Commit b6187392 authored by Mauro Carvalho Chehab's avatar Mauro Carvalho Chehab

[media] vsp1: fix CodingStyle violations on multi-line comments

Several multi-line comments added at the vsp1 patch series
violate the Kernel CodingStyle. Fix them.
Signed-off-by: default avatarMauro Carvalho Chehab <mchehab@s-opensource.com>
parent c9f49607
......@@ -242,7 +242,8 @@ static int bru_set_selection(struct v4l2_subdev *subdev,
goto done;
}
/* The compose rectangle top left corner must be inside the output
/*
* The compose rectangle top left corner must be inside the output
* frame.
*/
format = vsp1_entity_get_pad_format(&bru->entity, config,
......
......@@ -224,7 +224,8 @@ static void clu_configure(struct vsp1_entity *entity,
switch (params) {
case VSP1_ENTITY_PARAMS_INIT: {
/* The format can't be changed during streaming, only verify it
/*
* The format can't be changed during streaming, only verify it
* at setup time and store the information internally for future
* runtime configuration calls.
*/
......
......@@ -296,7 +296,8 @@ struct vsp1_dl_list *vsp1_dl_list_get(struct vsp1_dl_manager *dlm)
dl = list_first_entry(&dlm->free, struct vsp1_dl_list, list);
list_del(&dl->list);
/* The display list chain must be initialised to ensure every
/*
* The display list chain must be initialised to ensure every
* display list can assert list_empty() if it is not in a chain.
*/
INIT_LIST_HEAD(&dl->chain);
......@@ -315,7 +316,8 @@ static void __vsp1_dl_list_put(struct vsp1_dl_list *dl)
if (!dl)
return;
/* Release any linked display-lists which were chained for a single
/*
* Release any linked display-lists which were chained for a single
* hardware operation.
*/
if (dl->has_chain) {
......@@ -325,7 +327,8 @@ static void __vsp1_dl_list_put(struct vsp1_dl_list *dl)
dl->has_chain = false;
/* We can't free fragments here as DMA memory can only be freed in
/*
* We can't free fragments here as DMA memory can only be freed in
* interruptible context. Move all fragments to the display list
* manager's list of fragments to be freed, they will be
* garbage-collected by the work queue.
......@@ -437,7 +440,8 @@ static void vsp1_dl_list_fill_header(struct vsp1_dl_list *dl, bool is_last)
struct vsp1_dl_body *dlb;
unsigned int num_lists = 0;
/* Fill the header with the display list bodies addresses and sizes. The
/*
* Fill the header with the display list bodies addresses and sizes. The
* address of the first body has already been filled when the display
* list was allocated.
*/
......@@ -456,7 +460,8 @@ static void vsp1_dl_list_fill_header(struct vsp1_dl_list *dl, bool is_last)
dl->header->num_lists = num_lists;
/* If this display list's chain is not empty, we are on a list, where
/*
* If this display list's chain is not empty, we are on a list, where
* the next item in the list is the display list entity which should be
* automatically queued by the hardware.
*/
......@@ -482,7 +487,8 @@ void vsp1_dl_list_commit(struct vsp1_dl_list *dl)
if (dl->dlm->mode == VSP1_DL_MODE_HEADER) {
struct vsp1_dl_list *dl_child;
/* In header mode the caller guarantees that the hardware is
/*
* In header mode the caller guarantees that the hardware is
* idle at this point.
*/
......@@ -495,7 +501,8 @@ void vsp1_dl_list_commit(struct vsp1_dl_list *dl)
vsp1_dl_list_fill_header(dl_child, last);
}
/* Commit the head display list to hardware. Chained headers
/*
* Commit the head display list to hardware. Chained headers
* will auto-start.
*/
vsp1_write(vsp1, VI6_DL_HDR_ADDR(dlm->index), dl->dma);
......
......@@ -283,7 +283,8 @@ int vsp1_du_atomic_update(struct device *dev, unsigned int rpf_index,
cfg->pixelformat, cfg->pitch, &cfg->mem[0], &cfg->mem[1],
&cfg->mem[2], cfg->zpos);
/* Store the format, stride, memory buffer address, crop and compose
/*
* Store the format, stride, memory buffer address, crop and compose
* rectangles and Z-order position and for the input.
*/
fmtinfo = vsp1_get_format_info(vsp1, cfg->pixelformat);
......
......@@ -35,7 +35,7 @@ enum vsp1_entity_type {
VSP1_ENTITY_WPF,
};
/*
/**
* enum vsp1_entity_params - Entity configuration parameters class
* @VSP1_ENTITY_PARAMS_INIT - Initial parameters
* @VSP1_ENTITY_PARAMS_PARTITION - Per-image partition parameters
......
......@@ -136,7 +136,7 @@ static const struct vsp1_format_info vsp1_video_formats[] = {
3, { 8, 8, 8 }, false, true, 1, 1, false },
};
/*
/**
* vsp1_get_format_info - Retrieve format information for a 4CC
* @vsp1: the VSP1 device
* @fourcc: the format 4CC
......
......@@ -75,7 +75,8 @@ static void rpf_configure(struct vsp1_entity *entity,
unsigned int offsets[2];
struct v4l2_rect crop;
/* Source size and crop offsets.
/*
* Source size and crop offsets.
