staging: mt7621-pinctrl: Remove redundant owner assignment

Remove the owner assignment form the platform driver as
platform_driver_register() already initializes the owner.
Found using coccinelle.

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

staging: mt7621-pci: Remove redundant owner assignment

Remove the owner assignment form the platform driver as
platform_driver_register() already initializes the owner.
Found using coccinelle.

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

staging: mt7621-mmc: Remove redundant opcode check

The opcode check in msdc_command_start is redundant in case of
MMC_SET_RELATIVE_ADDR, this is 3 like SD_SEND_RELATIVE_ADDR, so the
second check can be removed.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
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: Fix: copy_from_user() returns a positive value

copy_from_user() returns a positive value in case of an error, to fix
this the check is turned around, also a better return value is chosen.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Suggested-by: Dan Carpenter <dan.carpenter@oracle.com>
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: Fix debug file world writable

Currently the debug file in proc fs is world writable, remove the
access for other.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
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 error message in debug

If the debug file in proc fs is not successfully created current code
prints an error message, this is removed. de is also not an error
pointer, so the IS_ERR() call leads to a static checker warning. The
unused return value of the msdc_debug_proc_init function is also removed.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
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 constant flag MSDC_HIGHSPEED

The MSDC_HIGHSPEED flag is always set and never unset, remove it to
clean up the code.

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 functions

The functions msdc_eirq_cd and msdc_eirq_sdio are defined but never
used, so they are removed. This also fixes two compiler warnings:

drivers/staging/mt7621-mmc/sd.c:517:13: warning: ‘msdc_eirq_cd’ defined but
not used [-Wunused-function]
 static void msdc_eirq_cd(void *data)
             ^~~~~~~~~~~~
drivers/staging/mt7621-mmc/sd.c:507:13: warning: ‘msdc_eirq_sdio’ defined but
not used [-Wunused-function]
 static void msdc_eirq_sdio(void *data)

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: msdc_6575_host

The mt6575_host variable is never used, only set and
exported. Removing it for code cleanup.

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 constant flag

The MSDC_SYS_SUSPEND flag is always set and is never unset. Removing
it cleans up the code.

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: Fix power management by removing old api

The mmc_suspend|resume_host functions have been decrepated in the
3.11 [1] release and were completely removed in 3.13 [2]. Removing the
calls to this functions as it was done for other mmc host drivers
before the old api was removed (eg. [3]) fixes compile errors in this
driver while CONFIG_PM is enabled.

The rest of the removed code is simply dead, the MSDC_SYS_SUSPEND flag
is always set and will be removed in a later patch.

[1] commit 986892ca78ee ("mmc: core: Initiate suspend|resume from mmc bus instead of mmc host")
[2] commit 3c0d22e8180b ("mmc: core: Remove deprecated mmc_suspend|resume_host APIs")
[3] commit 1e63d4854202 ("mmc: au1xmmc: Move away from using deprecated APIs")

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 duplicate defines

Remove defines that are both in board.h and sd.c from sd.c, this file
already includes board.h

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: Fix requesting of irq

The request of the irq line is not correctly handled by the current
code, platform_get_irq may fail and if the irq is requested via
devm_request_irq, its release is automatically done.

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: Fix dma_alloc_coherent should get device as first param

The dma_alloc_coherent (and also dma_free_coherent) should get the
device, the dma memory is allocated for as the first parameter.

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 dead codepath from probe function

The probe function contains lots of code, that is never used, because
the fields it checks for in the msdc_hw structure are never
initialized. Removing them should be safe.

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: Fix different caps for different devices

Current code saves the need for polling in a global variable, that
leeds to problems, when the driver should handle multiple devices with
different capabilities. By looking up the capability in the mmc_host
capabilities, they now have per device scope.

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

staging: mt7621-mmc: Remove unnecessary field from struct msdc_hw

Remove the unnecessary field data_pins from msdc_hw, later this
information should be taken from the device tree.

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

staging: mt7621-mmc: Remove unnecessary setting of dev.platform_data

In the probe function, pdev->dev.platform_data is set to &msdc0_hw,
but the only get access to that is also in the probe function. So that
assignment is removed. The msdc0_hw structure is still available to the
driver as part of the msdc_host structure.

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: Fix bug on dma_alloc_coherent fail

In case of dma_alloc_coherent failing the current code just called
BUG_ON. By adding error handling for that case this can be avoided.
This also fixes a memory leek in case of a fail later on in the probe
function.

