Commit aa86e0bb authored by Dafna Hirschfeld's avatar Dafna Hirschfeld Committed by Mauro Carvalho Chehab

media: staging: rkisp1: cap: don't set next buffer from rkisp1_vb2_buf_queue

The function 'rkisp1_vb2_buf_queue' sets the next buffer directly
in case the capture is already streaming but no frame yet arrived
from the sensor. This is an optimization that tries to avoid
dropping a frame.
The call atomic_read(&cap->rkisp1->isp.frame_sequence) is used
to check if a frame arrived. Reading the 'frame_sequence' should
be avoided outside irq handlers to avoid race conditions.

This patch removes this optimization. Dropping of the first
frames can be avoided if userspace queues the buffers before
start streaming. If userspace starts queueing buffers
only after calling 'streamon' he risks frame drops anyway.
Signed-off-by: default avatarDafna Hirschfeld <dafna.hirschfeld@collabora.com>
Acked-by: default avatarHelen Koike <helen.koike@collabora.com>
Signed-off-by: default avatarHans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: default avatarMauro Carvalho Chehab <mchehab+huawei@kernel.org>
parent 2cc6a9fe
......@@ -743,18 +743,7 @@ static void rkisp1_vb2_buf_queue(struct vb2_buffer *vb)
ispbuf->buff_addr[RKISP1_PLANE_CB]);
spin_lock_irqsave(&cap->buf.lock, flags);
/*
* If there's no next buffer assigned, queue this buffer directly
* as the next buffer, and update the memory interface.
*/
if (cap->is_streaming && !cap->buf.next &&
atomic_read(&cap->rkisp1->isp.frame_sequence) == -1) {
cap->buf.next = ispbuf;
rkisp1_set_next_buf(cap);
} else {
list_add_tail(&ispbuf->queue, &cap->buf.queue);
}
list_add_tail(&ispbuf->queue, &cap->buf.queue);
spin_unlock_irqrestore(&cap->buf.lock, flags);
}
......
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