openwrt/staging/blogic.git
9 years agoiwlwifi: mvm: add minimal multi-RXQ infrastructure
Johannes Berg [Fri, 22 May 2015 11:41:07 +0000 (13:41 +0200)]
iwlwifi: mvm: add minimal multi-RXQ infrastructure

Since the new multi-queue capability depends on a new firmware API,
we can already add some code for it. If the new API is present, a
new opmode ops struct is used that handles the new rx_rss method.

For now, only restructure the RX handling to distinguish between
the two. Future patches will convert the new infrastructure to
actually use the new RX descriptor layout.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: size firmware flags memory correctly
Johannes Berg [Wed, 16 Sep 2015 10:21:32 +0000 (12:21 +0200)]
iwlwifi: size firmware flags memory correctly

Instead of relying on a hard-coded constant of a maximum of 64 API and
capability bits, add a new enum value after the others that will then
always track the number of used bits in the API/capabilities. We thus
no longer need to maintain the maximum number, and on 32-bit platforms
even (currently) reduce the number of bits kept in memory.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: make threshold temperatures unsigned
Johannes Berg [Thu, 24 Sep 2015 20:19:43 +0000 (22:19 +0200)]
iwlwifi: mvm: make threshold temperatures unsigned

There's no need to have negative threshold temperatures, so make
them unsigned to avoid signedness warnings in debugfs.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: nvm: add nvm phy_sku section to debugfs
Moshe Harel [Wed, 2 Sep 2015 09:45:12 +0000 (12:45 +0300)]
iwlwifi: nvm: add nvm phy_sku section to debugfs

The only NVM section not captured in debugfs.

Signed-off-by: Moshe Harel <moshe.harel@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: fix signedness warnings in ToF debugfs
Johannes Berg [Thu, 24 Sep 2015 20:15:02 +0000 (22:15 +0200)]
iwlwifi: mvm: fix signedness warnings in ToF debugfs

Using an int* instead of u32* as the kstrtou32() output argument
obviously results in signedness warnings, change that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: rs: dynamically switch between 80MHz and 20MHz in some scenarios
Eyal Shapira [Sat, 19 Sep 2015 20:48:45 +0000 (23:48 +0300)]
iwlwifi: mvm: rs: dynamically switch between 80MHz and 20MHz in some scenarios

This is a tweak which has been shown to improve performance when
moving away from the AP while working in 80Mhz.
When RS decides to go down to 80MHz SISO MCS0 instead switch to 20MHz MCS4.
Go back to 80MHz MCS1 if RS can sustain 20MHz MCS5.

Signed-off-by: Eyal Shapira <eyalx.shapira@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: minor rx code cleanup
Johannes Berg [Tue, 22 Sep 2015 10:28:42 +0000 (12:28 +0200)]
iwlwifi: mvm: minor rx code cleanup

Clean up variable initialisation slightly.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: remove IWL3165_UCODE_API_OK and _MIN
Johannes Berg [Tue, 22 Sep 2015 08:51:17 +0000 (10:51 +0200)]
iwlwifi: remove IWL3165_UCODE_API_OK and _MIN

As the 3165 device uses the same firmware as 7265-D and currently
all 7000 series (including 3160/3165) use the same API versions
remove IWL3165_UCODE_API_OK and _MIN. We might have to put them
back if firmware support ever splits, but in that case might also
have to add a different MODULE_FIRMWARE statement.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: rs: fix success ratio comparison in rs_get_best_rate
Eyal Shapira [Mon, 21 Sep 2015 17:03:22 +0000 (20:03 +0300)]
iwlwifi: mvm: rs: fix success ratio comparison in rs_get_best_rate

success_ratio is actually 128 * SR in percentage while
IWL_MVM_RS_SR_NO_DECREASE is 85%. Fix this by using RS_PERCENT().
This bug caused the if branch to be always executed. This in turn
led to always selecting a rate, following a column switch, in which
the expected throughput would exceed the best expected current throughput.
In some scenarios where the success ratio isn't >85% such a rate
could be too aggressive leading us to avoid the new column.
This has the potential of causing sub optimal performance.

Reported-by: Moshe Harel <moshe.harel@intel.com>
Signed-off-by: Eyal Shapira <eyalx.shapira@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: rs: minor indentation fix
Eyal Shapira [Sat, 19 Sep 2015 20:54:38 +0000 (23:54 +0300)]
iwlwifi: mvm: rs: minor indentation fix

Indentation was off a bit. Fix it.

Signed-off-by: Eyal Shapira <eyalx.shapira@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: rs: remove overflowing debug message
Eyal Shapira [Sat, 19 Sep 2015 20:46:32 +0000 (23:46 +0300)]
iwlwifi: mvm: rs: remove overflowing debug message

This message isn't very useful and creates clutter.
Remove it.

Signed-off-by: Eyal Shapira <eyalx.shapira@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: rs: improve rate debug messages
Eyal Shapira [Sat, 19 Sep 2015 20:37:45 +0000 (23:37 +0300)]
iwlwifi: mvm: rs: improve rate debug messages

Pretty print the rate full details to ease debugging.

Signed-off-by: Eyal Shapira <eyalx.shapira@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: stop using DEVICE_POWER_FLAGS_CAM_MSK
Johannes Berg [Fri, 4 Sep 2015 09:27:25 +0000 (11:27 +0200)]
iwlwifi: mvm: stop using DEVICE_POWER_FLAGS_CAM_MSK

The firmware has always treated these two bits to mean that
powersave is enabled when POWER_SAVE_ENA is set and CAM is
clear; it doesn't use them in any non-combined way.

Therefore, it's pointless to send it two bits, and the API
should be cleaned up. Prepare the driver by removing the CAM
bit and using only POWER_SAVE_ENA to indicate whether PS is
enabled or not.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: support enabling a queue with a given ssn
Liad Kaufman [Wed, 22 Jul 2015 14:59:53 +0000 (17:59 +0300)]
iwlwifi: mvm: support enabling a queue with a given ssn

When enabling a queue, the default SSN is 0.

Allow determining what that SSN should be, if required. This
can happen, for example, if a queue gets reconfigured.

Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: support using multiple ACs on single HW queue
Liad Kaufman [Tue, 14 Jul 2015 10:36:18 +0000 (13:36 +0300)]
iwlwifi: mvm: support using multiple ACs on single HW queue

"DQA" is shorthand for "dynamic queue allocation", with the
idea of allocating queues per-RA/TID on-demand rather than
using shared queues statically allocated per vif. The goal
of this is to enable future features (like GO PM) and to
improve performance measurements of TX traffic.

