Introduce iommu_mm_data structure to keep sva information (pasid and the
related sva domains). Add iommu_mm pointer, pointing to an instance of
iommu_mm_data structure, to mm.

Signed-off-by: Tina Zhang <tina.zhang@intel.com>
---
 include/linux/iommu.h    | 5 +++++
 include/linux/mm_types.h | 2 ++
 2 files changed, 7 insertions(+)

diff --git a/include/linux/iommu.h b/include/linux/iommu.h
index ab8784dfdbd98..937f3abc26f2e 100644
--- a/include/linux/iommu.h
+++ b/include/linux/iommu.h
@@ -670,6 +670,11 @@ struct iommu_sva {
 	struct iommu_domain		*domain;
 };
 
+struct iommu_mm_data {
+	u32			pasid;
+	struct list_head	sva_domains;
+};
+
 int iommu_fwspec_init(struct device *dev, struct fwnode_handle *iommu_fwnode,
 		      const struct iommu_ops *ops);
 void iommu_fwspec_free(struct device *dev);
diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
index 5e74ce4a28cd6..3fd65b7537f0e 100644
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
@@ -595,6 +595,7 @@ struct mm_cid {
 #endif
 
 struct kioctx_table;
+struct iommu_mm_data;
 struct mm_struct {
 	struct {
 		/*
@@ -808,6 +809,7 @@ struct mm_struct {
 
 #ifdef CONFIG_IOMMU_SVA
 		u32 pasid;
+		struct iommu_mm_data *iommu_mm;
 #endif
 #ifdef CONFIG_KSM
 		/*
-- 
2.34.1

On 2023/8/27 16:43, Tina Zhang wrote:
> Introduce iommu_mm_data structure to keep sva information (pasid and the
> related sva domains). Add iommu_mm pointer, pointing to an instance of
> iommu_mm_data structure, to mm.
> 
> Signed-off-by: Tina Zhang<tina.zhang@intel.com>
> ---
>   include/linux/iommu.h    | 5 +++++
>   include/linux/mm_types.h | 2 ++
>   2 files changed, 7 insertions(+)

Nit:

iommu also has a per-device private pointer, it's defined as struct
dev_iommu and stored at dev->iommu. Is it valuable to align both?

Best regards,
baolu

Hi Baolu,

> -----Original Message-----
> From: Baolu Lu <baolu.lu@linux.intel.com>
> Sent: Thursday, August 31, 2023 10:45 AM
> To: Zhang, Tina <tina.zhang@intel.com>; Jason Gunthorpe <jgg@ziepe.ca>;
> Tian, Kevin <kevin.tian@intel.com>; Michael Shavit <mshavit@google.com>
> Cc: baolu.lu@linux.intel.com; iommu@lists.linux.dev; linux-
> kernel@vger.kernel.org
> Subject: Re: [PATCH v2 3/5] mm: Add structure to keep sva information
> 
> On 2023/8/27 16:43, Tina Zhang wrote:
> > Introduce iommu_mm_data structure to keep sva information (pasid and
> > the related sva domains). Add iommu_mm pointer, pointing to an
> > instance of iommu_mm_data structure, to mm.
> >
> > Signed-off-by: Tina Zhang<tina.zhang@intel.com>
> > ---
> >   include/linux/iommu.h    | 5 +++++
> >   include/linux/mm_types.h | 2 ++
> >   2 files changed, 7 insertions(+)
> 
> Nit:
> 
> iommu also has a per-device private pointer, it's defined as struct dev_iommu
> and stored at dev->iommu. Is it valuable to align both?
I'm not sure if I understand the idea correctly. This struct dev_iommu is used to describe a collection per-device IOMMU data. Is the idea about migrating some bits from this struct dev_iommu to iommu_mm_data structure?

Thanks,
-Tina
> 
> Best regards,
> baolu

On 2023/9/1 11:36, Zhang, Tina wrote:
>> -----Original Message-----
>> From: Baolu Lu<baolu.lu@linux.intel.com>
>> Sent: Thursday, August 31, 2023 10:45 AM
>> To: Zhang, Tina<tina.zhang@intel.com>; Jason Gunthorpe<jgg@ziepe.ca>;
>> Tian, Kevin<kevin.tian@intel.com>; Michael Shavit<mshavit@google.com>
>> Cc:baolu.lu@linux.intel.com;iommu@lists.linux.dev; linux-
>> kernel@vger.kernel.org
>> Subject: Re: [PATCH v2 3/5] mm: Add structure to keep sva information
>>
>> On 2023/8/27 16:43, Tina Zhang wrote:
>>> Introduce iommu_mm_data structure to keep sva information (pasid and
>>> the related sva domains). Add iommu_mm pointer, pointing to an
>>> instance of iommu_mm_data structure, to mm.
>>>
>>> Signed-off-by: Tina Zhang<tina.zhang@intel.com>
>>> ---
>>>    include/linux/iommu.h    | 5 +++++
>>>    include/linux/mm_types.h | 2 ++
>>>    2 files changed, 7 insertions(+)
>> Nit:
>>
>> iommu also has a per-device private pointer, it's defined as struct dev_iommu
>> and stored at dev->iommu. Is it valuable to align both?
> I'm not sure if I understand the idea correctly. This struct dev_iommu is used to describe a collection per-device IOMMU data. Is the idea about migrating some bits from this struct dev_iommu to iommu_mm_data structure?

Never mind. I just thought about this when I was reading the patch. This
does not constitute any suggestion.

Best regards,
baolu