From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by master.gitmailbox.com (Postfix) with ESMTP id 56E8A4ABE4 for ; Tue, 14 May 2024 21:06:02 +0000 (UTC) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 1DD9368D756; Wed, 15 May 2024 00:06:00 +0300 (EEST) Received: from mail-wm1-f42.google.com (mail-wm1-f42.google.com [209.85.128.42]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 95EFE68D73A for ; Wed, 15 May 2024 00:05:53 +0300 (EEST) Received: by mail-wm1-f42.google.com with SMTP id 5b1f17b1804b1-4200ee78f34so27244555e9.3 for ; Tue, 14 May 2024 14:05:53 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=jkqxz-net.20230601.gappssmtp.com; s=20230601; t=1715720753; x=1716325553; darn=ffmpeg.org; h=content-transfer-encoding:in-reply-to:from:references:to :content-language:subject:user-agent:mime-version:date:message-id :from:to:cc:subject:date:message-id:reply-to; bh=sGewc87m4zsIKoeSzReTG7mRuaBic1KBQtxTl68iJvw=; b=Mb6Dqjdy97eGuIQVfQduk7boQfwAhJtR5+Jv9lU45q2vuLlLQa5XqqY7t4fdaIVrvb 9yYa+kjLtxj3CsH1alpM9TWAglQzPdPFrQMaM5z5KLwbCe67qvUlEXmc78UdUSWoWK9b QZFyKxw3JD6L7D0ZwZelBmzo3Ck7NBnMAm4J0GqwpeyGoLr2ZMK0KGXuzbeEapHiK7TI Lb9R4YXAx43MWvWJlE/GFqLLeS20kxZu/dUfy6wm/95lJ7idncXxJo6s45fT7kd8EoUL E3Zd8E1bhba1ukDLKnvQGHIvASih6rQe5N8BI3zQJgnDtixw9Z8+WYOObIJhZ3hV7po1 JNiw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1715720753; x=1716325553; h=content-transfer-encoding:in-reply-to:from:references:to :content-language:subject:user-agent:mime-version:date:message-id :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=sGewc87m4zsIKoeSzReTG7mRuaBic1KBQtxTl68iJvw=; b=eMWyZSeucpRoZYZ/5u8vFQ1VIDMP09GlWou2qtU33RzDinGfE8ALgsYrAhClwrGEho 1MFODY6av/6cyF904T8EF7+5brOL7qkAqN9cmIAo9meei0i0MEwo+lZzptw3Auq47LBc 3unesoYajoy6MP3U2OHcUOWIwIrxV8WaJJ5AiauOsSvlL0gMLWdQF6k2pPaXLePiw+xN +tgBFdzagiAUrahsxqpLrvVMJqbapYd+0jpqN37XkE1S9Bk8MATb3Ii5pJHSLXC/V4Zf IaFjsyBntAYgZaWEv6fj8jTwHjH0q5gSzQlJ9h0rYGiG7gpt6ffFEMfL9g+jk4YxlEi3 pRSg== X-Gm-Message-State: AOJu0YyeWskrQO3wGvmZLJ1h3qtJBzRAWukBoN5uqSayVy5xQEFfntv5 tBBoqfc6NTWzRc8FPdj8cotjPdD/gaieS/fdW6goMMT+F+/+ikGajbKxCdQlJmsLaCXoUtXJ4zg i X-Google-Smtp-Source: AGHT+IHSElzU8SOxYn0K2DmsoFzx6eGApzYctPwKZM3pXaENp3uhoYnsJo2ZThWi8lnaCP3Y8qedWQ== X-Received: by 2002:a05:600c:450e:b0:41e:c9ad:c729 with SMTP id 5b1f17b1804b1-41feac59d29mr102187505e9.28.1715720752761; Tue, 14 May 2024 14:05:52 -0700 (PDT) Received: from [192.168.0.15] (cpc92302-cmbg19-2-0-cust1183.5-4.cable.virginm.net. [82.1.212.160]) by smtp.gmail.com with ESMTPSA id 5b1f17b1804b1-4201088fe8csm125875365e9.32.2024.05.14.14.05.52 for (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Tue, 14 May 2024 14:05:52 -0700 (PDT) Message-ID: <90573a46-11d1-42f8-8651-70ad339e94e4@jkqxz.net> Date: Tue, 14 May 2024 22:06:21 +0100 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Content-Language: en-US To: ffmpeg-devel@ffmpeg.org References: <20240418085910.547-1-tong1.wu@intel.com> <20240418085910.547-11-tong1.wu@intel.com> From: Mark Thompson In-Reply-To: <20240418085910.547-11-tong1.wu@intel.com> Subject: Re: [FFmpeg-devel] [PATCH v8 11/15] avcodec/vaapi_encode: extract a get_recon_format function to base layer X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Archived-At: List-Archive: List-Post: On 18/04/2024 09:59, tong1.wu-at-intel.com@ffmpeg.org wrote: > From: Tong Wu > > Surface size and block size parameters are also moved to base layer. > > Signed-off-by: Tong Wu > --- > libavcodec/hw_base_encode.c | 58 +++++++++++++++++++++++ > libavcodec/hw_base_encode.h | 12 +++++ > libavcodec/vaapi_encode.c | 81 ++++++++------------------------- > libavcodec/vaapi_encode.h | 10 ---- > libavcodec/vaapi_encode_av1.c | 10 ++-- > libavcodec/vaapi_encode_h264.c | 11 +++-- > libavcodec/vaapi_encode_h265.c | 25 +++++----- > libavcodec/vaapi_encode_mjpeg.c | 5 +- > libavcodec/vaapi_encode_vp9.c | 6 +-- > 9 files changed, 118 insertions(+), 100 deletions(-) > > diff --git a/libavcodec/hw_base_encode.c b/libavcodec/hw_base_encode.c > index a4223d90f0..af85bb99aa 100644 > --- a/libavcodec/hw_base_encode.c > +++ b/libavcodec/hw_base_encode.c > @@ -693,6 +693,64 @@ int ff_hw_base_init_gop_structure(AVCodecContext *avctx, uint32_t