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: Fix unmatched release_mem_regin

Current code calls release_mem_regin on driver remove without
requesting it explicit first. The region is only requested via
devm_ioremap_resource and that releases it automatically. Removing the
release_mem_region calls fixes this.

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

staging: mt7621-mmc: Remove code checking for old kernel versions

Remove code, that is checking for old kernel versions, leaving only
the case for current versions.

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

staging: mt7621-mmc: Fix memory leek in case of error in probe

If the base address is not successfully obtained in the probe
function, then the mmc_host struct is not freed. Adding an exit for
that case fixes the bug.

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: Refactor suspend, resume

Refactor msdc_drv_{suspend, resume} by adding a new function
msdc_drv_pm, that cleans up the code and removes double code.

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: Cleanup function mt_msdc_init

Cleanup the mt_msdc_init function, remove excceding printk.

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 datastructures

Remove unused datastructures and comments referring to their former use.

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

staging: mt7621-mmc: Remove code for different soc types

Remove code that is intended for soc other then the mt7621 because
this driver is only for that soc. This also removes some conditions.
However the remaining code is also at the wrong place, it just selects
the mode of the pins, this is a job for pinctl not for the sdxc driver.

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

staging: mt7621-mmc: Remove code guarded by nonexistent config option

Remove a piece of code that is guarded by a config option, that is
nonexistent, the function it calls is also nonexistent.

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

staging: mt7621-mmc: Replace seq_printf with seq_puts where possible

Replace seq_printf with seq_puts when the string is const.
Fix checkpatch warning: PREFER_SEQ_PUTS

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 redundant driver owner assignment

Remove the owner assignment form the platform driver as
platform_driver_register() already initializes the owner.

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: Fix style of braces

Fix checkpatch errors, warnings, checks about braces:
BRACES, ELSE_AFTER_BRACE, OPEN_BRACE

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

staging: mt7621-mmc: Remove global initialise to 0

Fix checkpatch error: GLOBAL_INITIALISERS

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

staging: mt7621-mmc: Fix trailing statements

Fix checkpatch error: TRAILING_STATEMENTS

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 unnecessary typedefs

Fixes checkpatch warning: NEW_TYPEDEFS

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

staging: mt7621-mmc: Fix Pointer Location

Fix checkpatch error: POINTER_LOCATION

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: Fix spacing

Fixes checkpatch errors, warnings, checks:
SPACING, LEADING_SPACE, LINE_SPACING

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

staging: mt7621-mmc: Fix indent for sd.c

This fixes the indentaions in the file sd.c

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

staging: mt7621-mmc: Fix indent for mt6575_sd.h

This fixes the indentaions in the file mt7621_sd.h

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

staging: mt7621-mmc: Fix indent for dbg.h

This fixes the indentaions in the file dbg.h

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: Fix indent for board.h

This fixes the indentaions in the file board.h

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: Fix indent for dbg.c

This fixes the indentaions in the file dbg.c

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

staging: mt7621-mmc: Remove whitespace errors in sd.c

The whitespace errors in the file sd.c are fixed by using the
cleanfile script. Indentations with whitespaces are not changed
in this patch.

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

staging: mt7621-mmc: Remove whitespace errors in mt6575_sd.h

The whitespace errors in the file mt6575_sd.h are fixed by using the
cleanfile script. Indentations with whitespaces are not changed
in this patch.

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

staging: mt7621-mmc: Remove whitespace errors in dbg.h

The whitespace errors in the file dbg.h are fixed by using the
cleanfile script. Indentations with whitespaces are not changed
in this patch.

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 whitespace errors in dbg.c

The whitespace errors in the file dbg.c are fixed by using the
cleanfile script. Indentations with whitespaces are not changed
in this patch.

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 whitespace errors in board.h

The whitespace errors in the file board.h are fixed by using the
cleanfile script. Indentations with whitespaces are not changed
in this patch.

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-gpio: prefer unsigned int to bare unsigned

This commit replaces some bare unsigned definitions in some
function parameters in favour of 'unsigned int' which is
preferred.

This also fix checkpatch warnings about this.

Signed-off-by: Sergio Paracuellos <sergio.paracuellos@gmail.com>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-pci: Hack 2 more PCI interrupts for gnubee1

The mt7621-pci support 3 pci devices and has 3 interrupts.
Each of these need to be enabled by the same sort of hack to
map hwirq number to virq number.
This is a hack which will go as soon as I understand how this is
supposed to work.

