staging: wfx: load the firmware faster

During the loading of the firmware, the WFX_DCA_GET register provide the
number available bytes in the receiving buffer. It is not necessary to
access to the WFX_DCA_GET after sent of each firmware fragment.

This patch allows to send the firmware:
  - in 64ms instead of 130ms using SDIO bus
  - in 78ms instead of 115ms using SPI bus

Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200701150707.222985-6-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wfx: add tracepoint "queues_stats"

It is useful to check which queue the driver choose to send to the
hardware.

Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200701150707.222985-5-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wfx: correctly retrieve vif ID from Tx confirmation

The device is able to send multiple Tx confirmations in the one reply.
In this case, there is only one vif identifier for all the
confirmations.

Unfortunately, to generate this kind of messages the device squashes all
the confirmations whatever their vif ID and use the vif ID of the first
confirmation. So, the driver cannot rely on the vif ID mentioned in the
header. Fortunately, using the packet_id, the driver can retrieve the Tx
request and the associated vif.

Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200701150707.222985-4-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wfx: check the vif ID of the Tx confirmations

When the driver has sent a frame on a virtual interface (vif), it
expects to receive the confirmation on the same vif.

This patch add a check for that.

Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200701150707.222985-3-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wfx: associate tx_queues to vifs

The device handles 4 queues (one per AC) for each virtual interface (and
maximum 4 virtual interfaces). Until now the driver unified the queue of
all interfaces and handled only 4 queues for whole device.

This architecture did not allow to balance the traffic between the vif. So,
this patch relocate the queues into the vif and change the API accordingly.

Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200701150707.222985-2-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: qlge: qlge_ethtool.c: Proper indentation.

Remove extra indentations from if-statement.

Signed-off-by: Suraj Upadhyay <usuraj35@gmail.com>
Link: https://lore.kernel.org/r/20200702084022.GA1586@blackclown
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8712: base_types: Remove unused macros

Those #define's are not used anywhere, get rid of them.

Signed-off-by: Mauro Dreissig <mukadr@gmail.com>
Link: https://lore.kernel.org/r/20200701214420.5566-3-mukadr@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8712: Replace FIELD_OFFSET() with offsetof()

Use the existing offsetof() macro instead of duplicating code.

Signed-off-by: Mauro Dreissig <mukadr@gmail.com>
Link: https://lore.kernel.org/r/20200701214420.5566-2-mukadr@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8712: use common ieee80211 constants

Many defined constants in wifi.h are unused and/or available from
<linux/ieee80211.h>, some with slightly different names. Use the
common ones, rename where necessary and remove unused.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200701164213.4205-1-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: Fix WARNINGs of Block comments

Fixed 5 WARNINGs of Block comments use * on subsequent lines.

Signed-off-by: Simon Fong <simon.fodin@gmail.com>
Link: https://lore.kernel.org/r/239f5c7f4761dd2ef0df8ee6966fca8ed2a13a9c.1593703689.git.simon.fodin@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: Fix CHECK of coding style

Fixed a CHECK of Lines should not end with a '('.

Signed-off-by: Simon Fong <simon.fodin@gmail.com>
Link: https://lore.kernel.org/r/9f1f9d7d4723aa8d9bc2d7149fd01aacc1191860.1593703689.git.simon.fodin@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: use common ieee80211 constants

Many defined constants in wifi.h are unused and/or available from
<linux/ieee80211.h>, some with slightly different names. Use the
common ones, rename where necessary and remove unused.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200701182957.7932-1-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: remove unused parameter

Remove unused parameter 'padapter' from rtw_os_xmit_resource_alloc().

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200701165459.8904-1-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wilc1000: remove obsolete TODO file

The movement of wilc1000 out of staging left an obsolete TODO file.

Remove that as it's no longer needed.

Cc: Kalle Valo <kvalo@codeaurora.org>
Cc: <Ajay.Kathat@microchip.com>
Cc: <Venkateswara.Kaja@microchip.com>
Cc: <Sripad.Balwadgi@microchip.com>
Cc: <Nicolas.Ferre@microchip.com>
Cc: <johannes@sipsolutions.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge branch 'wilc1000-move-out-of-staging' of git://git./linux/kernel/git/kvalo/wireless-drivers-next into staging-next

This is the movement of the wilc1000 driver out of staging, pulled in
here so that we do not end up doing duplicate work.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
* 'wilc1000-move-out-of-staging' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next:
  wilc1000: move wilc driver out of staging

staging: rtl8188eu: core: Fix coding style issue

Use %s and  __func__ in place of function names.
This solves following checkpatch.pl warning
WARNING: Prefer using '"%s...", __func__' to using 'rtw_do_join', this function's name, in a string

Signed-off-by: Puranjay Mohan <puranjay12@gmail.com>
Link: https://lore.kernel.org/r/20200701145457.9562-1-puranjay12@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: core: Fix WARNING of Block comments

Fixed a WARNING of Block comments use * on subsequent lines.

Signed-off-by: Simon Fong <simon.fodin@gmail.com>
Link: https://lore.kernel.org/r/20200630022228.GA580@simon-pc
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Move vchiq.h into include directory

To make the separation clear between vchiq's header files and vchiq.h,
which is to be used by services and is the 'public' API, move it into a
dedicated includes directory.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-48-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Move defines into core header

Those are only used in the core vchiq code, while present in vchiq's
'public' API header. Move them into the right place.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-47-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Use vchiq.h as the main header file for services

This used to be vchiq_if.h but vchiq.h is more concise for an include
file that will hopefully be in the future in the includes directory.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-46-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Move conditional barrier definition into vchiq_core.h

The barrier is only used by core code. So keep the barrier definition in
the core header.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-45-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Get rid of vchi

All the functions that vchi currently provides are a 1:1 mapping to its
vchiq counterparts. Get rid of vchi altogether and use vchiq's on all
services.

