Debugfs file location: ath9k/<wiphy>/dma
Contains values in DMA debug registers.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Make debugging configurable, and add a module parameter
to give the debug mask.
Add debug.c to hold all debug specific code.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Remove all the useless __func__ prefixes in debug messages,
and replace the DPRINTF macro with a function.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Hostapd now passes the HT parameters through the config()
callback, use these to set the appropriate channel in AP mode.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The TX channel width of the BSS can be obtained only after association.
In all cases, default to HT20 if HT is enabled, and set
chan width to HT40 only if the BSS supports it.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Merge core.c and base.c
Remove Antenna Diversity (unused now).
Remove unused chainmask handling code.
Comment, indentation scrub.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We would get an oops on RX on -ENOMEM by passing
NULL to the hardware on ath_rx_buf_link(). The oops
would look something like this:

ath_rx_tasklet+0x515/0x53b
ath9k_tasklet+0x48
tasklet_action
__do_softirq
irq_exit
do_IRQ

RIP: ath_rx_buf_link+0x3a

We correct this by handling the requeue directly on
the ath_rx_tasklet() and trying to allocate an skb
*prior* to sending up the last hardware processed
skb. If we run out of memory this gauranteees we have
skbs to work with while it simply drops new received
frames.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Update the rate control only with the tx status of first
AMPDU of an aggregation. This patch fixes frequent drops
in throughput.
Signed-off-by: NVasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

CHANNEL_CCK flag is set for all 2 Ghz channels, so IS_CHAN_CCK() would
turn out to be true for all channles in that band.
Use IS_CHAN_B() now, which checks the channel mode and not the channel
flags.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Also, use a helper function to setup RC status data
when processing completed TX descriptors.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Break down huge functions, use helper functions or
macros instead.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Replace TRUE/FALSE macros with VALID/INVALID macros.
Follow a consistent variable convention.
Remove unnecessary comments.
Add all RC phy macros into a single enum.
Merge functions into reasonably sized entities.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Maintaining two sets of rate tables is redundant, remove one
and use struct ath_rate_table exclusively.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Remove the hack using vif, and use rate_driver_data within
skb->cb to hold driver specific rate information.
Setup the rate series in the skb's tx control area and remove
all references to ath9k specific rate series ( using struct ath_rc_series ).
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Move the hw rate tables to ath_softc, and access them directly.
tx_triglevel_max is global, move it to ath_rate_node.
Now that ath_rate_softc is gone, rate control attach becomes simpler.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Rate control init is now confined to itself, using the
HT capabilites of the STA from rate_init().
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Remove a lot of old, crufty code and make
RX status reporting a bit sane and clean.

Do not do anything to the RX skb before unmapping.
So in ath_rx_tasklet(), move the skb_put() after PCI unmap.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Half/Quarter rate tables are needed only for legacy chipsets.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Setup legacy rates in ath_rate_init() itself.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Avoid casting of ath_tx_ratctrl and access the elements directly.
Signed-off-by: NSujith <Sujith.Manoharan@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

