Hauke Mehrtens [Mon, 26 Nov 2012 23:31:55 +0000 (00:31 +0100)]
ssb: extif: fix compile errors
If CONFIG_SSB_EMBEDDED or CONFIG_SSB_DRIVER_MIPS is set and
CONFIG_SSB_DRIVER_EXTIF is not set, it will cause compile problems
because of missing functions. This patch fixes these problems.
The mips driver now also uses ssb_chipco_available() instead of
checking bus->chipco.dev manually.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Hauke Mehrtens [Mon, 26 Nov 2012 23:30:04 +0000 (00:30 +0100)]
bcma: handle return value of pci_assign_resource
This fixes the following warning:
CC drivers/bcma/driver_pci_host.o
drivers/bcma/driver_pci_host.c: In function 'bcma_core_pci_fixup_addresses':
drivers/bcma/driver_pci_host.c:555:23: error: ignoring return value of
'pci_assign_resource', declared with attribute warn_unused_result
[-Werror=unused-result]
Reported-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Thu, 29 Nov 2012 00:25:20 +0000 (01:25 +0100)]
cfg80211: fix BSS struct IE access races
When a BSS struct is updated, the IEs are currently
overwritten or freed. This can lead to races if some
other CPU is accessing the BSS struct and using the
IEs concurrently.
Fix this by always allocating the IEs in a new struct
that holds the data and length and protecting access
to this new struct with RCU.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 28 Nov 2012 21:42:34 +0000 (22:42 +0100)]
cfg80211: fix cmp_hidden_bss
The cmp_bss() comparator function uses memcmp() to
compare the SSID. This means that cmp_hidden_bss()
needs to similarly return a number bigger than zero
(use 1) instead of -1 when ie1 is bigger than ie2,
which is the case if an ie2 byte is non-zero.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Marco Porsch [Fri, 23 Nov 2012 20:23:18 +0000 (12:23 -0800)]
mac80211: fix for mesh sync to indicate TBTT adjustment
Currently the mesh sync code checks, whether peers indicate TBTT adjustment,
but it never sets the corresponding flag itself.
By setting ifmsh->tbtt_adjusting to true, it will set the corresponding field
in the mesh configuration IE of own beacons.
This indication will be set in the current beacon. The TBTT adjustment will be
performed afterwards, affecting the next beacon. Thus, the first beacon with
stable TBTT will not indicate adjustment anymore and peers will continue
tracking the new offset.
Signed-off-by: Marco Porsch <marco.porsch@etit.tu-chemnitz.de> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Saravana [Wed, 28 Nov 2012 12:59:38 +0000 (18:29 +0530)]
mac80211: re-organize the rx rate calculation logic
Currently the logic to fill a struct rate_info with
a STA's last RX rate is accessible only in the cfg.c.
As the RX rate calculation might be needed elsewhere,
split this out into a separate function.
Signed-off-by: Saravana <saravanad@posedge.com>
[fix various whitespace issues, reword commit log] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Saravana [Thu, 29 Nov 2012 13:52:37 +0000 (19:22 +0530)]
mac80211: add debugfs file for current tx rate
Add a debugfs file showing the current tx rate.
The information available in the rc_stats file
doesn't evidently provides us the current tx rate.
This patch adds the support for the same.
Signed-off-by: Saravana <saravanad@posedge.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Helmut Schaa [Tue, 27 Nov 2012 17:23:06 +0000 (18:23 +0100)]
mac80211: reject setting masked mac addresses
If a driver registers an address mask we should ensure that no
interface gets an address assigned that isn't covered by the
registered address mask. This prevents invalid configurations
from reaching the device and causing problems.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
[change function flow to reduce indentation, fix locking] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 14 Nov 2012 14:17:28 +0000 (15:17 +0100)]
nl80211: support P2P GO powersave configuration
If a driver supports P2P GO powersave, allow it to
set the new feature flags for it and allow userspace
to configure the parameters for it. This can be done
at GO startup and later changed with SET_BSS.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 22 Nov 2012 13:11:39 +0000 (14:11 +0100)]
mac80211: support VHT association
Determine the VHT channel from the AP's VHT operation IE
(if present) and configure the hardware to that channel
if it is supported. If channel contexts cause a channel
to not be usable, try a smaller bandwidth.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 22 Nov 2012 15:59:45 +0000 (16:59 +0100)]
cfg80211: rework chandef checking and export it
Some of the chandef checking that we do in cfg80211
to check if a channel is supported or not is also
needed in mac80211, so rework that a bit and export
the functions that are needed.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The LMAC API states that the TSF clock value of
every rx'ed frame is a "usec accurate timestamp
of the hardware clock at the end of frame
(before OFDM SIFS EOF padding)".
