{"id":22201,"url":"https://patchwork.libcamera.org/api/1.1/patches/22201/?format=json","web_url":"https://patchwork.libcamera.org/patch/22201/","project":{"id":1,"url":"https://patchwork.libcamera.org/api/1.1/projects/1/?format=json","name":"libcamera","link_name":"libcamera","list_id":"libcamera_core","list_email":"libcamera-devel@lists.libcamera.org","web_url":"","scm_url":"","webscm_url":""},"msgid":"<20241206101344.767170-10-stefan.klug@ideasonboard.com>","date":"2024-12-06T10:13:31","name":"[v3,09/17] libcamera: converter: Add functions to adjust config","commit_ref":null,"pull_url":null,"state":"accepted","archived":false,"hash":"a22569f8bf35c636d0e699d22a32a77dbca10c08","submitter":{"id":184,"url":"https://patchwork.libcamera.org/api/1.1/people/184/?format=json","name":"Stefan Klug","email":"stefan.klug@ideasonboard.com"},"delegate":{"id":17,"url":"https://patchwork.libcamera.org/api/1.1/users/17/?format=json","username":"epaul","first_name":"Paul","last_name":"Elder","email":"paul.elder@ideasonboard.com"},"mbox":"https://patchwork.libcamera.org/patch/22201/mbox/","series":[{"id":4854,"url":"https://patchwork.libcamera.org/api/1.1/series/4854/?format=json","web_url":"https://patchwork.libcamera.org/project/libcamera/list/?series=4854","date":"2024-12-06T10:13:22","name":"rkisp1: Fix aspect ratio and ScalerCrop","version":3,"mbox":"https://patchwork.libcamera.org/series/4854/mbox/"}],"comments":"https://patchwork.libcamera.org/api/patches/22201/comments/","check":"pending","checks":"https://patchwork.libcamera.org/api/patches/22201/checks/","tags":{},"headers":{"Return-Path":"","X-Original-To":"parsemail@patchwork.libcamera.org","Delivered-To":"parsemail@patchwork.libcamera.org","Received":["from lancelot.ideasonboard.com (lancelot.ideasonboard.com\n\t[92.243.16.209])\n\tby patchwork.libcamera.org (Postfix) with ESMTPS id 15114BF415\n\tfor ;\n\tFri, 6 Dec 2024 10:14:21 +0000 (UTC)","from lancelot.ideasonboard.com (localhost [IPv6:::1])\n\tby lancelot.ideasonboard.com (Postfix) with ESMTP id 84B3966149;\n\tFri, 6 Dec 2024 11:14:20 +0100 (CET)","from perceval.ideasonboard.com (perceval.ideasonboard.com\n\t[IPv6:2001:4b98:dc2:55:216:3eff:fef7:d647])\n\tby lancelot.ideasonboard.com (Postfix) with ESMTPS id 4E56466147\n\tfor ;\n\tFri, 6 Dec 2024 11:14:14 +0100 (CET)","from ideasonboard.com (unknown\n\t[IPv6:2a00:6020:448c:6c00:3543:aebe:e043:ef86])\n\tby perceval.ideasonboard.com (Postfix) with ESMTPSA id DCEB874C;\n\tFri, 6 Dec 2024 11:13:44 +0100 (CET)"],"Authentication-Results":"lancelot.ideasonboard.com; dkim=pass (1024-bit key;\n\tunprotected) header.d=ideasonboard.com header.i=@ideasonboard.com\n\theader.b=\"rolPBWeI\"; dkim-atps=neutral","DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/simple; d=ideasonboard.com;\n\ts=mail; t=1733480025;\n\tbh=OVUhbjooxzBOWajQ6a2W1ABbAAkYO8boaBBU0dPBAzk=;\n\th=From:To:Cc:Subject:Date:In-Reply-To:References:From;\n\tb=rolPBWeIMZh+7iM5EjhVyuEHuH+dhcm03DZtZ7oKtvzSzV3jGtG859DqiAVKhgJu4\n\tquPcTCna6kdltih5OY3+9ifIRwXO6Z9Upe/acwji4FPuK+BD+JhDQIph11JtEOtWb9\n\tWYMSOZDDGXJJjEJoy1cF/gSmvjQQn6CCRxfkOOIk=","From":"Stefan Klug ","To":"libcamera-devel@lists.libcamera.org","Cc":"Jacopo Mondi ","Subject":"[PATCH v3 09/17] libcamera: converter: Add functions to adjust\n\tconfig","Date":"Fri, 6 Dec 2024 11:13:31 +0100","Message-ID":"<20241206101344.767170-10-stefan.klug@ideasonboard.com>","X-Mailer":"git-send-email 2.43.0","In-Reply-To":"<20241206101344.767170-1-stefan.klug@ideasonboard.com>","References":"<20241206101344.767170-1-stefan.klug@ideasonboard.com>","MIME-Version":"1.0","Content-Transfer-Encoding":"8bit","X-BeenThere":"libcamera-devel@lists.libcamera.org","X-Mailman-Version":"2.1.29","Precedence":"list","List-Id":"","List-Unsubscribe":",\n\t","List-Archive":"","List-Post":"","List-Help":"","List-Subscribe":",\n\t","Errors-To":"libcamera-devel-bounces@lists.libcamera.org","Sender":"\"libcamera-devel\" "},"content":"From: Jacopo Mondi \n\nAdd to the Converter interface two functions used by pipeline handlers\nto validate and adjust the converter input and output configurations\nby specifying the desired alignment for the adjustment.\n\nAdd the adjustInputSize() and adjustOutputSize() functions that allows\nto adjust the converter input/output sizes with the desired alignment.\n\nAdd a validateOutput() function meant to be used by the pipeline\nhandler implementations of validate(). The function adjusts a\nStreamConfiguration to a valid configuration produced by the Converter.\n\nSigned-off-by: Jacopo Mondi \n\n---\nChanges in v3:\n- Added this patch\n---\n include/libcamera/internal/converter.h | 17 ++\n .../internal/converter/converter_v4l2_m2m.h | 11 ++\n src/libcamera/converter.cpp | 43 +++++\n .../converter/converter_v4l2_m2m.cpp | 169 ++++++++++++++++++\n 4 files changed, 240 insertions(+)","diff":"diff --git a/include/libcamera/internal/converter.h b/include/libcamera/internal/converter.h\nindex ffbb6f345cd5..9213ae5b9c33 100644\n--- a/include/libcamera/internal/converter.h\n+++ b/include/libcamera/internal/converter.h\n@@ -41,6 +41,13 @@ public:\n \n \tusing Features = Flags;\n \n+\tenum class Alignment {\n+\t\tDown = 0,\n+\t\tUp = (1 << 0),\n+\t};\n+\n+\tusing Alignments = Flags;\n+\n \tConverter(MediaDevice *media, Features features = Feature::None);\n \tvirtual ~Converter();\n \n@@ -51,9 +58,19 @@ public:\n \tvirtual std::vector formats(PixelFormat input) = 0;\n \tvirtual SizeRange sizes(const Size &input) = 0;\n \n+\tvirtual Size adjustInputSize(const PixelFormat &pixFmt,\n+\t\t\t\t const Size &size,\n+\t\t\t\t Alignments align = Alignment::Down) = 0;\n+\tvirtual Size adjustOutputSize(const PixelFormat &pixFmt,\n+\t\t\t\t const Size &size,\n+\t\t\t\t Alignments align = Alignment::Down) = 0;\n+\n \tvirtual std::tuple\n \tstrideAndFrameSize(const PixelFormat &pixelFormat, const Size &size) = 0;\n \n+\tvirtual int validateOutput(StreamConfiguration *cfg, bool *adjusted,\n+\t\t\t\t Alignments align = Alignment::Down) = 0;\n+\n \tvirtual int configure(const StreamConfiguration &inputCfg,\n \t\t\t const std::vector> &outputCfgs) = 0;\n \tvirtual int exportBuffers(const Stream *stream, unsigned int count,\ndiff --git a/include/libcamera/internal/converter/converter_v4l2_m2m.h b/include/libcamera/internal/converter/converter_v4l2_m2m.h\nindex