ref_l0, uint32 > return 0; > } > > +int ff_hw_base_get_recon_format(AVCodecContext *avctx, const void *hwconfig, enum AVPixelFormat *fmt) > +{ > + HWBaseEncodeContext *ctx = avctx->priv_data; > + AVHWFramesConstraints *constraints = NULL; > + enum AVPixelFormat recon_format; > + int err, i; > + > + constraints = av_hwdevice_get_hwframe_constraints(ctx->device_ref, > + hwconfig); Does this mechanism actually make sense for D3D12? VAAPI is the currently the only implementation of this function with non-null hwconfig, and this is really relying on it to get useful information (otherwise the formats are just everything the device can allocate as a surface and the sizes are 0/INT_MAX). If D3D12 has something which would fit into the hwconfig method then this could work very nicely as well, but if it doesn't then presumably it does have some other calls to check things like the maximum frame size supported by the encoder and we should be using those rather than making this code generic? (Also consider Vulkan if possible; if two thirds of the cases want it then maybe we should do this even if it doesn't fit in one of them.) > + if (!constraints) { > + err = AVERROR(ENOMEM); > + goto fail; > + } > + > + // Probably we can use the input surface format as the surface format > + // of the reconstructed frames. If not, we just pick the first (only?) > + // format in the valid list and hope that it all works. > + recon_format = AV_PIX_FMT_NONE; > + if (constraints->valid_sw_formats) { > + for (i = 0; constraints->valid_sw_formats[i] != AV_PIX_FMT_NONE; i++) { > + if (ctx->input_frames->sw_format == > + constraints->valid_sw_formats[i]) { > + recon_format = ctx->input_frames->sw_format; > + break; > + } > + } > + if (recon_format == AV_PIX_FMT_NONE) { > + // No match. Just use the first in the supported list and > + // hope for the best. > + recon_format = constraints->valid_sw_formats[0]; > + } > + } else { > + // No idea what to use; copy input format. > + recon_format = ctx->input_frames->sw_format; > + } > + av_log(avctx, AV_LOG_DEBUG, "Using %s as format of " > + "reconstructed frames.\n", av_get_pix_fmt_name(recon_format)); > + > + if (ctx->surface_width < constraints->min_width || > + ctx->surface_height < constraints->min_height || > + ctx->surface_width > constraints->max_width || > + ctx->surface_height > constraints->max_height) { > + av_log(avctx, AV_LOG_ERROR, "Hardware does not support encoding at " > + "size %dx%d (constraints: width %d-%d height %d-%d).\n", > + ctx->surface_width, ctx->surface_height, > + constraints->min_width, constraints->max_width, > + constraints->min_height, constraints->max_height); > + err = AVERROR(EINVAL); > + goto fail; > + } > + > + *fmt = recon_format; > + err = 0; > +fail: > + av_hwframe_constraints_free(&constraints); > + return err; > +} > + > int ff_hw_base_encode_init(AVCodecContext *avctx) > { > HWBaseEncodeContext *ctx = avctx->priv_data; > diff --git a/libavcodec/hw_base_encode.h b/libavcodec/hw_base_encode.h > index d717f955d8..7686cf9501 100644 > --- a/libavcodec/hw_base_encode.h > +++ b/libavcodec/hw_base_encode.h > @@ -126,6 +126,16 @@ typedef struct HWBaseEncodeContext { > // Desired B frame reference depth. > int desired_b_depth; > > + // The required size of surfaces. This is probably the input > + // size (AVCodecContext.width|height) aligned up to whatever > + // block size is required by the codec. > + int surface_width; > + int surface_height; > + > + // The block size for slice calculations. > + int slice_block_width; > + int slice_block_height; > + > // The hardware device context. > AVBufferRef *device_ref; > AVHWDeviceContext *device; > @@ -210,6 +220,8 @@ int ff_hw_base_encode_receive_packet(AVCodecContext *avctx, AVPacket *pkt); > int ff_hw_base_init_gop_structure(AVCodecContext *avctx, uint32_t ref_l0, uint32_t ref_l1, > int flags, int prediction_pre_only); > > +int ff_hw_base_get_recon_format(AVCodecContext *avctx, const void *hwconfig, enum AVPixelFormat *fmt); > + > int ff_hw_base_encode_init(AVCodecContext *avctx); > > int ff_hw_base_encode_close(AVCodecContext *avctx); > ... Thanks, - Mark _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".