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Fri, 23 Nov 2012 22:46:19 +0000 (23:46 +0100)]
iwlwifi: initialize rx_status
The vendor radiotap patch added a few fields to
struct ieee80211_rx_status that need to be zero,
initialize the struct instead of using whatever
was left on the stack.
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Tested-by: Kalle Valo <kvalo@adurom.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Thu, 22 Nov 2012 11:58:16 +0000 (12:58 +0100)]
iwlegacy: initialize rx_status
The vendor radiotap patch added a few fields to
struct ieee80211_rx_status that need to be zero,
initialize the struct instead of using whatever
was left on the stack.
Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Fri, 23 Nov 2012 23:32:19 +0000 (00:32 +0100)]
mac80211: fix remain-on-channel (non-)cancelling
Felix Liao reported that when an interface is set DOWN
while another interface is executing a ROC, the warning
in ieee80211_start_next_roc() (about the first item on
the list having started already) triggers.
This is because ieee80211_roc_purge() calls it even if
it never actually changed the list of ROC items. To fix
this, simply remove the function call. If it is needed
then it will be done by the ieee80211_sw_roc_work()
function when the ROC item that is being removed while
active is cleaned up.
Cc: stable@vger.kernel.org Reported-by: Felix Liao <Felix.Liao@watchguard.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Fri, 16 Nov 2012 16:22:28 +0000 (17:22 +0100)]
cfg80211: fix some tracing output issues
In some cases, e.g. probe_status, there were spaces
missing so the trace output was confusing. Also make
it more like mac80211 when printing netdevs/wiphys
to make reading a combined log easier.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Fri, 9 Nov 2012 10:39:59 +0000 (11:39 +0100)]
mac80211: convert to channel definition struct
Convert mac80211 (and where necessary, some drivers a
little bit) to the new channel definition struct.
This will allow extending mac80211 for VHT, which is
currently restricted to channel contexts since there
are no drivers using that which makes it easier. As
I also don't care about VHT for drivers not using the
channel context API, I won't convert the previous API
to VHT support.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 8 Nov 2012 22:14:50 +0000 (23:14 +0100)]
nl80211/cfg80211: support VHT channel configuration
Change nl80211 to support specifying a VHT (or HT)
using the control channel frequency (as before) and
new attributes for the channel width and first and
second center frequency. The old channel type is of
course still supported for HT.
Also change the cfg80211 channel definition struct
to support these by adding the relevant fields to
it (and removing the _type field.)
This also adds new helper functions:
- cfg80211_chandef_create to create a channel def
struct given the control channel and channel type,
- cfg80211_chandef_identical to check if two channel
definitions are identical
- cfg80211_chandef_compatible to check if the given
channel definitions are compatible, and return the
wider of the two
This isn't entirely complete, but that doesn't matter
until we have a driver using it. In particular, it's
missing
- regulatory checks on the usable bandwidth (if that
even makes sense)
- regulatory TX power (database can't deal with it)
- a proper channel compatibility calculation for the
new channel types
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 8 Nov 2012 20:25:48 +0000 (21:25 +0100)]
cfg80211: pass a channel definition struct
Instead of passing a channel pointer and channel type
to all functions and driver methods, pass a new channel
definition struct. Right now, this struct contains just
the control channel and channel type, but for VHT this
will change.
Also, add a small inline cfg80211_get_chandef_type() so
that drivers don't need to use the _type field of the
new structure all the time, which will change.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 8 Nov 2012 17:31:02 +0000 (18:31 +0100)]
cfg80211: remove remain-on-channel channel type
As mwifiex (and mac80211 in the software case) are the
only drivers actually implementing remain-on-channel
with channel type, userspace can't be relying on it.
This is the case, as it's used only for P2P operations
right now.
Rather than adding a flag to tell userspace whether or
not it can actually rely on it, simplify all the code
by removing the ability to use different channel types.
