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 AF02BC021B1 for ; Thu, 20 Feb 2025 16:40:08 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:In-Reply-To:MIME-Version:References: Message-ID:Subject:Cc:To:From:Date:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=UJVOfh/pGaxNB+E32j2G2pVwHdrKyTjPF7Q1SJz9Fb4=; b=2NojToPj03Kg78 JtN9upvZsK9PmCImBvbdROajnIe1OreJZuX1triKfgNRYLsXsOfIEsVs6kb4ZCHaB/BeIgpCjHpVE hXkW1he2QvSX06sc9paYG5CM2vtblArWBT3Ut/57T7W9Q4wjF3ad2aHBRSbDctSaDHdTb5q0mWnw6 lXyzQsixUHenlBBqVdHdgsKeNq/l6bcHYqHOshIeQcL0tn/vGbG4lJEYY2TvGZtoF6/GCKTyn7Qmp 1qeNg6BKJCWYHhiCKyajzR1aD4uP8G5fiabQrMncQTnOwc3X/G6eWerpfHO+3f9sDT6J29dcg7Ff/ mNdLGrhF2Du1d/l4f9Sg==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.98 #2 (Red Hat Linux)) id 1tl9ak-00000001ue3-2frd; Thu, 20 Feb 2025 16:40:02 +0000 Received: from perceval.ideasonboard.com ([213.167.242.64]) by bombadil.infradead.org with esmtps (Exim 4.98 #2 (Red Hat Linux)) id 1tl9Zl-00000001uPb-1fRT for linux-rockchip@lists.infradead.org; Thu, 20 Feb 2025 16:39:02 +0000 Received: from pendragon.ideasonboard.com (81-175-209-231.bb.dnainternet.fi [81.175.209.231]) by perceval.ideasonboard.com (Postfix) with ESMTPSA id 953EC1193; Thu, 20 Feb 2025 17:37:36 +0100 (CET) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=ideasonboard.com; s=mail; t=1740069456; bh=NpVgzpSt57PVSSkjzpDrh/DNf40mgZQJo1PiwYskYnU=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=thGJhTqpVso78UUSwsWxI3jeLjd1QKzweP9dbUPIc/pZo6sbf4MsyA+VeBGnVMUes ei9Z6MzF6dMOhJwa3/+tQjMueCtvudqzmsS8/NnMLA7pyQu1aoeOLVQGwl8pAldMqz VE8mHbA9Y4abWBgVTohhqAIwSQZ7mf5dnfmRhUMM= Date: Thu, 20 Feb 2025 18:38:44 +0200 From: Laurent Pinchart To: Hans Verkuil Cc: Jacopo Mondi , Mauro Carvalho Chehab , Sakari Ailus , Dafna Hirschfeld , "open list:ROCKCHIP ISP V1 DRIVER" , Heiko Stuebner , "open list:ROCKCHIP ISP V1 DRIVER" , Adam Pigg Subject: Re: [PATCH v3] media: rkisp1: Remove min_queued_buffers Message-ID: <20250220163844.GI20111@pendragon.ideasonboard.com> References: <20241029082117.55385-1-jacopo.mondi@ideasonboard.com> <0d8d204f-c9b1-4152-a041-10d21f94af56@xs4all.nl> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <0d8d204f-c9b1-4152-a041-10d21f94af56@xs4all.nl> X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20250220_083901_583427_4670D573 X-CRM114-Status: GOOD ( 45.57 ) X-BeenThere: linux-rockchip@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: Upstream kernel work for Rockchip platforms List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: "Linux-rockchip" Errors-To: linux-rockchip-bounces+linux-rockchip=archiver.kernel.org@lists.infradead.org On Thu, Feb 20, 2025 at 04:33:28PM +0100, Hans Verkuil wrote: > On 2/20/25 15:22, Jacopo Mondi wrote: > > On Tue, Oct 29, 2024 at 09:21:16AM +0100, Jacopo Mondi wrote: > >> There apparently is no reason to require 3 queued buffers for RkISP1, > >> as the driver operates with a scratch buffer where data can be > >> directed to if there's no available buffer provided by userspace. > >> > >> Reduce the number of required buffers to 0 by removing the > >> initialization of min_queued_buffers, to allow applications to operate > >> by queueing capture buffers on-demand. > >> > >> Tested with libcamera, by operating with a single capture request. The > >> same request (and the associated capture buffer) gets recycled once > >> completed. This of course causes a frame rate drop but doesn't hinder > >> operations. > >> > >> Signed-off-by: Jacopo Mondi > > > > I just noticed v2 of this series: > > media: rkisp1: Reduce min_queued_buffers to 1 > > > > has been collected instead of this v3. > > > > And I noticed because a user complained to me about this. > > > > Now, I can provide an update based on the now merged v2, not a big > > deal, but this depresses me a bit as the discussion about > > implementing multi-commiter model is apparently (again) stalled. > > > > I