When the bus is in D0i3, we can't send regular commands to
the firmware. This means that we need to add a state to
remember what is our d0i3 state and make sure that only
d0i3 exit commands can be sent.
Add flags to CMD_ flags and transport status for this
purpose.
Commands with CMD_HIGH_PRIO set are queued at the head of
the command queue, behind other high priority commands.
Commands with CMD_SEND_IN_IDLE set can be sent while the
transport is idle (without taking rpm reference).
Commands with CMD_MAKE_TRANS_IDLE set indicate that command
completion should mark the transport as idle (and release
the bus).
Commands with CMD_WAKE_UP_TRANS set instruct the transport
to exit from idle when this command is completed.
The transport is marked as idle (STATUS_TRANS_IDLE) when
the FW enters D0i3 state. This bit is cleared when it
enters D0 state again.
Process only commands with CMD_SEND_IN_IDLE flag while the
transport is idle. Other enqueued commands will be
processed only later, right after exiting D0i3.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Add new enter_d0i3 and exit_d0i3 ops that
will be called by the transport on D0i3 enter/exit.
Each one of these ops will include the host commands
mentionned in the previous patch.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
D0i3 is bus power saving feature. It involves the
firmware - the driver needs to send a list of commands
to the firmware before entering this state. Wake up from
d0i3 also requires a few commands to the firmware.
The trigger to enter D0i3 is an idle timeout that will be
implemented later and will most probably rely on RUNTIME_PM
infrastructure.
In order to prevent entrance to D0i3 in critical flows, we
implement here a reference infrastructure. When a ref is
taken, we can't enter D0i3.
PCIe does't support D0i3.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The tx backoff settings used by the thermal throttling mechanism can
also be used for enforcing a limit on the power consumption of the module.
Handle the platform PCIe power limitation by translating the limit
(measured in mw) to its respective tx backoff value. The translation is
module specific.
The resulting tx backoff value is sent to the ucode, and also serves as the
minimal backoff value that can be set by the thermal throttling mechanism.
Signed-off-by: Ido Yariv <idox.yariv@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Some platforms may have power limitations on PCIe cards connected to
specific root ports.
This information is encoded as part of the ACPI tables, for instance:
<snip>
Name (SPLX, Package (0x02)
{
Zero,
Package (0x03)
{
0x07,
0x00000500,
0x80000000
}
})
Method (SPLC, 0, Serialized)
{
Return (SPLX)
}
</snip>
The structure returned contains the domain type, the default power
limitation and the default time window (reserved for future use).
Upon PCI probing, call the relevant ACPI method, parse the returned
structure, and save the power limitation.
Signed-off-by: Ido Yariv <idox.yariv@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Divide the maximal quota between all the data interfaces even in the
case of a single low latency binding without any other non low latency
interfaces, so that afterwards the quota allocation (which considers
the number of data interfaces) will be correct.
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>
Currently the quota remainder was added to the first binding, although
it is possible that this was not a data binding (only the P2P_DEVICE
interface is part of the binding).
Fix this by adding the remainder to the first binding that was actually
allocated quota.
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>
This makes the code a little bit longer as zero-extension
has to be done (mov vs. movzwl), but that's miniscule and
the space saving is significant, about 600 bytes in DVM
and 700 bytes in MVM, so the cache effect should be worth
the few bytes more code.
While at it, remove two spurious blank lines in variable
declaration blocks.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Eyal Shapira <eyal@wizery.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This will be useful during tests done on the physical layer.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Allow reading and setting bcast filtering configuration
through debugfs.
By default, mvm->bcast_filters is used for setting
the bcast filtering configuration (these filters
will be configured for each associated station).
For testing purposes, allow overriding this configuration,
and setting the bcast filtering configuration manually.
The following debugfs keys can be used:
* bcast_filtering/override - use debugfs values instead
of default configuration
* bcast_filtering/filters - set filters (+ attributes)
* bcast_filtering/macs - per-mac bcast filtering
configuration (policy + attached filters)
Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Add our ip as a new attribute to the bcast filtering
configuration (i.e. check the dest ip field of the
arp request).
Add bcast filter to pass incoming dhcp offer
broadcast frames as well (for sta vifs).
In order to support such dynamic configuration,
use the reserved1 field as a bitmap for driver internal
flags (which will indicate we want to configure the ip
in this attribute), and reconfigure the bcast
filtering on BSS_CHANGED_ARP_FILTER indication.
Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Configure arp request broadcast filter if this
option is enabled, in order to allow only arp
request broadcasts to pass-in.
(A following patch will make this filter even narrower
by limiting the arp request to our own ip)
Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Broadcast filtering allows dropping broadcast
frames that don't match the configured patterns.
Use predefined filters, and configure them for
each associated station vif.
There is no need to optimize and attach the same
filter to multiple vifs, as a following patch
will configure each filter to have per-vif unique
values.