*
* The crop offsets correspond to the location of the crop
* rectangle top left corner in the plane buffer. Only two
......@@ -84,7 +85,8 @@ static void rpf_configure(struct vsp1_entity *entity,
*/
crop = *vsp1_rwpf_get_crop(rpf, rpf->entity.config);
/* Partition Algorithm Control
/*
* Partition Algorithm Control
*
* The partition algorithm can split this frame into multiple
* slices. We must scale our partition window based on the pipe
......@@ -98,7 +100,8 @@ static void rpf_configure(struct vsp1_entity *entity,
struct vsp1_entity *wpf = &pipe->output->entity;
unsigned int input_width = crop.width;
/* Scale the partition window based on the configuration
/*
* Scale the partition window based on the configuration
* of the pipeline.
*/
output = vsp1_entity_get_pad_format(wpf, wpf->config,
......
......@@ -132,7 +132,8 @@ static int vsp1_rwpf_get_selection(struct v4l2_subdev *subdev,
struct v4l2_mbus_framefmt *format;
int ret = 0;
/* Cropping is only supported on the RPF and is implemented on the sink
/*
* Cropping is only supported on the RPF and is implemented on the sink
* pad.
*/
if (rwpf->entity.type == VSP1_ENTITY_WPF || sel->pad != RWPF_PAD_SINK)
......@@ -180,7 +181,8 @@ static int vsp1_rwpf_set_selection(struct v4l2_subdev *subdev,
struct v4l2_rect *crop;
int ret = 0;
/* Cropping is only supported on the RPF and is implemented on the sink
/*
* Cropping is only supported on the RPF and is implemented on the sink
* pad.
*/
if (rwpf->entity.type == VSP1_ENTITY_WPF || sel->pad != RWPF_PAD_SINK)
......
......@@ -205,7 +205,7 @@ static void vsp1_video_pipeline_setup_partitions(struct vsp1_pipeline *pipe)
pipe->partitions = DIV_ROUND_UP(format->width, div_size);
}
/*
/**
* vsp1_video_partition - Calculate the active partition output window
*
* @div_size: pre-determined maximum partition division size
......@@ -242,7 +242,8 @@ static struct v4l2_rect vsp1_video_partition(struct vsp1_pipeline *pipe,
modulus = format->width % div_size;
/* We need to prevent the last partition from being smaller than the
/*
* We need to prevent the last partition from being smaller than the
* *minimum* width of the hardware capabilities.
*
* If the modulus is less than half of the partition size,
......@@ -251,7 +252,8 @@ static struct v4l2_rect vsp1_video_partition(struct vsp1_pipeline *pipe,
* to prevents this: |1234|1234|1234|1234|1|.
*/
if (modulus) {
/* pipe->partitions is 1 based, whilst index is a 0 based index.
/*
* pipe->partitions is 1 based, whilst index is a 0 based index.
* Normalise this locally.
*/
unsigned int partitions = pipe->partitions - 1;
......@@ -371,7 +373,8 @@ static void vsp1_video_pipeline_run(struct vsp1_pipeline *pipe)
if (!pipe->dl)
pipe->dl = vsp1_dl_list_get(pipe->output->dlm);
/* Start with the runtime parameters as the configure operation can
/*
* Start with the runtime parameters as the configure operation can
* compute/cache information needed when configuring partitions. This
* is the case with flipping in the WPF.
*/
......@@ -391,13 +394,15 @@ static void vsp1_video_pipeline_run(struct vsp1_pipeline *pipe)
pipe->current_partition++) {
struct vsp1_dl_list *dl;
/* Partition configuration operations will utilise
/*
* Partition configuration operations will utilise
* the pipe->current_partition variable to determine
* the work they should complete.
*/
dl = vsp1_dl_list_get(pipe->output->dlm);
/* An incomplete chain will still function, but output only
/*
* An incomplete chain will still function, but output only
* the partitions that had a dl available. The frame end
* interrupt will be marked on the last dl in the chain.
*/
......@@ -818,7 +823,8 @@ static void vsp1_video_stop_streaming(struct vb2_queue *vq)
unsigned long flags;
int ret;
/* Clear the buffers ready flag to make sure the device won't be started
/*
* Clear the buffers ready flag to make sure the device won't be started
* by a QBUF on the video node on the other side of the pipeline.
*/
spin_lock_irqsave(&video->irqlock, flags);
......
......@@ -222,7 +222,8 @@ static void wpf_configure(struct vsp1_entity *entity,
unsigned int height = source_format->height;
unsigned int offset;
/* Cropping. The partition algorithm can split the image into
/*
* Cropping. The partition algorithm can split the image into
* multiple slices.
*/
if (pipe->partitions > 1)
......@@ -238,7 +239,8 @@ static void wpf_configure(struct vsp1_entity *entity,
if (pipe->lif)
return;
/* Update the memory offsets based on flipping configuration.
/*
* Update the memory offsets based on flipping configuration.
* The destination addresses point to the locations where the
* VSP starts writing to memory, which can be different corners
* of the image depending on vertical flipping.
......@@ -246,7 +248,8 @@ static void wpf_configure(struct vsp1_entity *entity,
if (pipe->partitions > 1) {
const struct vsp1_format_info *fmtinfo = wpf->fmtinfo;
/* Horizontal flipping is handled through a line buffer
/*
* Horizontal flipping is handled through a line buffer
* and doesn't modify the start address, but still needs
* to be handled when image partitioning is in effect to
* order the partitions correctly.
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment