From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id C7197CCA472 for ; Mon, 6 Oct 2025 09:59:07 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender:List-Subscribe:List-Help :List-Post:List-Archive:List-Unsubscribe:List-Id:In-Reply-To:Content-Type: MIME-Version:References:Message-ID:Subject:Cc:To:From:Date:Reply-To: Content-Transfer-Encoding:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=dvr3+xWGnZz9G0cdntzSKJQjrC5YwByuehU4ShGTxsg=; b=UsZdhZwcJl413Bp9VKfwDbo9q+ G4W0pr3qnzMX42s8Xnee+OoEcMa986+7Pj4O2v3dB13+WedJj2dISC6klFzD2VSrVK02dR03q4f+h Owpl1m//cqh8smLZLOxhJx9w1Ld1MalwRfiYzXL8xlsXeISb5VDP12OLL2DmCvALqZvsZLCqAFAyE 0sdHVTLoaNpW+j0h4F0AhdS4LJ4MQ/iac++amtpiSaxYGmdHbPG2QUL7VcrN2LE49ybEfAURMkHCT Y+3FVnIaLP/5iktLXfoheE8+b4wRWMXasRJe5NqkylLzzj3WC6zXrSW+bytwpOzSfzpmZcWV23zqX QMwbGJdg==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.98.2 #2 (Red Hat Linux)) id 1v5hzh-0000000HIOy-3T14; Mon, 06 Oct 2025 09:59:01 +0000 Received: from perceval.ideasonboard.com ([213.167.242.64]) by bombadil.infradead.org with esmtps (Exim 4.98.2 #2 (Red Hat Linux)) id 1v5hzf-0000000HIOY-2VAn; Mon, 06 Oct 2025 09:59:01 +0000 Received: from pendragon.ideasonboard.com (81-175-209-231.bb.dnainternet.fi [81.175.209.231]) by perceval.ideasonboard.com (Postfix) with UTF8SMTPSA id 4E2CC1741; Mon, 6 Oct 2025 11:57:25 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=ideasonboard.com; s=mail; t=1759744645; bh=kLrtsqAeDp4iPiBTibIAcBMqFwrsbVJb/iiWR+ZRV2U=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=vCYUaB18q2WsrdeqHwbl+K+ZyH7Nht9p87fF3kmDxGyH5L2KqM5upc3QyUGAjXKNj EQmilWMf+GeVuo2ObRFfabrGID0e5EHGUN9YwbyiotqJ5VJry9cD1TJtOEFtkAn7fp acCsxupaopO8NgIFl3ZIlouiwyYx4xh8UU7ZR+aQ= Date: Mon, 6 Oct 2025 12:58:50 +0300 From: Laurent Pinchart To: Dan Scally Cc: Jacopo Mondi , Dafna Hirschfeld , Keke Li , Mauro Carvalho Chehab , Heiko Stuebner , Sakari Ailus , Antoine Bouyer , linux-kernel@vger.kernel.org, linux-media@vger.kernel.org, linux-rockchip@lists.infradead.org, linux-arm-kernel@lists.infradead.org Subject: Re: [PATCH v5 1/8] media: uapi: Introduce V4L2 generic ISP types Message-ID: <20251006095850.GC5944@pendragon.ideasonboard.com> References: <20250915-extensible-parameters-validation-v5-0-e6db94468af3@ideasonboard.com> <20250915-extensible-parameters-validation-v5-1-e6db94468af3@ideasonboard.com> <20251005000602.GA13055@pendragon.ideasonboard.com> <20251006082751.GA16422@pendragon.ideasonboard.com> <246ae834-3077-4e38-b419-5c8773139e65@ideasonboard.com> <20251006090637.GA5944@pendragon.ideasonboard.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20251006_025859_777830_C10D3487 X-CRM114-Status: GOOD ( 50.35 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org On Mon, Oct 06, 2025 at 10:51:33AM +0100, Daniel Scally wrote: > On 06/10/2025 10:06, Laurent Pinchart wrote: > > On Mon, Oct 06, 2025 at 09:46:26AM +0100, Daniel Scally wrote: > >> On 06/10/2025 09:27, Laurent Pinchart wrote: > >>> On Mon, Oct 06, 2025 at 09:15:47AM +0100, Daniel Scally wrote: > >>>> On 05/10/2025 01:06, Laurent Pinchart wrote: > >>>>> On Mon, Sep 15, 2025 at 07:18:10PM +0200, Jacopo Mondi wrote: > >>>>>> Introduce v4l2-isp.h in the Linux kernel uAPI. > >>>>>> > >>>>>> The header includes types for generic ISP configuration parameters > >>>>>> and will be extended in future with support for generic ISP statistics > >>>>> > >>>>> s/in