Commit graph

388,156 commits

Author SHA1 Message Date
Lars-Peter Clausen
8abac3ba51 regmap: cache Fix regcache-rbtree sync
The last register block, which falls into the specified range, is not handled
correctly. The formula which calculates the number of register which should be
synced is inverse (and off by one). E.g. if all registers in that block should
be synced only one is synced, and if only one should be synced all (but one) are
synced. To calculate the number of registers that need to be synced we need to
subtract the number of the first register in the block from the max register
number and add one. This patch updates the code accordingly.

The issue was introduced in commit ac8d91c ("regmap: Supply ranges to the sync
operations").

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@vger.kernel.org
2013-03-13 19:07:19 +00:00
Mark Brown
76bf969e6f ASoC: arizona: Ensure we clock two channels for I2S mode
I2S requires stereo clocking even for mono data.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
2013-03-13 19:03:17 +00:00
Sachin Kamat
9cca117359 pinctrl: pinctrl-nomadik-stn8815: Fix checkpatch error
Fixes the following error:
ERROR: space required after that ',' (ctx:VxV)

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-03-13 19:50:20 +01:00
Haojian Zhuang
7cba5b3f5f pinctrl: single: correct argument for pinconf
pcs_pinconf_set() is always using "arg << shift" to configure two
parameters case. But pcs_add_conf2() didn't remove shift for config
argument. So correct it.

Signed-off-by: Haojian Zhuang <haojian.zhuang@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-03-13 19:50:19 +01:00
Sachin Kamat
bf4d7be57b pinctrl: generic: Fix compilation error
The function definition of pinconf_generic_dump_config is defined
under CONFIG_DEBUG_FS macro. Define the declaration too under this macro.

Without this patch we get the following build error:
drivers/built-in.o: In function `pcs_pinconf_config_dbg_show':
drivers/pinctrl/pinctrl-single.c:726: undefined reference to
`pinconf_generic_dump_config'

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-03-13 19:39:59 +01:00
Rafał Miłecki
4969b41798 b43: HT-PHY: enable basic TX power setup
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:53 -04:00
Rafał Miłecki
5949e04060 b43: HT-PHY: setup TX power control
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:52 -04:00
Rafał Miłecki
4409a23f6b b43: HT-PHY: implement RSSI polling
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:52 -04:00
Rafał Miłecki
396535e137 b43: HT-PHY: implement playing sample tone
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:51 -04:00
Rafał Miłecki
371ec465a3 b43: HT-PHY: implement stopping sample tone playback
It was another sequence I recognized in HT-PHY dump:
 phy_read(0x00c7) -> 0x0001
 phy_read(0x00c3) -> 0x0000
phy_write(0x00c3) <- 0x0002
 phy_read(0x00c3) -> 0x0000
phy_write(0x00c3) <- 0x0000
The difference to N-PHY is that it writes to 6 tables instead of a one
(after above).

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:50 -04:00
Rafał Miłecki
60e8fb9233 b43: HT-PHY: implement controlling TX power control
Don't enable it until we have (almost?) whole TX power management
figured out. It's similar to the N-PHY, the difference is that we call a
"fix" *before* disabling power control.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:50 -04:00
Rafał Miłecki
a51ab25811 b43: HT-PHY: implement PA override
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:49 -04:00
Rafał Miłecki
4cce095623 b43: HT-PHY: implement CCA reset
It was just another similar-to-N-PHY and easy-to-track routine:
write32 0xb0601408 <- 0x00002057
 phy_read(0x0001) -> 0x0000
phy_write(0x0001) <- 0x4000
phy_write(0x0001) <- 0x0000
write32 0xb0601408 <- 0x00002055
(b43_phy_ht_force_rf_sequence was moved up unmodified)

