On 30/01/2025 14:08, Florent Tomasin wrote:
> Allow mali-valhall-csf driver to retrieve a protected
> heap at probe time by passing the name of the heap
> as attribute to the device tree GPU node.
Please wrap commit message according to Linux coding style / submission
process (neither too early nor over the limit):
https://elixir.bootlin.com/linux/v6.4-rc1/source/Documentation/process/subm…
Why this cannot be passed by phandle, just like all reserved regions?
From where do you take these protected heaps? Firmware? This would
explain why no relation is here (no probe ordering, no device links,
nothing connecting separate devices).
Best regards,
Krzysztof
Even the kerneldoc says that with a zero timeout the function should not
wait for anything, but still return 1 to indicate that the fences are
signaled now.
Unfortunately that isn't what was implemented, instead of only returning
1 we also waited for at least one jiffies.
Fix that by adjusting the handling to what the function is actually
documented to do.
v2: improve code readability
Reported-by: Marek Olšák <marek.olsak(a)amd.com>
Reported-by: Lucas Stach <l.stach(a)pengutronix.de>
Signed-off-by: Christian König <christian.koenig(a)amd.com>
Cc: <stable(a)vger.kernel.org>
---
drivers/dma-buf/dma-resv.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c
index 5f8d010516f0..c78cdae3deaf 100644
--- a/drivers/dma-buf/dma-resv.c
+++ b/drivers/dma-buf/dma-resv.c
@@ -684,11 +684,13 @@ long dma_resv_wait_timeout(struct dma_resv *obj, enum dma_resv_usage usage,
dma_resv_iter_begin(&cursor, obj, usage);
dma_resv_for_each_fence_unlocked(&cursor, fence) {
- ret = dma_fence_wait_timeout(fence, intr, ret);
- if (ret <= 0) {
- dma_resv_iter_end(&cursor);
- return ret;
- }
+ ret = dma_fence_wait_timeout(fence, intr, timeout);
+ if (ret <= 0)
+ break;
+
+ /* Even for zero timeout the return value is 1 */
+ if (timeout)
+ timeout = ret;
}
dma_resv_iter_end(&cursor);
--
2.34.1
Even the kerneldoc says that with a zero timeout the function should not
wait for anything, but still return 1 to indicate that the fences are
signaled now.
Unfortunately that isn't what was implemented, instead of only returning
1 we also waited for at least one jiffies.
Fix that by adjusting the handling to what the function is actually
documented to do.
Reported-by: Marek Olšák <marek.olsak(a)amd.com>
Reported-by: Lucas Stach <l.stach(a)pengutronix.de>
Signed-off-by: Christian König <christian.koenig(a)amd.com>
Cc: <stable(a)vger.kernel.org>
---
drivers/dma-buf/dma-resv.c | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c
index 5f8d010516f0..ae92d9d2394d 100644
--- a/drivers/dma-buf/dma-resv.c
+++ b/drivers/dma-buf/dma-resv.c
@@ -684,11 +684,12 @@ long dma_resv_wait_timeout(struct dma_resv *obj, enum dma_resv_usage usage,
dma_resv_iter_begin(&cursor, obj, usage);
dma_resv_for_each_fence_unlocked(&cursor, fence) {
- ret = dma_fence_wait_timeout(fence, intr, ret);
- if (ret <= 0) {
- dma_resv_iter_end(&cursor);
- return ret;
- }
+ ret = dma_fence_wait_timeout(fence, intr, timeout);
+ if (ret <= 0)
+ break;
+
+ /* Even for zero timeout the return value is 1 */
+ timeout = min(timeout, ret);
}
dma_resv_iter_end(&cursor);
--
2.34.1
Am 22.01.25 um 12:04 schrieb Simona Vetter:
> On Tue, Jan 21, 2025 at 01:36:33PM -0400, Jason Gunthorpe wrote:
>> On Tue, Jan 21, 2025 at 05:11:32PM +0100, Simona Vetter wrote:
>>> On Mon, Jan 20, 2025 at 03:48:04PM -0400, Jason Gunthorpe wrote:
>>>> On Mon, Jan 20, 2025 at 07:50:23PM +0100, Simona Vetter wrote:
>>>>> On Mon, Jan 20, 2025 at 01:59:01PM -0400, Jason Gunthorpe wrote:
>>>>>> On Mon, Jan 20, 2025 at 01:14:12PM +0100, Christian König wrote:
>>>>>> What is going wrong with your email? You replied to Simona, but
>>>>>> Simona Vetter <simona.vetter(a)ffwll.ch> is dropped from the To/CC
>>>>>> list??? I added the address back, but seems like a weird thing to
>>>>>> happen.
