openwrt/staging/blogic.git
9 years agoath9k_htc: wmi: match wait_for_completion_timeout return type
Nicholas Mc Guire [Thu, 14 May 2015 16:43:41 +0000 (18:43 +0200)]
ath9k_htc: wmi: match wait_for_completion_timeout return type

Return type of wait_for_completion_timeout is unsigned long not int.
As time_left is exclusively used for wait_for_completion_timeout here its
type is simply changed to unsigned long.

API conformance testing for completions with coccinelle spatches are being
used to locate API usage inconsistencies:
./drivers/net/wireless/ath/ath9k/wmi.c:331
int return assigned to unsigned long

Patch was compile tested with x86_64_defconfig + CONFIG_ATH_CARDS=m,
CONFIG_ATH9K_HTC=m

Patch is against 4.1-rc3 (localversion-next is -next-20150514)

Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k_htc: drv_init: match wait_for_completion_timeout return type
Nicholas Mc Guire [Thu, 14 May 2015 16:25:26 +0000 (18:25 +0200)]
ath9k_htc: drv_init: match wait_for_completion_timeout return type

Return type of wait_for_completion_timeout is unsigned long not int.
As time_left is exclusively used for wait_for_completion_timeout here its
type is simply changed to unsigned long.

API conformance testing for completions with coccinelle spatches are being
used to locate API usage inconsistencies:
./drivers/net/wireless/ath/ath9k/htc_drv_init.c:81
int return assigned to unsigned long

Patch was compile tested with x86_64_defconfig + CONFIG_ATH_CARDS=m,
CONFIG_ATH9K_HTC=m

Patch is against 4.1-rc3 (localversion-next is -next-20150514)

Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k_htc: match wait_for_completion_timeout return type
Nicholas Mc Guire [Thu, 14 May 2015 15:29:54 +0000 (17:29 +0200)]
ath9k_htc: match wait_for_completion_timeout return type

Return type of wait_for_completion_timeout is unsigned long not int.
As time_left is exclusively used for wait_for_completion_timeout here its
type is simply changed to unsigned long.

API conformance testing for completions with coccinelle spatches are being
used to locate API usage inconsistencies:
./drivers/net/wireless/ath/ath9k/htc_hst.c:171
int return assigned to unsigned long
./drivers/net/wireless/ath/ath9k/htc_hst.c:277
int return assigned to unsigned long
./drivers/net/wireless/ath/ath9k/htc_hst.c:206
int return assigned to unsigned long

Patch was compile tested with x86_64_defconfig + CONFIG_ATH_CARDS=m,
CONFIG_ATH9K_HTC=m

Patch is against 4.1-rc3 (localversion-next is -next-20150514)

Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agortlwifi: misspelled code and comments corrected.
Cheolhyun Park [Sun, 12 Apr 2015 16:59:48 +0000 (16:59 +0000)]
rtlwifi: misspelled code and comments corrected.

Signed-off-by: Cheolhyun Park <pch851130@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath: Make ath_opmode_to_string understand OCB mode
Bertold Van den Bergh [Mon, 3 Aug 2015 13:45:41 +0000 (15:45 +0200)]
ath: Make ath_opmode_to_string understand OCB mode

Make ath_opmode_to_string return "OCB" for NL80211_IFTYPE_OCB. Currently
it will return "UNKNOWN".

Signed-off-by: Bertold Van den Bergh <bertold.vandenbergh@esat.kuleuven.be>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agobcma: fix access to host_pdev for PCIe devices
Hauke Mehrtens [Sun, 2 Aug 2015 18:26:52 +0000 (20:26 +0200)]
bcma: fix access to host_pdev for PCIe devices

bus->host_pdev is part of a union so bus->host_pdev != NULL is probably
also true for PCIe devices, because there it accesses bus->host_pci. If
we access the dev member at the offset defined in struct
platform_device in struct pci_dev instead we probably get something
else.

This patch adds a new function which returns the host dev struct and
NULL if we do not have a host dev. When this gets registered on MIPS
brcm47xx we do not have a host dev in some situations.
This function could also be used in other places.

This problem was introduced in this commit:
commit cae761b5a6bdc597ba476a040fdcd5b4bc559b85
Author: Rafa? Mi?ecki <zajec5@gmail.com>
Date:   Sun Jun 28 17:17:13 2015 +0200

    bcma: populate bus DT subnodes as platform_device-s

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomt7601u: lock out rx path and tx status reporting
Jakub Kicinski [Fri, 31 Jul 2015 13:04:49 +0000 (15:04 +0200)]
mt7601u: lock out rx path and tx status reporting

mac80211 requires that rx path does not run concurrently with
tx status reporting.  Add a spinlock which will ensure that.

Signed-off-by: Jakub Kicinski <kubakici@wp.pl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomt7601u: fix tx status reporting contexts
Jakub Kicinski [Fri, 31 Jul 2015 13:04:48 +0000 (15:04 +0200)]
mt7601u: fix tx status reporting contexts

mac80211 requires that rx path does not run concurrently with
tx status reporting.  Since rx path is run in driver tasklet,
tx status cannot be reported directly from interrupt context
(there would be no way to lock it out).

Add tasklet for tx and move all possible code from irq handler
there.

Note: tx tasklet is needed because workqueue is queued very
      rarely and that kills TCP performance.

Signed-off-by: Jakub Kicinski <kubakici@wp.pl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomt7601u: use correct ieee80211_rx variant
Jakub Kicinski [Fri, 31 Jul 2015 13:04:47 +0000 (15:04 +0200)]
mt7601u: use correct ieee80211_rx variant

Rx is run inside a tasklet so ieee80211_rx() should be used
instead of ieee80211_rx_ni().

Signed-off-by: Jakub Kicinski <kubakici@wp.pl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomt7601u: fix dma from stack address
Jakub Kicinski [Fri, 31 Jul 2015 13:04:46 +0000 (15:04 +0200)]
mt7601u: fix dma from stack address

DMA to variables located on the stack is a bad idea.
For simplicity and to avoid frequent allocations create
a buffer inside the device structure.  Protect this
buffer with vendor_req_mutex.  Don't protect vendor
requests which don't use this buffer.

Signed-off-by: Jakub Kicinski <kubakici@wp.pl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agowlcore: add p2p device support
Eliad Peller [Thu, 30 Jul 2015 19:38:22 +0000 (22:38 +0300)]
wlcore: add p2p device support

When starting a p2p mgmt interface, enable its device role. This
allows us to keep the sta role disabled and scan on the dev role.

In general, p2p management interfaces cannot send vif-specific commands
to FW, as the vif role id is invalid. Only off-channel data and scans
happen on this vif, so most ops are not relevant.

If the vif is a p2p mgmt vif, block some mac80211 ops.

Configure rate policies for p2p mgmt interface, as
otherwise p2p packets come out with arbitrary rates.

Since wpa_supplicant currently doesn't support standalone
p2p device mode (without another attached managed interface),
add p2p device to the allowed interface combinations without
decreasing the allowed station count.

Moreover, increase the station count in some cases, as AP
mode usually starts as station interface, and the AP interface
is now different from the p2p management one).

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agowlcore: add generic_cfg_feature command definitions
Eliad Peller [Thu, 30 Jul 2015 19:38:21 +0000 (22:38 +0300)]
wlcore: add generic_cfg_feature command definitions

Add definitions and function prototypes for generic_cfg
command.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agowl18xx: use long intervals in sched scan
Eliad Peller [Thu, 30 Jul 2015 19:38:20 +0000 (22:38 +0300)]
wl18xx: use long intervals in sched scan

Add support for long intervals on sched scan.