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:49 -04:00
Rafał Miłecki
d7bb7ca8e5 b43: HT-PHY: implement MAC reclocking
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:48 -04:00
Rafał Miłecki
2dacfe7c10 b43: HT-PHY: implement spurious tone avoidance
On N-PHY it's also done after TX power fix, so it was easy to spot.
Unfortunately the MMIO logs I have from ndsiwrapper include channels
1-12 only, so enabling code for 13 and 14 is just a N-PHY-based guess.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:47 -04:00
Rafał Miłecki
f6099f89e6 b43: HT-PHY: move TX fix to the separated function
On N-PHY after B43_PHY_B_TEST operation there is a call to TX power fix
function which iterates over available cores. It matches our HT-PHY code
which means it's probably also some TX fix.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:47 -04:00
Rafał Miłecki
b372afaed5 b43: HT-PHY: add classifier control function
After comparing operations on reg 0xB on N and HT it seems to be the
same register with similar ops. Implement them for HT-PHY.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:46 -04:00
Rafał Miłecki
476069224d b43: HT-PHY: rename AFE defines
It you take a look at N-PHY analog switch function it touches every core
on the chipset. It seems HT-PHY does they same, it just has 3 cores
instead of 2 (which make sense since BCM4331 is 3x3). Rename AFE defines
to include core id.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:45 -04:00
Vladimir Kondratiev
1a2780e0f3 wil6210: temperature measurement
Firmware got support for temperature measurement.
There are 2 temperature sensors: MAC and radio

"not available" temperature - reported by FW as 0 or ~0

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:27:35 -04:00
Vladimir Kondratiev
b80231773a wil6210: sync with new firmware
Adjust driver for changes in the FW API.
Noticeable changes in the FW are:
- temperature sensing
- infrastructure for multiple connections
- infrastructure for P2P
- signal strength indication

This commit introduces only changes that are required to support same functionality
as previous firmware, no new features.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:26:21 -04:00
Vladimir Kondratiev
55f7acdd24 wil6210: new SW reset
New firmware allows for shorter SW reset procedure.
After SW reset, FW raises "fw done" IRQ, at this
moment mailbox control structures are initialized, driver caches it.

New status bit wil_status_reset_done introduced to track completion
of the reset. It is set by "fw ready" irq, and required for WMI rx flow
to access control structures.

WMI Tx flow protected by other status bit, wil_status_fwready. It can't
be set before wil_status_reset_done is set by design.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:26:21 -04:00
Vladimir Kondratiev
c7996ef852 wil6210: headers clean-up
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:26:21 -04:00
Vladimir Kondratiev
acc9780d6e wil6210: use WLAN_CAPABILITY_DMG_TYPE_MASK
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:05 -04:00
Vladimir Kondratiev
de70ab87b1 wil6210: fix FW error notification
user space get notified through kobject_uevent_env(), that might sleep and thus
should run in thread context.

Move user space notification to the thread handler, while mark FW is non-functional
right in the hard IRQ.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:04 -04:00
Vladimir Kondratiev
102b1d99e5 wil6210: report all received mgmt frames
Pass to cfg80211 all management frames.
Used by wpa_supplicant.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:04 -04:00
Vladimir Kondratiev
a0f7845b7e wil6210: use cfg80211_inform_bss_frame()
Avoid unnecessary frame parsing

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:04 -04:00
Vladimir Kondratiev
d81079f170 wil6210: refactor connect_worker
Move wmi_connect_worker() to the main.c and change names for consistency.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:04 -04:00
Vladimir Kondratiev
03866e7d3f wil6210: Fix garbage sent to the FW with wmi_set_ie()
Extra reference was taken by mistake.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:03 -04:00
Vladimir Kondratiev
b1defa4d66 wil6210: do not set IE's for beacon
On the DMG band, there is no 'normal' beacon frame.
Instead, transmitted is short 'DMG beacon' frame, that do not include IE's
So, beacon IE's are not relevant for the DMG band.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:03 -04:00
Vladimir Kondratiev
249a382b8a wil6210: handle WMI_BA_STATUS_EVENTID
Firmware indicated block ack agreement status change.
For now, just log it.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:03 -04:00
Vladimir Kondratiev
3442a5048a wil6210: handle linkup/linkdown WMI events
Firmware indicates linkup/linkdown when data path becomes ready.

