platform/kernel/linux-rpi.git
7 years ago[media] media: ti-vpe: vpe: Make sure frame size dont exceed scaler capacity
Benoit Parrot [Fri, 18 Nov 2016 23:20:41 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Make sure frame size dont exceed scaler capacity

When scaler is to be used we need to make sure that the input and
output frame size do not exceed the maximum frame sizes that the
scaler h/w can handle otherwise streaming stall as the scaler
cannot proceed.

The scaler buffer is limited to 2047 pixels (i.e. 11 bits) when
attempting anything larger (2048 for example) the scaler stalls.

Realistically in an mem2mem device we can only check for this type
of issue when start_streaming is called. We can't do it during the
try_fmt/s_fmt because we do not have all of the info needed at that
point. So instead when start_streaming is called we need to check
that the input and output frames size do not exceed the scaler's
capability. The only time larger frame size are allowed is when
the input frame szie is the same as the output frame size.

Now in the case where we need to fail, start_streaming must return
all previously queued buffer back otherwise the vb2 framework
will issue kernel WARN messages.
In this case we also give an error message.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: scaler: Add debug support for multi-instance
Benoit Parrot [Fri, 18 Nov 2016 23:20:40 +0000 (21:20 -0200)]
[media] media: ti-vpe: scaler: Add debug support for multi-instance

Since there might be more then one instance it is better to
show the base address when dumping registers to help
with debugging.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: Make scaler library into its own module
Benoit Parrot [Fri, 18 Nov 2016 23:20:39 +0000 (21:20 -0200)]
[media] media: ti-vpe: Make scaler library into its own module

In preparation to add scaler support into VIP we need to
turn sc.c into its own kernel module.

Add support for multiple SC memory block as VIP contains
2 scaler instances.
This is done by passing the resource name to sc_create() and
modify the vpe invocation accordingly.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Enable DMABUF export
Benoit Parrot [Fri, 18 Nov 2016 23:20:38 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Enable DMABUF export

Allow VPE to be able to export DMA buffer.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Fix line stride for output motion vector
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:37 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Fix line stride for output motion vector

For deinterlacing operation, VPE hardware uses motion vectors.
MV calculated in the previous iteration are used for next interation.
Therefore driver allocates two motion vectors in ping-pong fashion.

For every transaction, one MV is DMAed in and one is DMAed out.
All the outbound DMAs (DMA to memory) use output parameters, but as
the motion vectors is generated purely out of input fields, it should
use the input parameters for DMA.

Fix the add_out_dtd to use source q_data for creating descriptor.
If the output size is greater than input stride, without this change,
MV DMA may overwrite the buffer causing memory corruption.

This CRITICAL fix ensures that the motion vector DMA descriptor is
created based on the attributes with which the buffer was allocated.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Ravikumar Kattekola <rk@ti.com>
Signed-off-by: Ravi Babu <ravibabu@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Use bidirectional cached buffers
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:36 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Use bidirectional cached buffers

VPDMA buffer will be used by CPU as well as by the VPDMA.
CPU will write/update the VPDMA descriptors containing data
about the video buffer DMA addresses.
VPDMA will write the "write descriptor" containing the
data about the DMA operation.

When mapping/unmapping the buffer, driver has to take care of
WriteBack and invalidation of the cache so that all the
coherency is maintained from both directions.

Use DMA_BIDIRECTIONAL to maintain coherency between CPU and VPDMA.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Fix race condition for firmware loading
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:35 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Fix race condition for firmware loading

vpdma_create API is supposed to allocated the struct vpdma_data and
return it to the driver. Also, it would call the callback function
when the VPDMA firmware is loaded.

Typically, VPE driver have following function call:
    dev->vpdma = vpdma_create(pdev, firmware_load_callback);
And the callback implementation would continue the probe further.
Also, the dev->vpdma is accessed from the callback implementation.

This may lead to race condition between assignment of dev->vpdma
and the callback function being triggered.
This would lead to kernel crash because of NULL pointer access.

Fix this by passing a driver wrapped &vpdma_data instead of allocating
inside vpdma_create.
Change the vpdma_create prototype accordingly and fix return paths.

Also, update the VPE driver to use the updated API and
initialize the dev->vpdma before hand so that the race condition
is avoided.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Fix vb2 buffer cleanup
Benoit Parrot [Fri, 18 Nov 2016 23:20:34 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Fix vb2 buffer cleanup

When stop_streaming is called we need to cleanup the queued
vb2 buffers properly.
This was not previously being done which caused kernel
warning when the application using the resources was killed.
Kernel warnings were also generated on successful completion
of a de-interlacing case as well as upon aborting a
conversion.

Make sure every vb2 buffers is properly handled in all cases.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: sc: Fix incorrect optimization
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:33 +0000 (21:20 -0200)]
[media] media: ti-vpe: sc: Fix incorrect optimization

Current scaler library implementation of sc_set_hs_coeffs and
sc_set_vs_coeffs tries to return immediately if the calculated
coefficient index is already being used.

As the same scaler block is going to be used for all the VPE contexts,
even if the calculated index is same, the parameters have to be
reconfigured for each of the context.

Because of this, when multiple contexts use the same coefficients,
all other contexts would have zero scaling coefficients.
Fix this and also remove the unnecessary hs_index and vs_index fields.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: RGB data type yield inverted data
Benoit Parrot [Fri, 18 Nov 2016 23:20:32 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: RGB data type yield inverted data

The VPDMA RGB data type definition have been updated
to match with Errata i839.

But some of the ARGB definition appeared to be wrong
in the document also. As they would yield RGBA instead.
They have been corrected based on experimentation.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Corrected YUV422 data type label
Benoit Parrot [Fri, 18 Nov 2016 23:20:31 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Corrected YUV422 data type label

The YUV data type definition below are taken from
both the TRM and i839 Errata information.
Use the correct data type considering byte
reordering of components.

Added the 2 missing YUV422 variant.
Also since the single use of "C" in the 422 case
to mean "Cr" (i.e. V component). It was decided
to explicitly label them CR to remove any confusion.
Bear in mind that the type label refer to the memory
packed order (LSB - MSB).

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Added MODULE_DEVICE_TABLE hint
Benoit Parrot [Fri, 18 Nov 2016 23:20:30 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Added MODULE_DEVICE_TABLE hint

ti_vpe module currently does not get loaded automatically.
Added MODULE_DEVICE_TABLE hint to the driver to assist.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: allocate and maintain hwlist
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:29 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: allocate and maintain hwlist

VPDMA block used in ti-vip and ti-vpe modules have support for
up to 8 hardware descriptor lists. A descriptor list can be
submitted to any of the 8 lists (as long as it's not busy).

When multiple clients want to transfer data in parallel, its easier
to allocate one list per client and let it use it. This way, the
list numbers need not be hard-coded into the driver.

Add support for allocating hwlist and maintain them with a priv data.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Add RGB565 and RGB5551 support
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:28 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Add RGB565 and RGB5551 support

VPE hardware can generate output in RGB565 or in RGB5551 format.
Add these formats in the supported format list for CAPTURE stream.
Also, for RGB5551 format, the alpha component is not processed,
so the alpha value is taken from the default color.
Set the default color to make alpha component full when the dst
format is of RGB color space.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Post next descriptor only for list complete IRQ
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:27 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Post next descriptor only for list complete IRQ

vpe_irq checks for the possible interrupt sources and prints the
errors for the DEI_ERROR and DS_UV interrupts. But it also post the
next descriptor list irrespective of whichever interrupt has occurred.

Because of this, driver may release the buffers even before DMA is
complete and also schedule next descriptor list.

Fix this by _actually_ handling the IRQ only when ListComplete IRQ
occurs.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Setup srcdst parameters in start_streaming
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:26 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Setup srcdst parameters in start_streaming

For deinterlacing operation, each operation needs 2 fields in the
history. This is achieved by holding three buffers in
ctx->src_vbs[0,1,2] (f,f-1,f-2)

This is achieved by using the ctx->sequence which gets reset via the
s_fmt ioctl.

These buffers are dequeued in stream OFF by calling free_vbs()
But the corresponding references aren't removed anywhere.

When application tries to stream ON and OFF continuously, s_fmt ioctl
won't be called and it won't setup the srcdst parameters.