a5286166f574..89bd2878b190 100644\n--- a/include/libcamera/internal/converter/converter_v4l2_m2m.h\n+++ b/include/libcamera/internal/converter/converter_v4l2_m2m.h\n@@ -47,6 +47,11 @@ public:\n \tstd::tuple\n \tstrideAndFrameSize(const PixelFormat &pixelFormat, const Size &size);\n \n+\tSize adjustInputSize(const PixelFormat &pixFmt,\n+\t\t\t const Size &size, Alignments align = Alignment::Down) override;\n+\tSize adjustOutputSize(const PixelFormat &pixFmt,\n+\t\t\t const Size &size, Alignments align = Alignment::Down) override;\n+\n \tint configure(const StreamConfiguration &inputCfg,\n \t\t const std::vector> &outputCfg);\n \tint exportBuffers(const Stream *stream, unsigned int count,\n@@ -55,6 +60,9 @@ public:\n \tint start();\n \tvoid stop();\n \n+\tint validateOutput(StreamConfiguration *cfg, bool *adjusted,\n+\t\t\t Alignments align = Alignment::Down) override;\n+\n \tint queueBuffers(FrameBuffer *input,\n \t\t\t const std::map &outputs);\n \n@@ -101,6 +109,9 @@ private:\n \t\tstd::pair inputCropBounds_;\n \t};\n \n+\tSize adjustSizes(const Size &size, const std::vector &ranges,\n+\t\t\t Alignments align);\n+\n \tstd::unique_ptr m2m_;\n \n \tstd::map> streams_;\ndiff --git a/src/libcamera/converter.cpp b/src/libcamera/converter.cpp\nindex aa16970cd446..c3da162b7de7 100644\n--- a/src/libcamera/converter.cpp\n+++ b/src/libcamera/converter.cpp\n@@ -50,6 +50,21 @@ LOG_DEFINE_CATEGORY(Converter)\n * \\brief A bitwise combination of features supported by the converter\n */\n \n+/**\n+ * \\enum Converter::Alignment\n+ * \\brief The alignment mode specified when adjusting the converter input or\n+ * output sizes\n+ * \\var Converter::Alignment::Down\n+ * \\brief Adjust the Converter sizes to a smaller valid size\n+ * \\var Converter::Alignment::Up\n+ * \\brief Adjust the Converter sizes to a larger valid size\n+ */\n+\n+/**\n+ * \\typedef Converter::Alignments\n+ * \\brief A bitwise combination of alignments supported by the converter\n+ */\n+\n /**\n * \\brief Construct a Converter instance\n * \\param[in] media The media device implementing the converter\n@@ -110,6 +125,24 @@ Converter::~Converter()\n * \\return A range of output image sizes\n */\n \n+/**\n+ * \\fn Converter::adjustInputSize()\n+ * \\brief Adjust the converter input \\a size to a valid value\n+ * \\param[in] pixFmt The pixel format of the converter input stream\n+ * \\param[in] size The converter input size to adjust to a valid value\n+ * \\param[in] align The desired alignment\n+ * \\return The adjusted converter input size\n+ */\n+\n+/**\n+ * \\fn Converter::adjustOutputSize()\n+ * \\brief Adjust the converter output \\a size to a valid value\n+ * \\param[in] pixFmt The pixel format of the converter output stream\n+ * \\param[in] size The converter output size to adjust to a valid value\n+ * \\param[in] align The desired alignment\n+ * \\return The adjusted converter output size\n+ */\n+\n /**\n * \\fn Converter::strideAndFrameSize()\n * \\brief Retrieve the output stride and frame size for an input configutation\n@@ -118,6 +151,16 @@ Converter::~Converter()\n * \\return A tuple indicating