When RA/TID streams can't be neatly sorted into different AC
queues, DQA allows sharing queues for the same RA. This means
that DQA allows different ACs may reach the same HW queue.

Update the code to allow such queue sharing by having a mapping
between the HW queue and the mac80211 queues using it (as this
could be more than one queue).

Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: transport: track number of allocated queues
Johannes Berg [Fri, 22 May 2015 09:33:12 +0000 (11:33 +0200)]
iwlwifi: transport: track number of allocated queues

As the transport will decide how many queues (and MSI-X vectors)
to allocate, add a field to indicate that to the op-mode so it
can size/allocate its own data structures appropriately.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: op-mode API: add rx_rss method
Johannes Berg [Fri, 22 May 2015 08:52:26 +0000 (10:52 +0200)]
iwlwifi: op-mode API: add rx_rss method

Upcoming hardware will have the ability to do L3 hashing for RSS,
directing data packets (and perhaps some associated metadata and
management notifications) to different MSI-X vectors.

In this case, it makes no sense to go through the full RX dispatch
since it's already known that only a subset of the possibilities
can come in, requiring a new receive method. In addition this must
know which queue the packet was received on.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove PHY RX from handlers
Johannes Berg [Thu, 3 Sep 2015 12:53:09 +0000 (14:53 +0200)]
iwlwifi: mvm: remove PHY RX from handlers

Treat PHY RX specially, since it's actually pretty frequent,
doesn't need all the notication etc. code, and will have a
different handler in future hardware.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: Improve debugfs tof robustness
Assaf Krauss [Wed, 16 Sep 2015 08:47:14 +0000 (11:47 +0300)]
iwlwifi: mvm: Improve debugfs tof robustness

Return a proper error when wrong parameters are passed to debugfs
tof_range_request.

Signed-off-by: Assaf Krauss <assaf.krauss@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: ToF - fill bssid of responder configuration
Gregory Greenman [Mon, 20 Jul 2015 04:44:44 +0000 (07:44 +0300)]
iwlwifi: mvm: ToF - fill bssid of responder configuration

The command needs to have the AP interfaces BSSID (which corresponds
to its address).

Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: Fix tof debugfs formats (dec vs. hex)
Assaf Krauss [Wed, 16 Sep 2015 08:44:55 +0000 (11:44 +0300)]
iwlwifi: mvm: Fix tof debugfs formats (dec vs. hex)

Make some input formats more natural, e.g. bandwidth and periods
are more natural in decimal than in hexadecimal.

Signed-off-by: Assaf Krauss <assaf.krauss@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: fix tof.h header guard
Nicolas Iooss [Sat, 12 Sep 2015 10:04:51 +0000 (12:04 +0200)]
iwlwifi: mvm: fix tof.h header guard

Commit ce7929186a39 ("iwlwifi: mvm: add basic Time of Flight (802.11mc
FTM) support") created drivers/net/wireless/iwlwifi/mvm/tof.h with a
broken header guard:

    #ifndef __tof
    #define __tof_h__

    ...

    #endif /* __tof_h__ */

Use __tof_h__ in the first line.

Signed-off-by: Nicolas Iooss <nicolas.iooss_linux@m4x.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: Correctly update MAC context on add/del station
Ilan Peer [Thu, 10 Sep 2015 09:54:38 +0000 (12:54 +0300)]
iwlwifi: mvm: Correctly update MAC context on add/del station

Commit "iwlwifi: mvm: don't ask beacons when AP vif and no
assoc sta" directly called iwl_mvm_mac_ctxt_cmd_ap() to update the
MAC context when adding/removing a station. However, this ignores
the case that the vif is actually a P2P GO.

Fix this by calling iwl_mvm_mac_ctxt_changed() that handles P2P GO
case as well.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: fix default disabled aggs in sta
Liad Kaufman [Tue, 1 Sep 2015 15:50:22 +0000 (18:50 +0300)]
iwlwifi: mvm: fix default disabled aggs in sta

For the ADD_STA command, when the flag for aggregation
disabling is set, there is a bitmap indicated what TIDs
are disabling aggregations and what aren't. Currently, by
default, all TIDs allow for aggregations since the value
we begin with is 0.

Change this default value to 0xffff so all TIDs don't
allow aggregations until explicitly turned on.

Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: enable tracing by default
Johannes Berg [Fri, 4 Sep 2015 18:54:35 +0000 (20:54 +0200)]
iwlwifi: enable tracing by default

Tracing, if disabled at runtime, has very low overhead with
great returns on debugging. It therefore makes sense to have
it enabled by default (if the kernel enables EVENT_TRACING).

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoMAINTAINERS: iwlwifi: update contact email
Johannes Berg [Tue, 8 Sep 2015 15:03:20 +0000 (17:03 +0200)]
MAINTAINERS: iwlwifi: update contact email

The ilw@linux.intel.com address is being phased out, replace
it with the new address.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: correct skip-over-DTIM implementation
Johannes Berg [Tue, 1 Sep 2015 12:15:17 +0000 (14:15 +0200)]
iwlwifi: mvm: correct skip-over-DTIM implementation

The formula used in D0i3 should also be used in D3, instead of
the hardcoded value.

Additionally, the formula is actually wrong - if the calculation
yields 0 then 1 should be used instead of disabling entirely.
Also need to add 1 since the firmware needs 3 to skip 2, etc.

To make all this clearer, centralize the calculation into a
single function.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: nvm: force 1x1 antenna in Series 8000
Moshe Harel [Sun, 6 Sep 2015 13:50:33 +0000 (16:50 +0300)]
iwlwifi: nvm: force 1x1 antenna in Series 8000

This is a workaround to an OTP bug. In Series 8000 1x1, the OTP
0xA052 defines 2x2 antenna configuration. This workaround overrides
the decision based on HW id and MIMO disabled bit which is correct
in the OTP and set to disabled.

Signed-off-by: Moshe Harel <moshe.harel@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: move DTS command and notification to new group
Aviya Erenfeld [Tue, 1 Sep 2015 16:34:38 +0000 (19:34 +0300)]
iwlwifi: mvm: move DTS command and notification to new group

Move the DTS measurement command and notification from short
command header to the new PHY command group for firmware
supporting the extended command headers.