If configured, the original request interval will
be used num_short_interval times, and then the
long interval will be used.

While on it, fix the scan command field names
to reflect the expected value is in ms (rather
than secs).

These values will be taken from the conf file,
so bump its version accordingly.

Signed-off-by: Eyal Shapira <eyal@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agowlcore/wl18xx : add time sync event handling
Machani, Yaniv [Thu, 30 Jul 2015 19:38:19 +0000 (22:38 +0300)]
wlcore/wl18xx : add time sync event handling

Added support for a new time sync event
the event data contains the WiLink TSF value.
To trigger the event, a HW modification is required,
so as a supporting firmware binary.

Signed-off-by: Yaniv Machani <yanivma@ti.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: correct TDLS link delete failure message priority
Amitkumar Karwar [Tue, 28 Jul 2015 14:22:20 +0000 (07:22 -0700)]
mwifiex: correct TDLS link delete failure message priority

Commit d8d2f19feb16 ("mwifiex: silence TDLS link delete failure
for nonexistent link") lowers the priority of error message when
TDLS link is already deleted. But it had got increased by
commit acebe8c10a6e ("mwifiex: change dbg print func to
mwifiex_dbg")

Reported-by: Nicholas Mc Guire <der.herr@hofr.at>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: fix build with CONFIG_ATH9K_STATION_STATISTICS=y
Felix Fietkau [Thu, 6 Aug 2015 10:44:30 +0000 (12:44 +0200)]
ath9k: fix build with CONFIG_ATH9K_STATION_STATISTICS=y

kbuild bot reported that commit 592fa228f213 ("ath9k: remove struct
ath_atx_ac") broke the build when CONFIG_ATH9K_STATION_STATISTICS is y.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Fixes: 592fa228f213 ("ath9k: remove struct ath_atx_ac")
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoMerge tag 'iwlwifi-next-for-kalle-2015-08-04' of https://git.kernel.org/pub/scm/linux...
Kalle Valo [Thu, 6 Aug 2015 07:27:59 +0000 (10:27 +0300)]
Merge tag 'iwlwifi-next-for-kalle-2015-08-04' of https://git./linux/kernel/git/iwlwifi/iwlwifi-next

* Deprecate -10.ucode
* Clean ups towards multiple Rx queues
* Add support for longer CMD IDs. This will be required by new
firmwares since we are getting close to the u8 limit.
* bugfixes for the D0i3 power state
* Add basic support for FTM
* More random that doesn't really stand out

9 years agowlcore: add antenna diversity reading comments
Guy Mishol [Mon, 27 Jul 2015 14:25:49 +0000 (17:25 +0300)]
wlcore: add antenna diversity reading comments

add comments to the antenna diversity reading

Signed-off-by: Guy Mishol <guym@ti.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: add missing skb_push() in mwifiex_check_uap_capabilties
Amitkumar Karwar [Mon, 27 Jul 2015 12:02:27 +0000 (05:02 -0700)]
mwifiex: add missing skb_push() in mwifiex_check_uap_capabilties

For PCIe/USB chipsets, preallocated skb buffers are reused for
event handling. mwifiex_check_uap_capabilties() performs
skb_pull(). This patch adds missing skb_push() to restore skb's
data pointer/length.

This bug was introduced by commit debfc6008169 ("mwifiex: update
AP WMM settings from BSS_START event")

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Nishant Sarmukadam <nishants@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agowl18xx: add dynamic fw traces
Guy Mishol [Mon, 27 Jul 2015 06:46:02 +0000 (09:46 +0300)]
wl18xx: add dynamic fw traces

add option to dynamically configure the fw
which debug traces to open

Signed-off-by: Guy Mishol <guym@ti.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: corrections in PCIe event skb handling
Amitkumar Karwar [Wed, 22 Jul 2015 11:53:45 +0000 (04:53 -0700)]
mwifiex: corrections in PCIe event skb handling

Preallocated event SKBs are getting reused for PCIe chipset.
Their physical addresses are shared with firmware so that
firmware can write data into them.

This patch makes sure that SKB is cleared and length is set to
default while submitting it to firmware.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Zhaoyang Liu <liuzy@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: fix system crash observed during initialisation
Zhaoyang Liu [Wed, 22 Jul 2015 11:53:44 +0000 (04:53 -0700)]
mwifiex: fix system crash observed during initialisation

System crash was observed if one of the driver initialisation
commands is timed out. The reason is our timeout handler triggers
firmware dump, meanwhile driver initialisation error paths have
already freed the adapter structure.

Firmware hasn't yet completely initialized. So collecting firmware
dump is not needed in this case. Command timeout handler is
modified in this patch to fix the crash issue.

Signed-off-by: Zhaoyang Liu <liuzy@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: fix command timeout for PCIe chipsets
Zhaoyang Liu [Wed, 22 Jul 2015 11:53:43 +0000 (04:53 -0700)]
mwifiex: fix command timeout for PCIe chipsets

When WLAN interface is up and running, driver unload and
load was causing command timeout error.

We enable Rx data by updating RX ring read pointer in
init_fw_port(). It should be done when FW is completely
intialialised. Command timeout is fixed in this patch by
moving init_fw_port() call to mwifiex_init_fw_complete().

Signed-off-by: Zhaoyang Liu <liuzy@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: using right aid value for tdls action frame
Xinming Hu [Wed, 22 Jul 2015 11:53:42 +0000 (04:53 -0700)]
mwifiex: using right aid value for tdls action frame

Variable pos is u8 here, so memcpy is needed to store u16 aid.
At the same time, aid should be platform independent, upper layer
utility(wpa_supplicant,etc.,) parse it as le16, so keep it le16
here.

Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: remove the sched field in struct ath_atx_tid
Felix Fietkau [Wed, 22 Jul 2015 11:06:14 +0000 (13:06 +0200)]
ath9k: remove the sched field in struct ath_atx_tid

Use list_empty(&tid->list) instead

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: remove struct ath_atx_ac
Felix Fietkau [Wed, 22 Jul 2015 11:06:13 +0000 (13:06 +0200)]
ath9k: remove struct ath_atx_ac

struct ath_atx_ac contains a list of active TIDs belonging to one WMM AC.
This patch changes the code to track active station TIDs in the txq directly.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: add fast-xmit support
Felix Fietkau [Wed, 22 Jul 2015 11:06:12 +0000 (13:06 +0200)]
ath9k: add fast-xmit support

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: setup rxfilter when offchannel
Janusz.Dziedzic@tieto.com [Tue, 21 Jul 2015 09:11:43 +0000 (11:11 +0200)]
ath9k: setup rxfilter when offchannel

Setup rxfiler correctly for offchannel ctx.

This fix problem we didn't configure rxfilter, next
didn't receive probe requests and next failed
p2p_find. This was seen when ath9k loaded with
use_chanctx=1

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: setup rxfilter for all chanctx
Janusz.Dziedzic@tieto.com [Tue, 21 Jul 2015 09:11:42 +0000 (11:11 +0200)]
ath9k: setup rxfilter for all chanctx

While mac80211 setup this per HW, set same
rxfilter configuration for all chanctx.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: handle RoC cancel correctly
Janusz.Dziedzic@tieto.com [Tue, 21 Jul 2015 09:11:41 +0000 (11:11 +0200)]
ath9k: handle RoC cancel correctly

In case we will get ROC cancel from mac80211 we
should not call ieee80211_remain_on_channel_expired().

