linux.git/drivers/media/platform/mtk-vcodec, branch v5.8

media: mtk-vpu: avoid unaligned access to DTCM buffer.

2020-03-12T15:25:33+00:00

Previously, vpu->recv_buf and send_buf are forced cast from
void __iomem *tcm. vpu->recv_buf->share_buf is passed to
vpu_ipi_desc.handler(). It's not able to do unaligned access. Otherwise
kernel would crash due to unable to handle kernel paging request.

struct vpu_run {
	u32 signaled;
	char fw_ver[VPU_FW_VER_LEN];
	unsigned int	dec_capability;
	unsigned int	enc_capability;
	wait_queue_head_t wq;
};

fw_ver starts at 4 byte boundary. If system enables
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS, strscpy() will do
read_word_at_a_time(), which tries to read 8-byte: *(unsigned long *)addr

vpu_init_ipi_handler() calls strscpy(), which would lead to crash.

vpu_init_ipi_handler() and several other handlers (eg.
vpu_dec_ipi_handler) only do read access to this data, so they can be
const, and we can use memcpy_fromio() to copy the buf to another non iomem
buffer then pass to handler.

Fixes: 85709cbf1524 ("media: replace strncpy() by strscpy()")
Signed-off-by: Hsin-Yi Wang 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: mtk-vcodec: reset segment data then trig decoder

2020-03-02T14:22:30+00:00

VP9 bitstream specification indicate segment data should reset to
default when meet key frames, intra only frames or enable error
resilience mode. So memset segmentation map buffer before every
decode process is not appropriate.

Reset segment data only when needed, then start decoder hardware

Signed-off-by: Rui Wang 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: media/platform: rename VFL_TYPE_GRABBER to _VIDEO

2020-02-24T15:54:14+00:00

'GRABBER' is a weird name, all other types map to the /dev
device names. Rename to 'VIDEO' to be consistent with the
other types.

Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: mtk-vcodec: use v4l2_m2m_buffer where appropriate

2020-01-09T13:46:41+00:00

Despite using M2M in both the decoder and encoder, this driver used
vb2_v4l2_buffer as its base buffer structure, and placed a list_head
right after the buffer declaration in order to match the layout of a
v4l2_m2m_buffer. This is very dangerous as it means the driver will
break should the layout of v4l2_m2m_buffer change.

Fix this by directly using v4l2_m2m_buffer and updating the sites that
accessed the buffer accordingly.

Signed-off-by: Alexandre Courbot 
Signed-off-by: Hans Verkuil 
[hverkuil-cisco@xs4all.nl: fix checkpatch warning]
Signed-off-by: Mauro Carvalho Chehab

media: mtk-vcodec: Remove extra area allocation in an input buffer on encoding

2019-12-16T09:19:42+00:00

MediaTek encoder allocates non pixel data area for an input buffer every
plane. As the input buffer should be read-only, the driver should not write
anything in the buffer. Therefore, the extra data should be unnecessary.

Signed-off-by: Hirokazu Honda 
Reviewed-by: Tomasz Figa 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: Revert "media: mtk-vcodec: Remove extra area allocation in an input buffer on encoding"

2019-11-10T17:10:11+00:00

This reverts commit 3192b2ca79b3f72fbc0eab9cd95432e3c317ab0d.

There are indications that this patch causes problems on some
platforms due to some hardware prefetch. Reverting this patch
for now until this is better understood.

Reported-by: Tomasz Figa 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: mtk-vcodec: Remove extra area allocation in an input buffer on encoding

2019-11-10T06:25:34+00:00

MediaTek encoder allocates non pixel data area for an input buffer every
plane. As the input buffer should be read-only, the driver should not write
anything in the buffer. Therefore, the extra data should be unnecessary.

Signed-off-by: Hirokazu Honda 
Reviewed-by: Tomasz Figa 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: platform: Use devm_platform_ioremap_resource() in two functions

2019-10-10T10:29:07+00:00

Simplify these function implementations by using a known wrapper function.

This issue was detected by using the Coccinelle software.

Signed-off-by: Markus Elfring 
Reviewed-by: Geert Uytterhoeven 
Tested-by: Niklas Söderlund 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: mtk-vcodec: vdec: clean up vidioc_vdec_s_fmt a bit

2019-10-01T19:46:42+00:00

Check for a potentially NULL pointer that was overlooked and use shorter
accessors to the same data. While we are at it, sprinkle a few comments.

Signed-off-by: Alexandre Courbot 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab

media: mtk-vcodec: vdec: set VPI IPI handler in one place

2019-10-01T19:45:25+00:00

Each of the supported decoder formats used to set the same RPC
interrupt handler by themselves, even though this could be done
by the IF init function itself. Move it to the right place and stop
making its symbol public.

Signed-off-by: Alexandre Courbot 
Signed-off-by: Hans Verkuil 
Signed-off-by: Mauro Carvalho Chehab