Signed-off-by: Aviya Erenfeld <aviya.erenfeld@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoMerge tag 'iwlwifi-next-for-kalle-2015-09-21' of git://git.kernel.org/pub/scm/linux...
Kalle Valo [Sat, 26 Sep 2015 18:25:18 +0000 (21:25 +0300)]
Merge tag 'iwlwifi-next-for-kalle-2015-09-21' of git://git./linux/kernel/git/iwlwifi/iwlwifi-next

* some debugfs improvements;
* fix signedness in beacon statistics;
* deinline some functions to reduce size when device tracing is enabled;
* filter beacons out in AP mode when no stations are associated;
* deprecate firmwares version -12;
* fix a runtime PM vs. legacy suspend race;
* one-liner fix for a ToF bug;
* clean-ups in the rx code;
* small debugging improvement;
* fix WoWLAN with new firmware versions;

9 years agoiwlwifi: mvm: add debug print for d0i3 exit indication
Eliad Peller [Sun, 6 Sep 2015 11:17:17 +0000 (14:17 +0300)]
iwlwifi: mvm: add debug print for d0i3 exit indication

In order to verify d0i3 flow, add debug print to indicate
d0i3 exit was completed (right after tx was re-enabled),
along with the wakeup reasons.

Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: configure wowlan configuration only if connected
Eliad Peller [Tue, 1 Sep 2015 11:16:00 +0000 (14:16 +0300)]
iwlwifi: mvm: configure wowlan configuration only if connected

Recent fw version added assert to make sure wowlan configuration
is configured only when a station is connected.

Change the driver behavior to pass this configuration only
if we indeed have ap station id (i.e. connected).

Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: move RX API into its own file
Johannes Berg [Wed, 2 Sep 2015 12:53:39 +0000 (14:53 +0200)]
iwlwifi: mvm: move RX API into its own file

The RX API is currently mixed up into the general fw-api.h
file, but we're going to need to extend it significantly in
the future, so move it to its own file.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove some unused defines from RX API
Johannes Berg [Wed, 2 Sep 2015 12:47:16 +0000 (14:47 +0200)]
iwlwifi: mvm: remove some unused defines from RX API

Remove some unused values from the RX API; these were used
with older firmware API that didn't have the RX energy API,
support for which was removed a long time ago.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove pointless cfg_phy_cnt length check
Johannes Berg [Wed, 2 Sep 2015 10:26:45 +0000 (12:26 +0200)]
iwlwifi: mvm: remove pointless cfg_phy_cnt length check

Since the driver can never configure the data here, this field
will always be reported as 0 by the firmware. Even if this was
not the case, however, it wouldn't matter since the extra data
would be added beyond the end of the phy_info structure we use
in the driver, so wouldn't harm anything in this code either.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove useless debug message from RX
Johannes Berg [Wed, 2 Sep 2015 09:51:23 +0000 (11:51 +0200)]
iwlwifi: mvm: remove useless debug message from RX

This message is useless - it's in the good case that always
happens so enabling it doesn't really help. Just remove it.
There are other ways to debug this (e.g. tracing) so there's
no need to add a message in the bad case.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: make sure AP is operating for ToF
Johannes Berg [Wed, 26 Aug 2015 13:55:04 +0000 (15:55 +0200)]
iwlwifi: mvm: make sure AP is operating for ToF

It's possible for an AP interface to be UP but not actually
operating (i.e. not beaconing etc.) - in this case it can't
actually do ToF, so check for it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove IWL_UCODE_TLV_API_STATS_V10 TLV flag
Emmanuel Grumbach [Wed, 26 Aug 2015 11:03:02 +0000 (14:03 +0300)]
iwlwifi: mvm: remove IWL_UCODE_TLV_API_STATS_V10 TLV flag

This flag is set in all supported firmwares.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luciano Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove IWL_UCODE_TLV_API_ASYNC_DTM TLV flag
Emmanuel Grumbach [Wed, 26 Aug 2015 10:57:37 +0000 (13:57 +0300)]
iwlwifi: mvm: remove IWL_UCODE_TLV_API_ASYNC_DTM TLV flag

This flag is set in all supported firmwares.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luciano Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove IWL_UCODE_TLV_API_SINGLE_SCAN_EBS TLV flag
Emmanuel Grumbach [Wed, 26 Aug 2015 10:55:11 +0000 (13:55 +0300)]
iwlwifi: mvm: remove IWL_UCODE_TLV_API_SINGLE_SCAN_EBS TLV flag

All the supported firmwares have this flag set.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luciano Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove IWL_UCODE_TLV_API_TX_POWER_DEV TLV flag
Emmanuel Grumbach [Wed, 26 Aug 2015 10:52:54 +0000 (13:52 +0300)]
iwlwifi: mvm: remove IWL_UCODE_TLV_API_TX_POWER_DEV TLV flag

All the supported firmwares use the new API.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luciano Coelho <luciano.coelho@intel.com>
9 years agoiwlwifi: mvm: remove IWL_UCODE_TLV_API_HDC_PHASE_0 TLV flag
Emmanuel Grumbach [Wed, 26 Aug 2015 10:50:59 +0000 (13:50 +0300)]
iwlwifi: mvm: remove IWL_UCODE_TLV_API_HDC_PHASE_0 TLV flag

All the supported firwmares support the new API.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luciano Coelho <luciano.coelho@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
9 years agoath9k_htc: introduce support for different fw versions
Oleksij Rempel [Sun, 6 Sep 2015 11:09:01 +0000 (13:09 +0200)]
ath9k_htc: introduce support for different fw versions

Current kernel support only one fw name with theoretically only one
fw version located in “firmware/htc_[9271|7010].fw”. Which is ok so far we
have only one fw version (1.3). After we realised new fw 1.4, we faced
compatibility problem which was decided to solve by firmware name and
location:
- new firmware is located now in
firmware/ath9k_htc/htc_[9271|7010]-1.4.0.fw
- old version 1.3 should be on old place, so old kernel have no issues
with it.
- new kernels including this patch should be able to try different
supported (min..max) fw version.
- new kernel should be able to support old fw location too. At least for
now.

At same time this patch will add new module option which should allow user
to play with development  fw version without replacing stable one. If user
will set “ath9k_htc use_dev_fw=1” module will try to find
firmware/ath9k_htc/htc_[9271|7010]-1.dev.0.fw first and if it fails, use
stable version: for example...1.4.0.fw.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
9 years agosch_dsmark: improve memory locality
Eric Dumazet [Thu, 17 Sep 2015 23:37:13 +0000 (16:37 -0700)]
sch_dsmark: improve memory locality

Memory placement in sch_dsmark is silly : Better place mask/value
in the same cache line.

Also, we can embed small arrays in the first cache line and
remove a potential cache miss.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'bcmgenet-irq-coalesce'
David S. Miller [Fri, 18 Sep 2015 05:17:14 +0000 (22:17 -0700)]
Merge branch 'bcmgenet-irq-coalesce'

Florian Fainelli says:

====================
net: bcmgenet: Interrupt coalescing

This patch series adds support for interrupt coalescing for GENET
adapters.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: bcmgenet: Implement RX coalescing control knobs
Florian Fainelli [Wed, 16 Sep 2015 23:47:40 +0000 (16:47 -0700)]
net: bcmgenet: Implement RX coalescing control knobs

Add support for the ethtool rx-frames coalescing parameter which allows
defining the number of RX interrupts per frames received. The RDMA
engine supports a configurable timeout with a resolution of
approximately 8.192 us.

We can no longer enable the BDONE/PDONE interrupts as those would
fire for each packet/buffer received, which would defeat the MBDONE
interrupt purpose. The MBDONE interrupt is guaranteed to correspond to a
PDONE/BDONE interrupt when the threshold is set to 1.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: bcmgenet: Implement TX coalescing control knobs
Florian Fainelli [Wed, 16 Sep 2015 23:47:39 +0000 (16:47 -0700)]
net: bcmgenet: Implement TX coalescing control knobs

Configuring the ethtool tx-frames property, which translates into N
packets before a TX interrupt is the simplest configuration scheme
because it requires no locking neither at the softare nor hardware
level, and is completely indepedent from the link speed. Since ethtool
does not allow per-tx queue coalescing parameters, we apply the same
setting to any transmit queue.

We can no longer enable the BDONE/PDONE interrupts as those would fire
for each packet/buffer received, which would defeat the MBDONE interrupt
purpose. The MBDONE interrupt is guaranteed to correspond to a
PDONE/BDONE interrupt when the threshold is set to 1, but offers
interrupt coalescing when the value is > 1.

Since the HW is configured to generate an interrupt when the ring
becomes emtpy, we have to deny any timeout/timer settings coming from
user-space to indicate we can only generate an interrupt very <N>
packets.

While we are at it, fix the DMA_INTR_THRESHOLD_MASK value which was off
by one bit (0xff vs. 0x1ff).

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agolan78xx: Remove not defined MAC_CR_GMII_EN_ bit from MAC_CR.
Woojung.Huh@microchip.com [Wed, 16 Sep 2015 23:41:19 +0000 (23:41 +0000)]
lan78xx: Remove not defined MAC_CR_GMII_EN_ bit from MAC_CR.

Remove not defined MAC_CR_GMII_EN_ bit from MAC_CR.

Signed-off-by: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agolan78xx: Create lan78xx_get_mdix_status() and lan78xx_set_mdix_status() for MDIX...
Woojung.Huh@microchip.com [Wed, 16 Sep 2015 23:41:14 +0000 (23:41 +0000)]
lan78xx: Create lan78xx_get_mdix_status() and lan78xx_set_mdix_status() for MDIX control.

Create lan78xx_get_mdix_status() and lan78xx_set_mdix_status() for MDIX control.

Signed-off-by: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agolan78xx: Remove phy defines in lan78xx.h and use defines in include/linux/microchipphy.h
Woojung.Huh@microchip.com [Wed, 16 Sep 2015 23:41:07 +0000 (23:41 +0000)]
lan78xx: Remove phy defines in lan78xx.h and use defines in include/linux/microchipphy.h

Remove phy defines in lan78xx.h and use defines in include/linux/microchipphy.h.

Signed-off-by: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agolan78xx: Update to use phylib instead of mii_if_info.
Woojung.Huh@microchip.com [Wed, 16 Sep 2015 23:40:54 +0000 (23:40 +0000)]
lan78xx: Update to use phylib instead of mii_if_info.

Update to use phylib instead of mii_if_info.

Signed-off-by: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agolan78xx: Add PHYLIB and MICROCHIP_PHY as default config.
Woojung.Huh@microchip.com [Wed, 16 Sep 2015 23:40:47 +0000 (23:40 +0000)]
lan78xx: Add PHYLIB and MICROCHIP_PHY as default config.

Add PHYLIB and MICROCHIP_PHY as default configuration for lan78xx.

Signed-off-by: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agolan78xx: Check device ready bit (PMT_CTL_READY_) after reset the PHY
Woojung.Huh@microchip.com [Wed, 16 Sep 2015 23:40:39 +0000 (23:40 +0000)]
lan78xx: Check device ready bit (PMT_CTL_READY_) after reset the PHY

Check device ready bit (PMT_CTL_READY_) after reset the PHY.
Device may not be ready even if PHY_RST_ is cleared depends on configuration.

Signed-off-by: Woojung Huh <woojung.huh@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: Initialize table in fib result
David Ahern [Wed, 16 Sep 2015 16:16:39 +0000 (10:16 -0600)]
net: Initialize table in fib result

Sergey, Richard and Fabio reported an oops in ip_route_input_noref. e.g., from Richard:

[    0.877040] BUG: unable to handle kernel NULL pointer dereference at 0000000000000056
[    0.877597] IP: [<ffffffff8155b5e2>] ip_route_input_noref+0x1a2/0xb00
[    0.877597] PGD 3fa14067 PUD 3fa6e067 PMD 0
[    0.877597] Oops: 0000 [#1] SMP
[    0.877597] Modules linked in: virtio_net virtio_pci virtio_ring virtio
[    0.877597] CPU: 1 PID: 119 Comm: ifconfig Not tainted 4.2.0+ #1
[    0.877597] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
[    0.877597] task: ffff88003fab0bc0 ti: ffff88003faa8000 task.ti: ffff88003faa8000
[    0.877597] RIP: 0010:[<ffffffff8155b5e2>]  [<ffffffff8155b5e2>] ip_route_input_noref+0x1a2/0xb00
[    0.877597] RSP: 0018:ffff88003ed03ba0  EFLAGS: 00010202
[    0.877597] RAX: 0000000000000046 RBX: 00000000ffffff8f RCX: 0000000000000020
[    0.877597] RDX: ffff88003fab50b8 RSI: 0000000000000200 RDI: ffffffff8152b4b8
[    0.877597] RBP: ffff88003ed03c50 R08: 0000000000000000 R09: 0000000000000000
[    0.877597] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88003fab6f00
[    0.877597] R13: ffff88003fab5000 R14: 0000000000000000 R15: ffffffff81cb5600
[    0.877597] FS:  00007f6de5751700(0000) GS:ffff88003ed00000(0000) knlGS:0000000000000000
[    0.877597] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    0.877597] CR2: 0000000000000056 CR3: 000000003fa6d000 CR4: 00000000000006e0
[    0.877597] Stack:
[    0.877597]  0000000000000000 0000000000000046 ffff88003fffa600 ffff88003ed03be0
[    0.877597]  ffff88003f9e2c00 697da8c0017da8c0 ffff880000000000 000000000007fd00
[    0.877597]  0000000000000000 0000000000000046 0000000000000000 0000000400000000
[    0.877597] Call Trace:
[    0.877597]  <IRQ>
[    0.877597]  [<ffffffff812bfa1f>] ? cpumask_next_and+0x2f/0x40
[    0.877597]  [<ffffffff8158e13c>] arp_process+0x39c/0x690
[    0.877597]  [<ffffffff8158e57e>] arp_rcv+0x13e/0x170
[    0.877597]  [<ffffffff8151feec>] __netif_receive_skb_core+0x60c/0xa00
[    0.877597]  [<ffffffff81515795>] ? __build_skb+0x25/0x100
[    0.877597]  [<ffffffff81515795>] ? __build_skb+0x25/0x100
[    0.877597]  [<ffffffff81521ff6>] __netif_receive_skb+0x16/0x70
[    0.877597]  [<ffffffff81522078>] netif_receive_skb_internal+0x28/0x90
[    0.877597]  [<ffffffff8152288f>] napi_gro_receive+0x7f/0xd0
[    0.877597]  [<ffffffffa0017906>] virtnet_receive+0x256/0x910 [virtio_net]
[    0.877597]  [<ffffffffa0017fd8>] virtnet_poll+0x18/0x80 [virtio_net]
[    0.877597]  [<ffffffff815234cd>] net_rx_action+0x1dd/0x2f0
[    0.877597]  [<ffffffff81053228>] __do_softirq+0x98/0x260
[    0.877597]  [<ffffffff8164969c>] do_softirq_own_stack+0x1c/0x30

The root cause is use of res.table uninitialized.

Thanks to Nikolay for noticing the uninitialized use amongst the maze of
gotos.

As Nikolay pointed out the second initialization is not required to fix
the oops, but rather to fix a related problem where a valid lookup should
be invalidated before creating the rth entry.

Fixes: b7503e0cdb5d ("net: Add FIB table id to rtable")
Reported-by: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Reported-by: Richard Alpe <richard.alpe@ericsson.com>
Reported-by: Fabio Estevam <festevam@gmail.com>
Tested-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'bpf_avoid_clone'
David S. Miller [Fri, 18 Sep 2015 04:09:07 +0000 (21:09 -0700)]
Merge branch 'bpf_avoid_clone'

Alexei Starovoitov says:

====================
bpf: performance improvements

v1->v2: dropped redundant iff_up check in patch 2

At plumbers we discussed different options on how to get rid of skb_clone
from bpf_clone_redirect(), the patch 2 implements the best option.
Patch 1 adds 'integrated exts' to cls_bpf to improve performance by
combining simple actions into bpf classifier.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobpf: add bpf_redirect() helper
Alexei Starovoitov [Wed, 16 Sep 2015 06:05:43 +0000 (23:05 -0700)]
bpf: add bpf_redirect() helper

Existing bpf_clone_redirect() helper clones skb before redirecting
it to RX or TX of destination netdev.
Introduce bpf_redirect() helper that does that without cloning.

Benchmarked with two hosts using 10G ixgbe NICs.
One host is doing line rate pktgen.
Another host is configured as:
$ tc qdisc add dev $dev ingress
$ tc filter add dev $dev root pref 10 u32 match u32 0 0 flowid 1:2 \
   action bpf run object-file tcbpf1_kern.o section clone_redirect_xmit drop
so it receives the packet on $dev and immediately xmits it on $dev + 1
The section 'clone_redirect_xmit' in tcbpf1_kern.o file has the program
that does bpf_clone_redirect() and performance is 2.0 Mpps

$ tc filter add dev $dev root pref 10 u32 match u32 0 0 flowid 1:2 \
   action bpf run object-file tcbpf1_kern.o section redirect_xmit drop
which is using bpf_redirect() - 2.4 Mpps

and using cls_bpf with integrated actions as:
$ tc filter add dev $dev root pref 10 \
  bpf run object-file tcbpf1_kern.o section redirect_xmit integ_act classid 1
performance is 2.5 Mpps

To summarize:
u32+act_bpf using clone_redirect - 2.0 Mpps
u32+act_bpf using redirect - 2.4 Mpps
cls_bpf using redirect - 2.5 Mpps

For comparison linux bridge in this setup is doing 2.1 Mpps
and ixgbe rx + drop in ip_rcv - 7.8 Mpps

Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocls_bpf: introduce integrated actions
Daniel Borkmann [Wed, 16 Sep 2015 06:05:42 +0000 (23:05 -0700)]
cls_bpf: introduce integrated actions

Often cls_bpf classifier is used with single action drop attached.
Optimize this use case and let cls_bpf return both classid and action.
For backwards compatibility reasons enable this feature under
TCA_BPF_FLAG_ACT_DIRECT flag.

Then more interesting programs like the following are easier to write:
int cls_bpf_prog(struct __sk_buff *skb)
{
  /* classify arp, ip, ipv6 into different traffic classes
   * and drop all other packets
   */
  switch (skb->protocol) {
  case htons(ETH_P_ARP):
    skb->tc_classid = 1;
    break;
  case htons(ETH_P_IP):
    skb->tc_classid = 2;
    break;
  case htons(ETH_P_IPV6):
    skb->tc_classid = 3;
    break;
  default:
    return TC_ACT_SHOT;
  }

  return TC_ACT_OK;
}

Joint work with Daniel Borkmann.

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: only check perm protocol when register proto
Junwei Zhang [Fri, 18 Sep 2015 04:00:05 +0000 (00:00 -0400)]
net: only check perm protocol when register proto

The permanent protocol nodes are at the head of the list,
So only need check all these nodes.

No matter the new node is permanent or not,
insert the new node after the last permanent protocol node,

If the new node conflicts with existing permanent node,
return error.

Signed-off-by: Martin Zhang <martinbj2008@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobonding: use l4 hash if available
Eric Dumazet [Tue, 15 Sep 2015 22:24:28 +0000 (15:24 -0700)]
bonding: use l4 hash if available

If skb carries a l4 hash, no need to perform a flow dissection.

Performance is slightly better :

lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.39012e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.39393e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.39988e+06

After patch :

lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.43579e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.44304e+06
lpaa5:~# ./super_netperf 200 -H lpaa6 -t TCP_RR -l 100
2.44312e+06

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Tom Herbert <tom@herbertland.com>
Cc: Mahesh Bandewar <maheshb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agotcp: provide skb->hash to synack packets
Eric Dumazet [Tue, 15 Sep 2015 22:24:20 +0000 (15:24 -0700)]
tcp: provide skb->hash to synack packets

In commit b73c3d0e4f0e ("net: Save TX flow hash in sock and set in skbuf
on xmit"), Tom provided a l4 hash to most outgoing TCP packets.

