The 24HD has 1 button per LED (first three buttons of each group).
We need a special treatment for them as it's not a uniq button that
switches between the LEDs.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The code for 21UX2 and 24HD makes the LED group 1 on the left, and
the group 0 on the right. The buttons are ordered in the other way,
but libwacom already exports those that way. So we simply can't reassign
LED group 0 to the left buttons, and have to quirk the incoming data...
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The EKR switches the LED directly, and there is no point in having
userspace handling the switch it self when it's easy enough to do
in the kernel.

The other benefit is that now userspace does not need to have root access
to the LED but need only to read them with user privileges.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

There is a bug (?) in devm_led_classdev_register() in which its increments
the refcount of the parent. If the parent is an input device, that means
the ref count never reaches 0 when devm_input_device_release() gets called.
This means that the LEDs and all the devres resources attached to the
input device are not released.

Manually force the release of the group so that the leds are released once
we are done using them.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The now obsolete sysfs files for LEDs and EKRemote are kept for backward
compatibility.
Both the EKR (read-only) and the regular Cintiqs and Intuos are now
sharing the same led API.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Or Gnome complains about an empty battery.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Previously, all the remotes attached to the same receiver would share the
same power_supply. That's not good as the remotes will constantly change
the battery information according to their own state.

To have something generic enough, we introduce struct wacom_battery
which regroups all the information we need for a battery.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Thanks to devres, we can now afford to create more than one input node
without having to overload the remove/failure paths. Having one input
node per remote is something which should have been implemented from start
but the probability of having users with several remotes is quite low.
Anyway, still, better looking at the future and implement things properly.

Remote input nodes will be freed/unregistered magically as they are
created in the devres group &remote->remotes[index].

We need to open the hid node now that the remotes are dynamically
allocated.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

No functional changes, just a prep patch for the one after.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

This will be useful when each remote will be assigned its own input device.
We won't need to unregister each input and sysfs and other elements one
at a time.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The wacom_remote_create_attr_group() and wacom_remote_destroy_attr_group()
functions were both allocating/destroying the sysfs groups but also
initializing the parameters for the remotes. Have proper functions
that can be called and extended.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Thanks to devres management, we don't need to remember a lot of failure
path. One or two is enough.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

If we want to have one input device per remote, it's better to have our
own struct wacom_remote which is dynamically allocated.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

wacom_remote_status_irq() sends information of addition/removal of EKR.
We want to allocate one input node per remote, so better having this
in a separate worker, not handled in the IRQ directly.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

We need to add an action to ensure wacom->led.groups is null when
wacom_led_control() gets called after the resources has been freed.

This also prevents to send a LED command when there is no support
from the device.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

wacom_release_shared_data() and wacom_remove_shared_data() are moved up
so they can be referenced in wacom_add_shared_data().

There is no point in explicitly setting wacom_wac1->shared->type to 0 in
wacom_wireless_work() (plus this would give an oops).
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

We started switching the driver to devres, so we should use it as much
as possible.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The sysfs group was indeed removed by kobject_put(wacom->remote_dir) in
wacom_remove(), but the name of the group was never freed.

Also remove the misplaced kobject_put(wacom->remote_dir) in the error
path of wacom_remote_create_attr_group().
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Use our own wacom_devm_sysfs_create_group() as there is currently no
generic one. It has been requested at least twice [1][2] but has been
always rejected.
However, in the Wacom case, for the wirelessly connected devices, we need
to be able to release the created sysfs files without removing the parent
kobject.

[1] https://patchwork.kernel.org/patch/7526551/
[2] https://lkml.org/lkml/2013/3/14/728Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

We currently have a complex clean_inputs() function while this can be
handled all by devres. Set a group that we can destroy in wireless_work().
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Simplifying the error code paths.
We need to keep wacom_clean_inputs() around for now as the wireless
module is using it to dynamically remove the inputs on disconnect.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Simplifying the error code paths.
We need to keep wacom_destroy_battery() around for now as the wireless
module and the remotes are using it to dynamically remove the battery
supply on disconnect.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Looks like the battery hijacked the wireless worker. That's not fair so
use a work queue per task.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Like remotes, LEDs should be handled by themself, not magically behind
the inputs as they have a complete different life.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

wacom->remote_dir has nothing to do with inputs, so better not magically
removing it when cleaning inputs.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The type is never set but we check for it in wacom_wireless_irq().
It looks like this is a big hack from the beginning, so fill in the gap
only.

