Andreas Kemnade [Sat, 30 Jan 2016 17:01:54 +0000 (18:01 +0100)]
libertas: go back to ps mode without commands pending
Removes the old todo block and checks only whether ieee powersave
mode is requested. We still have to check for being connected as
this powersave mode includes logic for regularly waking up and
checking for packets which only makes sense when connected.
For not being connected, another mode is needed.
Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Andreas Kemnade [Sat, 30 Jan 2016 17:01:53 +0000 (18:01 +0100)]
libertas: do not confirm sleep if commands are pending
If the main thread gets one PS AWAKE event and one PS SLEEP event
in one iteration over event_fifo there will never be checks for
commands to be processed, since psstate will always be
PS_STATE_SLEEP or PS_STATE_PRE_SLEEP
Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Andreas Kemnade [Sat, 30 Jan 2016 17:01:52 +0000 (18:01 +0100)]
libertas: check whether bus can do more than polling
If a sdio host does not support sdio irqs, polling is used
instead. That has an impact on performance. Some functionality
should not be enabled then. This add a variable in
libertas_priv to indicate that.
Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Andreas Kemnade [Sat, 30 Jan 2016 17:01:51 +0000 (18:01 +0100)]
libertas: fix pointer bugs for PS_MODE commands
struct cmd_ds_802_11_ps_mode
contains the command header and a pointer to it was
initialized with data points to the body which leads to
mis-interpretation of the cmd_ds_802_11_ps_mode.action member.
cmd[0] contains the header, &cmd[1] points beyond that.
cmdnode->cmdbuf is a pointer to the command buffer
This piece of code was unused since power saving was
not enabled.
Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Dan Carpenter [Tue, 2 Feb 2016 10:00:05 +0000 (13:00 +0300)]
mwifiex: fix a reversed condition
The NULL test here is reversed.
Fixes:
7d7f07d8c5d3 ('mwifiex: add wowlan net-detect support')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Nachiket Kukade [Mon, 1 Feb 2016 07:44:46 +0000 (23:44 -0800)]
mwifiex: fix bandwidth display problem
Instead of using HT info from beacon IEs, use HT info from
association response frame to update bandwidth in
cfg80211_get_channel handler.
Signed-off-by: Nachiket Kukade <kukaden@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Arnd Bergmann [Thu, 28 Jan 2016 21:58:28 +0000 (22:58 +0100)]
hostap: avoid uninitialized variable use in hfa384x_get_rid
The driver reads a value from hfa384x_from_bap(), which may fail,
and then assigns the value to a local variable. gcc detects that
in in the failure case, the 'rlen' variable now contains
uninitialized data:
In file included from ../drivers/net/wireless/intersil/hostap/hostap_pci.c:220:0:
drivers/net/wireless/intersil/hostap/hostap_hw.c: In function 'hfa384x_get_rid':
drivers/net/wireless/intersil/hostap/hostap_hw.c:842:5: warning: 'rec' may be used uninitialized in this function [-Wmaybe-uninitialized]
if (le16_to_cpu(rec.len) == 0) {
This restructures the function as suggested by Russell King, to
make it more readable and get more reliable error handling, by
handling each failure mode using a goto.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Sjoerd Simons [Mon, 25 Jan 2016 10:47:29 +0000 (11:47 +0100)]
brcmfmac: sdio: Increase the default timeouts a bit
On a Radxa Rock2 board with a Ampak AP6335 (Broadcom 4339 core) it seems
the card responds very quickly most of the time, unfortunately during
initialisation it sometimes seems to take just a bit over 2 seconds to
respond.
This results intialization failing with message like:
brcmf_c_preinit_dcmds: Retreiving cur_etheraddr failed, -52
brcmf_bus_start: failed: -52
brcmf_sdio_firmware_callback: dongle is not responding
Increasing the timeout to allow for a bit more headroom allows the
card to initialize reliably.
A quick search online after diagnosing/fixing this showed that Google
has a similar patch in their ChromeOS tree, so this doesn't seem
specific to the board I'm using.
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Reviewed-by: Julian Calaby <julian.calaby@gmail.com>
Acked-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Sun, 31 Jan 2016 11:14:34 +0000 (12:14 +0100)]
brcmfmac: treat NULL character in NVRAM as separator
Platform NVRAM (stored on a flash partition) has entries separated by a
NULL (\0) char. Our parsing code switches from VALUE state to IDLE
whenever it meets a NULL (\0). When that happens our IDLE handler should
simply consume it and analyze whatever is placed ahead.
This fixes harmless warnings spamming debugging output:
[ 155.165624] brcmfmac: brcmf_nvram_handle_idle warning: ln=1:col=20: ignoring invalid character
[ 155.180806] brcmfmac: brcmf_nvram_handle_idle warning: ln=1:col=44: ignoring invalid character
[ 155.195971] brcmfmac: brcmf_nvram_handle_idle warning: ln=1:col=63: ignoring invalid character
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 26 Jan 2016 16:57:05 +0000 (17:57 +0100)]
brcmfmac: add support for 14e4:4365 PCI ID with BCM4366 chipset
On Broadcom ARM routers BCM4366 cards are available with 14e4:4365 ID.
Unfortunately this ID was already used by Broadcom for cards with
BCM43142, a totally different chipset requiring SoftMAC driver. To avoid
a conflict between brcmfmac and bcma use more specific ID entry with
subvendor and subdevice specified.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 26 Jan 2016 16:57:04 +0000 (17:57 +0100)]
brcmfmac: access PMU registers using standalone PMU core if available
On recent Broadcom chipsets PMU is present as separated core and it
can't be accessed using ChipCommon anymore as it fails with e.g.:
[ 18.198412] Unhandled fault: imprecise external abort (0x1406) at 0xb6da200f
Add a new helper function that will return a proper core that should be
used for accessing PMU registers.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 26 Jan 2016 16:57:03 +0000 (17:57 +0100)]
brcmfmac: read extended capabilities of ChipCommon core
This is an extra bitfield with info about some present hardware.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 26 Jan 2016 16:57:02 +0000 (17:57 +0100)]
brcmfmac: allow storing PMU core without wrapper address
Separated PMU core can be found in new devices and should be used for
accessing PMU registers (which were routed through ChipCommon so far).
This core is one of exceptions that doesn't have or need wrapper address
to be still safely accessible.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 26 Jan 2016 16:57:01 +0000 (17:57 +0100)]
brcmfmac: analyze descriptors of current component only
So far we were looking for address descriptors without a check for
crossing current component border. In case of dealing with unsupported
descriptor or descriptor missing at all the code would incorrectly get
data from another component.
Consider this binary-described component from BCM4366 EROM:
4bf83b01 TAG==CI CID==0x83b
20080201 TAG==CI PORTS==0+1 WRAPPERS==0+1
18400035 TAG==ADDR SZ_SZD TYPE_SLAVE
00050000
18107085 TAG==ADDR SZ_4K TYPE_SWRAP
Driver was assigning invalid base address to this core:
brcmfmac: [6 ] core 0x83b:32 base 0x18109000 wrap 0x18107000
which came from totally different component defined in EROM:
43b36701 TAG==CI CID==0x367
00000201 TAG==CI PORTS==0+1 WRAPPERS==0+0
18109005 TAG==ADDR SZ_4K TYPE_SLAVE
This change will also allow us to support components without wrapper
address in the future.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 26 Jan 2016 16:02:09 +0000 (17:02 +0100)]
bcma: claim only 14e4:4365 PCI Dell card with SoftMAC BCM43142
It seems 14e4:4365 pattern is too generic as there are two devices:
1) 14e4:4365 1028:0016 with SoftMAC BCM43142 chipset
2) 14e4:4365 14e4:4365 with FullMAC BCM4366 chipset
The later one was found in D-Link DIR-885L router and we want to let
brcmfmac handle it.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Sun, 24 Jan 2016 15:37:33 +0000 (16:37 +0100)]
bcma: add support for BCM47094
It's another SoC with 32 GPIOs and simplified watchdog handling. It was
tested on D-Link DIR-885L.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Fri, 22 Jan 2016 17:02:54 +0000 (18:02 +0100)]
bcma: init serial console directly from ChipCommon code
UART is connected to and controlled over ChipCommon core. It doesn't
have much to do with MIPS core (where we initialize it currently)
except just existing on embedded systemms. There isn't point of such
cross-core initialization (and we needed #ifdef anyway) so just handle
it in ChipCommon.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Wed, 20 Jan 2016 15:46:04 +0000 (16:46 +0100)]
brcmfmac: fix setting primary channel for 80 MHz width
First of all it changes the way we calculate primary channel offset. If
we use e.g. 80 MHz channel with primary frequency 5180 MHz (which means
center frequency is 5210 MHz) it makes sense to calculate primary offset
as -30 MHz.