In the process also get rid of the vchi directory, as the only remaining
file was a TODO file, which now lives in the parent directory.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-44-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Move vchi_queue_kernel_message() into vchiq

We can't really merge it with vchiq_queue_message() as it has internal
users that will not benefit from the retry mechanism
vchiq_queue_kernel_message() uses. So, for the sake of getting rid of
vchi, move it into vchiq.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-43-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of vchi_bulk_queue_transmit()

Its vchiq counterpart, vchiq_bulk_transmit() is only used by vchi. We
can then merge both functions by moving vchi_bulk_queue_transmit()'s
retry mechanism into vchiq_bulk_transmit() and let services call the
later.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-42-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of vchi_bulk_queue_receive()

Its vchiq counterpart, vchiq_bulk_receive() is only used by vchi. We can
then merge both functions by moving vchi_bulk_queue_receive()'s retry
mechanism into vchiq_bulk_receive() and let services call the later.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-41-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Move definitions only used by core into core header

vchiq_if.h is used by service consumers and only the exported API should
be present. Move the rest of definitions, which were not exported by the
way, to vchiq_core.h where they belong.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-40-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Make vchiq_add_service() local

The function is being exported although there is no use for it outside
of vchiq's core code. Keep it local then.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-39-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Get rid of unnecessary definitions in vchiq_if.h

Those functions don't actually exist, nor have any use, nor the macros.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-38-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Don't include vchiq_core.h

This file is internal to vchiq.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-37-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of struct vchiq_instance forward declaration

It's already forward declared in vchiq_if.h. So no need to insist

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-36-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Unify fourcc definition mechanisms

There is already enough duplication with regard to fourcc generation in
the kernel. Let's at least use the same macro all over the vchiq
ecosystem.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-35-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Rework vchi_msg_hold() to match vchiq_msg_hold()

The services have access to struct vchiq_header's internals, so we can
let them get the data pointer. This pretty much makes both functions
exactly the same, which will allow us to make a switch to
vchiq_msg_hold() further down the road.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-34-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Pass vchiq's message when holding a message

vchi created an opaque structure to be held by services while they
process callback messages. The contents of this opaque structure are the
service handle, which all services already maintain, and a pointer to
vchiq's message structure, struct vchiq_header.

Let's get rid of this opaque structure and directly pass struct
vchiq_header, which is what ultimately vchiq consumes when handling
callback messages.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-33-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of struct vchi_service

The structure only contains a single parameter, which is the underlying
vchiq handle. Get rid of the struct and directly pass the handle around.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-32-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Use struct vchiq_service_params

For initialization, vchi has its own params structure, which is then
translated to vchiq's params structure. They are essentially the same,
so lets directly use vchiq's.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-31-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Don't use a typedef for vchiq_callback

Linux coding style says to avoid typdefs.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-30-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of vchiq_shim's message callback

As vchiq_shim's callback does nothing aside from pushing messages into
the service's queue, let's bypass it and jump directly to the service's
callbacks, letting them choose whether to use the message queue.

It turns out most services don't need to use the message queue, which
makes for simpler code in the end.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-29-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Export vchiq_msg_queue_push

vchiq consumer drivers may need to use this function in order to get the
benefits of vchiq's per service message queueing mechanism.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-28-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Export vchiq_get_service_userdata()

This is for service's callbacks to get their private data.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-27-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Expose struct vchi_service

This will make further changes easier. The struct will ultimately
disappear.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-26-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Get rid of vchiq_util.h

The header file only provides other includes. Move the relevant includes
to their respective C files and delete it for good.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-25-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Move message queue into struct vchiq_service

This has historically been handled by vchi, but there is no reason why
this couldn't be handled directly in vchiq.

The patch tries to avoid altering any behavior, with the exception of
the msg_queue size, which is now fixed to VCHIQ_MAX_SLOTS (it was set to
VCHIQ_MAX_SLOTS / 2). This is done to match vchiq's user_service message
queue, which could be merged with this one in the future.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-24-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq: Introduce vchiq_validate_params()

When adding a new service validate the configuration parameters
provided, and remove unnecessary checks in vchi, now that we have
validated service's config.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-23-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of effect less expression

It was probably there to trick compilers into ignoring unused variables,
which isn't needed in Linux.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-22-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Use vchiq's enum vchiq_reason

enum vchi_callback_reason maps 1:1 to enum vchiq_reason, in an effort to
simplify things, let's use the later, and get rid of the extra
indirection.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-21-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Use enum vchiq_bulk_mode instead of vchi's transmission flags

vchi has a set of transfer flags which almost map 1:1 to vchiq's own
transfer modes. For the sake of simplicity let's use the later and
delete vchi's.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-20-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of flags argument in vchi_msg_hold()

All users are ignoring the flags argument. So for the sake of
simplicity delete it.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-19-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vc04_services: Get rid of vchi_cfg.h

Nothing in it is being used.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-18-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of unnecessary defines

Those defines aren't used by anyone. Get rid of them.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-17-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi_common: Get rid of all unused definitions

There is a series of structures and enums defined but never used. Get
rid of them.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-16-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of vchi_msg_dequeue()

Nobody uses it. Get rid of it.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-15-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vc04_services: bcm2835-audio: Use vchi_msg_hold()

vchi_msg_dequeue() provides the same functionality as vchi_msg_hold()
except it copies the message data as opposed to the later which provides
the data in place.

