[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <27722665-a7b9-0d49-443d-3da2a20405de@collabora.com>
Date: Tue, 11 Apr 2023 15:39:06 +0200
From: Benjamin Gaignard <benjamin.gaignard@...labora.com>
To: Hans Verkuil <hverkuil-cisco@...all.nl>,
ezequiel@...guardiasur.com.ar, p.zabel@...gutronix.de,
mchehab@...nel.org, robh+dt@...nel.org,
krzysztof.kozlowski+dt@...aro.org, heiko@...ech.de,
nicolas.dufresne@...labora.com
Cc: linux-media@...r.kernel.org, linux-rockchip@...ts.infradead.org,
devicetree@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org, kernel@...labora.com
Subject: Re: [PATCH v5 09/13] media: verisilicon: Add Rockchip AV1 decoder
Le 11/04/2023 à 14:33, Hans Verkuil a écrit :
> Hi Benjamin,
>
> On 30/03/2023 17:40, Benjamin Gaignard wrote:
>> Implement AV1 stateless decoder for rockchip VPU981.
>> It decode 8 and 10 bits AV1 bitstreams.
>> AV1 scaling feature is done by the postprocessor.
>>
>> Signed-off-by: Benjamin Gaignard <benjamin.gaignard@...labora.com>
>> ---
>> drivers/media/platform/verisilicon/Makefile | 1 +
>> .../media/platform/verisilicon/hantro_hw.h | 64 +-
>> .../verisilicon/rockchip_vpu981_hw_av1_dec.c | 2024 +++++++++++++++++
>> .../verisilicon/rockchip_vpu981_regs.h | 477 ++++
>> 4 files changed, 2564 insertions(+), 2 deletions(-)
>> create mode 100644 drivers/media/platform/verisilicon/rockchip_vpu981_hw_av1_dec.c
>> create mode 100644 drivers/media/platform/verisilicon/rockchip_vpu981_regs.h
>>
> <snip>
>
>> +static void rockchip_vpu981_av1_dec_set_tile_info(struct hantro_ctx *ctx)
>> +{
>> + struct hantro_av1_dec_hw_ctx *av1_dec = &ctx->av1_dec;
>> + struct hantro_av1_dec_ctrls *ctrls = &av1_dec->ctrls;
>> + struct v4l2_av1_tile_info tile_info = ctrls->frame->tile_info;
> I get this warning:
>
> drivers/media/platform/verisilicon/rockchip_vpu981_hw_av1_dec.c: In function 'rockchip_vpu981_av1_dec_set_tile_info':
> drivers/media/platform/verisilicon/rockchip_vpu981_hw_av1_dec.c:635:1: warning: the frame size of 1080 bytes is larger than 1024 bytes [-Wframe-larger-than=]
> 635 | }
> | ^
>
> The cause is this tile_info struct that is on the stack.
>
> Does this have to be a copy? Or can it be a pointer to ctrls->frame->tile_info?
Using a pointer is a good solution.
Thanks for the finding.
I will include that in v6.
Regards,
Benjamin
>
>> + const struct v4l2_ctrl_av1_tile_group_entry *group_entry =
>> + ctrls->tile_group_entry;
>> + int context_update_y =
>> + tile_info.context_update_tile_id / tile_info.tile_cols;
>> + int context_update_x =
>> + tile_info.context_update_tile_id % tile_info.tile_cols;
>> + int context_update_tile_id =
>> + context_update_x * tile_info.tile_rows + context_update_y;
>> + u8 *dst = av1_dec->tile_info.cpu;
>> + struct hantro_dev *vpu = ctx->dev;
>> + int tile0, tile1;
>> +
>> + memset(dst, 0, av1_dec->tile_info.size);
>> +
>> + for (tile0 = 0; tile0 < tile_info.tile_cols; tile0++) {
>> + for (tile1 = 0; tile1 < tile_info.tile_rows; tile1++) {
>> + int tile_id = tile1 * tile_info.tile_cols + tile0;
>> + u32 start, end;
>> + u32 y0 =
>> + tile_info.height_in_sbs_minus_1[tile1] + 1;
>> + u32 x0 = tile_info.width_in_sbs_minus_1[tile0] + 1;
>> +
>> + // tile size in SB units (width,height)
>> + *dst++ = x0;
>> + *dst++ = 0;
>> + *dst++ = 0;
>> + *dst++ = 0;
>> + *dst++ = y0;
>> + *dst++ = 0;
>> + *dst++ = 0;
>> + *dst++ = 0;
>> +
>> + // tile start position
>> + start = group_entry[tile_id].tile_offset - group_entry[0].tile_offset;
>> + *dst++ = start & 255;
>> + *dst++ = (start >> 8) & 255;
>> + *dst++ = (start >> 16) & 255;
>> + *dst++ = (start >> 24) & 255;
>> +
>> + // # of bytes in tile data
>> + end = start + group_entry[tile_id].tile_size;
>> + *dst++ = end & 255;
>> + *dst++ = (end >> 8) & 255;
>> + *dst++ = (end >> 16) & 255;
>> + *dst++ = (end >> 24) & 255;
>> + }
>> + }
>> +
>> + hantro_reg_write(vpu, &av1_multicore_expect_context_update,
>> + !!(context_update_x == 0));
>> + hantro_reg_write(vpu, &av1_tile_enable,
>> + !!((tile_info.tile_cols > 1) || (tile_info.tile_rows > 1)));
>> + hantro_reg_write(vpu, &av1_num_tile_cols_8k, tile_info.tile_cols);
>> + hantro_reg_write(vpu, &av1_num_tile_rows_8k, tile_info.tile_rows);
>> + hantro_reg_write(vpu, &av1_context_update_tile_id,
>> + context_update_tile_id);
>> + hantro_reg_write(vpu, &av1_tile_transpose, 1);
>> + if (rockchip_vpu981_av1_tile_log2(tile_info.tile_cols) ||
>> + rockchip_vpu981_av1_tile_log2(tile_info.tile_rows))
>> + hantro_reg_write(vpu, &av1_dec_tile_size_mag, tile_info.tile_size_bytes - 1);
>> + else
>> + hantro_reg_write(vpu, &av1_dec_tile_size_mag, 3);
>> +
>> + hantro_write_addr(vpu, AV1_TILE_BASE, av1_dec->tile_info.dma);
>> +}
> Regards,
>
> Hans
>
Powered by blists - more mailing lists