Re: [PATCH v3 3/5] iommu: Add verisilicon IOMMU driver

On Fri, Jun 20, 2025 at 03:52:46PM +0200, Benjamin Gaignard wrote:

quoted

quoted

vsi_iommu_suspend() and vsi_iommu_resume() will also have to take the lock
before calling vsi_iommu_disable() and vsi_iommu_enable().

Yes, if they use iommu->domain that seems good

If the above locking is a problem then I'd use the group mutex instead
during resume/suspend. The attach functions are already called with
the group mutex held.

Does group mutex is also called when using vsi_iommu_map or vsi_iommu_unmap ?

No

Jason