Then it fixes values we compare primary_offset with. We were comparing
offset in MHz against -2 or 2 which was resulting in picking a wrong
primary channel.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 19 Jan 2016 07:45:26 +0000 (08:45 +0100)]
bcma: support PMU present as separated bus core
On recent Broadcom chipsets PMU is present as separated core and it
can't be accessed using ChipCommon anymore as it fails with e.g.:
[ 0.000577] Unhandled fault: external abort on non-linefetch (0x1008) at 0xf1000604
Solve it by using a new (PMU) core pointer set to ChipCommon or PMU
depending on the hardware capabilities.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Tue, 19 Jan 2016 07:45:25 +0000 (08:45 +0100)]
bcma: use _PMU_ in all names of PMU registers
PMU (Power Management Unit) seems to be a separated piece of hardware,
just accessed using ChipCommon core registers. In recent Broadcom
chipsets PMU is not bounded to CC but available as separated core.
To make code cleaner & easier to review (for a correct R/W access) use
clearer names.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Fri, 15 Jan 2016 23:48:53 +0000 (00:48 +0100)]
bcma: support chipsets with PMU and GCI cores (devices)
Both cores are another exceptions. They are not accessed in a standard
way and to they don't need or have wrapping addresses.
This fixes bus scanning after finding such core.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Fri, 15 Jan 2016 23:48:52 +0000 (00:48 +0100)]
bcma: identify bus cores (devices) found on BCM47189
Add missing defines and print proper names.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Rafał Miłecki [Fri, 15 Jan 2016 23:48:32 +0000 (00:48 +0100)]
bcma: support identifying MX25L25635F serial flash
It's a Macronix 32 MiB flash found on board with BCM47189 SoC.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Kalle Valo [Sat, 6 Feb 2016 11:01:25 +0000 (13:01 +0200)]
Merge tag 'iwlwifi-next-for-kalle-2016-01-31_2' of https://git./linux/kernel/git/iwlwifi/iwlwifi-next
Here's a first batch of patches for 4.6:
* continue the work on multiple Rx queues (Sara)
* add support for beacon storing
used in low power states (Sara)
* cleanups (Rodrigo, Johannes)
* fix the LED behavior for iwldvm (Hubert)
* Use the regular firmware image of WoWLAN (Matti)
* fix 8000 devices for Big Endian machines (Johannes)
* more firmware debug hooks (Golan)
* add support for P2P Client snoozing (Avri)
* make the beacon filtering for AP mode configurable (Andrei)
* fix transmit queues overflow with LSO
Andrei Otcheretianski [Tue, 26 Jan 2016 16:12:28 +0000 (18:12 +0200)]
iwlwifi: mvm: allow to disable beacon filtering for AP/GO interface
When in AP mode we need to filter in beacons from other APs to update HT
operation mode. As a power optimization the beacons are filtered out when
there are no associated stations. As a result, when there are no
associated stations, we will not update the HT operation mode until a
station connects.
Add a debugfs parameter that allows to disable this optimization.
Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Tue, 26 Jan 2016 10:35:13 +0000 (12:35 +0200)]
iwlwifi: pcie: update iwl_mpdu_desc fields
Final API of iwl_mpdu_desc has a change in the order of
the fields and does not include energy from the third
antenna (which is perfectly fine, since we don't have one).
Update the structure accordingly.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Mon, 25 Jan 2016 16:14:49 +0000 (18:14 +0200)]
iwlwifi: pcie: enable multi-queue rx path
Previous patches enabled new 9000 hardware DMA for one queue
only.
Enable the actual multi-queue path and configuration now.
This requires also per-queue NAPI struct.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Thu, 31 Dec 2015 09:49:18 +0000 (11:49 +0200)]
iwlwifi: mvm: support rss queues configuration command
9000 series supports multi-queue rx. The hardware needs
to be configured with the hash functions to perform and
indirection table that maps hash results to the relevant
CPUs\queues.
Support this configuration.
Add debugfs hook to configure the indirection table in
order to enable performance analysis. The configuration
is stateless, receives a partial or full pattern and sends
the command to the firmware.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Gregory Greenman [Tue, 26 Jan 2016 22:10:15 +0000 (00:10 +0200)]
iwlwifi: mvm: rs: fix TPC action decision algorithm
Decreasing Tx power is allowed only when success ratio is
above the threshold defined in the algorithm. Add this condition.
Signed-off-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Tue, 26 Jan 2016 11:17:47 +0000 (13:17 +0200)]
iwlwifi: mvm: add new ADD_STA command version
The 9000 hardware introduces the frame releaser, which
keeps track of the aggregation window and notifies host
of the window status. This requires in turn updating
the hardware with the RX BA session window size.
Firmware API was changed to enable that, update the driver
accordingly.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Avri Altman [Sun, 24 Jan 2016 13:56:11 +0000 (15:56 +0200)]
iwlwifi: mvm: Remove iwl_mvm_update_beacon_abort
It is only called from iwl_mvm_power_set_ba() so simplify things
by removing it.
Signed-off-by: Avri Altman <avri.altman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Avri Altman [Sun, 24 Jan 2016 13:18:59 +0000 (15:18 +0200)]
iwlwifi: mvm: Remove bf_vif from iwl_power_vifs
This member is actually not needed as beacon abort
is only allowed for a bss station.
Signed-off-by: Avri Altman <avri.altman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Johannes Berg [Sun, 24 Jan 2016 14:28:43 +0000 (15:28 +0100)]
iwlwifi: treat iwl_parse_nvm_data() MAC addr as little endian
The MAC address parameters passed to iwl_parse_nvm_data() are passed on
to iwl_set_hw_address_family_8000() which treats them as little endian.
Annotate them as such, and add the missing byte-swapping in mvm.
While at it, add the MAC address to the error to make debugging issues
with it easier.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Wed, 30 Dec 2015 21:58:29 +0000 (23:58 +0200)]
iwlwifi: mvm: add tlv for multi queue rx support
Previous patches enabled the multi-queue rx path based on
iwl_mvm_has_new_rx_api() which returned false by default.
Change it to return the actual value based on the firmware
TLV which is now defined.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Thu, 7 Jan 2016 14:50:45 +0000 (16:50 +0200)]
iwlwifi: mvm: change the check for ADD_STA status
The firmware will return the baid for BA session in the
ADD_STA command response.
This requires masking the check of the status, which is
actually only 8 bits, and not the whole 32 bits.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Max Stepanov [Mon, 4 Jan 2016 09:09:14 +0000 (11:09 +0200)]
iwlwifi: mvm: add debug print if scan config is ignored
Print a debug message in iwl_mvm_config_scan() if a scan configuration
data is decided not to be sent to FW.
Signed-off-by: Max Stepanov <Max.Stepanov@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Wed, 16 Dec 2015 19:17:06 +0000 (21:17 +0200)]
iwlwifi: mvm: change access to ieee80211_hdr
Make the code clearer.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Johannes Berg [Wed, 13 Jan 2016 14:01:00 +0000 (15:01 +0100)]
iwlwifi: mvm: support setting minimum quota from debugfs
For debug purposes, allow setting minimum quota (for a single
virtual interface) from debugfs. This is an absolute minimum,
so it can only be set up to 95%.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Johannes Berg [Wed, 6 Jan 2016 09:01:41 +0000 (10:01 +0100)]
iwlwifi: mvm: track low-latency sources separately
To be able to test low-latency behaviour properly, split the
different low-latency sources so that setting any one of them,
for example from debugfs, is sufficient; this avoids getting
the debug setting overwritten by other sources.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Tue, 29 Dec 2015 09:07:15 +0000 (11:07 +0200)]
iwlwifi: mvm: support beacon storing
Currently firmware is configured to filter out beacons. In case
a beacon was changed - it is waking the host.
However, some vendors change their IEs frequently without any
significant change, and redundant wakeups are triggered as a
result.
As a solution disable beacon filtering when entering d0i3.
Instead, firmware will store the latest beacon and upon exiting
d0i3 it will send it up to the host, so the host can act upon
changes (if there were any).