We'd like to provide one as well for SYNACK packets, so that all packets
of a given flow share same txhash, to later enable bonding driver to
also use skb->hash to perform slave selection.

Note that a SYNACK retransmit shuffles the tx hash, as Tom did
in commit 265f94ff54d62 ("net: Recompute sk_txhash on negative routing
advice") for established sockets.

This has nice effect making TCP flows resilient to some kind of black
holes, even at connection establish phase.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Tom Herbert <tom@herbertland.com>
Cc: Mahesh Bandewar <maheshb@google.com>
Acked-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'nf_hook_netns'
David S. Miller [Fri, 18 Sep 2015 00:18:38 +0000 (17:18 -0700)]
Merge branch 'nf_hook_netns'

Eric W. Biederman says:

====================
Passing net through the netfilter hooks

My primary goal with this patchset and it's follow ups is to cleanup the
network routing paths so that we do not look at the output device to
derive the network namespace.  My plan is to pass the network namespace
of the transmitting socket through the output path, to replace code that
looks at the output network device today.  Once that is done we can have
routes with output devices outside of the current network namespace.
Which should allow reception and transmission of packets in network
namespaces to be as fast as normal packet reception and transmission
with early demux disabled, because it will same code path.

Once skb_dst(skb)->dev is a little better under control I think it will
also be possible to use rcu to cleanup the ancient hack that sets
dst->dev to loopback_dev when a network device is removed.

The work to get there is a series of code cleanups.  I am starting with
passing net into the netfilter hooks and into the functions that are
called after the netfilter hooks.  This removes from netfilter the
need to guess which network namespace it is working on.

To get there I perform a series of minor prep patches so the big changes
at the end are possible to audit without getting lost in the noise.  In
particular I have a lot of patches computing net into a local variable
and then using it through out the function.

So this patchset encompases removing dead code, sorting out the _sk
functions that were added last time someone pushed a prototype change
through the post netfilter functions.  Cleaning up individual functions
use of the network namespace.  Passing net into the netfilter hooks.
Passing net into the post netfilter functions.  Using state->net in
the netfilter code where it is available and trivially usable.

Pablo, Dave I don't know whose tree this makes more sense to go
through.  I am assuming at least initially Pablos as netfilter is
involved.  From what I have seen there will be a lot of back and forth
between the netfilter code paths and the routing code paths.

The patches are also available (against 4.3-rc1) at:
git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/net-next.git master
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonetfilter: Add blank lines in callers of netfilter hooks
Eric W. Biederman [Thu, 17 Sep 2015 22:21:31 +0000 (17:21 -0500)]
netfilter: Add blank lines in callers of netfilter hooks

In code review it was noticed that I had failed to add some blank lines
in places where they are customarily used.  Taking a second look at the
code I have to agree blank lines would be nice so I have added them
here.

Reported-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonetfilter: Pass net into okfn
Eric W. Biederman [Wed, 16 Sep 2015 01:04:18 +0000 (20:04 -0500)]
netfilter: Pass net into okfn

This is immediately motivated by the bridge code that chains functions that
call into netfilter.  Without passing net into the okfns the bridge code would
need to guess about the best expression for the network namespace to process
packets in.

As net is frequently one of the first things computed in continuation functions
after netfilter has done it's job passing in the desired network namespace is in
many cases a code simplification.

To support this change the function dst_output_okfn is introduced to
simplify passing dst_output as an okfn.  For the moment dst_output_okfn
just silently drops the struct net.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonetfilter: Use nf_hook_state.net
Eric W. Biederman [Wed, 16 Sep 2015 01:04:17 +0000 (20:04 -0500)]
netfilter: Use nf_hook_state.net

Instead of saying "net = dev_net(state->in?state->in:state->out)"
just say "state->net".  As that information is now availabe,
much less confusing and much less error prone.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonetfilter: Pass struct net into the netfilter hooks
Eric W. Biederman [Wed, 16 Sep 2015 01:04:16 +0000 (20:04 -0500)]
netfilter: Pass struct net into the netfilter hooks

Pass a network namespace parameter into the netfilter hooks.  At the
call site of the netfilter hooks the path a packet is taking through
the network stack is well known which allows the network namespace to
be easily and reliabily.

This allows the replacement of magic code like
"dev_net(state->in?:state->out)" that appears at the start of most
netfilter hooks with "state->net".

In almost all cases the network namespace passed in is derived
from the first network device passed in, guaranteeing those
paths will not see any changes in practice.

The exceptions are:
xfrm/xfrm_output.c:xfrm_output_resume()         xs_net(skb_dst(skb)->xfrm)
ipvs/ip_vs_xmit.c:ip_vs_nat_send_or_cont()      ip_vs_conn_net(cp)
ipvs/ip_vs_xmit.c:ip_vs_send_or_cont()          ip_vs_conn_net(cp)
ipv4/raw.c:raw_send_hdrinc()                    sock_net(sk)
ipv6/ip6_output.c:ip6_xmit() sock_net(sk)
ipv6/ndisc.c:ndisc_send_skb()                   dev_net(skb->dev) not dev_net(dst->dev)
ipv6/raw.c:raw6_send_hdrinc()                   sock_net(sk)
br_netfilter_hooks.c:br_nf_pre_routing_finish() dev_net(skb->dev) before skb->dev is set to nf_bridge->physindev

In all cases these exceptions seem to be a better expression for the
network namespace the packet is being processed in then the historic
"dev_net(in?in:out)".  I am documenting them in case something odd
pops up and someone starts trying to track down what happened.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: Add br_netif_receive_skb remove netif_receive_skb_sk
Eric W. Biederman [Wed, 16 Sep 2015 01:04:15 +0000 (20:04 -0500)]
bridge: Add br_netif_receive_skb remove netif_receive_skb_sk

netif_receive_skb_sk is only called once in the bridge code, replace
it with a bridge specific function that calls netif_receive_skb.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: Cache net in br_nf_pre_routing_finish
Eric W. Biederman [Wed, 16 Sep 2015 01:04:14 +0000 (20:04 -0500)]
bridge: Cache net in br_nf_pre_routing_finish

This is prep work for passing net to the netfilter hooks.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: Pass net into br_nf_push_frag_xmit
Eric W. Biederman [Wed, 16 Sep 2015 01:04:13 +0000 (20:04 -0500)]
bridge: Pass net into br_nf_push_frag_xmit

When struct net starts being passed through the ipv4 and ipv6 fragment
routines br_nf_push_frag_xmit will need to take a net parameter.
Prepare br_nf_push_frag_xmit before that is needed and introduce
br_nf_push_frag_xmit_sk for the call sites that still need the old
calling conventions.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: Pass net into br_nf_ip_fragment
Eric W. Biederman [Wed, 16 Sep 2015 01:04:12 +0000 (20:04 -0500)]
bridge: Pass net into br_nf_ip_fragment

This is a prep work for passing struct net through ip_do_fragment and
later the netfilter okfn.   Doing this independently makes the later
code changes clearer.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Compute net once in raw6_send_hdrinc
Eric W. Biederman [Wed, 16 Sep 2015 01:04:11 +0000 (20:04 -0500)]
ipv6: Compute net once in raw6_send_hdrinc

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Cache net in ip6_output
Eric W. Biederman [Wed, 16 Sep 2015 01:04:10 +0000 (20:04 -0500)]
ipv6: Cache net in ip6_output

Keep net in a local variable so I can use it in NF_HOOK_COND
when I pass struct net to all of the netfilter hooks.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Only compute net once in ip6_finish_output2
Eric W. Biederman [Wed, 16 Sep 2015 01:04:09 +0000 (20:04 -0500)]
ipv6: Only compute net once in ip6_finish_output2

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Don't recompute net in ip6_rcv
Eric W. Biederman [Wed, 16 Sep 2015 01:04:08 +0000 (20:04 -0500)]
ipv6: Don't recompute net in ip6_rcv

Avoid silly redundant code

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: Remove dev_queue_xmit_sk
Eric W. Biederman [Wed, 16 Sep 2015 01:04:07 +0000 (20:04 -0500)]
net: Remove dev_queue_xmit_sk

A function with weird arguments that it will never use to accomdate a
netfilter callback prototype is absolutely in the core of the
networking stack.  Frankly it does not make sense and it causes a lot
of confusion as to why arguments that are never used are being passed
to the function.

As I am preparing to make a second change to arguments to the okfn even
the names stops making sense.

As I have removed the two callers of this function remove this confusion
from the networking stack.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agobridge: Introduce br_send_bpdu_finish
Eric W. Biederman [Wed, 16 Sep 2015 01:04:06 +0000 (20:04 -0500)]
bridge: Introduce br_send_bpdu_finish

The function dev_queue_xmit_skb_sk is unncessary and very confusing.
Introduce br_send_bpdu_finish to remove the need for dev_queue_xmit_skb_sk,
and have br_send_bpdu_finish call dev_queue_xmit.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoarp: Introduce arp_xmit_finish
Eric W. Biederman [Wed, 16 Sep 2015 01:04:05 +0000 (20:04 -0500)]
arp: Introduce arp_xmit_finish

The function dev_queue_xmit_skb_sk is unncessary and very confusing.
Introduce arp_xmit_finish to remove the need for dev_queue_xmit_skb_sk,
and have arp_xmit_finish call dev_queue_xmit.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv6: Only compute net once in ip6mr_forward2_finish
Eric W. Biederman [Wed, 16 Sep 2015 01:04:04 +0000 (20:04 -0500)]
ipv6: Only compute net once in ip6mr_forward2_finish

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Only compute net once in ipmr_forward_finish
Eric W. Biederman [Wed, 16 Sep 2015 01:04:03 +0000 (20:04 -0500)]
ipv4: Only compute net once in ipmr_forward_finish

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Only compute net once in ip_rcv_finish
Eric W. Biederman [Wed, 16 Sep 2015 01:04:02 +0000 (20:04 -0500)]
ipv4: Only compute net once in ip_rcv_finish

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Only compute net once in ip_finish_output2
Eric W. Biederman [Wed, 16 Sep 2015 01:04:01 +0000 (20:04 -0500)]
ipv4: Only compute net once in ip_finish_output2

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Explicitly compute net in ip_fragment
Eric W. Biederman [Wed, 16 Sep 2015 01:04:00 +0000 (20:04 -0500)]
ipv4: Explicitly compute net in ip_fragment

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Only compute net once in ip_do_fragment
Eric W. Biederman [Wed, 16 Sep 2015 01:03:59 +0000 (20:03 -0500)]
ipv4: Only compute net once in ip_do_fragment

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Don't recompute net in ipmr_queue_xmit
Eric W. Biederman [Wed, 16 Sep 2015 01:03:58 +0000 (20:03 -0500)]
ipv4: Don't recompute net in ipmr_queue_xmit

Calling dev_net(dev) for is just silly.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Remember the net in ip_output and ip_mc_output
Eric W. Biederman [Wed, 16 Sep 2015 01:03:57 +0000 (20:03 -0500)]
ipv4: Remember the net in ip_output and ip_mc_output

This is a prepatory patch to passing net int the netfilter hooks,
where net will be used again.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Compute net once in ip_rcv
Eric W. Biederman [Wed, 16 Sep 2015 01:03:56 +0000 (20:03 -0500)]
ipv4: Compute net once in ip_rcv

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Compute net once in ip_forward_finish
Eric W. Biederman [Wed, 16 Sep 2015 01:03:55 +0000 (20:03 -0500)]
ipv4: Compute net once in ip_forward_finish

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoipv4: Compute net once in ip_forward
Eric W. Biederman [Wed, 16 Sep 2015 01:03:54 +0000 (20:03 -0500)]
ipv4: Compute net once in ip_forward

Compute struct net from the input device in ip_forward before it is
used.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: Merge dst_output and dst_output_sk
Eric W. Biederman [Wed, 16 Sep 2015 01:03:53 +0000 (20:03 -0500)]
net: Merge dst_output and dst_output_sk

Add a sock paramter to dst_output making dst_output_sk superfluous.
Add a skb->sk parameter to all of the callers of dst_output
Have the callers of dst_output_sk call dst_output.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoxfrm: Remove unused afinfo method init_dst
Eric W. Biederman [Wed, 16 Sep 2015 01:03:52 +0000 (20:03 -0500)]
xfrm: Remove unused afinfo method init_dst

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonetfilter: Pass net to nf_hook_thresh
Eric W. Biederman [Wed, 16 Sep 2015 01:03:51 +0000 (20:03 -0500)]
netfilter: Pass net to nf_hook_thresh

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonetfilter: Store net in nf_hook_state
Eric W. Biederman [Wed, 16 Sep 2015 01:03:50 +0000 (20:03 -0500)]
netfilter: Store net in nf_hook_state

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonetfilter: Remove !CONFIG_NETFITLER definition of nf_hook_thresh
Eric W. Biederman [Wed, 16 Sep 2015 01:03:49 +0000 (20:03 -0500)]
netfilter: Remove !CONFIG_NETFITLER definition of nf_hook_thresh

The !CONFIG_NETFILTER definition of nf_hook_thresh calls okfn when
the CONFIG_NETFITLER defintion does not, making it buggy.