Leave only the validation of the attribute, so that if
we extend it again later (with the needed capability
flag), it can't break userspace sending invalid data.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Mon, 26 Nov 2012 10:57:41 +0000 (11:57 +0100)]
mac80211: fix managed mode channel flags handling
If ieee80211_prep_channel() decides that HT should be
disabled (because the HT IEs from the AP were invalid)
it will set the IEEE80211_STA_DISABLE_HT to not send
HT capabilities to the AP when associating. If this
happens during authentication, the flag will be lost
and we send HT frames, even if the channel config was
set up for non-HT. This can lead to issues.
Fix this by always resetting the ifmgd flags to zero
when the channel context is released so that the flag
resetting in ieee80211_mgd_assoc() isn't necessary.
To make the code a bit easier move the call to release
the channel in ieee80211_set_disassoc() to the end of
the function together with the flag resetting (which
needs to be at the end to avoid timers setting flags.)
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
mac80211: in ADHOC print debug message for every Auth message
The debug message has to be printed also for an Auth message with
auth_sequence != 1. This helps understanding whether the two Auth
messages are exchanged correctly or not.
Signed-off-by: Antonio Quartulli <antonio@open-mesh.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
mac80211: in ADHOC don't update last_rx if sta is not authorized
It does not make sense to keep a station alive if it is not authorised
at all. If IBSS/RSN is used it could also be the case that something
went wrong during the keys exchange and the stations ended up in a not
recoverable state.
By not updating last_rx we are giving the station a chance to be
deleted and to start the key exchange once again from scratch.
Signed-off-by: Antonio Quartulli <antonio@open-mesh.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Arend van Spriel [Sun, 25 Nov 2012 18:13:28 +0000 (19:13 +0100)]
cfg80211: change function signature of cfg80211_get_p2p_attr()
The function cfg80211_get_p2p_attr() can fail and returns
a negative error code. However, the return type is unsigned
int. The largest positive number is determined by desired_len
variable in the function, which is u16. So changing the return
type to int to allow easy error checking. Also change the type
for the attribute to enum for improved type checking.
Signed-off-by: Arend van Spriel <arend@broadcom.com>
[fix indentation, don't use u8 attr variable] Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Fri, 23 Nov 2012 13:23:30 +0000 (14:23 +0100)]
cfg80211: use DS or HT operation IEs to determine BSS channel
Currently, mac80211 checks the DS params IE if present and
uses it for the (primary) BSS channel, instead of the one
that the frame was received on. This is particularly useful
in the 2.4 GHz band since a frame is often received on one
of the adjacent channels due to overlap.
Move this code to cfg80211 so other drivers also do this.
Additionally, on 5 GHz, in particular with some (possibly)
upcoming changes in 802.11ai and duplicate transmissions
when wider channels are used, something similar happens.
So if present, also use the (primary) channel information
contained in the HT operation IE.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 22 Nov 2012 13:34:05 +0000 (14:34 +0100)]
mac80211: disable HT advertising unless AP supports it
If the AP doesn't support HT, or more importantly if
it does but we have to disable it because its IEs are
broken, don't advertise HT support in our association
request. Otherwise, we configure our channel to be a
20 MHz non-HT channel but the AP might still think we
support HT, or even 40 MHz.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 22 Nov 2012 13:17:23 +0000 (14:17 +0100)]
mac80211: fix RX chains configuration
If the driver doesn't support 40 MHz channels, then
mac80211 erroneously sets number of RX chains to one
although the number of chains is independent of the
support for 40 MHz channels.
Fix this by checking the 40 MHz support only for the
code that sets the 40 MHz channel not the complete
HT code block.
This also means the HT20 channel type will always be
set in the changed code block so there's no need to
set it in case we override the AP due to invalid IEs
in the probe response/beacon.
The indentation is a bit quirky, but I'm rewriting
this code for VHT support so this will change again
very soon.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Thu, 22 Nov 2012 09:20:58 +0000 (10:20 +0100)]
mac80211: fix radiotap vendor area skipping
The radiotap vendor area in the skb head must be skipped
and accounted for in a few functions until it is removed.
I missed this in my patch, so a few places use this data
as though it was the 802.11 header, fix these places.