This beacon will arrive as a dedicated notification - support it
as well.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Chaya Rachel Ivgi [Sun, 27 Dec 2015 11:45:42 +0000 (13:45 +0200)]
iwlwifi: mvm: add support for negative temperatures
The driver should support also negative temperatures.
So there is a need to separate between the return value and
temperature in order to be able to distinguish between
a negative temperature and error value.
Signed-off-by: Chaya Rachel Ivgi <chaya.rachel.ivgi@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Johannes Berg [Thu, 14 Jan 2016 09:55:22 +0000 (10:55 +0100)]
iwlwifi: mvm: fix debugfs signedness warning
Using kstrtouint() with a signed int isn't really right,
use kstrotoint() instead.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Johannes Berg [Thu, 14 Jan 2016 09:46:38 +0000 (10:46 +0100)]
iwlwifi: mvm: remove shadowing variable
The outer scope has a perfectly suitable 'i' variable,
use it instead of adding a shadowing one in the inner
scope.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Golan Ben-Ami [Wed, 23 Dec 2015 15:53:27 +0000 (17:53 +0200)]
iwlwifi: mvm: make collecting fw debug data optional
Slow platforms may have issues with dumping data upon
firmware assert. Make it easier to disable it for those
platform.
Signed-off-by: Golan Ben-Ami <golan.ben.ami@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Avri Altman [Wed, 25 Nov 2015 11:17:10 +0000 (13:17 +0200)]
iwlwifi: mvm: Add P2P client snoozing
Enable snoozing and U-APSD on P2P client. The firwmare will
support this only if the BSS vif is not associated.
Make this configurable by a constant variable and disable
it by default.
Signed-off-by: Avri Altman <avri.altman@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Golan Ben-Ami [Wed, 2 Sep 2015 09:34:23 +0000 (12:34 +0300)]
iwlwifi: mvm: add trigger for firmware dump upon TX response status
This will allow to collect the data when the firmware
sends a specific tx response status.
Signed-off-by: Golan Ben-Ami <golan.ben.ami@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Rodrigo Freire [Wed, 6 Jan 2016 22:24:01 +0000 (20:24 -0200)]
iwlwifi: Document missing module options
This patch documents two missing module options in the internal
code comment block.
Signed-off-by: Rodrigo Freire <rfreire@redhat.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Hubert Tarasiuk [Sun, 24 Jan 2016 12:35:06 +0000 (13:35 +0100)]
iwlwifi: dvm: handle zero brightness for wifi LED
In order to have the LED being OFF constantly when the
brightness is set to 0, we need to pass IWL_LED_SOLID to
iwl_led_cmd as the off parameter, otherwise the led will
stay on constantly.
This fixes
https://bugzilla.kernel.org/show_bug.cgi?id=110551
Signed-off-by: Hubert Tarasiuk <hubert.tarasiuk@gmail.com>
[reworked the commit message]
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Emmanuel Grumbach [Sun, 24 Jan 2016 21:04:22 +0000 (23:04 +0200)]
iwlwifi: various comments and code cleanups
No need to include net/ip6_checksum.h twice.
Remove TODOs.
Remove trailing space.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Matti Gottlieb [Thu, 31 Dec 2015 16:18:02 +0000 (18:18 +0200)]
iwlwifi: mvm: Do not switch to D3 image on suspend
Currently when the driver is configured with wowlan parameters, and enters
D3 mode, the driver switches the FW image to D3, and when it exists
suspend, it reloads the D0 image.
If the firmware supports the consolidation of the D0 & D3 images there is
no need to load the D3 image on suspend, and no need to reload the D0
image on resume.
Do not switch images on suspend / resume, for firmwares that support
consolidated images.
Signed-off-by: Matti Gottlieb <matti.gottlieb@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Luciano Coelho [Tue, 18 Aug 2015 13:02:38 +0000 (16:02 +0300)]
iwlwifi: pcie: add RTPM support when wifi is enabled
Enable runtime power management (RTPM) for PCIe devices and implement
the corresponding functions to enable D0i3 mode when the device is
idle.
Additionally, remove some unnecessary #ifdef's because the RTPM code
will not be called if runtime PM is not configured.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Luca Coelho [Wed, 6 Jan 2016 20:40:38 +0000 (18:40 -0200)]
iwlwifi: pcie: add initial RTPM support for PCI
Add an initial implementation of runtime power management (RTPM) for
PCI devices. With this patch, RTPM is only used when wifi is off
(i.e. the wifi interface is down). This implementation is behind a
new Kconfig flag, IWLWIFI_PCIE_RTPM.
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Wed, 23 Dec 2015 13:10:03 +0000 (15:10 +0200)]
iwlwifi: pcie: add 9000 series multi queue rx DMA support
The 9000 series introduces several changes in the device
DMA operation.
As the device now supports multi-queue rx, several DMA channels
should be configured.
The flows of providing the device with the allocated RBDs now
changes as well - the device maintains a separate table of used
and free table.
The hardware may use the free table to feed RBDs to any queue.
This requires maintaing a shared table to map returned RBDs to
the original RXB - for that purpose the VID is introduced - an
internal identifier of the RB placed in the lower 12 bits and
returned by HW in the used data.
Another change is the support of 64 bit DMA address.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Sara Sharon [Mon, 14 Dec 2015 15:44:11 +0000 (17:44 +0200)]
iwlwifi: pcie: add infrastructure for multi-queue rx
The 9000 series devices will support multi rx queues.
Current code has one static rx queue - change it to allocate
a number of queues per the device capability (pre-9000 devices
have the number of rx queues set to one).
Subsequent generalizations are:
Change the code to access an explicit numbered rx queue only
when the queue number is known - when handling interrupt, when
accessing the default queue and when iterating the queues.
The rest of the functions will receive the rx queue as a pointer.
Generalize the warning in allocation failure to consider the
allocator status instead of a single rx queue status.
Move the rx initial pool of memory buffers to be shared among
all the queues and allocated to the default queue on init.
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Emmanuel Grumbach [Thu, 14 Jan 2016 07:39:21 +0000 (09:39 +0200)]
iwlwifi: pcie: buffer packets to avoid overflowing Tx queues
When the Tx queues are full above a threshold, we
immediately stop the mac80211's queue to stop getting new
packets. This worked until TSO was enabled.
With TSO, one single packet from mac80211 can use many
descriptors since a large send needs to be split into
several segments.
This means that stopping mac80211's queues is not enough
and we also need to ensure that we don't overflow the Tx
queues with one single packet from mac80211.
Add code to transport layer to do just that. Stop
mac80211's queue as soon as the queue is full above the
same threshold as before, and keep pushing the current
packet along with its segments on the queue, but check
that we don't overflow. If that would happen, buffer the
segments, and send them when there is room in the Tx queue
again. Of course, we first need to send the buffered
segments and only then, wake up mac80211's queues.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Amitkumar Karwar [Wed, 13 Jan 2016 09:26:57 +0000 (01:26 -0800)]
mwifiex: use SYNC flag for canceling host sleep
Host sleep is cancelled in sdio resume() handler.
Cfg80211's resume handler is immediately called after
this. SYNC flag here ensures that host sleep handshake
gets completed and we have valid "adapter->nd_config"
before we report host wakeup reason in cfg80211's
resume handler.
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Ganapathi Bhat [Wed, 13 Jan 2016 09:26:56 +0000 (01:26 -0800)]
mwifiex: add RSSI support for net-detect
This patch adds support for waking up the device on
finding better RSSI. Threshold RSSI value will be
configured by application.
Signed-off-by: Ganapathi Bhat <gbhat@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
chunfan chen [Wed, 13 Jan 2016 09:26:55 +0000 (01:26 -0800)]
mwifiex: report wakeup reason to cfg80211
This patch adds code to report wakeup reason to cfg80211
when system is resumed.
Signed-off-by: chunfan chen <jeffc@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
chunfan chen [Wed, 13 Jan 2016 09:26:54 +0000 (01:26 -0800)]
mwifiex: add wowlan net-detect support
This patch adds support for wakeup when configured
network is detected.
Signed-off-by: chunfan chen <jeffc@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
chunfan chen [Wed, 13 Jan 2016 09:26:53 +0000 (01:26 -0800)]
mwifiex: add wowlan info messages
This patch adds informative messages in wake up on
magic packet, disconnect, pattern configuration paths.