Setting source/destination parameters in stream ON ioctl would make
sure that the context is re-initialized before it is being used by
the driver.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: configure line mode separately
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:25 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: configure line mode separately

Current driver configures the line mode of the DEI clients
from the open function directly. Even if the newly created context
is not yet scheduled, it updates some of the VPDMA registers.
This causes a problem in multi instance use case where just opening
the m2m device second time causes the running job to stall. This
happens especially if the source buffers used are NV12.

While all other configuration is being written to context specific
shadow registers, only line mode configuration is happening directly.

As there is no shadow register for line mode configuration, it's better
to separate the config_mode setting and line_mode setting. Call the
new "set_line_modes" functions only when actually loading the mmrs.
This makes sure that no non-running job will write to the registers
directly.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Clear IRQs for individual lists
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:24 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Clear IRQs for individual lists

VPDMA IRQs are registered for multiple lists
When clearing an IRQ for a list interrupt, all the
IRQs for the individual lists are to be cleared separately.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Make list post atomic operation
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:23 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Make list post atomic operation

Writing to the "VPDMA list attribute" register is considered as a list
post. This informs the VPDMA firmware to load the list from the address
which should be taken from the "VPDMA list address" register.

As these two register writes are dependent, it is important that the two
writes happen in atomic manner. This ensures multiple slices (which share
same VPDMA) can post lists asynchronously and all of them point to the
correct addresses.

Slightly modified to implementation for the original patch to use
spin_lock instead of mutex as the list post is also called from
interrupt context.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Add abort channel desc and cleanup APIs
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:22 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Add abort channel desc and cleanup APIs

Whenever VPDMA processes a data descriptor of a list, it processes it
and sets up the channel for the DMA transaction. List manager holds the
descriptor in the list until the DMA is complete. If sync_on_channel
descriptor, or another descriptor for the same channel is present in
the FIFO, list manager keeps them until the current channel is free.

When the capture stream is closed suddenly while there are pending
descriptors in the FIFO (streamON failed, application killed), it would
keep the VPDMA in a busy state. Any further list post would fail with
EBUSY.

To avoid this, drivers need to stop the current processing list and
cleanup all the resources VPDMA has taken and also clear the internal FSM
of list manager. The state machine is cleared by issuing channel specific
abort descriptor.

Therefore, the vpdma_list_cleanup accepts an array of channels for which
abort_channel descriptors should be posted. It is driver's responsibility
to post for all the channels or the channels which were used in the last
context.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Add support for setting max width height
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:21 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Add support for setting max width height

Add a helper function to be able to set the maximum
VPDMA transfer size to limit potential buffer overrun.

Added enums for max_width and max_height fields of the
outbound data descriptor.

Changed vpdma_add_out_dtd to accept two more arguments
for max width and height.

Make use of different max width & height sets for different
of capture module (i.e. slices).

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: Free vpdma buffers in vpe_release
Harinarayan Bhatta [Fri, 18 Nov 2016 23:20:20 +0000 (21:20 -0200)]
[media] media: ti-vpe: Free vpdma buffers in vpe_release

Free vpdma buffers in vpe_release. Otherwise it was generating random
backtrace.

Signed-off-by: Harinarayan Bhatta <harinarayan@ti.com>
Signed-off-by: Somnath Mukherjee <somnath@ti.com>
Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Return NULL for invalid buffer type
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:19 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Return NULL for invalid buffer type

get_q_data can be called with different values for type
e.g. vpe_try_crop calls it with the buffer type which gets passed
from user space

Framework doesn't check wheather its correct type or not
If user space passes wrong type, kernel should not crash.
Return NULL when the passed type is invalid.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: Increasing max buffer height and width
Harinarayan Bhatta [Fri, 18 Nov 2016 23:20:18 +0000 (21:20 -0200)]
[media] media: ti-vpe: Increasing max buffer height and width

Increasing max buffer height and width to allow for padded buffers.

Signed-off-by: Harinarayan Bhatta <harinarayan@ti.com>
Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: Add support for SEQ_TB buffers
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:17 +0000 (21:20 -0200)]
[media] media: ti-vpe: Add support for SEQ_TB buffers

The video source can generate the data in the SEQ_TB buffer format.
In the case of TI SoC, the IVA_HD can generate the interlaced content in
the SEQ_TB buffer format. This is the format where the top and bottom field
data can be contained in a single buffer. For example, for NV12, interlaced
format, the data in Y buffer will be arranged as Y-top followed by
Y-bottom. And likewise for UV plane.

Also, queuing one buffer of SEQ_TB is equivalent to queuing two different
buffers for top and bottom fields. Driver needs to take care of this when
handling source buffer lists.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpe: Do not perform job transaction atomically
Nikhil Devshatwar [Fri, 18 Nov 2016 23:20:16 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Do not perform job transaction atomically

Current VPE driver does not start the job until all the buffers for
a transaction are queued. When running in multiple context, this might
increase the processing latency.

Alternate solution would be to try to continue the same context as long as
buffers for the transaction are ready; else switch the context. This may
increase number of context switches but it reduces latency significantly.

In this approach, the job_ready always succeeds as long as there are
buffers on the CAPTURE and OUTPUT stream. Processing may start immediately
as the first 2 iterations don't need extra source buffers. Shift all the
source buffers after each iteration and remove the oldest buffer.

Also, with this removes the constraint of pre buffering 3 buffers before
call to STREAMON in case of de-interlacing.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: Use line average de-interlacing for first 2 frames
Archit Taneja [Fri, 18 Nov 2016 23:20:15 +0000 (21:20 -0200)]
[media] media: ti-vpe: Use line average de-interlacing for first 2 frames

The motion detection block requires 3 fields to create the motion vector
data. This means that using the default method the first progressive
frame is only generated after 3rd field is consumed.
Hence by default for N input field we would generate N - 2 progressive
frames.

In order to generate N progressive frames from N fields we use the
line averaging mode of the de-interlacer for the first 2 fields and then
revert back to the preferred Edge Directed Interpolation method (using
the motion vector).
Thus creating 2 line averaged frames + N - 2 motion based frames for a
total of N frames.

Signed-off-by: Archit Taneja <archit@ti.com>
Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Fix bus error when vpdma is writing a descriptor
Benoit Parrot [Fri, 18 Nov 2016 23:20:14 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Fix bus error when vpdma is writing a descriptor

On DRA7 since l3_noc event are being reported it was found that
when the write descriptor was being written it was consistently
causing bus error events.

The write address was improperly programmed.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Add helper to set a background color
Benoit Parrot [Fri, 18 Nov 2016 23:20:13 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Add helper to set a background color

Add a helper to set the background color during vpdma transfer.
This is needed when VPDMA is generating 32 bits RGB format
to have the Alpha channel set to an appropriate value.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Add multi-instance and multi-client support
Benoit Parrot [Fri, 18 Nov 2016 23:20:12 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Add multi-instance and multi-client support

The VPDMA (Video Port DMA) as found in devices such as DRA7xx is
used for both the Video Processing Engine (VPE) and the Video Input
Port (VIP). Some devices may have multiple VIP instances each with
its own VPDMA engine. Within VIP two slices can use a single VPDMA
engine simultaneously. So support for multi instances and multiple
clients has been added to VPDMA. Needed modification to the existing
helper functions were then reflected to VPE.

Multi-clients registers offset have also been added in preparation.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: ti-vpe: vpdma: Make vpdma library into its own module
Benoit Parrot [Fri, 18 Nov 2016 23:20:11 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Make vpdma library into its own module

The VPDMA (Video Port DMA) as found in devices such as DRA7xx is
used for both the Video Processing Engine (VPE) and the Video Input
Port (VIP).

In preparation for this we need to turn vpdma into its own
kernel module.

Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] serial_ir: fix reference to 8250 serial code
Mauro Carvalho Chehab [Tue, 22 Nov 2016 08:17:44 +0000 (06:17 -0200)]
[media] serial_ir: fix reference to 8250 serial code

While checking why we need i386 checking, I noticed that
the serial code referred at the driver was moved to another
place. Update it to make clear from where such code came from.