The copying is done on a local variable, so there is no need to keep the
message out the function's bounds, so use vchi_msg_hold() instead.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-14-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Unify struct shim_service and struct vchi_service_handle

The idea behind struct vchi_service_handle is to create an opaque handle
to struct shim_service. This can be achieved by doing a forward
declaration of struct shim_service, which will avoid unwarranted casts
and pointer play.

Ultimately as a rename is due all over the vchi user space, rename
struct shim_service into struvt vchi_service, which is more consistent
with the rest of the exposed API.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-13-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of struct vchi_instance_handle

The idea behind this was to create an opaque handle to struct
vchiq_instance. This can be achieved without creating a new type by
means of a forward declaration of struct vchiq_instance.

This saves us from a lot of useless casting and overall simplifies code.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-12-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of vchi_msg_peek()

There already is a function that covers most of the functionality
vchi_msg_peek() provides: vchi_msg_hold(). The main difference being
that the later removes the message from vchu's queue while the other
does it later on, while releasing the message.

There are no users of this function that can't be trivially converted to
vchi_msg_hold(). So, for the sake of removing duplicate code, get rid of
vchi_msg_peek().

Note that the opposite change could be performed as well. But
vchi_msg_peek()'s implementation was deemed less robust as messages have
to be released in order.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-11-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchi: Get rid of all useless callback reasons

They are neither produced nor expected, so just delete them.

Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-10-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vchiq_arm: Add a matching unregister call

All the registered children of vchiq have a corresponding call to
platform_device_unregister except bcm2835_audio. Fix that.

Fixes: 25c7597af20d ("staging: vchiq_arm: Register a platform device for audio")

Signed-off-by: Phil Elwell <phil@raspberrypi.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-9-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mmal-vchiq: Fix formatting errors in mmal_parameters.h

No functional changes in this commit.

- Remove erroneous whitespace.
- Remove _t postfix label on structs and enums.

Signed-off-by: Naushir Patuck <naush@raspberrypi.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-8-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mmal-vchiq: If the VPU returns an error, don't negate it

There is an enum for the errors that the VPU can return.
port_parameter_get was negating that value, but also using -EINVAL
from the Linux error codes.
Pass the VPU error code as positive values. Should the function
need to pass a Linux failure, then return that as negative.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-7-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mmal-vchiq: Always return the param size from param_get

mmal-vchiq is a reimplementation of the userland library for MMAL.
When getting a parameter, the client provides the storage and
the size of the storage. The VPU then returns the size of the
parameter that it wished to return, and as much as possible of
that parameter is returned to the client.

The implementation previously only returned the size provided
by the VPU should it exceed the buffer size. So for parameters
such as the supported encodings list the client had no idea
how much of the provided storage had been populated.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-6-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mmal-vchiq: Fix client_component for 64 bit kernel

The MMAL client_component field is used with the event
mechanism to allow the client to identify the component for
which the event is generated.
The field is only 32bits in size, therefore we can't use a
pointer to the component in a 64 bit kernel.

Component handles are already held in an array per VCHI
instance, so use the array index as the client_component handle
to avoid having to create a new IDR for this purpose.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-5-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mmal-vchiq: Fixup vchiq-mmal include ordering

There were dependencies on including the headers in the correct
order. Fix up the headers so that they include the other
headers that they depend on themselves.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-4-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mmal-vchiq: Make a mmal_buf struct for passing parameters

The callback from vchi_mmal to the client was growing lots of extra
parameters. Consolidate them into a single struct instead of
growing the list further.
The struct is associated with the client buffer, therefore there
are various changes to setup various containers for the struct,
and pass the appropriate members.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-3-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mmal-vchiq: Make timeout a defined parameter

The timeout period for VPU communications is a useful thing
to extend when debugging.
Set it via a define, rather than a magic number buried in the code.

Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.org>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Link: https://lore.kernel.org/r/20200629150945.10720-2-nsaenzjulienne@suse.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: kpc2000: kpc_dma: Remove additional goto statements

As 3 goto level referring to same common code, those can be
accomodated with a single goto level and renameing it to
unpin_pages. Set the -ERRNO when returning partial mapped
pages in more appropriate place.

When dma_map_sg() failed, the previously allocated memory was
not freed properly. This is corrected now.

Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Bharath Vedartham <linux.bhar@gmail.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/1593584264-16982-5-git-send-email-jrdr.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: kpc2000: kpc_dma: Convert get_user_pages() --> pin_user_pages()

In 2019, we introduced pin_user_pages*() and now we are converting
get_user_pages*() to the new API as appropriate. [1] & [2] could
be referred for more information. This is case 2 as per document [1].

[1] Documentation/core-api/pin_user_pages.rst

[2] "Explicit pinning of user-space pages":
        https://lwn.net/Articles/807108/

Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Bharath Vedartham <linux.bhar@gmail.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/1593584264-16982-4-git-send-email-jrdr.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: kpc2000: kpc_dma: Convert set_page_dirty() --> set_page_dirty_lock()

First, convert set_page_dirty() to set_page_dirty_lock()

Second, there is an interval in there after set_page_dirty() and
before put_page(), in which the device could be running and setting
pages dirty. Moving set_page_dirty_lock() after dma_unmap_sg().

Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
Suggested-by: John Hubbard <jhubbard@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Bharath Vedartham <linux.bhar@gmail.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/1593584264-16982-3-git-send-email-jrdr.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: kpc2000: kpc_dma: Unpin partial pinned pages

There is a bug, when get_user_pages() failed but partially pinned
pages are not unpinned and positive numbers are returned instead of
-ERRNO. Fixed it.

Also, int is more appropriate type for rv. Changed it.