Signed-off-by: chunfan chen <jeffc@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Xinming Hu [Wed, 13 Jan 2016 09:26:52 +0000 (01:26 -0800)]
mwifiex: add schedule scan support
This patch add sched scan support for mwifiex, include cfg80211
sched_scan_start/sched_scan_stop handler, corresponding bgscan
command path and event handler.
Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: chunfan chen <jeffc@marvell.com>
Signed-off-by: Cathy Luo <cluo@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Shengzhen Li [Tue, 12 Jan 2016 13:43:16 +0000 (05:43 -0800)]
mwifiex: fix power state out of sync problem
It's been observed that driver's power state goes out
of sync with firmware in some corner cases. When this
happens, driver tries to download the data to firmware
assuming it's awake which causes Tx data timeout.
Main thread will process interrupt as soon as interrupt
handler fills 'int_status' variable.
This patch fixes the race issue by updating 'int_status'
at the end of mwifiex_interrupt_status().
Signed-off-by: Shengzhen Li <szli@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Xinming Hu [Mon, 11 Jan 2016 10:16:40 +0000 (02:16 -0800)]
mwifiex: enable pcie MSIx interrupt mode support
Newer pcie chipsets (8997 onwards) support MSIx. This
patch enables it.
Signed-off-by: Xinming Hu <huxm@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Andrzej Hajda [Thu, 7 Jan 2016 13:28:50 +0000 (14:28 +0100)]
wlcore: fix error handling in wlcore_event_fw_logger
wlcore_read/wlcore_write can return negative values so it should
be assigned to signed variable.
The problem has been detected using proposed semantic patch
scripts/coccinelle/tests/unsigned_lesser_than_zero.cocci [1].
[1]: http://permalink.gmane.org/gmane.linux.kernel/2120705
Fixes:
3719c17e1816 ("wlcore/wl18xx: fw logger over sdio")
Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Xinming Hu [Thu, 7 Jan 2016 07:40:49 +0000 (23:40 -0800)]
mwifiex: add debugfs file to read chip information
This patch add 'verext' debugfs item, which can be used to
get detailed chip specific information from our firmware.
Examples:
echo "1" > /sys/kernel/debug/mwifiex/mlan0/verext
cat /sys/kernel/debug/mwifiex/mlan0/verext
Signed-off-by: Shengzhen Li <szli@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
chunfan chen [Thu, 7 Jan 2016 07:40:48 +0000 (23:40 -0800)]
mwifiex: fix IBSS data path issue.
The port_open flag is not applicable for IBSS mode. IBSS data
path was broken when port_open flag was introduced.
This patch fixes the problem by correcting the checks.
Fixes:
5c8946330abfa4c ("mwifiex: enable traffic only when port is open")
Signed-off-by: chunfan chen <jeffc@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
chunfan chen [Thu, 7 Jan 2016 07:40:47 +0000 (23:40 -0800)]
mwifiex: firmware download enhancements
Same chip is being used by WLAN as well as bluetooth
drivers. Each driver needs to check during initialisation
if firmware is already active or it needs to be freshly
downloaded. If one driver has started downloading the
firmware, other finds the winner flag as false.
mwifiex_check_fw_status() checks firmware status and also
check if WLAN is the winner for firmware downloading.
Once we detect that other interface is downloading
the firmware, we call this routine again with max
poll count to wait until firmware is ready.
This patch splits the routine to avoid checking
winner status unnecessarily multiple times and ensures
that correct messages are displayed to user.
Firmware status poll count is also increased to 150.
Signed-off-by: Chunfan Chen <jeffc@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Uri Mashiach [Wed, 30 Dec 2015 13:35:32 +0000 (15:35 +0200)]
wlcore/wl12xx: spi: add device tree support
Add DT support for the wl1271 SPI WiFi.
Add documentation file for the wl1271 SPI WiFi.
Signed-off-by: Uri Mashiach <uri.mashiach@compulab.co.il>
Acked-by: Rob Herring <robh@kernel.org>
Tested-By: Sebastian Reichel <sre@kernel.org>
Reviewed-by: Grygorii Strashko <grygorii.strashko@ti.com>
Acked-by: Igor Grinberg <grinberg@compulab.co.il>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Uri Mashiach [Wed, 30 Dec 2015 13:35:31 +0000 (15:35 +0200)]
wlcore/wl12xx: spi: add power operation function
The power function uses a consumer regulator access to update the WiFi
enable GPIO value.
Signed-off-by: Uri Mashiach <uri.mashiach@compulab.co.il>
Tested-By: Sebastian Reichel <sre@kernel.org>
Reviewed-by: Grygorii Strashko <grygorii.strashko@ti.com>
Acked-by: Igor Grinberg <grinberg@compulab.co.il>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
David S. Miller [Fri, 15 Jan 2016 23:36:23 +0000 (18:36 -0500)]
Merge git://git./linux/kernel/git/davem/net
Linus Torvalds [Fri, 15 Jan 2016 21:33:12 +0000 (13:33 -0800)]
Merge git://git./linux/kernel/git/davem/net
Pull networking fixes from David Miller:
"A quick set of bug fixes after there initial networking merge:
1) Netlink multicast group storage allocator only was tested with
nr_groups equal to 1, make it work for other values too. From
Matti Vaittinen.
2) Check build_skb() return value in macb and hip04_eth drivers, from
Weidong Wang.
3) Don't leak x25_asy on x25_asy_open() failure.
4) More DMA map/unmap fixes in 3c59x from Neil Horman.
5) Don't clobber IP skb control block during GSO segmentation, from
Konstantin Khlebnikov.
6) ECN helpers for ipv6 don't fixup the checksum, from Eric Dumazet.
7) Fix SKB segment utilization estimation in xen-netback, from David
Vrabel.
8) Fix lockdep splat in bridge addrlist handling, from Nikolay
Aleksandrov"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (26 commits)
bgmac: Fix reversed test of build_skb() return value.
bridge: fix lockdep addr_list_lock false positive splat
net: smsc: Add support h8300
xen-netback: free queues after freeing the net device
xen-netback: delete NAPI instance when queue fails to initialize
xen-netback: use skb to determine number of required guest Rx requests
net: sctp: Move sequence start handling into sctp_transport_get_idx()
ipv6: update skb->csum when CE mark is propagated
net: phy: turn carrier off on phy attach
net: macb: clear interrupts when disabling them
sctp: support to lookup with ep+paddr in transport rhashtable
net: hns: fixes no syscon error when init mdio
dts: hisi: fixes no syscon fault when init mdio
net: preserve IP control block during GSO segmentation
fsl/fman: Delete one function call "put_device" in dtsec_config()
hip04_eth: fix missing error handle for build_skb failed
3c59x: fix another page map/single unmap imbalance
3c59x: balance page maps and unmaps
x25_asy: Free x25_asy on x25_asy_open() failure.
mlxsw: fix SWITCHDEV_OBJ_ID_PORT_MDB
...
