build fix: ath_flushrecv_tgt uses ath_rx_buffer

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: sc_txbuf_held is ath_tx_buf

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: PHY_ADDR should be uint32

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: add ath_rx_bufhead and ath_tx_bufhead

Normally we have two working type of buffers tx and rx.
In some generic situations, for example ath_descdma_setup,
we do not really care about buffer type, here we can use
ath_bufhead.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: use ath_rx_desc for ah_setupRxDesc and ah_procRxDescFast, part 2

ar5416SetupRxDesc_20 and ar5416ProcRxDescFast_20 changed too.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

buid fix: probably typo, ieee80211_node instead of ieee80211_node_target

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: ath_tgt_tx_prepare should return ath_tx_buf

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: fix wrong conversation of ath_vap_target to ieee80211vap_target

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: ath_beacon_setup works only with ath_tx_buf

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: wmi_event and wmi_cmd_rsp should use void *buffer

we pass different buffer types to this functions.
It makes no sense to define some concrect type here

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: add type conversation for asf_tailq_next

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: ah_procRxDescFast should use ath_rx_desc

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: ah_setupRxDesc should use ath_rx_desc

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: sc_rxdesc_held should be ath_rx_desc

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: use struct ath_tx_buf for ath_dma_map and ath_dma_unmap

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: use arch safe uint32_t instead of unsigned long

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: change HAL_TXKEYIX_INVALID size

HAL_(R)XKEYIX_INVALID is a_uint8_t so i assume
HAL_(T)XKEYIX_INVALID should have same size.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: use our internal memcpy, for now.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: replace struct with static

replace struct with static, it probably was typo.
It is already struct by type deffinition.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: ath_hal_getrtsaggrlimit can be removed here

reason for this part of code:
"AR5416/AR5418 doesn't work with RTS + aggregates longer than 8KiB,
so the driver has to either limit aggregate sizes to 8KiB or less,
or disable RTS on longer aggregates. This was fixed in subsequent chips.
...
 it's quite possible that there's some carrier / enterprise solution
somewhere that uses the AR7010 as an offload CPU for an AR5418 device.
...
I'd eventually like us to support the ethernet offload mode -
even if it's just as a reference we can use internally
at work when demo'ing stuff."
  Adrian Chadd

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: remove unused variables from if_owl.c

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: remove unused dev_info in if_ath_pci.c

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: change return type of ath_detach_tgt

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: remove unused variables from if_ath.c

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: ds_link should be compared with 0, not NULL

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: remove firts definition of ATH_NODE_TARGET

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: predefine struct ath_tx_desc

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: if_llc.h: __packed__ is ignored if not after "}"

gcc ignores atribut __packed__ if it is not right after "}"

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: adf_os_export_symbol is undefined

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: ar5416_hw.c remove unused variables

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: magpie..init.c remove unused variables

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: fix array size

v2 - use #define..

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: remove byte order/swapping support

currently AH_BYTE_ORDER is not defined. So this parst is unused any way.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: src on memcpy is allways const

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: remove unused *ahp

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

build fix: Magpie_Sys_Commands_Tbl used only if MAGPIE_ENABLE_WLAN == 0

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>

Include licencing terms in the source files themselves.

Licence updates for the buf pool code.

Licence updates for the ADF code.

More licence file updates.

Mark files without an existing BSD licence as having a QCA ClearBSD
licence.

Signed-off-by: Adrian Chadd <adrian@freebsd.org>

Merge pull request #3 from KrasnikovEugene/master

Fix for issue #2 - make clean cause buid error

Move config files from toolchain folder

Signed-off-by: Eugene Krasnikov <k.eugene.e@gmail.com>

Update to clearbsd/isc licence; this is what it should've been.

Sorry!

Clarify the licence terms a little more. It seems people think
the firmware is (part, full) GPLv2, which is actually totally untrue.

Fix README - there are no AR9285+AR7010 cards.

Signed-off-by: Sujith Manoharan <sujith@msujith.org>

Add mailing lists and bug information.

Add local patches.

Add a README.

..oops, this is a shell script, not a makefile.  Silly me.

Initial cut of the open ath9k htc firmware.