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] lirc_serial: move out of staging and rename to serial_ir
Sean Young [Mon, 21 Nov 2016 21:55:53 +0000 (19:55 -0200)]
[media] lirc_serial: move out of staging and rename to serial_ir

Signed-off-by: Sean Young <sean@mess.org>
7 years agolirc_serial: make checkpatch happy
Mauro Carvalho Chehab [Tue, 22 Nov 2016 07:46:14 +0000 (05:46 -0200)]
lirc_serial: make checkpatch happy

There are a few checkpatch complains here. As we're about to
promote this driver out of staging, address them.

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] lirc_serial: use precision ktime rather than guessing
Sean Young [Mon, 21 Nov 2016 21:55:52 +0000 (19:55 -0200)]
[media] lirc_serial: use precision ktime rather than guessing

This makes transmission more reliable and the code much cleaner.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] lirc_serial: port to rc-core
Sean Young [Mon, 21 Nov 2016 21:55:51 +0000 (19:55 -0200)]
[media] lirc_serial: port to rc-core

Tested with a homebrew serial ir. Remove last remmants of the nslu2
which could not be enabled, and fix checkpatch warnings.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] cec: ignore messages that we initiated
Hans Verkuil [Mon, 21 Nov 2016 15:15:45 +0000 (13:15 -0200)]
[media] cec: ignore messages that we initiated

Some CEC adapters will receive messages that they initiated. Add a
check that will ignore such messages.

Most hardware behaves correctly in this respect, but I have seen
adapters that don't, so just filter this out in the framework.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] mtk-mdp: allocate video_device dynamically
Minghsiu Tsai [Tue, 15 Nov 2016 01:34:34 +0000 (23:34 -0200)]
[media] mtk-mdp: allocate video_device dynamically

It can fix known problems with embedded video_device structs.

Signed-off-by: Minghsiu Tsai <minghsiu.tsai@mediatek.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] zoran: fix spelling mistake in dprintk message
Colin Ian King [Sat, 12 Nov 2016 17:26:48 +0000 (15:26 -0200)]
[media] zoran: fix spelling mistake in dprintk message

Trivial fix to spelling mistake "unnsupported" to "unsupported"
in debug message.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] vivid: fix HDMI VSDB block in the EDID
Hans Verkuil [Sat, 12 Nov 2016 00:23:56 +0000 (22:23 -0200)]
[media] vivid: fix HDMI VSDB block in the EDID

The maximum 'Max TMDS Rate' in the HDMI VSDB block is 340 MHz, not 600.
Higher rates are advertised in the HDMI Forum VSDB block.

So lower the Max TMDS rate in the HDMI VSDB block that the vivid driver
uses to 300 MHz, which is typical of most HDMI 1.4b devices.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] atmel-isc: fix error return code in atmel_isc_probe()
Wei Yongjun [Fri, 11 Nov 2016 13:40:20 +0000 (11:40 -0200)]
[media] atmel-isc: fix error return code in atmel_isc_probe()

Fix to return error code -ENODEV from the error handling
case instead of 0, as done elsewhere in this function.

Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years agoMerge tag 'v4.9-rc6' into patchwork
Mauro Carvalho Chehab [Tue, 22 Nov 2016 07:20:06 +0000 (05:20 -0200)]
Merge tag 'v4.9-rc6' into patchwork

Linux 4.9-rc6

* tag 'v4.9-rc6': (305 commits)
  Linux 4.9-rc6
  ext4: sanity check the block and cluster size at mount time
  fscrypto: don't use on-stack buffer for key derivation
  fscrypto: don't use on-stack buffer for filename encryption
  i2c: i2c-mux-pca954x: fix deselect enabling for device-tree
  kvm: x86: merge kvm_arch_set_irq and kvm_arch_set_irq_inatomic
  KVM: x86: fix missed SRCU usage in kvm_lapic_set_vapic_addr
  KVM: async_pf: avoid recursive flushing of work items
  kvm: kvmclock: let KVM_GET_CLOCK return whether the master clock is in use
  KVM: Disable irq while unregistering user notifier
  KVM: x86: do not go through vcpu in __get_kvmclock_ns
  MAINTAINERS: Add LED subsystem co-maintainer
  crypto: algif_hash - Fix NULL hash crash with shash
  powerpc/mm: Fix missing update of HID register on secondary CPUs
  KVM: arm64: Fix the issues when guest PMCCFILTR is configured
  arm64: KVM: pmu: Fix AArch32 cycle counter access
  powerpc/mm/radix: Invalidate ERAT on tlbiel for POWER9 DD1
  i2c: digicolor: use clk_disable_unprepare instead of clk_unprepare
  ipmi/bt-bmc: change compatible node to 'aspeed, ast2400-ibt-bmc'
  Revert "drm/mediatek: set vblank_disable_allowed to true"
  ...

7 years ago[media] dvb-usb: remove another redundant #include <linux/kconfig.h>
Paul Bolle [Thu, 3 Nov 2016 09:38:01 +0000 (07:38 -0200)]
[media] dvb-usb: remove another redundant #include <linux/kconfig.h>

Kernel source files need not include <linux/kconfig.h> explicitly
because the top Makefile forces to include it with:

  -include $(srctree)/include/linux/kconfig.h

Remove another reduntdant include, that managed to sneak by commit
97139d4a6f26 ("treewide: remove redundant #include <linux/kconfig.h>").

Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] lirc: use-after free while reading from device and unplugging
Sean Young [Mon, 31 Oct 2016 17:52:27 +0000 (15:52 -0200)]
[media] lirc: use-after free while reading from device and unplugging

Many lirc drivers have their own receive buffers which are freed on
unplug (e.g. ir_lirc_unregister). This means that ir->buf->wait_poll
will be freed directly after unplug so do not remove yourself from the
wait queue.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] lirc: prevent use-after free
Sean Young [Mon, 31 Oct 2016 17:52:26 +0000 (15:52 -0200)]
[media] lirc: prevent use-after free

If you unplug an lirc device while reading from it, you will get an
use after free as the cdev is freed while still in use.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] lirc: might sleep error in lirc_dev_fop_read
Sean Young [Mon, 31 Oct 2016 17:52:25 +0000 (15:52 -0200)]
[media] lirc: might sleep error in lirc_dev_fop_read

[  101.457944] ------------[ cut here ]------------
[  101.457954] WARNING: CPU: 3 PID: 1819 at kernel/sched/core.c:7708 __might_sleep+0x7e/0x80
[  101.457960] do not call blocking ops when !TASK_RUNNING; state=1 set at [<ffffffffc0364bc2>] lirc_dev_fop_read+0x292/0x4e0 [lirc_dev]

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] redrat3: increase set size for lengths to maximum
Sean Young [Mon, 31 Oct 2016 17:52:24 +0000 (15:52 -0200)]
[media] redrat3: increase set size for lengths to maximum

In learning mode, you can get much longer messages which can run out
of lengths. The usb message will slightly larger.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] redrat3: enable carrier reports using wideband receiver
Sean Young [Mon, 31 Oct 2016 21:13:11 +0000 (19:13 -0200)]
[media] redrat3: enable carrier reports using wideband receiver

The wideband receiver is a little awkward on the redrat3. Data arrives
on a different endpoint, and the learning command must be reissued
every time data is learned.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] redrat3: fix error paths in probe
Sean Young [Mon, 31 Oct 2016 17:52:22 +0000 (15:52 -0200)]
[media] redrat3: fix error paths in probe

If redrat3_delete() is called, ensure ep_in and udev members are set
up so we don't dereference null in the error path. Also ensure that
rc dev device exists before we enable the receiver and that the
led urb exists before we create the led device.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] redrat3: remove dead code and pointless messages
Sean Young [Mon, 31 Oct 2016 17:52:21 +0000 (15:52 -0200)]
[media] redrat3: remove dead code and pointless messages

Cleanup the error logic, removing checks for things that
should be always initialized when the routines are called,
and remove some bogus messages.

[mchehab@s-opensource.com: fix some merge conflicts]
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] redrat3: don't include vendor/product id in name
Sean Young [Mon, 31 Oct 2016 17:52:20 +0000 (15:52 -0200)]
[media] redrat3: don't include vendor/product id in name

No need to duplicate these in the rc name.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] winbond-cir: use name without space for pnp driver
Sean Young [Mon, 31 Oct 2016 17:52:19 +0000 (15:52 -0200)]
[media] winbond-cir: use name without space for pnp driver

Rename the pnp driver in sysfs from /sys/bus/pnp/drivers/Winbond CIR
to /sys/bus/pnp/drivers/winbond-cir

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] c8sectpfe: fix error return code in c8sectpfe_probe()
Wei Yongjun [Sun, 30 Oct 2016 01:53:10 +0000 (23:53 -0200)]
[media] c8sectpfe: fix error return code in c8sectpfe_probe()

Fix to return error code -ENODEV from the error handling
case instead of 0, as done elsewhere in this function.

Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Patrice Chotard <patrice.chotard@st.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dibusb: fix possible memory leak in dibusb_rc_query()
Wei Yongjun [Sun, 30 Oct 2016 01:36:24 +0000 (23:36 -0200)]
[media] dibusb: fix possible memory leak in dibusb_rc_query()

'buf' is malloced in dibusb_rc_query() and should be freed before
leaving from the error handling cases, otherwise it will cause
memory leak.

Fixes: ff1c123545d7 ("[media] dibusb: handle error code on RC query")

Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years agoLinux 4.9-rc6
Linus Torvalds [Sun, 20 Nov 2016 21:52:19 +0000 (13:52 -0800)]
Linux 4.9-rc6

7 years agoMerge branch 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm
Linus Torvalds [Sun, 20 Nov 2016 18:27:39 +0000 (10:27 -0800)]
Merge branch 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm

Pull ARM fixes from Russell King:
 "A few more ARM fixes:

   - the assembly backtrace code suffers problems with the new printk()
     implementation which assumes that kernel messages without KERN_CONT
     should have newlines inserted between them. Fix this.
   - fix a section naming error - ".init.text" rather than ".text.init"
   - preallocate DMA debug memory at core_initcall() time rather than
     fs_initcall(), as we have some core drivers that need to use DMA
     mapping - and that triggers a kernel warning from the DMA debug
     code.
   - fix XIP kernels after the ro_after_init changes made this data
     permanently read-only"

* 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm:
  ARM: Fix XIP kernels
  ARM: 8628/1: dma-mapping: preallocate DMA-debug hash tables in core_initcall
  ARM: 8624/1: proc-v7m.S: fix init section name
  ARM: fix backtrace

7 years agoMerge tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Linus Torvalds [Sun, 20 Nov 2016 02:40:47 +0000 (18:40 -0800)]
Merge tag 'armsoc-fixes' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Olof Johansson:
 "Again a set of smaller fixes across several platforms (OMAP, Marvell,
  Allwinner, i.MX, etc).

  A handful of typo fixes and smaller missing contents from device
  trees, with some tweaks to OMAP mach files to deal with CPU feature
  print misformatting, potential NULL ptr dereference and one setup
  issue with UARTs"

* tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ipmi/bt-bmc: change compatible node to 'aspeed, ast2400-ibt-bmc'
  ARM: dts: STiH410-b2260: Fix typo in spi0 chipselect definition
  ARM: dts: omap5: board-common: fix wrong SMPS6 (VDD-DDR3) voltage
  ARM: omap3: Add missing memory node in SOM-LV
  arm64: dts: marvell: add unique identifiers for Armada A8k SPI controllers
  arm64: dts: marvell: fix clocksource for CP110 slave SPI0
  arm64: dts: marvell: Fix typo in label name on Armada 37xx
  ASoC: omap-abe-twl6040: fix typo in bindings documentation
  dts: omap5: board-common: enable twl6040 headset jack detection
  dts: omap5: board-common: add phandle to reference Palmas gpadc
  ARM: OMAP2+: avoid NULL pointer dereference
  ARM: OMAP2+: PRM: initialize en_uart4_mask and grpsel_uart4_mask
  ARM: dts: omap3: Fix memory node in Torpedo board
  ARM: AM43XX: Select OMAP_INTERCONNECT in Kconfig
  ARM: OMAP3: Fix formatting of features printed
  ARM: dts: imx53-qsb: Fix regulator constraints
  ARM: dts: sun8i: fix the pinmux for UART1

7 years agoMerge tag 'ext4_for_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso...
Linus Torvalds [Sun, 20 Nov 2016 02:33:50 +0000 (18:33 -0800)]
Merge tag 'ext4_for_stable' of git://git./linux/kernel/git/tytso/ext4

Pull ext4 fixes from Ted Ts'o:
 "A security fix (so a maliciously corrupted file system image won't
  panic the kernel) and some fixes for CONFIG_VMAP_STACK"

* tag 'ext4_for_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: sanity check the block and cluster size at mount time
  fscrypto: don't use on-stack buffer for key derivation
  fscrypto: don't use on-stack buffer for filename encryption

7 years agoext4: sanity check the block and cluster size at mount time
Theodore Ts'o [Fri, 18 Nov 2016 18:00:24 +0000 (13:00 -0500)]
ext4: sanity check the block and cluster size at mount time

If the block size or cluster size is insane, reject the mount.  This
is important for security reasons (although we shouldn't be just
depending on this check).

Ref: http://www.securityfocus.com/archive/1/539661
Ref: https://bugzilla.redhat.com/show_bug.cgi?id=1332506
Reported-by: Borislav Petkov <bp@alien8.de>
Reported-by: Nikolay Borisov <kernel@kyup.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org
7 years agofscrypto: don't use on-stack buffer for key derivation
Eric Biggers [Mon, 14 Nov 2016 01:41:09 +0000 (20:41 -0500)]
fscrypto: don't use on-stack buffer for key derivation

With the new (in 4.9) option to use a virtually-mapped stack
(CONFIG_VMAP_STACK), stack buffers cannot be used as input/output for
the scatterlist crypto API because they may not be directly mappable to
struct page.  get_crypt_info() was using a stack buffer to hold the
output from the encryption operation used to derive the per-file key.
Fix it by using a heap buffer.

This bug could most easily be observed in a CONFIG_DEBUG_SG kernel
because this allowed the BUG in sg_set_buf() to be triggered.

Cc: stable@vger.kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
7 years agofscrypto: don't use on-stack buffer for filename encryption
Eric Biggers [Mon, 14 Nov 2016 01:35:52 +0000 (20:35 -0500)]
fscrypto: don't use on-stack buffer for filename encryption

With the new (in 4.9) option to use a virtually-mapped stack
(CONFIG_VMAP_STACK), stack buffers cannot be used as input/output for
the scatterlist crypto API because they may not be directly mappable to
struct page.  For short filenames, fname_encrypt() was encrypting a
stack buffer holding the padded filename.  Fix it by encrypting the
filename in-place in the output buffer, thereby making the temporary
buffer unnecessary.

This bug could most easily be observed in a CONFIG_DEBUG_SG kernel
because this allowed the BUG in sg_set_buf() to be triggered.

Cc: stable@vger.kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
7 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sat, 19 Nov 2016 21:35:09 +0000 (13:35 -0800)]
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "Some I2C driver bugfixes (and one documentation fix)"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: i2c-mux-pca954x: fix deselect enabling for device-tree
  i2c: digicolor: use clk_disable_unprepare instead of clk_unprepare
  i2c: mux: fix up dependencies
  i2c: Documentation: i2c-topology: fix minor whitespace nit
  i2c: mux: demux-pinctrl: make drivers with no pinctrl work again

7 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Sat, 19 Nov 2016 21:31:40 +0000 (13:31 -0800)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull KVM fixes from Radim Krčmář:
 "ARM:
   - Fix handling of the 32bit cycle counter
   - Fix cycle counter filtering

  x86:
   - Fix a race leading to double unregistering of user notifiers
   - Amend oversight in kvm_arch_set_irq that turned Hyper-V code dead
   - Use SRCU around kvm_lapic_set_vapic_addr
   - Avoid recursive flushing of asynchronous page faults
   - Do not rely on deferred update in KVM_GET_CLOCK, which fixes #GP
   - Let userspace know that KVM_GET_CLOCK is useful with master clock;
     4.9 changed the return value to better match the guest clock, but
     didn't provide means to let guests take advantage of it"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  kvm: x86: merge kvm_arch_set_irq and kvm_arch_set_irq_inatomic
  KVM: x86: fix missed SRCU usage in kvm_lapic_set_vapic_addr
  KVM: async_pf: avoid recursive flushing of work items
  kvm: kvmclock: let KVM_GET_CLOCK return whether the master clock is in use
  KVM: Disable irq while unregistering user notifier
  KVM: x86: do not go through vcpu in __get_kvmclock_ns
  KVM: arm64: Fix the issues when guest PMCCFILTR is configured
  arm64: KVM: pmu: Fix AArch32 cycle counter access

7 years agoi2c: i2c-mux-pca954x: fix deselect enabling for device-tree
Alex Hemme [Sat, 19 Nov 2016 09:48:38 +0000 (10:48 +0100)]
i2c: i2c-mux-pca954x: fix deselect enabling for device-tree

Deselect functionality can be ignored for device-trees with
"i2c-mux-idle-disconnect" entries if no platform_data is available.
By enabling the deselect functionality outside the platform_data
block the logic works as it did in previous kernels.