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

staging: mt7621-eth: Fix sparse warning in ethtool.c

This fixes the following sparse warning:

drivers/staging/mt7621-eth/ethtool.c:213:6: warning: symbol
'mtk_set_ethtool_ops' was not declared. Should it be static?

Signed-off-by: Chris Coffey <cmc@babblebit.net>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: remove unnecessary line breaks in function definition.

This patch corrects the function definition style of CARDvSafeResetRx.
Issue found by checkpatch.

CHECK: Lines should not end with a '('

Signed-off-by: Danilo Alves <daniloalves@riseup.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: add handling memory leak on vnt_start()

There was no code for handling memory leaks of device_init_rings() and
request_irq(). It needs to free allocated memory in the device_init_rings()
, when request_irq() would be failed. Add freeing sequences of irq and
device init rings.

Signed-off-by: Ji-Hun Kim <ji_hun.kim@samsung.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655: check for memory allocation failures

There are no null pointer checking on rd_info and td_info values which
are allocated by kzalloc. It has potential null pointer dereferencing
issues. Implement error handling code on device_init_rd*, device_init_td*
and vnt_start for the allocation failures.

Signed-off-by: Ji-Hun Kim <ji_hun.kim@samsung.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: add container_of_safe()

Luster has a container_of0() function which is similar to
container_of() but passes an IS_ERR_OR_NULL() pointer through
unchanged.
This could be generally useful: bcache at last has a similar function.

Naming is hard, but the precedent set by hlist_entry_safe() suggests
a _safe suffix might be most consistent.

So add container_of_safe() to kernel.h, and replace all occurrences of
container_of0() with one of
  - list_first_entry, list_next_entry, when that is a better fit,
  - container_of(), when the pointer is used as a validpointer in
    surrounding code,
  - container_of_safe() when there is no obviously better alternative.

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

staging: lustre: rearrange placement of CPU partition management code.

Currently the code for cpu-partition tables lives in various places.
The non-SMP code is partly in libcfs/libcfs_cpu.h as static inlines,
and partly in lnet/libcfs/libcfs_cpu.c - some of the functions are
tiny and could well be inlines.

The SMP code is all in lnet/libcfs/linux/linux-cpu.c.

This patch moves all the trivial non-SMP functions into
libcfs_cpu.h as inlines, and all the SMP functions into libcfs_cpu.c
with the non-trival !SMP code.

Now when you go looking for some function, it is easier to find both
versions together when neither is trivial.

There is no code change here - just code movement.

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

staging: lustre: remove include/linux/libcfs/linux/linux-cpu.h

This include file contains definitions used when CONFIG_SMP
is in effect.  Other includes contain corresponding definitions
for when it isn't.
This can be hard to follow, so move the definitions to the one place.

As HAVE_LIBCFS_CPT is defined precisely when CONFIG_SMP, we discard
that macro and just use CONFIG_SMP when needed.

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

staging: lustre: convert osc_quota hash to rhashtable

As this is indexed by an integer, an extensible array
or extensible bitmap would be better.
If/when xarray lands, we should change to use that.

For now, just a simple conversion to rhashtable.

When removing an entry, we need to hold rcu_read_lock()
across the lookup and remove in case we race with another thread
performing a removal.  This means we need to use call_rcu()
to free the quota info so we need an rcu_head in there, which
unfortunately doubles the size of the structure.

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

staging: lustre: convert obd uuid hash to rhashtable

The rhashtable data type is a perfect fit for the
export uuid hash table, so use that instead of
cfs_hash (which will eventually be removed).

As rhashtable supports lookups and insertions in atomic
context, there is no need to drop a spinlock while
inserting a new entry, which simplifies code quite a bit.

As there are no simple lookups on this hash table (only
insertions which might fail and take a spinlock), there is
no need to use rcu to free the exports.

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

staging: lustre: convert lov_pool to use rhashtable

The pools hashtable can be implemented using
the rhashtable implementation in lib.
This has the benefit that lookups are lock-free.

We need to use kfree_rcu() to free a pool so
that a lookup racing with a deletion will not access
freed memory.

rhashtable has no combined lookup-and-delete interface,
but as the lookup is lockless and the chains are short,
this brings little cost.  Even if a lookup finds a pool,
we must be prepared for the delete to fail to find it,
as we might race with another thread doing a delete.

