Sara Sharon [Thu, 17 Nov 2016 11:57:32 +0000 (13:57 +0200)]
iwlwifi: mvm: support new binding API
For a000 devices the binding API needs to include relevant
lmac ID - support the new API.
The new API should be used regardless if the device had CDB or
not. If there is no actual CDB support the binding is bound
to first lmac regardless of the band.
There are some functionality changes in binding restrictions
and quota allocations that will be handled in future patches.
When IEEE80211_CHAN_RADAR is set by cfg80211, passive scan must be
performed. In mwifiex,active scan is performed even though flag is
set from cfg80211. mwifiex_reg_apply_radar_flags() function added
in this patch correctly uses radar flag.
rtlwifi: btcoex: 23b 2ant: new fw use h2c to control GNT_BT
Earlier versions of the FW did not support h2c to set GNT_BT, but later
versions have that capability. Hence we check the FW version and decide
whether to use h2c or just write to the register.
rtlwifi: btcoex: 23b 2ant: add comments to describe how duration adjusted
Since btcoex uses static variables to store the cumulative information
on the wifi status, some tricks are used that are a bit obscure. We add
some comments about the criteria we use to adjust wifi duration.
rtlwifi: btcoex: 23b 2ant: finer bt power adjustment
The bt can specify a power level to decrease. Rather than decreasing by a
fixed value (usually 2), this change makes the driver fit to the
environment more quickly.
rtlwifi: btcoex: 23b 2ant: adjust wifi duration for bt a2dp
The larger the bt a2dp bit pool is, the more time bt needs to receive
them. If we do not adjust the wifi duration, the voice quality will be
low. Hence we reduce the time that wifi holds, to improve the a2dp
service.
rtlwifi: btcoex: 23b 2ant: monitor wifi counter for allocate tdma time
If there are too many CRC error packets, it means the wifi is operating
under low quality circumstances. If so, we need to reallocate the
resources of wifi and bt
rtlwifi: btcoex: 23b 2ant: add btc8723b2ant_limited_rx to reduce agg size
For wifi and bt coexistence, if the aggregation size of wifi is too
large, the transmission time of the aggregated packet could be too long
and the bt packets might "cut off" the wifi packet. We need to reduce
the aggregation size to improve wifi performance.
mwifiex: add missing IEs related to TDLS operation
In mwifiex,IEs such as supported channels, supported operating classes
20/40 BSS COexistence are missing and also extra QOS capabilities IE
is added during TDLS discovery response, TDLS setup request and
TDLS setupresponse.
This patch adds require IEs and also removes extra IE.
Brian Norris [Tue, 28 Mar 2017 23:59:33 +0000 (16:59 -0700)]
mwifiex: catch mwifiex_fw_dpc() errors properly in reset
When resetting the device, we take a synchronous firmware-loading code
path, which borrows a lot from the asynchronous path used at probe time.
We don't catch errors correctly though, which means that in the PCIe
driver, we may try to dereference the 'adapter' struct after
mwifiex_fw_dpc() has freed it. See this (erronous) print in
mwifiex_pcie_reset_notify():
Let's instead refactor the synchronous (or "!req_fw_nowait") path so
that we propagate errors and handle them properly.
This fixes a use-after-free issue in the PCIe driver, as well as a
misleading debug message ("successful"). It looks like the SDIO driver
doesn't have these problems, since it doesn't do anything after
mwifiex_reinit_sw().
Fixes: 4c5dae59d2e9 ("mwifiex: add PCIe function level reset support") Signed-off-by: Brian Norris <[email protected]> Signed-off-by: Kalle Valo <[email protected]>
Brian Norris [Tue, 28 Mar 2017 23:59:32 +0000 (16:59 -0700)]
mwifiex: fix use-after-free for FW reinit errors
If we fail to reinit the FW when resetting the device (in the
synchronous version of mwifiex_init_hw_fw() -> mwifiex_fw_dpc()),
mwifiex_fw_dpc() will tear down the interface and free up the adapter.
But we don't actually check for all failure cases of mwifiex_fw_dpc(),
so some of them fall through and dereference adapter->fw_done with a
freed adapter, causing a use-after-free bug.
