staging: mt7621-eth: Add missing blank lines after declarations

Add missing blank lines after declarations to solve checkpatch.pl errors.

Signed-off-by: Kamal Heib <kamalheib1@gmail.com>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-eth: Remove unnecessary blank lines

Remove un-necessary blank lines to solve errors found by checkpatch.pl.

Signed-off-by: Kamal Heib <kamalheib1@gmail.com>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-eth: Fix compilation error

Fix the following comilation error by making sure that "phy_ring_head" is
defined as "dma_addr_t" and change "phy_ring_tail" to match it as both
should be "dma_addr_t".

error: passing argument 3 of ‘dma_alloc_coherent’ from incompatible pointer type

Signed-off-by: Kamal Heib <kamalheib1@gmail.com>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: o2iblnd: Enable Multiple OPA Endpoints between Nodes

OPA driver optimizations are based on the MPI model where it is
expected to have multiple endpoints between two given nodes. To
enable this optimization for Lustre, we need to make it possible,
via an LND-specific tuneable, to create multiple endpoints and to
balance the traffic over them.

Both sides of a connection must have this patch for it to work.
Only the active side of the connection (usually the client)
needs to have the new tuneable set > 1.

Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8943
Reviewed-on: https://review.whamcloud.com/25168
Reviewed-by: Amir Shehata <amir.shehata@intel.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Doug Oucharek <dougso@me.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: ks7010: Remove unnecessary limit checks

uwrq is an unsigned 32-bit integer, it cannot be less than zero.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: usb: remove local variable

This patch removes the local variable dev that is used to store the pointer
to the usb_device whenever it is used only once.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: usb: fix usb_disconnect race condition

The functions usb_disconnect and usb_sndbulkpipe are racing for the struct
usb_device, which might cause a null pointer dereference exception. This
patch fixes this race condition by protecting the critical section inside
the function hdm_enque with the io_mutex.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: usb: don't set URB_ZERO_PACKET flag for synchronous data

This patch avoids setting the URB_ZERO_PACKET transfer flag for synchronous
data. This is needed to prevent the host from sending an empty packet when
data is aligned to an endpoint packet boundary.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: dim2: use device to allocate coherent memory

On several architectures the allocation of coherent memory needs a device
that has the dma_ops structure properly initialized. This patch enables
the DIM2 platform to be used to allocate this type of memory.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: dim2: read clock speed from the device

This patch implemets reading of the clock speed from DT.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: dim2: use device tree

This patch removes the dependency to platform specific source files
that do platform specific initialization and supply the IRQ number.
Instead DT code is added

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: cdev: fix race condition

This patch fixes a race condition between the functions disconnect and poll.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: dim2: fix startup sequence

Platform specific initialization (data->init) has to be done before
calling dim_startup to start the DIM2 IP.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: cdev: fix function return value

The function ch_get_mbo declares its return value as type bool,
but returns a pointer to mbo.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: usb: add ep number to log

This patch adds the endpoint number of the USB pipe that reports to be
broken into the log message. It is needed to make debugging for
applications more comfortable.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: cdev: fix chrdev_region leak

The function unregister_chrdev_region is called with a different counter
as the alloc_chrdev_region. To fix this, this patch introduces the
constant CHRDEV_REGION_SIZE that is used in both functions.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: cdev: avoid warning about potentially uninitialized variable

This patch avoids the warning that the pointer mbo might be used
uninitialized that some environmens throw.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: sound: call snd_card_new with struct device

This patch is needed as function snd_card_new needs a valid
parent device. Passing a NULL pointer leads to kernel Ooops.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: make interface drivers allocate coherent memory

On arm64/aarch64 architectures the allocation of coherent memory needs a
device that has the dma_ops properly set. That's why the core module of
the MOST driver is no longer able to allocate this type or memory. This
patch moves the allocation process down to the interface drivers where
the proper devices exist (e.g. platform device or USB system software).