Fixes: 7fcac9807175 ("i2c: i2c-mux-pca954x: convert to use an explicit i2c mux core")
Cc: <stable@vger.kernel.org> # v4.7+
Signed-off-by: Alex Hemme <ahemme@cisco.com>
Signed-off-by: Ziyang Wu <ziywu@cisco.com>
[touched up a few minor issues /peda]
Signed-off-by: Peter Rosin <peda@axentia.se>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
7 years agoMerge tag 'powerpc-4.9-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sat, 19 Nov 2016 19:21:59 +0000 (11:21 -0800)]
Merge tag 'powerpc-4.9-5' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:
 "Fixes marked for stable:
   - fix system reset interrupt winkle wakeups
   - fix setting of AIL in hypervisor mode

  Fixes for code merged this cycle:
   - fix exception vector build with 2.23 era binutils
   - fix missing update of HID register on secondary CPUs

  Other:
   - fix missing pr_cont()s
   - invalidate ERAT on tlbiel for POWER9 DD1"

* tag 'powerpc-4.9-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/mm: Fix missing update of HID register on secondary CPUs
  powerpc/mm/radix: Invalidate ERAT on tlbiel for POWER9 DD1
  powerpc/64: Fix setting of AIL in hypervisor mode
  powerpc/oops: Fix missing pr_cont()s in instruction dump
  powerpc/oops: Fix missing pr_cont()s in show_regs()
  powerpc/oops: Fix missing pr_cont()s in print_msr_bits() et. al.
  powerpc/oops: Fix missing pr_cont()s in show_stack()
  powerpc: Fix exception vector build with 2.23 era binutils
  powerpc/64s: Fix system reset interrupt winkle wakeups

7 years agoMerge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Sat, 19 Nov 2016 19:15:45 +0000 (11:15 -0800)]
Merge branch 'linus' of git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fixes from Herbert Xu:
 "This fixes the following issues:

   - Compiler warning in caam driver that was the last one remaining

   - Do not register aes-xts in caam drivers on unsupported platforms

   - Regression in algif_hash interface that may lead to an oops"

* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: algif_hash - Fix NULL hash crash with shash
  crypto: caam - fix type mismatch warning
  crypto: caam - do not register AES-XTS mode on LP units

7 years agoMerge tag 'leds_4.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/j.anaszewsk...
Linus Torvalds [Sat, 19 Nov 2016 19:09:28 +0000 (11:09 -0800)]
Merge tag 'leds_4.9-rc6' of git://git./linux/kernel/git/j.anaszewski/linux-leds

Pull LED subsystem update from Jacek Anaszewski:
 "I'd like to announce a new co-maintainer - Pavel Machek"

* tag 'leds_4.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/j.anaszewski/linux-leds:
  MAINTAINERS: Add LED subsystem co-maintainer

7 years agoMerge tag 'dmaengine-fix-4.9-rc6' of git://git.infradead.org/users/vkoul/slave-dma
Linus Torvalds [Sat, 19 Nov 2016 19:05:47 +0000 (11:05 -0800)]
Merge tag 'dmaengine-fix-4.9-rc6' of git://git.infradead.org/users/vkoul/slave-dma

Pull dmaengine fixes from Vinod Koul:
 "Some driver fixes which we pending in my tree:

   - return error code fix in edma driver
   - Kconfig fix for genric allocator in mmp_tdma
   - fix uninitialized value in sun6i
   - Runtime pm fixes for cppi"

* tag 'dmaengine-fix-4.9-rc6' of git://git.infradead.org/users/vkoul/slave-dma:
  dmaengine: cppi41: More PM runtime fixes
  dmaengine: cpp41: Fix handling of error path
  dmaengine: cppi41: Fix unpaired pm runtime when only a USB hub is connected
  dmaengine: cppi41: Fix list not empty warning on module removal
  dmaengine: sun6i: fix the uninitialized value for v_lli
  dmaengine: mmp_tdma: add missing select GENERIC_ALLOCATOR in Kconfig
  dmaengine: edma: Fix error return code in edma_alloc_chan_resources()

7 years agokvm: x86: merge kvm_arch_set_irq and kvm_arch_set_irq_inatomic
Paolo Bonzini [Thu, 17 Nov 2016 14:55:47 +0000 (15:55 +0100)]
kvm: x86: merge kvm_arch_set_irq and kvm_arch_set_irq_inatomic

kvm_arch_set_irq is unused since commit b97e6de9c96.  Merge
its functionality with kvm_arch_set_irq_inatomic.

Reported-by: Jiang Biao <jiang.biao2@zte.com.cn>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
7 years agoKVM: x86: fix missed SRCU usage in kvm_lapic_set_vapic_addr
Paolo Bonzini [Thu, 17 Nov 2016 14:55:46 +0000 (15:55 +0100)]
KVM: x86: fix missed SRCU usage in kvm_lapic_set_vapic_addr

Reported by syzkaller:

    [ INFO: suspicious RCU usage. ]
    4.9.0-rc4+ #47 Not tainted
    -------------------------------
    ./include/linux/kvm_host.h:536 suspicious rcu_dereference_check() usage!

    stack backtrace:
    CPU: 1 PID: 6679 Comm: syz-executor Not tainted 4.9.0-rc4+ #47
    Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
     ffff880039e2f6d0 ffffffff81c2e46b ffff88003e3a5b40 0000000000000000
     0000000000000001 ffffffff83215600 ffff880039e2f700 ffffffff81334ea9
     ffffc9000730b000 0000000000000004 ffff88003c4f8420 ffff88003d3f8000
    Call Trace:
     [<     inline     >] __dump_stack lib/dump_stack.c:15
     [<ffffffff81c2e46b>] dump_stack+0xb3/0x118 lib/dump_stack.c:51
     [<ffffffff81334ea9>] lockdep_rcu_suspicious+0x139/0x180 kernel/locking/lockdep.c:4445
     [<     inline     >] __kvm_memslots include/linux/kvm_host.h:534
     [<     inline     >] kvm_memslots include/linux/kvm_host.h:541
     [<ffffffff8105d6ae>] kvm_gfn_to_hva_cache_init+0xa1e/0xce0 virt/kvm/kvm_main.c:1941
     [<ffffffff8112685d>] kvm_lapic_set_vapic_addr+0xed/0x140 arch/x86/kvm/lapic.c:2217

Reported-by: Dmitry Vyukov <dvyukov@google.com>
Fixes: fda4e2e85589191b123d31cdc21fd33ee70f50fd
Cc: Andrew Honig <ahonig@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
7 years agoKVM: async_pf: avoid recursive flushing of work items
Paolo Bonzini [Thu, 17 Nov 2016 14:55:45 +0000 (15:55 +0100)]
KVM: async_pf: avoid recursive flushing of work items

This was reported by syzkaller:

    [ INFO: possible recursive locking detected ]
    4.9.0-rc4+ #49 Not tainted
    ---------------------------------------------
    kworker/2:1/5658 is trying to acquire lock:
     ([ 1644.769018] (&work->work)
    [<     inline     >] list_empty include/linux/compiler.h:243
    [<ffffffff8128dd60>] flush_work+0x0/0x660 kernel/workqueue.c:1511

    but task is already holding lock:
     ([ 1644.769018] (&work->work)
    [<ffffffff812916ab>] process_one_work+0x94b/0x1900 kernel/workqueue.c:2093

    stack backtrace:
    CPU: 2 PID: 5658 Comm: kworker/2:1 Not tainted 4.9.0-rc4+ #49
    Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
    Workqueue: events async_pf_execute
     ffff8800676ff630 ffffffff81c2e46b ffffffff8485b930 ffff88006b1fc480
     0000000000000000 ffffffff8485b930 ffff8800676ff7e0 ffffffff81339b27
     ffff8800676ff7e8 0000000000000046 ffff88006b1fcce8 ffff88006b1fccf0
    Call Trace:
    ...
    [<ffffffff8128ddf3>] flush_work+0x93/0x660 kernel/workqueue.c:2846
    [<ffffffff812954ea>] __cancel_work_timer+0x17a/0x410 kernel/workqueue.c:2916
    [<ffffffff81295797>] cancel_work_sync+0x17/0x20 kernel/workqueue.c:2951
    [<ffffffff81073037>] kvm_clear_async_pf_completion_queue+0xd7/0x400 virt/kvm/async_pf.c:126
    [<     inline     >] kvm_free_vcpus arch/x86/kvm/x86.c:7841
    [<ffffffff810b728d>] kvm_arch_destroy_vm+0x23d/0x620 arch/x86/kvm/x86.c:7946
    [<     inline     >] kvm_destroy_vm virt/kvm/kvm_main.c:731
    [<ffffffff8105914e>] kvm_put_kvm+0x40e/0x790 virt/kvm/kvm_main.c:752
    [<ffffffff81072b3d>] async_pf_execute+0x23d/0x4f0 virt/kvm/async_pf.c:111
    [<ffffffff8129175c>] process_one_work+0x9fc/0x1900 kernel/workqueue.c:2096
    [<ffffffff8129274f>] worker_thread+0xef/0x1480 kernel/workqueue.c:2230
    [<ffffffff812a5a94>] kthread+0x244/0x2d0 kernel/kthread.c:209
    [<ffffffff831f102a>] ret_from_fork+0x2a/0x40 arch/x86/entry/entry_64.S:433

The reason is that kvm_put_kvm is causing the destruction of the VM, but
the page fault is still on the ->queue list.  The ->queue list is owned
by the VCPU, not by the work items, so we cannot just add list_del to
the work item.

Instead, use work->vcpu to note async page faults that have been resolved
and will be processed through the done list.  There is no need to flush
those.

Cc: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
7 years agokvm: kvmclock: let KVM_GET_CLOCK return whether the master clock is in use
Paolo Bonzini [Wed, 9 Nov 2016 16:48:15 +0000 (17:48 +0100)]
kvm: kvmclock: let KVM_GET_CLOCK return whether the master clock is in use

Userspace can read the exact value of kvmclock by reading the TSC
and fetching the timekeeping parameters out of guest memory.  This
however is brittle and not necessary anymore with KVM 4.11.  Provide
a mechanism that lets userspace know if the new KVM_GET_CLOCK
semantics are in effect, and---since we are at it---if the clock
is stable across all VCPUs.

Cc: Radim Krčmář <rkrcmar@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
7 years agoKVM: Disable irq while unregistering user notifier
Ignacio Alvarado [Fri, 4 Nov 2016 19:15:55 +0000 (12:15 -0700)]
KVM: Disable irq while unregistering user notifier

Function user_notifier_unregister should be called only once for each
registered user notifier.

Function kvm_arch_hardware_disable can be executed from an IPI context
which could cause a race condition with a VCPU returning to user mode
and attempting to unregister the notifier.

Signed-off-by: Ignacio Alvarado <ikalvarado@google.com>
Cc: stable@vger.kernel.org
Fixes: 18863bdd60f8 ("KVM: x86 shared msr infrastructure")
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
7 years agoKVM: x86: do not go through vcpu in __get_kvmclock_ns
Paolo Bonzini [Wed, 16 Nov 2016 17:31:30 +0000 (18:31 +0100)]
KVM: x86: do not go through vcpu in __get_kvmclock_ns

Going through the first VCPU is wrong if you follow a KVM_SET_CLOCK with
a KVM_GET_CLOCK immediately after, without letting the VCPU run and
call kvm_guest_time_update.

To fix this, compute the kvmclock value ourselves, using the master
clock (tsc, nsec) pair as the base and the host CPU frequency as
the scale.

Reported-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
7 years agoMerge tag 'kvm-arm-for-4.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git...
Radim Krčmář [Sat, 19 Nov 2016 17:02:07 +0000 (18:02 +0100)]
Merge tag 'kvm-arm-for-4.9-rc6' of git://git./linux/kernel/git/kvmarm/kvmarm

KVM/ARM updates for v4.9-rc6

- Fix handling of the 32bit cycle counter
- Fix cycle counter filtering

7 years ago[media] media: mtk-mdp: mark PM functions as __maybe_unused
Arnd Bergmann [Wed, 26 Oct 2016 09:43:43 +0000 (07:43 -0200)]
[media] media: mtk-mdp: mark PM functions as __maybe_unused

A previous patch tried to fix a build error, but introduced another
warning:

drivers/media/platform/mtk-mdp/mtk_mdp_core.c:71:13: error: ‘mtk_mdp_clock_off’ defined but not used [-Werror=unused-function]
drivers/media/platform/mtk-mdp/mtk_mdp_core.c:62:13: error: ‘mtk_mdp_clock_on’ defined but not used [-Werror=unused-function]

This marks all the PM functions as __maybe_unused and removes
the #ifdef around them, as that will always do the right thing.

Fixes: 1b06fcf56aa6 ("[media] media: mtk-mdp: fix build error")

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years agoMerge tag 'acpi-4.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sat, 19 Nov 2016 01:21:58 +0000 (17:21 -0800)]
Merge tag 'acpi-4.9-rc6' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI fixes from Rafael Wysocki:
 "They fix an ACPI thermal management regression introduced by a recent
  FADT handling cleanup, an ACPI tools build issue introduced by a
  recent ACPICA commit and a PCC mailbox initialization bug causing
  lockdep to complain loudly.

  Specifics:

   - Revert a recent ACPICA cleanup that attempted to get rid of all
     FADT version 2 legacy, but broke ACPI thermal management on at
     least one system (Rafael Wysocki).

   - Fix cross-compiled builds of ACPI tools that stopped working after
     a recent cleanup related to the handling of header files in ACPICA
     (Lv Zheng).

   - Fix a locking issue in the PCC channel initialization code that
     invokes devm_request_irq() under a spinlock (among other things)
     and causes lockdep to complain (Hoan Tran)"

* tag 'acpi-4.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  tools/power/acpi: Remove direct kernel source include reference
  mailbox: PCC: Fix lockdep warning when request PCC channel
  Revert "ACPICA: FADT support cleanup"

7 years agoMerge branch 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild
Linus Torvalds [Sat, 19 Nov 2016 00:45:21 +0000 (16:45 -0800)]
Merge branch 'rc-fixes' of git://git./linux/kernel/git/mmarek/kbuild

Pull kbuild fixes from Michal Marek:
 "Here are some regression fixes for kbuild:

   - modversion support for exported asm symbols (Nick Piggin). The
     affected architectures need separate patches adding
     asm-prototypes.h.

   - fix rebuilds of lib-ksyms.o (Nick Piggin)

   - -fno-PIE builds (Sebastian Siewior and Borislav Petkov). This is
     not a kernel regression, but one of the Debian gcc package.
     Nevertheless, it's quite annoying, so I think it should go into
     mainline and stable now"

* 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  kbuild: Steal gcc's pie from the very beginning
  kbuild: be more careful about matching preprocessed asm ___EXPORT_SYMBOL
  x86/kexec: add -fno-PIE
  scripts/has-stack-protector: add -fno-PIE
  kbuild: add -fno-PIE
  kbuild: modversions for EXPORT_SYMBOL() for asm
  kbuild: prevent lib-ksyms.o rebuilds

7 years agoMerge tag 'nfsd-4.9-2' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Sat, 19 Nov 2016 00:32:21 +0000 (16:32 -0800)]
Merge tag 'nfsd-4.9-2' of git://linux-nfs.org/~bfields/linux

Pull nfsd bugfix from Bruce Fields:
 "Just one fix for an NFS/RDMA crash"

* tag 'nfsd-4.9-2' of git://linux-nfs.org/~bfields/linux:
  sunrpc: svc_age_temp_xprts_now should not call setsockopt non-tcp transports

7 years agoMAINTAINERS: Add LED subsystem co-maintainer
Pavel Machek [Tue, 15 Nov 2016 10:12:05 +0000 (11:12 +0100)]
MAINTAINERS: Add LED subsystem co-maintainer

Mark me as a co-maintainer of LED subsystem.

