Re: [PATCH v19 12/18] vfio/pci: Add dmabuf TPH metadata storage and fd query helper

[Alex Williamson <alex@shazbot.org>]

On Tue, 14 Jul 2026 18:46:51 +0800
fengchengwen <fengchengwen@huawei.com> wrote:
> On 7/10/2026 2:50 AM, Alex Williamson wrote:
> > 
> > Additionally, dma_buf_get_pci_tph() defines that the provided values
> > are only valid until invalidate_mappings, but this series has no
> > plumbing to honor that requirement.
> > 
> > This needs significant work if we need to track the lifecycle of every
> > dmabuf related TPH value.  Thanks,  
> 
> Is it possible to use this as a risk warning? After all, even if the
> invalid mapping is performed, the ST only causes performance loss to
> the system, and there will be no function problems.

Seems like the sort of thing that should have been resolved BEFORE
sending out v20.  By the same argument, maybe it's sufficient to zero
invalidated STs automatically, but we can only do that for IV mode
in-kernel and rely on userspace to perform the same for DS mode, but
regardless, no mechanism is proposed to perform such invalidation.

IMO, this brings into question the supportability of STs provided
through dmabufs.  It would seem vfio-pci needs to become an importer of
the dmabuf, rather than simply scraping TPH data from it, but maybe it
can only be supported in IV mode.  We don't have a means to represent
that.

Alex