the stride and frame size or an empty tuple on error\n */\n \n+/**\n+ * \\fn Converter::validateOutput()\n+ * \\brief Validate and possibily adjust \\a cfg to a valid converter output\n+ * \\param[inout] cfg The StreamConfiguration to validate and adjust\n+ * \\param[out] adjusted Set to true if \\a cfg has been adjusted\n+ * \\param[in] align The desired alignment\n+ * \\return 0 if \\a cfg is valid or has been adjusted, a negative error code\n+ * otherwise if \\a cfg cannot be adjusted\n+ */\n+\n /**\n * \\fn Converter::configure()\n * \\brief Configure a set of output stream conversion from an input stream\ndiff --git a/src/libcamera/converter/converter_v4l2_m2m.cpp b/src/libcamera/converter/converter_v4l2_m2m.cpp\nindex bd7e5cce600d..6857472b29f2 100644\n--- a/src/libcamera/converter/converter_v4l2_m2m.cpp\n+++ b/src/libcamera/converter/converter_v4l2_m2m.cpp\n@@ -8,6 +8,7 @@\n \n #include \"libcamera/internal/converter/converter_v4l2_m2m.h\"\n \n+#include \n #include \n \n #include \n@@ -400,6 +401,127 @@ V4L2M2MConverter::strideAndFrameSize(const PixelFormat &pixelFormat,\n \treturn std::make_tuple(format.planes[0].bpl, format.planes[0].size);\n }\n \n+/**\n+ * \\copydoc libcamera::Converter::adjustInputSize\n+ */\n+Size V4L2M2MConverter::adjustInputSize(const PixelFormat &pixFmt,\n+\t\t\t\t const Size &size, Alignments align)\n+{\n+\tauto formats = m2m_->output()->formats();\n+\tV4L2PixelFormat v4l2PixFmt = m2m_->output()->toV4L2PixelFormat(pixFmt);\n+\n+\tauto it = formats.find(v4l2PixFmt);\n+\tif (it == formats.end()) {\n+\t\tLOG(Converter, Info)\n+\t\t\t<< \"Unsupported pixel format \" << pixFmt;\n+\t\treturn {};\n+\t}\n+\n+\treturn adjustSizes(size, it->second, align);\n+}\n+\n+/**\n+ * \\copydoc libcamera::Converter::adjustOutputSize\n+ */\n+Size V4L2M2MConverter::adjustOutputSize(const PixelFormat &pixFmt,\n+\t\t\t\t\tconst Size &size, Alignments align)\n+{\n+\tauto formats = m2m_->capture()->formats();\n+\tV4L2PixelFormat v4l2PixFmt = m2m_->capture()->toV4L2PixelFormat(pixFmt);\n+\n+\tauto it = formats.find(v4l2PixFmt);\n+\tif (it == formats.end()) {\n+\t\tLOG(Converter, Info)\n+\t\t\t<< \"Unsupported pixel format \" << pixFmt;\n+\t\treturn {};\n+\t}\n+\n+\treturn adjustSizes(size, it->second, align);\n+}\n+\n+Size V4L2M2MConverter::adjustSizes(const Size &cfgSize,\n+\t\t\t\t const std::vector &ranges,\n+\t\t\t\t Alignments align)\n+{\n+\tSize size = cfgSize;\n+\n+\tif (ranges.size() == 1) {\n+\t\t/*\n+\t\t * The device supports either V4L2_FRMSIZE_TYPE_CONTINUOUS or\n+\t\t * V4L2_FRMSIZE_TYPE_STEPWISE.\n+\t\t */\n+\t\tconst SizeRange &range = *ranges.begin();\n+\n+\t\tsize.width = std::clamp(size.width, range.min.width,\n+\t\t\t\t\trange.max.width);\n+\t\tsize.height = std::clamp(size.height, range.min.height,\n+\t\t\t\t\t range.max.height);\n+\n+\t\t/*\n+\t\t * Check if any alignment is needed. If the sizes are already\n+\t\t * aligned, or the device supports V4L2_FRMSIZE_TYPE_CONTINUOUS\n+\t\t * with hStep and vStep equal to 1, we're done here.