Signed-off-by: Souptick Joarder <jrdr.linux@gmail.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Cc: Bharath Vedartham <linux.bhar@gmail.com>
Link: https://lore.kernel.org/r/1593584264-16982-2-git-send-email-jrdr.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: qlge: qlge.h: remove unnecessary spaces

Spaces aren't necessary before function pointer arguments.

Signed-off-by: Diego Bienz <diego.bienz@gmail.com>
Link: https://lore.kernel.org/r/20200630113324.GA10534@diegob-nb
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

clk: staging: Specify IOMEM dependency for Xilinx Clocking Wizard driver

The Xilinx Clocking Wizard driver uses the devm_ioremap_resource
function, but does not specify a dependency on IOMEM in Kconfig. This
causes a build failure on architectures without IOMEM, for example, UML
(notably with make allyesconfig).

Fix this by making CONFIG_COMMON_CLK_XLNX_CLKWZRD depend on CONFIG_IOMEM.

Signed-off-by: David Gow <davidgow@google.com>
Link: https://lore.kernel.org/r/20200630044518.1084468-1-davidgow@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: Replace function name with __func__

Fix the following checkpatch warning:
WARNING: Prefer using '"%s...", __func__' to using 'rtw_get_bcn_info', this function's name, in a string

Signed-off-by: Brooke Basile <brookebasile@gmail.com>
Link: https://lore.kernel.org/r/20200629173711.5158-1-brookebasile@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8712: remove some ieee80211 constants from wifi.h

The constants IEEE80211_BAR_CTRL_ACK_POLICY_NORMAL and
IEEE80211_BAR_CTRL_CBMTID_COMPRESSED_BA are not used in the driver
code and, if ever needed, available from 'include/linux/ieee80211.h'.
Remove them from wifi.h.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200629171405.7711-3-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723bs: remove some ieee80211 constants from wifi.h

The constants IEEE80211_BAR_CTRL_ACK_POLICY_NORMAL and
IEEE80211_BAR_CTRL_CBMTID_COMPRESSED_BA are not used in the driver
code and, if ever needed, available from 'include/linux/ieee80211.h'.
Remove them from wifi.h.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200629171405.7711-2-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: remove some ieee80211 constants from wifi.h

The constants IEEE80211_BAR_CTRL_ACK_POLICY_NORMAL and
IEEE80211_BAR_CTRL_CBMTID_COMPRESSED_BA are not used in the driver
code and, if ever needed, available from 'include/linux/ieee80211.h'.
Remove them from wifi.h.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200629171405.7711-1-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: add spaces around operators

Add spaces around operators to improve readability and clear
checkpatch issues.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200629161255.28371-2-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: use compound assignment operators

Use compound assignment operators to simplify the code and clear
missing spaces around operators checkpatch issues.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Link: https://lore.kernel.org/r/20200629161255.28371-1-straube.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: qlge/qlge_main.c: use generic power management

Drivers should not use legacy power management as they have to manage power
states and related operations, for the device, themselves. This driver was
handling them with the help of PCI helper functions like
pci_save/restore_state(), pci_enable/disable_device(), etc.

With generic PM, all essentials will be handled by the PCI core. Driver
needs to do only device-specific operations.

The driver was also using pci_enable_wake(...,..., 0) to disable wake. Use
device_wakeup_disable() instead.

Compile-tested only.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Link: https://lore.kernel.org/r/20200629173459.262075-2-vaibhavgupta40@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: vt6655/device_main.c: use generic power management

Drivers should not use legacy power management as they have to manage power
states and related operations, for the device, themselves. This driver was
handling them with the help of PCI helper functions like
pci_save/restore_state(), pci_enable/disable_device(), etc.

With generic PM, all essentials will be handled by the PCI core. Driver
needs to do only device-specific operations.

The driver was also using pci_enable_wake(...,..., 0) to disable wake. Use
device_wakeup_disable() instead.

Compile-tested only.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Link: https://lore.kernel.org/r/20200629082819.216405-5-vaibhavgupta40@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rts5208/rtsx.c: use generic power management

Drivers should not use legacy power management as they have to manage power
states and related operations, for the device, themselves. This driver was
handling them with the help of PCI helper functions like
pci_save/restore_state(), pci_enable/disable_device(), etc.

With generic PM, all essentials will be handled by the PCI core. Driver
needs to do only device-specific operations.

The driver was also using pci_enable_wake(...,..., 0) to disable wake. Use
device_wakeup_disable() instead.

Compile-tested only.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Link: https://lore.kernel.org/r/20200629082819.216405-4-vaibhavgupta40@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8192e: use generic power management

The structure of working of PM hooks for source files is:
    drivers/staging/rtl8192e/rtl8192e/rtl_pm.h   : callbacks declared
    drivers/staging/rtl8192e/rtl8192e/rtl_pm.c   : callbacks defined
    drivers/staging/rtl8192e/rtl8192e/rtl_core.c : callbacks used

Drivers should not use legacy power management as they have to manage power
states and related operations, for the device, themselves. This driver was
handling them with the help of PCI helper functions like
pci_save/restore_state(), pci_enable/disable_device(), etc.

With generic PM, all essentials will be handled by the PCI core. Driver
needs to do only device-specific operations.

The driver was also using pci_enable_wake(...,..., 0) to disable wake. Use
device_wakeup_disable() instead. Use device_set_wakeup_enable() where WOL
is decided by the value of a variable during runtime.

Compile-tested only.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Link: https://lore.kernel.org/r/20200629082819.216405-3-vaibhavgupta40@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: include: odm.h: fixed a blank space coding style issue.

added blank space and enclosed a complex valued macro within parentheses for improved code readability.

