Since polling for tx completion is handled whenever target to host
messages are received, removing the unnecessary polling mechanism for
send completion at HTC level.
Reviewed-by: Michal Kazior
Signed-off-by: Rajkumar Manoharan
---
drivers/net/wireless/ath/ath10k/hif.h | 9 +++--
drivers/n
Since polling for received messages not supported, remove unused
dl_is_polled.
Reviewed-by: Michal Kazior
Signed-off-by: Rajkumar Manoharan
---
drivers/net/wireless/ath/ath10k/hif.h | 7 +++
drivers/net/wireless/ath/ath10k/htc.c | 7 +++
drivers/net/wireless/ath/ath10k/htc.h | 1 -
d
Register receive callbacks for every copy engines (CE) separately
instead of having common receive handler. Some of the copy engines
receives different type of messages (i.e HTT/HTC/pktlog) from target.
Hence to service them accordingly, register per copy engine receive
callbacks.
Reviewed-by: Mic
Register send completion callbacks for every copy engines (CE) separately
instead of having common completion handler. Since some of the copy
engines delivers different type of messages, per-CE callbacks help to
service them differently.
Reviewed-by: Michal Kazior
Signed-off-by: Rajkumar Manohara
Some special copy engines delivers messages directly to HTT by
bypassing HTC layer. Hence exporting tx_completion and rx_handler
for delivering the data to HTT layer.
Reviewed-by: Michal Kazior
Signed-off-by: Rajkumar Manoharan
---
drivers/net/wireless/ath/ath10k/htt.h| 1 +
drivers/net/wir
Currently target to host (T2H) HTT messages are received at copy engine 1.
These messages are processed by HTC layer in both host and target.
To avoid HTC level processing overhead in both host and target,
the unused copy engine 5 is being used for receiving HTT T2H messages.
This will speedup the
Export HTC layer tx and rx handlers. This will be used by HIF layer
for per-CE data processing. Instead of callback mechanism, HIF will
call appropriate upper layers API directly.
Reviewed-by: Michal Kazior
Signed-off-by: Rajkumar Manoharan
---
drivers/net/wireless/ath/ath10k/htc.c | 8
Below patchset adds fast path support for uplink traffic by bypassing
HTC layer processing. This is enabled by making use of unused copy
engine 5 to receive HTT messages directly from HIF layer. From initial
validation in VHT80/5G mode TCP UL is improved to 900Mbps from ~840Mbps
in conducted test.
The function returns always non-negative values.
The problem has been detected using proposed semantic patch
scripts/coccinelle/tests/assign_signed_to_unsigned.cocci [1].
[1]: http://permalink.gmane.org/gmane.linux.kernel/2046107
Signed-off-by: Andrzej Hajda
---
v2: added missing declaration ch
On Thu, Sep 24, 2015 at 08:45:04PM +0200, Arend van Spriel wrote:
> A number of end-users reported issues with BCM43602 on the 2015
> MacBook Pro. It turned out the previously released firmware did
> not properly support that device [1]. Hence an update of the STA
> firmware and while at it refresh
From: Johannes Berg
Date: Wed, 23 Sep 2015 10:44:14 +0200
> It's been pretty quiet for this cycle since the sizeof() mess, but I
> have two more fixes now I'd like to get in.
>
> Let me know if there's any problem.
Pulled, thanks a lot.
--
To unsubscribe from this list: send the line "unsubscri
Hello!
I'm trying to figure out why IBSS with AMSDU breaks on CT firmware (ath10k).
First, I have no idea if it works on any other firmware
The problem I see is that the first frame of the AMPDU
has a single frame, and it looks fine and is received up the stack.
Subsequent frames in the AM
ath10k driver is using dma_pool_alloc per packet and dma_pool_free
in coresponding at Tx completion.
Use of pre-allocated DMA buffer in Tx will improve saving CPU resource
by 5% while it consumes about 56KB memory more as trade off.
Signed-off-by: Peter Oh
---
drivers/net/wireless/ath/ath10k/htt
ath9k inserts padding between the 802.11 header and the data area (to
align it). Since it didn't declare this extra required headroom, this
led to some nasty issues like randomly dropped packets in some setups.
