[Skiboot] [PATCH 9/9] [RFC] dts: Possible FIX for set trip bits

[Cédric Le Goater]

On 03/19/2018 09:12 AM, Cyril Bur wrote:
> On Mon, 2018-03-19 at 09:06 +0100, Cédric Le Goater wrote:
>> On 03/19/2018 05:54 AM, Cyril Bur wrote:
>>> It is possible that fixing CID 209503 will fix these FIXMEs. I do not
>>> know this code well at all and am in absolutely no position to test.
>>> Removing the FIXME needs much more scrutiny!!
>>
>> When first support for the POWER8 DTS was added a couple of years ago, 
>> the trip bits always had some 'bogus' value, which raised useless 
>> warnings at the userspace level in hwmon. Is this fixed now ? 
>>
>> What is 209503 about ? 
> 
> 209503 is about the dts parameter to these functions coming from
> uninitialised memory which may explain bits always being set. 

hmm, I don't see where the dts could be uninitialized. Anyhow, 
you can check the raw DTS values from userspace with this command :

	getscom <0x10000000 | 0x50000 | (core_id << 24)>

which returns something like 023f023f022f0000 on a palmetto.

C.