In other case I hit such warning on MIPS and
p2p negotiation failed (tested with use_chanctx=1).

ath: phy0: Starting RoC period
ath: phy0: Channel definition created: 2412 MHz
ath: phy0: Assigned next_chan to 2412 MHz
ath: phy0: Offchannel duration for chan 2412 MHz : 506632
ath: phy0: ath_chanctx_set_next: current: 2412 MHz, next: 2412 MHz
ath: phy0: Stopping current chanctx: 2412
ath: phy0: Flush timeout: 200
ath: phy0: ath_chanctx_set_next: Set channel 2412 MHz
ath: phy0: Set channel: 2412 MHz width: 0
ath: phy0: Reset to 2412 MHz, HT40: 0 fastcc: 0
ath: phy0: cur_chan: 2412 MHz, event: ATH_CHANCTX_EVENT_TSF_TIMER, state: ATH_CHANCTX_STATE_IDLE
ath: phy0: ath_offchannel_channel_change: offchannel state: ATH_OFFCHANNEL_ROC_START
ath: phy0: cur_chan: 2412 MHz, event: ATH_CHANCTX_EVENT_SWITCH, state: ATH_CHANCTX_STATE_IDLE
ath: phy0: Cancel RoC
ath: phy0: RoC aborted
ath: phy0: RoC request on vif: 00:03:7f:4e:a0:cd, type: 1 duration: 500
ath: phy0: Starting RoC period
ath: phy0: Channel definition created: 2412 MHz
ath: phy0: Assigned next_chan to 2412 MHz
ath: phy0: Offchannel duration for chan 2412 MHz : 506705
ath: phy0: ath_chanctx_set_next: current: 2412 MHz, next: 2412 MHz
ath: phy0: ath_offchannel_channel_change: offchannel state: ATH_OFFCHANNEL_ROC_START
ath: phy0: cur_chan: 2412 MHz, event: ATH_CHANCTX_EVENT_SWITCH, state: ATH_CHANCTX_STATE_IDLE
------------[ cut here ]------------
WARNING: CPU: 0 PID: 3312 at drivers/net/wireless/ath/ath9k/main.c:2319
Modules linked in: ath9k ath9k_common ath9k_hw ath mac80211 cfg80211

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoath9k: advertise p2p dev support when chanctx
Janusz.Dziedzic@tieto.com [Tue, 21 Jul 2015 09:11:40 +0000 (11:11 +0200)]
ath9k: advertise p2p dev support when chanctx

Advertise p2p device support when ath9k loaded with
use_chanctx=1.

This will fix problem, when first interface is an AP
and next we would like to run p2p_find.
Before p2p find (scan phase) failed with EOPNOTSUPP.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@tieto.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: simplify mwifiex_complete_cmd
Andreas Fenkart [Fri, 17 Jul 2015 07:13:06 +0000 (09:13 +0200)]
mwifiex: simplify mwifiex_complete_cmd

600f5d909a54("mwifiex: cleanup ioctl wait queue and abstraction layer")
introduced the wakeup_interruptible suppression in mwifiex_complete_cmd
b1a47aa5e1e1("mwifiex: fix system hang issue in cmd timeout error case")
then added wakup_interruptible to mwifiex_cmd_timeout_func the single
place setting a status of ETIMEDOUT.
Instead of doing extra work, using the standard call-chain will have the
same effect:
mwifiex_cancel_pending_ioctl
-> mwifiex_recycle_cmd_node
-> mwifiex_insert_cmd_to_free_q
-> mwifiex_complete_cmd
-> wake_up_interruptible

The difference is that previously the condition was not set to true,
but that's probably just an oversight in b1a47aa5e1e1 and shouldn't
have any consequence

Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
Acked-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: remove CMD_F_CANCELED flag
Andreas Fenkart [Fri, 17 Jul 2015 07:13:05 +0000 (09:13 +0200)]
mwifiex: remove CMD_F_CANCELED flag

CMD_F_CANCELED was used to abort mwifiex_process_cmdresp in
case it already started or starts processing the cmd.
But this was probably not working the way intended:
- it is racy: mwifiex_process_cmdresp might already have passed that
  test and is continuing to use the cmd node being recycled
- mwifiex_process_cmdresp repeatedly uses adapter->curr_cmd which
  we just set to NULL
- mwifiex_recycle_cmd_node will clear the flag

The reason why it probably works is that mwifiex_cancel_pending_ioctl
is only called from mwifiex_cmd_timeout_func, where the there is little
chance of a command response still arriving

Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
Acked-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: remove redundant reset of cmd_wait_q status
Andreas Fenkart [Fri, 17 Jul 2015 07:13:04 +0000 (09:13 +0200)]
mwifiex: remove redundant reset of cmd_wait_q status

mwifiex_cancel_pending_ioctl is called only from
mwifiex_cmd_timeout_func. There the wait_q status is set to
-ETIMEDWAIT before calling this function. Whether we reset the status
to -1 or leave it at -ETIMEDWAIT at end doesn't matter since both
are != 0 hence mean failure

Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
Acked-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: remove explicit mwifiex_complete_cmd calls
Andreas Fenkart [Fri, 17 Jul 2015 07:13:03 +0000 (09:13 +0200)]
mwifiex: remove explicit mwifiex_complete_cmd calls

standard call chain when releasing a cmd node:
mwifiex_recycle_cmd_node
-> mwifiex_insert_cmd_to_free_q
-> mwifiex_complete_cmd, if wait_q_enabled

calling mwifiex_complete_cmd explicitly and setting
wait_q_enabled = false is redundant

Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
Acked-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agomwifiex: sdio: reset adapter using mmc_hw_reset
Andreas Fenkart [Thu, 16 Jul 2015 16:50:01 +0000 (18:50 +0200)]
mwifiex: sdio: reset adapter using mmc_hw_reset

Since 1fb654fd97ff("mmc: sdio: add reset callback to bus operations"),
sdio cards can be power cycled using mmc_hw_reset.
The use mmc_remove_host/mmc_add_host is discouraged, because these are
internal functions to the mmc core and should only be used by mmc hosts

Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
Acked-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next
David S. Miller [Wed, 5 Aug 2015 06:57:45 +0000 (23:57 -0700)]
Merge git://git./linux/kernel/git/pablo/nf-next

Pablo Neira Ayuso says:

====================
Netfilter updates for net-next

The following patchset contains Netfilter updates for net-next, they are:

1) A couple of cleanups for the netfilter core hook from Eric Biederman.

2) Net namespace hook registration, also from Eric. This adds a dependency with
   the rtnl_lock. This should be fine by now but we have to keep an eye on this
   because if we ever get the per-subsys nfnl_lock before rtnl we have may
   problems in the future. But we have room to remove this in the future by
   propagating the complexity to the clients, by registering hooks for the init
   netns functions.

3) Update nf_tables to use the new net namespace hook infrastructure, also from
   Eric.

4) Three patches to refine and to address problems from the new net namespace
   hook infrastructure.

5) Switch to alternate jumpstack in xtables iff the packet is reentering. This
   only applies to a very special case, the TEE target, but Eric Dumazet
   reports that this is slowing down things for everyone else. So let's only
   switch to the alternate jumpstack if the tee target is in used through a
   static key. This batch also comes with offline precalculation of the
   jumpstack based on the callchain depth. From Florian Westphal.