Linus Torvalds [Fri, 15 Jan 2016 21:28:12 +0000 (13:28 -0800)]
Merge git://git./linux/kernel/git/davem/sparc
Pull sparc fixes from David Miller:
"Two sparc bug fixes"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
sparc64: Fix numa node distance initialization
sparc64: fix incorrect sign extension in sys_sparc64_personality
Linus Torvalds [Fri, 15 Jan 2016 21:18:47 +0000 (13:18 -0800)]
Merge tag 'powerpc-4.5-1' of git://git./linux/kernel/git/powerpc/linux
Pull powerpc updates from Michael Ellerman:
"Core:
- Ground work for the new Power9 MMU from Aneesh Kumar K.V
- Optimise FP/VMX/VSX context switching from Anton Blanchard
Misc:
- Various cleanups from Krzysztof Kozlowski, John Ogness, Rashmica
Gupta, Russell Currey, Gavin Shan, Daniel Axtens, Michael Neuling,
Andrew Donnellan
- Allow wrapper to work on non-english system from Laurent Vivier
- Add rN aliases to the pt_regs_offset table from Rashmica Gupta
- Fix module autoload for rackmeter & axonram drivers from Luis de
Bethencourt
- Include KVM guest test in all interrupt vectors from Paul Mackerras
- Fix DSCR inheritance over fork() from Anton Blanchard
- Make value-returning atomics & {cmp}xchg* & their atomic_ versions
fully ordered from Boqun Feng
- Print MSR TM bits in oops messages from Michael Neuling
- Add TM signal return & invalid stack selftests from Michael Neuling
- Limit EPOW reset event warnings from Vipin K Parashar
- Remove the Cell QPACE code from Rashmica Gupta
- Append linux_banner to exception information in xmon from Rashmica
Gupta
- Add selftest to check if VSRs are corrupted from Rashmica Gupta
- Remove broken GregorianDay() from Daniel Axtens
- Import Anton's context_switch2 benchmark into selftests from
Michael Ellerman
- Add selftest script to test HMI functionality from Daniel Axtens
- Remove obsolete OPAL v2 support from Stewart Smith
- Make enter_rtas() private from Michael Ellerman
- PPR exception cleanups from Michael Ellerman
- Add page soft dirty tracking from Laurent Dufour
- Add support for Nvlink NPUs from Alistair Popple
- Add support for kexec on 476fpe from Alistair Popple
- Enable kernel CPU dlpar from sysfs from Nathan Fontenot
- Copy only required pieces of the mm_context_t to the paca from
Michael Neuling
- Add a kmsg_dumper that flushes OPAL console output on panic from
Russell Currey
- Implement save_stack_trace_regs() to enable kprobe stack tracing
from Steven Rostedt
- Add HWCAP bits for Power9 from Michael Ellerman
- Fix _PAGE_PTE breaking swapoff from Aneesh Kumar K.V
- Fix _PAGE_SWP_SOFT_DIRTY breaking swapoff from Hugh Dickins
- scripts/recordmcount.pl: support data in text section on powerpc
from Ulrich Weigand
- Handle R_PPC64_ENTRY relocations in modules from Ulrich Weigand
cxl:
- cxl: Fix possible idr warning when contexts are released from
Vaibhav Jain
- cxl: use correct operator when writing pcie config space values
from Andrew Donnellan
- cxl: Fix DSI misses when the context owning task exits from Vaibhav
Jain
- cxl: fix build for GCC 4.6.x from Brian Norris
- cxl: use -Werror only with CONFIG_PPC_WERROR from Brian Norris
- cxl: Enable PCI device ID for future IBM CXL adapter from Uma
Krishnan
Freescale:
- Freescale updates from Scott: Highlights include moving QE code out
of arch/powerpc (to be shared with arm), device tree updates, and
minor fixes"
* tag 'powerpc-4.5-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (149 commits)
powerpc/module: Handle R_PPC64_ENTRY relocations
scripts/recordmcount.pl: support data in text section on powerpc
powerpc/powernv: Fix OPAL_CONSOLE_FLUSH prototype and usages
powerpc/mm: fix _PAGE_SWP_SOFT_DIRTY breaking swapoff
powerpc/mm: Fix _PAGE_PTE breaking swapoff
cxl: Enable PCI device ID for future IBM CXL adapter
cxl: use -Werror only with CONFIG_PPC_WERROR
cxl: fix build for GCC 4.6.x
powerpc: Add HWCAP bits for Power9
powerpc/powernv: Reserve PE#0 on NPU
powerpc/powernv: Change NPU PE# assignment
powerpc/powernv: Fix update of NVLink DMA mask
powerpc/powernv: Remove misleading comment in pci.c
powerpc: Implement save_stack_trace_regs() to enable kprobe stack tracing
powerpc: Fix build break due to paca mm_context_t changes
cxl: Fix DSI misses when the context owning task exits
MAINTAINERS: Update Scott Wood's e-mail address
powerpc/powernv: Fix minor off-by-one error in opal_mce_check_early_recovery()
powerpc: Fix style of self-test config prompts
powerpc/powernv: Only delay opal_rtc_read() retry when necessary
...
David S. Miller [Fri, 15 Jan 2016 21:07:13 +0000 (16:07 -0500)]
bgmac: Fix reversed test of build_skb() return value.
Fixes:
f1640c3ddeec ("bgmac: fix a missing check for build_skb")
Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Fri, 15 Jan 2016 21:01:01 +0000 (13:01 -0800)]
Merge tag 'vfio-v4.5-rc1' of git://github.com/awilliam/linux-vfio
Pull VFIO updates from Alex Williamson:
- Fixes in AMD xgbe reset, spapr structure padding, type 1 flags (Dan
Carpenter, Alexey Kardashevskiy, Pierre Morel)
- Re-introduce no-iommu mode, with a user this time (Alex Williamson)
* tag 'vfio-v4.5-rc1' of git://github.com/awilliam/linux-vfio:
vfio/iommu_type1: make use of info.flags
vfio: Include No-IOMMU mode
vfio: Add explicit alignments in vfio_iommu_spapr_tce_create
VFIO: platform: reset: fix a warning message condition
Linus Torvalds [Fri, 15 Jan 2016 20:49:44 +0000 (12:49 -0800)]
Merge tag 'nfsd-4.5' of git://linux-nfs.org/~bfields/linux
Pull nfsd updates from Bruce Fields:
"Smaller bugfixes and cleanup, including a fix for a failures of
kerberized NFSv4.1 mounts, and Scott Mayhew's work addressing ACK
storms that can affect some high-availability NFS setups"
* tag 'nfsd-4.5' of git://linux-nfs.org/~bfields/linux:
nfsd: add new io class tracepoint
nfsd: give up on CB_LAYOUTRECALLs after two lease periods
nfsd: Fix nfsd leaks sunrpc module references
lockd: constify nlmsvc_binding structure
lockd: use to_delayed_work
nfsd: use to_delayed_work
Revert "svcrdma: Do not send XDR roundup bytes for a write chunk"
lockd: Register callbacks on the inetaddr_chain and inet6addr_chain
nfsd: Register callbacks on the inetaddr_chain and inet6addr_chain
sunrpc: Add a function to close temporary transports immediately
nfsd: don't base cl_cb_status on stale information
nfsd4: fix gss-proxy 4.1 mounts for some AD principals
nfsd: fix unlikely NULL deref in mach_creds_match
nfsd: minor consolidation of mach_cred handling code
nfsd: helper for dup of possibly NULL string
svcrpc: move some initialization to common code
nfsd: fix a warning message
nfsd: constify nfsd4_callback_ops structure
nfsd: recover: constify nfsd4_client_tracking_ops structures
svcrdma: Do not send XDR roundup bytes for a write chunk
Linus Torvalds [Fri, 15 Jan 2016 20:41:32 +0000 (12:41 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs
Pull vfs regression fix from Al Viro:
"Fix for braino introduced in vfs.git#work.misc"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
amdkfd: Copy from the proper user command pointer
Nikolay Aleksandrov [Fri, 15 Jan 2016 18:03:54 +0000 (19:03 +0100)]
bridge: fix lockdep addr_list_lock false positive splat
After promisc mode management was introduced a bridge device could do
dev_set_promiscuity from its ndo_change_rx_flags() callback which in
turn can be called after the bridge's addr_list_lock has been taken
(e.g. by dev_uc_add). This causes a false positive lockdep splat because
the port interfaces' addr_list_lock is taken when br_manage_promisc()
runs after the bridge's addr list lock was already taken.
To remove the false positive introduce a custom bridge addr_list_lock
class and set it on bridge init.