future/in the future/ > >>>>> > >>>>> (and you can reflow the commit message) > >>>>> > >>>>>> formats. > >>>>>> > >>>>>> Generic ISP parameters support is provided by introducing two new > >>>>>> types that represent an extensible and versioned buffer of ISP > >>>>>> configuration parameters. > >>>>>> > >>>>>> The v4l2_params_block_header structure represents the header to be > >>>>>> prepend to each ISP configuration block and the v4l2_params_buffer type > >>>>>> represents the base type for the configuration parameters buffer. > >>>>> > >>>>> The second part of the sentence describes the same structure as the next > >>>>> paragraph. > >>>>> > >>>>>> The v4l2_params_buffer represents the container for the ISP > >>>>>> configuration data block. The generic type is defined with a 0-sized > >>>>>> data member that the ISP driver implementations shall properly size > >>>>>> according to their capabilities. > >>>>> > >>>>> This will be easier to understand if you describe v4l2_params_buffer > >>>>> first. > >>>>> > >>>>> The commit message would benefit from being rewritten. > >>>>> > >>>>>> [Add v4l2_params_buffer_size()] > >>>>>> Signed-off-by: Daniel Scally > >>>>>> Reviewed-by: Daniel Scally > >>>>>> Signed-off-by: Jacopo Mondi > >>>>>> --- > >>>>>> MAINTAINERS | 6 +++ > >>>>>> include/uapi/linux/media/v4l2-isp.h | 100 ++++++++++++++++++++++++++++++++++++ > >>>>>> 2 files changed, 106 insertions(+) > >>>>>> > >>>>>> diff --git a/MAINTAINERS b/MAINTAINERS > >>>>>> index ee8cb2db483f6a5e96b62b6f2edd05b1427b69f5..e82c3d0758d6033fe8fcd56ffde2c03c4319fd11 100644 > >>>>>> --- a/MAINTAINERS > >>>>>> +++ b/MAINTAINERS > >>>>>> @@ -26410,6 +26410,12 @@ F: drivers/media/i2c/vd55g1.c > >>>>>> F: drivers/media/i2c/vd56g3.c > >>>>>> F: drivers/media/i2c/vgxy61.c > >>>>>> > >>>>>> +V4L2 GENERIC ISP PARAMETERS AND STATISTIC FORMATS > >>>>>> +M: Jacopo Mondi > >>>>>> +L: linux-media@vger.kernel.org > >>>>>> +S: Maintained > >>>>>> +F: include/uapi/linux/media/v4l2-isp.h > >>>>>> + > >>>>>> VF610 NAND DRIVER > >>>>>> M: Stefan Agner > >>>>>> L: linux-mtd@lists.infradead.org > >>>>>> diff --git a/include/uapi/linux/media/v4l2-isp.h b/include/uapi/linux/media/v4l2-isp.h > >>>>>> new file mode 100644 > >>>>>> index 0000000000000000000000000000000000000000..b838555dce2b290a14136ab09ea4d2dfdc95b26b > >>>>>> --- /dev/null > >>>>>> +++ b/include/uapi/linux/media/v4l2-isp.h > >>>>>> @@ -0,0 +1,100 @@ > >>>>>> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ > >>>>>> +/* > >>>>>> + * Video4Linux2 generic ISP parameters and statistics support > >>>>>> + * > >>>>>> + * Copyright (C) 2025 Ideas On Board Oy > >>>>>> + * Author: Jacopo Mondi > >>>>>> + */ > >>>>>> + > >>>>>> +#ifndef _UAPI_V4L2_ISP_H_ > >>>>>> +#define _UAPI_V4L2_ISP_H_ > >>>>>> + > >>>>>> +#include > >>>>>> +#include > >>>>>> + > >>>>>> +#define V4L2_PARAMS_FL_BLOCK_DISABLE (1U << 0) > >>>>>> +#define V4L2_PARAMS_FL_BLOCK_ENABLE (1U << 1) > >>>>>> + > >>>>>> +/* > >>>>>> + * Reserve the first 8 bits for V4L2_PARAMS_FL_* flag. > >>>>>> + * > >>>>>> + * Driver-specific flags should be defined as: > >>>>>> + * #define PLATFORM_SPECIFIC_FLAG0 ((1U << V4L2_PARAMS_FL_DRIVER_FLAGS(0)) > >>>>> > >>>>> s/PLATFORM/DRIVER/ > >>>>> > >>>>>> + * #define PLATFORM_SPECIFIC_FLAG1 ((1U << V4L2_PARAMS_FL_DRIVER_FLAGS(1)) > >>>>>> + */ > >>>>>> +#define V4L2_PARAMS_FL_DRIVER_FLAGS(n) ((n) + 