 [ 6133.670329] BUG: unable to handle kernel NULL pointer dereference at 0000000000000010
 [ 6133.672802] IP: [<ffffffffa030fcf6>] ieee80211_stop_queues+0x26/0x40 [mac80211]
 [ 6133.672802] PGD 759dc067 PUD 74f1d067 PMD 0
 [ 6133.672802] Oops: 0002 [#1] PREEMPT SMP
 [ 6133.672802] last sysfs file: /sys/class/backlight/acpi_video0/brightness
 [ 6133.672802] CPU 0
 [ 6133.672802] Modules linked in: ath9k(-) mac80211 pciehp pci_hotplug arc4 ecb joydev pcmcia ppdev lp ppp_generic psmouse sg pcspkr s]
 [ 6133.735830] Pid: 4445, comm: rmmod Tainted: G        W  2.6.28-rc5-wl #1
 [ 6133.735830] RIP: 0010:[<ffffffffa030fcf6>]  [<ffffffffa030fcf6>] ieee80211_stop_queues+0x26/0x40 [mac80211]
 [ 6133.735830] RSP: 0018:ffff88007d1efd10  EFLAGS: 00010246
 [ 6133.735830] RAX: 0000000000000000 RBX: ffff880074f41aa0 RCX: 0000000000000000
 [ 6133.735830] RDX: 0000000000000010 RSI: 0000000000000000 RDI: ffff880074f40340
 [ 6133.735830] RBP: ffff880074990000 R08: 0000000000000000 R09: 000000000000224d
 [ 6133.735830] R10: 0000000000000000 R11: ffffffff8031dc70 R12: 0000000000000000
 [ 6133.735830] R13: 0000000000000001 R14: ffff880074f46c9c R15: 0000000000000000
 [ 6133.735830] FS:  00007f1e2e0bc6f0(0000) GS:ffffffff805e0b80(0000) knlGS:0000000000000000
 [ 6133.735830] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
 [ 6133.735830] CR2: 0000000000000010 CR3: 0000000075593000 CR4: 00000000000006e0
 [ 6133.735830] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 [ 6133.735830] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
 [ 6133.735830] Process rmmod (pid: 4445, threadinfo ffff88007d1ee000, task ffff88007c0c8000)
 [ 6133.735830] Stack:
 [ 6133.735830]  ffffffffa034d583 ffff88007c7d9410 ffff88007c7d9410 ffff88007c7d9410
 [ 6133.735830]  ffffffff80481dab ffff880074f41aa0 00000000fffffff0 0000000000000000
 [ 6133.735830]  0000000000000001 0000000000000001 ffffffffa034d8a5 ffff88007c7d9400
 [ 6133.735830] Call Trace:
 [ 6133.735830]  [<ffffffffa034d583>] ? ath_radio_disable+0x33/0x150 [ath9k]
 [ 6133.735830]  [<ffffffff80481dab>] ? __mutex_lock_slowpath+0x20b/0x2a0
 [ 6133.735830]  [<ffffffffa034d8a5>] ? ath_sw_toggle_radio+0x65/0xa0 [ath9k]
 [ 6133.735830]  [<ffffffffa019d1f4>] ? rfkill_toggle_radio+0x74/0x140 [rfkill]
 [ 6133.735830]  [<ffffffffa019d597>] ? rfkill_remove_switch+0x67/0x80 [rfkill]
 [ 6133.735830]  [<ffffffffa019d955>] ? rfkill_unregister+0x25/0x50 [rfkill]
 [ 6133.735830]  [<ffffffffa034bf75>] ? ath_detach+0xf5/0x140 [ath9k]
 [ 6133.735830]  [<ffffffffa034bfe9>] ? ath_pci_remove+0x29/0x80 [ath9k]
 [ 6133.735830]  [<ffffffff8035263c>] ? pci_device_remove+0x2c/0x60
 [ 6133.735830]  [<ffffffff803c3829>] ? __device_release_driver+0x99/0x100
 [ 6133.735830]  [<ffffffff803c3950>] ? driver_detach+0xc0/0xd0
 [ 6133.735830]  [<ffffffff803c296e>] ? bus_remove_driver+0x8e/0xd0
 [ 6133.735830]  [<ffffffff80352916>] ? pci_unregister_driver+0x36/0xa0
 [ 6133.735830]  [<ffffffffa0356ad4>] ? exit_ath_pci+0x10/0x29 [ath9k]
 [ 6133.735830]  [<ffffffff8026bb1b>] ? sys_delete_module+0x1cb/0x2d0
 [ 6133.735830]  [<ffffffff802960d9>] ? do_munmap+0x349/0x390
 [ 6133.735830]  [<ffffffff80342d01>] ? __up_write+0x21/0x150
 [ 6133.735830]  [<ffffffff8020c45b>] ? system_call_fastpath+0x16/0x1b
 [ 6133.735830] Code: c3 0f 1f 40 00 0f b7 57 5e 0f b7 47 5c 01 c2 74 30 31 c9 66 90 48 8b 57 78 0f b7 c1 48 c1 e0 07 48 03 82 00 03 00
 [ 6133.735830] RIP  [<ffffffffa030fcf6>] ieee80211_stop_queues+0x26/0x40 [mac80211]
Signed-off-by: NVasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This fixes the poor wireless connection which happens even
if we are very well in the range.