In any case, mwifiex_fw_dpc() will always signal FW completion -- in the
error OR success case -- so at best, this was repeat work. Let's not do
it.
Arend Van Spriel [Tue, 28 Mar 2017 10:43:28 +0000 (11:43 +0100)]
brcmfmac: no need for d11inf instance in brcmf_pno_start_sched_scan()
In brcmf_pno_start_sched_scan() a local variable is declared and
assigned for struct brcmu_d11inf. However, there is no other reference
to it so it is unnecessary.
Arend Van Spriel [Tue, 28 Mar 2017 10:43:27 +0000 (11:43 +0100)]
brcmfmac: restore bus state when enter_D3 fails
In brcmf_pcie_suspend() we inform the firmware on the device that
it will enter in D3 state. Before this is done we already bring down
the bus state. However, When entering D3 fails we abort the suspend
and the bus state need to be restored.
Arend Van Spriel [Tue, 28 Mar 2017 10:43:26 +0000 (11:43 +0100)]
brcmfmac: add support to move wiphy instance into network namespace
To support network namespace the driver must assure all created
network interfaces are in the same namespace as the wiphy instance
and flag the support using WIPHY_FLAG_NETNS_OK.
Verified using two terminals:
Terminal 1 Terminal 2
-------------------------- ---------------------------------
# ip netns add brcm-wifi # iw dev
phy#0
Interface wlan3
ifindex 11
wdev 0x1
# ip netns exec brcm-wifi bash
# iw dev
# echo $$
20337 # iw phy0 set netns 20337
# iw dev
phy#0
Interface wlan3
ifindex 11
wdev 0x2
# iw phy0 interface add wl3.ap type __ap
# iw dev
phy#0
Interface wl3.ap
ifindex 2
wdev 0x3
Interface wlan3
ifindex 11
wdev 0x2
# iw dev
# iw phy0 set netns 1
# iw dev
# iw dev
phy#0
Interface wl3.ap
ifindex 2
wdev 0x5
Interface wlan3
ifindex 11
wdev 0x4
Note:
increasing wdev identifier above indicates issue in
cfg80211 which is addressed separately.
Franky Lin [Tue, 28 Mar 2017 10:43:24 +0000 (11:43 +0100)]
brcmfmac: wrap brcmf_fws_init into bcdc layer
Create a new protocol layer interface brcmf_proto_init_cb for protocol
layer to finish initialzation after core module components(fweh and
etc.) are initialized.
Ganapathi Bhat [Mon, 27 Mar 2017 12:57:44 +0000 (18:27 +0530)]
mwifiex: Support USB interrupt endpoint for command response/event
USB firmware added support for sending command response/event through
interrupt endpoint, to enhance RX throughput. Added corresponding changes
required to support this feature. This change takes care of backward
compatibility with older firmware.
Rafał Miłecki [Thu, 23 Mar 2017 10:29:28 +0000 (11:29 +0100)]
brcmfmac: update BRCMFMAC symbol description
For quite some time now brcmfmac supports 802.11ac chipsets and it's
not limited to embedded devices only. There are even standalone PCIe
cards based on BCM43602 or BCM4366.
Roman Yeryomin [Mon, 20 Mar 2017 23:43:00 +0000 (00:43 +0100)]
rt2x00: add support for MT7620
Basic support for MT7620 built-in wireless radio was added to
OpenWrt in r41441. It has seen some heavy cleaning and refactoring
since in order to match the Kernel's code quality standards.
Johan Hovold [Mon, 13 Mar 2017 12:44:21 +0000 (13:44 +0100)]
zd1211rw: fix NULL-deref at probe
Make sure to check the number of endpoints to avoid dereferencing a
NULL-pointer or accessing memory beyond the endpoint array should a
malicious device lack the expected endpoints.
Fixes: a1030e92c150 ("[PATCH] zd1211rw: Convert installer CDROM device into WLAN device") Cc: Daniel Drake <[email protected]> Signed-off-by: Johan Hovold <[email protected]> Signed-off-by: Kalle Valo <[email protected]>
Franky Lin [Fri, 10 Mar 2017 21:17:04 +0000 (21:17 +0000)]
brcmfmac: wrap brcmf_fws_add_interface into bcdc layer
fwsignal is only used by bcdc. Create a new protocol interface function
brcmf_proto_add_if for core module to notify protocol layer upon a new
interface is created.