6) Minimal SCTP multihoming support for our conntrack helper, from Michal
   Kubecek.

7) Reduce nf_bridge_info per skbuff scratchpad area to 32 bytes, from Florian
   Westphal.

8) Fix several checkpatch errors in bridge netfilter, from Bernhard Thaler.

9) Get rid of useless debug message in ip6t_REJECT, from Subash Abhinov.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoiwlwifi: mvm: clean up fw-api-scan.h
David Spinadel [Wed, 24 Jun 2015 08:03:56 +0000 (11:03 +0300)]
iwlwifi: mvm: clean up fw-api-scan.h

Remove outdated and unused definitions

Signed-off-by: David Spinadel <david.spinadel@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: add wide firmware command support for debug triggers
Sara Sharon [Mon, 13 Jul 2015 11:23:59 +0000 (14:23 +0300)]
iwlwifi: mvm: add wide firmware command support for debug triggers

Add support for extended command id in triggers handling.
Extended command id header contains group id in addition to command id.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: add wide firmware command infrastructure for RX
Avraham Stern [Thu, 9 Jul 2015 14:17:03 +0000 (17:17 +0300)]
iwlwifi: mvm: add wide firmware command infrastructure for RX

Add support for extended firmware event header that contains
a group id as well as the command id.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: add wide firmware command support for notifications
Sara Sharon [Mon, 13 Jul 2015 11:50:47 +0000 (14:50 +0300)]
iwlwifi: add wide firmware command support for notifications

Add support for extended command id in notification system.
Extended command id header contains group id in addition to command id.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: make various functions void in the file rs.c
Nicholas Krause [Wed, 8 Jul 2015 15:02:15 +0000 (11:02 -0400)]
iwlwifi: make various functions void in the file rs.c

This makes various functions in the file rs.c void due to these
functions never returning a error code to signal to their callers
if and how they have failed to complete their intended work.

Signed-off-by: Nicholas Krause <xerofoify@gmail.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: move TX PN assignment for CCMP to the driver
Johannes Berg [Mon, 20 Apr 2015 15:54:54 +0000 (17:54 +0200)]
iwlwifi: mvm: move TX PN assignment for CCMP to the driver

Move the TX PN assignment (for CCMP only) to the driver. This prepares
the driver for future DSO (driver segmentation offload) where it will
split an SKB into multiple MPDUs by itself.

For TDLS, split out the CCMP TX command handling so that it won't get
a PN assigned, the firmware assigns the PN in that case.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agonetfilter: ip6t_REJECT: Remove debug messages from reject_tg6()
Subash Abhinov Kasiviswanathan [Thu, 30 Jul 2015 16:53:45 +0000 (16:53 +0000)]
netfilter: ip6t_REJECT: Remove debug messages from reject_tg6()

Make it similar to reject_tg() in ipt_REJECT.

Suggested-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
9 years agoMerge branch 'cxgb4-next'
David S. Miller [Tue, 4 Aug 2015 08:24:35 +0000 (01:24 -0700)]
Merge branch 'cxgb4-next'

Hariprasad Shenai says:

====================
add meminfo, bist status and misc. fixes

This patch series adds the following.
Add support to dump memory address range of various hw modules
Add support to dump edc bist status during ecc error
Read correct bits of who am i register for T6 adapter
and update T6 register range

This patch series has been created against net-next tree and includes
patches on cxgb4 and cxgb4vf driver.

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

V2: PATCH 3/4 ("cxgb4/cxgb4vf: read the correct bits of PL Who Am I
    register") Fix switch statement in get_chip_type() and some more style
    fixes based on review comment by Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4: Update T6 register ranges
Hariprasad Shenai [Tue, 4 Aug 2015 09:06:20 +0000 (14:36 +0530)]
cxgb4: Update T6 register ranges

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4/cxgb4vf: read the correct bits of PL Who Am I register
Hariprasad Shenai [Tue, 4 Aug 2015 09:06:19 +0000 (14:36 +0530)]
cxgb4/cxgb4vf: read the correct bits of PL Who Am I register

Read the correct bits of PL Who Am I for the Source PF field which has
changed in T6

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4: Add support to dump edc bist status
Hariprasad Shenai [Tue, 4 Aug 2015 09:06:18 +0000 (14:36 +0530)]
cxgb4: Add support to dump edc bist status

Add support to dump edc bist status for ECC data errors

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocxgb4: Add debugfs support to dump meminfo
Hariprasad Shenai [Tue, 4 Aug 2015 09:06:17 +0000 (14:36 +0530)]
cxgb4: Add debugfs support to dump meminfo

Add debug support to dump memory address ranges of various hardware
modules of the adapter.

Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoiwlwifi: pcie: reset write pointer on ict reset
Eliad Peller [Thu, 16 Jul 2015 17:17:42 +0000 (20:17 +0300)]
iwlwifi: pcie: reset write pointer on ict reset

Since the CSR_DRAM_INIT_TBL_WRITE_POINTER bit wasn't set
on ict reset, in some flows (like disable ict followed by
immediate reset ict) the driver and hardware went out
of sync (the driver cleared the ict_index, while the hw
kept it intact).

Fix it by setting the flag when resetting ict.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: call d3_suspend/resume in d0i3 case as well
Eliad Peller [Thu, 9 Jul 2015 11:17:24 +0000 (14:17 +0300)]
iwlwifi: call d3_suspend/resume in d0i3 case as well

Some CSR registers have to be configured also
in case of suspend/resume with unified image
(which doesn't includes reconfiguration flow).

Reuse the existing d3_suspend/d3_resume trans ops,
while making sure some configurations are a bit
different, according to the wowlan type.

After this change, we no longer need the special
wowlan_d0i3 configurations done in iwl_pci_resume,
as they are already being done in the d3_resume op.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: ignore CQM when setting beacon filtering in D0i3 enter flow
Gregory Greenman [Mon, 20 Jul 2015 09:55:34 +0000 (12:55 +0300)]
iwlwifi: mvm: ignore CQM when setting beacon filtering in D0i3 enter flow

CQM overwrites a few thresholds in the bf command. On the other hand,
when entering D0i3 the thresholds are set to higher values on purpose,
so ignore CQM in this case.

Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: fix beacon filtering temperature thresholds for D0i3
Gregory Greenman [Wed, 22 Jul 2015 12:46:37 +0000 (15:46 +0300)]
iwlwifi: mvm: fix beacon filtering temperature thresholds for D0i3

The slow filtering threshold should be higher in D0i3 case.

Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: Enable Rx Checksum hw
Avri Altman [Sat, 27 Dec 2014 07:09:47 +0000 (09:09 +0200)]
iwlwifi: mvm: Enable Rx Checksum hw

TCP software implementation on the host requires extensive computing
power.  Offloading even some of the TCP/IP stack to the NIC might save
a significant overhead. In order to enable this feature on our hw,
we need to configure it first. Once done, we mark this capability,
to be advertised later to the OS via ieee80211_register_hw.
The driver Rx indications for TCP Checksum is integrated within the
standard Rx status. The driver responds to those indications as follows:
If the frame was tested by hw and checksum ok report CHECKSUM_UNNECESSARY.
Otherwise, report CHECKSUM_NONE.