Signed-off-by: B K Karthik <karthik.bk2000@live.com>
Link: https://lore.kernel.org/r/20200629080748.l4ufcpuk4cg2m725@pesu-pes-edu
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: ks7010: fix ks_wlan_start_xmit()'s return type

The method ndo_start_xmit() is defined as returning an 'netdev_tx_t',
which is a typedef for an enum type, but the implementation in this
driver returns an 'int'.

Fix this by returning 'netdev_tx_t' in this driver too and
usind 'NETDEV_TX_OK' instead of 0 accordingly.

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Link: https://lore.kernel.org/r/20200628183926.74908-1-luc.vanoostenryck@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: include: hal8188e_rate_adaptive.h: fixed a blank space coding style issue

added blank space around arithmetic operators to improve readability

Signed-off-by: B K Karthik <karthik.bk2000@live.com>
Link: https://lore.kernel.org/r/20200629053442.eob7oixlxbs5rh33@pesu-pes-edu
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge 5.8-rc3 into staging-next

We want the staging fixes in here as well.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Linux 5.8-rc3

Merge tag 'arm-omap-fixes-5.8-1' of git://git./linux/kernel/git/soc/soc

Pull ARM OMAP fixes from Arnd Bergmann:
 "The OMAP developers are particularly active at hunting down
  regressions, so this is a separate branch with OMAP specific
  fixes for v5.8:

  As Tony explains
    "The recent display subsystem (DSS) related platform data changes
     caused display related regressions for suspend and resume. Looks
     like I only tested suspend and resume before dropping the legacy
     platform data, and forgot to test it after dropping it. Turns out
     the main issue was that we no longer have platform code calling
     pm_runtime_suspend for DSS like we did for the legacy platform data
     case, and that fix is still being discussed on the dri-devel list
     and will get merged separately. The DSS related testing exposed a
     pile other other display related issues that also need fixing
     though":

   - Fix ti-sysc optional clock handling and reset status checks for
     devices that reset automatically in idle like DSS

   - Ignore ti-sysc clockactivity bit unless separately requested to
     avoid unexpected performance issues

   - Init ti-sysc framedonetv_irq to true and disable for am4

   - Avoid duplicate DSS reset for legacy mode with dts data

   - Remove LCD timings for am4 as they cause warnings now that we're
     using generic panels

  Other OMAP changes from Tony include:

   - Fix omap_prm reset deassert as we still have drivers setting the
     pm_runtime_irq_safe() flag

   - Flush posted write for ti-sysc enable and disable

   - Fix droid4 spi related errors with spi flags

   - Fix am335x USB range and a typo for softreset

   - Fix dra7 timer nodes for clocks for IPU and DSP

   - Drop duplicate mailboxes after mismerge for dra7

   - Prevent pocketgeagle header line signal from accidentally setting
     micro-SD write protection signal by removing the default mux

   - Fix NFSroot flakeyness after resume for duover by switching the
     smsc911x gpio interrupt to back to level sensitive

   - Fix regression for omap4 clockevent source after recent system
     timer changes

   - Yet another ethernet regression fix for the "rgmii" vs "rgmii-rxid"
     phy-mode

   - One patch to convert am3/am4 DT files to use the regular sdhci-omap
     driver instead of the old hsmmc driver, this was meant for the
     merge window but got lost in the process"

* tag 'arm-omap-fixes-5.8-1' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (21 commits)
  ARM: dts: am5729: beaglebone-ai: fix rgmii phy-mode
  ARM: dts: Fix omap4 system timer source clocks
  ARM: dts: Fix duovero smsc interrupt for suspend
  ARM: dts: am335x-pocketbeagle: Fix mmc0 Write Protect
  Revert "bus: ti-sysc: Increase max softreset wait"
  ARM: dts: am437x-epos-evm: remove lcd timings
  ARM: dts: am437x-gp-evm: remove lcd timings
  ARM: dts: am437x-sk-evm: remove lcd timings
  ARM: dts: dra7-evm-common: Fix duplicate mailbox nodes
  ARM: dts: dra7: Fix timer nodes properly for timer_sys_ck clocks
  ARM: dts: Fix am33xx.dtsi ti,sysc-mask wrong softreset flag
  ARM: dts: Fix am33xx.dtsi USB ranges length
  bus: ti-sysc: Increase max softreset wait
  ARM: OMAP2+: Fix legacy mode dss_reset
  bus: ti-sysc: Fix uninitialized framedonetv_irq
  bus: ti-sysc: Ignore clockactivity unless specified as a quirk
  bus: ti-sysc: Use optional clocks on for enable and wait for softreset bit
  ARM: dts: omap4-droid4: Fix spi configuration and increase rate
  bus: ti-sysc: Flush posted write on enable and disable
  soc: ti: omap-prm: use atomic iopoll instead of sleeping one
  ...

Merge tag 'arm-fixes-5.8-1' of git://git./linux/kernel/git/soc/soc

Pull ARM SoC fixes from Arnd Bergmann:
 "Here are a couple of bug fixes, mostly for devicetree files

  NXP i.MX:
   - Use correct voltage on some i.MX8M board device trees to avoid
     hardware damage
   - Code fixes for a compiler warning and incorrect reference counting,
     both harmless.
   - Fix the i.MX8M SoC driver to correctly identify imx8mp
   - Fix watchdog configuration in imx6ul-kontron device tree.

