Date: Wed, 2 Aug 2017 08:54:40 -0500 (CDT)
From: Christopher Lameter <cl@linux.com>
To: "Huang, Ying" <ying.huang@intel.com>
cc: Peter Zijlstra <peterz@infradead.org>, linux-kernel@vger.kernel.org,
        Tejun Heo <tj@kernel.org>, Joerg Roedel <joro@8bytes.org>,
        Ingo Molnar <mingo@kernel.org>, Michael Ellerman <mpe@ellerman.id.au>,
        Borislav Petkov <bp@suse.de>, Thomas Gleixner <tglx@linutronix.de>,
        Juergen Gross <jgross@suse.com>, Aaron Lu <aaron.lu@intel.com>
Subject: Re: [PATCH 0/3] IPI: Avoid to use 2 cache lines for one
 call_single_data
In-Reply-To: <20170802085220.4315-1-ying.huang@intel.com>
Message-ID: <alpine.DEB.2.20.1708020851150.23779@nuc-kabylake>
References: <20170802085220.4315-1-ying.huang@intel.com>
User-Agent: Alpine 2.20 (DEB 67 2015-01-07)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 504
Lines: 15

On Wed, 2 Aug 2017, Huang, Ying wrote:

> To allocate cache line size aligned percpu memory dynamically,
> alloc_percpu_aligned() is introduced and used in iova drivers too.

alloc_percpu() already aligns objects as specified when they are declared.

Moreover the function is improperly named since it aligns
to a cacheline(). If you want this then you would use

alloc_percpu_cacheline_aligned()

But then the alignment can already be requested by adding
__cacheline_aligned to the per cpu definition.