On Tue, 2020-07-21 at 19:52 -0500, Brian King wrote: > > > > As of today, there seems to be nothing like that happening in the > > driver I am testing. > > I spoke to Brian King on slack, and he mentioned that at the point DDW > > is created there should be no allocations in place. > > I think there are a couple of scenarios here. One is where there is a DMA > allocation prior to a call to set the DMA mask. Second scenario is if the > driver makes multiple calls to set the DMA mask. I would argue that a properly > written driver should tell the IOMMU subsystem what DMA mask it supports prior > to allocating DMA memroy. Documentation/core-api/dma-api-howto.rst should > describe what is legal and what is not. > > It might be reasonable to declare its not allowed to allocate DMA memory > and then later change the DMA mask and clearly call this out in the > documentation > if its not already. > > -Brian
Thank you for the feedback Brian! That makes sense to me. I will try to have this in mind for the next patchset. Best regards,