  Broadcom:
   - A small regression fix for the Raspberry-Pi firmware driver
   - A Kconfig change to use the correct timer driver on Northstar
   - A DT fix for the Luxul XWC-2000 machine
   - Two more DT fixes for NSP SoCs

  STmicroelectronics STI
   - Revert one broken patch for L2 cache configuration

  ARM Versatile Express:
   - Fix a regression by reverting a broken DT cleanup

  TEE drivers:
   - MAINTAINERS: change tee mailing list"

* tag 'arm-fixes-5.8-1' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:
  Revert "ARM: sti: Implement dummy L2 cache's write_sec"
  soc: imx8m: fix build warning
  ARM: imx6: add missing put_device() call in imx6q_suspend_init()
  ARM: imx5: add missing put_device() call in imx_suspend_alloc_ocram()
  soc: imx8m: Correct i.MX8MP UID fuse offset
  ARM: dts: imx6ul-kontron: Change WDOG_ANY signal from push-pull to open-drain
  ARM: dts: imx6ul-kontron: Move watchdog from Kontron i.MX6UL/ULL board to SoM
  arm64: dts: imx8mm-beacon: Fix voltages on LDO1 and LDO2
  arm64: dts: imx8mn-ddr4-evk: correct ldo1/ldo2 voltage range
  arm64: dts: imx8mm-evk: correct ldo1/ldo2 voltage range
  ARM: dts: NSP: Correct FA2 mailbox node
  ARM: bcm2835: Fix integer overflow in rpi_firmware_print_firmware_revision()
  MAINTAINERS: change tee mailing list
  ARM: dts: NSP: Disable PL330 by default, add dma-coherent property
  ARM: bcm: Select ARM_TIMER_SP804 for ARCH_BCM_NSP
  ARM: dts: BCM5301X: Add missing memory "device_type" for Luxul XWC-2000
  arm: dts: vexpress: Move mcc node back into motherboard node

Merge tag 'timers-urgent-2020-06-28' of git://git./linux/kernel/git/tip/tip

Pull timer fix from Ingo Molnar:
 "A single DocBook fix"

* tag 'timers-urgent-2020-06-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  timekeeping: Fix kerneldoc system_device_crosststamp & al

Merge tag 'perf-urgent-2020-06-28' of git://git./linux/kernel/git/tip/tip

Pull perf fix from Ingo Molnar:
 "A single Kbuild dependency fix"

* tag 'perf-urgent-2020-06-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/rapl: Fix RAPL config variable bug

Merge tag 'efi-urgent-2020-06-28' of git://git./linux/kernel/git/tip/tip

Pull EFI fixes from Ingo Molnar:

 - Fix build regression on v4.8 and older

 - Robustness fix for TPM log parsing code

 - kobject refcount fix for the ESRT parsing code

 - Two efivarfs fixes to make it behave more like an ordinary file
   system

 - Style fixup for zero length arrays

 - Fix a regression in path separator handling in the initrd loader

 - Fix a missing prototype warning

 - Add some kerneldoc headers for newly introduced stub routines

 - Allow support for SSDT overrides via EFI variables to be disabled

 - Report CPU mode and MMU state upon entry for 32-bit ARM

 - Use the correct stack pointer alignment when entering from mixed mode

* tag 'efi-urgent-2020-06-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  efi/libstub: arm: Print CPU boot mode and MMU state at boot
  efi/libstub: arm: Omit arch specific config table matching array on arm64
  efi/x86: Setup stack correctly for efi_pe_entry
  efi: Make it possible to disable efivar_ssdt entirely
  efi/libstub: Descriptions for stub helper functions
  efi/libstub: Fix path separator regression
  efi/libstub: Fix missing-prototype warning for skip_spaces()
  efi: Replace zero-length array and use struct_size() helper
  efivarfs: Don't return -EINTR when rate-limiting reads
  efivarfs: Update inode modification time for successful writes
  efi/esrt: Fix reference count leak in esre_create_sysfs_entry.
  efi/tpm: Verify event log header before parsing
  efi/x86: Fix build with gcc 4

Merge tag 'sched_urgent_for_5.8_rc3' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Borislav Petkov:
 "The most anticipated fix in this pull request is probably the horrible
  build fix for the RANDSTRUCT fail that didn't make -rc2. Also included
  is the cleanup that removes those BUILD_BUG_ON()s and replaces it with
  ugly unions.

  Also included is the try_to_wake_up() race fix that was first
  triggered by Paul's RCU-torture runs, but was independently hit by
  Dave Chinner's fstest runs as well"

* tag 'sched_urgent_for_5.8_rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/cfs: change initial value of runnable_avg
  smp, irq_work: Continue smp_call_function*() and irq_work*() integration
  sched/core: s/WF_ON_RQ/WQ_ON_CPU/
  sched/core: Fix ttwu() race
  sched/core: Fix PI boosting between RT and DEADLINE tasks
  sched/deadline: Initialize ->dl_boosted
  sched/core: Check cpus_mask, not cpus_ptr in __set_cpus_allowed_ptr(), to fix mask corruption
  sched/core: Fix CONFIG_GCC_PLUGIN_RANDSTRUCT build fail

Merge tag 'x86_urgent_for_5.8_rc3' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Borislav Petkov:

 - AMD Memory bandwidth counter width fix, by Babu Moger.

 - Use the proper length type in the 32-bit truncate() syscall variant,
   by Jiri Slaby.

 - Reinit IA32_FEAT_CTL during wakeup to fix the case where after
   resume, VMXON would #GP due to VMX not being properly enabled, by
   Sean Christopherson.

 - Fix a static checker warning in the resctrl code, by Dan Carpenter.

 - Add a CR4 pinning mask for bits which cannot change after boot, by
   Kees Cook.

 - Align the start of the loop of __clear_user() to 16 bytes, to improve
   performance on AMD zen1 and zen2 microarchitectures, by Matt Fleming.

