There exist mismatch between the size used for pci_map and
pci_unmap on command skb. Correcting it.
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
All the command buffers are freed in mwifiex_free_cmd_buffer()
and hence there is no need to kfree the current command buffer
again. This might ends up freeing memory allocated by some other
kernel code.
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
PCI regions are associated with the device using
pci_request_region() call. Hence use pci_release_region()
instead of pci_release_regions().
Cc: <stable@vger.kernel.org> # 3.2+
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rotate bss prio list, so the bss next to the one served, will come first
in the list of bss' with equal priority. This way we pick bss nodes in a
round robin fashion. Using list rotation instead of a cur ptr simplifies
iteration to calling list_for_each_entry. List rotation is done via
list_move, where the head itself is temporarily removed and then
re-inserted after the bss just served.
Signed-off-by: Andreas Fenkart <andreas.fenkart@streamunlimited.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
After a packet is successfully transmitted, ra list is rotated, so the ra
next to the one transmitted, will be the first in the list. This way we
pick the ra' in a round robin fashion. This significantly simplifies
iteration in mwifiex_wmm_get_highest_priolist_ptr to a call to
list_for_each_entry.
List rotation is done via list_move, where the head itself is temporarily
removed and then re-inserted after the item just transferred.
Signed-off-by: Andreas Fenkart <andreas.fenkart@streamunlimited.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Initialization of bss_descriptor is unnecessary as the entire
structure will be overwritten by a memcpy followed by.
Initialize disable_11ac flag properly by setting it to true in
mwifiex_fill_new_bss_desc().
Reported-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Paul Stewart <pstew@chromium.org>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
beacon_ie buffer is allocated in mwifiex_fill_new_bss_desc()
and the buffer pointer is saved in bss_desc->beacon_buf.
beacon_ie is freed before the function returns. However,
bss_desc->beacon_buf is still being accessed afterwards.
Fix it by freeing beacon_ie (bss_desc->beacon_buf) in
caller's scope.
Reviewed-by: Doug Anderson <dianders@chromium.org>
Reviewed-by: Paul Stewart <pstew@chromium.org>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We have blocked association attempts on interfaces configured in
AP and AD-HOC modes. P2P mode should be blocked too.
Furthermore, an error code must be returned if we are unable to
associate.
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
priv->bss_mode uses NL80211_IFTYPE_* definitions.
HostCmd_BSS_MODE_IBSS is used in ad-hoc start/join command between
driver and firmware.
Coincidentally both HostCmd_BSS_MODE_IBSS and NL80211_IFTYPE_STATION
are defined as 2. That explains why nobody complained.
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
There is a typo in mwifiex_cleanup_rxq_ring() which uses
PCI_DMA_TODEVICE while unmapping PCI memory.
We should actually use PCI_DMA_FROMDEVICE.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Send P2P_MODE_CFG cmd to firmware when p2p interface is created.
Without proper p2p configuration firmware may behave incorrectly
while handling commands sent through this interface.
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Stone Piao <piaoyun@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
p2p devices should identify themselves as such to userspace at
startup, so the connection manager can decide which interface
to start wpa_supplicant instances on.
Signed-off-by: Paul Stewart <pstew@chromium.org>
Reviewed-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Not locking ra_list when dequeuing packets creates race conditions.
When adding a packet 'tx_pkts_queued' is modified before setting
highest_priority_queue. If in-between the main loop starts, it will
see a packet queued (tx_pkts_queued > 0) but will not find it, since
max prio is not set yet. Depending on the scheduling, the thread
trying to add the packet could complete and restore the situation.
But this is not something to rely on.
Another race condition exists, if a new packet, exceeding current
max prio is added. If concurrently a packet is dequeued, the newly
set max prio will be overwritten with the value of the dequeued
packet. This can occur, because selecting a packet and modifying
the max prio is not atomic. The result in an infinite loop unless,
a new packet is added that has at least the priority of the hidden
packet.
Same applies to bss_prio_tbl. Forward iteration is no proper
lock-free technique and provides no protection from calls to
list_del. Although BSS are currently not added/removed dynamically,
this must not be the case in the future. Hence always hold proper
locks when accessing those lists.
Signed-off-by: Andreas Fenkart <andreas.fenkart@streamunlimited.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Using NO_PKT_PRIO_TID and tx_pkts_queued to check for an empty
state, can lead to a contradictory state, resulting in an
infinite loop. Currently queueing and dequeuing of packets is
not synchronized, and can happen concurrently. While tx_pkts_queued
is incremented when adding a packet, max prio is set to NO_PKT when
the WMM list is empty. If a packet is added right after the check
for empty, but before setting max prio to NO_PKT, that packet is
trapped and creates an infinite loop.