Reported-by: Wojciech Dubowik <Wojciech.Dubowik@neratec.com> Tested-by: Wojciech Dubowik <Wojciech.Dubowik@neratec.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com>
John W. Linville [Wed, 21 Nov 2012 19:59:46 +0000 (14:59 -0500)]
rtl8723ae: fix build break from "mac80211: support RX_FLAG_MACTIME_END"
CC drivers/net/wireless/rtlwifi/rtl8723ae/trx.o
drivers/net/wireless/rtlwifi/rtl8723ae/trx.c: In function ‘rtl8723ae_rx_query_desc’:
drivers/net/wireless/rtlwifi/rtl8723ae/trx.c:324:21: error: ‘RX_FLAG_MACTIME_MPDU’ undeclared (first use in this function)
drivers/net/wireless/rtlwifi/rtl8723ae/trx.c:324:21: note: each undeclared identifier is reported only once for each function it appears in
make[3]: *** [drivers/net/wireless/rtlwifi/rtl8723ae/trx.o] Error 1
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith Manoharan [Wed, 21 Nov 2012 12:43:15 +0000 (18:13 +0530)]
ath9k: Remove 'stations' debugfs file
The 'stations' debugfs file has multiple issues. It doesn't scale
to an arbitrary number of associated stations and allocating
64K is not elegant either. Now that changes have been made in
mac80211 to support dynamic creation/deletion of driver-specific
debugfs files on station addition/removal, remove this file and
make use of the mac80211 hooks (which will be done in a sebsequent
patch).
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith Manoharan [Wed, 21 Nov 2012 12:43:13 +0000 (18:13 +0530)]
ath9k: Fill remove_sta_debugfs() callback
Remove the rate control statistics debugfs file properly
via remove_sta_debugfs(). Also, check for both MAC80211_DEBUGFS
and ATH9K_DEBUGFS config options.
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith Manoharan [Wed, 21 Nov 2012 12:43:11 +0000 (18:13 +0530)]
ath9k: Fix the 'xmit' debugfs file
The 'xmit' debugfs file has become big and unwieldy, fix
multiple issues with its usage:
* Store TX counters/statistics only for the 4 Access Categories.
Use IEEE80211_NUM_ACS instead of ATH9K_NUM_TX_QUEUES.
* Move various utility macros to debug.h, they can be reused
elsewhere.
* Remove tx_complete_poll_work_seen.
* Remove code that accesses various internal queue-specific
variables without any locking whatsoever. HW/SW queue details
will be handled in a subsequent patch.
* Do not print internal values like txq_headidx and txq_headidx.
They were mostly unused anyway, considering code like:
PRX("txq_tailidx: ", txq_headidx);
* Handle 'txprocdesc' for EDMA too.
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
ath9k: stomp audio profiles on weak signal strength
On lower WLAN signal strength, WLAN downlink traffic might suffer
from retransmissions. At the mean time, playing SCO/A2DP profiles
is affecting WLAN stability. In such scenario, by stomping SCO/A2DP
BT traffic completely for a BTCOEX period, gives WLAN traffic an
oppertunity to recover PHY rate. It also improves WLAN stability at
lower RSSI without sacificing BT traffic.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch reverts the commit "ath9k_hw: Wait BT calibration to complete"
and bail out from MCI interrupt routine for chip reset. The above commit
stalls the WLAN TCP traffic while bringing up and down the BT interface
iteratively. Fixing this properly by queueing up chip reset and bailing
out properly from tasklet routine.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Sujith Manoharan [Tue, 20 Nov 2012 03:16:02 +0000 (08:46 +0530)]
mac80211: Add debugfs callbacks for station addition/removal
Provide drivers with hooks to create debugfs files when
a new station is added. This would help drivers to take
advantage of mac80211's station list infrastructure and not maintain
tedious station management code internally.
John W. Linville [Mon, 19 Nov 2012 16:04:28 +0000 (11:04 -0500)]
brcmfmac: include linux/vmalloc.h from usb.c
This avoids build failures on some architectures...
Signed-off-by: John W. Linville <linville@tuxdriver.com> Reported-by: Fengguang Wu <fengguang.wu@intel.com> Acked-by: Arend van Spriel <arend@broadcom.com>
John W. Linville [Mon, 19 Nov 2012 15:53:32 +0000 (10:53 -0500)]
brcmfmac: check return from kzalloc in brcmf_fweh_process_event
Signed-off-by: John W. Linville <linville@tuxdriver.com> Reported-by: Fengguang Wu <fengguang.wu@intel.com> Acked-by: Arend van Spriel <arend@broadcom.com>
Seth Forshee [Thu, 15 Nov 2012 14:08:12 +0000 (08:08 -0600)]
brcmsmac: Remove some noisy and uninformative debug messages
These messages clutter up the trace buffer without adding any useful
information.
Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:08:11 +0000 (08:08 -0600)]
brcmsmac: Add tracepoint for AMPDU session information
Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:08:10 +0000 (08:08 -0600)]
brcmsmac: Add tracepoint for macintstatus
Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:08:09 +0000 (08:08 -0600)]
brcmsmac: Improve tx trace and debug support
Add the brcmsmac_tx trace system for tx debugging. Existing code to dump
tx status and descriptors are converted to using tracepoints, allowing
for more efficient collection and post-processing of this data. These
tracepoints are placed to collect data for all tx frames instead of only
on errors. Logging of tx errors is also improved.
Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:08:03 +0000 (08:08 -0600)]
brcmsmac: Use debug macros for general error and debug statements
Convert most uses of wiphy_* and pr_* for general error and debug
messages to use the internal debug macros instead. Most code used only
for initialization still use wiphy_err(), as well as some locations
which are executed too early to use the debug macros. Some debug
messages which are redundant or not useful are removed.
Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:08:02 +0000 (08:08 -0600)]
brcmsmac: Add support for writing debug messages to the trace buffer
Add a new brcmsmac_msg trace system to enable writing of debug messages
to the trace buffer, and add brcms_* macros for storing device debug
messages in the trace buffer in addition to the printk log buffer.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by: Arend van Spriel <arend@broadcom.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:08:00 +0000 (08:08 -0600)]
brcm80211: Convert log message levels to debug levels
In preparation for enhancements to debug and trace support, convert the
message levels to debug levels which will be used for enabling
categories of debug messages. The two message levels are little-used
anyway and are combined into the BRCM_DL_INFO debug level.
Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:07:59 +0000 (08:07 -0600)]
brcm80211: Allow trace support to be enabled separately from debug
Since the runtime overhead of trace support is small when tracing is
disabled, users may be interested in turning on trace support while
leaving other debug features off. Add a new config option named
CONFIG_BRCM_TRACING for this purpose.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by: Arend van Spriel <arend@broadcom.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:07:58 +0000 (08:07 -0600)]
brcmsmac: Reduce number of entries in tx DMA rings
Currently up to 256 frames can be queued for each DMA ring. This is
excessive, and now that we have better flow control we can get by with
less. Experimentation has shown 64 to work well.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by: Arend van Spriel <arend@broadcom.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:07:57 +0000 (08:07 -0600)]
brcmsmac: Use correct descriptor count when calculating next rx descriptor
nextrxd() is calling txd(), which means that the tx descriptor count is
used to determine when to wrap for determining the next ring buffer
entry. This has worked so far since the driver has been using the same
number of rx and tx descriptors, but it's obviously going to be a
problem if different numbers of descriptors are used.
Acked-by: Arend van Spriel <arend@broadcom.com> Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:07:56 +0000 (08:07 -0600)]
brcmsmac: Remove internal tx queue
The brcmsmac internal tx buffering is problematic. The amount of
buffering is excessive (228 packets in addition to the 256 slots in each
DMA ring), and frames may be dropped due to a lack of flow control.
This patch reworks the transmit code path to remove the internal
buffering. Frames are immediately handed off to the DMA support rather
than passing through an intermediate queue. Non-aggregate frames are
queued immediately into the tx rings, and aggregate frames are queued
temporarily in an AMPDU session until ready for transmit.
Transmit flow control is also added to avoid dropping packets when the
tx rings are full. Conceptually this is a separate change, but it's
included in this commit because removing the tx queue without adding
flow control could cause significant problems.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by: Arend van Spriel <arend@broadcom.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Seth Forshee [Thu, 15 Nov 2012 14:07:55 +0000 (08:07 -0600)]
brcmsmac: Use IEEE 802.11 AC levels for pktq precedence levels
The mac80211 tx queues and brcmsmac DMA fifos both map directly to AC
levels. Therefore it's much more straightforward to queue tx frames and
choose the tx fifo based on the mac80211 queue instead of mapping 802.1D
priority tags to precedence levels then back to AC levels. mac80211
already maps the 802.1D levels to the appropriate AC levels and queues
management frames at the maximum priority, so the results should be
identical.
One functional change resulting from this patch is that AMPDU retries no
longer get a priority boost to queue them ahead of packets with the same
priority already in the tx queue. This behavior will be restored (in
effect at least) in a later patch when the tx queue is removed.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com> Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com> Reviewed-by: Arend van Spriel <arend@broadcom.com> Tested-by: Daniel Wagner <wagi@monom.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
projects / mirror_ubuntu-artful-kernel.git / log