Date: Tue, 22 Nov 2016 15:38:24 -0800
From: Shaohua Li <shli@fb.com>
To: Tejun Heo <tj@kernel.org>
CC: <linux-block@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
        <Kernel-team@fb.com>, <axboe@fb.com>, <vgoyal@redhat.com>
Subject: Re: [PATCH V4 05/15] blk-throttle: add downgrade logic
Message-ID: <20161122233824.GD14628@shli-mbp.local>
References: <cover.1479161136.git.shli@fb.com>
 <d5670c98a3e4239202c4dad8f16add803a629f15.1479161136.git.shli@fb.com>
 <20161122212121.GC17534@htj.duckdns.org>
 <20161122214200.GE17534@htj.duckdns.org>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
In-Reply-To: <20161122214200.GE17534@htj.duckdns.org>
User-Agent: Mutt/1.6.1 (2016-04-27)
SpamDiagnosticOutput: 1:99
SpamDiagnosticMetadata: NSPM
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 22 Nov 2016 23:38:31.2899 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BN6PR15MB1635
X-OriginatorOrg: fb.com
X-Proofpoint-Spam-Reason: safe
X-FB-Internal: Safe
X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10432:,, definitions=2016-11-22_13:,,
 signatures=0
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
Content-Length: 1332
Lines: 29

On Tue, Nov 22, 2016 at 04:42:00PM -0500, Tejun Heo wrote:
> Hello,
> 
> On Tue, Nov 22, 2016 at 04:21:21PM -0500, Tejun Heo wrote:
> > 1. A cgroup and its high and max limits don't have much to do with
> >    other cgroups and their limits.  I don't get how the choice between
> >    high and max limits can be a td-wide state.
> 
> Ah, okay, this combines with idle cgroup detection to determine
> whether the cgroups should be allowed to exceed high limits.  It makes
> more sense to me now.  In that case, for the high/max limit range
> issues, the enforced high/max limits can simply follow what's implied
> by the configuration.  e.g. if high=100 max=80, just behave as if both
> high and max are 80.
> 
> > 2. Comparing parent's and child's limits and saying that either can be
> >    ignored because one is higher than the other isn't correct.  A
> >    parent's limit doesn't apply to each child separately.  It has to
> >    be aggregated.  e.g. you can ignore a parent's setting if the sum
> >    of all children's limits is smaller than the parent's but then
> >    again there could still be a lower limit higher up the tree, so
> >    they would still have to be examined.
> 
> This part still seems weird tho.  What am I misunderstanding?

You are right, the checks are unncessary. I'll delete them.

Thanks,
Shaohua