linux-kernel - Potential issue with smb word operations for tmp461 device in tmp401 driver

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: <SA1PR09MB7440BF952778F0DB8138747DE7B99@SA1PR09MB7440.namprd09.prod.outlook.com>
Date:   Fri, 15 Oct 2021 17:43:54 +0000
From:   "Wilson, David T. (GSFC-5870)" <david.wilson@...a.gov>
To:     "linux@...ck-us.net" <linux@...ck-us.net>
CC:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-hwmon@...r.kernel.org" <linux-hwmon@...r.kernel.org>
Subject: Potential issue with smb word operations for tmp461 device in tmp401
 driver

Hi,

I am reporting what I believe is a potential issue in the tmp401 driver for the tmp461 device specifically. I am new to reporting issues, so I apologize in advance if I've provided insufficient information for an issue report.

The problem I'm encountering is that when I use the tmp401 linux driver to read temperature values from the tmp461, all of the read temperature values end with 996 (e.g. 33996, 38996, etc...).

Looking further into the tmp401 commit messages, I see that the driver was changed to use smb word operations instead of separate byte operations. Although the other supported devices (i.e. tmp432, etc...) are noted to support 16-bit read operations in their respective datasheets, I see no indications of 16-bit read support in the tmp461 datasheet, which is supported by my inquiry in the TI forums (https://e2e.ti.com/support/sensors-group/sensors/f/sensors-forum/1044935/tmp461-linux-driver-support-and-16-bit-temperature-register-reads).

Reverting the driver to the commit before the smb word change, I am then able to read temperature values that do not end only with 996. As a result, I believe that the tmp461 support may be partially broken by the switch to smb word operations.

Thanks,
David Wilson