staging: rtl8723au: Remove unused pointer in rtw_wdev_free()

Pointer 'pwdev_priv' in function rtw_wdev_free() is unused - thus remove it.

Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: fix some confusing indenting

The break and the "unknown++" are at the same indent level so my static
checker complains.  The if statement should be indented more.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused struct odm_ra_info

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Since struct odm_sta_info isn't used, remove it completely

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused variable IOTPeer

For bizarre reason we have unused variables IOTPeer both in struct
sta_info and struct odm_sta_info.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: struct odm_mac_info isn't used

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: struct phy_info and struct odm_phy_info are identical

We don't need two copies of the same struct, it just leads to
pointless typecasts.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove empty *_free_xmit_priv() and friends

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Eliminate no-op function rtl8723au_init_recvbuf()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove another unused queue from struct recv_priv

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove some unused lists from struct sta_priv

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_pwrctrl.c: Reduce spaghettification

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: pattrib->priority can't be > 15, no need for CHKBIT()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: qos_asm23a(): Don't pass through an invalid priority

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: set_qos() keep userpriority in a u8

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Another case of missing 'tid' bounds checking.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: 'keyid' is always 0-3, no need for CHKBIT()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: No need for CHKBIT since 16 + 12 is still < 32

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: tid is 4 bits, so we need a 16 bit value to hold a bitmap

Remove unused add_enable_bitmap and candidate_tid_bitmap from struct
mlme_ext_info.

In addition we don't need to use CHKBIT when we know we are dealing
with a value in the range 0-15.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove local #define BIT()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Call c2h_handler_8723a() directly instead of via HAL interface

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: The driver only has one c2h_id_filter

Since we only have one function for this, there is no point in keeping
the HAL silliness.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: c2h_evt_hdl() is always called with 3rd argument = NULL

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Declare c2h_evt_hdl static

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Move rtw_do_join23a() prototype to rtw_ioctl_set.h

rtw_do_join23a() is declared in rtw_ioctl_set.c so move the prototype
to it's correct place, and remove duplicate protoype from rtw_mlme.c

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Move {WPA,RSN}_TKIP_CIPHER23A to rtw_wlan_util.c and mark them static

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Move REALTEK96B_IE23A to rtw_mlme_ext.c and mark it static

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused ADPT_{FMT,AGR} macros

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Get rid of FUNC_NDEV_{FMT,ARG}

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove FUNC_ADPT_{FMT,ARG}

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Get rid of NDEV_FMT/NDEV_ARG macros

Silly use-once macros that simply obfuscates the code.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove some unused prototypes

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Eliminate RTW_STATUS_CODE23a()

Inline the parsing of status codes. Long term we should pass back the
proper status codes and get rid of the ugly _FAIL/_SUCCESS mess. This
is one small step in that direction.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Eliminate _rtw_queue_empty23a()

This was just a silly wrapper around list_empty()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_free_recvframe23a() is always passed precvpriv->free_recv_queue

No point in passing in the pointer to free_recv_queue, when we always
pass in the same value.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused semaphore 'allrxreturnevt'

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove ugly vmalloc() wrappers

Finally with the last user converted, get rid of ugly vmalloc wrappers

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Use kmalloc instead of vmalloc to allocate recv_frame pool

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl723au: Remove some obsolete misleading comments

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

stating: rtl8723au: Use kmalloc instead of vmalloc to allocate pool of xmit_frame_ext

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Use kmalloc() rather than vmalloc() to allocate xmit_frames

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Declare rtw_alloc_xmitframe() static

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: fix potential leak in update_bcn_wps_ie()

Fix a potential leak in the error path of function update_bcn_wps_ie().
Move the affected input verification to the beginning of the function so
that it may return directly without leaking already allocated memory.
Detected by Coverity - CID 1077718.

Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Add GFP argument to rtw_alloc_stainfo23a()

No need to allocate GFP_ATOMIC when we don't need to.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Use kernel memory allocator for sta_priv allocations

The kernel has a great memory allocator, use it instead of hacking up
something in the driver.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Eliminate stainfo_by_offset spaghetti

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: add GFP argument to rtw_alloc_network()

This allows us to use GFP_KERNEL when calling from a work handler.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_createbss_cmd23a_callback(): Relax holding of pmlmepriv->lock

We only really need to hold pmlmepriv->lock while calling
rtw_indicate_connect23a(). rtw_get_stainfo23a() and
rtw_alloc_stainfo23a() rely on pstapriv->sta_hash_lock and the non
WIFI_AP_STATE path relies on pmlmepvi->scanned_queue.lock, except we
need to used clr_fwstate() instead of _clr_fwstate_().

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_alloc_stainfo23a(): Remove unused variable tmp_aid

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_sta_mgt.c: Fix some formatting badness

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Get rid of second set of custom error codes

One set of custom error codes for the driver should more than
suffice. This allows us to get rid of odm_types.h as well.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove some unused clutter from odm_types.h

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused struct rtl8723a_priv

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove a bunch of unused clutter from the ODM code

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused ODM spin lock functions

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused ODM malloc/free functions

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Use random ethaddr if EEPROM address is corrupted

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: wifi.h: remove more unused #defines

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Use struct ieee80211_pspoll to obtain 'aid'

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Fold get_hdr_bssid() into update_recvframe_phyinfo()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: update_recvframe_phyinfo(): Don't set variables twice

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: validate_recv_data_frame() use fctl knowledge to obtain bssid

Use the knowledge we already have from parsing the TODS/FROMDS bits in
hdr->frame_control to obtain the bssid.

Note that get_hdr_bssid() would never return NULL as handling 4
combinations of a 2 bit word leaves little space for falling through
to the 'default' value.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove to_fr_ds packet attribute

This is only used in one place - do the work there properly.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Declare validate_recv_*_frame() static

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove redundant check

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_wlan_util.c: Clean up the code

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtl8723a_recv.h: Don't put parenthesis around constants

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove a bunch of always true #ifdefs

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove a pile of unused firmware image file names

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_ioctl_set.h: Remove unused struct bssid_info

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused rtw_ioctl.h

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Eliminate struct qos_priv

Having struct qos_priv containing a single integer, and carrying a
dedicated header file for it, is just plain silly. Move the integer
into struct mlme_priv, which is the only place qos_priv was used
anyway, and get rid of the header file.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_event.h remove some unused structs and random blank lines

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Make event callbacks take a const u8 *

This avoids an ugly cast in mlme_evt_hdl23a()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Fix endian abnormality in mlme_evt_hdl23a()

Every other place uses C2HEvent_Header() for events. Given the struct
is endian dependant, use it here too to retrieve data from the parm
buffer.

Note the length field is not set/read in le order - question is
whether it's simply an opaque field?

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Move data byte-swap closer to register access

This makes sparse happy

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused + write-only entries from struct evt_priv

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Get rid of ugly cbuf interface

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Use workqueue to handle interrupt complete processing

Split the old work handler into a workqueue for processing
usb_read_interrupt_complete() events that require more than ust
clearing the event, and leave the old handler to just handle event
clearing.

This means we can get rid of the hacks with magic pointers to
determine what actions needs to be taken in the work handler, and
as an extra bonus obsoletes the bizarre cbuf code.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Name cmd workqueue appropriately to allow for more workqueues

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove rtw_free_cmd_priv23a()

This function no longer does anything, so get rid of it.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rsp_buf wasn't used for anything - don't allocate it

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: pass the cmd parm buffer directly to the cmd handler

commit 0348dc74f6689825c16db40fbe0ce6ad2da5bab9 ensured that the parm
buffer passed to the cmd handlers is not being over-written. Hence
there is no need to make a copy of the parm buffer just to pass it
into the cmd handler.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_ieee80211.c: Reduce some excessive parenthesis usage

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_ieee80211.c mark a couple of function static

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge branch 'imx-drm-fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm into staging-next

imx-drm fixes from Russell

Merge tag 'iio-for-3.16b' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-next

Jonathan writes:

Second set of IIO new drivers, cleanups and functionality for the 3.16 cycle.