Signed-off-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Jacek Anaszewski <j.anaszewski@samsung.com>
7 years agoRevert "[media] dvb_frontend: merge duplicate dvb_tuner_ops.release implementations"
Mauro Carvalho Chehab [Fri, 18 Nov 2016 22:30:51 +0000 (20:30 -0200)]
Revert "[media] dvb_frontend: merge duplicate dvb_tuner_ops.release implementations"

While this patch sounded a good idea, unfortunately, it causes
bad dependencies, as drivers that would otherwise work without
the DVB core will now break:

ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/tea5767.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/tea5761.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/tda827x.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/tda18218.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/qt1010.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/mt2266.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/mt20xx.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/mt2060.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/mc44s803.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/fc0013.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/fc0012.ko] undefined!
ERROR: "dvb_tuner_simple_release" [drivers/media/tuners/fc0011.ko] undefined!

So, we have to revert it.

Note: as the argument for the release ops changed from "int"
to "void", we needed to change it at the revert patch, to
avoid compilation issues like:
drivers/media/tuners/tea5767.c:437:23: error: initialization from incompatible pointer type [-Werror=incompatible-pointer-types]
  .release           = tea5767_release,
                       ^~~~~~~~~~~~~~~

This reverts commit 22a613e89825ea7a3984a968463cc6d425bd8856.

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years agoMerge branches 'acpica-fixes', 'acpi-cppc-fixes' and 'acpi-tools-fixes'
Rafael J. Wysocki [Fri, 18 Nov 2016 20:34:42 +0000 (21:34 +0100)]
Merge branches 'acpica-fixes', 'acpi-cppc-fixes' and 'acpi-tools-fixes'

* acpica-fixes:
  Revert "ACPICA: FADT support cleanup"

* acpi-cppc-fixes:
  mailbox: PCC: Fix lockdep warning when request PCC channel

* acpi-tools-fixes:
  tools/power/acpi: Remove direct kernel source include reference

7 years ago[media] Kconfig: fix breakages when DVB_CORE is not selected
Mauro Carvalho Chehab [Fri, 18 Nov 2016 19:47:41 +0000 (17:47 -0200)]
[media] Kconfig: fix breakages when DVB_CORE is not selected

On some weird randconfigs, it is possible to select DVB
drivers, without having the DVB_CORE:

CONFIG_DVB_AU8522=m
CONFIG_DVB_AU8522_V4L=m
CONFIG_DVB_TUNER_DIB0090=m

This was never supposed to work, but changeset 22a613e89825
("[media] dvb_frontend: merge duplicate dvb_tuner_ops.release implementations")
caused it to be exposed:

   drivers/built-in.o: In function `fc0011_attach':
   (.text+0x1598fb): undefined reference to `dvb_tuner_simple_release'
   drivers/built-in.o:(.rodata+0x55e58): undefined reference to `dvb_tuner_simple_release'
   drivers/built-in.o:(.rodata+0x57398): undefined reference to `dvb_tuner_simple_release'

Fixes: 22a613e89825 ("[media] dvb_frontend: merge duplicate dvb_tuner_ops.release implementations")
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: rc: nuvoton: replace usage of spin_lock_irqsave in ISR
Heiner Kallweit [Tue, 25 Oct 2016 19:23:52 +0000 (17:23 -0200)]
[media] media: rc: nuvoton: replace usage of spin_lock_irqsave in ISR

Kernel takes care that interrupts from one source are serialized.
So there's no need to use spinlock_irq_save.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: rc: nuvoton: rename spinlock nvt_lock
Heiner Kallweit [Tue, 25 Oct 2016 19:23:49 +0000 (17:23 -0200)]
[media] media: rc: nuvoton: rename spinlock nvt_lock

Spinlock nvt_lock is a member of struct nvt_dev and there's no need
to prefix it with nvt_. So remove this prefix.

[mchehab@s-opensource.org: change the prefix also at the open function,
 as the patch removing it were not applied (yet?)]
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: rc: nuvoton: eliminate nvt->tx.lock
Heiner Kallweit [Tue, 25 Oct 2016 19:23:45 +0000 (17:23 -0200)]
[media] media: rc: nuvoton: eliminate nvt->tx.lock

Using a separate spinlock to protect access to substruct tx of struct
nvt_dev doesn't provide any actual benefit. We can use spinlock
nvt_lock to protect all access to struct nvt_dev and get rid of
nvt->tx.lock.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: rc: nuvoton: eliminate member pdev from struct nvt_dev
Heiner Kallweit [Tue, 25 Oct 2016 19:23:37 +0000 (17:23 -0200)]
[media] media: rc: nuvoton: eliminate member pdev from struct nvt_dev

Member pdev of struct nvt_dev is needed only to access &pdev->dev.
We can get rid of this it by using rdev->dev.parent instead
(both point to the same struct device).

Setting rdev->dev.parent can be removed from the probe function
as this is done by devm_rc_allocate_device now.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] drivers/media/media-device: fix double free bug in _unregister()
Max Kellermann [Tue, 9 Aug 2016 21:33:03 +0000 (18:33 -0300)]
[media] drivers/media/media-device: fix double free bug in _unregister()

While removing all interfaces in media_device_unregister(), all
media_interface pointers are freed.  This is illegal and results in
double kfree() if any media_interface is still linked at this point;
maybe because a userspace process still has a file handle.  Once the
process closes the file handle, dvb_media_device_free() gets called,
which frees the dvb_device.intf_devnode again.

This patch removes the unnecessary kfree() call, and documents who's
responsible for really freeing it.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media-entity: clear media_gobj.mdev in _destroy()
Max Kellermann [Tue, 9 Aug 2016 21:32:57 +0000 (23:32 +0200)]
[media] media-entity: clear media_gobj.mdev in _destroy()

media_gobj_destroy() may be called twice on one instance - once by
media_device_unregister() and again by dvb_media_device_free().  The
function media_remove_intf_links() establishes and documents the
convention that mdev==NULL means that the object is not registered,
but nobody ever NULLs this variable.  So this patch really implements
this behavior, and adds another mdev==NULL check to
media_gobj_destroy() to protect against double removal.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb_frontend: move kref to struct dvb_frontend
Max Kellermann [Tue, 9 Aug 2016 21:32:51 +0000 (18:32 -0300)]
[media] dvb_frontend: move kref to struct dvb_frontend

This commit amends my old commit fe35637b0a9f ("[media] dvb_frontend:
eliminate blocking wait in dvb_unregister_frontend()"), which added
kref to struct dvb_frontend_private.  It turned out that there are
several use-after-free bugs left, which affect the struct
dvb_frontend.  Protecting it with kref also protects struct
dvb_frontend_private, so we can simply move it.

