Hi Xia,
On Fri, Apr 03, 2020 at 05:40:24PM +0800, Xia Jiang wrote:
> Call pm_runtime_get_sync() before starting a frame and then
> pm_runtime_put() after completing it. This can save power for the time
> between processing two frames.
>
> Signed-off-by: Xia Jiang <xia.jiang@xxxxxxxxxxxx>
> ---
> .../media/platform/mtk-jpeg/mtk_jpeg_core.c | 27 +++++--------------
> 1 file changed, 6 insertions(+), 21 deletions(-)
>
Thank you for the patch. Please see my comments inline.
> diff --git a/drivers/media/platform/mtk-jpeg/mtk_jpeg_core.c b/drivers/media/platform/mtk-jpeg/mtk_jpeg_core.c
> index a536fa95b3d6..dd5cadd101ef 100644
> --- a/drivers/media/platform/mtk-jpeg/mtk_jpeg_core.c
> +++ b/drivers/media/platform/mtk-jpeg/mtk_jpeg_core.c
> @@ -710,23 +710,6 @@ static struct vb2_v4l2_buffer *mtk_jpeg_buf_remove(struct mtk_jpeg_ctx *ctx,
> return v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx);
> }
>
> -static int mtk_jpeg_start_streaming(struct vb2_queue *q, unsigned int count)
> -{
> - struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(q);
> - struct vb2_v4l2_buffer *vb;
> - int ret = 0;
> -
> - ret = pm_runtime_get_sync(ctx->jpeg->dev);
> - if (ret < 0)
> - goto err;
> -
> - return 0;
> -err:
> - while ((vb = mtk_jpeg_buf_remove(ctx, q->type)))
> - v4l2_m2m_buf_done(vb, VB2_BUF_STATE_QUEUED);
> - return ret;
> -}
> -
> static void mtk_jpeg_stop_streaming(struct vb2_queue *q)
> {
> struct mtk_jpeg_ctx *ctx = vb2_get_drv_priv(q);
> @@ -751,8 +734,6 @@ static void mtk_jpeg_stop_streaming(struct vb2_queue *q)
>
> while ((vb = mtk_jpeg_buf_remove(ctx, q->type)))
> v4l2_m2m_buf_done(vb, VB2_BUF_STATE_ERROR);
> -
> - pm_runtime_put_sync(ctx->jpeg->dev);
> }
>
> static const struct vb2_ops mtk_jpeg_qops = {
> @@ -761,7 +742,6 @@ static const struct vb2_ops mtk_jpeg_qops = {
> .buf_queue = mtk_jpeg_buf_queue,
> .wait_prepare = vb2_ops_wait_prepare,
> .wait_finish = vb2_ops_wait_finish,
> - .start_streaming = mtk_jpeg_start_streaming,
> .stop_streaming = mtk_jpeg_stop_streaming,
> };
>
> @@ -812,7 +792,7 @@ static void mtk_jpeg_device_run(void *priv)
> struct mtk_jpeg_src_buf *jpeg_src_buf;
> struct mtk_jpeg_bs bs;
> struct mtk_jpeg_fb fb;
> - int i;
> + int i, ret;
>
> src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
> dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx);
> @@ -832,6 +812,10 @@ static void mtk_jpeg_device_run(void *priv)
> return;
> }
>
> + ret = pm_runtime_get_sync(jpeg->dev);
> + if (ret < 0)
> + goto dec_end;
> +
> mtk_jpeg_set_dec_src(ctx, &src_buf->vb2_buf, &bs);
> if (mtk_jpeg_set_dec_dst(ctx, &jpeg_src_buf->dec_param, &dst_buf->vb2_buf, &fb))
> goto dec_end;
> @@ -957,6 +941,7 @@ static irqreturn_t mtk_jpeg_dec_irq(int irq, void *priv)
> v4l2_m2m_buf_done(src_buf, buf_state);
> v4l2_m2m_buf_done(dst_buf, buf_state);
> v4l2_m2m_job_finish(jpeg->m2m_dev, ctx->fh.m2m_ctx);
> + pm_runtime_put_sync(ctx->jpeg->dev);
The _sync variant explicitly waits until the asynchronous PM operation
completes. This is usually undesired, because the CPU stays blocked for
no good reason. In this context it is actually a bug, because this is an
interrupt handler and it's not allowed to sleep. I wonder why this
actually didn't crash in your testing. Please change to the regular
pm_runtime_put().
Best regards,
Tomasz