8) > >>>>>> + > >>>>>> +/** > >>>>>> + * struct v4l2_params_block_header - V4L2 extensible parameters block header > >>>>>> + * > >>>>>> + * This structure represents the common part of all the ISP configuration > >>>>>> + * blocks. Each parameters block shall embed an instance of this structure type > >>>>>> + * as its first member, followed by the block-specific configuration data. The > >>>>>> + * driver inspects this common header to discern the block type and its size and > >>>>>> + * properly handle the block content. > >>>>> > >>>>> The last sentence is not relevant for the UAPI. > >>>>> > >>>>>> + * > >>>>>> + * The @type field is an ISP driver-specific value that identifies the block > >>>>>> + * type. The @size field specifies the size of the parameters block. > >>>>>> + * > >>>>>> + * The @flags field is a bitmask of per-block flags V4L2_PARAMS_FL_* and > >>>>>> + * driver-specific flags specified by the driver header. > >>>>>> + * > >>>>>> + * @type: The parameters block type (driver-specific) > >>>>>> + * @flags: A bitmask of block flags (driver-specific) > >>>>>> + * @size: Size (in bytes) of the parameters block, including this header > >>>>> > >>>>> I think the fields usually go right after the structure name, followed > >>>>> by the rest of the documentation. > >>>>> > >>>>>> + */ > >>>>>> +struct v4l2_params_block_header { > >>>>>> + __u16 type; > >>>>>> + __u16 flags; > >>>>>> + __u32 size; > >>>>>> +} __attribute__((aligned(8))); > >>>>>> + > >>>>>> +/** > >>>>>> + * v4l2_params_buffer_size - Calculate size of v4l2_params_buffer for a platform > >>>>>> + * > >>>>>> + * Users of the v4l2 extensible parameters will have differing sized data arrays > >>>>>> + * depending on their specific parameter buffers. Drivers and userspace will > >>>>>> + * need to be able to calculate the appropriate size of the struct to > >>>>>> + * accommodate all ISP configuration blocks provided by the platform. > >>>>>> + * This macro provides a convenient tool for the calculation. > >>>>>> + * > >>>>>> + * @max_params_size: The total size of the ISP configuration blocks > >>>>>> + */ > >>>>>> +#define v4l2_params_buffer_size(max_params_size) \ > >>>>>> + (offsetof(struct v4l2_params_buffer, data) + (max_params_size)) > >>>>> > >>>>> This isn't used in this series as far as I can tell, and neither is it > >>>>> used in your libcamera implementation. I'd drop the macro (as well as > >>>>> the mention in the commit message). > >>>> > >>>> This is because the rkisp1 and c3 ISP implementations are already merged with a custom parameters > >>>> buffer struct defined at [1] and [2]. There the array size is set to the max size macro. This series > >>>> just asserts that the header part is a size match for struct v4l2_params_buffer to guarantee > >>>> compatibility, so throughout those drivers they can use sizeof(struct c3_isp_params_cfg) or the > >>>> rkisp equivalent and it's fine. > >>>> > >>>> For new users that don't have a custom struct in their uAPI and are relying on struct > >>>> v4l2_params_buffer we can't just do sizeof(), so the idea was for this to provide a canonical way > >>>> for the size calculation to be done across both the kernel and userspace. > >>>> > >>>> If we decide that it's worth keeping but want a user in this series to justify its inclusion, it > >>>> could be used in patches 2 and 