Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Demi Marie Obenour]

On 1/24/25 7:42 PM, Demi Marie Obenour wrote:
> On 1/8/25 12:05 PM, Simona Vetter wrote:
>> On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
>>>
>>> On 2024/12/22 9:59, Demi Marie Obenour wrote:
>>>> On 12/20/24 10:35 AM, Simona Vetter wrote:
>>>>> On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
>>>>>> From: Honglei Huang <Honglei1.Huang@amd.com>
>>>>>>
>>>>>> A virtio-gpu userptr is based on HMM notifier.
>>>>>> Used for let host access guest userspace memory and
>>>>>> notice the change of userspace memory.
>>>>>> This series patches are in very beginning state,
>>>>>> User space are pinned currently to ensure the host
>>>>>> device memory operations are correct.
>>>>>> The free and unmap operations for userspace can be
>>>>>> handled by MMU notifier this is a simple and basice
>>>>>> SVM feature for this series patches.
>>>>>> The physical PFNS update operations is splited into
>>>>>> two OPs in here. The evicted memories won't be used
>>>>>> anymore but remap into host again to achieve same
>>>>>> effect with hmm_rang_fault.
>>>>>
>>>>> So in my opinion there are two ways to implement userptr that make sense:
>>>>>
>>>>> - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
>>>>>    notifier
>>>>>
>>>>> - unpinnned userptr where you entirely rely on userptr and do not hold any
>>>>>    page references or page pins at all, for full SVM integration. This
>>>>>    should use hmm_range_fault ideally, since that's the version that
>>>>>    doesn't ever grab any page reference pins.
>>>>>
>>>>> All the in-between variants are imo really bad hacks, whether they hold a
>>>>> page reference or a temporary page pin (which seems to be what you're
>>>>> doing here). In much older kernels there was some justification for them,
>>>>> because strange stuff happened over fork(), but with FOLL_LONGTERM this is
>>>>> now all sorted out. So there's really only fully pinned, or true svm left
>>>>> as clean design choices imo.
>>>>>
>>>>> With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
>>>>> you?
>>>>
>>>> +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
>>>> in complexity that pinning everything avoids.  Furthermore, this avoids the
>>>> host having to take action in response to guest memory reclaim requests.
>>>> This avoids additional complexity (and thus attack surface) on the host side.
>>>> Furthermore, since this is for ROCm and not for graphics, I am less concerned
>>>> about supporting systems that require swappable GPU VRAM.
>>>
>>> Hi Sima and Demi,
>>>
>>> I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
>>> version.
>>>
>>> And for the first pin variants implementation, the MMU notifier is also
>>> needed I think.Cause the userptr feature in UMD generally used like this:
>>> the registering of userptr always is explicitly invoked by user code like
>>> "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
>>> there is no explicit API for it, at least in hsakmt/KFD stack. User just
>>> need call system call "free(userptrAddr)", then kernel driver will release
>>> the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
>>> user has been free the userptr except for MMU notifior.And in UMD theres is
>>> no way to get the free() operation is invoked by user.The only way is use
>>> MMU notifier in virtio-GPU driver and free the corresponding data in host by
>>> some virtio CMDs as far as I can see.
>>>
>>> And for the second way that is use hmm_range_fault, there is a predictable
>>> issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
>>> may migrate when GPU/device is working. In bare metal, when memory is
>>> migrating KFD driver will pause the compute work of the device in
>>> mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
>>> memories to GPU then restore the compute work of device to ensure the
>>> correction of the data. But in virtio-GPU driver the migration happen in
>>> guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
>>> can be used for notify host but as lack of mmap_write_lock protection in
>>> host kernel, host will hold invalid data for a short period of time, this
>>> may lead to some issues. And it is hard to fix as far as I can see.
>>>
>>> I will extract some APIs into helper according to your request, and I will
>>> refactor the whole userptr implementation, use some callbacks in page
>>> getting path, let the pin method and hmm_range_fault can be choiced
>>> in this series patches.
>>
>> Ok, so if this is for svm, then you need full blast hmm, or the semantics
>> are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
>> not work.
> 
> Is this still broken in the virtualized case?  Page migration between host
> and device memory is completely transparent to the guest kernel, so pinning
> guest memory doesn't interfere with the host KMD at all.  In fact, the host
> KMD is not even aware of it.

To elaborate further:

Memory in a KVM guest is *not* host physical memory, or even host kernel
memory.  It is host *userspace* memory, and in particular, *it is fully pageable*.
There *might* be a few exceptions involving structures that are accessed by
the (physical) CPU, but none of these are relevant here.

