On Tue, Dec 28, 2021 at 05:41:37PM +0800, Yunfei Dong wrote: > From: Yunfei Dong <yunfei.dong@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> [...] > diff --git a/drivers/media/platform/mtk-vcodec/mtk_vcodec_dec.c b/drivers/media/platform/mtk-vcodec/mtk_vcodec_dec.c > index 130ecef2e766..87891ebd7246 100644 > --- a/drivers/media/platform/mtk-vcodec/mtk_vcodec_dec.c > +++ b/drivers/media/platform/mtk-vcodec/mtk_vcodec_dec.c > @@ -466,6 +466,8 @@ static int vidioc_vdec_s_fmt(struct file *file, void *priv, > } > ctx->state = MTK_STATE_INIT; > } > + } else { > + ctx->capture_fourcc = fmt->fourcc; > } > > /* > diff --git a/drivers/media/platform/mtk-vcodec/mtk_vcodec_drv.h b/drivers/media/platform/mtk-vcodec/mtk_vcodec_drv.h > index a23a7646437c..95e07cf2cd3e 100644 > --- a/drivers/media/platform/mtk-vcodec/mtk_vcodec_drv.h > +++ b/drivers/media/platform/mtk-vcodec/mtk_vcodec_drv.h > @@ -277,6 +277,7 @@ struct vdec_pic_info { > * to be used with encoder and stateful decoder. > * @is_flushing: set to true if flushing is in progress. > * @current_codec: current set input codec, in V4L2 pixel format > + * @capture_fourcc: capture queue type in V4L2 pixel format > * > * @colorspace: enum v4l2_colorspace; supplemental to pixelformat > * @ycbcr_enc: enum v4l2_ycbcr_encoding, Y'CbCr encoding > @@ -322,6 +323,7 @@ struct mtk_vcodec_ctx { > bool is_flushing; > > u32 current_codec; > + u32 capture_fourcc; What is the purpose of capture_fourcc? It is not used. > +/** > + * struct vdec_ap_ipi_get_param - for AP_IPIMSG_SET_PARAM > + * @msg_id : AP_IPIMSG_DEC_START > + * @inst_id : instance ID. Used if the ABI version >= 2. > + * @data : picture information > + * @param_type : get param type > + * @codec_type : Codec fourcc > + */ > +struct vdec_ap_ipi_get_param { > + uint32_t msg_id; > + uint32_t inst_id; > + uint32_t data[4]; > + uint32_t param_type; > + uint32_t codec_type; > +}; Are AP_IPIMSG_SET_PARAM and AP_IPIMSG_DEC_START typos? > +/** > + * struct vdec_vpu_ipi_init_ack - for VPU_IPIMSG_DEC_INIT_ACK > + * @msg_id : VPU_IPIMSG_DEC_INIT_ACK > + * @status : VPU exeuction result > + * @ap_inst_addr : AP vcodec_vpu_inst instance address > + * @data : picture information from SCP. > + * @param_type : get param type > + */ > +struct vdec_vpu_ipi_get_param_ack { > + uint32_t msg_id; > + int32_t status; > + uint64_t ap_inst_addr; > + uint32_t data[4]; > + uint32_t param_type; > + uint32_t reserved; > +}; Same here: is VPU_IPIMSG_DEC_INIT_ACK a typo? What is the purpose of the "reserved" field? > diff --git a/drivers/media/platform/mtk-vcodec/vdec_vpu_if.c b/drivers/media/platform/mtk-vcodec/vdec_vpu_if.c [...] > +static void handle_get_param_msg_ack( > + const struct vdec_vpu_ipi_get_param_ack *msg) > +{ > + struct vdec_vpu_inst *vpu = (struct vdec_vpu_inst *) > + (unsigned long)msg->ap_inst_addr; Does it really need to cast twice? > + > + mtk_vcodec_debug(vpu, "+ ap_inst_addr = 0x%llx", msg->ap_inst_addr); > + > + /* param_type is enum vdec_get_param_type */ > + switch(msg->param_type) { > + case 2: What is 2? Is it GET_PARAM_PIC_INFO? > +int vpu_dec_get_param(struct vdec_vpu_inst *vpu, uint32_t *data, > + unsigned int len, unsigned int param_type) > +{ > + struct vdec_ap_ipi_get_param msg; > + int i; > + int err; > + > + mtk_vcodec_debug_enter(vpu); > + > + if (len > ARRAY_SIZE(msg.data)) { > + mtk_vcodec_err(vpu, "invalid len = %d\n", len); > + return -EINVAL; > + } > + > + memset(&msg, 0, sizeof(msg)); > + msg.msg_id = AP_IPIMSG_DEC_GET_PARAM; > + msg.inst_id = vpu->inst_id; > + for (i = 0; i < len; i++) > + msg.data[i] = data[i]; Would it be more concise if using memcpy? > diff --git a/drivers/media/platform/mtk-vcodec/vdec_vpu_if.h b/drivers/media/platform/mtk-vcodec/vdec_vpu_if.h > index 4cb3c7f5a3ad..963f8d4877b7 100644 > --- a/drivers/media/platform/mtk-vcodec/vdec_vpu_if.h > +++ b/drivers/media/platform/mtk-vcodec/vdec_vpu_if.h > @@ -28,6 +28,8 @@ struct mtk_vcodec_ctx; > * @wq : wait queue to wait VPU message ack > * @handler : ipi handler for each decoder > * @codec_type : use codec type to separate different codecs > + * @capture_type : used capture type to separate different capture format > + * @fb_sz : frame buffer size of each plane > */ > struct vdec_vpu_inst { > int id; > @@ -42,6 +44,8 @@ struct vdec_vpu_inst { > wait_queue_head_t wq; > mtk_vcodec_ipi_handler handler; > unsigned int codec_type; > + unsigned int capture_type; > + unsigned int fb_sz[2]; > }; capture_type is not used in the patch. Does fb_sz take effect in later patches?