Re: port-arm/53171 (Broadcom (and other ARM SoC i2c drivers) mis-handle indirect configuration of devices)

[Jason Thorpe <thorpej%me.com@localhost>]

> On Apr 30, 2018, at 1:38 PM, Jared McNeill <jmcneill%invisible.ca@localhost> wrote:
> 
> On Apr 30, 2018, at 5:21 PM, Jared McNeill <jmcneill%invisible.ca@localhost> wrote:
>> 
>> 
>>> On Apr 30, 2018, at 5:16 PM, Jared McNeill <jmcneill%invisible.ca@localhost> wrote:
>>> 
>>> I thought we had some drivers that relied on this but when I went to look I couldn’t find any..
>> 
>> Ok I found one! It’s clearly broken..
>> 
>> https://nxr.netbsd.org/xref/src/sys/dev/i2c/ibmhawk.c#96
> 
> I found three more, and apparently I’m the idiot responsible for all three of them. Checked in fixes for titemp, tcakp, and act8846pm.

Hey man, I wasn’t going to say anything… :-)

Anyway, I double-checked, and iic_search() already disallows IICCF_ADDR_DEFAULT, so that’s good.  “ibmhawk” is particularly egregious.  It looks like it’s expected at 0x37, so we can filter it out there, I suppose.  It would be even better if there were an ACPI node for it, or something.

(Speaking of ACPI — geez, that thing is just crying out to be converted to FDT at run-time :-) )

> 
> 

-- thorpej