>>>>> Might also be funny mailing list stuff, depending how you get these. I
>>>>> read mails over lore and pretty much ignore cc (unless it's not also on
>>>>> any list, since those tend to be security issues) because I get cc'ed on
>>>>> way too much stuff for that to be a useful signal.
>>>> Oh I see, you are sending a Mail-followup-to header that excludes your
>>>> address, so you don't get any emails at all.. My mutt is dropping you
>>>> as well.
I'm having all kind of funny phenomena with AMDs mail servers since
coming back from xmas vacation.
From the news it looks like Outlook on Windows has a new major security
issue where just viewing a mail can compromise the system and my
educated guess is that our IT guys went into panic mode because of this
and has changed something.
>> [SNIP]
>> I have been assuming that dmabuf mmap remains unchanged, that
>> exporters will continue to implement that mmap() callback as today.
That sounds really really good to me because that was my major concern
when you noted that you want to have PFNs to build up KVM page tables.
But you don't want to handle mmap() on your own, you basically don't
want to have a VMA for this stuff at all, correct?
>> My main interest has been what data structure is produced in the
>> attach APIs.
>>
>> Eg today we have a struct dma_buf_attachment that returns a sg_table.
>>
>> I'm expecting some kind of new data structure, lets call it "physical
>> list" that is some efficient coding of meta/addr/len tuples that works
>> well with the new DMA API. Matthew has been calling this thing phyr..
I would not use a data structure at all. Instead we should have
something like an iterator/cursor based approach similar to what the new
DMA API is doing.
>> So, I imagine, struct dma_buf_attachment gaining an optional
>> feature negotiation and then we have in dma_buf_attachment:
>>
>> union {
>> struct sg_table *sgt;
>> struct physical_list *phyr;
>> };
>>
>> That's basicaly it, an alternative to scatterlist that has a clean
>> architecture.
I would rather suggest something like dma_buf_attachment() gets offset
and size to map and returns a cursor object you can use to get your
address, length and access attributes.
And then you can iterate over this cursor and fill in your importer data
structure with the necessary information.
This way neither the exporter nor the importer need to convert their
data back and forth between their specific representations of the
information.
>> Now, if you are asking if the current dmabuf mmap callback can be
>> improved with the above? Maybe? phyr should have the neccessary
>> information inside it to populate a VMA - eventually even fully
>> correctly with all the right cachable/encrypted/forbidden/etc flags.
That won't work like this.
See the exporter needs to be informed about page faults on the VMA to
eventually wait for operations to end and sync caches.
Otherwise we either potentially allow access to freed up or re-used
memory or run into issues with device cache coherency.
>> So, you could imagine that exporters could just have one routine to
>> generate the phyr list and that goes into the attachment, goes into
>> some common code to fill VMA PTEs, and some other common code that
>> will convert it into the DMABUF scatterlist. If performance is not a
>> concern with these data structure conversions it could be an appealing
>> simplification.
>>
>> And yes, I could imagine the meta information being descriptive enough
>> to support the private interconnect cases, the common code could
>> detect private meta information and just cleanly fail.
> I'm kinda leaning towards entirely separate dma-buf interfaces for the new
> phyr stuff, because I fear that adding that to the existing ones will only
> make the chaos worse. But that aside sounds all reasonable, and even that
> could just be too much worry on my side and mixing phyr into existing
> attachments (with a pile of importer/exporter flags probably) is fine.
I lean into the other direction.
Dmitry and Thomas have done a really good job at cleaning up all the
interaction between dynamic and static exporters / importers.
Especially that we now have consistent locking for map_dma_buf() and
unmap_dma_buf() should make that transition rather straight forward.
> For the existing dma-buf importers/exporters I'm kinda hoping for a pure
> dma_addr_t based list eventually. Going all the way to a phyr based
> approach for everyone might be too much churn, there's some real bad cruft
> there. It's not going to work for every case, but it covers a lot of them
> and might be less pain for existing importers.