Signed-off-by: Avri Altman <avri.altman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: pcie: provide a way to stop configuration if it is forbidden
Haim Dreyfuss [Mon, 20 Jul 2015 11:16:21 +0000 (14:16 +0300)]
iwlwifi: pcie: provide a way to stop configuration if it is forbidden

The firmware debug infrastructure allows the user to
provide a firmware that will toggle a few registers to
configure the debugging capabilities.
On certain devices, certain operations are forbidden.
Executing a forbidden operation will cause the hardware to
die in a way that only driver unload / load will bring it
back to life.
Fortunately, there is a way to know in advance if those
operations will be accepted by the device. This is where
the new PRPH_BLOCKBIT operation plays its role. If the bit
X from PRPH register Y is set, then we should prevent any
further register configuration. When that happens, drop a
line in the kernel log since this is really an error state:
the user won't have his device configured as he expected.
Add operations that will be used in the future:
INDIRECT_ASSIGN, INDIRECT_SETBIT, and INDIRECT_CLEARBIT.

Other debugging configurations (such as destination
configuration for the monitor) will take place in any case.

Signed-off-by: Haim Dreyfuss <haim.dreyfuss@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: Use the AP station for non_sta transmit
Ilan Peer [Wed, 24 Jun 2015 06:23:01 +0000 (09:23 +0300)]
iwlwifi: mvm: Use the AP station for non_sta transmit

In iwl_mvm_tx_skb_non_sta(), in case of managed interface,
use the AP station for multicast frames instead of the auxiliary
station as otherwise the frames can be sent to an absent P2P GO as
the FW does not block transmissions for the auxiliary station
since it is not associated with the station MAC context.

Note that this is not possible for unicast frames, as a TDLS
discovery response is sent without a station entry, and in this
case the P2P GO NoA should not block transmission to the peer.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: update comment of power_scheme module parameter
Avri Altman [Sun, 28 Jun 2015 05:10:46 +0000 (08:10 +0300)]
iwlwifi: mvm: update comment of power_scheme module parameter

Signed-off-by: Avri Altman <avri.altman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: return error if d0i3 was aborted
Eliad Peller [Thu, 4 Jun 2015 18:59:32 +0000 (21:59 +0300)]
iwlwifi: return error if d0i3 was aborted

Allow the transport layer to return an error upon suspend.

Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Reviewed-by: Luciano Coelho <luciano.coelho@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoRevert "iwlwifi: mvm: move deferred d0i3 exit to resume_complete op"
Eliad Peller [Wed, 17 Jun 2015 12:32:37 +0000 (15:32 +0300)]
Revert "iwlwifi: mvm: move deferred d0i3 exit to resume_complete op"

This reverts commit 088070a2f6575402d3dd82e1c5a4a8e1941805f6.

When working in d0i3_on_idle mode, we explicitly go out
of d0i3 on resume (so other potential commands could
be sent).

However, D0I3_DEFER_WAKEUP is currently cleared on
resume complete (which happens only later on), causing
d0i3 exit to timeout.

Since mac80211 was modified to accept incoming frames
once drv_resume was called, we can safely revert this
patch, and handle the pending work on iwl_mvm_resume().

Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: move existing UMAC commands to group 1
Johannes Berg [Tue, 30 Jun 2015 13:31:22 +0000 (15:31 +0200)]
iwlwifi: mvm: move existing UMAC commands to group 1

Existing UMAC commands already use the long header, but are sent
with group 0 and the long header inserted manually. Move them to
the group 1 to take advantage of the header building in the low-
level transport.

Existing firmware ignores the group_id field (it's reserved) and
the first firmware that really supports long command headers can
parse all commands in both group 0 (with short header) and group
1 (with long header.)

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: add wide firmware command infrastructure for TX
Aviya Erenfeld [Tue, 9 Jun 2015 13:45:52 +0000 (16:45 +0300)]
iwlwifi: add wide firmware command infrastructure for TX

As the firmware is slowly running out of command IDs and grouping of
commands is desirable anyway, the firmware is extending the command
header from 4 bytes to 8 bytes to introduce a group (in place of the
former flags field, since that's always 0 on commands and thus can
be easily used to distinguish between the two.

In order to support this most easily in the driver widen the command
command ID used in the command sending functions and encode the new
values (group and version) in the ID. That way existing code doesn't
have to be changed (since the higher bits are 0 automatically) and
newer code can easily use the new ID generation function to create a
value to use in place of just the command ID.

Signed-off-by: Aviya Erenfeld <aviya.erenfeld@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: add basic Time of Flight (802.11mc FTM) support
Gregory Greenman [Tue, 2 Jun 2015 15:06:16 +0000 (18:06 +0300)]
iwlwifi: mvm: add basic Time of Flight (802.11mc FTM) support

ToF is a time based method for measurement of the WiFi device
location within a WiFi environment. The driver functionality provided
by this patch is the interface for communication with FW and receiving
location related updates from the FW. The interface provided by this
patch is via debugfs.

Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: remove IWL_UCODE_TLV_API_BASIC_DWELL
Sara Sharon [Sun, 7 Jun 2015 07:36:23 +0000 (10:36 +0300)]
iwlwifi: mvm: remove IWL_UCODE_TLV_API_BASIC_DWELL

All the supported firmwares support this API.
This includes removing dwell per band, as band is no longer a factor
in calculating the dwell. Only basic dwell is used and FW will calculate
the actual dwell time.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: remove command header flags field
Johannes Berg [Tue, 30 Jun 2015 12:45:40 +0000 (14:45 +0200)]
iwlwifi: remove command header flags field

The 'flags' field really has been reserved in the firmware API for a
very long time, probably since 4965. As a consequence, the field is
always 0 and checking for a IWL_CMD_FAILED_MSK flag makes no sense.

Rename the field to 'reserved', get rid of IWL_CMD_FAILED_MSK and
all the code for it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: pcie: don't warn on long MPDUs when supported
Emmanuel Grumbach [Thu, 2 Jul 2015 11:53:02 +0000 (14:53 +0300)]
iwlwifi: pcie: don't warn on long MPDUs when supported

In iwlmvm firmwares, the Byte count written in the scheduler
byte count table is in DWORDs and not in bytes.
We should check that this value fits in the 12 bits and
the value can be either in bits of in DWORD or bytes
depending on the firmware. Check the value after the
translation to DWORDs is done (if needed).

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: pcie: add missing calls to synchronize_irq()
Emmanuel Grumbach [Thu, 25 Jun 2015 09:55:45 +0000 (12:55 +0300)]
iwlwifi: pcie: add missing calls to synchronize_irq()

In a few places, we were disabling interrupts but didn't
make sure that the interrupt handler has finished running.
Add calls to synchronize_irq() to ensure we finish handling
the interrupts before we free resources or other things that
could lead to a crash if the interrupt were to be handled
later.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: pcie: cancel Tx timer upon firmware crash
Emmanuel Grumbach [Wed, 24 Jun 2015 11:58:13 +0000 (14:58 +0300)]
iwlwifi: pcie: cancel Tx timer upon firmware crash

When the firmware crashes, we can't expect the Tx queues to
progress. Cancel their timer.

Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: Do not sample the device time for session protection
Ilan Peer [Wed, 24 Jun 2015 07:25:26 +0000 (10:25 +0300)]
iwlwifi: mvm: Do not sample the device time for session protection

Since the time-event is sent with the immediate flag set, there is
no need to sample the device time.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: remove command and return value from opmode RX
Johannes Berg [Tue, 23 Jun 2015 19:58:17 +0000 (21:58 +0200)]
iwlwifi: remove command and return value from opmode RX

With the previous patch series, no opmode continues using the
command or handler_status (i.e. the return value from the RX)
so it can be removed now.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: remove command/return value from RX handlers
Johannes Berg [Tue, 23 Jun 2015 19:22:09 +0000 (21:22 +0200)]
iwlwifi: mvm: remove command/return value from RX handlers

In the mvm driver, neither the old command nor the return value
are used, so remove them.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: dvm: remove command/return value from RX handlers
Johannes Berg [Tue, 23 Jun 2015 19:22:09 +0000 (21:22 +0200)]
iwlwifi: dvm: remove command/return value from RX handlers

After the previous patches, the command that's passed in nor the
return value are used any more, so can be removed.

While at it, make some functions static.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: dvm: remove ADD_STA prints relying on station ID
Johannes Berg [Tue, 23 Jun 2015 19:01:44 +0000 (21:01 +0200)]
iwlwifi: dvm: remove ADD_STA prints relying on station ID

This makes the logging a little less useful, but as they're mostly
synchronous commands it won't matter much. It gets rid of the
dependency on the input command, which this is the only user of.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: dvm: move ADD_STA response handling to sync command
Johannes Berg [Tue, 23 Jun 2015 18:56:45 +0000 (20:56 +0200)]
iwlwifi: dvm: move ADD_STA response handling to sync command

This driver currently has some very confusing ADD_STA response handling
that runs asynchronously in the background for all of the commands, but
is only really necessary for synchronous ones (the really asynchronous
ones can only be done for already existing stations), and for the sync
ones it actually waits for the RX handler to return a status code.

Rework this to keep the debug printing in the handler, but do the code
that's supposed to have an effect only for sync commands in the command
sending function.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: mvm: LRU-assign key offsets
Johannes Berg [Tue, 16 Jun 2015 15:09:18 +0000 (17:09 +0200)]
iwlwifi: mvm: LRU-assign key offsets

The current key offset assignment algorithm always uses the lowest
unused key offset, which will potentially lead to issues when the
firmware will change to take the key material for TX from the key
table rather than from the TX command.

In order to avoid those issues (and avoid forgetting about them)
change the key offset allocation algorithm now to avoid reusing key
offsets quickly.

The new algorithm always picks as the next offset the least recently
freed offset, i.e. the offset that has been unused for the longest
amount of time. This is implemented by having a generation counter
for each key offset that is incremented every time a key is deleted,
except for the one that's deleted, which is reset to zero. Thus the
highest counter is the key that's been unused longest.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agoiwlwifi: pcie: Set scheduler to work on auto mode
Haim Dreyfuss [Sun, 14 Jun 2015 08:17:07 +0000 (11:17 +0300)]
iwlwifi: pcie: Set scheduler to work on auto mode

During NIC initialization shared HW is reset and this disables the
scheduler. Some HW platforms do not activate the scheduler after it.
Consequently all HCMD sent by the driver stay at the queues which cause
to queue stuck.
Set the scheduler to work on auto active mode so it would be activated upon
change over one of the queues' write pointer.

Signed-off-by: Haim Dreyfuss <haim.dreyfuss@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
9 years agompls: Use definition for reserved label checks
Robert Shearman [Mon, 3 Aug 2015 16:50:04 +0000 (17:50 +0100)]
mpls: Use definition for reserved label checks

In multiple locations there are checks for whether the label in hand
is a reserved label or not using the arbritray value of 16. Factor
this out into a #define for better maintainability and for
documentation.

Signed-off-by: Robert Shearman <rshearma@brocade.com>
Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'lwtunnel-encap-local'
David S. Miller [Tue, 4 Aug 2015 05:26:14 +0000 (22:26 -0700)]
Merge branch 'lwtunnel-encap-local'

Robert Shearman says:

====================
lwtunnel: encap locally-generated ipv4 packets

Locally-generated IPv4 packets, such as from applications running on
the host or traceroute/ping currently don't have lwtunnel output
redirected encap applied. However, they should do in the same way as
for forwarded packets and this patch series addresses that.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: apply lwtunnel encap for locally-generated packets
Robert Shearman [Mon, 3 Aug 2015 16:39:21 +0000 (17:39 +0100)]
ipv4: apply lwtunnel encap for locally-generated packets

lwtunnel encap is applied for forwarded packets, but not for
locally-generated packets. This is because the output function is not
overridden in __mkroute_output, unlike it is in __mkroute_input.

The lwtunnel state is correctly set on the rth through the call to
rt_set_nexthop, so all that needs to be done is to override the dst
output function to be lwtunnel_output if there is lwtunnel state
present and it requires output redirection.

Signed-off-by: Robert Shearman <rshearma@brocade.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agolwtunnel: set skb protocol and dev
Robert Shearman [Mon, 3 Aug 2015 16:39:20 +0000 (17:39 +0100)]
lwtunnel: set skb protocol and dev

In the locally-generated packet path skb->protocol may not be set and
this is required for the lwtunnel encap in order to get the lwtstate.

This would otherwise have been set by ip_output or ip6_output so set
skb->protocol prior to calling the lwtunnel encap
function. Additionally set skb->dev in case it is needed further down
the transmit path.

Signed-off-by: Robert Shearman <rshearma@brocade.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: mdb: fix vlan_enabled access when vlans are not configured
Nikolay Aleksandrov [Mon, 3 Aug 2015 23:19:58 +0000 (01:19 +0200)]
bridge: mdb: fix vlan_enabled access when vlans are not configured

Instead of trying to access br->vlan_enabled directly use the provided
helper br_vlan_enabled().

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoact_bpf: properly support late binding of bpf action to a classifier
Daniel Borkmann [Mon, 3 Aug 2015 14:21:57 +0000 (16:21 +0200)]
act_bpf: properly support late binding of bpf action to a classifier

Since the introduction of the BPF action in d23b8ad8ab23 ("tc: add BPF
based action"), late binding was not working as expected. I.e. setting
the action part for a classifier only via 'bpf index <num>', where <num>
is the index of an existing action, is being rejected by the kernel due
to other missing parameters.

It doesn't make sense to require these parameters such as BPF opcodes
etc, as they are not going to be used anyway: in this case, they're just
allocated/parsed and then freed again w/o doing anything meaningful.

Instead, parse and verify the remaining parameters *after* the test on
tcf_hash_check(), when we really know that we're dealing with creation
of a new action or replacement of an existing one and where late binding
is thus irrelevant.

After patch, test case is now working:

  FOO="1,6 0 0 4294967295,"
  tc actions add action bpf bytecode "$FOO"
  tc filter add dev foo parent 1: bpf bytecode "$FOO" flowid 1:1 action bpf index 1
  tc actions show action bpf
    action order 0: bpf bytecode '1,6 0 0 4294967295' default-action pipe
    index 1 ref 2 bind 1
  tc filter show dev foo
    filter protocol all pref 49152 bpf
    filter protocol all pref 49152 bpf handle 0x1 flowid 1:1 bytecode '1,6 0 0 4294967295'
    action order 1: bpf bytecode '1,6 0 0 4294967295' default-action pipe
    index 1 ref 2 bind 1

Late binding of a BPF action can be useful for preloading maps (e.g. before
they hit traffic) in case of eBPF programs, or to share a single eBPF action
with multiple classifiers.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: dsa: mv88e6xxx: call _mv88e6xxx_stats_wait with SMI lock held
Vivien Didelot [Mon, 3 Aug 2015 13:17:44 +0000 (09:17 -0400)]
net: dsa: mv88e6xxx: call _mv88e6xxx_stats_wait with SMI lock held

At switch setup, _mv88e6xxx_stats_wait was called without holding the
SMI mutex. Fix this by requesting the lock for this call.