know, sh*t happens (TM) and hiccups are expected in the process, > > we all make mistakes and I'm not even sure through which path the > > patch has been collected, but I could have handled this one easily, > > and instead what we have is: > > > > 1) an unhappy user that will likely have to wait for the next release > > 2) me having to send an additional (rather trivial) patch > > 3) Someone will have to review, collect, PR etc etc > > > > (and I'm not even mentioning this patch is 3 lines) > > > > Issues like this one seems to be considered a fact of life we decided > > is fine to live with, while every possible corner case of the proposed > > multi-committer model is analyzed with great concern like we're > > trading a perfect model for something that has to be equally perfect. > > > > And while I agree the biggest reason for the proverbial v4l2 slow pace > > is the reviewers scarcity and the limited maintainers bandwidth, now > > that we have everything in place to reduce the system clogginess > > it still seems we're not all sold for it. I really don't get it, sorry. > > The main gitlab blocker is that we need a merge bot. Ricardo said in the > last meeting that he hopes to have that in a month. The other blocker is > of course the gitlab freedesktop migration to another provider. > > Without a merge bot even with just two committers Mauro and myself are > frequently stepping on each others toes, so that really is needed. Heck, if > I have two CI pipelines in flight, I'm stepping on my own toes! Hopefully a > merge bot should fix this issue. > > Otherwise it is going quite well on the gitlab front. > > The other is the "Document the new media-committer's model" series. I > think Laurent scared Mauro off with his last set of comments. I think there may be less fundamental disagreements that one may fear, and that part of the trouble is caused by misunderstandings. E-mail communication makes it worse, I'd like to reiterate an invitation to discuss this in a video call. Decisions will still be made on the list, with everybody involved, but explaining our positions live will I think improve mutual understanding. > >> --- > >> v2->v3: > >> - Remove min_queued_buffers initialization > >> > >> v1->v2: > >> The first version of this patch set min_queued_buffers to 1, but setting it > >> to 0 doesn't compromise operations and it's even better as it allows application > >> to queue buffers to the capture devices on-demand. If a buffer is not provided > >> to the DMA engines, image data gets directed to the driver's internal scratch > >> buffer. > >> --- > >> drivers/media/platform/rockchip/rkisp1/rkisp1-capture.c | 3 --- > >> 1 file changed, 3 deletions(-) > >> > >> diff --git a/drivers/media/platform/rockchip/rkisp1/rkisp1-capture.c b/drivers/media/platform/rockchip/rkisp1/rkisp1-capture.c > >> index 2bddb4fa8a5c..2f0c610e74b9 100644 > >> --- a/drivers/media/platform/rockchip/rkisp1/rkisp1-capture.c > >> +++ b/drivers/media/platform/rockchip/rkisp1/rkisp1-capture.c > >> @@ -35,8 +35,6 @@ > >> #define RKISP1_SP_DEV_NAME RKISP1_DRIVER_NAME "_selfpath" > >> #define RKISP1_MP_DEV_NAME RKISP1_DRIVER_NAME "_mainpath" > >> > >> -#define RKISP1_MIN_BUFFERS_NEEDED 3 > >> - > >> enum rkisp1_plane { > >> RKISP1_PLANE_Y = 0, > >> RKISP1_PLANE_CB = 1, > >> @@ -1563,7 +1561,6 @@ static int rkisp1_register_capture(struct rkisp1_capture *cap) > >> q->ops = &rkisp1_vb2_ops; > >> q->mem_ops = &vb2_dma_contig_memops; > >> q->buf_struct_size = sizeof(struct rkisp1_buffer); > >> - q->min_queued_buffers = RKISP1_MIN_BUFFERS_NEEDED; > >> q->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC; > >> q->lock = &node->vlock; > >> q->dev = cap->rkisp1->dev; -- Regards, Laurent Pinchart _______________________________________________ Linux-rockchip mailing list Linux-rockchip@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-rockchip