3 - I'll reply to patch 2 in a second to indicate where. > >>> > >>> Will it be used by userspace too for the Mali C55, or only by the kernel > >> > >> I have used it in userspace too. > > > > Any pointer to the code ? > > I didn't post my commits to adapt the mali-c55 IPA to the extensible parameters, but it's basically > the same as Jacopo's patch here: > https://lists.libcamera.org/pipermail/libcamera-devel/2025-October/053567.html > > And the change I made using the macro was in IPAMaliC55::fillParams(): > > - memset(params, 0, sizeof(mali_c55_params_buffer)); > + memset(params, 0, v4l2_params_buffer_size(MALI_C55_PARAMS_MAX_SIZE)); Given that the buffer should be sized for the worst case, wouldn't it be better to memset .planes()[0].data().size() ? > >>> driver ? > >>> > >>>> [1] https://elixir.bootlin.com/linux/v6.17/source/include/uapi/linux/rkisp1-config.h#L1675 > >>>> [2] https://elixir.bootlin.com/linux/v6.17/source/include/uapi/linux/media/amlogic/c3-isp-config.h#L558> > >>>> > >>>>>> + > >>>>>> +/** > >>>>>> + * struct v4l2_params_buffer - V4L2 extensible parameters configuration > >>>>>> + * > >>>>>> + * This struct contains the configuration parameters of the ISP algorithms, > >>>>> > >>>>> s/struct/structure/ > >>>>> > >>>>>> + * serialized by userspace into a data buffer. Each configuration parameter > >>>>>> + * block is represented by a block-specific structure which contains a > >>>>>> + * :c:type:`v4l2_params_block_header` entry as first member. Userspace populates > >>>>>> + * the @data buffer with configuration parameters for the blocks that it intends > >>>>>> + * to configure. As a consequence, the data buffer effective size changes > >>>>>> + * according to the number of ISP blocks that userspace intends to configure and > >>>>>> + * is set by userspace in the @data_size field. > >>>>>> + * > >>>>>> + * The parameters buffer is versioned by the @version field to allow modifying > >>>>>> + * and extending its definition. Userspace shall populate the @version field to > >>>>>> + * inform the driver about the version it intends to use. The driver will parse > >>>>>> + * and handle the @data buffer according to the data layout specific to the > >>>>>> + * indicated version and return an error if the desired version is not > >>>>>> + * supported. > >>>>>> + * > >>>>>> + * For each ISP block that userspace wants to configure, a block-specific > >>>>>> + * structure is appended to the @data buffer, one after the other without gaps > >>>>>> + * in between nor overlaps. Userspace shall populate the @data_size field with > >>>>> > >>>>> I think you can drop "nor overlaps", nobody in their right mind should > >>>>> think the blocks could be overlayed :-) > >>>>> > >>>>>> + * the effective size, in bytes, of the @data buffer. > >>>>>> + * > >>>>>> + * @version: The parameters buffer version (driver-specific) > >>>>>> + * @data_size: The configuration data effective size, excluding this header > >>>>>> + * @data: The configuration data > >>>>> > >>>>> Move the fields up just after the structure name. > >>>>> > >>>>> Reviewed-by: Laurent Pinchart > >>>>> > >>>>>> + */ > >>>>>> +struct v4l2_params_buffer { > >>>>>> + __u32 version; > >>>>>> + __u32 data_size; > >>>>>> + __u8 data[] __counted_by(data_size); > >>>>>> +}; > >>>>>> + > >>>>>> +#endif /* _UAPI_V4L2_ISP_H_ */ -- Regards, Laurent Pinchart