* tag 'x86_urgent_for_5.8_rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/asm/64: Align start of __clear_user() loop to 16-bytes
  x86/cpu: Use pinning mask for CR4 bits needing to be 0
  x86/resctrl: Fix a NULL vs IS_ERR() static checker warning in rdt_cdp_peer_get()
  x86/cpu: Reinitialize IA32_FEAT_CTL MSR on BSP during wakeup
  syscalls: Fix offset type of ksys_ftruncate()
  x86/resctrl: Fix memory bandwidth counter width for AMD

Merge tag 'rcu_urgent_for_5.8_rc3' of git://git./linux/kernel/git/tip/tip

Pull RCU-vs-KCSAN fixes from Borislav Petkov:
 "A single commit that uses "arch_" atomic operations to avoid the
  instrumentation that comes with the non-"arch_" versions.

  In preparation for that commit, it also has another commit that makes
  these "arch_" atomic operations available to generic code.

  Without these commits, KCSAN uses can see pointless errors"

* tag 'rcu_urgent_for_5.8_rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  rcu: Fixup noinstr warnings
  locking/atomics: Provide the arch_atomic_ interface to generic code

Merge tag 'objtool_urgent_for_5.8_rc3' of git://git./linux/kernel/git/tip/tip

Pull objtool fixes from Borislav Petkov:
 "Three fixes from Peter Zijlstra suppressing KCOV instrumentation in
  noinstr sections.

  Peter Zijlstra says:
    "Address KCOV vs noinstr. There is no function attribute to
     selectively suppress KCOV instrumentation, instead teach objtool
     to NOP out the calls in noinstr functions"

  This cures a bunch of KCOV crashes (as used by syzcaller)"

* tag 'objtool_urgent_for_5.8_rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  objtool: Fix noinstr vs KCOV
  objtool: Provide elf_write_{insn,reloc}()
  objtool: Clean up elf_write() condition

Merge tag 'x86_entry_for_5.8' of git://git./linux/kernel/git/tip/tip

Pull x86 entry fixes from Borislav Petkov:
 "This is the x86/entry urgent pile which has accumulated since the
  merge window.

  It is not the smallest but considering the almost complete entry core
  rewrite, the amount of fixes to follow is somewhat higher than usual,
  which is to be expected.

  Peter Zijlstra says:
   'These patches address a number of instrumentation issues that were
    found after the x86/entry overhaul. When combined with rcu/urgent
    and objtool/urgent, these patches make UBSAN/KASAN/KCSAN happy
    again.

    Part of making this all work is bumping the minimum GCC version for
    KASAN builds to gcc-8.3, the reason for this is that the
    __no_sanitize_address function attribute is broken in GCC releases
    before that.

    No known GCC version has a working __no_sanitize_undefined, however
    because the only noinstr violation that results from this happens
    when an UB is found, we treat it like WARN. That is, we allow it to
    violate the noinstr rules in order to get the warning out'"

* tag 'x86_entry_for_5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/entry: Fix #UD vs WARN more
  x86/entry: Increase entry_stack size to a full page
  x86/entry: Fixup bad_iret vs noinstr
  objtool: Don't consider vmlinux a C-file
  kasan: Fix required compiler version
  compiler_attributes.h: Support no_sanitize_undefined check with GCC 4
  x86/entry, bug: Comment the instrumentation_begin() usage for WARN()
  x86/entry, ubsan, objtool: Whitelist __ubsan_handle_*()
  x86/entry, cpumask: Provide non-instrumented variant of cpu_is_offline()
  compiler_types.h: Add __no_sanitize_{address,undefined} to noinstr
  kasan: Bump required compiler version
  x86, kcsan: Add __no_kcsan to noinstr
  kcsan: Remove __no_kcsan_or_inline
  x86, kcsan: Remove __no_kcsan_or_inline usage

sched/cfs: change initial value of runnable_avg

Some performance regression on reaim benchmark have been raised with
  commit 070f5e860ee2 ("sched/fair: Take into account runnable_avg to classify group")

The problem comes from the init value of runnable_avg which is initialized
with max value. This can be a problem if the newly forked task is finally
a short task because the group of CPUs is wrongly set to overloaded and
tasks are pulled less agressively.

Set initial value of runnable_avg equals to util_avg to reflect that there
is no waiting time so far.

Fixes: 070f5e860ee2 ("sched/fair: Take into account runnable_avg to classify group")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200624154422.29166-1-vincent.guittot@linaro.org

smp, irq_work: Continue smp_call_function*() and irq_work*() integration

Instead of relying on BUG_ON() to ensure the various data structures
line up, use a bunch of horrible unions to make it all automatic.

Much of the union magic is to ensure irq_work and smp_call_function do
not (yet) see the members of their respective data structures change
name.

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lkml.kernel.org/r/20200622100825.844455025@infradead.org

sched/core: s/WF_ON_RQ/WQ_ON_CPU/

Use a better name for this poorly named flag, to avoid confusion...

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Mel Gorman <mgorman@suse.de>
Link: https://lkml.kernel.org/r/20200622100825.785115830@infradead.org

sched/core: Fix ttwu() race

Paul reported rcutorture occasionally hitting a NULL deref:

  sched_ttwu_pending()
    ttwu_do_wakeup()
      check_preempt_curr() := check_preempt_wakeup()
        find_matching_se()
          is_same_group()
            if (se->cfs_rq == pse->cfs_rq) <-- *BOOM*

Debugging showed that this only appears to happen when we take the new
code-path from commit:

  2ebb17717550 ("sched/core: Offload wakee task activation if it the wakee is descheduling")

and only when @cpu == smp_processor_id(). Something which should not
be possible, because p->on_cpu can only be true for remote tasks.
Similarly, without the new code-path from commit:

  c6e7bd7afaeb ("sched/core: Optimize ttwu() spinning on p->on_cpu")

this would've unconditionally hit:

  smp_cond_load_acquire(&p->on_cpu, !VAL);

and if: 'cpu == smp_processor_id() && p->on_cpu' is possible, this
would result in an instant live-lock (with IRQs disabled), something
that hasn't been reported.

