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: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NSeth Forshee <seth.forshee@canonical.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Functions for flow control exist but remain unimplemented. Remove these
in advance of adding real flow control.
Signed-off-by: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Use this helper function rather than open-coding the same calculation in
multiple places.
Signed-off-by: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

According to the comments this "reduces rate lag," but in reality the
only way this value is used is for determining whether or not any frames
remain to be transmitted. Therefore there's no reason for AMPDU packets
to receive any weighting.
Signed-off-by: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

AMPDU session allows MPDUs to be temporarily queued until either a full
AMPDU has been collected or circumstances dictate that transmission
should start with a partial AMPDU. Packets are added to the session by
calling brcms_c_ampdu_add_frame(). brcms_c_ampdu_finalize() should be
called to fix up the tx headers in the first and last packet before
adding the packets to the DMA ring. brmcs_c_sendampdu() is converted to
using AMPDU sessions.

This patch has no real value on it's own, but is needed in preparation
for elimination of the tx packet queue from brcmsmac.
Signed-off-by: NSeth Forshee <seth.forshee@canonical.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NArend van Spriel <arend@broadcom.com>
Tested-by: NDaniel Wagner <wagi@monom.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Samuel says:

"This is the 2nd NFC pull request for 3.8.

With this one we have:

- A few HCI improvements in preparation for an upcoming HCI chipset support.
- A pn544 code cleanup after the old driver was removed.
- An LLCP improvement for notifying user space when one peer stops ACKing I
  frames."
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Drivers are allowed to modify the sent skb and thus we need to make a copy
of it before passing it to the driver. Without this fix, LLCP Tx skbs were
not queued properly as the ptype check was failing due to e.g. the pn533
driver skb_pushing the Tx skb.
Reported-by: NThierry Escande <thierry.escande@linux.intel.com>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

When the tx pending queues and/or the socket tx queue is getting too deep,
we have to let userspace know. We won't be queueing any more frames until
the congestion is fixed.
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

Using the userspace IO vector directly is wrong, we should copy it from
user space first.
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

The platform data header for PN544 based NFC devices should
also be mentioned here.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

The pn544.h just provides the platform data struct and defines and
nothing else. So move it to to linux/platform_data/ now.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

The majority of the defines and structures from pn544.h are no
longer in use. So just remove them.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

Some HCI drivers will need it.
Signed-off-by: NEric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

Signed-off-by: NEric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

A chip with pre-opened gates may send events on a gate that nobody
has opened in the handset host. Discard those events.
Signed-off-by: NEric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

NFC_HCI_ID_MGMT_VERSION_SW and NFC_HCI_ID_MGMT_VERSION_HW are optional
registers for gate NFC_HCI_ID_MGMT_GATE in standard HCI. When chip
doesn't implement, just leave all the information as zeros.
Signed-off-by: NEric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

In some cases, pre-opened pipes don't stay open when a clear all pipes
command is sent. They stay created however. Therefore, one can never
assume that such a pipe is already open. As re-opening a pipe seems not
to be a problem, we do that now.
Signed-off-by: NEric Lapuyade <eric.lapuyade@intel.com>
Signed-off-by: NSamuel Ortiz <sameo@linux.intel.com>

The debugfs file for dumping btcoex parameters unconditionally
assumes a MCI-based device. This will not work for older btcoex
chips. Fix this by branching out the routine into separate
functions.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

2484 Mhz (Japan) usage requires filter coefficients to
be programmed in the CCK TX FIR registers. This is required
for AR9331, AR9485 and AR9462. Fix this and also remove
a few useless macros and a duplicate variable.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

There's no need to ask the user about lib80211
since it will be selected by drivers requiring
it, hide it from Kconfig.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The driver doesn't use any lib80211 symbols
so it shouldn't select it in Kconfig.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Acked-by: NBing Zhao <bzhao@marvell.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Using eth_zero_addr() to assign zero address insetad of
memset() or an inefficient copy from a static array.
Signed-off-by: NWei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Remove unnecessary if statements because libipw_set_geo always
returns success.  Also change function's return value from int
to void.
Signed-off-by: NStanislav Yakovlev <stas.yakovlev@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Removing obsolete functions and prototypes. Moving (and renaming)
defines to place with similar definitions. Removing unnecessary
includes.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When debug is turned on for fwil then the whole data buffer is
dumped. In some cases this gives excessive amount of debug. With
this patch the dumps are limited to 64 bytes.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

On sdio module unload followed by load (without removing the
device) the access window should be moved back to enumeration
space. Force this by removing initialisation of sbwad during
probe.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

brcmf_ops_sdio_probe used the private_date func->card->dev to
store device data of brcmfmac sdio. This is not a good place to
store the data. Use dev of func and use func->card->sdio_func
to group the functions the driver is using.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

on sdio remove the bus_if should be configured for close, so
new data from higher layers will be blocked. Also the access
to bus_if in the watchdog should be checked for null pointer
access on sdio remove.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

when brcmf_sdbrcm_probe_attach results in error then cleanup
will result in null pointer access. In brcmf_sdbrcm_release and
in brcmf_ops_sdio_remove. This patch fixes order of init and
de-init.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In brcmf_sdbrcm_probe only error ELINK is seen as error. However
brcmf_bus_start can return many more error codes and all should
result in failed init of driver.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The IF event need special care. It can be either an ADD, DEL, or
CHANGE. For an ADD we need to call brcmf_add_if() before the
event handler call. Upon a DEL we need to call brcmf_del_if()
after the event handler call. CHANGE does not require special
attention.
Reviewed-by: NHante Meuleman <meuleman@broadcom.com>
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Firmware fires IF event to add the primary interface but that
is already created in the driver.
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Replaced <linux/unaligned/access_ok.h> by <asm/unaligned.h>
to make it work on ARM and other architectures.
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

brcmf_fweh_detach can be called while ifp is already NULL, due to
init error. Fix NULL pointer access by checking ifp.
Reviewed-by: NArend Van Spriel <arend@broadcom.com>
Signed-off-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The mac_addr field in ifp object is always valid so no need to
validate.
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The function brcmf_add_if() is called with mac address set to NULL
for the primary interface. When handling IF ADD events the firmware
provides a address mask in the event to derive its mac address from
the primary mac address. Rename the parameter and use it as a mask.
Reviewed-by: NPieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: NHante Meuleman <meuleman@broadcom.com>
Signed-off-by: NArend van Spriel <arend@broadcom.com>
Signed-off-by: NFranky Lin <frankyl@broadcom.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>