lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20240208-fd-migrate-mdp5-v4-1-945d08ef3fa8@linaro.org>
Date: Thu, 08 Feb 2024 17:01:08 +0200
From: Dmitry Baryshkov <dmitry.baryshkov@...aro.org>
To: Rob Clark <robdclark@...il.com>, 
 Abhinav Kumar <quic_abhinavk@...cinc.com>, Sean Paul <sean@...rly.run>, 
 Marijn Suijten <marijn.suijten@...ainline.org>, 
 David Airlie <airlied@...il.com>, Daniel Vetter <daniel@...ll.ch>
Cc: linux-arm-msm@...r.kernel.org, dri-devel@...ts.freedesktop.org, 
 freedreno@...ts.freedesktop.org, linux-kernel@...r.kernel.org
Subject: [PATCH v4 1/4] drm/msm/mdss: generate MDSS data for MDP5 platforms

Older (mdp5) platforms do not use per-SoC compatible strings. Instead
they use a single compat entry 'qcom,mdss'. To facilitate migrating
these platforms to the DPU driver provide a way to generate the MDSS /
UBWC data at runtime, when the DPU driver asks for it.

It is not possible to generate this data structure at the probe time,
since some platforms might not have MDP_CLK enabled, which makes reading
HW_REV register useless and prone to possible crashes.

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@...aro.org>
---
 drivers/gpu/drm/msm/msm_mdss.c | 51 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 51 insertions(+)

diff --git a/drivers/gpu/drm/msm/msm_mdss.c b/drivers/gpu/drm/msm/msm_mdss.c
index 35423d10aafa..65657230bbff 100644
--- a/drivers/gpu/drm/msm/msm_mdss.c
+++ b/drivers/gpu/drm/msm/msm_mdss.c
@@ -3,6 +3,7 @@
  * Copyright (c) 2018, The Linux Foundation
  */
 
+#include <linux/bitfield.h>
 #include <linux/clk.h>
 #include <linux/delay.h>
 #include <linux/interconnect.h>
@@ -213,6 +214,49 @@ static void msm_mdss_setup_ubwc_dec_40(struct msm_mdss *msm_mdss)
 	}
 }
 
+#define MDSS_HW_MAJ_MIN		GENMASK(31, 16)
+
+#define MDSS_HW_MSM8996		0x1007
+#define MDSS_HW_MSM8937		0x100e
+#define MDSS_HW_MSM8953		0x1010
+#define MDSS_HW_MSM8998		0x3000
+#define MDSS_HW_SDM660		0x3002
+#define MDSS_HW_SDM630		0x3003
+
+/*
+ * MDP5 platforms use generic qcom,mdp5 compat string, so we have to generate this data
+ */
+static const struct msm_mdss_data *msm_mdss_generate_mdp5_mdss_data(struct msm_mdss *mdss)
+{
+	struct msm_mdss_data *data;
+	u32 hw_rev;
+
+	data = devm_kzalloc(mdss->dev, sizeof(*data), GFP_KERNEL);
+	if (!data)
+		return NULL;
+
+	hw_rev = readl_relaxed(mdss->mmio + HW_REV);
+	hw_rev = FIELD_GET(MDSS_HW_MAJ_MIN, hw_rev);
+
+	if (hw_rev == MDSS_HW_MSM8996 ||
+	    hw_rev == MDSS_HW_MSM8937 ||
+	    hw_rev == MDSS_HW_MSM8953 ||
+	    hw_rev == MDSS_HW_MSM8998 ||
+	    hw_rev == MDSS_HW_SDM660 ||
+	    hw_rev == MDSS_HW_SDM630) {
+		data->ubwc_dec_version = UBWC_1_0;
+		data->ubwc_enc_version = UBWC_1_0;
+	}
+
+	if (hw_rev == MDSS_HW_MSM8996 ||
+	    hw_rev == MDSS_HW_MSM8998)
+		data->highest_bank_bit = 2;
+	else
+		data->highest_bank_bit = 1;
+
+	return data;
+}
+
 const struct msm_mdss_data *msm_mdss_get_mdss_data(struct device *dev)
 {
 	struct msm_mdss *mdss;
@@ -222,6 +266,13 @@ const struct msm_mdss_data *msm_mdss_get_mdss_data(struct device *dev)
 
 	mdss = dev_get_drvdata(dev);
 
+	/*
+	 * We could not do it at the probe time, since hw revision register was
+	 * not readable. Fill data structure now for the MDP5 platforms.
+	 */
+	if (!mdss->mdss_data && mdss->is_mdp5)
+		mdss->mdss_data = msm_mdss_generate_mdp5_mdss_data(mdss);
+
 	return mdss->mdss_data;
 }
 

-- 
2.39.2


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