Cc: sta...@vger.kernel.org
Signed-off-by: Felix Fietkau
---
drivers/net/wireless/ath/
With this .config: http://busybox.net/~vda/kernel_config_ALLYES_Os,
after deinlining these functions have sizes and callsite counts
as follows:
iwl_write8: 315 bytes, 3 calls
iwl_write32: 296 bytes, 90 calls
iwl_read32: 278 bytes, 51 calls
Total size reduction is about 40 kbytes.
Signed-off-by:
On Thu, 2015-09-24 at 14:59 +0200, Felix Fietkau wrote:
> When using software queueing, tx sequence number assignment happens
> at
> ieee80211_tx_dequeue time, so the fast-xmit codepath must not do
> that.
>
Applied, but to mac80211-next since there's no driver in 4.3 yet that
actually uses ieee
On Thu, 2015-09-24 at 14:59 +0200, Felix Fietkau wrote:
> Fixes dropped packets in the tx path in case a non-PS station
> triggers
> the tx filter.
>
Applied, but without the stable tag since no driver in 4.2 that uses
IEEE80211_TX_CTL_CLEAR_PS_FILT also had FAST_XMIT.
johannes
--
To unsubscribe
From: Maharaja Kennadyrajan
It is noticed that pci wakeup time is exceeding current timeout (10ms)
randomly which is tested on QCA988x. So, the wake up time is increased
to 30 ms and added debug prints to log total timeout.
Signed-off-by: Maharaja Kennadyrajan
---
drivers/net/wireless/ath/ath1
From: Johannes Berg
There's not all that much value in having a software PS-Poll
implementation, particularly since it's very old and can't
deal with multiple virtual interfaces.
The code still supports a very basic software implementation
of powersave using null-data packets.
Signed-off-by: Jo
From: Johannes Berg
Instead of int with 0/1, use bool with false/true for the
powersave argument to ieee80211_send_nullfunc().
Signed-off-by: Johannes Berg
---
net/mac80211/ieee80211_i.h | 2 +-
net/mac80211/mlme.c| 10 +-
net/mac80211/offchannel.c | 6 +++---
net/mac80211/u
From: Johannes Berg
Instead of nesting two if statements, inline the second
check into the first if statement and to indentation.
Signed-off-by: Johannes Berg
---
net/mac80211/mlme.c | 44
1 file changed, 20 insertions(+), 24 deletions(-)
diff --gi
On Thu, 2015-09-24 at 16:00 +0200, Andrzej Hajda wrote:
> The function returns always non-negative values.
>
> The problem has been detected using proposed semantic patch
> scripts/coccinelle/tests/assign_signed_to_unsigned.cocci [1].
>
You should at least compile your patches.
johannes
--
To un
The function returns always non-negative values.
The problem has been detected using proposed semantic patch
scripts/coccinelle/tests/assign_signed_to_unsigned.cocci [1].
[1]: http://permalink.gmane.org/gmane.linux.kernel/2046107
Signed-off-by: Andrzej Hajda
---
Hi,
To avoid problems with too
Fixes dropped packets in the tx path in case a non-PS station triggers
the tx filter.
Cc: sta...@vger.kernel.org # 4.2
Signed-off-by: Felix Fietkau
---
net/mac80211/status.c | 1 +
net/mac80211/tx.c | 7 +--
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/net/mac80211/stat
When using software queueing, tx sequence number assignment happens at
ieee80211_tx_dequeue time, so the fast-xmit codepath must not do that.
Signed-off-by: Felix Fietkau
---
net/mac80211/tx.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/net/mac80211/tx.c b/net/mac80211/
mutex txq_cs is never used in the driver. txq_cs, txq_critical_section and
txq_lock have same pointer so just delete them and it's related codes.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 5 -
drivers/staging/wilc1000/wilc_wfi_netdevice.h | 1 -
drivers/stag
This patch removes function pointer os_wait which is pointer of
linux_wlan_lock_timeout and just call the real name function directly.
Remove also static from linux_wlan_lock_timeout declaration.