We use atomic_inc_not_zero() after finding a pool in the
hash table and if that fails, we must have raced with a
deletion, so we treat the lookup as a failure.

Use hashlen_string() rather than a hand-crafted hash
function.
Note that the pool_name, and the search key, are
guaranteed to be nul terminated.

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

staging: lustre: ptlrpc: convert conn_hash to rhashtable

Linux has a resizeable hashtable implementation in lib,
so we should use that instead of having one in libcfs.

This patch converts the ptlrpc conn_hash to use rhashtable.
In the process we gain lockless lookup.

As connections are never deleted until the hash table is destroyed,
there is no need to count the reference in the hash table.  There
is also no need to enable automatic_shrinking.

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

staging: lustre: remove libcfs/linux/libcfs.h

This include file is only included in one place,
and only contains a list of other include directives.
So just move all those to the place where this file
is included, and discard the file.

One include directive uses a local name ("linux-cpu.h"), so
that needs to be given a proper path.

Probably many of these should be remove from here, and moved to
just the files that need them.

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

staging: lustre: move stack-check macros to libcfs_debug.h

CDEBUG_STACK() and CHECK_STACK() are macros to help with
debugging, so move them from
   drivers/staging/lustre/include/linux/libcfs/linux/libcfs.h
to
   drivers/staging/lustre/include/linux/libcfs/libcfs_debug.h

This seems a more fitting location, and is a step towards
removing linux/libcfs.h and simplifying the include file structure.

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

staging: lustre: llite: correct removexattr detection

In ll_xattr_set_common() detect the removexattr() case correctly by
testing for a NULL value as well as XATTR_REPLACE.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-10787
Reviewed-on: https://review.whamcloud.com/
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: use xattr_handler name for ACLs

If struct xattr_handler has a name member then use it (rather than
prefix) for the ACL xattrs. This avoids a bug where ACL operations
failed for some kernels.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-10785
Reviewed-on: https://review.whamcloud.com/
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: style changes in xattr.c

Small style changes to match more the kernel code standard
and it make it more readable.

Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9183
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: cleanup xattr code comments

Add proper punctuation to the comments. Change buf_size to size
for comment in ll_listxattr() since buf_size doesn't exit which
will confuse someone reading the code.

Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9183
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: use proper types in the xattr code

Convert __uXX types to uXX types since this is kernel code.
The function ll_lov_user_md_size() returns ssize_t so change
lum_size from int to ssize_t.

Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9183
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: cleanup posix acl xattr code

Having an extra ifdef makes the code harder to read. For the case
of ll_xattr_get_common() we have a variable initialized at the
start of the function but it is only used in XATTR_ACL_ACCESS_T
code block. Lets move that variable to that location since its
only used there and make the code look cleaner.

Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9183
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: record in stats attempted removal of lma/link xattr

Keep track of attempted deletions as well as changing of the
lma/link xattrs.

Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9183
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: remove newline in fullname strings

In creating the full name of a xattr a new line was added that
was seen by the remote MDS server which confused it. Remove the
newline.

Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9183
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: fix invalid size test in ll_setstripe_ea()

The size check at the start of ll_setstripe_ea() is only
valid for a directory. Move that check to the section of
code handling the S_ISDIR case.

Signed-off-by: James Simmons <uja.ornl@yahoo.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9183
Reviewed-on: https://review.whamcloud.com/27240
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: eat -EEXIST on setting trusted.lov

Tools like rsync, tar, cp may copy and restore the xattrs on a file.
The client previously ignored the setting of trusted.lov/lustre.lov
if the layout had already been specified, to avoid causing these
tools to fail for no reason.

For PFL files we still need to silently eat -EEXIST on setting these
attributes to avoid problems.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9484
Reviewed-on: https://review.whamcloud.com/27126
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: return from ll_adjust_lum() if lump is NULL

No need to check several times if lump is NULL. Just test once and
return 0 if NULL.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9484
Reviewed-on: https://review.whamcloud.com/27126
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: break up ll_setstripe_ea function

Place all the handling of information of trusted.lov that
is not stripe related into the new function ll_adjust_lum().
Now ll_setstripe_ea() only handles striping information.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9484
Reviewed-on: https://review.whamcloud.com/27126
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Niu Yawei <yawei.niu@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: add simple comment about lustre.lov xattrs

Simple comment added to ll_xattr_set.

