From patchwork Mon Jan 22 10:35:44 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Maxime Ripard X-Patchwork-Id: 10178159 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 8C2F160224 for ; Mon, 22 Jan 2018 10:37:23 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 7D21427F82 for ; Mon, 22 Jan 2018 10:37:23 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 71E3527F88; Mon, 22 Jan 2018 10:37:23 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-4.2 required=2.0 tests=BAYES_00, RCVD_IN_DNSWL_MED autolearn=ham version=3.3.1 Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id F010C27F82 for ; Mon, 22 Jan 2018 10:37:22 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 46A766E1D2; Mon, 22 Jan 2018 10:36:36 +0000 (UTC) X-Original-To: dri-devel@lists.freedesktop.org Delivered-To: dri-devel@lists.freedesktop.org Received: from mail.free-electrons.com (mail.free-electrons.com [62.4.15.54]) by gabe.freedesktop.org (Postfix) with ESMTP id 3DF996E1C6 for ; Mon, 22 Jan 2018 10:36:21 +0000 (UTC) Received: by mail.free-electrons.com (Postfix, from userid 110) id 9E8EB20724; Mon, 22 Jan 2018 11:36:20 +0100 (CET) Received: from localhost (LStLambert-657-1-97-87.w90-63.abo.wanadoo.fr [90.63.216.87]) by mail.free-electrons.com (Postfix) with ESMTPSA id 5B16B2072C; Mon, 22 Jan 2018 11:35:56 +0100 (CET) From: Maxime Ripard To: Chen-Yu Tsai , Maxime Ripard , daniel.vetter@intel.com, jani.nikula@linux.intel.com, seanpaul@chromium.org Subject: [PATCH v2 15/19] drm/sun4i: backend: Check for the number of alpha planes Date: Mon, 22 Jan 2018 11:35:44 +0100 Message-Id: <7371f62a1385f2cbe3ed75dfca2e746338eb2286.1516617243.git-series.maxime.ripard@free-electrons.com> X-Mailer: git-send-email 2.14.3 In-Reply-To: References: In-Reply-To: References: Cc: linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, thomas@vitsch.nl X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.18 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" X-Virus-Scanned: ClamAV using ClamSMTP Due to the way the composition is done in hardware, we can only have a single alpha-enabled plane active at a time, placed in the second (highest priority) pipe. Make sure of that in our atomic_check to not end up in an impossible scenario. Signed-off-by: Maxime Ripard Reviewed-by: Chen-Yu Tsai --- drivers/gpu/drm/sun4i/sun4i_backend.c | 50 ++++++++++++++++++++++++++++- drivers/gpu/drm/sun4i/sun4i_backend.h | 2 +- drivers/gpu/drm/sun4i/sun4i_layer.c | 23 +------------- 3 files changed, 53 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/sun4i/sun4i_backend.c b/drivers/gpu/drm/sun4i/sun4i_backend.c index c4986054909b..eb1749d2c0d5 100644 --- a/drivers/gpu/drm/sun4i/sun4i_backend.c +++ b/drivers/gpu/drm/sun4i/sun4i_backend.c @@ -329,6 +329,8 @@ static int sun4i_backend_atomic_check(struct sunxi_engine *engine, struct drm_atomic_state *state = crtc_state->state; struct drm_device *drm = state->dev; struct drm_plane *plane; + unsigned int num_planes = 0; + unsigned int num_alpha_planes = 0; unsigned int num_frontend_planes = 0; DRM_DEBUG_DRIVER("Starting checking our planes\n"); @@ -341,6 +343,7 @@ static int sun4i_backend_atomic_check(struct sunxi_engine *engine, drm_atomic_get_plane_state(state, plane); struct sun4i_layer_state *layer_state = state_to_sun4i_layer_state(plane_state); + struct drm_framebuffer *fb = plane_state->fb; if (sun4i_backend_plane_uses_frontend(plane_state)) { DRM_DEBUG_DRIVER("Using the frontend for plane %d\n", @@ -351,6 +354,50 @@ static int sun4i_backend_atomic_check(struct sunxi_engine *engine, } else { layer_state->uses_frontend = false; } + + DRM_DEBUG_DRIVER("Plane FB format is %s\n", + drm_get_format_name(fb->format->format, + &format_name)); + if (fb->format->has_alpha) + num_alpha_planes++; + + num_planes++; + } + + /* + * The