The NULL deref can be explained however if the task_cpu(p) load at the
beginning of try_to_wake_up() returns an old value, and this old value
happens to be smp_processor_id(). Further assume that the p->on_cpu
load accurately returns 1, it really is still running, just not here.

Then, when we enqueue the task locally, we can crash in exactly the
observed manner because p->se.cfs_rq != rq->cfs_rq, because p's cfs_rq
is from the wrong CPU, therefore we'll iterate into the non-existant
parents and NULL deref.

The closest semi-plausible scenario I've managed to contrive is
somewhat elaborate (then again, actual reproduction takes many CPU
hours of rcutorture, so it can't be anything obvious):

X->cpu = 1
rq(1)->curr = X

CPU0 CPU1 CPU2

// switch away from X
LOCK rq(1)->lock
smp_mb__after_spinlock
dequeue_task(X)
  X->on_rq = 9
switch_to(Z)
  X->on_cpu = 0
UNLOCK rq(1)->lock

// migrate X to cpu 0
LOCK rq(1)->lock
dequeue_task(X)
set_task_cpu(X, 0)
  X->cpu = 0
UNLOCK rq(1)->lock

LOCK rq(0)->lock
enqueue_task(X)
  X->on_rq = 1
UNLOCK rq(0)->lock

// switch to X
LOCK rq(0)->lock
smp_mb__after_spinlock
switch_to(X)
  X->on_cpu = 1
UNLOCK rq(0)->lock

// X goes sleep
X->state = TASK_UNINTERRUPTIBLE
smp_mb(); // wake X
ttwu()
  LOCK X->pi_lock
  smp_mb__after_spinlock

  if (p->state)

  cpu = X->cpu; // =? 1

  smp_rmb()

// X calls schedule()
LOCK rq(0)->lock
smp_mb__after_spinlock
dequeue_task(X)
  X->on_rq = 0

  if (p->on_rq)

  smp_rmb();

  if (p->on_cpu && ttwu_queue_wakelist(..)) [*]

  smp_cond_load_acquire(&p->on_cpu, !VAL)

  cpu = select_task_rq(X, X->wake_cpu, ...)
  if (X->cpu != cpu)
switch_to(Y)
  X->on_cpu = 0
UNLOCK rq(0)->lock

However I'm having trouble convincing myself that's actually possible
on x86_64 -- after all, every LOCK implies an smp_mb() there, so if ttwu
observes ->state != RUNNING, it must also observe ->cpu != 1.

(Most of the previous ttwu() races were found on very large PowerPC)

Nevertheless, this fully explains the observed failure case.

Fix it by ordering the task_cpu(p) load after the p->on_cpu load,
which is easy since nothing actually uses @cpu before this.

Fixes: c6e7bd7afaeb ("sched/core: Optimize ttwu() spinning on p->on_cpu")
Reported-by: Paul E. McKenney <paulmck@kernel.org>
Tested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lkml.kernel.org/r/20200622125649.GC576871@hirez.programming.kicks-ass.net

sched/core: Fix PI boosting between RT and DEADLINE tasks

syzbot reported the following warning:

 WARNING: CPU: 1 PID: 6351 at kernel/sched/deadline.c:628
 enqueue_task_dl+0x22da/0x38a0 kernel/sched/deadline.c:1504

At deadline.c:628 we have:

 623 static inline void setup_new_dl_entity(struct sched_dl_entity *dl_se)
 624 {
 625  struct dl_rq *dl_rq = dl_rq_of_se(dl_se);
 626  struct rq *rq = rq_of_dl_rq(dl_rq);
 627
 628  WARN_ON(dl_se->dl_boosted);
 629  WARN_ON(dl_time_before(rq_clock(rq), dl_se->deadline));
        [...]
     }

Which means that setup_new_dl_entity() has been called on a task
currently boosted. This shouldn't happen though, as setup_new_dl_entity()
is only called when the 'dynamic' deadline of the new entity
is in the past w.r.t. rq_clock and boosted tasks shouldn't verify this
condition.

Digging through the PI code I noticed that what above might in fact happen
if an RT tasks blocks on an rt_mutex hold by a DEADLINE task. In the
first branch of boosting conditions we check only if a pi_task 'dynamic'
deadline is earlier than mutex holder's and in this case we set mutex
holder to be dl_boosted. However, since RT 'dynamic' deadlines are only
initialized if such tasks get boosted at some point (or if they become
DEADLINE of course), in general RT 'dynamic' deadlines are usually equal
to 0 and this verifies the aforementioned condition.

Fix it by checking that the potential donor task is actually (even if
temporary because in turn boosted) running at DEADLINE priority before
using its 'dynamic' deadline value.

Fixes: 2d3d891d3344 ("sched/deadline: Add SCHED_DEADLINE inheritance logic")
Reported-by: syzbot+119ba87189432ead09b4@syzkaller.appspotmail.com
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Daniel Bristot de Oliveira <bristot@redhat.com>
Tested-by: Daniel Wagner <dwagner@suse.de>
Link: https://lkml.kernel.org/r/20181119153201.GB2119@localhost.localdomain