wfx_bss_info_changed() is not called from atomic contexts. So, it is not
necessary to raise an asynchronous work to change ERP. Thus
wfx_set_cts_work() become useless.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-27-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
setbssparams_done was here to ensure that the firmware does not enable
powersave before to get the first beacon. However, mac80211 already
ensures it gets a beacon before to associate to the BSS. And even, if
it won't, the firmware wake up at least on every DTIM, which is
sufficient to finalize the association.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-26-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current code keeps a copy of bss_conf->cqm_rssi_thold in wfx_vif. There
is no sane reason for that.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-25-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It seems that current code try to save calls to hif_beacon_transmit() by
keeping a copy of the previous value of bss_conf->enable_beacon.
However, hif_beacon_transmit() does not cost so much and mac80211
already take care to not send useless events.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-24-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It is not necessary to keep a copy of dtim_period in wfx_vif. Prefer to
just rely on bss_conf->dtim_period.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-23-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current code parse the TIM and retrieve the DTIM period. It is far
easier to rely on bss_info_changed() for this job.
It is no more necessary to run task asynchronously. So
set_beacon_wakeup_period_work is now useless.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-22-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current code of wfx_flush() force to drop packets in some contexts.
However, there is no obvious reasons to do that. It looks like a
workaround for a bug with the old implementation of __wfx_flush().
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-21-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
wfx_tx_queues_clear() only clear not yet sent requests. So, it always
necessary to wait for tx_queue_stats.wait_link_id_empty whatever the
value of "drop" argument.
In add, it is not necessary to return with tx queue locked since all
calls to __wfx_flush() unlock the tx queue just after the call to
__wfx_tx_flush().
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-20-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This function built probe response from data retrieved in beacon. Yet,
this job can be done with ieee80211_proberesp_get(). So, we can simplify
that code (and fix bugs like inclusion of TIM in probe responses).
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-18-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In fact, wfx_upload_beacon() uploads beacon and probe response. So,
rename it in wfx_upload_ap_templates().
The call to wfx_fwd_probe_req() has nothing to do with template
uploading, so relocate it.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-17-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mac80211 and the device are both able to control tx power per vif. But,
the current code retrieve tx power from wfx_config(). So, it does not
allow to setup the tx power independently for each vif. Driver just has
to rely on bss_conf->txpower to get the correct information.
In add, it is no more necessary to protect access to wdev->output_power
with scan_lock.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-15-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
wvif->ht_info contains useless copies of sta->ht_cap and
bss_conf->ht_operation_mode. Prefer to retrieve information from the
original structs instead of rely on wvif->ht_info.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-13-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Most of the commands that are sent to device should take struct in
argument. In the current code, when this struct is binary compatible
with a __le32, the driver use a __le32. This behavior is error prone.
This patch fixes that and uses the specialized structs instead.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-11-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The structure hif_req_start come from hardware API. It is not intended
to be manipulated in upper layers of the driver.
In add, current code for hif_start() is too dumb. It should pack data
with hardware representation instead of leaving all work to the caller.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-10-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The structure hif_mib_arp_ip_addr_table come from hardware API. It is
not intended to be manipulated in upper layers of the driver.
In add, current code for hif_set_arp_ipv4_filter() is too dumb. It
should pack data using the hardware representation instead of leaving
all work to the caller.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-9-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The structure hif_mib_rcpi_rssi_threshold come from hardware API. It is
not intended to be manipulated in upper layers of the driver.
In add, current code for hif_set_rcpi_rssi_threshold() is dumb. It
should pack data using the hardware representation instead of leaving
all work to the caller.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-8-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Hardware API use 10th of dBm for output power unit. Upper layers should
use same units than mac80211 and the conversion should be done by low
level layer of the driver (hif_set_output_power())
In add, current code of hif_set_output_power() use a __le32 while the
device API specify a specific structure for this.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-7-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The structure hif_mib_set_tx_rate_retry_policy come from hardware
API. It is not intended to be manipulated in upper layers of the driver.
So, this patch relocate handling of this structure to
hif_set_tx_rate_retry_policy() (the low level function).
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-6-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Rate policies (aka. tx_rate_retry_policy in hardware API) are sent to
device asynchronously from tx requests. So, the device maintains a list
of active rate policies and the tx requests only reference an existent
rate policy.
The device API allows to send multiple rate policies at once. However,
this property is very rarely used. We prefer to send rate policies one
by one and simplify the architecture.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-5-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
hif_scan() return max number of jiffies to wait before the completion
indication. However, if this value is negative, an error has occurred.
Reword the code to reflect that behavior.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-3-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It appears that commit 8c7128c4cf ("staging: align to fix warnings of
line over 80 characters") do slightly more than what is explained in
commit log.
Especially, it changes the output of the file rx_stats from debugfs.
From some point of view, this file can be considered as a part of the
API. Any change on it should be clearly announced.
Since the change introduced does not seems to have any justification,
revert it.
Reported-by: Pascal Prime <pascal.prime@silabs.com>
Cc: Jules Irenge <jbi.octave@gmail.com>
Fixes: 8c7128c4cf ("staging: align to fix warnings of line over 80 characters")
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200115135338.14374-2-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is no need to add "ccflags-y += -I $(srctree)/drivers/staging"
just for including <most/most.h>.
Use the #include "..." directive with the correct relative path.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Link: https://lore.kernel.org/r/20200115164451.13203-1-masahiroy@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Checkpatch reports 'WARNING: printk() should include KERN_<LEVEL>
facility level'. Fix this by specifying a relevant KERN_<LEVEL> value
for each line in which it was missing.