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:03 -04:00
Vladimir Kondratiev
3b0378a88b wil6210: Remove local implementation of dynamic hexdump
This functionality now integrated in kernel, local hack not needed any more

Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:02 -04:00
Wei Yongjun
d95f1d20ab wil6210: remove unused including <linux/version.h>
Remove including <linux/version.h> that don't need it.

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Acked-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:02 -04:00
Larry Finger
a9fac7399b ssb: pci: Fix flipping of MAC address
Since commit e565275 entitled "ssb: pci: Standardize a function to get mac
address", the SPROM readout of the MAC has had the values flipped so that
00:11:22:33:44:55 became 11:00:33:22:55:44. The fix has been tested on both
little- and big-endian architectures.

Reported-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:23:02 -04:00
Larry Finger
9437a248e7 rtlwifi: rtl8192cu: Fix problem that prevents reassociation
The driver was failing to clear the BSSID when a disconnect happened. That
prevented a reconnection. This problem is reported at
https://bugzilla.redhat.com/show_bug.cgi?id=789605,
https://bugzilla.redhat.com/show_bug.cgi?id=866786,
https://bugzilla.redhat.com/show_bug.cgi?id=906734, and
https://bugzilla.kernel.org/show_bug.cgi?id=46171.

Thanks to Jussi Kivilinna for making the critical observation
that led to the solution.

Reported-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
Tested-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
Tested-by: Alessandro Lannocca <alessandro.lannocca@gmail.com>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:18:53 -04:00
John Crispin
5818a46a99 rt2x00: fix rt2x00 to work with the new ralink SoC config symbols
Since v3.9-rc1 the kernel has basic support for Ralink WiSoC. The config symbols
are named slightly different than before. Fix the rt2x00 to match the new
symbols.

The commit causing this breakage is:
commit ae2b5bb657
Author: John Crispin <blogic@openwrt.org>
Date:   Sun Jan 20 22:05:30 2013 +0100
MIPS: ralink: adds Kbuild files

Signed-off-by: John Crispin <blogic@openwrt.org>
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-03-13 14:18:53 -04:00
Laurent Pinchart
53ded8191e pinctrl: Print the correct information in debugfs pinconf-state file
A bad copy&paste resulted in the debugfs pinconf-state file printing the
pin name instead of the state name. Fix it.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-03-13 19:12:07 +01:00
Axel Lin
6a40cdd544 pinctrl: abx500: Fix checking if pin use AlternateFunction register
It's pointless to check "af.alt_bit1 == UNUSED" twice.
This looks like a copy-paste bug, I think what we want is to check if *both*
af.alt_bit1 and af.alt_bit2 are UNUSED.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
2013-03-13 19:11:56 +01:00
John W. Linville
d15591b008 Merge branch 'for-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth 2013-03-13 13:29:56 -04:00
Frederic Weisbecker
d9a3c9823a sched: Lower chances of cputime scaling overflow
Some users have reported that after running a process with
hundreds of threads on intensive CPU-bound loads, the cputime
of the group started to freeze after a few days.

This is due to how we scale the tick-based cputime against
the scheduler precise execution time value.

We add the values of all threads in the group and we multiply
that against the sum of the scheduler exec runtime of the whole
group.

This easily overflows after a few days/weeks of execution.

A proposed solution to solve this was to compute that multiplication
on stime instead of utime:
   62188451f0
   ("cputime: Avoid multiplication overflow on utime scaling")

The rationale behind that was that it's easy for a thread to
spend most of its time in userspace under intensive CPU-bound workload
but it's much harder to do CPU-bound intensive long run in the kernel.

This postulate got defeated when a user recently reported he was still
seeing cputime freezes after the above patch. The workload that
triggers this issue relates to intensive networking workloads where
most of the cputime is consumed in the kernel.