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8998
Reviewed-on: https://review.whamcloud.com/24851
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: refactor lustre.lov xattr handling

The function ll_xattr_set() contains special code to handle
the lustre specific xattr lustre.lov. Move all this code to
a new function ll_setstripe_ea().

Signed-off-by: Bobi Jam <bobijam.xu@intel.com>
Signed-off-by: Niu Yawei <yawei.niu@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8998
Reviewed-on: https://review.whamcloud.com/24851
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@gmail.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: Remove filtering of seclabel xattr

The security.capability xattr is used to implement File
Capabilities in recent Linux versions. Capabilities are a
fine grained approach to granting executables elevated
privileges. eg. /bin/ping can have capabilities
cap_net_admin, cap_net_raw+ep instead of being setuid root.

This xattr has long been filtered out by llite, initially for
stability reasons (b15587), and later over performance
concerns as this xattr is read for every file with eg.
'ls --color'. Since LU-2869 xattr's are cached on clients,
alleviating most performance concerns.

Removing llite's filtering of the security.capability xattr
enables using Lustre as a root filesystem, which is used on
some large clusters.

Signed-off-by: Robin Humble <plaguedbypenguins@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-9562
Reviewed-on: https://review.whamcloud.com/27292
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: handle xattr cache refill race

In ll_xattr_cache_refill() if the xattr cache was invalid (and no
request was sent) then return -EAGAIN so that ll_getxattr_common()
caller will fetch the xattr from the MDT.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-10132
Reviewed-on: https://review.whamcloud.com/29654
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: ldlm: xattr locks are lost on mdt

On the server side mdt_intent_getxattr() can return EFAULT if a
buffer cannot be found, it is returned after lock_replace, where a
new lock is installed into lockp. An error forces ldlm_lock_enqueue()
to destroy the original lock, but ldlm_handle_enqueue0() drops the
reference on the new lock. The xattr client code implied intent
error is returned under a lock, which is immediately cancelled.
Check if a lock obtained and cancel it properly for error cases.
Note: we should support both cases for interop needs, an intent
error under a lock and with a lock abort. Keep returning a lock
with an intent error for interop purposes for now, to be dropped
later when client will get old enough. make all intent ops to
work through md_intent_lock: getxattr and layout, which should
extract the intent error.

Signed-off-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7433
Seagate-bug-id: MRP-3072 MRP-3137
Reviewed-on: http://review.whamcloud.com/17220
Reviewed-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: obd: change debug reporting in lmv_enqueue()

Remove LL_IT2STR(it) from debug macros in lmv_enqueue(). The
removal makes it possible to simplify the md_enqueue() functions.

Signed-off-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7433
Seagate-bug-id: MRP-3072 MRP-3137
Reviewed-on: http://review.whamcloud.com/17220
Reviewed-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: obd: create it_has_reply_body()

The lookup_intent it_op fields in many cases will be compared
to the settings of IT_OPEN | IT_UNLINK | IT_LOOKUP | IT_GETATTR.
Create a simple inline function for this common case.

Signed-off-by: Vitaly Fertman <vitaly.fertman@seagate.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7433
Seagate-bug-id: MRP-3072 MRP-3137
Reviewed-on: http://review.whamcloud.com/17220
Reviewed-by: Andrew Perepechko <andrew.perepechko@seagate.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@seagate.com>
Tested-by: Elena V. Gryaznova <elena.gryaznova@seagate.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: llite: initialize xattr->xe_namelen

When the allocation of xattr->xe_name was moved to kstrdup()
setting xattr->xe_namelen was dropped. This field is used
in several parts of the xattr cache code so it broke xattr
handling. Initialize xattr->xe_namelen when allocating
xattr->xe_name succeeds. Also change the debugging statement
to really report the xattr name instead of its length which
wasn't event being set.

Fixes: b3dd8957c23a ("staging: lustre: lustre: llite: Use kstrdup"
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: libcfs: use dynamic minors for /dev/{lnet, obd}

Request dynamic minor allocation when registering /dev/lnet and
/dev/obd.

Signed-off-by: John L. Hammond <john.hammond@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-100086
Reviewed-on: https://review.whamcloud.com/29741
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jian Yu <jian.yu@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: NeilBrown <neilb@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: remove cruft from libcfs/linux/libcfs.h

These defines are unused or nearly unused, and do not
help at all.