Franky Lin [Fri, 10 Mar 2017 21:17:02 +0000 (21:17 +0000)]
brcmfmac: move brcmf_txflowblock to bcdc layer
brcmf_txflowblock is invoked by sdio and usb bus module which are using
bcdc protocol. This patch makes it a bcdc API instead of a core module
function.
Hans de Goede [Wed, 8 Mar 2017 13:50:17 +0000 (14:50 +0100)]
brcmfmac: Handle status == BRCMF_E_STATUS_ABORT in cfg80211_escan_handler
If a scan gets aborted BRCMF_SCAN_STATUS_BUSY gets cleared in
cfg->scan_status and when we receive an abort event from the firmware
the BRCMF_SCAN_STATUS_BUSY check in the cfg80211_escan_handler will
trigger resulting in multiple errors getting logged.
Check for a status of BRCMF_E_STATUS_ABORT and in this case simply
cleanly exit the cfg80211_escan_handler. This also avoids a
BRCMF_E_STATUS_ABORT event arriving after a new scan has been started
causing the new scan to complete prematurely without any data.
Hans de Goede [Wed, 8 Mar 2017 13:50:15 +0000 (14:50 +0100)]
brcmfmac: Do not print the firmware version as an error
Using pr_err for things which are not errors is a bad idea. E.g. it
will cause the plymouth bootsplash screen to drop back to the text
console so that the user can see the error, which is not what we
normally want to happen.
Rafał Miłecki [Fri, 24 Feb 2017 16:32:46 +0000 (17:32 +0100)]
brcmfmac: always print error when PSM's watchdog fires
So far we were attaching BRCMF_E_PSM_WATCHDOG event listener in
brcmf_debug_attach which gets compiled only with CONFIG_BRCMDBG. This
event means something went wrong and firmware / hardware usually can't
be expected to work (reliably).
Such a problem is significant for user experience so I believe we should
print an error unconditionally (even with debugging disabled). What can
be indeed optional is dumping bus memory as this is clearly part of
debugging process.
In the future we may also try to extend this listener by trying to
recover from the error or at least signal it to the cfg80211.
Rafał Miłecki [Fri, 3 Mar 2017 10:33:30 +0000 (11:33 +0100)]
bcma: drop unneeded check for CONFIG_OF_IRQ
We already have the same check in bcma_of_get_irq which really calls
symbols available with CONFIG_OF_IRQ only. It appears this duplicated
check was accidentally added in commit c58d900cc96a ("bcma: fix building
without OF_IRQ"). The rest of code in bcma_of_fill_device should work
fine without CONFIG_OF_IRQ.
Larry Finger [Thu, 23 Feb 2017 17:19:56 +0000 (11:19 -0600)]
rtlwifi: Add code to read new versions of firmware
Changes in the drivers for RTL8723BE and RTL8821AE require corresponding
changes in the firmware. This new firmware has been accepted into the
Linux firmware repo. To handle the case where the kernel has been
updated before the firmware, the new versions have been given new names.
The code will attempt to read the new name, and fall back to the old one
if the new one is not available.
Brian Norris [Thu, 16 Mar 2017 21:36:07 +0000 (14:36 -0700)]
mwifiex: fix kernel crash after shutdown command timeout
We observed a SHUTDOWN command timeout during reboot stress test due to
a corner case firmware bug. It can lead to either a use-after-free +
OOPS (on either the adapter structure, or the 'card' structure) or an
abort (where, e.g., the PCI device is "disabled" before we're done
dumping the FW).
We can avoid this by canceling/flushing the FW dump work:
(a) after we've terminated all other work queues (e.g., for processing
commands which could time out)
(b) after we've disabled all interrupts (which could also queue more
work for us)
(c) after we've unregistered the netdev and wiphy structures (and
implicitly, and debugfs entries which could manually trigger FW dumps)
(d) before we've actually disabled the device (e.g.,
pci_device_disable())
Altogether, this means no card->work will be scheduled if we sync at
a point that satisfies the above. This can be done at the beginning of
the .cleanup_if() callback.