The point is we have use cases that won't work without exchanging DMA
addresses any more.
For example we have cases with multiple devices are in the same IOMMU
domain and re-using their DMA address mappings.
> But in theory it should be possible to use phyr everywhere eventually, as
> long as there's no obviously api-rules-breaking way to go from a phyr back to
> a struct page even when that exists.
I would rather say we should stick to DMA addresses as much as possible.
What we can do is to add an address space description to the addresses,
e.g. if it's a PCIe BUS addr in IOMMU domain X, or of it's a device
private bus addr or in the case of sharing with iommufd and KVM PFNs.
Regards,
Christian.
>>> At least the device mapping / dma_buf_attachment
>>> side should be doable with just the pfn and the new dma-api?
>> Yes, that would be my first goal post. Figure out some meta
>> information and a container data structure that allows struct
>> page-less P2P mapping through the new DMA API.
>>
>>>> I'm hoping we can get to something where we describe not just how the
>>>> pfns should be DMA mapped, but also can describe how they should be
>>>> CPU mapped. For instance that this PFN space is always mapped
>>>> uncachable, in CPU and in IOMMU.
>>> I was pondering whether dma_mmap and friends would be a good place to
>>> prototype this and go for a fully generic implementation. But then even
>>> those have _wc/_uncached variants.
>> Given that the inability to correctly DMA map P2P MMIO without struct
>> page is a current pain point and current source of hacks in dmabuf
>> exporters, I wanted to make resolving that a priority.
>>
>> However, if you mean what I described above for "fully generic [dmabuf
>> mmap] implementation", then we'd have the phyr datastructure as a
>> dependency to attempt that work.
>>
>> phyr, and particularly the meta information, has a number of
>> stakeholders. I was thinking of going first with rdma's memory
>> registration flow because we are now pretty close to being able to do
>> such a big change, and it can demonstrate most of the requirements.
>>
>> But that doesn't mean mmap couldn't go concurrently on the same agreed
>> datastructure if people are interested.
> Yeah cpu mmap needs a lot more, going with a very limited p2p use-case
> first only makes sense.
>
>>>> We also have current bugs in the iommu/vfio side where we are fudging
>>>> CC stuff, like assuming CPU memory is encrypted (not always true) and
>>>> that MMIO is non-encrypted (not always true)
>>> tbf CC pte flags I just don't grok at all. I've once tried to understand
>>> what current exporters and gpu drivers do and just gave up. But that's
>>> also a bit why I'm worried here because it's an enigma to me.
>> For CC, inside the secure world, is some information if each PFN
>> inside the VM is 'encrypted' or not. Any VM PTE (including the IOPTEs)
>> pointing at the PFN must match the secure world's view of
>> 'encrypted'. The VM can ask the secure world to change its view at
>> runtime.
>>
>> The way CC has been bolted on to the kernel so far laregly hides this
>> from drivers, so it is troubled to tell in driver code if the PFN you
>> have is 'encrypted' or not. Right now the general rule (that is not
>> always true) is that struct page CPU memory is encrypted and
>> everything else is decrypted.
>>
>> So right now, you can mostly ignore it and the above assumption
>> largely happens for you transparently.
>>
>> However, soon we will have encrypted P2P MMIO which will stress this
>> hiding strategy.
> It's already breaking with stuff like virtual gpu drivers, vmwgfx is
> fiddling around with these bits (at least last I tried to understand this
> all) and I think a few others do too.
>
>>>>> I thought iommuv2 (or whatever linux calls these) has full fault support
>>>>> and could support current move semantics. But yeah for iommu without
>>>>> fault support we need some kind of pin or a newly formalized revoke model.
>>>> No, this is HW dependent, including PCI device, and I'm aware of no HW
>>>> that fully implements this in a way that could be useful to implement
>>>> arbitary move semantics for VFIO..
>>> Hm I thought we've had at least prototypes floating around of device fault
>>> repair, but I guess that only works with ATS/pasid stuff and not general
>>> iommu traffic from devices. Definitely needs some device cooperation since
>>> the timeouts of a full fault are almost endless.
>> Yes, exactly. What all real devices I'm aware have done is make a
>> subset of their traffic work with ATS and PRI, but not all their
>> traffic. Without *all* traffic you can't make any generic assumption
>> in the iommu that a transient non-present won't be fatal to the
>> device.
