On Fri, 2016-04-15 at 02:07 +, Borislav Petkov wrote:
> Hi,
>
> so I'm seeing this when wlan0 tries to associate. On 4.6-rc2 + tip.
> After that, wifi is dead in the water. Anyone have a clue?
>
> And 4.5 seems fine, I'm typing from it as we speak.
> ...
> [ 661.142657] [
Hi,
so I'm seeing this when wlan0 tries to associate. On 4.6-rc2 + tip.
After that, wifi is dead in the water. Anyone have a clue?
And 4.5 seems fine, I'm typing from it as we speak.
...
[2.111418] iwlwifi :02:00.0: loaded firmware version 21.302800.0
op_mode iwlmvm
...
[2.220954]
From: Jes Sorensen
The 8188RU does not like PAPE to be enabled, while all the other gen1
parts seem to require it.
This makes the RTL8188RU able to associate for me.
Signed-off-by: Jes Sorensen
---
From: Jes Sorensen
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 118 ---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.h | 26 +
2 files changed, 130 insertions(+), 14 deletions(-)
From: Jes Sorensen
This implements an 8192eu specific enable_rf() function. The 8192eu is
not a combo device, so no need for doing the BT specific bits needed
by the 8723bu.
Signed-off-by: Jes Sorensen
---
From: Jes Sorensen
Load the RF table in init_phy_rf(), which allows for applying device
specific RF hacks in the same place. Getting rid of more ugly if ()
clutter.
Signed-off-by: Jes Sorensen
---
From: Jes Sorensen
In particular set APS_FSMCO_WLON_RESET in the right register, and do
not overwrite too much of REG_CR.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 16
1 file changed,
From: Jes Sorensen
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 2 +-
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.h | 2 +-
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_regs.h | 2 +-
3 files
From: Jes Sorensen
This driver relies on driver_info in struct usb_device_id, so allowing
adding a device via /sys/bus/usb/drivers/rtl8xxxu/new_id will cause a
NULL pointer dereference.
Set .no_dynamic_id = 1 to disable hot add of USB IDs.
Reported-by: Xose Vazquez
From: Jes Sorensen
This patch was being missed since rtl_chip will never match RTL8188C
if hi_pa is true.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
From: Jes Sorensen
This fixes a long standing bug where mac80211 would send disconnect
packets to the device, after we had shut down the device.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 14
From: Jes Sorensen
This dongle was tested successfully by Andrea Merello
Reported-by: Andrea Merello
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 9 +++--
1 file changed, 7
From: Jes Sorensen
All the disable_rf() functions were setting REG_TXPAUSE to 0xff to
stop transmission. Do it centrally before calling disable_rf()
instead.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 6
From: Jes Sorensen
Per the vendor driver, devices with an external PA needs limiting it's
TX power to 0x20.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 7 +++
1 file changed, 7 insertions(+)
diff
From: Colin Ian King
several functions are not initializing a return status in ret
resulting in garbage to be returned instead of 0 for success.
Currently, the calls to these functions are not checking the
return, however, it seems prudent to return the correct status
From: Jes Sorensen
The 8723bu is a combo WiFi/BT dongle, and path B is not used for WiFi,
so no point in calibrating it.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 76 +---
1 file
From: Jes Sorensen
The old code incorrectly wiped out bits 0-23 by mistake when setting
the RX path for 1T parts.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 2 +-
1 file changed, 1 insertion(+), 1
From: Jes Sorensen
Hi,
This set of patches should be what is required to get rtl8192eu
support going.
I have some additional fixes for 8188ru and other generic fixes for
the driver. I'll send those out once these make it in.