Also, return the _mv88e6xxx_stats_wait code, since it may fail.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: mdb: add/del entry on all vlans if vlan_filter is enabled and vid is 0
Satish Ashok [Mon, 3 Aug 2015 11:29:16 +0000 (13:29 +0200)]
bridge: mdb: add/del entry on all vlans if vlan_filter is enabled and vid is 0

Before this patch when a vid was not specified, the entry was added with
vid 0 which is useless when vlan_filtering is enabled. This patch makes
the entry to be added on all configured vlans when vlan filtering is
enabled and respectively deleted from all, if the entry vid is 0.
This is also closer to the way fdb works with regard to vid 0 and vlan
filtering.

Example:
Setup:
$ bridge vlan add vid 256 dev eth4
$ bridge vlan add vid 1024 dev eth4
$ bridge vlan add vid 64 dev eth3
$ bridge vlan add vid 128 dev eth3
$ bridge vlan
port vlan ids
eth3  1 PVID Egress Untagged
 64
 128

eth4  1 PVID Egress Untagged
 256
 1024
$ echo 1 > /sys/class/net/br0/bridge/vlan_filtering

Before:
$ bridge mdb add dev br0 port eth3 grp 239.0.0.1
$ bridge mdb
dev br0 port eth3 grp 239.0.0.1 temp

After:
$ bridge mdb add dev br0 port eth3 grp 239.0.0.1
$ bridge mdb
dev br0 port eth3 grp 239.0.0.1 temp vid 1
dev br0 port eth3 grp 239.0.0.1 temp vid 128
dev br0 port eth3 grp 239.0.0.1 temp vid 64

Signed-off-by: Satish Ashok <sashok@cumulusnetworks.com>
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'stacked-vlan-TSO'
David S. Miller [Mon, 3 Aug 2015 21:24:50 +0000 (14:24 -0700)]
Merge branch 'stacked-vlan-TSO'

Toshiaki Makita says:

====================
Stacked vlan TSO for virtual devices

Basically virtual devices do not need to segment double tagged packets.
This patch set adds TSO feature for double tagged packets to several
virtual devices, which can be realized by simply setting
.ndo_features_check to passthru_features_check.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agotuntap: Don't segment multiple tagged packets on tap device
Toshiaki Makita [Fri, 31 Jul 2015 06:03:27 +0000 (15:03 +0900)]
tuntap: Don't segment multiple tagged packets on tap device

Tap devices don't need to segment multiple tagged packets.

Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: Don't segment multiple tagged packets on bridge device
Toshiaki Makita [Fri, 31 Jul 2015 06:03:26 +0000 (15:03 +0900)]
bridge: Don't segment multiple tagged packets on bridge device

Bridge devices don't need to segment multiple tagged packets since thier
ports can segment them.

Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoveth: Don't segment multiple tagged packets on veth device
Toshiaki Makita [Fri, 31 Jul 2015 06:03:25 +0000 (15:03 +0900)]
veth: Don't segment multiple tagged packets on veth device

Veth devices don't need to segment multiple tagged packets.

Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agomacvlan: Don't segment multiple tagged packets on macvlan device
Toshiaki Makita [Fri, 31 Jul 2015 06:03:24 +0000 (15:03 +0900)]
macvlan: Don't segment multiple tagged packets on macvlan device

Macvlan/macvtap devices don't need to segment multiple tagged packets
since the lower devices can segment them.

Signed-off-by: Toshiaki Makita <makita.toshiaki@lab.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agovirtio_net: add gro capability
Eric Dumazet [Fri, 31 Jul 2015 16:25:17 +0000 (18:25 +0200)]
virtio_net: add gro capability

Straightforward patch to add GRO processing to virtio_net.

napi_complete_done() usage allows more aggressive aggregation,
opted-in by setting /sys/class/net/xxx/gro_flush_timeout

Tested:

Setting /sys/class/net/xxx/gro_flush_timeout to 1000 nsec,
Rick Jones reported following results.

One VM of each on a pair of OpenStack compute nodes with E5-2650Lv3 CPUs
and Intel 82599ES-based NICs. So, two "before" and two "after" VMs.
The OpenStack compute nodes were running OpenStack Kilo, with VxLAN
encapsulation being used through OVS so no GRO coming-up the host
stack.  The compute nodes themselves were running a 3.14-based kernel.

Single-stream netperf, CPU utilizations and thus service demands are
based on intra-guest reported CPU.

Throughput Mbit/s, bigger is better
        Min     Median  Average Max
4.2.0-rc3+      1364    1686    1678    1938
4.2.0-rc3+flush1k       1824    2269    2275    2647

Send Service Demand, smaller is better
        Min     Median  Average Max
4.2.0-rc3+      0.236   0.558   0.524   0.802
4.2.0-rc3+flush1k       0.176   0.503   0.471   0.738

Receive Service Demand, smaller is better.
        Min     Median  Average Max
4.2.0-rc3+      1.906   2.188   2.191   2.531
4.2.0-rc3+flush1k       0.448   0.529   0.533   0.692

Signed-off-by: Eric Dumazet <edumazet@google.com>
Tested-by: Rick Jones <rick.jones2@hp.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: linearize skb in case frags would not fit into tx descriptor
Jiri Pirko [Sun, 2 Aug 2015 18:56:38 +0000 (20:56 +0200)]
rocker: linearize skb in case frags would not fit into tx descriptor

Suggested-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agorocker: enable support for scattered packets
Ido Schimmel [Sun, 2 Aug 2015 18:56:37 +0000 (20:56 +0200)]
rocker: enable support for scattered packets

rocker supports the transmission of scattered packets, so let the kernel
know about it by setting the NETIF_F_SG bit in the device's features.

Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Scott Feldman <sfeldma@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoebpf: add skb->hash to offset map for usage in {cls, act}_bpf or filters
Daniel Borkmann [Fri, 31 Jul 2015 22:46:29 +0000 (00:46 +0200)]
ebpf: add skb->hash to offset map for usage in {cls, act}_bpf or filters

Add skb->hash to the __sk_buff offset map, so it can be accessed from
an eBPF program. We currently already do this for classic BPF filters,
but not yet on eBPF, it might be useful as a demuxer in combination with
helpers like bpf_clone_redirect(), toy example:

  __section("cls-lb") int ingress_main(struct __sk_buff *skb)
  {
    unsigned int which = 3 + (skb->hash & 7);
    /* bpf_skb_store_bytes(skb, ...); */
    /* bpf_l{3,4}_csum_replace(skb, ...); */
    bpf_clone_redirect(skb, which, 0);
    return -1;
  }

I was thinking whether to add skb_get_hash(), but then concluded the
raw skb->hash seems fine in this case: we can directly access the hash
w/o extra eBPF helper function call, it's filled out by many NICs on
ingress, and in case the entropy level would not be sufficient, people
can still implement their own specific sw fallback hash mix anyway.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobnx2x: Correct logic for pvid configuration.
Yuval Mintz [Sun, 2 Aug 2015 09:42:41 +0000 (12:42 +0300)]
bnx2x: Correct logic for pvid configuration.