This is how the use-after-free looks like in KASAN:

    BUG: KASAN: use-after-free in string+0x60/0xb1 at addr ffff880033bd9fc0
    Read of size 1 by task kworker/0:2/617
    CPU: 0 PID: 617 Comm: kworker/0:2 Not tainted 4.8.0-rc1-hosting+ #60
    Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
    Workqueue: usb_hub_wq hub_event
     0000000000000000 ffff880033757218 ffffffff81394e50 ffff880033bd9fd0
     ffff880035c03b00 ffff880033757240 ffffffff811f271d ffff880033bd9fc0
     1ffff1000677b3f8 ffffed000677b3f8 ffff8800337572b8 ffffffff811f2afe
    Call Trace:
     [...]
     [<ffffffff813a2d2f>] vsnprintf+0x39d/0x7e9
     [<ffffffff813993f9>] add_uevent_var+0x10f/0x1dc
     [<ffffffff814fe5ca>] rc_dev_uevent+0x55/0x6f
     [<ffffffff814438f8>] dev_uevent+0x2e1/0x316
     [<ffffffff81399744>] kobject_uevent_env+0x27e/0x701
     [<ffffffff81399bd2>] kobject_uevent+0xb/0xd
     [<ffffffff81443445>] device_del+0x322/0x383
     [<ffffffff81500c0c>] rc_unregister_device+0x98/0xc3
     [<ffffffff81508fb4>] dvb_usb_remote_exit+0x7a/0x90
     [<ffffffff81506157>] dvb_usb_exit+0x1d/0xe5
     [<ffffffff81506e90>] dvb_usb_device_exit+0x69/0x7d
     [<ffffffff8150a181>] pctv452e_usb_disconnect+0x7b/0x80
     [...]
    Object at ffff880033bd9fc0, in cache kmalloc-16 size: 16
    Allocated:
     [...]
    Freed:
    PID = 617
     [...]
     [<ffffffff811f034c>] kfree+0xd9/0x166
     [<ffffffff814fe513>] ir_free_table+0x2f/0x51
     [<ffffffff81500bc1>] rc_unregister_device+0x4d/0xc3
     [<ffffffff81508fb4>] dvb_usb_remote_exit+0x7a/0x90
     [<ffffffff81506157>] dvb_usb_exit+0x1d/0xe5
     [<ffffffff81506e90>] dvb_usb_device_exit+0x69/0x7d
     [<ffffffff8150a181>] pctv452e_usb_disconnect+0x7b/0x80

Another one:

    BUG: KASAN: use-after-free in do_sys_poll+0x336/0x6b8 at addr ffff88003563fcc0
    Read of size 8 by task tuner on fronte/1042
    CPU: 1 PID: 1042 Comm: tuner on fronte Tainted: G    B           4.8.0-rc1-hosting+ #60
    Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
     0000000000000000 ffff88003353f910 ffffffff81394e50 ffff88003563fd80
     ffff880035c03200 ffff88003353f938 ffffffff811f271d ffff88003563fc80
     1ffff10006ac7f98 ffffed0006ac7f98 ffff88003353f9b0 ffffffff811f2afe
    Call Trace:
     [...]
     [<ffffffff812289b3>] do_sys_poll+0x336/0x6b8
     [...]
     [<ffffffff81228ed9>] SyS_poll+0xa9/0x194
     [...]
    Object at ffff88003563fc80, in cache kmalloc-256 size: 256
    Allocated:
     [...]
    Freed:
    PID = 617
     [...]
     [<ffffffff811f034c>] kfree+0xd9/0x166
     [<ffffffff814eb60d>] dvb_unregister_device+0xd6/0xe5
     [<ffffffff814fa4ed>] dvb_unregister_frontend+0x4b/0x66
     [<ffffffff8150810b>] dvb_usb_adapter_frontend_exit+0x69/0xac
     [<ffffffff8150617d>] dvb_usb_exit+0x43/0xe5
     [<ffffffff81506e90>] dvb_usb_device_exit+0x69/0x7d
     [<ffffffff8150a181>] pctv452e_usb_disconnect+0x7b/0x80

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] stb0899: move code to "detach" callback
Max Kellermann [Tue, 9 Aug 2016 21:32:46 +0000 (18:32 -0300)]
[media] stb0899: move code to "detach" callback

Ensure that STB0899_POSTPROC_GPIO_POWER is set synchronously.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb_frontend: add "detach" callback
Max Kellermann [Tue, 9 Aug 2016 21:32:41 +0000 (18:32 -0300)]
[media] dvb_frontend: add "detach" callback

Prepare for making "release" asynchronous (via kref).  Some operations
may need to be run synchronously in dvb_frontend_detach(), and that's
why we need a "detach" callback.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb_frontend: merge the two dvb_frontend_detach() versions
Max Kellermann [Tue, 9 Aug 2016 21:32:36 +0000 (18:32 -0300)]
[media] dvb_frontend: merge the two dvb_frontend_detach() versions

This code duplication is confusing and error prone.  Let's merge them
by moving the release/dvb_detach call into one function with one

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb_frontend: tuner_ops.release returns void
Max Kellermann [Tue, 9 Aug 2016 21:32:31 +0000 (18:32 -0300)]
[media] dvb_frontend: tuner_ops.release returns void

It is not clear what this return value means.  All implemenations
return 0, and the one caller ignores the value.  Let's remove this
useless return value completely.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb_frontend: merge duplicate dvb_tuner_ops.release implementations
Max Kellermann [Tue, 9 Aug 2016 21:32:26 +0000 (18:32 -0300)]
[media] dvb_frontend: merge duplicate dvb_tuner_ops.release implementations

Most release callback functions are identical: free the "tuner_priv"
and clear it.  Let's eliminate some bloat by providing this simple
implementation in the dvb_frontend library.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb-core/en50221: use dvb_remove_device()
Max Kellermann [Tue, 9 Aug 2016 21:32:16 +0000 (18:32 -0300)]
[media] dvb-core/en50221: use dvb_remove_device()

Commit da677fe14364 ("[media] dvb-core/en50221: use kref to manage
struct dvb_ca_private") moved the dvb_unregister_device() call to the
kref callback, but that left lots of stale device state visible to
userspace (e.g. in sysfs).  By using dvb_remove_device() and
dvb_free_device() instead of dvb_unregister_device(), we can avoid
that.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvbdev: split dvb_unregister_device()
Max Kellermann [Tue, 9 Aug 2016 21:32:11 +0000 (18:32 -0300)]
[media] dvbdev: split dvb_unregister_device()

dvb_unregister_device() has a major problem: it combines unregistering
with memory disposal.  Sometimes, it is necessary to unregister a
device, but no memory can be freed yet, because a process still has a
(stale) file handle.  Therefore, we need to split
dvb_unregister_device().  This will allow sanitizing a few callers.

With my new design, dvb_unregister_device() appears misnamed, but to
reduce patch noise, I'm not renaming it just yet.

Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb: make DVB frontend *_ops instances "const"
Max Kellermann [Tue, 9 Aug 2016 21:32:21 +0000 (18:32 -0300)]
[media] dvb: make DVB frontend *_ops instances "const"

These are immutable.  Making them "const" allows the compiler to move
them to the "rodata" section.

Note that cxd2841er_t_c_ops cannot be made "const", because
cxd2841er_attach() modifies it.  Ouch!

[mchehab@s-opensource.com: fix merge conflicts]
Signed-off-by: Max Kellermann <max.kellermann@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years agoMerge tag 'sound-4.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Fri, 18 Nov 2016 16:56:47 +0000 (08:56 -0800)]
Merge tag 'sound-4.9-rc6' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "Three trivial fixes:

  A regression fix for ASRock mobo, a use-after-free fix at hot-unplug
  of USB-audio, and a quirk for new Thinkpad models"

* tag 'sound-4.9-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: usb-audio: Fix use-after-free of usb_device at disconnect
  ALSA: hda - Fix mic regression by ASRock mobo fixup
  ALSA: hda - add a new condition to check if it is thinkpad

7 years agoMerge tag 'gpio-v4.9-4' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux...
Linus Torvalds [Fri, 18 Nov 2016 16:47:47 +0000 (08:47 -0800)]
Merge tag 'gpio-v4.9-4' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO fixes from Linus Walleij:
 "These are hopefully the last GPIO fixes for v4.9. The most important
  is that it fixes the UML randconfig builds that have been nagging me
  for some time and me being confused about where the problem was really
  sitting, now this fix give this nice feeling that everything is solid
  and builds fine.

  Summary:

   - Finally, after being puzzled by a bunch of recurrent UML build
     failures on randconfigs from the build robot, Keno Fischer nailed
     it: GPIO_DEVRES is optional and depends on HAS_IOMEM even though
     many users just unconditionally rely on it to be available. And it
     *should* be available: garbage collection is nice for this and it
     *certainly* has nothing to do with having IOMEM. So we got rid of
     it, and now the UML builds should JustWork(TM).

   - Do not call .get_direction() on sleeping GPIO chips on the fastpath
     when locking GPIOs for interrupts: it is done from atomic context,
     no way.

   - Some driver fixes"

* tag 'gpio-v4.9-4' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  gpio: Remove GPIO_DEVRES option
  gpio: tc3589x: fix up .get_direction()
  gpio: do not double-check direction on sleeping chips
  gpio: pca953x: Move memcpy into mutex lock for set multiple
  gpio: pca953x: Fix corruption of other gpios in set_multiple.