From: Stewart Smith <stewart@linux.vnet.ibm.com>
To: Cyril Bur <cyrilbur@gmail.com>,
        Shilpasri G Bhat <shilpa.bhat@linux.vnet.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org,
        benh@kernel.crashing.org, paulus@samba.org, mpe@ellerman.id.au,
        ego@linux.vnet.ibm.com, svaidy@linux.vnet.ibm.com
Subject: Re: [PATCH V8 1/3] powernv: powercap: Add support for powercap framework
In-Reply-To: <1501205974.3324.1.camel@gmail.com>
References: <1501045509-21732-1-git-send-email-shilpa.bhat@linux.vnet.ibm.com> <1501045509-21732-2-git-send-email-shilpa.bhat@linux.vnet.ibm.com> <1501205974.3324.1.camel@gmail.com>
Date: Fri, 28 Jul 2017 17:18:55 +1000
MIME-Version: 1.0
Content-Type: text/plain
Message-Id: <877eyt1128.fsf@linux.vnet.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 853
Lines: 19

Cyril Bur <cyrilbur@gmail.com> writes:
> This is more for MPE and Stewart:
> If there is some mutual exclusion that needs to happen, it should be
> done in skiboot. We've had problems in the past with double entering
> skiboot and hard to interpret errors ending up in userspace, this
> solves that but it seems more like a coverup than an actual solution.

Yeah, I'd like us to be in a position where we don't force mutual
exclusion on such things out to the OS.

The real issue this papers over is the async token stuff you're
reworking. There's that, plus for this specific thing, you *could* go
and set a differnt powercap 180 times concurrently and we should do the
right thing in skiboot... but then you're sitting in skiboot rather than
sitting in linux being able to go run some other task on the thread.

-- 
Stewart Smith
OPAL Architect, IBM.