A simple way to reproduce this is with the following:
$ brctl addbr br0
$ ip l add l br0 br0.100 type vlan id 100
$ ip l set br0 up
$ ip l set br0.100 up
$ echo 1 > /sys/class/net/br0/bridge/vlan_filtering
$ brctl addif br0 eth0
Splat:
[ 43.684325] =============================================
[ 43.684485] [ INFO: possible recursive locking detected ]
[ 43.684636] 4.4.0-rc8+ #54 Not tainted
[ 43.684755] ---------------------------------------------
[ 43.684906] brctl/1187 is trying to acquire lock:
[ 43.685047] (_xmit_ETHER){+.....}, at: [<
ffffffff8150169e>] dev_set_rx_mode+0x1e/0x40
[ 43.685460] but task is already holding lock:
[ 43.685618] (_xmit_ETHER){+.....}, at: [<
ffffffff815072a7>] dev_uc_add+0x27/0x80
[ 43.686015] other info that might help us debug this:
[ 43.686316] Possible unsafe locking scenario:
[ 43.686743] CPU0
[ 43.686967] ----
[ 43.687197] lock(_xmit_ETHER);
[ 43.687544] lock(_xmit_ETHER);
[ 43.687886] *** DEADLOCK ***
[ 43.688438] May be due to missing lock nesting notation
[ 43.688882] 2 locks held by brctl/1187:
[ 43.689134] #0: (rtnl_mutex){+.+.+.}, at: [<
ffffffff81510317>] rtnl_lock+0x17/0x20
[ 43.689852] #1: (_xmit_ETHER){+.....}, at: [<
ffffffff815072a7>] dev_uc_add+0x27/0x80
[ 43.690575] stack backtrace:
[ 43.690970] CPU: 0 PID: 1187 Comm: brctl Not tainted 4.4.0-rc8+ #54
[ 43.691270] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.8.1-20150318_183358- 04/01/2014
[ 43.691770]
ffffffff826a25c0 ffff8800369fb8e0 ffffffff81360ceb ffffffff826a25c0
[ 43.692425]
ffff8800369fb9b8 ffffffff810d0466 ffff8800369fb968 ffffffff81537139
[ 43.693071]
ffff88003a08c880 0000000000000000 00000000ffffffff 0000000002080020
[ 43.693709] Call Trace:
[ 43.693931] [<
ffffffff81360ceb>] dump_stack+0x4b/0x70
[ 43.694199] [<
ffffffff810d0466>] __lock_acquire+0x1e46/0x1e90
[ 43.694483] [<
ffffffff81537139>] ? netlink_broadcast_filtered+0x139/0x3e0
[ 43.694789] [<
ffffffff8153b5da>] ? nlmsg_notify+0x5a/0xc0
[ 43.695064] [<
ffffffff810d10f5>] lock_acquire+0xe5/0x1f0
[ 43.695340] [<
ffffffff8150169e>] ? dev_set_rx_mode+0x1e/0x40
[ 43.695623] [<
ffffffff815edea5>] _raw_spin_lock_bh+0x45/0x80
[ 43.695901] [<
ffffffff8150169e>] ? dev_set_rx_mode+0x1e/0x40
[ 43.696180] [<
ffffffff8150169e>] dev_set_rx_mode+0x1e/0x40
[ 43.696460] [<
ffffffff8150189c>] dev_set_promiscuity+0x3c/0x50
[ 43.696750] [<
ffffffffa0586845>] br_port_set_promisc+0x25/0x50 [bridge]
[ 43.697052] [<
ffffffffa05869aa>] br_manage_promisc+0x8a/0xe0 [bridge]
[ 43.697348] [<
ffffffffa05826ee>] br_dev_change_rx_flags+0x1e/0x20 [bridge]
[ 43.697655] [<
ffffffff81501532>] __dev_set_promiscuity+0x132/0x1f0
[ 43.697943] [<
ffffffff81501672>] __dev_set_rx_mode+0x82/0x90
[ 43.698223] [<
ffffffff815072de>] dev_uc_add+0x5e/0x80
[ 43.698498] [<
ffffffffa05b3c62>] vlan_device_event+0x542/0x650 [8021q]
[ 43.698798] [<
ffffffff8109886d>] notifier_call_chain+0x5d/0x80
[ 43.699083] [<
ffffffff810988b6>] raw_notifier_call_chain+0x16/0x20
[ 43.699374] [<
ffffffff814f456e>] call_netdevice_notifiers_info+0x6e/0x80
[ 43.699678] [<
ffffffff814f4596>] call_netdevice_notifiers+0x16/0x20
[ 43.699973] [<
ffffffffa05872be>] br_add_if+0x47e/0x4c0 [bridge]
[ 43.700259] [<
ffffffffa058801e>] add_del_if+0x6e/0x80 [bridge]
[ 43.700548] [<
ffffffffa0588b5f>] br_dev_ioctl+0xaf/0xc0 [bridge]
[ 43.700836] [<
ffffffff8151a7ac>] dev_ifsioc+0x30c/0x3c0
[ 43.701106] [<
ffffffff8151aac9>] dev_ioctl+0xf9/0x6f0
[ 43.701379] [<
ffffffff81254345>] ? mntput_no_expire+0x5/0x450
[ 43.701665] [<
ffffffff812543ee>] ? mntput_no_expire+0xae/0x450
[ 43.701947] [<
ffffffff814d7b02>] sock_do_ioctl+0x42/0x50
[ 43.702219] [<
ffffffff814d8175>] sock_ioctl+0x1e5/0x290
[ 43.702500] [<
ffffffff81242d0b>] do_vfs_ioctl+0x2cb/0x5c0
[ 43.702771] [<
ffffffff81243079>] SyS_ioctl+0x79/0x90
[ 43.703033] [<
ffffffff815eebb6>] entry_SYSCALL_64_fastpath+0x16/0x7a
CC: Vlad Yasevich <vyasevic@redhat.com>
CC: Stephen Hemminger <stephen@networkplumber.org>
CC: Bridge list <bridge@lists.linux-foundation.org>
CC: Andy Gospodarek <gospo@cumulusnetworks.com>
CC: Roopa Prabhu <roopa@cumulusnetworks.com>
Fixes:
2796d0c648c9 ("bridge: Automatically manage port promiscuous mode.")
Reported-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Fri, 15 Jan 2016 20:28:00 +0000 (12:28 -0800)]
Merge tag 'md/4.5' of git://neil.brown.name/md
Pull md updates from Neil Brown:
"Mostly clustered-raid1 and raid5 journal updates. one Y2038 fix and
other minor stuff.
One patch removes me from the MAINTAINERS file and adds a record of my
md maintainership to Credits"
Many thanks to Neil, who has been around for a _looong_ time.
* tag 'md/4.5' of git://neil.brown.name/md: (26 commits)
md/raid: only permit hot-add of compatible integrity profiles
Remove myself as MD Maintainer, and add to Credits.
raid5-cache: handle journal hotadd in quiesce
MD: add journal with array suspended
md: set MD_HAS_JOURNAL in correct places
md: Remove 'ready' field from mddev.
md: remove unnecesary md_new_event_inintr
raid5: allow r5l_io_unit allocations to fail
raid5-cache: use a mempool for the metadata block
raid5-cache: use a bio_set
raid5-cache: add journal hot add/remove support
drivers: md: use ktime_get_real_seconds()
md: avoid warning for 32-bit sector_t
raid5-cache: free meta_page earlier
raid5-cache: simplify r5l_move_io_unit_list
md: update comment for md_allow_write
md-cluster: update comments for MD_CLUSTER_SEND_LOCKED_ALREADY
md-cluster: Protect communication with mutexes
md-cluster: Defer MD reloading to mddev->thread
md-cluster: update the documentation
...
Linus Torvalds [Fri, 15 Jan 2016 20:14:47 +0000 (12:14 -0800)]
Merge tag 'regulator-v4.5' of git://git./linux/kernel/git/broonie/regulator
Pull regulator updates from Mark Brown:
"Aside from a fix for a spurious warning (which caused more problems
than it fixed in the fixing really) this is all driver updates,
including new drivers for Dialog PV88060/90 and TI LM363x and TPS65086
devices. The qcom_smd driver has had PM8916 and PMA8084 support
added"
* tag 'regulator-v4.5' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator: (36 commits)
regulator: core: remove some dead code
regulator: core: use dev_to_rdev
regulator: lp872x: Get rid of duplicate reference to DVS GPIO
regulator: lp872x: Add missing of_match in regulators descriptions
regulator: axp20x: Fix GPIO LDO enable value for AXP22x
regulator: lp8788: constify regulator_ops structures
regulator: wm8*: constify regulator_ops structures
regulator: da9*: constify regulator_ops structures
regulator: mt6311: Use REGCACHE_RBTREE
regulator: tps65917/palmas: Add bypass ops for LDOs with bypass capability
regulator: qcom-smd: Add support for PMA8084
regulator: qcom-smd: Add PM8916 support
soc: qcom: documentation: Update SMD/RPM Docs
regulator: pv88090: logical vs bitwise AND typo
regulator: pv88090: Fix irq leak
regulator: pv88090: new regulator driver
regulator: wm831x-ldo: Use platform_register/unregister_drivers()
regulator: wm831x-dcdc: Use platform_register/unregister_drivers()
regulator: lp8788-ldo: Use platform_register/unregister_drivers()
regulator: core: Fix nested locking of supplies
...