Untested.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Since fd5f92b6 ("HID: wacom: reuse wacom_parse_and_register() in
wireless_work"), wacom->shared->type is not set.
Send the information of the battery if we have one.
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: NPing Cheng <pingc@wacom.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Pull HID updates from Jiri Kosina:

 - new hid-alps driver for ALPS Touchpad-Stick device, from Masaki Ota

 - much improved and generalized HID led handling, and merge of
   specialized hid-thingm driver into this generic hid-led one, from
   Heiner Kallweit

 - i2c-hid power management improvements from Fu Zhonghui and Guohua
   Zhong

 - uhid initialization race fix from Roderick Colenbrander

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid: (21 commits)
  HID: add usb device id for Apple Magic Keyboard
  HID: hid-led: fix Delcom support on big endian systems
  HID: hid-led: add support for Greynut Luxafor
  HID: hid-led: add support for Delcom Visual Signal Indicator G2
  HID: hid-led: remove report id from struct hidled_config
  HID: alps: a few cleanups
  HID: remove ThingM blink(1) driver
  HID: hid-led: add support for ThingM blink(1)
  HID: hid-led: add support for reading from LED devices
  HID: hid-led: add support for devices with multiple independent LEDs
  HID: i2c-hid: set power sleep before shutdown
  HID: alps: match alps devices in core
  HID: thingm: simplify debug output code
  HID: alps: pass correct sizes to hid_hw_raw_request()
  HID: alps: struct u1_dev *priv is internal to the driver
  HID: add Alps I2C HID Touchpad-Stick support
  HID: led: fix config
  usb: misc: remove outdated USB LED driver
  HID: migrate USB LED driver from usb misc to hid
  HID: i2c_hid: enable i2c-hid devices to suspend/resume asynchronously
  ...

Pull trivial tree updates from Jiri Kosina.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial:
  fat: fix error message for bogus number of directory entries
  fat: fix typo s/supeblock/superblock/
  ASoC: max9877: Remove unused function declaration
  dw2102: don't output spurious blank lines to the kernel log
  init: fix Kconfig text
  ARM: io: fix comment grammar
  ocfs: fix ocfs2_xattr_user_get() argument name
  scsi/qla2xxx: Remove erroneous unused macro qla82xx_get_temp_val1()

Pull quota update from Jan Kara:
 "time64 support for quota"

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  quota: use time64_t internally

Pull random driver fix from Ted Ts'o:
 "Fix a boot failure on systems with non-contiguous NUMA id's"

* tag 'random_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/random:
  random: use for_each_online_node() to iterate over NUMA nodes

Pull vfs updates from Al Viro:
 "Assorted cleanups and fixes.

  Probably the most interesting part long-term is ->d_init() - that will
  have a bunch of followups in (at least) ceph and lustre, but we'll
  need to sort the barrier-related rules before it can get used for
  really non-trivial stuff.

  Another fun thing is the merge of ->d_iput() callers (dentry_iput()
  and dentry_unlink_inode()) and a bunch of ->d_compare() ones (all
  except the one in __d_lookup_lru())"

* 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (26 commits)
  fs/dcache.c: avoid soft-lockup in dput()
  vfs: new d_init method
  vfs: Update lookup_dcache() comment
  bdev: get rid of ->bd_inodes
  Remove last traces of ->sync_page
  new helper: d_same_name()
  dentry_cmp(): use lockless_dereference() instead of smp_read_barrier_depends()
  vfs: clean up documentation
  vfs: document ->d_real()
  vfs: merge .d_select_inode() into .d_real()
  unify dentry_iput() and dentry_unlink_inode()
  binfmt_misc: ->s_root is not going anywhere
  drop redundant ->owner initializations
  ufs: get rid of redundant checks
  orangefs: constify inode_operations
  missed comment updates from ->direct_IO() prototype change
  file_inode(f)->i_mapping is f->f_mapping
  trim fsnotify hooks a bit
  9p: new helper - v9fs_parent_fid()
  debugfs: ->d_parent is never NULL or negative
  ...

This changes the vfs dentry hashing to mix in the parent pointer at the
_beginning_ of the hash, rather than at the end.

That actually improves both the hash and the code generation, because we
can move more of the computation to the "static" part of the dcache
setup, and do less at lookup runtime.

It turns out that a lot of other hash users also really wanted to mix in
a base pointer as a 'salt' for the hash, and so the slightly extended
interface ends up working well for other cases too.

Users that want a string hash that is purely about the string pass in a
'salt' pointer of NULL.

* merge branch 'salted-string-hash':
  fs/dcache.c: Save one 32-bit multiply in dcache lookup
  vfs: make the string hashes salt the hash

Conflicts:
	drivers/hid/hid-thingm.c

Merge branches 'for-4.8/alps', 'for-4.8/apple', 'for-4.8/i2c-hid', 'for-4.8/uhid-offload-hid-device-add' and 'for-4.8/upstream' into for-linus

This fixes a crash on s390 with fake NUMA enabled.
Reported-by: NHeiko Carstens <heiko.carstens@de.ibm.com>
Fixes: 1e7f583a ("random: make /dev/urandom scalable for silly userspace programs")
Signed-off-by: NTheodore Ts'o <tytso@mit.edu>

Some of our "for_each_xyz()" macro constructs make gcc unhappy about
lack of braces around if-statements inside or outside the loop, because
the loop construct itself has a "if-then-else" statement inside of it.

