Re: [PATCH 3/3] vfio/type1: batch page pinning

2021-02-18 Thread Daniel Jordan
Alex Williamson writes: > This might not be the first batch we fill, I think this needs to unwind > rather than direct return. So it does, well spotted. And it's the same thing with the ENODEV case farther up. > Series looks good otherwise. Thanks for going through it!

Re: [PATCH 3/3] vfio/type1: batch page pinning

2021-02-18 Thread Alex Williamson
On Wed, 3 Feb 2021 15:47:56 -0500 Daniel Jordan wrote: > Pinning one 4K page at a time is inefficient, so do it in batches of 512 > instead. This is just an optimization with no functional change > intended, and in particular the driver still calls iommu_map() with the > largest physically cont