Cheers,
Jes
Jes Sorensen (26):
From: Jes Sorensen
This caught a bug where too little memory was allocated for RX urbs
for parts using 24 byte RX descriptors
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 38 ++--
From: Jes Sorensen
Presumably 8192eu devices do not have leds, so do not enable them.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git
From: Jes Sorensen
No actual code flow change, but no need to warn about something that
isn't a prioblem.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 2 --
1 file changed, 2 deletions(-)
diff --git
From: Jes Sorensen
This should get the order right and avoid patching something that is
later overwritten.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 28
1 file changed, 14
From: Jes Sorensen
Like the 8723bu, the vendor driver does not set FPGA0_TX_INFO for
8192eu in the init sequence.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 2 +-
1 file changed, 1 insertion(+), 1
From: Jes Sorensen
Rather than scattering the code with #ifdefs, use the fileops
structure to hold device specific PBP values.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 17 ++---
From: Jes Sorensen
Instead use tx_desc_size() to distinguish between gen1
(8723a/8192c/8188c) and gen2 (8723b/8192e) parts.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 7 ---
1 file changed, 4
From: Jes Sorensen
This moves the loading of the AGC table into init_phy_bb() and reduces
the if() clutter.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 26 +++-
1 file changed, 12
From: Jes Sorensen
8192eu is a 2T part, so setting TX power for path A only, as done by
rtl8723bu_set_tx_power() is not sufficient.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 78 +++-
From: Jes Sorensen
This reduce the use of magic values a little.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 4 +++-
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_regs.h | 4
2 files changed, 7
From: Jes Sorensen
Move setting it to rtl8723au_init_phy_bb()
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 22 +-
1 file changed, 9 insertions(+), 13 deletions(-)
diff --git
From: Jes Sorensen
Revert "rtl8xxxu: Temporarily disable 8192eu device init"
This reverts commit ccfe1e85322090649d2fae599e55300c1512bf15.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 4
1 file
From: Jes Sorensen
Write the adjusted value back to the correct register
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
From: Jes Sorensen
This removes another case of ugly if () clutter
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 11 +--
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.h | 1 +
2 files changed, 6
From: Jes Sorensen
Another flow order change to match the vendor driver.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 42
1 file changed, 21 insertions(+), 21 deletions(-)
diff
From: Jes Sorensen
This further reorganizes the init code flow to match that of the
8192eu vendor driver. This helps diffing the register write log
against that of the vendor driver.
Signed-off-by: Jes Sorensen
---
From: Jes Sorensen
This removes an unneeded hack for 8192eu, and allows for initializing
REG_FPGA0_XAB_RF_SW_CTRL at the same point as it is done for all other
parts.
Signed-off-by: Jes Sorensen
---
From: Jes Sorensen
This removes a bunch of if () spaghetti and re-applies the USB bus
quirks for 8188/8192 that had gotten lost.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 95 ++--
From: Jes Sorensen
This was expired by the vendor driver, but we never ended up using the
backed up value.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 3 ---
From: Jes Sorensen
Longer term we should switch all the chips over to use this function
instead of the random chip specific ifdef hacks.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 75
From: Jes Sorensen
In order to debug 8192eu support, reorder some init code to match the
flow of the vendor driver.
Signed-off-by: Jes Sorensen
---
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu.c | 35 ++--
1 file changed,
From: Johannes Berg
Date: Thu, 14 Apr 2016 09:27:17 +0200
> Since I didn't get anything else, and this has been pending for a week,
> here's the other part of the nl80211 socket problem fix (the netlink
> family URELEASE was the first part.)
>
> Let me know if there's
On 04/14/2016 07:48 AM, David Müller wrote:
The for-loop condition does not work correctly on architectures where "char"
is unsigned. Fix it by using an "int", which may also result in more
efficient code.
Signed-off-by: David Müller
---
The semaphore was being released twice, once at the beginning of the
thread and then again when the thread is about to close.
The semaphore is acquired immediately after creating the thread so we
should be releasing it when the thread ends.
Signed-off-by: Sudip Mukherjee
These variables were havnig the pointer returned by wiphy_priv() but
they were never used.
Signed-off-by: Sudip Mukherjee
---
drivers/staging/wilc1000/linux_wlan.c | 8
1 file changed, 8 deletions(-)
diff --git a/drivers/staging/wilc1000/linux_wlan.c
Dan Carpenter writes:
> Smatch complains that since "ev->peer_id" comes from skb->data that
> means we can't trust it and have to do a bounds check on it to prevent
> an array overflow.
>
> Fixes: 6942726f7f7b ('ath10k: add fast peer_map lookup')
> Signed-off-by: Dan
Kalle Valo writes:
> This has ben deprecated years ago, I haven't heard anyone using it since and
> most likely it won't even work anymore. So just remove all of it.
>
> Signed-off-by: Kalle Valo
[...]
> ---
> During system suspend, there is a kernel WARNING issued if there
> is a pending command present. By marking the wait queue disabled
> after calling the command completion routine fixes it.
>
> Signed-off-by: Vishal Thanki
Thanks, applied to wireless-drivers-next.git.
> This patch adds missing break statement at the end of
> PCIE_DEVICE_ID_MARVELL_88W8897 switch section.