This means that memory management works very differently than in the
non-virtualized case.  The host KMD can migrate pages between host memory
and device memory without either the guest kernel or host userspace being
aware that such migration has happened.  This means that pin(FOLL_LONGTERM)
in the guest doesn't pin memory on the host.  Instead, it pins memory in the
*guest*.  The host will continue to migrate pages between host and device
as needed.  I’m no expert on SVM, but I suspect this is the desired behavior.

Xen is significantly trickier, because most guest memory is provided by
the Xen toolstack via the hypervisor and is _not_ pageable.  Therefore,
it cannot be mapped into the GPU without using Xen grant tables.  Since
Xen grants do not support non-cooperative revocation, this requires a
FOLL_LONGTERM pin *anyway*.  Furthermore, granted pages _cannot_ be
migrated from host to device, so unless the GPU is an iGPU all of its
accesses will need to cross the PCI bus.  This will obviously be slow.

The guest can avoid this problem by migrating userptr memory to virtio-GPU
blob objects _before_ pinning it.  Virtio-GPU blob objects are backed by
host userspace memory, so the host can migrate them between device and host
memory just like in the KVM case.  Under KVM, such migration would be be
slightly wasteful but otherwise harmless in the common case.  In the case
where PCI passthrough is also in use, however, it might be necessary even
for KVM guests.  This is because PCI passthrough requires pinned memory,
and pinned memory cannot be migrated to the device.

Since AMD’s automotive use-case uses Xen, and since KVM might also need
page migration, I recommend that the initial implementation _always_
migrate pages to blob objects no matter what the hypervisor is.  Direct
GPU access to guest memory can be implemented as a KVM-specific optimization
later.

Also worth noting is that only pages that have been written need to be
migrated.  If a page hasn't been written, it should not be migrated, because
unwritten pages of a blob objects will read as zero.  However, the migration
should almost certainly be done in 2M chunks, rather than 4K ones.  This is
because the TLBs of at least AMD GPU are optimized for 2M pages, and GPU access
to 4K pages takes a 30% performance penalty.  This nicely matches the penalty
that AMD observed.
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Demi Marie Obenour]

On 1/8/25 12:05 PM, Simona Vetter wrote:
> On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
>>
>> On 2024/12/22 9:59, Demi Marie Obenour wrote:
>>> On 12/20/24 10:35 AM, Simona Vetter wrote:
>>>> On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
>>>>> From: Honglei Huang <Honglei1.Huang@amd.com>
>>>>>
>>>>> A virtio-gpu userptr is based on HMM notifier.
>>>>> Used for let host access guest userspace memory and
>>>>> notice the change of userspace memory.
>>>>> This series patches are in very beginning state,
>>>>> User space are pinned currently to ensure the host
>>>>> device memory operations are correct.
>>>>> The free and unmap operations for userspace can be
>>>>> handled by MMU notifier this is a simple and basice
>>>>> SVM feature for this series patches.
>>>>> The physical PFNS update operations is splited into
>>>>> two OPs in here. The evicted memories won't be used
>>>>> anymore but remap into host again to achieve same
>>>>> effect with hmm_rang_fault.
>>>>
>>>> So in my opinion there are two ways to implement userptr that make sense:
>>>>
>>>> - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
>>>>    notifier
>>>>
>>>> - unpinnned userptr where you entirely rely on userptr and do not hold any
>>>>    page references or page pins at all, for full SVM integration. This
>>>>    should use hmm_range_fault ideally, since that's the version that
>>>>    doesn't ever grab any page reference pins.
>>>>
>>>> All the in-between variants are imo really bad hacks, whether they hold a
>>>> page reference or a temporary page pin (which seems to be what you're
>>>> doing here). In much older kernels there was some justification for them,
>>>> because strange stuff happened over fork(), but with FOLL_LONGTERM this is
>>>> now all sorted out. So there's really only fully pinned, or true svm left
>>>> as clean design choices imo.
>>>>
>>>> With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
>>>> you?
>>>
>>> +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
>>> in complexity that pinning everything avoids.  Furthermore, this avoids the
>>> host having to take action in response to guest memory reclaim requests.
>>> This avoids additional complexity (and thus attack surface) on the host side.
>>> Furthermore, since this is for ROCm and not for graphics, I am less concerned
>>> about supporting systems that require swappable GPU VRAM.
>>
>> Hi Sima and Demi,
>>
>> I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
>> version.
>>
>> And for the first pin variants implementation, the MMU notifier is also
>> needed I think.Cause the userptr feature in UMD generally used like this:
>> the registering of userptr always is explicitly invoked by user code like
>> "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
>> there is no explicit API for it, at least in hsakmt/KFD stack. User just
>> need call system call "free(userptrAddr)", then kernel driver will release
>> the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
>> user has been free the userptr except for MMU notifior.And in UMD theres is
>> no way to get the free() operation is invoked by user.The only way is use
>> MMU notifier in virtio-GPU driver and free the corresponding data in host by
>> some virtio CMDs as far as I can see.
>>
>> And for the second way that is use hmm_range_fault, there is a predictable
>> issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
>> may migrate when GPU/device is working. In bare metal, when memory is
>> migrating KFD driver will pause the compute work of the device in
>> mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
>> memories to GPU then restore the compute work of device to ensure the
>> correction of the data. But in virtio-GPU driver the migration happen in
>> guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
>> can be used for notify host but as lack of mmap_write_lock protection in
>> host kernel, host will hold invalid data for a short period of time, this
>> may lead to some issues. And it is hard to fix as far as I can see.
>>
>> I will extract some APIs into helper according to your request, and I will
>> refactor the whole userptr implementation, use some callbacks in page
>> getting path, let the pin method and hmm_range_fault can be choiced
>> in this series patches.
> 
> Ok, so if this is for svm, then you need full blast hmm, or the semantics
> are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
> not work.
> 
> The other option is that hsakmt/kfd api is completely busted, and that's
> kinda not a kernel problem.
> -Sima

