svn commit: r323465 - head/usr.sbin/i2c

[Andriy Gapon]

On 10/10/2017 19:12, Ian Lepore wrote:
> i2c -s is not a thing that's done routinely in a production system or
> normal system operations... it's something a person does manually when
> trying to configure or debug a system.  In that situation, there is
> more harm in being told there are no working devices on the bus when in
> fact everything is fine, than there is some some hypothetical device
> doing some hypothetical "bad thing" in response to a read command.  In
> all my years of working with i2c stuff I've never seen a device doing
> anything more harmful than hanging the bus, requiring a reset (and even
> causing that requires worse behavior than an unexpected read).  On the
> other hand, I've seen a lot of people frustrated that i2c -s on freebsd
> says there are no devices, while the equivelent command on linux shows
> that everything is fine.

Okay.

However, I will just mention that in the past I used to own a system where
scanning the bus would make a slave that controlled CPU frequency to change it
to some garbage.  The system "just" crashed, but theoretically the damage could
have been worse.
Also, I own a system right now where scanning the bus results in something like
what you mentioned, but a little bit worse, the hanging bus that can be brought
back only by a power cycle (not even a warm reset).

-- 
Andriy Gapon