>>
>> Stuff like dmabuf move semantics rely on transient non-present being
>> non-disruptive...
> Ah now I get it, at the iommu level you have to pessimistically assume
> whether a device can handle a fault, and none can for all traffic. I was
> thinking too much about the driver level where generally the dma-buf you
> importer are only used for the subset of device functions that can cope
> with faults on many devices.
>
> Cheers, Sima
On Thu, Jan 23, 2025 at 03:41:58PM +0800, Xu Yilun wrote:
> I don't have a complete idea yet. But the goal is not to make any
> existing driver seamlessly work with secure device. It is to provide a
> generic way for bind/attestation/accept, and may save driver's effort
> if they don't care about this startup process. There are plenty of
> operations that a driver can't do to a secure device, FLR is one of
> them. The TDISP SPEC has described some general rules but some are even
> device specific.
You can FLR a secure device, it just has to be re-secured and
re-attested after. Otherwise no VFIO for you.
> So I think a driver (including VFIO) expects change to support trusted
> device, but may not have to cover bind/attestation/accept flow.
I expect changes, but not fundamental ones. VFIO will still have to
FLR devices as part of it's security architecture.
The entire flow needs to have options for drivers to be involved in
the flow, somehow.
Jason
On Wed, Jan 22, 2025 at 12:04:19PM +0100, Simona Vetter wrote:
> I'm kinda leaning towards entirely separate dma-buf interfaces for the new
> phyr stuff, because I fear that adding that to the existing ones will only
> make the chaos worse.
Lets see when some patches come up, if importers have to deal with
several formats a single attach interface would probably be simpler
for them..
> For the existing dma-buf importers/exporters I'm kinda hoping for a pure
> dma_addr_t based list eventually.
IMHO the least churn would be to have the old importers call a helper
(perhaps implicitly in the core code) to convert phyr into a dma
mapped scatterlist and then just keep using their existing code
exactly as is.
Performance improvement would come from importers switching to use the
new dma api internally and avoiding the scatterlist allocation, but
even the extra alocation is not so bad.
Then, perhaps, and I really hesitate to say this, but perhaps to ease
the migration we could store a dma mapped list in a phyr using the
meta information. That would allow a dmabuf scatterlist exporter to be
converted to a phyr with a helper. The importer would have to have a
special path to detect the dma mapped mode and skip the normal
mapping.
The point would be to let maintained drivers use the new data
structures and flows easily and still interoperate with older
drivers. Otherwise we have to somehow build parallel scatterlist and
phyr code flows into importers :\
> Going all the way to a phyr based
> approach for everyone might be too much churn, there's some real bad cruft
> there. It's not going to work for every case, but it covers a lot of them
> and might be less pain for existing importers.
I'd hope we can reach every case.. But I don't know what kind of
horrors you guys have :)
> But in theory it should be possible to use phyr everywhere eventually, as
> long as there's no obviously api-rules-breaking way to go from a phyr back to
> a struct page even when that exists.
I'd say getting a struct page is a perfectly safe operation, from a
phyr API perspective.
The dmabuf issue seems to be entirely about following the locking
rules, an importer is not allowed to get a struct page and switch from
reservation locking to page refcount locking.
I get the appeal for DRM of blocking struct page use because that
directly prevents the above. But, in RDMA I want to re-use the same
phyr datastructure for tracking pin_user_pages() CPU memory, and I
must get the struct page back so I can put_user_page() it when done.
Perhaps we can find some compromise where the phyr data structure has
some kind of flag 'disable struct page' and then the
phyr_entry_to_page() API will WARN_ON() or something like that.
> > However, soon we will have encrypted P2P MMIO which will stress this
> > hiding strategy.
>
> It's already breaking with stuff like virtual gpu drivers, vmwgfx is
> fiddling around with these bits (at least last I tried to understand this
> all) and I think a few others do too.