On further thought, I believe the driver needs to migrate the pages to
device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
pin on them.  The reason is that it isn’t possible to migrate these pages
back to "host" memory without unmapping them from the GPU.  For the reasons
I mention in [1], I believe that temporarily revoking access to virtio-GPU
blob objects is not feasible.  Instead, the pages must be treated as if
they are permanently in device memory until guest userspace unmaps them
from the GPU, after which they must be migrated back to host memory.

The problems with other approaches are most obvious if one considers a Xen
guest using a virtio-GPU backend that is not all-powerful.  Normal guest
memory is not accessible to the GPU, and Xen uses the IOMMU to enforce this
restriction.  Therefore, the guest must migrate pages to virtio-GPU blob
objects before the GPU can access them.  From Xen’s perspective, virtio-GPU
blob objects belong to the backend domain, so Xen allows the GPU to access
them.  However, the pages in blob objects _cannot_ be used in Xen grant table
operations, because Xen doesn’t consider them to belong to the guest!
Similarly, if the guest has an assigned PCI device, that device will not
be able to access the blob object’s pages.

I’m no expert on Linux memory management, so I’m not sure how to implement
this behavior.  What I _can_ say is that a blob object is I/O memory, and
behaves somewhat similar to a PCI BAR in a system with no P2PDMA support:
CPU access works, but DMA from other devices does not.  Furthermore, the
memory can’t be used for page tables or granted to other Xen guests, and it
will go away if the device is hot-unplugged.  In fact, if the PCI transport
is used, the blob object is located in the BAR of an (emulated) device.
There are non-PCI transports, though, so assuming that blob objects are
located in a PCI BAR is not a good idea.

The reason that pinning the objects in "device" memory is a reasonable
approach is that the host (or backend, in the Xen case) can still migrate
pages between device and host memory and not allocate backing store for
pages that are never accessed.  Therefore, it is not necessary for every
CPU access to go across the PCIe bus even for dGPUs.  Instead, if guest
CPU accesses are much more frequent than device accesses, the memory will
be migrated to the host side.  It’s up to the virtio-GPU backend
implementation to make sure that this happens.  For KVM, this should be
automatic, but for Xen, this might need additional Xen patches so that
the backend domain is notified when pages are accessed or dirtied.

[1]: https://lore.kernel.org/dri-devel/9572ba57-5552-4543-a3b0-6097520a12a3@gmail.com
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Demi Marie Obenour]

[-- Attachment #1: Type: text/plain, Size: 7902 bytes --]