Once they are fixed, checkpatch reports 'WARNING: Prefer [subsystem eg:
netdev]_dbg([subsystem]dev, ... then dev_dbg(dev, ... then
pr_debug(... to printk(KERN_DEBUG ...'. Fix this by replacing
relevant printk_<level> statements with their netdev_<level>
equivalent.
Signed-off-by: Paulo Miguel Almeida <paulo.miguel.almeida.rodenas@gmail.com>
Link: https://lore.kernel.org/r/20200115101208.GA683742@localhost.localdomain
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The variable ret is being initialized with a value that
is never read and is being re-assigned later on. The
assignment is redundant and hence can be removed.
Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20200114171145.375356-1-colin.king@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The variable result is being initialized with a value that
is never read and is being re-assigned later on. The assignment
is redundant and hence can be removed.
Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20200114172720.376286-1-colin.king@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently when the call to prism2sta_ifst fails a netdev_err error
is reported, error return variable result is set to -1 but the
function always returns 0 for success. Fix this by returning
the error value in variable result rather than 0.
Addresses-Coverity: ("Unused value")
Fixes: 00b3ed1685 ("Staging: add wlan-ng prism2 usb driver")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Cc: stable <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20200114181604.390235-1-colin.king@canonical.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replaced strcpy call with safer strscpy call with given length.
This elimates the need to manually null-terminate the given string,
since strscpy will null terminate the destination anyway.:
Signed-off-by: Ole Wiedemann <ole.wiedemann@fau.de>
Co-developed-by: Sebastian Scherbel <sebastian.scherbel@fau.de>
Signed-off-by: Sebastian Scherbel <sebastian.scherbel@fau.de>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Acked-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Acked-by: Todd Kjos <tkjos@google.com>
Link: https://lore.kernel.org/r/20191213131032.22579-1-ole.wiedemann@fau.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Function rtw_hal_antdiv_before_linked() returns boolean values, so
change the return type from u8 to bool.
Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200114134422.13598-2-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Refactor rtw_hal_antdiv_before_linked() to clear checkpatch warnings.
WARNING: line over 80 characters
WARNING: else is not generally useful after a break or return
Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200114134422.13598-1-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Function ODM_GetRightChnlPlaceforIQK() returns non-zero values only
for channels > 14. According to the TODO code valid only for 5 GHz
should be removed.
- find and remove remaining code valid only for 5 GHz. Most of the
obvious ones have been removed, but things like channel > 14 still
exist.
Remove ODM_GetRightChnlPlaceforIQK() and replace the uses of the
return value with zero.
Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200110210456.13178-1-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Two fixes for RISC-V:
- Clear FP registers during boot when FP support is present, rather than
when they aren't present
- Move the header files associated with the SiFive L2 cache controller
to drivers/soc (where the code was recently moved)
-----BEGIN PGP SIGNATURE-----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=qvPe
-----END PGP SIGNATURE-----
Merge tag 'riscv/for-v5.5-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux
Pull RISC-V fixes from Paul Walmsley:
"Two fixes for RISC-V:
- Clear FP registers during boot when FP support is present, rather
than when they aren't present
- Move the header files associated with the SiFive L2 cache
controller to drivers/soc (where the code was recently moved)"
* tag 'riscv/for-v5.5-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
riscv: Fixup obvious bug for fp-regs reset
riscv: move sifive_l2_cache.h to include/soc
CSR_MISA is defined in Privileged Architectures' spec: 3.1.1 Machine
ISA Register misa. Every bit:1 indicate a feature, so we should beqz
reset_done when there is no F/D bit in csr_misa register.
Signed-off-by: Guo Ren <ren_guo@c-sky.com>
[paul.walmsley@sifive.com: fix typo in commit message]
Fixes: 9e80635619 ("riscv: clear the instruction cache and all registers when booting")
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
The commit 9209fb5189 ("riscv: move sifive_l2_cache.c to drivers/soc")
moves the sifive L2 cache driver to driver/soc. It did not move the
header file along with the driver. Therefore this patch moves the header
file to driver/soc
Signed-off-by: Yash Shah <yash.shah@sifive.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
[paul.walmsley@sifive.com: updated to fix the include guard]
Fixes: 9209fb5189 ("riscv: move sifive_l2_cache.c to drivers/soc")
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Including:
- Two fixes for VT-d and generic IOMMU code to fix teardown on
error handling code paths.
- Patch for the Intel VT-d driver to fix handling of non-PCI
devices
- Fix W=1 compile warning in dma-iommu code
-----BEGIN PGP SIGNATURE-----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=rTxu
-----END PGP SIGNATURE-----
Merge tag 'iommu-fixes-v5.5-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu
Pull iommu fixes from Joerg Roedel:
- Two fixes for VT-d and generic IOMMU code to fix teardown on error
handling code paths.
- Patch for the Intel VT-d driver to fix handling of non-PCI devices
- Fix W=1 compile warning in dma-iommu code
* tag 'iommu-fixes-v5.5-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
iommu/dma: fix variable 'cookie' set but not used
iommu/vt-d: Unlink device if failed to add to group
iommu: Remove device link to group on failure
iommu/vt-d: Fix adding non-PCI devices to Intel IOMMU
Pull i2c fixes from Wolfram Sang:
"Two driver bugfixes, a documentation fix, and a removal of a spec
violation for the bus recovery algorithm in the core"
* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
i2c: fix bus recovery stop mode timing
i2c: bcm2835: Store pointer to bus clock
dt-bindings: i2c: at91: fix i2c-sda-hold-time-ns documentation for sam9x60
i2c: at91: fix clk_offset for sam9x60