To reduce much more the opportunities for multiplication overflow,
lets reduce the multiplication factors to the remainders of the division
between sched exec runtime and cputime. Assuming the difference between
these shouldn't ever be that large, it could work on many situations.

This gets the same results as in the upstream scaling code except for
a small difference: the upstream code always rounds the results to
the nearest integer not greater to what would be the precise result.
The new code rounds to the nearest integer either greater or not
greater. In practice this difference probably shouldn't matter but
it's worth mentioning.

If this solution appears not to be enough in the end, we'll
need to partly revert back to the behaviour prior to commit
     0cf55e1ec0
     ("sched, cputime: Introduce thread_group_times()")

Back then, the scaling was done on exit() time before adding the cputime
of an exiting thread to the signal struct. And then we'll need to
scale one-by-one the live threads cputime in thread_group_cputime(). The
drawback may be a slightly slower code on exit time.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
2013-03-13 18:18:14 +01:00
Takashi Iwai
ba615b86d6 ALSA: hda - Don't apply EAPD power filter as default
So far, the driver doesn't power down the widget at going down to D3
when the widget node has an EAPD capability and EAPD is actually set
on all codecs unless codec->power_filter is set explicitly.
This caused a problem on some Conexant codecs, leading to click
noises, and we set it as NULL there.  But it is very unlikely that the
problem hits only these codecs.

Looking back at the development history, this workaround for EAPD was
introduced just for some laptops with STAC9200 codec, then we applied
it blindly for all.  Now, since it's revealed to have an ill effect,
we should disable this workaround per default and apply only for the
known requiring systems.

The EAPD workaround is implemented now as snd_hda_codec_eapd_power_filter(),
and this has to be set explicitly by the codec driver when needed.
As of now, only patch_stac9200() sets this one.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
2013-03-13 18:07:05 +01:00
Takashi Iwai
bce0d2a80e ALSA: hda - Allow unlimited pins and converters in patch_hdmi.c
Use the dynamic array allocations for pins, converters and PCM arrays
instead of the fixed size arrays.  The modern HDMI codecs get more and
more pins, and we don't know the sensitive limit.

Most of the patch are spent for the straight conversions from the
fixed array access to snd_array helpers.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
2013-03-13 18:07:04 +01:00
Takashi Iwai
5265fd9a9f ALSA: hda - Drop explicit memset() by reallocation with __GFP_ZERO
Signed-off-by: Takashi Iwai <tiwai@suse.de>
2013-03-13 18:06:59 +01:00
Frederic Weisbecker
f792685006 math64: New div64_u64_rem helper
Provide an extended version of div64_u64() that
also returns the remainder of the division.

We are going to need this to refine the cputime
scaling code.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
2013-03-13 18:03:27 +01:00
Jean-Christophe PLAGNIOL-VILLARD
9317960fa3 ARM: at91: renamme rm9200 dt file
Rename the board-rm9200-dt.c file so that we follow the pattern for
Device Tree board files: board-dt-<name of SoC sharing same core>.c

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
[nicolas.ferre@atmel.com: modify commit message]
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
2013-03-13 17:28:14 +01:00
Jean-Christophe PLAGNIOL-VILLARD
4afcd1db82 ARM: at91: rename board-dt to more specific name board-dt-sam9
We will produce a board-dt file per SoC core type. That will ease code
readability and will prevent from including superfluous code for supporting
machines that will never be compiled together (particularly the ARM9 and C-A5
upcoming SoCs).

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
[nicolas.ferre@atmel.com: modify commit message]
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
2013-03-13 17:18:00 +01:00
Jean-Christophe PLAGNIOL-VILLARD
3ae8dbdcfa ARM: at91: move non DT Kconfig to Kconfig.non_dt
This is the legacy platform support

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
2013-03-13 17:16:15 +01:00
Dan Williams
3f8bc5e4da qcserial: bind to DM/DIAG port on Gobi 1K devices
Turns out we just need altsetting 1 and then we can talk to it.