On Wed, Jan 29, 2025 at 03:54:59PM -0500, Demi Marie Obenour wrote:
> On 1/8/25 12:05 PM, Simona Vetter wrote:
> > On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
> >>
> >> On 2024/12/22 9:59, Demi Marie Obenour wrote:
> >>> On 12/20/24 10:35 AM, Simona Vetter wrote:
> >>>> On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
> >>>>> From: Honglei Huang <Honglei1.Huang@amd.com>
> >>>>>
> >>>>> A virtio-gpu userptr is based on HMM notifier.
> >>>>> Used for let host access guest userspace memory and
> >>>>> notice the change of userspace memory.
> >>>>> This series patches are in very beginning state,
> >>>>> User space are pinned currently to ensure the host
> >>>>> device memory operations are correct.
> >>>>> The free and unmap operations for userspace can be
> >>>>> handled by MMU notifier this is a simple and basice
> >>>>> SVM feature for this series patches.
> >>>>> The physical PFNS update operations is splited into
> >>>>> two OPs in here. The evicted memories won't be used
> >>>>> anymore but remap into host again to achieve same
> >>>>> effect with hmm_rang_fault.
> >>>>
> >>>> So in my opinion there are two ways to implement userptr that make sense:
> >>>>
> >>>> - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
> >>>>    notifier
> >>>>
> >>>> - unpinnned userptr where you entirely rely on userptr and do not hold any
> >>>>    page references or page pins at all, for full SVM integration. This
> >>>>    should use hmm_range_fault ideally, since that's the version that
> >>>>    doesn't ever grab any page reference pins.
> >>>>
> >>>> All the in-between variants are imo really bad hacks, whether they hold a
> >>>> page reference or a temporary page pin (which seems to be what you're
> >>>> doing here). In much older kernels there was some justification for them,
> >>>> because strange stuff happened over fork(), but with FOLL_LONGTERM this is
> >>>> now all sorted out. So there's really only fully pinned, or true svm left
> >>>> as clean design choices imo.
> >>>>
> >>>> With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
> >>>> you?
> >>>
> >>> +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
> >>> in complexity that pinning everything avoids.  Furthermore, this avoids the
> >>> host having to take action in response to guest memory reclaim requests.
> >>> This avoids additional complexity (and thus attack surface) on the host side.
> >>> Furthermore, since this is for ROCm and not for graphics, I am less concerned
> >>> about supporting systems that require swappable GPU VRAM.
> >>
> >> Hi Sima and Demi,
> >>
> >> I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
> >> version.
> >>
> >> And for the first pin variants implementation, the MMU notifier is also
> >> needed I think.Cause the userptr feature in UMD generally used like this:
> >> the registering of userptr always is explicitly invoked by user code like
> >> "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
> >> there is no explicit API for it, at least in hsakmt/KFD stack. User just
> >> need call system call "free(userptrAddr)", then kernel driver will release
> >> the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
> >> user has been free the userptr except for MMU notifior.And in UMD theres is
> >> no way to get the free() operation is invoked by user.The only way is use
> >> MMU notifier in virtio-GPU driver and free the corresponding data in host by
> >> some virtio CMDs as far as I can see.
> >>
> >> And for the second way that is use hmm_range_fault, there is a predictable
> >> issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
> >> may migrate when GPU/device is working. In bare metal, when memory is
> >> migrating KFD driver will pause the compute work of the device in
> >> mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
> >> memories to GPU then restore the compute work of device to ensure the
> >> correction of the data. But in virtio-GPU driver the migration happen in
> >> guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
> >> can be used for notify host but as lack of mmap_write_lock protection in
> >> host kernel, host will hold invalid data for a short period of time, this
> >> may lead to some issues. And it is hard to fix as far as I can see.
> >>
> >> I will extract some APIs into helper according to your request, and I will
> >> refactor the whole userptr implementation, use some callbacks in page
> >> getting path, let the pin method and hmm_range_fault can be choiced
> >> in this series patches.
> > 
> > Ok, so if this is for svm, then you need full blast hmm, or the semantics
> > are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
> > not work.
> > 
> > The other option is that hsakmt/kfd api is completely busted, and that's
> > kinda not a kernel problem.
> > -Sima
> 
> On further thought, I believe the driver needs to migrate the pages to
> device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
> pin on them.  The reason is that it isn’t possible to migrate these pages
> back to "host" memory without unmapping them from the GPU.  For the reasons
> I mention in [1], I believe that temporarily revoking access to virtio-GPU
> blob objects is not feasible.  Instead, the pages must be treated as if
> they are permanently in device memory until guest userspace unmaps them
> from the GPU, after which they must be migrated back to host memory.

Discussion on IRC indicates that migration isn't reliable.  This is
because Linux core memory management is largely lock-free for
performance reasons, so there is no way to prevent temporary elevation
of a page's reference count.  A page with an elevated reference count
cannot be migrated.

The only alternative I can think of is for the hypervisor to perform the
migration.  The hypervisor can revoke the guest's access to the page
without the guest's consent or involvement.  The host can then replace
the page with one of its own pages, which might be on the CPU or GPU.
Further migration between the CPU and GPU is controlled by the host
kernel-mode driver (KMD) and host kernel memory management.  The guest
kernel driver must take a FOLL_LONGTERM pin before telling the host to
use the pages, but that is all.

On KVM, this should be essentially automatic, as guest memory really is
just host userspace memory.  On Xen, this requires that the backend
domain can revoke fronted access to _any_ frontend page, or at least
frontend pages that have been granted to the backend.  The backend will
then need to be able to handle page faults for the frontend pages, and
replace the frontend pages with its own pages at will.  Furthermore,
revoking pages that the backend has installed into the frontend must
never fail, because the backend will panic if it does fail.

