Re: [RFC PATCH 0/6] iommufd: harden DMA-BUF PAL setup and release

[Karl Mehltretter]

On Tue, Oct 20, 2026 at 03:20:29PM +0100, Jason Gunthorpe wrote:
> On Tue, Oct 20, 2026 at 10:16:26PM +1100, Alistair Popple wrote:
> 
> > I haven't had a chance to look at these patches yet, but if some of them are
> > fixing bugs in Jason's series it might be best for us to fold them into that
> > series, with correct attribution of course. Anyway I will take a closer look in
> > the morning.
> 
> Are any fixing bugs in tree right now? I think one of them is already
> applied from someone else..
> 

Good catch. None of these fixes a bug in the current mainline tree. I
based the series on the stale c37cf2824740d topic and missed two fixes
that landed in mainline afterward:

- patch 4 is already covered more broadly by f2d70dbd3dce
  ("iommufd: Destroy the pages content after detaching from dmabuf");
- patch 5 duplicates Ankit Soni's e745cd2c749e
  ("iommufd: Take dma_resv lock before dma_buf_unpin() in release path").

Patches 1-3 address lifecycle issues introduced by the PAL conversion,
and patch 6 tests those paths.

My main goal with this RFC was to put the concrete lifecycle issues and
tests on the table and get the discussion moving.

I am not concerned about preserving these as separate commits for
attribution purposes. Please fold or rework any useful pieces into the
mapping-type series as makes sense.

Thanks,
Karl