Amitkumar Karwar [Thu, 16 Mar 2017 10:46:11 +0000 (16:16 +0530)]
mwifiex: send fewer channels to scan while connected
Application triggers periodic background scans when device is connected.
We will scan less number of channels per scan command so that data
traffic won't get affected.
Devidas Puranik [Thu, 9 Mar 2017 08:36:17 +0000 (14:06 +0530)]
mwifiex: fix for unaligned reads
Using the accessor function e.g. get_unaligned_le32 instead of
le32_to_cpu to avoid the unaligned access. This is for the
architectures that don't handle the unaligned memory access
Daniel Mentz [Thu, 9 Mar 2017 08:36:16 +0000 (14:06 +0530)]
mwifiex: Use accessors routines for unaligned values
Synopsys' ARCompact architecture does not support loading from or
storing values to unaligned memory locations. We saw a series of
misaligned access exceptions on ARC. To work around this issue, we bulk
replaced le16_to_cpu and le32_to_cpu with get_unaligned_le16 and
get_unaligned_le32, respectively. We also added le16_unaligned_add_cpu
which is similar to le16_add_cpu but works with unaligned values.
Adding qualifier "__packed" indicates that no padding should be
performed on the qualified object for alignment.
This patch adds qualifier __packed to the required firmware
structures in mwifiex driver.
Brian Norris [Tue, 7 Mar 2017 20:47:46 +0000 (12:47 -0800)]
mwifiex: pcie: clean up error prints in mwifiex_pcie_reset_notify()
We shouldn't be printing a kernel pointer as a decimal integer. But we
really shouldn't be printing this case at all; we should never get here
with NULL drvdata. We've eliminated this unnecessary conditional in
several other places, so kill it here too.
Similarly, there's no need to check for '!pdev'; we are guaranteed to
have a real device here.
And finally, use dev_err() instead of pr_err().
This yields (for failed PCIe resets):
[ 68.286586] mwifiex_pcie 0000:01:00.0: mwifiex_pcie_reset_notify: adapter structure is not valid
instead of:
[ 82.932658] mwifiex_pcie: mwifiex_pcie_reset_notify: Card or adapter structure is not valid (-270880688088)
Jeffy Chen [Fri, 24 Feb 2017 06:24:31 +0000 (14:24 +0800)]
mwifiex: wake system up when receives a wake irq
Currrently we are disabling this wake irq after receiving it. If this
happens before we finish suspend and the pm event check is disabled,
the system will continue suspending, and this irq would not work again.
We may need to abort system suspend to avoid that.
Channels 34/38/42/46 can only be used for compatibility with old devices sold
in Japan. Modern products, such as AR6003/AR6004 don't support these channels.
Keeping them in the upstream is error prone and requires full network stack
support. A custom patch should be used in case such compatibility is required.
Without this one, a user is able to start an AP using wpa_supplicant,
for example, on one of these channels (34/38/42/46), without getting
any warning/error from the cfg80211 or ath6kl - which is correct
(since these channels match regdom rules). However, the AR6003 and its
firmware (we're using v3.4.0.225) will fail and return
"WMI_CMDERROR_EVENTID" with "INVALID_PARAM" error code.
Kalle Valo [Thu, 16 Mar 2017 09:11:02 +0000 (11:11 +0200)]
ath10k: fix warnings from an earlier commit
I failed to notice that commit 523f6701dbab ("ath10k: update available channel
list for 5G radio") added two new warnings:
drivers/net/wireless/ath/ath10k/mac.c:3129:6: warning: symbol 'ath10k_mac_update_channel_list' was not declared. Should it be static?