IMHO, most places I've seen touching this out side arch code feel very
hacky. :\
Jason
On Wed, Jan 22, 2025 at 12:32:56PM +0800, Xu Yilun wrote:
> On Tue, Jan 21, 2025 at 01:43:03PM -0400, Jason Gunthorpe wrote:
> > On Tue, Jun 25, 2024 at 05:12:10AM +0800, Xu Yilun wrote:
> >
> > > When VFIO works as a TEE user in VM, it means an attester (e.g. PCI
> > > subsystem) has already moved the device to RUN state. So VFIO & DPDK
> > > are all TEE users, no need to manipulate TDISP state between them.
> > > AFAICS, this is the most preferred TIO usage in CoCo-VM.
> >
> > No, unfortunately. Part of the motivation to have the devices be
> > unlocked when the VM starts is because there is an expectation that a
> > driver in the VM will need to do untrusted operations to boot up the
>
> I assume these operations are device specific.
Yes
> > device before it can be switched to the run state.
> >
> > So any vfio use case needs to imagine that VFIO starts with an
> > untrusted device, does stuff to it, then pushes everything through to
>
> I have concern that VFIO has to do device specific stuff. Our current
> expectation is a specific device driver deals with the untrusted
> operations, then user writes a 'bind' device sysfs node which detaches
> the driver for untrusted, do the attestation and accept, and try match
> the driver for trusted (e.g. VFIO).
I don't see this as working, VFIO will FLR the device which will
destroy anything that was done prior.
VFIO itself has to do the sequence and the VFIO userspace has to
contain the device specific stuff.
The bind/unbind dance for untrusted->trusted would need to be
internalized in VFIO without unbinding. The main motivation for the
bind/unbind flow was to manage the DMA API, which VFIO does not use.
Jason
On Tue, Jun 25, 2024 at 05:12:10AM +0800, Xu Yilun wrote:
> When VFIO works as a TEE user in VM, it means an attester (e.g. PCI
> subsystem) has already moved the device to RUN state. So VFIO & DPDK
> are all TEE users, no need to manipulate TDISP state between them.
> AFAICS, this is the most preferred TIO usage in CoCo-VM.
No, unfortunately. Part of the motivation to have the devices be
unlocked when the VM starts is because there is an expectation that a
driver in the VM will need to do untrusted operations to boot up the
device before it can be switched to the run state.
So any vfio use case needs to imagine that VFIO starts with an
untrusted device, does stuff to it, then pushes everything through to
run. The exact mirror as what a kernel driver should be able to do.
How exactly all this very complex stuff works, I have no idea, but
this is what I've understood is the target. :\
Jason
On Tue, Jan 21, 2025 at 05:11:32PM +0100, Simona Vetter wrote:
> On Mon, Jan 20, 2025 at 03:48:04PM -0400, Jason Gunthorpe wrote:
> > On Mon, Jan 20, 2025 at 07:50:23PM +0100, Simona Vetter wrote:
> > > On Mon, Jan 20, 2025 at 01:59:01PM -0400, Jason Gunthorpe wrote:
> > > > On Mon, Jan 20, 2025 at 01:14:12PM +0100, Christian König wrote:
> > > > What is going wrong with your email? You replied to Simona, but
> > > > Simona Vetter <simona.vetter(a)ffwll.ch> is dropped from the To/CC
> > > > list??? I added the address back, but seems like a weird thing to
> > > > happen.
> > >
> > > Might also be funny mailing list stuff, depending how you get these. I
> > > read mails over lore and pretty much ignore cc (unless it's not also on
> > > any list, since those tend to be security issues) because I get cc'ed on
> > > way too much stuff for that to be a useful signal.
> >
> > Oh I see, you are sending a Mail-followup-to header that excludes your
> > address, so you don't get any emails at all.. My mutt is dropping you
> > as well.
> >
> > > Yeah I'm not worried about cpu mmap locking semantics. drm/ttm is a pretty
> > > clear example that you can implement dma-buf mmap with the rules we have,
> > > except the unmap_mapping_range might need a bit fudging with a separate
> > > address_space.
> >
> > From my perspective the mmap thing is a bit of a side/DRM-only thing
> > as nothing I'm interested in wants to mmap dmabuf into a VMA.
>
> I guess we could just skip mmap on these pfn exporters, but it also means
> a bit more boilerplate.
I have been assuming that dmabuf mmap remains unchanged, that
exporters will continue to implement that mmap() callback as today.
My main interest has been what data structure is produced in the
attach APIs.
Eg today we have a struct dma_buf_attachment that returns a sg_table.