Because of the new packet, tx_pkts_queued is at least 1, indicating
wmm lists are not empty. Opposing that max prio is NO_PKT, which
means "skip this wmm queue, it has no packets". The infinite loop
results, because the main loop checks the wmm lists for not empty
via tx_pkts_queued, but for dequeing it uses max_prio to see if it
can skip current list. This will never end, unless a new packet is
added which will restore max prio to the level of the trapped packet.
The solution here is to rely on tx_pkts_queued solely for checking
wmm queue to be empty, and drop the NO_PKT define. It does not
address the locking issue.
Signed-off-by: Andreas Fenkart <andreas.fenkart@streamunlimited.com>
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
ra_list_spinlock is used to protect struct mwifiex_wmm_desc and
embedded structures such as ra_list. tid_tbl_lock while more fine
grained, is not used but in one function. That function is not
called reentrantly. To protect ra_list from concurrent modification
ra_list_spinlock must be held.
Signed-off-by: Andreas Fenkart <andreas.fenkart@streamunlimited.com>
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
adapter->bss_prio_tbl list has already been checked in outer loop.
The inner loop works with priv_tmp->wmm.tid_tbl_ptr list. Also the
lock taken, gives hint that this is likely a copy-paste error.
Signed-off-by: Andreas Fenkart <andreas.fenkart@streamunlimited.com>
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
cmd_pending is increased in mwifiex_wait_queue_complete() and
decreased in mwifiex_complete_cmd() currently.
If there are two or more commands in the cmd_pending_q the main
worker thread will pick up next command from cmd_pending_q
automatically after finishing current command. As a result
mwifiex_wait_queue_complete() will not be called because
the command is alreay completed. This leads to a negative
number in cmd_pending count.
Fix it by increasing cmd_pending when a cmd is queued into
cmd_pending_q and decreasing when that cmd is recycled. For scan
commands we don't perform inc/dec operations until it's moved
from scan_pending_q to cmd_pending_q. This covers both
synchronous and asynchronous commands.
Reported-by: Daniel Drake <dsd@laptop.org>
Tested-by: Daniel Drake <dsd@laptop.org>
Tested-by: Marco Cesarano <marco@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We are waiting on first scan command of internal scan request
before association, so we should complete on last internal scan
command response.
Cc: <stable@vger.kernel.org> # 3.8
Tested-by: Daniel Drake <dsd@laptop.org>
Tested-by: Marco Cesarano <marco@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Newer 11ac enabled chipsets have more TX and RX buffers in FW
and hardware; so they may support larger TX and RX window sizes
for BA. Reset BA settings during association, adhoc join/start
or start_ap() if we are joining/creating 11ac network.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Sagar Bijwe <bsagar@marvell.com>
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch fixes an issue where RX throughput values observed
were substantially lower than TX counterparts for PCIe8897 STA.
PCIe8897 supports larger rx_win_size. After changing these values
we see big improvement for TX and RX throughput values.
Different tx_win_size and rx_win_size are used for AP mode.
All BA setup related initialization has been moved to separate
function.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Sagar Bijwe <bsagar@marvell.com>
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Limit the channel number in scan request, or the driver scan
config structure memory will be overflowed.
Cc: <stable@vger.kernel.org> # 3.5+
Signed-off-by: Stone Piao <piaoyun@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
FW can be woken up even by accessing device registers; we need
not explicitily enable interrupts for doing this. Future PCIe
devices will not be woken up by writing to host registers.
This patch enables driver to wake up device by reading FW status
register.
Also devices with sleep cookie enabled need some more time before
proceeding with processing. Handle this by adding a delay loop.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We have received interrupt from device means FW is not sleeping.
In this case make sure wakeup handler for PCIe is not invoked by
setting adapter->pm_wakeup_fw_try to false.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Enabling PCIe host interrupt may accidently wake up the card
when it's in sleep.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Currently, default VHT configuration from the firmware is used
for the VHT operations. Adding vhtcfg command to configure the
firmware based on input received from cfg.
Enable VHT for AP mode only when cfg80211_ap_settings has a
VHT IE i.e., when ieee80211ac is set to 1 in the hostapd.conf.
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch is a bug fix for an issue wherein power save was not
working for PCIe. This happens because for processing power save
sleep confirm command we pull skb so that skb->data points ahead
of interface header. We use same skb to get other cmda responses
as well. So if we don't push skb after processing cmd response,
it results into reduction in skb->len and finally skb->len reaches
zero. This causes failure in processing sleep command response.
Fix this by pushing skb by INTF_HEADER_LEN at the end of command
response processing.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Firmware returned VHT cap has the same format that cfg80211
expects. There is no need to parse the vht cap from the firmware
and then set it to ieee80211_sta_vht_cap. Just copying is
sufficient.
