Re: [PATCH] i2c: mpc: Use atomic read and fix break condition

[wsa@xxxxxxxxxx]

> we'd hit the 100us timeout in the poll). But I see no evidence of that 
> actually happening (and no idea what arbitration lost means w.r.t i2c).

On a bus with multiple masters, it means the other master has won the
arbitration because the address it wants to talk to contains more 0 bits.

> I don't know that there is a maximum clock stretch time (we certainly 
> know there are misbehaving devices that hold SCL low forever). The SMBUS 
> protocol adds some timeouts but as far as I know i2c says nothing about 
> how long a remote device can hold SCL.

The above is all correct.

Even with the unclear situation about the 100us, I think this should go
to for-current soon, right?

Attachment: signature.asc
Description: PGP signature