I'm expecting some kind of new data structure, lets call it "physical
list" that is some efficient coding of meta/addr/len tuples that works
well with the new DMA API. Matthew has been calling this thing phyr..
So, I imagine, struct dma_buf_attachment gaining an optional
feature negotiation and then we have in dma_buf_attachment:
union {
struct sg_table *sgt;
struct physical_list *phyr;
};
That's basicaly it, an alternative to scatterlist that has a clean
architecture.
Now, if you are asking if the current dmabuf mmap callback can be
improved with the above? Maybe? phyr should have the neccessary
information inside it to populate a VMA - eventually even fully
correctly with all the right cachable/encrypted/forbidden/etc flags.
So, you could imagine that exporters could just have one routine to
generate the phyr list and that goes into the attachment, goes into
some common code to fill VMA PTEs, and some other common code that
will convert it into the DMABUF scatterlist. If performance is not a
concern with these data structure conversions it could be an appealing
simplification.
And yes, I could imagine the meta information being descriptive enough
to support the private interconnect cases, the common code could
detect private meta information and just cleanly fail.
> At least the device mapping / dma_buf_attachment
> side should be doable with just the pfn and the new dma-api?
Yes, that would be my first goal post. Figure out some meta
information and a container data structure that allows struct
page-less P2P mapping through the new DMA API.
> > I'm hoping we can get to something where we describe not just how the
> > pfns should be DMA mapped, but also can describe how they should be
> > CPU mapped. For instance that this PFN space is always mapped
> > uncachable, in CPU and in IOMMU.
>
> I was pondering whether dma_mmap and friends would be a good place to
> prototype this and go for a fully generic implementation. But then even
> those have _wc/_uncached variants.
Given that the inability to correctly DMA map P2P MMIO without struct
page is a current pain point and current source of hacks in dmabuf
exporters, I wanted to make resolving that a priority.
However, if you mean what I described above for "fully generic [dmabuf
mmap] implementation", then we'd have the phyr datastructure as a
dependency to attempt that work.
phyr, and particularly the meta information, has a number of
stakeholders. I was thinking of going first with rdma's memory
registration flow because we are now pretty close to being able to do
such a big change, and it can demonstrate most of the requirements.
But that doesn't mean mmap couldn't go concurrently on the same agreed
datastructure if people are interested.
> > We also have current bugs in the iommu/vfio side where we are fudging
> > CC stuff, like assuming CPU memory is encrypted (not always true) and
> > that MMIO is non-encrypted (not always true)
>
> tbf CC pte flags I just don't grok at all. I've once tried to understand
> what current exporters and gpu drivers do and just gave up. But that's
> also a bit why I'm worried here because it's an enigma to me.
For CC, inside the secure world, is some information if each PFN
inside the VM is 'encrypted' or not. Any VM PTE (including the IOPTEs)
pointing at the PFN must match the secure world's view of
'encrypted'. The VM can ask the secure world to change its view at
runtime.
The way CC has been bolted on to the kernel so far laregly hides this
from drivers, so it is troubled to tell in driver code if the PFN you
have is 'encrypted' or not. Right now the general rule (that is not
always true) is that struct page CPU memory is encrypted and
everything else is decrypted.
So right now, you can mostly ignore it and the above assumption
largely happens for you transparently.
However, soon we will have encrypted P2P MMIO which will stress this
hiding strategy.
> > > I thought iommuv2 (or whatever linux calls these) has full fault support
> > > and could support current move semantics. But yeah for iommu without
> > > fault support we need some kind of pin or a newly formalized revoke model.
> >
> > No, this is HW dependent, including PCI device, and I'm aware of no HW
> > that fully implements this in a way that could be useful to implement
> > arbitary move semantics for VFIO..
>
> Hm I thought we've had at least prototypes floating around of device fault
> repair, but I guess that only works with ATS/pasid stuff and not general
> iommu traffic from devices. Definitely needs some device cooperation since
> the timeouts of a full fault are almost endless.
Yes, exactly. What all real devices I'm aware have done is make a
subset of their traffic work with ATS and PRI, but not all their
traffic. Without *all* traffic you can't make any generic assumption
in the iommu that a transient non-present won't be fatal to the
device.
Stuff like dmabuf move semantics rely on transient non-present being
non-disruptive...
Jason