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: reduce parameters inconsistency

Currently, there are two module parameters for the i2c driver:
  - polling_req: boolean irq/polling mode;
  - scan_rate: polling rate, that is used in the case where the polling
    mode is active

This model is misconfiguration-prone.  For example, it is possible to
select polling mode with the zero polling rate or configure non-zero
polling rate in a combination with the IRQ mode.

This patch replaces the 'polling_req' and 'scan_rate' by the
'polling_rate', where the value zero means the interrupt driven mode and
other values are used as the polling rate in the polling mode.

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: remove redundant list_mutex

The elements of the dev->rx.list are consumed in the pending_rx_work and
populated in the function enqueue() that cancels the pending_rx_work.

The function enqueue() and poison_channel() do not race anyway.

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: remove redundant is_open

The variable is_open is checked only in the work function
pending_rx_work() that is only active between the calls
configure_channel() and poison_channel().

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: trace real polling rate

The real polling rate depends on the CONFIG_HZ and may differ from the
required polling rate.

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: prevent zero delay polling

This patch avoids that a configured scan_rate of more than MSEC_PER_SEC
might result in a polling delay of zero.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: avoid polling in case of misconfig

This patch prevents the driver from falling back to polling mode
in case of IRQ misconfiguration.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: do not wait in work function

This patch removes the function wait_event_interruptible from the
work function to avoid waiting.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: shorten lifetime of IRQ handler

Currently the IRQ handler used for the rx channel lives between the
functions i2c_probe and i2c_remove. This patch shortens the lifetime
and keeps the handler alive only between the functions configure_channel
and poison_channel.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: aim-sound: add flexible format support

Currently, the only supported PCM formats are 1x8", "2x16", "2x24",
"2x32" or "6x16".

This adds support for the format "Nx{8,16,24,32}" that also includes the
exotic PCM formats like "4x16", "5x8", etc.

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: add channel property dbr_size

This patch adds the channel property dbr_size to control the corresponding
buffer size of the channels of the DIM2 interface.

Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: remove unnecessary poison_channel call

This removes call of the poison_channel that is:
  - not allowed after most_deregister_interface;
  - is made during the most_deregister_interface call.

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: i2c: prevent division by zero

This prevents division by zero scan_rate.

The zero scan_rate does not need any special action as it actually means
"never poll again".

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: dim2: remove clock speed processing from the HDM

This removes the module parameter clock_speed from the HDM code.

Instead, the platform-dependent clock speed must be delivered by the
platform driver with the help of the dim2_platform_data.clk_speed.

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: most: allocate only all requested memory

This prohibits the allocation of the memory for the MBOs if only the
part of the MBOs, requested by the application, may be allocated.  The
function arm_mbo_chain, if cannot allocate all requested MBO, frees all
prior allocated memory and returns 0.

Signed-off-by: Andrey Shvetsov <andrey.shvetsov@k2l.de>
Signed-off-by: Christian Gromm <christian.gromm@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: ion: Remove unnecessary blank line

Fixes a checkpatch.pl warning.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Acked-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: ion: Fix license identifier comment format

checkpatch.pl complains these are invalid because the rules in
Documentation/process/license-rules.rst state that C headers should
have "/* */" style comments.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Acked-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: vsoc: Fix ending '(' warnings in vsoc_probe_device

Fixes checkpatch.pl warnings about lines ending with parentheses.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: vsoc: Fix ending '(' warnings in vsoc_ioctl

Fixes checkpatch.pl warnings about lines ending with parentheses.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: vsoc: Fix ending '(' warnings in do_destroy_fd_scoped_permission

Fixes checkpatch.pl warnings about lines ending with parentheses.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: vsoc: Fix ending '(' warnings in function defintions

Fixes checkpatch.pl warnings about lines ending with parentheses.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: Clean up license identifiers

Add the identifiers when missing and fix the ones already present
according to checkpatch.pl.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: Kconfig; Remove excessive hyphens

Fixes the following checkpatch.pl warning:
"prefer 'help' over '---help---' for new help texts"

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fix error deref in ll_splice_alias().

d_splice_alias() can return an ERR_PTR().
If it does while debugging is enabled, the following
CDEBUG() will dereference that error and crash.