Signed-off-by: Dan Williams <dcbw@redhat.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-13 08:50:42 -07:00
Michal Hocko
be871b7e54 device: separate all subsys mutexes
ca22e56d (driver-core: implement 'sysdev' functionality for regular
devices and buses) has introduced bus_register macro with a static
key to distinguish different subsys mutex classes.

This however doesn't work for different subsys which use a common
registering function. One example is subsys_system_register (and
mce_device and cpu_device).

In the end this leads to the following lockdep splat:
[  207.271924] ======================================================
[  207.271932] [ INFO: possible circular locking dependency detected ]
[  207.271942] 3.9.0-rc1-0.7-default+ #34 Not tainted
[  207.271948] -------------------------------------------------------
[  207.271957] bash/10493 is trying to acquire lock:
[  207.271963]  (subsys mutex){+.+.+.}, at: [<ffffffff8134af27>] bus_remove_device+0x37/0x1c0
[  207.271987]
[  207.271987] but task is already holding lock:
[  207.271995]  (cpu_hotplug.lock){+.+.+.}, at: [<ffffffff81046ccf>] cpu_hotplug_begin+0x2f/0x60
[  207.272012]
[  207.272012] which lock already depends on the new lock.
[  207.272012]
[  207.272023]
[  207.272023] the existing dependency chain (in reverse order) is:
[  207.272033]
[  207.272033] -> #4 (cpu_hotplug.lock){+.+.+.}:
[  207.272044]        [<ffffffff810ae329>] lock_acquire+0xe9/0x120
[  207.272056]        [<ffffffff814ad807>] mutex_lock_nested+0x37/0x360
[  207.272069]        [<ffffffff81046ba9>] get_online_cpus+0x29/0x40
[  207.272082]        [<ffffffff81185210>] drain_all_stock+0x30/0x150
[  207.272094]        [<ffffffff811853da>] mem_cgroup_reclaim+0xaa/0xe0
[  207.272104]        [<ffffffff8118775e>] __mem_cgroup_try_charge+0x51e/0xcf0
[  207.272114]        [<ffffffff81188486>] mem_cgroup_charge_common+0x36/0x60
[  207.272125]        [<ffffffff811884da>] mem_cgroup_newpage_charge+0x2a/0x30
[  207.272135]        [<ffffffff81150531>] do_wp_page+0x231/0x830
[  207.272147]        [<ffffffff8115151e>] handle_pte_fault+0x19e/0x8d0
[  207.272157]        [<ffffffff81151da8>] handle_mm_fault+0x158/0x1e0
[  207.272166]        [<ffffffff814b6153>] do_page_fault+0x2a3/0x4e0
[  207.272178]        [<ffffffff814b2578>] page_fault+0x28/0x30
[  207.272189]
[  207.272189] -> #3 (&mm->mmap_sem){++++++}:
[  207.272199]        [<ffffffff810ae329>] lock_acquire+0xe9/0x120
[  207.272208]        [<ffffffff8114c5ad>] might_fault+0x6d/0x90
[  207.272218]        [<ffffffff811a11e3>] filldir64+0xb3/0x120
[  207.272229]        [<ffffffffa013fc19>] call_filldir+0x89/0x130 [ext3]
[  207.272248]        [<ffffffffa0140377>] ext3_readdir+0x6b7/0x7e0 [ext3]
[  207.272263]        [<ffffffff811a1519>] vfs_readdir+0xa9/0xc0
[  207.272273]        [<ffffffff811a15cb>] sys_getdents64+0x9b/0x110
[  207.272284]        [<ffffffff814bb599>] system_call_fastpath+0x16/0x1b
[  207.272296]
[  207.272296] -> #2 (&type->i_mutex_dir_key#3){+.+.+.}:
[  207.272309]        [<ffffffff810ae329>] lock_acquire+0xe9/0x120
[  207.272319]        [<ffffffff814ad807>] mutex_lock_nested+0x37/0x360
