From patchwork Tue Oct 17 14:47:23 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Benjamin Gaignard X-Patchwork-Id: 736083 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 8C2A8C41513 for ; Tue, 17 Oct 2023 14:48:32 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1344307AbjJQOsb (ORCPT ); Tue, 17 Oct 2023 10:48:31 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:53468 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1344234AbjJQOsZ (ORCPT ); Tue, 17 Oct 2023 10:48:25 -0400 Received: from madras.collabora.co.uk (madras.collabora.co.uk [46.235.227.172]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id E752D11A; Tue, 17 Oct 2023 07:48:16 -0700 (PDT) Received: from benjamin-XPS-13-9310.. (unknown [IPv6:2a01:e0a:120:3210:7205:da49:a7e8:59f8]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) (Authenticated sender: benjamin.gaignard) by madras.collabora.co.uk (Postfix) with ESMTPSA id 54E126607375; Tue, 17 Oct 2023 15:48:15 +0100 (BST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=collabora.com; s=mail; t=1697554095; bh=XncDOWWMuPZSxD68Y9kDXDjEfVt/yc45C6+oTQD6xKw=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=eugusbxbRhpH6t3IJEtRO10Utlf0w1h9mdOPq3QxSjGG4IQpCzzJ7//dscMwfNzA+ hqC73V8yfPK+JiOrY3cCPS9xG/d69omFGiN0r3ZjPF+Mp2i3+7qhGYzzoUf0CPZL9I toBF4q5txeaIziTdACljNvBHmTFM6dknaxipMlOoEKG1EL4Ix3wnBotEMmmTFRDMj3 0+/JNrEisC0ml2sF62HT89yQVvC23PIEkYsLfQeoCn7CfiIhXtOzfK2pZblujEBMu0 djACjsvHNL/BTCpWmJGhUA3O2GAMgF2isfyysq9YATncOnO69BmawkJvlCkMEnaQwy SiHpLPD9ahdOQ== From: Benjamin Gaignard To: mchehab@kernel.org, tfiga@chromium.org, m.szyprowski@samsung.com, ming.qian@nxp.com, ezequiel@vanguardiasur.com.ar, p.zabel@pengutronix.de, gregkh@linuxfoundation.org, hverkuil-cisco@xs4all.nl, nicolas.dufresne@collabora.com Cc: linux-media@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-mediatek@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-rockchip@lists.infradead.org, linux-staging@lists.linux.dev, kernel@collabora.com, Benjamin Gaignard , Sergey Kozlov , Abylay Ospan Subject: [PATCH v12 23/56] media: pci: cx18: Stop direct calls to queue num_buffers field Date: Tue, 17 Oct 2023 16:47:23 +0200 Message-Id: <20231017144756.34719-24-benjamin.gaignard@collabora.com> X-Mailer: git-send-email 2.39.2 In-Reply-To: <20231017144756.34719-1-benjamin.gaignard@collabora.com> References: <20231017144756.34719-1-benjamin.gaignard@collabora.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-media@vger.kernel.org Use vb2_get_num_buffers() to avoid using queue num_buffers field directly. This allows us to change how the number of buffers is computed in the futur. Signed-off-by: Benjamin Gaignard CC: Sergey Kozlov CC: Abylay Ospan --- drivers/media/pci/cx18/cx18-streams.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/media/pci/cx18/cx18-streams.c b/drivers/media/pci/cx18/cx18-streams.c index 597472754c4c..cfbc4a907802 100644 --- a/drivers/media/pci/cx18/cx18-streams.c +++ b/drivers/media/pci/cx18/cx18-streams.c @@ -104,6 +104,7 @@ static int cx18_queue_setup(struct vb2_queue *vq, unsigned int *nbuffers, unsigned int *nplanes, unsigned int sizes[], struct device *alloc_devs[]) { + unsigned int q_num_bufs = vb2_get_num_buffers(vq); struct cx18_stream *s = vb2_get_drv_priv(vq); struct cx18 *cx = s->cx; unsigned int szimage; @@ -121,8 +122,8 @@ static int cx18_queue_setup(struct vb2_queue *vq, * Let's request at least three buffers: two for the * DMA engine and one for userspace. */ - if (vq->num_buffers + *nbuffers < 3) - *nbuffers = 3 - vq->num_buffers; + if (q_num_bufs + *nbuffers < 3) + *nbuffers = 3 - q_num_bufs; if (*nplanes) { if (*nplanes != 1 || sizes[0] < szimage)