As the !CONFIG_NETFILTER defintion of nf_hook_thresh is not used remove
it rather than fix it.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Thu, 17 Sep 2015 23:50:36 +0000 (16:50 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates 2015-09-15

This series contains updates to ixgbe and fm10k.

Don fixes a ixgbe issue by adding checks for systems that do not have
SFP's to avoid incorrectly acting on interrupts that are falsely
interpreted as SFP events.

Alex Williamson adds a fix for ixgbe to disable SR-IOV prior to
unregistering the netdev to avoid issues with guest OS's which do not
support hot-unplug or their hot-unplug is broken.

Alex Duyck update the lowest limit for adaptive interrupt interrupt
moderation to about 12K interrupts per second for ixgbe.  This change
increases the performance for ixgbe.  Also fixed up fm10k to remove
the optimization that assumed that all fragments would be limited to
page size, since that assumption is incorrect as the TCP allocator can
provide up to a 32K page fragment.  Updated fm10k to add the MAC
address to the list of values recorded on driver load.  Fixes fm10k
so that we only trigger the data path reset if the fabric is ready to
handle traffic to avoid triggering the reset unless the switch API is
ready for us.

Jacob updates the fm10k driver to disable the service task during
suspend and re-enable it after we resume. If we don't do this, the
device could be UP when you suspend and come back from resume as
DOWN.  Also update fm10k to prevent the removal of default VID rules,
 and correctly remove the stack layers information of the VLAN, but then
return to forwarding that VID as untagged frames.  If we deleted the VID
rules here, we would begin dropping traffic due to VLAN membership
violations.  Fixed fm10k to use pcie_get_minimum_link(), which is useful
in cases where we connect to a slot at Gen3, but the slot is behind a bus
which is only connected at Gen2.  Updated fm10k to update the netdev
permanent address during reinit instead of up to enable users to
immediately see the new MAC address on the VF even if the device is not
up.  Adds the creation of VLAN interfaces on a device, even while the
device is down for fm10k.  Fixed an issue where we request the incorrect
MAC/VLAN combinations, and prevents us from accidentally reporting some
frames as VLAN tagged.  Provided a couple of trivial fixes for fm10k
to fix code style and typos in code comments.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet-sysfs: get_netdev_queue_index() cleanup
Thadeu Lima de Souza Cascardo [Tue, 15 Sep 2015 21:28:00 +0000 (18:28 -0300)]
net-sysfs: get_netdev_queue_index() cleanup

Redo commit ed1acc8cd8c22efa919da8d300bab646e01c2dce.

Commit 822b3b2ebfff8e9b3d006086c527738a7ca00cd0 ("net: Add max rate tx queue
attribute") moved get_netdev_queue_index around, but kept the old version.
Probably because of a reuse of the original patch from before Eric's change to
that function.

Remove one inline keyword, and no need for a loop to find
an index into a table.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@redhat.com>
Fixes: 822b3b2ebfff ("net: Add max rate tx queue attribute")
Acked-by: Or Gerlitz <ogerlitz@mellanox.com>
Acked-by: John Fastabend <john.r.fastabend@intel.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: smc91x: convert pxa dma to dmaengine
Robert Jarzmik [Thu, 10 Sep 2015 19:26:04 +0000 (21:26 +0200)]
net: smc91x: convert pxa dma to dmaengine

Convert the dma transfers to be dmaengine based, now pxa has a dmaengine
slave driver. This makes this driver a bit more PXA agnostic.

The driver was tested on pxa27x (mainstone) and pxa310 (zylonite),
ie. only pxa platforms.

Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agonet: fix cdc-phonet.c dependency and build error
Randy Dunlap [Wed, 16 Sep 2015 18:32:41 +0000 (11:32 -0700)]
net: fix cdc-phonet.c dependency and build error

Fix build error caused by missing Kconfig dependency:

ERROR: "cdc_parse_cdc_header" [drivers/net/usb/cdc-phonet.ko] undefined!

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agocdc: add header guards
Stephen Rothwell [Wed, 16 Sep 2015 01:10:16 +0000 (11:10 +1000)]
cdc: add header guards

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
9 years agofm10k: fix iov_msg_mac_vlan_pf VID checks
Jacob Keller [Wed, 24 Jun 2015 20:34:50 +0000 (13:34 -0700)]
fm10k: fix iov_msg_mac_vlan_pf VID checks

The VF will send a message to request multicast addresses with the
default VID. In the current code, if the PF has statically assigned a
VLAN to a VF, then the VF will not get the multicast addresses. Fix up
all of the various VLAN messages to use identical checks (since each
check was different). Also use set as a variable, so that it simplifies
our check for whether VLAN matches the pf_vid.

The new logic will allow set of a VLAN if it is zero, automatically
converting to the default VID. Otherwise it will allow setting the PF
VID, or any VLAN if PF has not statically assigned a VLAN. This is
consistent behavior, and allows VF to request either 0 or the
default_vid without silently failing.

Note that we need the check for zero since VFs might not get the default
VID message in time to actually request non-zero VLANs.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agofm10k: Only trigger data path reset if fabric is up
Alexander Duyck [Wed, 24 Jun 2015 20:34:49 +0000 (13:34 -0700)]
fm10k: Only trigger data path reset if fabric is up

This change makes it so that we only trigger the data path reset if the
fabric is ready to handle traffic.  The general idea is to avoid
triggering the reset unless the switch API is ready for us.  Otherwise
we can just postpone the reset until we receive a switch ready
notification.

Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
9 years agofm10k: re-enable VF after a full reset on detection of a Malicious event
Jacob Keller [Wed, 24 Jun 2015 20:34:48 +0000 (13:34 -0700)]
fm10k: re-enable VF after a full reset on detection of a Malicious event

Modify behavior of Malicious Driver Detection events. Presently, the
hardware disables the VF queues and re-assigns them to the PF. This
causes the VF in question to continuously Tx hang, because it assumes
that it can transmit over the queues in question. For transient events,
this results in continuous logging of malicious events.

New behavior is to reset the LPORT and VF state, so that the VF will
have to reset and re-enable itself. This does mean that malicious VFs
will possibly be able to continue and attempt malicious events again.
However, it is expected that system administrators will step in and
manually remove or disable the VF in question.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>