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
curr_cmd points to the command that is in processing or waiting
for its command response from firmware. If the function shutdown
happens to occur at this time we should cancel the cmd timer and
put the command back to free queue.
Cc: <stable@vger.kernel.org> # 3.8
Tested-by: Marco Cesarano <marco@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
During rmmod mwifiex_sdio processing FUNC_SHUTDOWN command is
sent to firmware. Firmware expcets only FUNC_INIT once WLAN
function is shut down.
Any command pending in the command queue should be ignored and
freed.
Cc: <stable@vger.kernel.org> # 3.8
Tested-by: Daniel Drake <dsd@laptop.org>
Tested-by: Marco Cesarano <marco@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Running the following script repeatedly on XO-4 with SD8787
produces command timeout and system lockup.
insmod mwifiex_sdio.ko
sleep 1
ifconfig eth0 up
iwlist eth0 scan &
sleep 0.5
rmmod mwifiex_sdio
mwifiex_send_cmd_async() is called for sync as well as async
commands. (mwifiex_send_cmd_sync() internally calls it for
sync command.)
"adapter->cmd_queued" gets filled inside mwifiex_send_cmd_async()
routine for both types of commands. But it is used only for sync
commands in mwifiex_wait_queue_complete(). This could lead to a
race when two threads try to queue a sync command with another
sync/async command simultaneously.
Get rid of global variable and pass command node as a parameter
to mwifiex_wait_queue_complete() to fix the problem.
Cc: <stable@vger.kernel.org> # 3.8
Reported-by: Daniel Drake <dsd@laptop.org>
Tested-by: Daniel Drake <dsd@laptop.org>
Tested-by: Marco Cesarano <marco@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
tx_timeout doesn't always lead to a cmd_timeout. There are
occurrences where cmd_timeout never gets triggered for a long
time and we encounter a kernel crash. In this patch, we track
the consecutive timeouts (tx_timeout_cnt). When tx_timeout_cnt
exceeds the threshold, trigger a card reset thereby avoiding a
kernel crash.
Signed-off-by: Ashok Nagarajan <asnagarajan@chromium.org>
Signed-off-by: Paul Stewart <pstew@chromium.org>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Host sleep wakeup condition is configured using this command.
Supports Wake-on: pumb
For examples:
wake-on any unicast packets:
ethtool -s mlan0 wol u
wake-on multicast/broadcast packet:
ethtool -s mlan0 wol mb
wake-on unicast packets and MAC events:
ethtool -s mlan0 wol pu
wake-on unicast/multicast/broadcast packets and MAC events:
ethtool -s mlan0 wol pmbu
disable all wake-on options:
ethtool -s mlan0 wol d
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
As we are adding a few more macros in this category in next
patch, this cleanup work is required.
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We allocate SKB buffers of 4K size to make sure that we process
RX AMSDU of 4K. So when skb->len is lesser than 4K; we should
modify skb->truesize. This resolves an issue where kernel has
allocated packets with 2K assumption and starts dropping packets
for large size data transfer.
This fix is already present for USB; extend it to PCIE.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
drivers/net/wireless/mwifiex/pcie.c:1157:9: warning:
'desc2' may be used uninitialized in this function
[-Wmaybe-uninitialized]
drivers/net/wireless/mwifiex/pcie.c:1048:31: note:
'desc2' was declared here
drivers/net/wireless/mwifiex/pcie.c:1159:9: warning:
'desc' may be used uninitialized in this function
[-Wmaybe-uninitialized]
drivers/net/wireless/mwifiex/pcie.c:1047:32: note:
'desc' was declared here
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
drivers/net/wireless/mwifiex/pcie.c:204:12: warning:
'mwifiex_pcie_resume' defined but not used [-Wunused-function]
drivers/net/wireless/mwifiex/pcie.c:166:12: warning:
'mwifiex_pcie_suspend' defined but not used [-Wunused-function]
The suspend/resume handlers ought to be under CONFIG_PM directive.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
move functions up just to avoid static forward declaration for
mwifiex_pcie_resume.
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Maximum delay for waking up card is 50 ms. Because of typo in
counter, this delay goes to 500ms. This patch fixes the bug.
Cc: <stable@vger.kernel.org> # 3.2+
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Unload sequence for mwifiex PCIE driver is as follows:
1. Invoking cleanup module from kernel results into
pci_unregister_driver
2. Kernel invokes PCIE remove() handler which disconnects all
interfaces.
3. One step during disconnect is to clean PCIE TX rings.
During this we read txbd_rdptr from FW.
While loading driver next time, we see pci_enable_device()
results into system freeze. This may have happened because we
accessed PCI device after unregistering from bus driver.
Removing this ioread() operation resolves this bug.
Signed-off-by: Avinash Patil <patila@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>