Sima, is putting guest pages under host kernel control the only option?
I thought that this could be avoided by leaving the pages on the CPU if
migration fails, but that won't work because there will be no way to
migrate them to the GPU later, causing performance problems that would
be impossible to debug.  Is waiting (possibly forever) on migration to
finish an option?  Otherwise, this might mean extra complexity in the
Xen hypervisor, as I do not believe the primitives needed are currently
available.  Specifically, in addition to the primitives discussed at Xen
Project Summit 2024, the backend also needs to intercept access to, and
replace the contents of, arbitrary frontend-controlled pages.
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Huang, Honglei1]

On 2025/1/31 8:33, Demi Marie Obenour wrote:
> On Wed, Jan 29, 2025 at 03:54:59PM -0500, Demi Marie Obenour wrote:
>> On 1/8/25 12:05 PM, Simona Vetter wrote:
>>> On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
>>>>
>>>> On 2024/12/22 9:59, Demi Marie Obenour wrote:
>>>>> On 12/20/24 10:35 AM, Simona Vetter wrote:
>>>>>> On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
>>>>>>> From: Honglei Huang <Honglei1.Huang@amd.com>
>>>>>>>
>>>>>>> A virtio-gpu userptr is based on HMM notifier.
>>>>>>> Used for let host access guest userspace memory and
>>>>>>> notice the change of userspace memory.
>>>>>>> This series patches are in very beginning state,
>>>>>>> User space are pinned currently to ensure the host
>>>>>>> device memory operations are correct.
>>>>>>> The free and unmap operations for userspace can be
>>>>>>> handled by MMU notifier this is a simple and basice
>>>>>>> SVM feature for this series patches.
>>>>>>> The physical PFNS update operations is splited into
>>>>>>> two OPs in here. The evicted memories won't be used
>>>>>>> anymore but remap into host again to achieve same
>>>>>>> effect with hmm_rang_fault.
>>>>>>
>>>>>> So in my opinion there are two ways to implement userptr that make sense:
>>>>>>
>>>>>> - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
>>>>>>     notifier
>>>>>>
>>>>>> - unpinnned userptr where you entirely rely on userptr and do not hold any
>>>>>>     page references or page pins at all, for full SVM integration. This
>>>>>>     should use hmm_range_fault ideally, since that's the version that
>>>>>>     doesn't ever grab any page reference pins.
>>>>>>
>>>>>> All the in-between variants are imo really bad hacks, whether they hold a
>>>>>> page reference or a temporary page pin (which seems to be what you're
>>>>>> doing here). In much older kernels there was some justification for them,
>>>>>> because strange stuff happened over fork(), but with FOLL_LONGTERM this is
>>>>>> now all sorted out. So there's really only fully pinned, or true svm left
>>>>>> as clean design choices imo.
>>>>>>
>>>>>> With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
>>>>>> you?
>>>>>
>>>>> +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
>>>>> in complexity that pinning everything avoids.  Furthermore, this avoids the
>>>>> host having to take action in response to guest memory reclaim requests.
>>>>> This avoids additional complexity (and thus attack surface) on the host side.
>>>>> Furthermore, since this is for ROCm and not for graphics, I am less concerned
>>>>> about supporting systems that require swappable GPU VRAM.
>>>>
>>>> Hi Sima and Demi,
>>>>
>>>> I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
>>>> version.
>>>>
>>>> And for the first pin variants implementation, the MMU notifier is also
>>>> needed I think.Cause the userptr feature in UMD generally used like this:
>>>> the registering of userptr always is explicitly invoked by user code like
>>>> "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
>>>> there is no explicit API for it, at least in hsakmt/KFD stack. User just
>>>> need call system call "free(userptrAddr)", then kernel driver will release
>>>> the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
>>>> user has been free the userptr except for MMU notifior.And in UMD theres is
>>>> no way to get the free() operation is invoked by user.The only way is use
>>>> MMU notifier in virtio-GPU driver and free the corresponding data in host by
>>>> some virtio CMDs as far as I can see.
>>>>
>>>> And for the second way that is use hmm_range_fault, there is a predictable
>>>> issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
>>>> may migrate when GPU/device is working. In bare metal, when memory is
>>>> migrating KFD driver will pause the compute work of the device in
>>>> mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
>>>> memories to GPU then restore the compute work of device to ensure the
>>>> correction of the data. But in virtio-GPU driver the migration happen in
>>>> guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
>>>> can be used for notify host but as lack of mmap_write_lock protection in
>>>> host kernel, host will hold invalid data for a short period of time, this
>>>> may lead to some issues. And it is hard to fix as far as I can see.
>>>>
>>>> I will extract some APIs into helper according to your request, and I will
>>>> refactor the whole userptr implementation, use some callbacks in page
>>>> getting path, let the pin method and hmm_range_fault can be choiced
>>>> in this series patches.
>>>
>>> Ok, so if this is for svm, then you need full blast hmm, or the semantics
>>> are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
>>> not work.
>>>
>>> The other option is that hsakmt/kfd api is completely busted, and that's
>>> kinda not a kernel problem.
>>> -Sima
>>
>> On further thought, I believe the driver needs to migrate the pages to
>> device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
>> pin on them.  The reason is that it isn’t possible to migrate these pages
>> back to "host" memory without unmapping them from the GPU.  For the reasons
>> I mention in [1], I believe that temporarily revoking access to virtio-GPU
>> blob objects is not feasible.  Instead, the pages must be treated as if
>> they are permanently in device memory until guest userspace unmaps them
>> from the GPU, after which they must be migrated back to host memory.
> 
> Discussion on IRC indicates that migration isn't reliable.  This is
> because Linux core memory management is largely lock-free for
> performance reasons, so there is no way to prevent temporary elevation
> of a page's reference count.  A page with an elevated reference count
> cannot be migrated.
> 
> The only alternative I can think of is for the hypervisor to perform the
> migration.  The hypervisor can revoke the guest's access to the page
> without the guest's consent or involvement.  The host can then replace
> the page with one of its own pages, which might be on the CPU or GPU.
> Further migration between the CPU and GPU is controlled by the host
> kernel-mode driver (KMD) and host kernel memory management.  The guest
> kernel driver must take a FOLL_LONGTERM pin before telling the host to
> use the pages, but that is all.
> 
> On KVM, this should be essentially automatic, as guest memory really is
> just host userspace memory.  On Xen, this requires that the backend
> domain can revoke fronted access to _any_ frontend page, or at least
> frontend pages that have been granted to the backend.  The backend will
> then need to be able to handle page faults for the frontend pages, and
> replace the frontend pages with its own pages at will.  Furthermore,
> revoking pages that the backend has installed into the frontend must
> never fail, because the backend will panic if it does fail.
> 
> Sima, is putting guest pages under host kernel control the only option?
> I thought that this could be avoided by leaving the pages on the CPU if
> migration fails, but that won't work because there will be no way to
> migrate them to the GPU later, causing performance problems that would
> be impossible to debug.  Is waiting (possibly forever) on migration to
> finish an option?  Otherwise, this might mean extra complexity in the
> Xen hypervisor, as I do not believe the primitives needed are currently
> available.  Specifically, in addition to the primitives discussed at Xen
> Project Summit 2024, the backend also needs to intercept access to, and
> replace the contents of, arbitrary frontend-controlled pages.