drivers/net/wireless/ath/ath10k/mac.c:3170: Alignment should match open parenthesis
ath10k: fix fetching channel during potential radar detection
Fetch target operating channel during potential radar detection when
the interface is just brought up, but no channel is assigned from
userspace. In this scenario rx_channel may not be having a valid pointer
hence fetch the target operating channel to avoid warnings as below
which can be triggered by the commands with DFS testing over longer run
Once the CAC expires, current channel context will be removed and
we are only left with the fallback option of using 'target operating
channel'
Firmware and driver log:
ath: phy1: DFS: radar found on freq=5300: id=1, pri=1125, count=5,
count_false=4
ath: phy1: DFS: radar found on freq=5260: id=5, pri=3151, count=6,
count_false=11
ath: phy1: DFS: radar found on freq=5280: id=1, pri=1351, count=6,
count_false=4
ath: phy1: DFS: radar found on freq=5300: id=1, pri=1125, count=5,
count_false=4
ath10k_pci 0001:01:00.0: failed to derive channel for radar pulse,
treating as radar
ath10k_pci 0001:01:00.0: failed to derive channel for radar pulse,
treating as radar
Call trace:
WARNING: CPU: 1 PID: 2145 at
backports-20161201-3.14.77-9ab3068/net/wireless/chan.c:265
cfg80211_set_dfs_state+0x3c/0x88 [cfg80211]()
Workqueue: phy1 ieee80211_dfs_radar_detected_work
[mac80211]
[<c0320770>] (warn_slowpath_null) from [<bf79b90c>]
(cfg80211_set_dfs_state+0x3c/0x88 [cfg80211])
[<bf79b90c>] (cfg80211_set_dfs_state [cfg80211]) from
[<bf79697c>] (cfg80211_radar_event+0xc4/0x140 [cfg80211])
[<bf79697c>] (cfg80211_radar_event [cfg80211]) from
[<bf83c058>] (ieee80211_dfs_radar_detected_work+0xa8/0xb4 [mac80211])
[<bf83c058>] (ieee80211_dfs_radar_detected_work
[mac80211]) from [<c0339518>] (process_one_work+0x298/0x4a4)
Colin Ian King [Tue, 28 Feb 2017 11:14:39 +0000 (11:14 +0000)]
ath10k: remove redundant error check
The check on ret for an error is redundant because it is already been
checked for non-zero earlier on and ret is never non-zero at this point.
Fix this by removing the redundant check and error message.
Detected by CoverityScan, CID#1357170 ("Logically Dead Code")
Tamizh chelvam [Thu, 23 Feb 2017 13:18:22 +0000 (18:48 +0530)]
ath10k: update available channel list for 5G radio
If a 5 GHz radio is calibrated for operation in both
the low band (channels 36 to 64) and high band(channels 100 to 169),
hardware allows operations in all the listed channels. However,
if the chip has been calibrated only for the low/high band and
a high/low band channel is configured, due to lack of calibration
there will be potentially invalid signal on those non calibrated channels.
To avoid this problem this patch sets IEEE80211_CHAN_DISABLED flag for
those non calibrated channels by using low_5ghz_chan and high_5ghz_chan
values which we get from target through wmi service ready event.
Driver initialized flags are getting re initialized in handle_channel
in cfg80211. So calling the function to disable the non supported channel
from reg_notifier().
ath10k: disallow DFS simulation if DFS channel is not enabled
If DFS is not enabled in hostapd (ieee80211h=0) DFS channels shall
not be available for use even though the hardware may have the capability
to support DFS. With this configuration (DFS disabled in hostapd) trying to
bring up ath10k device in DFS channel for AP mode fails and trying to
simulate DFS in ath10k debugfs results in a warning in cfg80211 complaining
invalid channel and this should be avoided in the driver itself rather than
false propogating RADAR detection to mac80211/cfg80211. Fix this by
checking for the first vif 'is_started' state(should work for client mode
as well) as all the vifs shall be configured for the same channel
Lior David [Wed, 8 Mar 2017 11:52:19 +0000 (13:52 +0200)]
wil6210: add oob_mode for AP certification
Add a new value to the oob_mode module parameter for
supporting AP certification.
All enabled values of oob_mode (>0) are intended only
for debugging and diagnostics.
Dedy Lansky [Wed, 8 Mar 2017 11:52:18 +0000 (13:52 +0200)]
wil6210: correctly report locally generated disconnect in STA mode
Driver always invoke cfg80211_disconnected() with locally_generated as
false.
