Received-SPF: pass (google.com: domain of linux-kernel+bounces-183084-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Message-ID: <6b0529a9-b1c6-44ba-8c45-ad233de6484e@linux.intel.com>
Date: Sun, 19 May 2024 18:07:42 +0800
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Cc: baolu.lu@linux.intel.com, "iommu@lists.linux.dev"
 <iommu@lists.linux.dev>,
 "virtualization@lists.linux-foundation.org"
 <virtualization@lists.linux-foundation.org>,
 "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
 Jason Gunthorpe <jgg@nvidia.com>
Subject: Re: [PATCH v5 1/9] iommu: Introduce domain attachment handle
To: "Tian, Kevin" <kevin.tian@intel.com>, Jason Gunthorpe <jgg@ziepe.ca>,
 Joerg Roedel <joro@8bytes.org>, Will Deacon <will@kernel.org>,
 Robin Murphy <robin.murphy@arm.com>,
 Jean-Philippe Brucker <jean-philippe@linaro.org>,
 Nicolin Chen <nicolinc@nvidia.com>, "Liu, Yi L" <yi.l.liu@intel.com>,
 Jacob Pan <jacob.jun.pan@linux.intel.com>,
 Joel Granados <j.granados@samsung.com>
References: <20240430145710.68112-1-baolu.lu@linux.intel.com>
 <20240430145710.68112-2-baolu.lu@linux.intel.com>
 <BN9PR11MB5276F0DAE4B5F63513C38A158CEC2@BN9PR11MB5276.namprd11.prod.outlook.com>
Content-Language: en-US
From: Baolu Lu <baolu.lu@linux.intel.com>
In-Reply-To: <BN9PR11MB5276F0DAE4B5F63513C38A158CEC2@BN9PR11MB5276.namprd11.prod.outlook.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

On 5/15/24 3:17 PM, Tian, Kevin wrote:
>> From: Lu Baolu <baolu.lu@linux.intel.com>
>> Sent: Tuesday, April 30, 2024 10:57 PM
>>
>> +/* Add an attach handle to the group's pasid array. */
>> +static struct iommu_attach_handle *
>> +iommu_attach_handle_set(struct iommu_domain *domain,
>> +			struct iommu_group *group, ioasid_t pasid)
>> +{
>> +	struct iommu_attach_handle *handle;
>> +	void *curr;
>> +
>> +	handle = kzalloc(sizeof(*handle), GFP_KERNEL);
>> +	if (!handle)
>> +		return ERR_PTR(-ENOMEM);
>> +
>> +	handle->domain = domain;
>> +	curr = xa_cmpxchg(&group->pasid_array, pasid, NULL, handle,
>> GFP_KERNEL);
> 
> this could be just xa_insert() which returns -EBUSY if the entry isn't NULL.

Yes, exactly.

> 
>> +	if (curr) {
>> +		kfree(handle);
>> +		return xa_err(curr) ? curr : ERR_PTR(-EBUSY);
> 
> 'curr' is not an error pointer. You need ERR_PTR(xa_err(curr)).

Fixed.

> 
>>   int iommu_attach_group(struct iommu_domain *domain, struct
>> iommu_group *group)
>>   {
>> +	struct iommu_attach_handle *handle;
>>   	int ret;
>>
>>   	mutex_lock(&group->mutex);
>> +	handle = iommu_attach_handle_set(domain, group,
>> IOMMU_NO_PASID);
>> +	if (IS_ERR(handle)) {
>> +		ret = PTR_ERR(handle);
>> +		goto out_unlock;
>> +	}
>>   	ret = __iommu_attach_group(domain, group);
>> +	if (ret)
>> +		goto out_remove_handle;
>>   	mutex_unlock(&group->mutex);
> 
> It's slightly better to set the handler after __iommu_attach_group().
> 
> doing so is aligned to the sequence in iommu_group_replace_domain().

I did it in this way because erasing an attach handle is easier than
rolling back the group to the previous domain from the perspective of
coding. I didn't realize attaching the group and storing the attach
handle had to be done in a specific order.

> 
>> @@ -2211,13 +2307,33 @@ EXPORT_SYMBOL_GPL(iommu_attach_group);
>>   int iommu_group_replace_domain(struct iommu_group *group,
>>   			       struct iommu_domain *new_domain)
>>   {
>> +	struct iommu_domain *old_domain = group->domain;
>> +	struct iommu_attach_handle *handle;
>>   	int ret;
>>
>>   	if (!new_domain)
>>   		return -EINVAL;
>>
>> +	if (new_domain == old_domain)
>> +		return 0;
>> +
> 
> __iommu_group_set_domain() already does the check.

Removed.

Best regards,
baolu