linux-kernel - Re: [PATCH 2/4] venus: update calculation for dpb buffers

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <68ee1330-3319-d39e-8628-8ebc1b856a21@linaro.org>
Date:   Thu, 4 May 2023 19:30:15 +0200
From:   Konrad Dybcio <konrad.dybcio@...aro.org>
To:     Dikshita Agarwal <quic_dikshita@...cinc.com>,
        linux-media@...r.kernel.org, stanimir.k.varbanov@...il.com,
        quic_vgarodia@...cinc.com, agross@...nel.org, andersson@...nel.org,
        mchehab@...nel.org
Cc:     linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH 2/4] venus: update calculation for dpb buffers



On 4.05.2023 19:29, Konrad Dybcio wrote:
> 
> 
> On 4.05.2023 12:36, Dikshita Agarwal wrote:
>> Use dpb color format, width and height of output port
>> for calculating buffer size of dpb buffers.
>>
>> Signed-off-by: Dikshita Agarwal <quic_dikshita@...cinc.com>
>> ---
> Looks sane but I'm not exactly an expert on this
> 
> Acked-by: Konrad Dybcio <konrad.dybcio@...aro.org>
> 
> Konrad
>>  drivers/media/platform/qcom/venus/helpers.c          | 4 ++++
>>  drivers/media/platform/qcom/venus/hfi_plat_bufs.h    | 3 +++
>>  drivers/media/platform/qcom/venus/hfi_plat_bufs_v6.c | 8 +++++++-
>>  3 files changed, 14 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/media/platform/qcom/venus/helpers.c b/drivers/media/platform/qcom/venus/helpers.c
>> index 5946def..4ad6232 100644
>> --- a/drivers/media/platform/qcom/venus/helpers.c
>> +++ b/drivers/media/platform/qcom/venus/helpers.c
>> @@ -641,12 +641,16 @@ static int platform_get_bufreq(struct venus_inst *inst, u32 buftype,
>>  	if (is_dec) {
>>  		params.width = inst->width;
>>  		params.height = inst->height;
>> +		params.out_width = inst->out_width;
>> +		params.out_height = inst->out_height;
>>  		params.codec = inst->fmt_out->pixfmt;
>>  		params.hfi_color_fmt = to_hfi_raw_fmt(inst->fmt_cap->pixfmt);
>>  		params.dec.max_mbs_per_frame = mbs_per_frame_max(inst);
>>  		params.dec.buffer_size_limit = 0;
>>  		params.dec.is_secondary_output =
>>  			inst->opb_buftype == HFI_BUFFER_OUTPUT2;
>> +		if (params.dec.is_secondary_output)
>> +			params.hfi_dpb_color_fmt = inst->dpb_fmt;
>>  		params.dec.is_interlaced =
>>  			inst->pic_struct != HFI_INTERLACE_FRAME_PROGRESSIVE;
>>  	} else {
>> diff --git a/drivers/media/platform/qcom/venus/hfi_plat_bufs.h b/drivers/media/platform/qcom/venus/hfi_plat_bufs.h
>> index 52a51a3..25e6074 100644
>> --- a/drivers/media/platform/qcom/venus/hfi_plat_bufs.h
>> +++ b/drivers/media/platform/qcom/venus/hfi_plat_bufs.h
>> @@ -12,8 +12,11 @@
>>  struct hfi_plat_buffers_params {
>>  	u32 width;
>>  	u32 height;
>> +	u32 out_width;
>> +	u32 out_height;
>>  	u32 codec;
>>  	u32 hfi_color_fmt;
>> +	u32 hfi_dpb_color_fmt;
>>  	enum hfi_version version;
>>  	u32 num_vpp_pipes;
>>  	union {
>> diff --git a/drivers/media/platform/qcom/venus/hfi_plat_bufs_v6.c b/drivers/media/platform/qcom/venus/hfi_plat_bufs_v6.c
>> index ea25c45..3855b04 100644
>> --- a/drivers/media/platform/qcom/venus/hfi_plat_bufs_v6.c
>> +++ b/drivers/media/platform/qcom/venus/hfi_plat_bufs_v6.c
>> @@ -1185,6 +1185,7 @@ static int bufreq_dec(struct hfi_plat_buffers_params *params, u32 buftype,
>>  	enum hfi_version version = params->version;
>>  	u32 codec = params->codec;
>>  	u32 width = params->width, height = params->height, out_min_count;
>> +	u32 out_width = params->out_width, out_height = params->out_height;
>>  	struct dec_bufsize_ops *dec_ops;
>>  	bool is_secondary_output = params->dec.is_secondary_output;
>>  	bool is_interlaced = params->dec.is_interlaced;
>> @@ -1235,7 +1236,12 @@ static int bufreq_dec(struct hfi_plat_buffers_params *params, u32 buftype,
>>  		bufreq->count_min = out_min_count;
>>  		bufreq->size =
>>  			venus_helper_get_framesz_raw(params->hfi_color_fmt,
>> -						     width, height);
>> +						     out_width, out_height);
>> +		if (buftype == HFI_BUFFER_OUTPUT &&
>> +		    params->dec.is_secondary_output)
I suppose this line could be unbroken

Konrad
>> +			bufreq->size =
>> +				venus_helper_get_framesz_raw(params->hfi_dpb_color_fmt,
>> +							     out_width, out_height);
>>  	} else if (buftype == HFI_BUFFER_INTERNAL_SCRATCH(version)) {
>>  		bufreq->size = dec_ops->scratch(width, height, is_interlaced);
>>  	} else if (buftype == HFI_BUFFER_INTERNAL_SCRATCH_1(version)) {