linux-kernel - Re: [PATCH] hwmod: (pmbus) disable PEC if not enabled

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]

Date:   Tue, 19 Apr 2022 16:46:38 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     wujek dev <dev_public@...ek.eu>
Cc:     Jean Delvare <jdelvare@...e.com>, linux-hwmon@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] hwmod: (pmbus) disable PEC if not enabled

On 4/19/22 15:10, wujek dev wrote:
> ------- Original Message -------
> On Wednesday, April 20th, 2022 at 00:00, Guenter Roeck <linux@...ck-us.net> wrote:
>>
>>
>> On 4/19/22 13:53, Adam Wujek wrote:
>>
>>> Explicitly disable PEC when the client does not support it.
>>> Without the explicit disable, when the device with the PEC support is removed
>>> later when a device without PEC support is inserted into the same address,
>>> the driver uses the old value of client->flags which contains the I2C_CLIENT_PEC
>>> flag. As a consequence the PEC is used when it should not.
>>
>>
>> How can that happen ? I would assume the I2C device gets deleted and re-created
>> in that case, which should clear the PEC flag.
>>
>> Guenter
> In my case it was when I unloaded the driver for the I2C slave, changed the advertised PEC value in PMBUS_CAPABILITY register on slave. Then loaded the driver. When the switch was from disable->enable it worked as expected (this case was already covered), but when the PEC was set in the slave from enabled->disabled it was still using PEC to communicate.

So it is really the same device, only you unload the driver, change the
device configuration (presumably with i2cset commands), and load it
again. Please explain that in more detail in the commit description.

Thanks,
Guenter