linux-kernel - [PATCH v2 1/2] net: dsa: LAN9303: Add early read to sync

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 2 Sep 2022 16:30:20 -0500
From:   Jerry Ray <jerry.ray@...rochip.com>
To:     Andrew Lunn <andrew@...n.ch>,
        Vivien Didelot <vivien.didelot@...il.com>,
        Florian Fainelli <f.fainelli@...il.com>,
        Vladimir Oltean <olteanv@...il.com>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>, <netdev@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>, Jerry Ray <jerry.ray@...rochip.com>
Subject: [PATCH v2 1/2] net: dsa: LAN9303: Add early read to sync

Add initial BYTE_ORDER read to sync the 32-bit accesses over the 16-bit
mdio bus to improve driver robustness.

The lan9303 expects two mdio read transactions back-to-back to read a
32-bit register. The first read transaction causes the other half of the
32-bit register to get latched.  The subsequent read returns the latched
second half of the 32-bit read. The BYTE_ORDER register is an exception to
this rule. As it is a constant value, there is no need to latch the second
half. We read this register first in case there were reads during the boot
loader process that might have occurred prior to this driver taking over
ownership of accessing this device.

This patch has been tested on the SAMA5D3-EDS with a LAN9303 RMII daughter
card.

Signed-off-by: Jerry Ray <jerry.ray@...rochip.com>
---
 drivers/net/dsa/lan9303-core.c | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

diff --git a/drivers/net/dsa/lan9303-core.c b/drivers/net/dsa/lan9303-core.c
index e03ff1f267bb..9d5302001abf 100644
--- a/drivers/net/dsa/lan9303-core.c
+++ b/drivers/net/dsa/lan9303-core.c
@@ -32,6 +32,7 @@
 #define LAN9303_INT_EN 0x17
 # define LAN9303_INT_EN_PHY_INT2_EN BIT(27)
 # define LAN9303_INT_EN_PHY_INT1_EN BIT(26)
+#define LAN9303_BYTE_ORDER 0x19
 #define LAN9303_HW_CFG 0x1D
 # define LAN9303_HW_CFG_READY BIT(27)
 # define LAN9303_HW_CFG_AMDX_EN_PORT2 BIT(26)
@@ -851,10 +852,6 @@ static int lan9303_check_device(struct lan9303 *chip)
 	if (ret) {
 		dev_err(chip->dev, "failed to read chip revision register: %d\n",
 			ret);
-		if (!chip->reset_gpio) {
-			dev_dbg(chip->dev,
-				"hint: maybe failed due to missing reset GPIO\n");
-		}
 		return ret;
 	}
 
@@ -1349,6 +1346,7 @@ static int lan9303_probe_reset_gpio(struct lan9303 *chip,
 int lan9303_probe(struct lan9303 *chip, struct device_node *np)
 {
 	int ret;
+	u32 reg;
 
 	mutex_init(&chip->indirect_mutex);
 	mutex_init(&chip->alr_mutex);
@@ -1359,6 +1357,19 @@ int lan9303_probe(struct lan9303 *chip, struct device_node *np)
 
 	lan9303_handle_reset(chip);
 
+	/* First read to the device.  This is a Dummy read to ensure MDIO */
+	/* access is in 32-bit sync. */
+	ret = lan9303_read(chip->regmap, LAN9303_BYTE_ORDER, &reg);
+	if (ret) {
+		dev_err(chip->dev, "failed to access the device: %d\n",
+			ret);
+		if (!chip->reset_gpio) {
+			dev_dbg(chip->dev,
+				"hint: maybe failed due to missing reset GPIO\n");
+		}
+		return ret;
+	}
+
 	ret = lan9303_check_device(chip);
 	if (ret)
 		return ret;
-- 
2.17.1