Hi Demi,

I agree that to achieve the complete SVM feature in virtio-GPU, it is 
necessary to have the hypervisor deeply involved and add new features.
It needs solid design, I saw the detailed reply in a another thread, it
is very helpful,looking forward to the response from the Xen/hypervisor 
experts.

So for the current virito-GPU userptr implementation, It can not support 
the full SVM feature, it just can only let GPU access the user space 
memory, maybe can be called by userptr feature. I think I will finish 
this small part firstly and then to try to complete the whole SVM feature.

Regards,
Honglei

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Demi Marie Obenour]

[-- Attachment #1: Type: text/plain, Size: 11250 bytes --]

On Thu, Feb 06, 2025 at 06:53:55PM +0800, Huang, Honglei1 wrote:
> On 2025/1/31 8:33, Demi Marie Obenour wrote:
> > On Wed, Jan 29, 2025 at 03:54:59PM -0500, Demi Marie Obenour wrote:
> > > On 1/8/25 12:05 PM, Simona Vetter wrote:
> > > > On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
> > > > > 
> > > > > On 2024/12/22 9:59, Demi Marie Obenour wrote:
> > > > > > On 12/20/24 10:35 AM, Simona Vetter wrote:
> > > > > > > On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
> > > > > > > > From: Honglei Huang <Honglei1.Huang@amd.com>
> > > > > > > > 
> > > > > > > > A virtio-gpu userptr is based on HMM notifier.
> > > > > > > > Used for let host access guest userspace memory and
> > > > > > > > notice the change of userspace memory.
> > > > > > > > This series patches are in very beginning state,
> > > > > > > > User space are pinned currently to ensure the host
> > > > > > > > device memory operations are correct.
> > > > > > > > The free and unmap operations for userspace can be
> > > > > > > > handled by MMU notifier this is a simple and basice
> > > > > > > > SVM feature for this series patches.
> > > > > > > > The physical PFNS update operations is splited into
> > > > > > > > two OPs in here. The evicted memories won't be used
> > > > > > > > anymore but remap into host again to achieve same
> > > > > > > > effect with hmm_rang_fault.
> > > > > > > 
> > > > > > > So in my opinion there are two ways to implement userptr that make sense:
> > > > > > > 
> > > > > > > - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
> > > > > > >     notifier
> > > > > > > 
> > > > > > > - unpinnned userptr where you entirely rely on userptr and do not hold any
> > > > > > >     page references or page pins at all, for full SVM integration. This
> > > > > > >     should use hmm_range_fault ideally, since that's the version that
> > > > > > >     doesn't ever grab any page reference pins.
> > > > > > > 
> > > > > > > All the in-between variants are imo really bad hacks, whether they hold a
> > > > > > > page reference or a temporary page pin (which seems to be what you're
> > > > > > > doing here). In much older kernels there was some justification for them,
> > > > > > > because strange stuff happened over fork(), but with FOLL_LONGTERM this is
> > > > > > > now all sorted out. So there's really only fully pinned, or true svm left
> > > > > > > as clean design choices imo.
> > > > > > > 
> > > > > > > With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
> > > > > > > you?
> > > > > > 
> > > > > > +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
> > > > > > in complexity that pinning everything avoids.  Furthermore, this avoids the
> > > > > > host having to take action in response to guest memory reclaim requests.
> > > > > > This avoids additional complexity (and thus attack surface) on the host side.
> > > > > > Furthermore, since this is for ROCm and not for graphics, I am less concerned
> > > > > > about supporting systems that require swappable GPU VRAM.
> > > > > 
> > > > > Hi Sima and Demi,
> > > > > 
> > > > > I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
> > > > > version.
> > > > > 
> > > > > And for the first pin variants implementation, the MMU notifier is also
> > > > > needed I think.Cause the userptr feature in UMD generally used like this:
> > > > > the registering of userptr always is explicitly invoked by user code like
> > > > > "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
> > > > > there is no explicit API for it, at least in hsakmt/KFD stack. User just
> > > > > need call system call "free(userptrAddr)", then kernel driver will release
> > > > > the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
> > > > > user has been free the userptr except for MMU notifior.And in UMD theres is
> > > > > no way to get the free() operation is invoked by user.The only way is use
> > > > > MMU notifier in virtio-GPU driver and free the corresponding data in host by
