Message-ID: <5A5F45DA.5000104@rock-chips.com>
Date: Wed, 17 Jan 2018 20:47:22 +0800
From: JeffyChen <jeffy.chen@rock-chips.com>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:19.0) Gecko/20130126 Thunderbird/19.0
MIME-Version: 1.0
To: Robin Murphy <robin.murphy@arm.com>, linux-kernel@vger.kernel.org
CC: jcliang@chromium.org, xxm@rock-chips.com, tfiga@chromium.org,
        Heiko Stuebner <heiko@sntech.de>,
        linux-rockchip@lists.infradead.org,
        iommu@lists.linux-foundation.org, Joerg Roedel <joro@8bytes.org>,
        linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH v2 09/13] iommu/rockchip: Use iommu_group_get_for_dev()
 for add_device
References: <20180116132540.18939-1-jeffy.chen@rock-chips.com> <20180116132540.18939-10-jeffy.chen@rock-chips.com> <4c1b2469-9a32-c6ca-522f-73143b0fa618@arm.com>
In-Reply-To: <4c1b2469-9a32-c6ca-522f-73143b0fa618@arm.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org

Hi Robin,

On 01/17/2018 08:31 PM, Robin Murphy wrote:
> On 16/01/18 13:25, Jeffy Chen wrote:
>> IOMMU drivers are supposed to call this function instead of manually
>> creating a group in their .add_device callback. This behavior is not
>> strictly required by ARM DMA mapping implementation, but ARM64 already
>> relies on it. This patch fixes the rockchip-iommu driver to comply with
>> this requirement.
>
> FWIW that's not 100% true: what arm64 relies on is the group having a
> default DMA ops domain. Technically, you *could* open-code that in the
> driver's group allocation, but obviously using the appropriate existing
> API is nicer :)
ok, will rewrite the commit message.
>
> [...]
>> @@ -1182,6 +1164,29 @@ static void rk_iommu_remove_device(struct
>> device *dev)
>>       iommu_group_remove_device(dev);
>>   }
>> +static struct iommu_group *rk_iommu_device_group(struct device *dev)
>> +{
>> +    struct iommu_group *group;
>> +    int ret;
>> +
>> +    group = iommu_group_get(dev);
>> +    if (!group) {
>
> This check is pointless - if dev->iommu_group were non-NULL you wouldn't
> have been called in the first place.
right, it's allocated in the probe.
>
> Robin.