Yoshinori Sato [Fri, 15 Jan 2016 15:05:09 +0000 (00:05 +0900)]
net: smsc: Add support h8300
Add H8/300 platform support for smc91x
Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
Borislav Petkov [Fri, 15 Jan 2016 18:26:44 +0000 (19:26 +0100)]
amdkfd: Copy from the proper user command pointer
8f1d57c17248 ("amdkfd: don't open-code memdup_user()") mistakenly uses
an uninitialized local pointer, gcc complains:
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c: In function ‘kfd_ioctl_dbg_address_watch’:
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c:562:12: warning: ‘args_buff’ may be used uninitialized in this function [-Wmaybe-uninitialized]
args_buff = memdup_user(args_buff,
^
Fix it.
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Linus Torvalds [Fri, 15 Jan 2016 20:13:58 +0000 (12:13 -0800)]
Merge branch 'mailbox-for-next' of git://git.linaro.org/landing-teams/working/fujitsu/integration
Pull mailbox fixlet from Jussi Brar.
* 'mailbox-for-next' of git://git.linaro.org/landing-teams/working/fujitsu/integration:
mailbox: constify mbox_chan_ops structure
David S. Miller [Fri, 15 Jan 2016 20:13:19 +0000 (15:13 -0500)]
Merge branch 'xen-netback-fixes'
David Vrabel says:
====================
xen-netback: use skb to determine number of required (etc.)
"xen-netback: use skb to determine number of required" plus two other
minor fixes I found down the back of the sofa.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
David Vrabel [Fri, 15 Jan 2016 14:55:36 +0000 (14:55 +0000)]
xen-netback: free queues after freeing the net device
If a queue still has a NAPI instance added to the net device, freeing
the queues early results in a use-after-free.
The shouldn't ever happen because we disconnect and tear down all queues
before freeing the net device, but doing this makes it obviously safe.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David Vrabel [Fri, 15 Jan 2016 14:55:35 +0000 (14:55 +0000)]
xen-netback: delete NAPI instance when queue fails to initialize
When xenvif_connect() fails it may leave a stale NAPI instance added to
the device. Make sure we delete it in the error path.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David Vrabel [Fri, 15 Jan 2016 14:55:34 +0000 (14:55 +0000)]
xen-netback: use skb to determine number of required guest Rx requests
Using the MTU or GSO size to determine the number of required guest Rx
requests for an skb was subtly broken since these value may change at
runtime.
After
1650d5455bd2dc6b5ee134bd6fc1a3236c266b5b (xen-netback: always
fully coalesce guest Rx packets) we always fully pack a packet into
its guest Rx slots. Calculating the number of required slots from the
packet length is then easy.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Geert Uytterhoeven [Fri, 15 Jan 2016 13:44:31 +0000 (14:44 +0100)]
net: sctp: Move sequence start handling into sctp_transport_get_idx()
net/sctp/proc.c: In function ‘sctp_transport_get_idx’:
net/sctp/proc.c:313: warning: ‘obj’ may be used uninitialized in this function
This is currently a false positive, as all callers check for a zero
offset first, and handle this case in the exact same way.
Move the check and handling into sctp_transport_get_idx() to kill the
compiler warning, and avoid future bugs.
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Fri, 15 Jan 2016 12:56:56 +0000 (04:56 -0800)]
ipv6: update skb->csum when CE mark is propagated
When a tunnel decapsulates the outer header, it has to comply
with RFC 6080 and eventually propagate CE mark into inner header.
It turns out IP6_ECN_set_ce() does not correctly update skb->csum
for CHECKSUM_COMPLETE packets, triggering infamous "hw csum failure"
messages and stack traces.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Fri, 15 Jan 2016 19:51:51 +0000 (11:51 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jack/linux-fs
Pull UDF fixes and quota cleanups from Jan Kara:
"Several UDF fixes and some minor quota cleanups"
* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
udf: Check output buffer length when converting name to CS0
udf: Prevent buffer overrun with multi-byte characters
quota: constify qtree_fmt_operations structures
udf: avoid uninitialized variable use
udf: Fix lost indirect extent block
udf: Factor out code for creating indirect extent
udf: limit the maximum number of indirect extents in a row
udf: limit the maximum number of TD redirections
fs: make quota/dquot.c explicitly non-modular
fs: make quota/netlink.c explicitly non-modular
Sjoerd Simons [Thu, 14 Jan 2016 20:57:18 +0000 (21:57 +0100)]
net: phy: turn carrier off on phy attach
The operstate of a networking device initially IF_OPER_UNKNOWN aka
"unknown", updated on carrier state changes (with carrier state being on
by default). This means it will stay unknown unless the carrier state
goes to off at some point, which is not the case if the phy is already
up/connected at startup.
Explicitly turn off the carrier on phy attach, leaving the phy state
machine to turn the carrier on when it has done the initial negotiation.
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Nathan Sullivan [Thu, 14 Jan 2016 19:27:27 +0000 (13:27 -0600)]
net: macb: clear interrupts when disabling them
Disabling interrupts with the IDR register does not stop the macb hardware
from asserting its interrupt line if there are interrupts pending. Always
clear the interrupts using ISR, and be sure to write it on hardware that
is not read-to-clear, like Zynq. Not doing so will cause interrupts when
the driver doesn't expect them.
Signed-off-by: Nathan Sullivan <nathan.sullivan@ni.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Fri, 15 Jan 2016 19:41:44 +0000 (11:41 -0800)]
Merge branch 'akpm' (patches from Andrew)
Merge first patch-bomb from Andrew Morton:
- A few hotfixes which missed 4.4 becasue I was asleep. cc'ed to
-stable
- A few misc fixes
- OCFS2 updates
- Part of MM. Including pretty large changes to page-flags handling
and to thp management which have been buffered up for 2-3 cycles now.
I have a lot of MM material this time.
[ It turns out the THP part wasn't quite ready, so that got dropped from
this series - Linus ]
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (117 commits)
zsmalloc: reorganize struct size_class to pack 4 bytes hole
mm/zbud.c: use list_last_entry() instead of list_tail_entry()
zram/zcomp: do not zero out zcomp private pages
zram: pass gfp from zcomp frontend to backend
zram: try vmalloc() after kmalloc()
zram/zcomp: use GFP_NOIO to allocate streams
mm: add tracepoint for scanning pages
drivers/base/memory.c: fix kernel warning during memory hotplug on ppc64
mm/page_isolation: use macro to judge the alignment
mm: fix noisy sparse warning in LIBCFS_ALLOC_PRE()
mm: rework virtual memory accounting
include/linux/memblock.h: fix ordering of 'flags' argument in comments
mm: move lru_to_page to mm_inline.h
Documentation/filesystems: describe the shared memory usage/accounting
memory-hotplug: don't BUG() in register_memory_resource()
hugetlb: make mm and fs code explicitly non-modular
mm/swapfile.c: use list_for_each_entry_safe in free_swap_count_continuations
mm: /proc/pid/clear_refs: no need to clear VM_SOFTDIRTY in clear_soft_dirty_pmd()
mm: make sure isolate_lru_page() is never called for tail page
vmstat: make vmstat_updater deferrable again and shut down on idle
...
Xin Long [Thu, 14 Jan 2016 05:49:34 +0000 (13:49 +0800)]
sctp: support to lookup with ep+paddr in transport rhashtable
Now, when we sendmsg, we translate the ep to laddr by selecting the
first element of the list, and then do a lookup for a transport.
But sctp_hash_cmp() will compare it against asoc addr_list, which may
be a subset of ep addr_list, meaning that this chosen laddr may not be
there, and thus making it impossible to find the transport.
So we fix it by using ep + paddr to lookup transports in hashtable. In
sctp_hash_cmp, if .ep is set, we will check if this ep == asoc->ep,
or we will do the laddr check.
Fixes:
d6c0256a60e6 ("sctp: add the rhashtable apis for sctp global transport hashtable")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reported-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Weijie Yang [Thu, 14 Jan 2016 23:22:40 +0000 (15:22 -0800)]
zsmalloc: reorganize struct size_class to pack 4 bytes hole
Reoder the pages_per_zspage field in struct size_class which can
eliminate the 4 bytes hole between it and stats field.
Signed-off-by: Weijie Yang <weijie.yang@samsung.com>
Reviewed-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Geliang Tang [Thu, 14 Jan 2016 23:22:38 +0000 (15:22 -0800)]
mm/zbud.c: use list_last_entry() instead of list_tail_entry()
list_last_entry*( has been defined in list.h, so replace
list_tail_entry() with it.