This set contains a change to the ABI for the hid-sensors drivers to bring them
in line with the long published documentation.  Unfortunately, rather than
reporting true scale and offset values via sysfs they were reporting
some magic numbers that could only be converted to anything useful using
the HID sensors specification. I missed this entirely through the introduction
of a number of drivers, only picking up on it recently.  Srinivas has had
user feedback about this as well. The patch set is too large to go as a fix
at this stage in the cycle and is not a regression fix as this was never
right and so will have to wait for the next merge window. Srinivas assures
me that there are relatively few pieces of hardware out there and he has
had a number of people contact him to point out that the drivers did not
obey the ABI.  Hence hopefully the fallout of this, if any will be minor.
If we don't fix it now, it will only get worse going forward. There is no
sensible way of maintaining the incorrect ABI as it is simply returning
the wrong values through the standard interfaces.

Non IIO elements
* Introduce devm_kmemdup.  Does what it says on the tin.

New drivers:
* hid-sensors rotation devices (output as quaternion)
* Freescale MPL115A2 presure and temperature sensor.
* Melexis mlx90614 contactless infrared sensor.
* Freescale MMA8452Q 3-axis accelerometer.

New functionality:
* Addition of multiple element callback to allow for sysfs interfaces to access
  elements such as quaternions which have no useful meaning if all 4 elements
  are not presented together.  Other future usecases for this include
  rotation matrices.
* Support for multiple element buffer entries for exactly the same uses as
  the sysfs related elements described above.
* Quaternion support via the quaternion IIO modifier.
* TEMP_AMBIENT and TEMP_OBJECT modifiers to distinguish cases with thermopile
  devices.
* hid-sensors gain sysfs access to the sensor readings. Previously these
  drivers used the buffered interface only.  This change involves some
  additional hid-sensors core support to read poll values back from the devices
  to allow the drivers to know roughly how long to wait for a result when
  polling the sensor.  There is also an associated hid-sensors abi to allow
  the devices to be turned off between reads and powered up on demand.

Cleanups and fixes
* Hid sensors fix as described above. Result is to make the _scale and _offset
  attributes applicable in the same way as for all other IIO drivers.
* Some additional documentation - mostly covering stuff that graduated from
  staging without managing to take it's ABI docs with it.
* A series of little tidy ups to the exynos_adc driver that make the code
  nicer to read and improve handling of some corner cases.
* A tidy up to mag3110 (logical fix rather than a real one ;). Also enable
  user offset calibration for this device.
* Drop some left over IS_ERR() checks from ad799x that snuck through during
  the cleanup in the last IIO patch set.
* Fix a naming issue from clashing patches in ak8975 - note the clash only
  occured in the last IIO patch set, hence the fix needs to go through this
  tree.
* A format string missmatch fix in ad7280.c. Unlikely to have ever had an
  impact so not worth rushing through.

iio: hid-sensors: Pressure: Raw read support

Added support for raw reading of channel. If the sensor is powered
off, it will turn on for reading value.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Inclinometer 3D: Raw read support

Added support for raw reading of channel. If the sensor is powered
off, it will turn on for reading value.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Compass 3D: Raw read support

Added support for raw reading of channel. If the sensor is powered
off, it will turn on for reading value

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Proximity: Raw read support

Added support for raw reading of channel. If the sensor is powered
off, it will turn on for reading value.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: ALS: Raw read support

Added support for raw reading of channel. If the sensor is powered
off, it will turn on for reading value.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Gyro 3D: Raw read support

Added support for raw reading of channel. If the sensor is powered
off, it will turn on for reading value.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Accelerometer 3D: Raw read support

Added support for raw reading of channel. If the sensor is powered
off, it will turn on for reading value.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Add API to power on/off

Added an API to allow client drivers to turn ON and OFF sensors for
quick read. Added data_read as counting varaible instead of boolean,
so that sensor is powered off only when last user released it.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Pressure: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Inclinometer 3D: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Compass 3D: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: ALS: adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

iio: hid-sensors: Gyro 3D : adjust scale and offset

Using units and unit exponent to calculate scale which is compliant
to IIO ABI.

Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>