[  207.272329]        [<ffffffff8119c254>] link_path_walk+0x6f4/0x9a0
[  207.272339]        [<ffffffff8119e7fa>] path_openat+0xba/0x470
[  207.272349]        [<ffffffff8119ecf8>] do_filp_open+0x48/0xa0
[  207.272358]        [<ffffffff8118d81c>] file_open_name+0xdc/0x110
[  207.272369]        [<ffffffff8118d885>] filp_open+0x35/0x40
[  207.272378]        [<ffffffff8135c76e>] _request_firmware+0x52e/0xb20
[  207.272389]        [<ffffffff8135cdd6>] request_firmware+0x16/0x20
[  207.272399]        [<ffffffffa03bdb91>] request_microcode_fw+0x61/0xd0 [microcode]
[  207.272416]        [<ffffffffa03bd554>] microcode_init_cpu+0x104/0x150 [microcode]
[  207.272431]        [<ffffffffa03bd61c>] mc_device_add+0x7c/0xb0 [microcode]
[  207.272444]        [<ffffffff8134a419>] subsys_interface_register+0xc9/0x100
[  207.272457]        [<ffffffffa04fc0f4>] 0xffffffffa04fc0f4
[  207.272472]        [<ffffffff81000202>] do_one_initcall+0x42/0x180
[  207.272485]        [<ffffffff810bbeff>] load_module+0x19df/0x1b70
[  207.272499]        [<ffffffff810bc376>] sys_init_module+0xe6/0x130
[  207.272511]        [<ffffffff814bb599>] system_call_fastpath+0x16/0x1b
[  207.272523]
[  207.272523] -> #1 (umhelper_sem){++++.+}:
[  207.272537]        [<ffffffff810ae329>] lock_acquire+0xe9/0x120
[  207.272548]        [<ffffffff814ae9c4>] down_read+0x34/0x50
[  207.272559]        [<ffffffff81062bff>] usermodehelper_read_trylock+0x4f/0x100
[  207.272575]        [<ffffffff8135c7dd>] _request_firmware+0x59d/0xb20
[  207.272587]        [<ffffffff8135cdd6>] request_firmware+0x16/0x20
[  207.272599]        [<ffffffffa03bdb91>] request_microcode_fw+0x61/0xd0 [microcode]
[  207.272613]        [<ffffffffa03bd554>] microcode_init_cpu+0x104/0x150 [microcode]
[  207.272627]        [<ffffffffa03bd61c>] mc_device_add+0x7c/0xb0 [microcode]
[  207.272641]        [<ffffffff8134a419>] subsys_interface_register+0xc9/0x100
[  207.272654]        [<ffffffffa04fc0f4>] 0xffffffffa04fc0f4
[  207.272666]        [<ffffffff81000202>] do_one_initcall+0x42/0x180
[  207.272678]        [<ffffffff810bbeff>] load_module+0x19df/0x1b70
[  207.272690]        [<ffffffff810bc376>] sys_init_module+0xe6/0x130
[  207.272702]        [<ffffffff814bb599>] system_call_fastpath+0x16/0x1b
[  207.272715]
[  207.272715] -> #0 (subsys mutex){+.+.+.}:
[  207.272729]        [<ffffffff810ae002>] __lock_acquire+0x13b2/0x15f0
[  207.272740]        [<ffffffff810ae329>] lock_acquire+0xe9/0x120
[  207.272751]        [<ffffffff814ad807>] mutex_lock_nested+0x37/0x360
[  207.272763]        [<ffffffff8134af27>] bus_remove_device+0x37/0x1c0
[  207.272775]        [<ffffffff81349114>] device_del+0x134/0x1f0
[  207.272786]        [<ffffffff813491f2>] device_unregister+0x22/0x60
[  207.272798]        [<ffffffff814a24ea>] mce_cpu_callback+0x15e/0x1ad
[  207.272812]        [<ffffffff814b6402>] notifier_call_chain+0x72/0x130
[  207.272824]        [<ffffffff81073d6e>] __raw_notifier_call_chain+0xe/0x10
[  207.272839]        [<ffffffff81498f76>] _cpu_down+0x1d6/0x350
[  207.272851]        [<ffffffff81499130>] cpu_down+0x40/0x60
[  207.272862]        [<ffffffff8149cc55>] store_online+0x75/0xe0
[  207.272874]        [<ffffffff813474a0>] dev_attr_store+0x20/0x30
