Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 17 Feb 2020 00:03:39 -0800
From:   Christoph Hellwig <hch@infradead.org>
To:     "Isaac J. Manjarres" <isaacm@codeaurora.org>
Cc:     iommu@lists.linux-foundation.org, linux-kernel@vger.kernel.org,
        Liam Mark <lmark@codeaurora.org>, joro@8bytes.org,
        pratikp@codeaurora.org, kernel-team@android.com
Subject: Re: [RFC PATCH] iommu/iova: Add a best-fit algorithm
Message-ID: <20200217080339.GC10342@infradead.org>
References: <1581721602-17010-1-git-send-email-isaacm@codeaurora.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1581721602-17010-1-git-send-email-isaacm@codeaurora.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, Feb 14, 2020 at 03:06:42PM -0800, Isaac J. Manjarres wrote:
> From: Liam Mark <lmark@codeaurora.org>
> 
> Using the best-fit algorithm, instead of the first-fit
> algorithm, may reduce fragmentation when allocating
> IOVAs.

As we like to say in standards groups:  may also implies may not.
Please provide numbers showing that this helps, and preferably and
explanation how it doesn't hurt as well.

> + * Should be called prior to using dma-apis.
> + */
> +int iommu_dma_enable_best_fit_algo(struct device *dev)
> +{
> +	struct iommu_domain *domain;
> +	struct iova_domain *iovad;
> +
> +	domain = iommu_get_domain_for_dev(dev);
> +	if (!domain || !domain->iova_cookie)
> +		return -EINVAL;
>  
> +	iovad = &((struct iommu_dma_cookie *)domain->iova_cookie)->iovad;
> +	iovad->best_fit = true;
>  	return 0;
>  }
> +EXPORT_SYMBOL(iommu_dma_enable_best_fit_algo);

Who is going to call this?  Patches adding exports always need a user
that goes along with the export.  Also drivers have no business calling
directly into dma-iommu.