\n+\t\t */\n+\t\tint widthR = size.width % range.hStep;\n+\t\tint heightR = size.height % range.vStep;\n+\n+\t\t/* Align up or down according to the caller request. */\n+\n+\t\tif (widthR != 0)\n+\t\t\tsize.width = size.width - widthR\n+\t\t\t\t + ((align == Alignment::Up) ? range.hStep : 0);\n+\n+\t\tif (heightR != 0)\n+\t\t\tsize.height = size.height - heightR\n+\t\t\t\t + ((align == Alignment::Up) ? range.vStep : 0);\n+\t} else {\n+\t\t/*\n+\t\t * The device supports V4L2_FRMSIZE_TYPE_DISCRETE, find the\n+\t\t * size closer to the requested output configuration.\n+\t\t *\n+\t\t * The size ranges vector is not ordered, so we sort it first.\n+\t\t * If we align up, start from the larger element.\n+\t\t */\n+\t\tstd::vector sizes(ranges.size());\n+\t\tstd::transform(ranges.begin(), ranges.end(), std::back_inserter(sizes),\n+\t\t\t [](const SizeRange &range) { return range.max; });\n+\t\tstd::sort(sizes.begin(), sizes.end());\n+\n+\t\tif (align == Alignment::Up)\n+\t\t\tstd::reverse(sizes.begin(), sizes.end());\n+\n+\t\t/*\n+\t\t * Return true if s2 is valid according to the desired\n+\t\t * alignment: smaller than s1 if we align down, larger than s1\n+\t\t * if we align up.\n+\t\t */\n+\t\tauto nextSizeValid = [](const Size &s1, const Size &s2, Alignments a) {\n+\t\t\treturn a == Alignment::Down\n+\t\t\t\t? (s1.width > s2.width && s1.height > s2.height)\n+\t\t\t\t: (s1.width < s2.width && s1.height < s2.height);\n+\t\t};\n+\n+\t\tSize newSize;\n+\t\tfor (const Size &sz : sizes) {\n+\t\t\tif (!nextSizeValid(size, sz, align))\n+\t\t\t\tbreak;\n+\n+\t\t\tnewSize = sz;\n+\t\t}\n+\n+\t\tif (newSize.isNull()) {\n+\t\t\tLOG(Converter, Error)\n+\t\t\t\t<< \"Cannot adjust \" << cfgSize\n+\t\t\t\t<< \" to a supported converter size\";\n+\t\t\treturn {};\n+\t\t}\n+\n+\t\tsize = newSize;\n+\t}\n+\n+\treturn size;\n+}\n+\n /**\n * \\copydoc libcamera::Converter::configure\n */\n@@ -507,6 +629,53 @@ void V4L2M2MConverter::stop()\n \t\titer.second->stop();\n }\n \n+/**\n+ * \\copydoc libcamera::Converter::validateOutput\n+ */\n+int V4L2M2MConverter::validateOutput(StreamConfiguration *cfg, bool *adjusted,\n+\t\t\t\t Alignments align)\n+{\n+\tV4L2VideoDevice *capture = m2m_->capture();\n+\tV4L2VideoDevice::Formats fmts = capture->formats();\n+\n+\tif (adjusted)\n+\t\t*adjusted = false;\n+\n+\tPixelFormat fmt = cfg->pixelFormat;\n+\tV4L2PixelFormat v4l2PixFmt = capture->toV4L2PixelFormat(fmt);\n+\n+\tauto it = fmts.find(v4l2PixFmt);\n+\tif (it == fmts.end()) {\n+\t\tit = fmts.begin();\n+\t\tv4l2PixFmt = it->first;\n+\t\tcfg->pixelFormat = v4l2PixFmt.toPixelFormat();\n+\n+\t\tif (adjusted)\n+\t\t\t*adjusted = true;\n+\n+\t\tLOG(Converter, Info)\n+\t\t\t<< \"Converter output pixel format adjusted to \"\n+\t\t\t<< cfg->pixelFormat;\n+\t}\n+\n+\tconst Size cfgSize = cfg->size;\n+\tcfg->size = adjustSizes(cfgSize, it->second, align);\n+\n+\tif (cfg->size.isNull())\n+\t\treturn -EINVAL;\n+\n+\tif (cfg->size.width != cfgSize.width ||\n+\t cfg->size.height != cfgSize.height) {\n+\t\tLOG(Converter, Info)\n+\t\t\t<< \"Converter size adjusted to \"\n+\t\t\t<< cfg->size;\n+\t\tif (adjusted)\n+\t\t\t*adjusted = true;\n+\t}\n+\n+\treturn 0;\n+}\n+\n /**\n * \\copydoc libcamera::Converter::queueBuffers\n */\n","prefixes":["v3","09/17"]}