> > > > > some virtio CMDs as far as I can see.
> > > > > 
> > > > > And for the second way that is use hmm_range_fault, there is a predictable
> > > > > issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
> > > > > may migrate when GPU/device is working. In bare metal, when memory is
> > > > > migrating KFD driver will pause the compute work of the device in
> > > > > mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
> > > > > memories to GPU then restore the compute work of device to ensure the
> > > > > correction of the data. But in virtio-GPU driver the migration happen in
> > > > > guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
> > > > > can be used for notify host but as lack of mmap_write_lock protection in
> > > > > host kernel, host will hold invalid data for a short period of time, this
> > > > > may lead to some issues. And it is hard to fix as far as I can see.
> > > > > 
> > > > > I will extract some APIs into helper according to your request, and I will
> > > > > refactor the whole userptr implementation, use some callbacks in page
> > > > > getting path, let the pin method and hmm_range_fault can be choiced
> > > > > in this series patches.
> > > > 
> > > > Ok, so if this is for svm, then you need full blast hmm, or the semantics
> > > > are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
> > > > not work.
> > > > 
> > > > The other option is that hsakmt/kfd api is completely busted, and that's
> > > > kinda not a kernel problem.
> > > > -Sima
> > > 
> > > On further thought, I believe the driver needs to migrate the pages to
> > > device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
> > > pin on them.  The reason is that it isn’t possible to migrate these pages
> > > back to "host" memory without unmapping them from the GPU.  For the reasons
> > > I mention in [1], I believe that temporarily revoking access to virtio-GPU
> > > blob objects is not feasible.  Instead, the pages must be treated as if
> > > they are permanently in device memory until guest userspace unmaps them
> > > from the GPU, after which they must be migrated back to host memory.
> > 
> > Discussion on IRC indicates that migration isn't reliable.  This is
> > because Linux core memory management is largely lock-free for
> > performance reasons, so there is no way to prevent temporary elevation
> > of a page's reference count.  A page with an elevated reference count
> > cannot be migrated.
> > 
> > The only alternative I can think of is for the hypervisor to perform the
> > migration.  The hypervisor can revoke the guest's access to the page
> > without the guest's consent or involvement.  The host can then replace
> > the page with one of its own pages, which might be on the CPU or GPU.
> > Further migration between the CPU and GPU is controlled by the host
> > kernel-mode driver (KMD) and host kernel memory management.  The guest
> > kernel driver must take a FOLL_LONGTERM pin before telling the host to
> > use the pages, but that is all.
> > 
> > On KVM, this should be essentially automatic, as guest memory really is
> > just host userspace memory.  On Xen, this requires that the backend
> > domain can revoke fronted access to _any_ frontend page, or at least
> > frontend pages that have been granted to the backend.  The backend will
> > then need to be able to handle page faults for the frontend pages, and
> > replace the frontend pages with its own pages at will.  Furthermore,
> > revoking pages that the backend has installed into the frontend must
> > never fail, because the backend will panic if it does fail.
> > 
> > Sima, is putting guest pages under host kernel control the only option?
> > I thought that this could be avoided by leaving the pages on the CPU if
> > migration fails, but that won't work because there will be no way to
> > migrate them to the GPU later, causing performance problems that would
> > be impossible to debug.  Is waiting (possibly forever) on migration to
> > finish an option?  Otherwise, this might mean extra complexity in the
> > Xen hypervisor, as I do not believe the primitives needed are currently
> > available.  Specifically, in addition to the primitives discussed at Xen
> > Project Summit 2024, the backend also needs to intercept access to, and
> > replace the contents of, arbitrary frontend-controlled pages.
> 
> Hi Demi,
> 
> I agree that to achieve the complete SVM feature in virtio-GPU, it is
> necessary to have the hypervisor deeply involved and add new features.
> It needs solid design, I saw the detailed reply in a another thread, it
> is very helpful,looking forward to the response from the Xen/hypervisor
> experts.