[  207.272886]        [<ffffffff812090d9>] sysfs_write_file+0xd9/0x150
[  207.272900]        [<ffffffff8118e10b>] vfs_write+0xcb/0x130
[  207.272911]        [<ffffffff8118e924>] sys_write+0x64/0xa0
[  207.272923]        [<ffffffff814bb599>] system_call_fastpath+0x16/0x1b
[  207.272936]
[  207.272936] other info that might help us debug this:
[  207.272936]
[  207.272952] Chain exists of:
[  207.272952]   subsys mutex --> &mm->mmap_sem --> cpu_hotplug.lock
[  207.272952]
[  207.272973]  Possible unsafe locking scenario:
[  207.272973]
[  207.272984]        CPU0                    CPU1
[  207.272992]        ----                    ----
[  207.273000]   lock(cpu_hotplug.lock);
[  207.273009]                                lock(&mm->mmap_sem);
[  207.273020]                                lock(cpu_hotplug.lock);
[  207.273031]   lock(subsys mutex);
[  207.273040]
[  207.273040]  *** DEADLOCK ***
[  207.273040]
[  207.273055] 5 locks held by bash/10493:
[  207.273062]  #0:  (&buffer->mutex){+.+.+.}, at: [<ffffffff81209049>] sysfs_write_file+0x49/0x150
[  207.273080]  #1:  (s_active#150){.+.+.+}, at: [<ffffffff812090c2>] sysfs_write_file+0xc2/0x150
[  207.273099]  #2:  (x86_cpu_hotplug_driver_mutex){+.+.+.}, at: [<ffffffff81027557>] cpu_hotplug_driver_lock+0x17/0x20
[  207.273121]  #3:  (cpu_add_remove_lock){+.+.+.}, at: [<ffffffff8149911c>] cpu_down+0x2c/0x60
[  207.273140]  #4:  (cpu_hotplug.lock){+.+.+.}, at: [<ffffffff81046ccf>] cpu_hotplug_begin+0x2f/0x60
[  207.273158]
[  207.273158] stack backtrace:
[  207.273170] Pid: 10493, comm: bash Not tainted 3.9.0-rc1-0.7-default+ #34
[  207.273180] Call Trace:
[  207.273192]  [<ffffffff810ab373>] print_circular_bug+0x223/0x310
[  207.273204]  [<ffffffff810ae002>] __lock_acquire+0x13b2/0x15f0
[  207.273216]  [<ffffffff812086b0>] ? sysfs_hash_and_remove+0x60/0xc0
[  207.273227]  [<ffffffff810ae329>] lock_acquire+0xe9/0x120
[  207.273239]  [<ffffffff8134af27>] ? bus_remove_device+0x37/0x1c0
[  207.273251]  [<ffffffff814ad807>] mutex_lock_nested+0x37/0x360
[  207.273263]  [<ffffffff8134af27>] ? bus_remove_device+0x37/0x1c0
[  207.273274]  [<ffffffff812086b0>] ? sysfs_hash_and_remove+0x60/0xc0
[  207.273286]  [<ffffffff8134af27>] bus_remove_device+0x37/0x1c0
[  207.273298]  [<ffffffff81349114>] device_del+0x134/0x1f0
[  207.273309]  [<ffffffff813491f2>] device_unregister+0x22/0x60
[  207.273321]  [<ffffffff814a24ea>] mce_cpu_callback+0x15e/0x1ad
[  207.273332]  [<ffffffff814b6402>] notifier_call_chain+0x72/0x130
[  207.273344]  [<ffffffff81073d6e>] __raw_notifier_call_chain+0xe/0x10
[  207.273356]  [<ffffffff81498f76>] _cpu_down+0x1d6/0x350
[  207.273368]  [<ffffffff81027557>] ? cpu_hotplug_driver_lock+0x17/0x20
[  207.273380]  [<ffffffff81499130>] cpu_down+0x40/0x60
[  207.273391]  [<ffffffff8149cc55>] store_online+0x75/0xe0
[  207.273402]  [<ffffffff813474a0>] dev_attr_store+0x20/0x30
[  207.273413]  [<ffffffff812090d9>] sysfs_write_file+0xd9/0x150
[  207.273425]  [<ffffffff8118e10b>] vfs_write+0xcb/0x130
[  207.273436]  [<ffffffff8118e924>] sys_write+0x64/0xa0
[  207.273447]  [<ffffffff814bb599>] system_call_fastpath+0x16/0x1b