Commit 05cc5a39ddb7 ("bnx2x: add vlan filtering offload") has introduced
an incorrect logic for checking whether pvid should be configured for
a vf, causing the hypervisor driver to send unneeded ramrods for all of
the vfs each time a pvid has changed.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Yuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: Ariel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
David S. Miller [Sat, 1 Aug 2015 06:52:20 +0000 (23:52 -0700)]
Merge git://git./linux/kernel/git/davem/net

Conflicts:
arch/s390/net/bpf_jit_comp.c
drivers/net/ethernet/ti/netcp_ethss.c
net/bridge/br_multicast.c
net/ipv4/ip_fragment.c

All four conflicts were cases of simple overlapping
changes.

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Sat, 1 Aug 2015 00:10:56 +0000 (17:10 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Must teardown SR-IOV before unregistering netdev in igb driver, from
    Alex Williamson.

 2) Fix ipv6 route unreachable crash in IPVS, from Alex Gartrell.

 3) Default route selection in ipv4 should take the prefix length, table
    ID, and TOS into account, from Julian Anastasov.

 4) sch_plug must have a reset method in order to purge all buffered
    packets when the qdisc is reset, likewise for sch_choke, from WANG
    Cong.

 5) Fix deadlock and races in slave_changelink/br_setport in bridging.
    From Nikolay Aleksandrov.

 6) mlx4 bug fixes (wrong index in port even propagation to VFs,
    overzealous BUG_ON assertion, etc.) from Ido Shamay, Jack
    Morgenstein, and Or Gerlitz.

 7) Turn off klog message about SCTP userspace interface compat that
    makes no sense at all, from Daniel Borkmann.

 8) Fix unbounded restarts of inet frag eviction process, causing NMI
    watchdog soft lockup messages, from Florian Westphal.

 9) Suspend/resume fixes for r8152 from Hayes Wang.

10) Fix busy loop when MSG_WAITALL|MSG_PEEK is used in TCP recv, from
    Sabrina Dubroca.

11) Fix performance regression when removing a lot of routes from the
    ipv4 routing tables, from Alexander Duyck.

12) Fix device leak in AF_PACKET, from Lars Westerhoff.

13) AF_PACKET also has a header length comparison bug due to signedness,
    from Alexander Drozdov.

14) Fix bug in EBPF tail call generation on x86, from Daniel Borkmann.

15) Memory leaks, TSO stats, watchdog timeout and other fixes to
    thunderx driver from Sunil Goutham and Thanneeru Srinivasulu.

16) act_bpf can leak memory when replacing programs, from Daniel
    Borkmann.

17) WOL packet fixes in gianfar driver, from Claudiu Manoil.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (79 commits)
  stmmac: fix missing MODULE_LICENSE in stmmac_platform
  gianfar: Enable device wakeup when appropriate
  gianfar: Fix suspend/resume for wol magic packet
  gianfar: Fix warning when CONFIG_PM off
  act_pedit: check binding before calling tcf_hash_release()
  net: sk_clone_lock() should only do get_net() if the parent is not a kernel socket
  net: sched: fix refcount imbalance in actions
  r8152: reset device when tx timeout
  r8152: add pre_reset and post_reset
  qlcnic: Fix corruption while copying
  act_bpf: fix memory leaks when replacing bpf programs
  net: thunderx: Fix for crash while BGX teardown
  net: thunderx: Add PCI driver shutdown routine
  net: thunderx: Fix crash when changing rss with mutliple traffic flows
  net: thunderx: Set watchdog timeout value
  net: thunderx: Wakeup TXQ only if CQE_TX are processed
  net: thunderx: Suppress alloc_pages() failure warnings
  net: thunderx: Fix TSO packet statistic
  net: thunderx: Fix memory leak when changing queue count
  net: thunderx: Fix RQ_DROP miscalculation
  ...

9 years agoMerge branch 'ipv6-auto-flow-labels'
David S. Miller [Sat, 1 Aug 2015 00:07:12 +0000 (17:07 -0700)]
Merge branch 'ipv6-auto-flow-labels'

Tom Herbert says:

====================
ipv6: Turn on auto IPv6 flow labels by default

BSD (MacOS) has already turned on flow labels by default and this does
not seem to be causing any problems in the Internet. Let's go ahead
and turn them on by default. We'll continue to monitor for any devices
start choking on them.

Flow labels are important since they are the desired solution for
network devices to perform ECMP and RSS (RFC6437 and RFC6438).
Traditionally, devices perform a 5-tuple hash on packets that
includes port numbers. For the most part, these devices can only
compute 5-tuple hashes for TCP and UDP. This severely limits our ability
to get good network load balancing for other protocols (IPIP, GRE,ESP,
etc.), and hence we are limited in using other protocols. Unfortunately,
this method is accepted as the de facto standard to the extent that
there are several proposals to encapsulate protocols in UDP _just_ for
the purposes for getting ECMP to work. With hosts generating flow labels
and devices taking them as input into ECMP (several already do), we can
start to fix this fundamental problem.

This patch set:
 - Changes IPV6_FLOWINFO sockopt to be opt-out of flow labels for
   connections rather than opt-in
 - Disable flow label state ranges sysctl by default
 - Enable auto flow labels sysctl by default

v2:
  - Added functions to create an skb->hash based on flowi4 and flowi6.
    These are called in output path when creating a packet
  - Call skb_get_hash_flowi6 in ip6_make_flowlabel
  - Implement the auto_flowlabels sysctl as a mode for auto flowlabels.
    There are four modes which correspond to flow labels being enabled
    and whether socket option can be used to opt in or opt out of
    using them
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Enable auto flow labels by default
Tom Herbert [Fri, 31 Jul 2015 23:52:14 +0000 (16:52 -0700)]
ipv6: Enable auto flow labels by default

Initialize auto_flowlabels to one. This enables automatic flow labels,
individual socket may disable them using the IPV6_AUTOFLOWLABEL socket
option.

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Disable flowlabel state ranges by default
Tom Herbert [Fri, 31 Jul 2015 23:52:13 +0000 (16:52 -0700)]
ipv6: Disable flowlabel state ranges by default

Per RFC6437 stateful flow labels (e.g. labels set by flow label manager)
cannot "disturb" nodes taking part in stateless flow labels. While the
ranges only reduce the flow label entropy by one bit, it is conceivable
that this might bias the algorithm on some routers causing a load
imbalance. For best results on the Internet we really need the full
20 bits.

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Implement different admin modes for automatic flow labels
Tom Herbert [Fri, 31 Jul 2015 23:52:12 +0000 (16:52 -0700)]
ipv6: Implement different admin modes for automatic flow labels

Change the meaning of net.ipv6.auto_flowlabels to provide a mode for
automatic flow labels generation. There are four modes:

0: flow labels are disabled
1: flow labels are enabled, sockets can opt-out
2: flow labels are allowed, sockets can opt-in
3: flow labels are enabled and enforced, no opt-out for sockets

np->autoflowlabel is initialized according to the sysctl value.

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Call skb_get_hash_flowi6 to get skb->hash in ip6_make_flowlabel
Tom Herbert [Fri, 31 Jul 2015 23:52:11 +0000 (16:52 -0700)]
ipv6: Call skb_get_hash_flowi6 to get skb->hash in ip6_make_flowlabel

We can't call skb_get_hash here since the packet is not complete to do
flow_dissector. Create hash based on flowi6 instead.

Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>