Signed-off-by: Don.breslin@atheros.com
Signed-off-by: NVasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Make the DMAable mameory consistent with pci_set_consistent_dma_mask().
The DMA-mapping.txt Documentation recommends this but for PCI-X
considerations and on strange architecture like SGI SN2, not sure
why it would fix an issue but lets see if it does, just in case.

Before this, this driver was tested with x86_64 with about
7 GB of RAM, not sure if this is really needed.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Use DMA_32BIT_MASK to clarify we only want 32-bit DMA
memory. What was there before is also 32-bit but this makes it
clearer
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Race condition causes RX buffers to be accessed even before it is
initialized. The RX and TX buffers are initialized immediately after
the hardware is registered with mac80211. The mac80211 start callback
is ready to be fired once the device is registered for a case when the
wpa_supplicant is also running at the same time.

The same race condition is also possible for RKFILL registration
as RFKILL init happens after the device registration with mac80211
and it is possible that rfkill_register would be called even before
it is initialized.
Signed-off-by: NSenthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

ath9k frees irq even before IRQs are disabled and existing IRQs
are flushed when rfkill_register() fails.
Signed-off-by: NSenthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Currently, ath9k builds RFKILL feature only when the RFKILL subsystem
is built part of the kernel. Build RFKILL feature regardless of whether
RFKILL subsystem is built as a MODULE or part of the kernel.
Signed-off-by: NSenthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This patch fixes incorrect noise floor threshold values.
Signed-off-by: NSenthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

init values update for various atheros chipsets.
Signed-off-by: NSenthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We should only tell the hardware its capable of DMA'ing
to us only what we asked dev_alloc_skb(). Prior to this
it is possible a large RX'd frame could have corrupted
DMA data but for us but we were saved only because we
were previously also pci_map_single()'ing the same large
value. The issue prior to this though was we were unmapping
a smaller amount which the prior DMA patch fixed.
Signed-off-by: NBennyam Malavazi <Bennyam.Malavazi@atheros.com>
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This should fix the SW-IOMMU bounce buffer starvation
seen ok kernel.org bugzilla 11811:

http://bugzilla.kernel.org/show_bug.cgi?id=11811

Users on MacBook Pro 3.1/MacBook v2 would see something like:

DMA: Out of SW-IOMMU space for 4224 bytes at device 0000:0b:00.0

Unfortunately its only easy to trigger on MacBook Pro 3.1/MacBook v2
so far so its difficult to debug (even with swiotlb=force).

We were pci_unmap_single()'ing less bytes than what we called
for with pci_map_single() and as such we were starving
the swiotlb from its 64MB amount of bounce buffers. We remain
consistent and now always use sc->rxbufsize for RX. While at
it we update the beacon DMA maps as well to only use the data
portion of the skb, previous to this we were pci_map_single()'ing
more data for beaconing than what we tell the hardware it can use,
therefore pushing more iotlb abuse.

Still not sure why this is so easily triggerable on
MacBook Pro 3.1, it may be the hardware configuration
tends to use more memory > 3GB mark for DMA.
Signed-off-by: NMaciej Zenczykowski <zenczykowski@gmail.com>
Signed-off-by: NBennyam Malavazi <Bennyam.Malavazi@atheros.com>
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

ARRAY_SIZE is more concise to use when the size of an array is divided by
the size of its type or the size of its first element.

The semantic patch that makes this change is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@i@
@@

#include <linux/kernel.h>

@depends on i using "paren.iso"@
type T;
T[] E;
@@

- (sizeof(E)/sizeof(E[...]))
+ ARRAY_SIZE(E)
// </smpl>
Signed-off-by: NJulia Lawall <julia@diku.dk>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>