As os_wait is deleted, structure wilc_wlan_os_func_t is useless. Delete
wilc_wlan_os_func_t, os_func an
This patch removes variable cif_func and use mac_cfg directly. No need to have
another pointer variable.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/wilc_wlan.c | 13 +
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/drivers/staging/wilc1000/wilc_wlan.c
b/driv
This patch removes os_debug and call linux_wlan_dbg function instead of
os_debug. Delete static from the linux_wlan_dbg.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 3 +--
drivers/staging/wilc1000/wilc_wfi_netdevice.h | 1 +
drivers/staging/wilc1000/wilc_wlan.c
This patch removes two variables rx_buffer_size and use LINUX_RX_SIZE as
argument.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 3 ---
drivers/staging/wilc1000/wilc_wlan.c| 9 ++---
drivers/staging/wilc1000/wilc_wlan_if.h | 3 ---
3 files changed, 2 insertions(+)
Remove hif_lock and hif_critical_section which are pointer of
g_linux_wlan->hif_cs. Remove also it's related codes.
Just use mutex varialbe g_linux_wlan->hif_cs.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 1 -
drivers/staging/wilc1000/wilc_wlan.c| 10 --
d
This patch removes unnecessary two variables tx_buffer_size and just use
LINUX_TX_SIZE as argument.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 1 -
drivers/staging/wilc1000/wilc_wlan.c| 6 ++
drivers/staging/wilc1000/wilc_wlan_if.h | 1 -
3 files changed, 2 ins
Use mutex variable g_linux_wlan->rxq_cs instead of pointer varialbe.
No need to make extra pointer variable for the mutex. Remove
rxq_critical_section, rxq_lock and it's related codes.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 1 -
drivers/staging/wilc1000/wilc_wlan.
This patch removes txq_add_to_head_critical_section and txq_add_to_head_lock
which are pointer of g_linux_wlan->txq_add_to_head_cs.
Just use g_linux_wlan->txq_add_to_head_cs as argument of function.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 3 ---
drivers/staging/wil
Remove variable txq_wait and txq_wait_event which are pointer varialbe
of g_linux_wlan->txq_event. No need to have extra pointer varialbe. Just use
g_linux_wlan->txq_event. Remove os_context.txq_wait_event, txq_wait and it's
related codes.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/lin
Use spinlock variable g_linux_wlan->txq_spinlock itself instead of
g_wlan.txq_spinlock which is pointer of g_linux_wlan->txq_spinlock.
Delete os_context.txq_spin_lock, g_wlan.txq_spinlock and it's related codes.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 2 --
d
just call the function linux_wlan_rx_complete directly. No need for a pointer
to the functions. Remove rx_complete, wilc_wlan_net_func_t and net_func which
are not used anymore.
Finally remove static from the function linux_wlan_rx_complete.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/l
This patch removes cfg_wait and cfg_wait_event which are pointer of
g_linux_wlan->cfg_event. No need to have pointer varialbe of it. Just use
g_linux_wlan->cfg_event.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 2 --
drivers/staging/wilc1000/wilc_wlan.c| 14 +++
This patch removes function pointer rx_indicate and call the function
frmw_to_linux directly.
Signed-off-by: Glen Lee
---
drivers/staging/wilc1000/linux_wlan.c | 2 --
drivers/staging/wilc1000/wilc_wfi_netdevice.h | 1 +
drivers/staging/wilc1000/wilc_wlan.c | 11 +--
d
No need for a pointer to a function. Just call linux_wlan_mac_indicate.
Remove mac_indicate and also wilc_wlan_indicate_func_t since no members in it.
Variable indicate_func is not used so delete it and related codes.
Finally remove static from the function linux_wlan_mac_indicate.
Signed-off-by:
Fixing Sparse warnings in rtw_security.c. When checking crc, both
actual and expected value was converted to cpu endianness before
comparing, causing sparse warnings as below. Since the values are
read from the buffer in correct byte order the extra conversions
should not be needed.
Thanks to Larr
On Sat, May 9, 2015 at 3:10 PM, Daniel J Blueman
wrote:
After gathering more data, we find this lockout issue occurs only on
5GHz networks, and on average with 15 hours of association (high
variance). We see wl_timer consuming ~19% system time in ksoftirq [1]
and no association or scan data. Sof
42 matches
Mail list logo