The resulting warnings look something like this:

  drivers/gpu/drm/i915/i915_debugfs.c: In function ‘i915_dump_lrc’:
  drivers/gpu/drm/i915/i915_debugfs.c:2103:6: warning: suggest explicit braces to avoid ambiguous ‘else’ [-Wparentheses]
     if (ctx != dev_priv->kernel_context)
        ^

even if the code itself is fine.

Since the warning is fairly easy to avoid by adding a braces around the
if-statement near the for_each_xyz() construct, do so, rather than
disabling the otherwise potentially useful warning.

(The if-then-else statements used in the "for_each_xyz()" constructs are
designed to be inherently safe even with no braces, but in this case
it's quite understandable that gcc isn't really able to tell that).

This finally leaves the standard "allmodconfig" build with just a
handful of remaining warnings, so new and valid warnings hopefully will
stand out.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Newer versions of gcc warn about the use of __builtin_return_address()
with a non-zero argument when "-Wall" is specified:

  kernel/trace/trace_irqsoff.c: In function ‘stop_critical_timings’:
  kernel/trace/trace_irqsoff.c:433:86: warning: calling ‘__builtin_return_address’ with a nonzero argument is unsafe [-Wframe-address]
     stop_critical_timing(CALLER_ADDR0, CALLER_ADDR1);
  [ .. repeats a few times for other similar cases .. ]

It is true that a non-zero argument is somewhat dangerous, and we do not
actually have very many uses of that in the kernel - but the ftrace code
does use it, and as Stephen Rostedt says:

 "We are well aware of the danger of using __builtin_return_address() of
  > 0.  In fact that's part of the reason for having the "thunk" code in
  x86 (See arch/x86/entry/thunk_{64,32}.S).  [..] it adds extra frames
  when tracking irqs off sections, to prevent __builtin_return_address()
  from accessing bad areas.  In fact the thunk_32.S states: 'Trampoline to
  trace irqs off.  (otherwise CALLER_ADDR1 might crash)'."

For now, __builtin_return_address() with a non-zero argument is the best
we can do, and the warning is not helpful and can end up making people
miss other warnings for real problems.

So disable the frame-address warning on compilers that need it.
Acked-by: NSteven Rostedt <rostedt@goodmis.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Pull HSI updates from Sebastian Reichel:

 - proper runtime pm support for omap-ssi and ssi-protocol

 - misc fixes

* tag 'hsi-for-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-hsi: (24 commits)
  HSI: omap_ssi: drop pm_runtime_irq_safe
  HSI: omap_ssi_port: use rpm autosuspend API
  HSI: omap_ssi: call msg->complete() from process context
  HSI: omap_ssi_port: ensure clocks are kept enabled during transfer
  HSI: omap_ssi_port: replace pm_runtime_put_sync with non-sync variant
  HSI: omap_ssi_port: avoid calling runtime_pm_*_sync inside spinlock
  HSI: omap_ssi_port: avoid pm_runtime_get_sync in ssi_start_dma and ssi_start_pio
  HSI: omap_ssi_port: switch to threaded pio irq
  HSI: omap_ssi_core: remove pm_runtime_get_sync call from tasklet
  HSI: omap_ssi_core: use pm_runtime_put instead of pm_runtime_put_sync
  HSI: omap_ssi_port: prepare start_tx/stop_tx for blocking pm_runtime calls
  HSI: core: switch port event notifier from atomic to blocking
  HSI: omap_ssi_port: replace wkin_cken with atomic bitmap operations
  HSI: omap_ssi: convert cawake irq handler to thread
  HSI: ssi_protocol: fix ssip_xmit invocation
  HSI: ssi_protocol: replace spin_lock with spin_lock_bh
  HSI: ssi_protocol: avoid ssi_waketest call with held spinlock
  HSI: omap_ssi: do not reset module
  HSI: omap_ssi_port: remove useless newline
  hsi: Only descend into hsi directory when CONFIG_HSI is set
  ...

Pull random driver updates from Ted Ts'o:
 "A number of improvements for the /dev/random driver; the most
  important is the use of a ChaCha20-based CRNG for /dev/urandom, which
  is faster, more efficient, and easier to make scalable for
  silly/abusive userspace programs that want to read from /dev/urandom
  in a tight loop on NUMA systems.

  This set of patches also improves entropy gathering on VM's running on
  Microsoft Azure, and will take advantage of a hw random number
  generator (if present) to initialize the /dev/urandom pool"

(It turns out that the random tree hadn't been in linux-next this time
around, because it had been dropped earlier as being too quiet.  Oh
well).

* tag 'random_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/random:
  random: strengthen input validation for RNDADDTOENTCNT
  random: add backtracking protection to the CRNG
  random: make /dev/urandom scalable for silly userspace programs
  random: replace non-blocking pool with a Chacha20-based CRNG
  random: properly align get_random_int_hash
  random: add interrupt callback to VMBus IRQ handler
  random: print a warning for the first ten uninitialized random users
  random: initialize the non-blocking pool via add_hwgenerator_randomness()