lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <445dd26796e2242666e5256e7e36bcbe@codeaurora.org>
Date:   Tue, 08 Dec 2020 22:07:35 +0530
From:   dikshita@...eaurora.org
To:     Hans Verkuil <hverkuil-cisco@...all.nl>
Cc:     linux-media@...r.kernel.org, nicolas@...fresne.ca,
        stanimir.varbanov@...aro.org, linux-kernel@...r.kernel.org,
        linux-arm-msm@...r.kernel.org, vgarodia@...eaurora.org
Subject: Re: [PATCH v3 3/3] venus: venc: Add support for frame-specific
 min/max qp controls

On 2020-12-02 19:06, Hans Verkuil wrote:
> On 18/11/2020 07:54, Dikshita Agarwal wrote:
>> Add support for frame type specific min and max qp controls
>> for encoder.
>> 
>> Signed-off-by: Dikshita Agarwal <dikshita@...eaurora.org>
>> ---
>>  drivers/media/platform/qcom/venus/core.h       | 18 +++++++++
>>  drivers/media/platform/qcom/venus/venc.c       | 21 ++++++++---
>>  drivers/media/platform/qcom/venus/venc_ctrls.c | 51 
>> ++++++++++++++++++++++++++
>>  3 files changed, 85 insertions(+), 5 deletions(-)
>> 
>> diff --git a/drivers/media/platform/qcom/venus/core.h 
>> b/drivers/media/platform/qcom/venus/core.h
>> index 3bc129a..6a764c9 100644
>> --- a/drivers/media/platform/qcom/venus/core.h
>> +++ b/drivers/media/platform/qcom/venus/core.h
>> @@ -230,10 +230,28 @@ struct venc_controls {
>>  	u32 h264_b_qp;
>>  	u32 h264_min_qp;
>>  	u32 h264_max_qp;
>> +	u32 h264_i_min_qp;
>> +	u32 h264_i_max_qp;
>> +	u32 h264_p_min_qp;
>> +	u32 h264_p_max_qp;
>> +	u32 h264_b_min_qp;
>> +	u32 h264_b_max_qp;
>>  	u32 h264_loop_filter_mode;
>>  	s32 h264_loop_filter_alpha;
>>  	s32 h264_loop_filter_beta;
>> 
>> +	u32 hevc_i_qp;
>> +	u32 hevc_p_qp;
>> +	u32 hevc_b_qp;
>> +	u32 hevc_min_qp;
>> +	u32 hevc_max_qp;
>> +	u32 hevc_i_min_qp;
>> +	u32 hevc_i_max_qp;
>> +	u32 hevc_p_min_qp;
>> +	u32 hevc_p_max_qp;
>> +	u32 hevc_b_min_qp;
>> +	u32 hevc_b_max_qp;
>> +
>>  	u32 vp8_min_qp;
>>  	u32 vp8_max_qp;
>> 
>> diff --git a/drivers/media/platform/qcom/venus/venc.c 
>> b/drivers/media/platform/qcom/venus/venc.c
>> index 0bf92cc..f2f5a85 100644
>> --- a/drivers/media/platform/qcom/venus/venc.c
>> +++ b/drivers/media/platform/qcom/venus/venc.c
>> @@ -668,17 +668,28 @@ static int venc_set_properties(struct venus_inst 
>> *inst)
>>  		return ret;
>> 
>>  	ptype = HFI_PROPERTY_PARAM_VENC_SESSION_QP;
>> -	quant.qp_i = ctr->h264_i_qp;
>> -	quant.qp_p = ctr->h264_p_qp;
>> -	quant.qp_b = ctr->h264_b_qp;
>> +	if (inst->fmt_cap->pixfmt == V4L2_PIX_FMT_HEVC) {
>> +		quant.qp_i = ctr->hevc_i_qp;
>> +		quant.qp_p = ctr->hevc_p_qp;
>> +		quant.qp_b = ctr->hevc_b_qp;
>> +	} else {
>> +		quant.qp_i = ctr->h264_i_qp;
>> +		quant.qp_p = ctr->h264_p_qp;
>> +		quant.qp_b = ctr->h264_b_qp;
>> +	}
>>  	quant.layer_id = 0;
>>  	ret = hfi_session_set_property(inst, ptype, &quant);
>>  	if (ret)
>>  		return ret;
>> 
>>  	ptype = HFI_PROPERTY_PARAM_VENC_SESSION_QP_RANGE;
>> -	quant_range.min_qp = ctr->h264_min_qp;
>> -	quant_range.max_qp = ctr->h264_max_qp;
>> +	if (inst->fmt_cap->pixfmt == V4L2_PIX_FMT_HEVC) {
>> +		quant_range.min_qp = ctr->hevc_min_qp;
>> +		quant_range.max_qp = ctr->hevc_max_qp;
>> +	} else {
>> +		quant_range.min_qp = ctr->h264_min_qp;
>> +		quant_range.max_qp = ctr->h264_max_qp;
>> +	}
>>  	quant_range.layer_id = 0;
>>  	ret = hfi_session_set_property(inst, ptype, &quant_range);
>>  	if (ret)
>> diff --git a/drivers/media/platform/qcom/venus/venc_ctrls.c 
>> b/drivers/media/platform/qcom/venus/venc_ctrls.c
>> index 0708b3b..cd131e3 100644
>> --- a/drivers/media/platform/qcom/venus/venc_ctrls.c
>> +++ b/drivers/media/platform/qcom/venus/venc_ctrls.c
>> @@ -125,9 +125,60 @@ static int venc_op_s_ctrl(struct v4l2_ctrl *ctrl)
>>  	case V4L2_CID_MPEG_VIDEO_H264_MIN_QP:
>>  		ctr->h264_min_qp = ctrl->val;
>>  		break;
>> +	case V4L2_CID_MPEG_VIDEO_H264_I_FRAME_MIN_QP:
>> +		ctr->h264_i_min_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_H264_P_FRAME_MIN_QP:
>> +		ctr->h264_p_min_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_H264_B_FRAME_MIN_QP:
>> +		ctr->h264_b_min_qp = ctrl->val;
>> +		break;
>>  	case V4L2_CID_MPEG_VIDEO_H264_MAX_QP:
>>  		ctr->h264_max_qp = ctrl->val;
>>  		break;
>> +	case V4L2_CID_MPEG_VIDEO_H264_I_FRAME_MAX_QP:
>> +		ctr->h264_i_max_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_H264_P_FRAME_MAX_QP:
>> +		ctr->h264_p_max_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_H264_B_FRAME_MAX_QP:
>> +		ctr->h264_b_max_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_I_FRAME_QP:
>> +		ctr->hevc_i_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_P_FRAME_QP:
>> +		ctr->hevc_p_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_B_FRAME_QP:
>> +		ctr->hevc_b_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_MIN_QP:
>> +		ctr->hevc_min_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_I_FRAME_MIN_QP:
>> +		ctr->hevc_i_min_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_P_FRAME_MIN_QP:
>> +		ctr->hevc_p_min_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_B_FRAME_MIN_QP:
>> +		ctr->hevc_b_min_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_MAX_QP:
>> +		ctr->hevc_max_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_I_FRAME_MAX_QP:
>> +		ctr->hevc_i_max_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_P_FRAME_MAX_QP:
>> +		ctr->hevc_p_max_qp = ctrl->val;
>> +		break;
>> +	case V4L2_CID_MPEG_VIDEO_HEVC_B_FRAME_MAX_QP:
>> +		ctr->hevc_b_max_qp = ctrl->val;
>> +		break;
>>  	case V4L2_CID_MPEG_VIDEO_MULTI_SLICE_MODE:
>>  		ctr->multi_slice_mode = ctrl->val;
>>  		break;
>> 
> 
> This looks incomplete: the new controls aren't actually added to the
> driver with v4l2_ctrl_new_std(). Did you test this?
Oh, Sorry. Looks like I forgot to add some changes.
will send a new patch.

Thanks,
Dikshita
> 
> Regards,
> 
> 	Hans

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