Fix this by reporting true whenever the disconnect is triggered from
upper layers (cfg80211) or from within the driver itself (reset,
deinit).
Dedy Lansky [Wed, 8 Mar 2017 11:52:17 +0000 (13:52 +0200)]
wil6210: use WMI_DISCONNECT_CMDID upon connect timeout
Upon connect timeout driver invokes _wil6210_disconnect() which iterates
over sta array and disconnects each connected sta. In practice, because
the connection is still ongoing and because cid is not yet allocated,
disconnect is not actually happening. This leaves FW in connecting
state while driver is in disconnected state.
To fix this, upon connect timeout, explicitly send WMI_DISCONNECT_CMDID
to FW to make sure it gets disconnected.
Hamad Kadmany [Wed, 8 Mar 2017 11:52:16 +0000 (13:52 +0200)]
wil6210: protect list of pending wmi events during flush
When flush is done, pending events list is manipulated
without taking the proper spinlock, which could lead to
memory corruption if list is manipulated by wmi worker
or by interrupt routine.
Dedy Lansky [Wed, 8 Mar 2017 11:52:15 +0000 (13:52 +0200)]
wil6210: store bss object and use cfg80211_connect_bss()
In a fast disconnect/connect sequence, cfg80211_connect_result() can
fail to find the bss object which the driver is connecting to. Detailed
sequence of events:
* Driver is connected in STA mode
* Disconnect request arrives from user space. Driver disconnects and
calls cfg80211_disconnected() which adds new event to the
cfg80211_wq worker thread
* Connect request arrives from user space. cfg80211_connect() stores
ssid/ssid_len and calls rdev_connect()
* __cfg80211_disconnected() runs in worker thread and zero
wdev->ssid_len
* Connect succeeds. Driver calls cfg80211_connect_result() which fails
to find the bss because wdev->ssid_len is zero
To overcome this, upon connect request, store the bss object in the
driver and upon connect completion pass it to kernel using
cfg80211_connect_bss().
Maya Erez [Wed, 8 Mar 2017 11:52:14 +0000 (13:52 +0200)]
wil6210: missing reinit_completion in HALP voting
After setting HALP ICR bit, we keep it set until HALP unvote.
Masking HALP ICR should protect the driver from hitting the HALP ICR
over and over again. However, in case there is another MISC ICR
we will read the HALP ICR and issue a completion. This can lead to
a case where HALP voting is completed immediately, as the completion
is already set.
Reinit the HALP completion before the actual vote will clear previous
completions and protect from such cases.
Dedy Lansky [Wed, 8 Mar 2017 11:52:13 +0000 (13:52 +0200)]
wil6210: use print_hex_dump_debug instead of print_hex_dump_bytes
Some dynamic debug printouts in driver are using print_hex_dump_bytes.
However, with dynamic debug disabled, print_hex_dump_bytes outputs to
log unconditionally.
Use print_hex_dump_debug instead to prevent log pollution when dynamic
debug disabled.
The driver uses the bus_request platform operation to
request resources from the platform for a specific bandwidth.
Currently the driver requests resources for the maximum
theoretical bandwidth, when interface is brought up.
Refine this process a bit: now the driver will request a
small amount of resources when interface is up, and will only
issue the maximum request when connected.
This mechanism will be improved further in the future to make
more refined requests based on actual bandwidth.
ath10k: fix a warning during channel switch with multiple vaps
Doing a channel switch via hostapd_cli seems to update
the new channel context for each VAP's appropriately as below
in 'ath10k_mac_update_vif_chan', hence we can safely suppress the
warning that shows up during this operation and dump the warning only
if no vaps are available for channel switch
Ryan Hsu [Wed, 8 Mar 2017 11:52:04 +0000 (13:52 +0200)]
ath10k: improve the firmware download time for QCA6174
Len Brown reported the system resume time is taking more than 2 seconds in
bug - https://bugzilla.kernel.org/show_bug.cgi?id=185621.
The reason of the 2 seconds is due to the firmware download time.
The chip is booted up in the default reference clock speed to handle the
firmware download to chip memory and advanced to the support higher speed
clock to run the firmware after all. The default reference clock in the
hardware is slow so that the firmware download time is taking up to 2
seconds for a 600KB firmware file.