Signed-off-by: Geliang Tang <geliangtang@163.com>
Cc: Seth Jennings <sjennings@variantweb.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sergey Senozhatsky [Thu, 14 Jan 2016 23:22:35 +0000 (15:22 -0800)]
zram/zcomp: do not zero out zcomp private pages
Do not __GFP_ZERO allocated zcomp ->private pages. We keep allocated
streams around and use them for read/write requests, so we supply a
zeroed out ->private to compression algorithm as a scratch buffer only
once -- the first time we use that stream. For the rest of IO requests
served by this stream ->private usually contains some temporarily data
from the previous requests.
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Minchan Kim [Thu, 14 Jan 2016 23:22:32 +0000 (15:22 -0800)]
zram: pass gfp from zcomp frontend to backend
Each zcomp backend uses own gfp flag but it's pointless because the
context they could be called is driven by upper layer(ie, zcomp
frontend). As well, zcomp frondend could call them in different
context. One context(ie, zram init part) is it should be better to make
sure successful allocation other context(ie, further stream allocation
part for accelarating I/O speed) is just optional so let's pass gfp down
from driver (ie, zcomp frontend) like normal MM convention.
[sergey.senozhatsky@gmail.com: add missing __vmalloc zero and highmem gfps]
Signed-off-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Kyeongdon Kim [Thu, 14 Jan 2016 23:22:29 +0000 (15:22 -0800)]
zram: try vmalloc() after kmalloc()
When we're using LZ4 multi compression streams for zram swap, we found
out page allocation failure message in system running test. That was
not only once, but a few(2 - 5 times per test). Also, some failure
cases were continually occurring to try allocation order 3.
In order to make parallel compression private data, we should call
kzalloc() with order 2/3 in runtime(lzo/lz4). But if there is no order
2/3 size memory to allocate in that time, page allocation fails. This
patch makes to use vmalloc() as fallback of kmalloc(), this prevents
page alloc failure warning.
After using this, we never found warning message in running test, also
It could reduce process startup latency about 60-120ms in each case.
For reference a call trace :
Binder_1: page allocation failure: order:3, mode:0x10c0d0
CPU: 0 PID: 424 Comm: Binder_1 Tainted: GW 3.10.49-perf-g991d02b-dirty #20
Call trace:
dump_backtrace+0x0/0x270
show_stack+0x10/0x1c
dump_stack+0x1c/0x28
warn_alloc_failed+0xfc/0x11c
__alloc_pages_nodemask+0x724/0x7f0
__get_free_pages+0x14/0x5c
kmalloc_order_trace+0x38/0xd8
zcomp_lz4_create+0x2c/0x38
zcomp_strm_alloc+0x34/0x78
zcomp_strm_multi_find+0x124/0x1ec
zcomp_strm_find+0xc/0x18
zram_bvec_rw+0x2fc/0x780
zram_make_request+0x25c/0x2d4
generic_make_request+0x80/0xbc
submit_bio+0xa4/0x15c
__swap_writepage+0x218/0x230
swap_writepage+0x3c/0x4c
shrink_page_list+0x51c/0x8d0
shrink_inactive_list+0x3f8/0x60c
shrink_lruvec+0x33c/0x4cc
shrink_zone+0x3c/0x100
try_to_free_pages+0x2b8/0x54c
__alloc_pages_nodemask+0x514/0x7f0
__get_free_pages+0x14/0x5c
proc_info_read+0x50/0xe4
vfs_read+0xa0/0x12c
SyS_read+0x44/0x74
DMA: 3397*4kB (MC) 26*8kB (RC) 0*16kB 0*32kB 0*64kB 0*128kB 0*256kB
0*512kB 0*1024kB 0*2048kB 0*4096kB = 13796kB
[minchan@kernel.org: change vmalloc gfp and adding comment about gfp]
[sergey.senozhatsky@gmail.com: tweak comments and styles]
Signed-off-by: Kyeongdon Kim <kyeongdon.kim@lge.com>
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sergey Senozhatsky [Thu, 14 Jan 2016 23:22:26 +0000 (15:22 -0800)]
zram/zcomp: use GFP_NOIO to allocate streams
We can end up allocating a new compression stream with GFP_KERNEL from
within the IO path, which may result is nested (recursive) IO
operations. That can introduce problems if the IO path in question is a
reclaimer, holding some locks that will deadlock nested IOs.
Allocate streams and working memory using GFP_NOIO flag, forbidding
recursive IO and FS operations.
An example:
inconsistent {IN-RECLAIM_FS-W} -> {RECLAIM_FS-ON-W} usage.
git/20158 [HC0[0]:SC0[0]:HE1:SE1] takes:
(jbd2_handle){+.+.?.}, at: start_this_handle+0x4ca/0x555
{IN-RECLAIM_FS-W} state was registered at:
__lock_acquire+0x8da/0x117b
lock_acquire+0x10c/0x1a7
start_this_handle+0x52d/0x555
jbd2__journal_start+0xb4/0x237
__ext4_journal_start_sb+0x108/0x17e
ext4_dirty_inode+0x32/0x61
__mark_inode_dirty+0x16b/0x60c
iput+0x11e/0x274
__dentry_kill+0x148/0x1b8
shrink_dentry_list+0x274/0x44a
prune_dcache_sb+0x4a/0x55
super_cache_scan+0xfc/0x176
shrink_slab.part.14.constprop.25+0x2a2/0x4d3
shrink_zone+0x74/0x140
kswapd+0x6b7/0x930
kthread+0x107/0x10f
ret_from_fork+0x3f/0x70
irq event stamp: 138297
hardirqs last enabled at (138297): debug_check_no_locks_freed+0x113/0x12f
hardirqs last disabled at (138296): debug_check_no_locks_freed+0x33/0x12f
softirqs last enabled at (137818): __do_softirq+0x2d3/0x3e9
softirqs last disabled at (137813): irq_exit+0x41/0x95
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(jbd2_handle);
<Interrupt>
lock(jbd2_handle);
*** DEADLOCK ***
5 locks held by git/20158:
#0: (sb_writers#7){.+.+.+}, at: [<
ffffffff81155411>] mnt_want_write+0x24/0x4b
#1: (&type->i_mutex_dir_key#2/1){+.+.+.}, at: [<
ffffffff81145087>] lock_rename+0xd9/0xe3
#2: (&sb->s_type->i_mutex_key#11){+.+.+.}, at: [<
ffffffff8114f8e2>] lock_two_nondirectories+0x3f/0x6b
#3: (&sb->s_type->i_mutex_key#11/4){+.+.+.}, at: [<
ffffffff8114f909>] lock_two_nondirectories+0x66/0x6b
#4: (jbd2_handle){+.+.?.}, at: [<
ffffffff811e31db>] start_this_handle+0x4ca/0x555
stack backtrace:
CPU: 2 PID: 20158 Comm: git Not tainted 4.1.0-rc7-next-
20150615-dbg-00016-g8bdf555-dirty #211
Call Trace:
dump_stack+0x4c/0x6e
mark_lock+0x384/0x56d
mark_held_locks+0x5f/0x76
lockdep_trace_alloc+0xb2/0xb5
kmem_cache_alloc_trace+0x32/0x1e2
zcomp_strm_alloc+0x25/0x73 [zram]
zcomp_strm_multi_find+0xe7/0x173 [zram]
zcomp_strm_find+0xc/0xe [zram]
zram_bvec_rw+0x2ca/0x7e0 [zram]
zram_make_request+0x1fa/0x301 [zram]
generic_make_request+0x9c/0xdb
submit_bio+0xf7/0x120
ext4_io_submit+0x2e/0x43
ext4_bio_write_page+0x1b7/0x300
mpage_submit_page+0x60/0x77
mpage_map_and_submit_buffers+0x10f/0x21d
ext4_writepages+0xc8c/0xe1b
do_writepages+0x23/0x2c
__filemap_fdatawrite_range+0x84/0x8b
filemap_flush+0x1c/0x1e
ext4_alloc_da_blocks+0xb8/0x117
ext4_rename+0x132/0x6dc
? mark_held_locks+0x5f/0x76
ext4_rename2+0x29/0x2b
vfs_rename+0x540/0x636
SyS_renameat2+0x359/0x44d
SyS_rename+0x1e/0x20
entry_SYSCALL_64_fastpath+0x12/0x6f
[minchan@kernel.org: add stable mark]
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Cc: Kyeongdon Kim <kyeongdon.kim@lge.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>