adds a debugfs node for that purpose.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com> Cc: Wei Liu <wei.liu2@citrix.com> Acked-by: Wei Liu <wei.liu2@citrix.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Edward Cree [Wed, 17 Aug 2016 15:00:10 +0000 (16:00 +0100)]
sfc: avoid division by zero
The division is already being done properly in efx_ef10_get_timer_config
which returns zero-on-success, unlike the old efx_ef10_get_sysclk_freq.
Fixes: d95e329a55ba ("sfc: get timer configuration from adapter") Signed-off-by: Edward Cree <ecree@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Wed, 17 Aug 2016 14:48:36 +0000 (07:48 -0700)]
tcp: defer sacked assignment
While chasing tcp_xmit_retransmit_queue() kasan issue, I found
that we could avoid reading sacked field of skb that we wont send,
possibly removing one cache line miss.
Very minor change in slow path, but why not ? ;)
Signed-off-by: Eric Dumazet <edumazet@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 19 Aug 2016 06:18:42 +0000 (23:18 -0700)]
Merge branch 'bridge-vlan-stats-with-flags'
Nikolay Aleksandrov says:
====================
net: bridge: export vlan stats per-port with flags
This set adds the ability to export vlan stats per-port. Patch 01 makes
that possible by consolidating the bridge and port linkxstats calls. Then
patch 02 allows to dump the vlan entry flags in order to be able to
distinguish between bridge and port vlan entries when dumping the master
device vlan stats. That is needed because that call was implemented when
the stats API didn't have slave dumping capabilities and it dumps all vlan
stats (for both bridge and port entries). We also need it in order to print
the vlan flags when dumping the stats.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Use one of the vlan xstats padding fields to export the vlan flags. This is
needed in order to be able to distinguish between master (bridge) and port
vlan entries in user-space when dumping the bridge vlan stats.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by: David S. Miller <davem@davemloft.net>
net: bridge: consolidate bridge and port linkxstats calls
In the bridge driver we usually have the same function working for both
port and bridge. In order to follow that logic and also avoid code
duplication, consolidate the bridge_ and brport_ linkxstats calls into
one since they share most of their code. As a side effect this allows us
to dump the vlan stats also via the slave call which is in preparation for
the upcoming per-port vlan stats and vlan flag dumping.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 19 Aug 2016 06:13:14 +0000 (23:13 -0700)]
Merge branch 'flow-dissector-vlan-tag'
Hadar Hen Zion says:
====================
net_sched, flow_dissector, flower: Introduce vlan tag support
This patchset introduce vlan tag support to the flower classifier and the flow
dissector. In addition to adding vlan priority to act vlan.
The first 2 patches are dealing with flow-dissector:
- The first patch is a fix, in case the vlan was already stripped from the
skb, take it from skb->vlan_tci.
- The second patch adds support for vlan priority.
The next 2 patches are dealing with flower:
- The first patch is a fix, sets flow dissector 'used_keys' according to the
mask value of each key.
- The secound patch adds vlan tag support to the flower classifier, user space
patches will be sent later to complete it.
The last patch adds vlan priority to act vlan since only vlan id is currently supported.
Changes from V1:
- A new patch was added to this series "net_sched: flower: Avoid dissection of unmasked keys"
- Adding u16 padding to struct flow_dissector_key_vlan
- change flow_label field in struct flow_dissector_key_tags form 20 bits field to u32
- Remove 'if (v->tcfv_push_prio)' check from tcf_vlan_dump function
- Add support to un-stripped vlan skb and skb with multipale vlans in __skb_flow_dissect
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:12 +0000 (13:36 +0300)]
net_sched: flower: Avoid dissection of unmasked keys
The current flower implementation checks the mask range and set all the
keys included in that range as "used_keys", even if a specific key in
the range has a zero mask.
This behavior can cause a false positive return value of
dissector_uses_key function and unnecessary dissection in
__skb_flow_dissect.
This patch checks explicitly the mask of each key and "used_keys" will
be set accordingly.
Fixes: 77b9900ef53a ('tc: introduce Flower classifier') Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Signed-off-by: Jiri Pirko <jiri@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:11 +0000 (13:36 +0300)]
flow_dissector: Get vlan priority in addition to vlan id
Add vlan priority check to the flow dissector by adding new flow
dissector struct, flow_dissector_key_vlan which includes vlan tag
fields.
vlan_id and flow_label fields were under the same struct
(flow_dissector_key_tags). It was a convenient setting since struct
flow_dissector_key_tags is used by struct flow_keys and by setting
vlan_id and flow_label under the same struct, we get precisely 24 or 48
bytes in flow_keys from flow_dissector_key_basic.
Now, when adding vlan priority support, the code will be cleaner if
flow_label and vlan tag won't be under the same struct anymore.
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Acked-by: Jiri Pirko <jiri@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Hadar Hen Zion [Wed, 17 Aug 2016 10:36:10 +0000 (13:36 +0300)]
flow_dissector: For stripped vlan, get vlan info from skb->vlan_tci
Early in the datapath skb_vlan_untag function is called, stripped
the vlan from the skb and set skb->vlan_tci and skb->vlan_proto fields.
The current dissection doesn't handle stripped vlan packets correctly.
In some flows, vlan doesn't exist in skb->data anymore when applying
flow dissection on the skb, fix that.
In case vlan info wasn't stripped before applying flow_dissector (RPS
flow for example), or in case of skb with multiple vlans (e.g. 802.1ad),
get the vlan info from skb->data. The flow_dissector correctly skips
any number of vlans and stores only the first level vlan.
Fixes: 0744dd00c1b1 ('net: introduce skb_flow_dissect()') Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com> Acked-by: Jiri Pirko <jiri@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Fri, 19 Aug 2016 06:04:40 +0000 (23:04 -0700)]
Merge branch 'qed-link-fixes'
Yuval Mintz says:
====================
qed*: Fix ethtool issues relating to link
This series addresses two issues that were introduced when adding
support for ethtool's link_ksettings support - the first fixes a
regression and second incorrect functionallity in the submission.
Although these are fixes, as the feature currently exists only in
'next-next' I'm aiming them for it.
Dave, please consider applying this series to 'net-next'.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
projects / mirror_ubuntu-zesty-kernel.git / log