So add appropriate checking, and provide a separate
debug message for the error case.

Reported-and-tested-by: James Simmons <jsimmons@infradead.org>
Fixes: e9d4f0b9f559 ("staging: lustre: llite: use d_splice_alias for directories.")
Reviewed-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: move remaining code from linux-module.c to module.c

There is no longer any need to keep this code separate,
and now we can remove linux-module.c

Reviewed-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: move misc-device registration closer to related code.

The ioctl handler for the misc device is in  lnet/libcfs/module.c
but is it registered in lnet/libcfs/linux/linux-module.c.

Keeping related code together make maintenance easier, so move the
code.

Reviewed-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: remove redundant lookup in dump_pgcache

Both the 'next' and the 'show' functions for the dump_page_cache
seqfile perform a lookup based on the current file index.  This is
needless duplication.

The reason appears to be that the state that needs to be communicated
from "next" to "show" is two pointers, but seq_file only provides for
a single pointer to be returned from next and passed to show.

So make use of the new 'seq_private' structure to store the extra
pointer.
So when 'next' (or 'start') find something, it returns the page and
stores the clob in the private area.
'show' accepts the page as an argument, and finds the clob where it
was stored.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: use more private data in dump_pgcache

The dump_page_cache debugfs file allocates and frees an 'env' in each
call to vvp_pgcache_start,next,show.  This is likely to be fast, but
does introduce the need to check for errors.

It is reasonable to allocate a single 'env' when the file is opened,
and use that throughout.

So create 'seq_private' structure which stores the sbi, env, and
refcheck, and attach this to the seqfile.

Then use it throughout instead of allocating 'env' repeatedly.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fold lu_object_new() into lu_object_find_at()

lu_object_new() duplicates a lot of code that is in
lu_object_find_at().
There is no real need for a separate function, it is simpler just
to skip the bits of lu_object_find_at() that we don't
want in the LOC_F_NEW case.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: lu_object: move retry logic inside htable_lookup

The current retry logic, to wait when a 'dying' object is found,
spans multiple functions.  The process is attached to a waitqueue
and set TASK_UNINTERRUPTIBLE in htable_lookup, and this status
is passed back through lu_object_find_try() to lu_object_find_at()
where schedule() is called and the process is removed from the queue.

This can be simplified by moving all the logic (including
hashtable locking) inside htable_lookup(), which now never returns
EAGAIN.

Note that htable_lookup() is called with the hash bucket lock
held, and will drop and retake it if it needs to schedule.

I made this a 'goto' loop rather than a 'while(1)' loop as the
diff is easier to read.

Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: lu_object: discard extra lru count.

lu_object maintains 2 lru counts.
One is a per-bucket lsb_lru_len.
The other is the per-cpu ls_lru_len_counter.

The only times the per-bucket counters are use are:
- a debug message when an object is added
- in lu_site_stats_get when all the counters are combined.

The debug message is not essential, and the per-cpu counter
can be used to get the combined total.

So discard the per-bucket lsb_lru_len.

Reviewed-by: Andreas Dilger <andreas.dilger@dilger.ca>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: make struct lu_site_bkt_data private

This data structure only needs to be public so that
various modules can access a wait queue to wait for object
destruction.
If we provide a function to get the wait queue, rather than the
whole bucket, the structure can be made private.

Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: greybus: Remove unused local variable

Fixes the following W=1 warning: variable ‘intf_id’ set but
not used [-Wunused-but-set-variable]

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: board: Replace license boilerplate with SPDX identifiers

This satisfies a checkpatch.pl warning and is the preferred method for
notating the license due to its lack of ambiguity.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: Remove unnecessary array index check

This statment triggers GCC's -Wtype-limit since key_index is an
unsigned integer so it cannot be less than zero.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: Remove useless function

GCC warns that 'wid' is unused in wilc_remove_key and it's correct; the
variable is only local. Get rid of the function (since it just returns
zero) and shuffle the remaining code into one if statement.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: Remove unused variables

GCC warns these variables are all set but never used so remove them.

Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: r8822be: fix typo in variable name "offest" -> "offset"

There are multiple occurrances of typos of "offest" that should be
"offset". Fix these.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-dts: update nor-flash info for gnubee1

The GNUBEE has 32MB flash, so set partitions accordingly.
Also remove "m25p,chunked-io" which isn't documented or
used anywhere (outside of freewrt).

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-pci: remove some dead code.

Some code is dead because it is commented out.
Some is dead because it is uninteresting printks.
Some is dead because it declares unused functions.

Remove it all.

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-pci: remove unnecessary resource details.

These resources are extracted from devicetree, so they aren't
needed here.

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-pci: remove conditional compilation.

Code currently defines:

#define CONFIG_PCIE_PORT0
#define CONFIG_PCIE_PORT1
#define CONFIG_PCIE_PORT2
#define GPIO_PERST

and then compiles code only if they are defined.
We might want to disable some of these via devicetree one
day, but for now just remove the #defines and the
conditions - all the code for different ports is
easy to identify.

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-pci: white-space cleanups.

- remove white space at end of line.
- no more than 2 blank line at a time
- remove spaces before tabs
- use tabs to line things up
- re-indent some #define do{}while(0)

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-pci: improve interrupt mapping

As the Interrupts for the PCI adapters are listed in
devicetree we shouldn't need to have them explicit in the code.

The simplest way to do this is to use of_irq_parse_and_map_pci()
and specify an interrupt-map which identifies the different
PCI hosts by bus/slot numbers.
This has the advantage that the hwirq number are mapped to virq
numbers for us, so the ugly hack can go.

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-spi: remove unused lock.

This lock is never initialized, locked once, and never unlocked.
Clearly it is pointless - so remove it.

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-eth: Lock is never unlocked.

mtk_phy_link_adjust takes a spinlock and disables
interrupts, but never unlocks.
This can leave interrupts disabled on one CPU and
various things stop working.

Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused barrier macros

The current code has some barrier macros in it, which are already '#if
0' out, so just remove them.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Cleanup includes

This commit cleans the includes in the sd.c file. Those which are not
needed are removed, the remaining ones are sorted alphabetically.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused macro HOST_MAX_NUM

The macro HOST_MAX_NUM is never used, so remove it.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Always use linked DMA mode

Current code uses linked DMA mode only when there is more than one
entry in the scatterlist. But the overhead of the linked DMA is very
small, so the code is easier when only using linked DMA and this is
also the way, the upstream driver mtk-sd handles it.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Change return value of msdc_dma_config to void

The msdc_dma_config function currently returns always 0. Remove the
return and change to void.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Don't disables DMA

Current code disables the DMA after the transfer. That only changes
the transfer mode to non-DMA and does not save any power. This is not
necessary any more, because now DMA transfer is always used. The macro
for disabling DMA transfer is also removed, because it is not used any
more.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Inline function msdc_set_blknum

The function msdc_set_blknum consists of one (real) line of code and
is only called once, so inline it makes the code shorter and more
readable.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove initialisation in msdc_dma_config

Current code initialises the variable sg at the beginning of the
msdc_dma_config function. This is not necessary, because the variable
is assigned by the for_each_sg macro later on.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove redundant check for card status

Current code checks the card status at the beginning of
msdc_ops_request. This is not necessary because mmc core always checks
the card status before calling this operation.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Do not BUG() if mrq set in msdc_ops_request

Currently the code BUG()'s, if host->mrq is set at the beginning of
msdc_ops_request. This shoould normally not happen, but it is not that
critical, because the critical sections are protected by a spin lock
and in the worst case, some commands to the card are lost, so it is
sufficient to just WARN_ON().