>
> Signed-off-by: Amitkumar Karwar
Thanks, 5 patches applied to wireless-drivers-next.git:
c865a70098d9 mwifiex: missing break statement
e87650bce9a5 mwifiex: add
> From: Hante Meuleman
>
> When the event_msgs iovar is set an array is used to configure the
> enabled events. This arrays needs to nulled before configuring
> otherwise unhandled events will be enabled. This solves a problem
> where in case of wowl the host got
> From: Jes Sorensen
>
> The 8192eu requires a different MAC init table. Add the missing table
> and specify the table to use in the fileops structure.
>
> Signed-off-by: Jes Sorensen
Thanks, 18 patches applied to wireless-drivers-next.git:
The for-loop condition does not work correctly on architectures where "char"
is unsigned. Fix it by using an "int", which may also result in more
efficient code.
Signed-off-by: David Müller
---
drivers/net/wireless/realtek/rtlwifi/rtl8821ae/phy.c | 2 +-
1 file changed, 1
Hi,
This patchset disables qdiscs for drivers
using software queuing and performs fq_codel-like
dequeuing on txqs.
I've reworked it as per Avery's suggestion (and
more).
I've (re)tested it against ath10k with and without
DQL (my ath10k RFC which is limited but sufficient
for some proofing
Qdiscs are designed with no regard to 802.11
aggregation requirements and hand out
packet-by-packet with no guarantee they are
destined to the same tid. This does more bad than
good no matter how fairly a given qdisc may behave
on an ethernet interface.
Software queuing used per-AC netdev
There is no other limit other than a global
packet count limit when using software queuing.
This means a single flow queue can grow insanely
long. This is particularly bad for TCP congestion
algorithms which requires a little more
sophisticated frame dropping scheme than a mere
headdrop on limit
This adds a few debugfs entries to make it easier
to test, debug and experiment.
Signed-off-by: Michal Kazior
---
net/mac80211/debugfs.c | 14 ++
net/mac80211/ieee80211_i.h | 2 ++
net/mac80211/tx.c | 21 ++---
3 files changed,
mac80211's software queues were designed to work
very closely with device tx queues. They are
required to make use of 802.11 packet aggregation
easily and efficiently.
Due to the way 802.11 aggregation is designed it
only makes sense to keep fair queuing as close to
hardware as possible to reduce
This adds a few debugfs entries and a module
parameter to make it easier to test, debug and
experiment.
Signed-off-by: Michal Kazior
---
net/mac80211/debugfs.c| 77 +++
net/mac80211/debugfs_netdev.c | 28 +++-
The commit fb724ed5c617 ("wlcore: Fix regression in
wlcore_set_partition()") fixed wl12xx functionality.
However, it reverted the support in fw logger
over sdio in wl18xx.
This patch reverts the changes made and also fixes
the original functionality issue introduced in wl12xx.
Fixes:
On Mon, 2016-04-11 at 15:16 +0530, Kanchanapally, Vidyullatha wrote:
> From: "Kanchanapally, Vidyullatha"
>
> Since cfg80211 maintains separate BSS table entries for APs if the
> same
> BSSID, SSID pair is seen on multiple channels, it is possible that it
> can map the
On Thu, 2016-04-07 at 19:59 +0530, Mohammed Shafi Shajakhan wrote:
> From: Mohammed Shafi Shajakhan
>
> Add support for new netlink attribute 'NL80211_STA_INFO_RX_DURATION'.
> 'NL80211_STA_INFO_RX_DURATION' accounts for aggregate PPDU duration
> in microseconds for all
Hi,
I know this is a LONG time ago, but nonetheless...
> -static struct ieee80211_iface_combination hwsim_if_comb = {
> - .limits = hwsim_if_limits,
> - .n_limits = ARRAY_SIZE(hwsim_if_limits),
> - .max_interfaces = 2048,
> - .num_different_channels = 1,
> +static const struct
On Wed, 2016-04-13 at 14:13 +0300, Kalle Valo wrote:
> Fix checkpatch warnings about use of spaces with operators:
> spaces preferred around that '*' (ctx:VxV)
> This has been recently added to checkpatch.
Trivia: That's not a particularly recent change.
> diff --git
Hi Dave,
Since I didn't get anything else, and this has been pending for a week,
here's the other part of the nl80211 socket problem fix (the netlink
family URELEASE was the first part.)
Let me know if there's any problem.
Thanks,
johannes
The following changes since commit
62 matches
Mail list logo