Configure the bcast filtering on assoc changes.
Add a new IWLWIFI_BCAST_FILTERING Kconfig option
in order to enable broadcast filtering.
Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
As a debug tool, we dump the SRAM from the device when an
error occurs. The main users of this want it in a different
format, so change the format to suit their needs.
Also - add a short delay between the prints to make sure
that the user space logger can catch up.
This happens only when the firmware asserts, and only when
fw_restart is set to 0 which is typically a testing
configuration.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Based on the Bluetooth activity grading, we can stop using
the shared antenna and ask the stations to honor the new
SMPS state.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
If a vif is in low latency mode, it should be in primary
channel.
Also tell BT Coex about the change when a vif enters or
exits low latency mode.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Limit the scheduling duration of bindings without a low-latency
interface in the firmware, this prevents those bindings from
occupying the medium for a period of time longer than what we
want for the other interfaces in low-latency mode.
As older firmware doesn't do anything with the max_duration field
and ignores it completely, there's no need for a firmware flag.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
If there is/are interface(s) in low-latency mode, reserve a
percentage (currently 64%) of the quota for that binding to
improve the quality of service for those interfaces. However,
if there's more than one binding that has low-latency, then
give up and don't reserve, we can't allocate more than 100%.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
While an interface is in low-latency mode, for now powersave
should be disabled for it, so take low-latency into account
in the powersave code and force powersave recalculation when
low-latency mode changes.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
For various traffic use cases, we want to be able to treat multi-
channel scenarios differently. Introduce a low-latency framework
that currently only has a debugfs file to enable low-latency mode,
but can later be extended.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Notify scan completed if fw_restart flow isn't going to be run.
Otherwise, the scan will stay stack forever and mac80211 will
not be able to remove the interface.
Signed-off-by: David Spinadel <david.spinadel@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Don't stop scheduled scan before reporting HW restart;
mac80211 was changed to reschedule it after reconfigure.
Signed-off-by: David Spinadel <david.spinadel@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
In iwl_pcie_int_cause_non_ict, trans_pcie is used for lockdep
purposes only. Since this might not be enabled, trans_pcie
finds itself without user leading to a complaint from gcc.
Avoid using trans_pcie by inlining IWL_TRANS_GET_PCIE_TRANS.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Since we use IWL_MVM_STATION_COUNT all over the driver, we
need to make sure that it is the right constant to look at.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
We somtimes need to fetch the iwl_mvm_sta structure from a
station index - provide a helper to do that.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
iwlwifi-7260-8.ucode has been release. Warn if it is not
on the file system.
iwlwifi-7260-7.ucode is still supported for another kernel
version.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The code seems fine, as buf won't be assigned when an error
is returned, but checking for the error first is easier to
understand.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Also handle the bypass mode in which the second CPU doesn't
interfere.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This configuration is invalid for this family.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Reviewed-by: Dor Shaish <dor.shaish@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
This register is not present in 8000 family devices.
There is prph register instead.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Reviewed-by: Dor Shaish <dor.shaish@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
APMG HW block was removed in this NIC, hence, no need to
configure it.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
The identification of the hardware section in the NVM
of new devices has been changed, hence the need to add it
to iwl_cfg and adapt the code that uses this value
accordingly.
Signed-off-by: Eran Harary <eran.harary@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Newer firmware will support uAPSD clients in AP/GO mode, so complete
the driver support for it. The way it works is described in comments
in the code, but basically the driver just has to pass down all the
mac80211 requests and do accounting on agg/non-agg queues properly.
For older firmware, this doesn't change anything as it ignores the
fields used by the new firmware, and we only advertise uAPSD support
when the firmware does.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Pull networking updates from David Miller:
1) BPF debugger and asm tool by Daniel Borkmann.
2) Speed up create/bind in AF_PACKET, also from Daniel Borkmann.
3) Correct reciprocal_divide and update users, from Hannes Frederic
Sowa and Daniel Borkmann.
4) Currently we only have a "set" operation for the hw timestamp socket
ioctl, add a "get" operation to match. From Ben Hutchings.
5) Add better trace events for debugging driver datapath problems, also
from Ben Hutchings.
6) Implement auto corking in TCP, from Eric Dumazet. Basically, if we
have a small send and a previous packet is already in the qdisc or
device queue, defer until TX completion or we get more data.
7) Allow userspace to manage ipv6 temporary addresses, from Jiri Pirko.
8) Add a qdisc bypass option for AF_PACKET sockets, from Daniel
Borkmann.
9) Share IP header compression code between Bluetooth and IEEE802154
layers, from Jukka Rissanen.
10) Fix ipv6 router reachability probing, from Jiri Benc.
11) Allow packets to be captured on macvtap devices, from Vlad Yasevich.
12) Support tunneling in GRO layer, from Jerry Chu.
13) Allow bonding to be configured fully using netlink, from Scott
Feldman.