[76796.349701] ath10k_pci : boot uploading firmware image len 688691
[76798.334612] ath10k_pci : htt tx max num pending tx 1056
The resolution here is to enable the higher speed clock if the hardware
supported before the firmware download at BMI stage, so that the hardware
can handle the firmare download in a more efficient way. This can help to
improve the firmware download time from 2 seconds to around 500ms for the
same 600KB firmware file.
[322858.577919] ath10k_pci boot uploading firmware image len 688691
[322859.093094] ath10k_pci htt tx max num pending tx 1056
The steps to advance to the higher speed clock is very hardware specific,
so adding the hardware ops for the hardware that can support this.
Nils Holland [Sun, 19 Feb 2017 23:59:28 +0000 (00:59 +0100)]
rtl8187: Enable monitor mode to fix multicast reception
The rtl8187 cards don't seem to receive multicast frames, which,
among other things, makes them fail to receive RAs in IPv6 networks.
The cause seems to be that the RTL818X_RX_CONF_MULTICAST flag doesn't
have the desired effect.
Fix the issue by setting RTL818X_RX_CONF_MONITOR instead, which puts the
card into monitor mode and resolves the problem so that multicast frames
are sucessfully passed to the kernel.
The existence of the problem and the effectiveness of the solution has
originally been confirmed on an 8187B based card with the USB id of
0bda:8197. Subsequent testing by Larry Finger on an 8187L based card,
which follows the second (8187, i.e. "non-b") code path in the driver,
has confirmed that the fix does not cause any noticeable regresssions
there either.
Even if we do not set AMPDU bit in TXWI, device still can aggregate
frame and send it with rate not corresponding to requested. That mean
we can do not sent probe frames with requested rate. To prevent that
use TXOP_BACKOFF for probe frames.
If we do not match skb entry, provide tx status via nomatch procedure.
Currently in that case we do rt2x00lib_txdone_noinfo(TXDONE_NOINFO),
which actually assume that entry->skb was posted without retries and
provide rate saved in skb desc as successful. Patch changed that to
rate read from TX_STAT_FIFO, however still do not provide correct
number of retries.
On SoC/PCI devices we keep providing status via standard txdone
procedure, no change in those devices, though we should thing about it.
Gabor Juhos [Wed, 15 Feb 2017 09:25:04 +0000 (10:25 +0100)]
rt2x00: rt2800lib: move rt2800_drv_data declaration into rt2800lib.h
The rt2800_drv_data structure contains driver specific
information. Move the declaration into the rt2800lib.h
header which is a more logical place for it. Also fix
the comment style to avoid checkpatch warning.
The patch contains no functional changes, it is in
preparation for the next patch.
Rafał Miłecki [Tue, 14 Feb 2017 22:03:46 +0000 (23:03 +0100)]
bcma: gpio: set of_node regardless of the host type
DT allows describing many device types, not only platform ones. If e.g.
bcma is hosted on PCI(e) and it has its of_node, let's pass it to the
GPIO subsystem. This allows GPIO code to handle more hardware details
not only for bcma on a SoC.
Joey Zhong [Tue, 28 Feb 2017 18:51:01 +0000 (13:51 -0500)]
net: ethernet: bgmac: driver power manangement
Implement suspend/resume callbacks in the bgmac driver. This makes sure
that we de-initialize and re-initialize the hardware correctly before
entering suspend and when resuming.
Jon Mason [Tue, 28 Feb 2017 18:51:00 +0000 (13:51 -0500)]
net: ethernet: bgmac: unify code of the same family
BCM471X and BCM535X are of the same family (from what I can derive from
internal documents). Group them into the case statement together, which
results in more code reuse.
Also, use existing helper variables to make the code a little more
readable too.
Jon Mason [Tue, 28 Feb 2017 18:50:59 +0000 (13:50 -0500)]
net: ethernet: bgmac: use #defines for MAX size
The maximum frame size is really just the standard ethernet frame size
and FCS. So use those existing defines to make the code a little more
beautiful.
projects / linux.git / log