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused sdio irq code

Currently the code for the sdio irq is never used, because the flags
for it (MSDC_(EXT_)SDIO_IRQ) are never set. So the whole code for it
can be removed.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused field abort from msdc_host

The field abort of msdc_host is only set, but never read, so it can be
removed.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Fix error number assignment

Currently the value that data->error is set to is converted to an
unsigned int, but it is a usual error number, so it should be
negative.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused variable dma_xfer from msdc_host

The field dma_xfer of the struct msdc_host is not used anymore, remove
it.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove non-DMA transfer code

Currently the driver is capable of DMA and non-DMA transfer. But the
option to choose non-DMA transfer has already been removed. Now remove
also the code for the non-DMA transfer.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused global dma_size

The global variable dma_size was previously used to select DMA or
non-DMA transfer mode based on the size of the data that should be
transferred. This option was removed preivously, so the variable is not
used any more and can be removed.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove transfer mode setting variable

Currently the transfer mode can be chosen between DMA, a size
dependent mode and non-DMA by writing to a proc file. The upstream
driver mtk-sd uses DMA all times. The previous patch removed the
ability to set that option.

Now the remaining uses of the transfer mode setting variable are
cleaned up, because it cannot be changed any more.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove transfer mode setting from proc

Currently the transfer mode can be chosen between DMA, a size
dependent mode and non-DMA by writing to a proc file. The upstream
driver mtk-sd uses DMA all times.

There is no apparent reason why somebody would like to change the
transfer mode and the position of the setting in the debug part of the
driver also indicates, that the option was used for debugging
purposes. So it is removed to clean up the driver and bring it more in
line with the upstream one.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused static msdc_regs

The static variable msdc_regs is set once, but never used, so remove
it.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Replace macro sdr_get_field with function

Currently sdr_get_field is a macro, to bring the code in line with the
upstream driver mtk-sd, it is changed to a function.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Replace macro sdr_set_field with function

Currently sdr_set_field is a macro, to bring the code in line with the
upstream driver mtk-sd, it is changed to a function.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove function uffs

The function uffs that is implemented by this driver returns the same
values as the kernel function ffs. So the uffs function is removed and
the calls to it are replaced with calls to ffs.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused macros sdr_{read,write}16

The sdr_{read,write}16 macros are never used and does not provide any
information about the device, so they are removed.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused macro msdc_init_gpd_ex

The macro msdc_init_gpd_ex is never used and does not provied any
information about the hardware, so it is removed.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused field clk_drv from msdc_hw

The field clk_drv of msdc_hw is set to a constant and only used once,
replace that position with the constant and remove the unused field.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused field cmd_drv from msdc_hw

The field cmd_drv of msdc_hw is set to a constant and only used once,
replace that position with the constant and remove the unused field.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused field dat_drv from msdc_hw

The field dat_drv of msdc_hw is set to a constant and only used once,
replace that position with the constant and remove the unused field.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused field data_edge from msdc_hw

The field data_edge of msdc_hw is set to a constant and only used once,
replace that position with the constant and remove the unused field.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Remove unused field cmd_edge from msdc_hw

The field cmd_edge of msdc_hw is set to a constant and only used once,
replace that position with the constant and remove the unused field.

Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Fix typo in function parameters

The type pm_message does not exist in the kernel, the correct type is
pm_message_t, so the type of the parameter is corrected.

Fixes: 9673d9f6f44b ("staging: mt7621-mmc: Refactor suspend, resume")
Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vc04_services: Remove stack hog issue from TODO

Since there is no vchiq function listed by "make checkstack" anymore,
we can remove the stack hog issue from the TODO.

Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq_shim: Remove unnecessary break

This removes an unnecessary break which has been reported by checkpatch.pl.

Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq_arm: remove unnecessary return

This removing an unnecessary return which has been reported by checkpatch.pl.

Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq_2835_arm: Fix whitespace issue in create_pagelist

This fixes a whitespace issue reported by checkpatch:

ERROR: code indent should use tabs where possible

Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vc04_services: Remove defines from Makefile

The definition of USE_VCHIQ_ARM and VCOS_VERIFY_BKPTS isn't really necessary
so remove them.

Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq_arm: Make index variable more self-explaining

The chance to mixup i and j is very high. So rename variable j to a more
explaining one.

Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>