Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH 4.19 27/38] iommu/exynos: add missing put_device() call
 in exynos_iommu_of_xlate()
To:     Pavel Machek <pavel@denx.de>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc:     linux-kernel@vger.kernel.org, stable@vger.kernel.org,
        Yu Kuai <yukuai3@huawei.com>, Joerg Roedel <jroedel@suse.de>,
        Sasha Levin <sashal@kernel.org>
From:   Marek Szyprowski <m.szyprowski@samsung.com>
Message-ID: <5b869c86-8d35-e834-4fec-6b63a942e484@samsung.com>
Date:   Tue, 13 Oct 2020 10:54:15 +0200
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0)
        Gecko/20100101 Thunderbird/78.3.2
MIME-Version: 1.0
In-Reply-To: <20201007094737.GA12593@duo.ucw.cz>
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Content-Type: text/plain; charset="utf-8"
CMS-TYPE: 201P
References: <20201005142108.650363140@linuxfoundation.org>
        <20201005142109.977461657@linuxfoundation.org>
        <CGME20201007095256eucas1p150311eeced01b2cc66f6a9ef7061e6ff@eucas1p1.samsung.com>
        <20201007094737.GA12593@duo.ucw.cz>
Precedence: bulk

Hi Pavel,

On 07.10.2020 11:47, Pavel Machek wrote:
>> From: Yu Kuai <yukuai3@huawei.com>
>>
>> [ Upstream commit 1a26044954a6d1f4d375d5e62392446af663be7a ]
>>
>> if of_find_device_by_node() succeed, exynos_iommu_of_xlate() doesn't have
>> a corresponding put_device(). Thus add put_device() to fix the exception
>> handling for this function implementation.
> Okay, this looks reasonable, but...
>
> Do we miss put_device() in normal path, too? I'd expect another
> put_device at end of exynos_iommu_of_xlate() or perhaps in release
> path somewhere...

Frankly, there is no release path, so there is no need for put_device. 
Once initialized, Exynos IOMMU stays in the system forever. There is no 
point to remove IOMMU nor the API for that. Keeping increased refcount 
for its device just matches this behavior.

If the missing put_device() is really a problem, then we can move it 
from the error path just after data = platform_get_drvdata(sysmmu) 
assignment. Feel free to send a patch if you think this is a more 
appropriate approach.

Best regards

-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland