linux-kernel - [PATCH] thermal/x86_pkg_temp_thermal: Add lower bound check for sysfs input

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-next>] [day] [month] [year] [list]

Message-Id: <20230308030711.4362-1-rui.zhang@intel.com>
Date:   Wed,  8 Mar 2023 11:07:11 +0800
From:   Zhang Rui <rui.zhang@...el.com>
To:     linux-pm@...r.kernel.org, rafael.j.wysocki@...el.com,
        daniel.lezcano@...aro.org
Cc:     linux-kernel@...r.kernel.org, error27@...il.com,
        srinivas.pandruvada@...el.com
Subject: [PATCH] thermal/x86_pkg_temp_thermal: Add lower bound check for sysfs input

When setting a trip point temperature from sysfs, there is an upper
bound check on the user input, but no lower bound check.

As hardware register has 7 bits for a trip point temperature, the offset
to tj_max of the input temperature must be equal to/less than 0x7f.
Or else,
1. bogus temperature is updated into the trip temperature bits.
2. the upper bits of the register can be polluted.

For example,
$ rdmsr 0x1b2
2000003
$ echo -180000 > /sys/class/thermal/thermal_zone1/trip_point_1_temp
$ rdmsr 0x1b2
3980003

Not only the trip point temp is set to 76C on this platform (tj_max is
100), the Power Notification (Bit 24) is also enabled erronously.

Fix the problem by adding lower bound check for sysfs input.

Reported-by: Dan Carpenter <error27@...il.com>
Link: https://lore.kernel.org/all/add7a378-4d50-4ba1-81d3-a0c17db25a0b@kili.mountain/
Signed-off-by: Zhang Rui <rui.zhang@...el.com>
---
Note that I didn't add a Fixes tag here because this problem always
exists, and even the location of this file has been changed.

Note that I didn't cc -stable because this patch doesn't apply to older
kernels due to recent changes. If Rafael thinks this is a stable
material, we can rebase this fix on previous kernel after this patch
hits upstream.
---
 drivers/thermal/intel/x86_pkg_temp_thermal.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/thermal/intel/x86_pkg_temp_thermal.c b/drivers/thermal/intel/x86_pkg_temp_thermal.c
index 1c2de84742df..128ee02596ed 100644
--- a/drivers/thermal/intel/x86_pkg_temp_thermal.c
+++ b/drivers/thermal/intel/x86_pkg_temp_thermal.c
@@ -124,14 +124,16 @@ sys_set_trip_temp(struct thermal_zone_device *tzd, int trip, int temp)
 {
 	struct zone_device *zonedev = tzd->devdata;
 	u32 l, h, mask, shift, intr;
-	int tj_max, ret;
+	int tj_max, val, ret;
 
 	tj_max = intel_tcc_get_tjmax(zonedev->cpu);
 	if (tj_max < 0)
 		return tj_max;
 	tj_max *= 1000;
 
-	if (trip >= MAX_NUMBER_OF_TRIPS || temp >= tj_max)
+	val = (tj_max - temp)/1000;
+
+	if (trip >= MAX_NUMBER_OF_TRIPS || val < 0 || val > 0x7f)
 		return -EINVAL;
 
 	ret = rdmsr_on_cpu(zonedev->cpu, MSR_IA32_PACKAGE_THERM_INTERRUPT,
@@ -156,7 +158,7 @@ sys_set_trip_temp(struct thermal_zone_device *tzd, int trip, int temp)
 	if (!temp) {
 		l &= ~intr;
 	} else {
-		l |= (tj_max - temp)/1000 << shift;
+		l |= val << shift;
 		l |= intr;
 	}
 
-- 
2.25.1