linux-kernel - [RFC PATCH 1/2] firmware: arm_scmi: Filter out negative results on scmi_reading

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: <20211220174155.40239-2-cristian.marussi@arm.com>
Date:   Mon, 20 Dec 2021 17:41:54 +0000
From:   Cristian Marussi <cristian.marussi@....com>
To:     linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org
Cc:     sudeep.holla@....com, f.fainelli@...il.com,
        souvik.chakravarty@....com, nicola.mazzucato@....com,
        Cristian Marussi <cristian.marussi@....com>
Subject: [RFC PATCH 1/2] firmware: arm_scmi: Filter out negative results on scmi_reading_get

When the backing SCMI Platform supports a Sensor protocol whose version is
greater than SCMI v2.0 (0x10000), the underlying SCMI SENSOR_READING_GET
command will report sensors readings no more as unsigned but as signed
values.

A new scmi_reading_get_timestamped operation was added to properly handle
this and other changes like timestamps and multi-axis sensors; on the other
side the Sensor scmi_reading_get protocol operation was kepts as it was for
backward compatibility and so it stil reports values as unsigned, but no
check was added to detect the situation in which a newer FW could try to
report negative values over an unsigned quantity.

Filter out negative values in scmi_reading_get, reporting an error, when
the SCMI FW version is greater than V2.0 and a negative value has been
received.

Reported-by: Nicola Mazzucato <nicola.mazzucato@....com>
Signed-off-by: Cristian Marussi <cristian.marussi@....com>
---
Note that, till now, HWMON SCMI driver was silently interpreting u64 as
s64, so after this change whenever a FW >2.0 is used it won't be possible
anymore to report negative values in HWMON.
---
 drivers/firmware/arm_scmi/sensors.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/firmware/arm_scmi/sensors.c b/drivers/firmware/arm_scmi/sensors.c
index cdbb287bd8bc..f3952f1d9f61 100644
--- a/drivers/firmware/arm_scmi/sensors.c
+++ b/drivers/firmware/arm_scmi/sensors.c
@@ -730,6 +730,14 @@ static int scmi_sensor_reading_get(const struct scmi_protocol_handle *ph,
 			*value = get_unaligned_le64(t->rx.buf);
 	}

+	if (!ret && si->version > SCMIv2_SENSOR_PROTOCOL && *value & BIT(63)) {
+		dev_warn_once(ph->dev,
+			      "SCMI FW Sensor version:0x%X reported negative value %ld\n",
+			      si->version, (long)*value);
+		*value = 0;
+		ret = -EIO;
+	}
+
 	ph->xops->xfer_put(ph, t);
 	return ret;
 }
-- 
2.17.1