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

staging: lustre: discard linux-time.h

Very little is left in linux-time.h.
Move CFS_TICK to libcfs.h discard the test.

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

staging: lustre: discard cfs_time_before_64()

cfs_time_before_64 is the same as time_before64()
similarly cfs_time_beforeq_64() matsches time_before_eq64()
So just use the standard interfaces.

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

staging: lustre: discard cfs_duration_sec()

cfs_duration_sec() simply divides by HZ.
It is mostly used to report durations in debug messages.
Remove and just use X/HZ.

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

staging: lustre: libcfs: remove cfs_timeout_cap()

This wrapper is only used once, so open-code it as max().

This allows us to remove the libcfs_time.h include file.

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

staging: lustre: libcfs: discard cfs_time_after()

cfs_time_after() behaves exactly like time_after()
similarly cfs_time_aftereq() matches time_after_eq()

so discard the cfs versions.

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

staging: lustre: libcfs: discard cfs_time_shift().

This function simply multiplies by HZ and adds jiffies.
This is simple enough to be opencoded, and doing so
makes the code easier to read.

Same for cfs_time_shift_64()

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

staging: lustre: libcfs: discard cfs_time_add/sub

cfs_time_add adds its arguments.
cfs_time_sub subtracts finds the difference.
Discard these and use '+' and '-' directly.

This change highlighted a type error.  The structure field
cr_queued_time was used to store jiffies, but was declared
as time_t (meant for seconds).  So the time is changed to
"unsigned long".

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

staging: lustre: cfs_time_current() -> jiffies.

Discard cfs_time_current() and cfs_time_current64()
and use jiffies and get_jiffies_64() like the rest of the kernel.

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

staging: Android: Add 'vsoc' driver for cuttlefish.

The cuttlefish system is a virtual SoC architecture based on QEMU. It
uses the QEMU ivshmem feature to share memory regions between guest and
host with a custom protocol.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Arve Hjønnevåg <arve@android.com>
Cc: Todd Kjos <tkjos@android.com>
Cc: Martijn Coenen <maco@android.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Cc: devel@driverdev.osuosl.org
Cc: kernel-team@android.com
Signed-off-by: Greg Hartman <ghartman@google.com>
[astrachan: rebased against 4.16, added TODO, fixed checkpatch issues]
Signed-off-by: Alistair Strachan <astrachan@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: android: ion: remove duplicate buffer field initializes

As a result of various previous patches, ion_buffer_create() now
has two sets of identical statements for initializing two fields of
the buffer struct, next to each other.  Remove one set.

Move the initialization of these two fields together with the
statements that initialize the other two fields from the function
parameters, prior to the heap allocate() call, for consistency.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Acked-by: Sumit Semwal <sumit.semwal@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: ldlm: free resource when ldlm_lock_create() fails.

ldlm_lock_create() gets a resource, but don't put it on
all failure paths. It should.

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

staging: lustre: tidy up ldlm_resource_putref()

1/ the return value of ldlm_resource_putref() is never
  used, so change it to return 'void'.
2/ Move all of the code to run on the last putref to
   __ldlm_resource_putref_final().  This means a lock
   is taken in one function and dropped in another, but
   that isn't too uncommon, and will disappear in a future
   patch.
   Now that the code it together, it becomes apparent that
   we are dropping a ref on the namespace *before* the last
   use.  So keep the ref until after.

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

staging: lustre: remove unneeded include.

cfs_hash usage was removed from cl_object.c
some time ago, so remove the #include as well.

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

staging: lustre: remove unused hash tables.

exp_lock_hash and exp_flock_hash are unused in
the client, so remove all references.

Also remove unused hashtable size definitions.

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

staging: lustre: remove unused ldlm_completion_ast_async()

This function is unused.

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

staging: lustre: remove invariant in cl_io_read_ahead()

This invariant can fail as ci_type is somtimes CIS_INIT.

According to Jinshan Xiong, or of the most recent significant
change to this code, "We should get rid of it."

So do that.

Fixes: 1e1db2a97be5 ("staging: lustre: clio: Revise read ahead implementation")
Cc: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: NeilBrown <neilb@suse.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: nvec: convert to use GPIO descriptors

Use GPIO descriptors instead of relying on the old method.

Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Marc Dietrich <marvin24@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: nvec: cleanup use of dev in probe function

This cleanups the probe function a bit by using a dev variable instead
of &pdev-dev.

Signed-off-by: Marc Dietrich <marvin24@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>