linux-kernel - [PATCH v2 2/2] i2c: i801: don't instantiate spd5118 under SPD Write Disable

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: <20250430-for-upstream-i801-spd5118-no-instantiate-v2-2-2f54d91ae2c7@canonical.com>
Date: Wed, 30 Apr 2025 19:26:17 +0800
From: "Yo-Jung (Leo) Lin" <leo.lin@...onical.com>
To: Jean Delvare <jdelvare@...e.com>, Andi Shyti <andi.shyti@...nel.org>, 
 Wolfram Sang <wsa+renesas@...g-engineering.com>
Cc: Guenter Roeck <linux@...ck-us.net>, 
 "Chia-Lin Kao (AceLan)" <acelan.kao@...onical.com>, 
 linux-i2c@...r.kernel.org, linux-kernel@...r.kernel.org, 
 "Yo-Jung Lin (Leo)" <leo.lin@...onical.com>
Subject: [PATCH v2 2/2] i2c: i801: don't instantiate spd5118 under SPD
 Write Disable

If SPD Write Disable bit in the SMBus is enabled, writing data to
addresses from 0x50 to 0x57 is forbidden. This may lead to the
following issues for spd5118 devices:

  1) Writes to the sensor hwmon sysfs attributes will always result
     in ENXIO.

  2) During system-wide resume, errors may occur during regcache sync,
     resulting in the following error messages:

     kernel: spd5118 1-0050: failed to write b = 0: -6
     kernel: spd5118 1-0050: pm: dpm_run_callback(): spd5118_resume [spd5118] returns -6
     kernel: spd5118 1-0050: pm: failed to resume async: error -6

  3) nvmem won't be usable, because writing to the page selector becomes
     impossible.

Also, BIOS vendors may choose to set the page to a value != 0 after a board
reset. This will make the sensor not functional unless its MR11 register
can be changed, which is impossible due to writes being disabled.

To address these issues, don't instantiate it at all if the SPD Write Disable
bit is set.

Signed-off-by: Yo-Jung (Leo) Lin <leo.lin@...onical.com>
---
 drivers/i2c/busses/i2c-i801.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/drivers/i2c/busses/i2c-i801.c b/drivers/i2c/busses/i2c-i801.c
index a7f89946dad4..88474409e82d 100644
--- a/drivers/i2c/busses/i2c-i801.c
+++ b/drivers/i2c/busses/i2c-i801.c
@@ -1177,10 +1177,12 @@ static void i801_probe_optional_targets(struct i801_priv *priv)
 		dmi_walk(dmi_check_onboard_devices, &priv->adapter);
 
 	/* Instantiate SPD EEPROMs unless the SMBus is multiplexed */
-#ifdef CONFIG_I2C_I801_MUX
-	if (!priv->mux_pdev)
-#endif
-		i2c_register_spd_write_enable(&priv->adapter);
+	if (!IS_ENABLED(CONFIG_I2C_I801_MUX) || !priv->mux_pdev) {
+		if (priv->original_hstcfg & SMBHSTCFG_SPD_WD)
+			i2c_register_spd_write_disable(&priv->adapter);
+		else
+			i2c_register_spd_write_enable(&priv->adapter);
+	}
 }
 #else
 static void __init input_apanel_init(void) {}
@@ -1283,7 +1285,10 @@ static int i801_notifier_call(struct notifier_block *nb, unsigned long action,
 		return NOTIFY_DONE;
 
 	/* Call i2c_register_spd for muxed child segments */
-	i2c_register_spd_write_enable(to_i2c_adapter(dev));
+	if (priv->original_hstcfg & SMBHSTCFG_SPD_WD)
+		i2c_register_spd_write_disable(to_i2c_adapter(dev));
+	else
+		i2c_register_spd_write_enable(to_i2c_adapter(dev));
 
 	return NOTIFY_OK;
 }

-- 
2.43.0