Re: [PATCH v12 00/16] iommu: Prepare to deliver page faults to user space

From: Baolu Lu
Date: Thu Feb 08 2024 - 21:40:04 EST


On 2024/2/8 17:03, Zhangfei Gao wrote:
Hi, Baolu

On Wed, 7 Feb 2024 at 09:39, Lu Baolu <baolu.lu@xxxxxxxxxxxxxxx <mailto:baolu.lu@xxxxxxxxxxxxxxx>> wrote:

When a user-managed page table is attached to an IOMMU, it is necessary
to deliver IO page faults to user space so that they can be handled
appropriately. One use case for this is nested translation, which is
currently being discussed in the mailing list.

I have posted a RFC series [1] that describes the implementation of
delivering page faults to user space through IOMMUFD. This series has
received several comments on the IOMMU refactoring, which I am trying to
address in this series.

The major refactoring includes:

- [PATCH 01 ~ 04] Move include/uapi/linux/iommu.h to
  include/linux/iommu.h. Remove the unrecoverable fault data
definition.
- [PATCH 05 ~ 06] Remove iommu_[un]register_device_fault_handler().
- [PATCH 07 ~ 10] Separate SVA and IOPF. Make IOPF a generic page fault
  handling framework.
- [PATCH 11 ~ 16] Improve iopf framework.

This is also available at github [2].

[1]
https://lore.kernel.org/linux-iommu/20230530053724.232765-1-baolu.lu@xxxxxxxxxxxxxxx/ <https://lore.kernel.org/linux-iommu/20230530053724.232765-1-baolu.lu@xxxxxxxxxxxxxxx/>
[2]
https://github.com/LuBaolu/intel-iommu/commits/preparatory-io-pgfault-delivery-v12 <https://github.com/LuBaolu/intel-iommu/commits/preparatory-io-pgfault-delivery-v12>


Wandering are these patches dropped now,

[PATCH v2 2/6] iommufd: Add iommu page fault uapi data
https://lore.kernel.org/lkml/20231026024930.382898-3-baolu.lu@xxxxxxxxxxxxxxx/raw <https://lore.kernel.org/lkml/20231026024930.382898-3-baolu.lu@xxxxxxxxxxxxxxx/raw>

[PATCH v2 4/6] iommufd: Deliver fault messages to user space
https://lore.kernel.org/lkml/20231026024930.382898-5-baolu.lu@xxxxxxxxxxxxxxx/ <https://lore.kernel.org/lkml/20231026024930.382898-5-baolu.lu@xxxxxxxxxxxxxxx/>

Above patches are part of another series named "IOMMUFD: Deliver IO page
faults to user space", which is now updated to v3.

https://lore.kernel.org/linux-iommu/20240122073903.24406-1-baolu.lu@xxxxxxxxxxxxxxx/

And does iouring still be used in user space?

iouring is not related to this series. For uapi of iommufd, it's still
recommended to use iouring to speed up the handling of faults.

Best regards,
baolu