14) Allow AF_PACKET users to obtain the VLAN TPID, just like they can
already get the TCI. From Atzm Watanabe.
15) New "Heavy Hitter" qdisc, from Terry Lam.
16) Significantly improve the IPSEC support in pktgen, from Fan Du.
17) Allow ipv4 tunnels to cache routes, just like sockets. From Tom
Herbert.
18) Add Proportional Integral Enhanced packet scheduler, from Vijay
Subramanian.
19) Allow openvswitch to mmap'd netlink, from Thomas Graf.
20) Key TCP metrics blobs also by source address, not just destination
address. From Christoph Paasch.
21) Support 10G in generic phylib. From Andy Fleming.
22) Try to short-circuit GRO flow compares using device provided RX
hash, if provided. From Tom Herbert.
The wireless and netfilter folks have been busy little bees too.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (2064 commits)
net/cxgb4: Fix referencing freed adapter
ipv6: reallocate addrconf router for ipv6 address when lo device up
fib_frontend: fix possible NULL pointer dereference
rtnetlink: remove IFLA_BOND_SLAVE definition
rtnetlink: remove check for fill_slave_info in rtnl_have_link_slave_info
qlcnic: update version to 5.3.55
qlcnic: Enhance logic to calculate msix vectors.
qlcnic: Refactor interrupt coalescing code for all adapters.
qlcnic: Update poll controller code path
qlcnic: Interrupt code cleanup
qlcnic: Enhance Tx timeout debugging.
qlcnic: Use bool for rx_mac_learn.
bonding: fix u64 division
rtnetlink: add missing IFLA_BOND_AD_INFO_UNSPEC
sfc: Use the correct maximum TX DMA ring size for SFC9100
Add Shradha Shah as the sfc driver maintainer.
net/vxlan: Share RX skb de-marking and checksum checks with ovs
tulip: cleanup by using ARRAY_SIZE()
ip_tunnel: clear IPCB in ip_tunnel_xmit() in case dst_link_failure() is called
net/cxgb4: Don't retrieve stats during recovery
...
The threshold values for RX interrupt mitigation
are different for AR9003 and AR9002 families.
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The commit "ath9k: Process GTT interrupts" accidentally
had a line that was commented out.
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Use the right function to update frequency value.
If rx skb is probe response or beacon, the wrong frequency value can
cause problem that bss info can't be updated when it should be.
Cc: <stable@vger.kernel.org>
Fixes: 8318d78a44 ("cfg80211 API for channels/bitrates, mac80211
and driver conversion")
Signed-off-by: ZHAO Gang <gamerh2o@gmail.com>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
8 bytes preamble
14 bytes src/dst/eth_type
6 bytes 0xff:0xff..
http://en.wikipedia.org/wiki/Wake-on-LAN#Magic_packethttp://en.wikipedia.org/wiki/EtherType
This will fail if we VLAN or the magic packet is encapsulated
as a UDP packet...
Signed-off-by: Andreas Fenkart <afenkart@gmail.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Pull trivial tree updates from Jiri Kosina:
"Usual rocket science stuff from trivial.git"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (39 commits)
neighbour.h: fix comment
sched: Fix warning on make htmldocs caused by wait.h
slab: struct kmem_cache is protected by slab_mutex
doc: Fix typo in USB Gadget Documentation
of/Kconfig: Spelling s/one/once/
mkregtable: Fix sscanf handling
lp5523, lp8501: comment improvements
thermal: rcar: comment spelling
treewide: fix comments and printk msgs
IXP4xx: remove '1 &&' from a condition check in ixp4xx_restart()
Documentation: update /proc/uptime field description
Documentation: Fix size parameter for snprintf
arm: fix comment header and macro name
asm-generic: uaccess: Spelling s/a ny/any/
mtd: onenand: fix comment header
doc: driver-model/platform.txt: fix a typo
drivers: fix typo in DEVTMPFS_MOUNT Kconfig help text
doc: Fix typo (acces_process_vm -> access_process_vm)
treewide: Fix typos in printk
drivers/gpu/drm/qxl/Kconfig: reformat the help text
...
None of the devices supported by iwldvm have support for
shadow registers. This means that we wake the NIC
when we increment the write pointer on Tx ring.
This happened even before my bad commit mentionned below.
Since my commit below, we wake up the NIC when we put a
host command on the ring regardless of shadow register
support. This means that in iwldvm (when the NIC doesn't
support shadow register), we wake up the NIC twice:
pcie_enqueue_hcmd:
wake up the NIC
iwl_pcie_txq_inc_wr_ptr:
wake up the NIC - no shadow reg support
Since waking up the NIC means that we need to acquire a
spinlock, this obviously leads to a recursive spinlock
and hence a freeze.
Fixes: b943949105 ("iwlwifi: pcie: keep the NIC awake when commands are in flight")
Reported-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Arik Nemtsov