hardware is a bit unusual here. + * + * Even though it supports 4 layers, it does the composition + * in two separate steps. + * + * The first one is assigning a layer to one of its two + * pipes. If more that 1 layer is assigned to the same pipe, + * and if pixels overlaps, the pipe will take the pixel from + * the layer with the highest priority. + * + * The second step is the actual alpha blending, that takes + * the two pipes as input, and uses the eventual alpha + * component to do the transparency between the two. + * + * This two steps scenario makes us unable to guarantee a + * robust alpha blending between the 4 layers in all + * situations, since this means that we need to have one layer + * with alpha at the lowest position of our two pipes. + * + * However, we cannot even do that, since the hardware has a + * bug where the lowest plane of the lowest pipe (pipe 0, + * priority 0), if it has any alpha, will discard the pixel + * entirely and just display the pixels in the background + * color (black by default). + * + * Since means that we effectively have only three valid + * configurations with alpha, all of them with the alpha being + * on pipe1 with the lowest position, which can be 1, 2 or 3 + * depending on the number of planes and their zpos. + */ + if (num_alpha_planes > SUN4I_BACKEND_NUM_ALPHA_LAYERS) { + DRM_DEBUG_DRIVER("Too many planes with alpha, rejecting...\n"); + return -EINVAL; } if (num_frontend_planes > SUN4I_BACKEND_NUM_FRONTEND_LAYERS) { @@ -358,6 +405,9 @@ static int sun4i_backend_atomic_check(struct sunxi_engine *engine, return -EINVAL; } + DRM_DEBUG_DRIVER("State valid with %u planes, %u alpha, %u video\n", + num_planes, num_alpha_planes, num_frontend_planes); + return 0; } diff --git a/drivers/gpu/drm/sun4i/sun4i_backend.h b/drivers/gpu/drm/sun4i/sun4i_backend.h index 04a4f11b87a8..52e77591186a 100644 --- a/drivers/gpu/drm/sun4i/sun4i_backend.h +++ b/drivers/gpu/drm/sun4i/sun4i_backend.h @@ -146,6 +146,8 @@ #define SUN4I_BACKEND_HWCCOLORTAB_OFF 0x4c00 #define SUN4I_BACKEND_PIPE_OFF(p) (0x5000 + (0x400 * (p))) +#define SUN4I_BACKEND_NUM_LAYERS 4 +#define SUN4I_BACKEND_NUM_ALPHA_LAYERS 1 #define SUN4I_BACKEND_NUM_FRONTEND_LAYERS 1 struct sun4i_backend { diff --git a/drivers/gpu/drm/sun4i/sun4i_layer.c b/drivers/gpu/drm/sun4i/sun4i_layer.c index fbf25d59cf88..900e716443b8 100644 --- a/drivers/gpu/drm/sun4i/sun4i_layer.c +++ b/drivers/gpu/drm/sun4i/sun4i_layer.c @@ -201,32 +201,11 @@ struct drm_plane **sun4i_layers_init(struct drm_device *drm, struct sun4i_backend *backend = engine_to_sun4i_backend(engine); int i; - planes = devm_kcalloc(drm->dev, ARRAY_SIZE(sun4i_backend_planes) + 1, + planes = devm_kcalloc(drm->dev, SUN4I_BACKEND_NUM_LAYERS, sizeof(*planes), GFP_KERNEL); if (!planes) return ERR_PTR(-ENOMEM); - /* - * The hardware is a bit unusual here. - * - * Even though it supports 4 layers, it does the composition - * in two separate steps. - * - * The first one is assigning a layer to one of its two - * pipes. If more that 1 layer is assigned to the same pipe, - * and if pixels overlaps, the pipe will take the pixel from - * the layer with the highest priority. - * - * The second step is the actual alpha blending, that takes - * the two pipes as input, and uses the eventual alpha - * component to do the transparency between the two. - * - * This two steps scenario makes us unable to guarantee a - * robust alpha blending between the 4 layers in all - * situations. So we just expose two layers, one per pipe. On - * SoCs that support it, sprites could fill the need for more - * layers. - */ for (i = 0; i < ARRAY_SIZE(sun4i_backend_planes); i++) { const struct sun4i_plane_desc *plane = &sun4i_backend_planes[i]; struct sun4i_layer *layer;