Which reports a false possitive deadlock because it sees:
1) load_module -> subsys_interface_register -> mc_deveice_add (*) -> subsys->p->mutex -> link_path_walk -> lookup_slow -> i_mutex
2) sys_write -> _cpu_down -> cpu_hotplug_begin -> cpu_hotplug.lock -> mce_cpu_callback -> mce_device_remove(**) -> device_unregister -> bus_remove_device -> subsys mutex
3) vfs_readdir -> i_mutex -> filldir64 -> might_fault -> might_lock_read(mmap_sem) -> page_fault -> mmap_sem -> drain_all_stock -> cpu_hotplug.lock

but
1) takes cpu_subsys subsys (*) but 2) takes mce_device subsys (**) so
the deadlock is not possible AFAICS.

The fix is quite simple. We can pull the key inside bus_type structure
because they are defined per device so the pointer will be unique as
well. bus_register doesn't need to be a macro anymore so change it
to the inline. We could get rid of __bus_register as there is no other
caller but maybe somebody will want to use a different key so keep it
around for now.

Reported-by: Li Zefan <lizefan@huawei.com>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2013-03-13 08:48:28 -07:00
Bjørn Mork
b701f16dd4 net: qmi_wwan: set correct altsetting for Gobi 1K devices
commit bd877e4 ("net: qmi_wwan: use a single bind function for
all device types") made Gobi 1K devices fail probing.

Using the number of endpoints in the default altsetting to decide
whether the function use one or two interfaces is wrong.  Other
altsettings may provide more endpoints.

With Gobi 1K devices, USB interface #3's altsetting is 0 by default, but
altsetting 0 only provides one interrupt endpoint and is not sufficent
for QMI.  Altsetting 1 provides all 3 endpoints required for qmi_wwan
and works with QMI. Gobi 1K layout for intf#3 is:

    Interface Descriptor:  255/255/255
      bInterfaceNumber        3
      bAlternateSetting       0
      Endpoint Descriptor:  Interrupt IN
    Interface Descriptor:  255/255/255
      bInterfaceNumber        3
      bAlternateSetting       1
      Endpoint Descriptor:  Interrupt IN
      Endpoint Descriptor:  Bulk IN
      Endpoint Descriptor:  Bulk OUT

Prior to commit bd877e4, we would call usbnet_get_endpoints
before giving up finding enough endpoints. Removing the early
endpoint number test and the strict functional descriptor
requirement allow qmi_wwan_bind to continue until
usbnet_get_endpoints has made the final attempt to collect
endpoints.  This restores the behaviour from before commit
bd877e4 without losing the added benefit of using a single bind
function.

The driver has always required a CDC Union functional descriptor
for two-interface functions. Using the existence of this
descriptor to detect two-interface functions is the logically
correct method.

Reported-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Tested-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-03-13 11:33:42 -04:00