From further discussion with Sima, I suspect that virtio-GPU cannot
support SVM with reasonable performance.  Native contexts have such good
performance for graphics workloads because graphics workloads very rarely
perform blocking waits for host GPU operations to complete, so one can
make all frequently-used operations asynchronous and therefore hide the
guest <=> host latency.  SVM seems to require many synchronous GPU
operations, and I believe those will severely harm performance with
virtio-GPU.

If you need full SVM for your workloads, I recommend using hardware
SR-IOV.  This should allow the guest to perform GPU virtual memory
operations without host involvement, which I expect will be much faster
than paravirtualizing these operations.  Scalable I/O virtualization
might also work, but it might also require paravirtualizing the
performance-critical address-space operations unless the hardware has
stage 2 translation tables.

> So for the current virito-GPU userptr implementation, It can not support the
> full SVM feature, it just can only let GPU access the user space memory,
> maybe can be called by userptr feature. I think I will finish this small
> part firstly and then to try to complete the whole SVM feature.

I think you will still have problems if the host is able to migrate
pages in any way.  This requires that the host install an MMU notifier
for the pages it has received from the guest, which in turn implies that
the host must be able to prevent the guest from accessing the pages.
If the pages are used in grant table operations, this isn't possible.

If you are willing to have the pages be pinned on the host side things
are much simpler.  Such pages will always be in system memory, and will
never be able to migrate to VRAM.  This will result in a performance
penalty and will likely require explicit prefetching by programs using
ROCm, but this may be acceptable for your use-cases.  The performance
penalty is the same as that with XNACK disabled, which is the case for
all RDNA2+ GPUs, so all code that aims to be portable to recent consumer
hardware will have to account for it anyway.
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Huang, Honglei1]

On 2025/2/7 2:21, Demi Marie Obenour wrote:
> On Thu, Feb 06, 2025 at 06:53:55PM +0800, Huang, Honglei1 wrote:
>> On 2025/1/31 8:33, Demi Marie Obenour wrote:
>>> On Wed, Jan 29, 2025 at 03:54:59PM -0500, Demi Marie Obenour wrote:
>>>> On 1/8/25 12:05 PM, Simona Vetter wrote:
>>>>> On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
>>>>>>
>>>>>> On 2024/12/22 9:59, Demi Marie Obenour wrote:
>>>>>>> On 12/20/24 10:35 AM, Simona Vetter wrote:
>>>>>>>> On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
>>>>>>>>> From: Honglei Huang <Honglei1.Huang@amd.com>
>>>>>>>>>
>>>>>>>>> A virtio-gpu userptr is based on HMM notifier.
>>>>>>>>> Used for let host access guest userspace memory and
>>>>>>>>> notice the change of userspace memory.
>>>>>>>>> This series patches are in very beginning state,
>>>>>>>>> User space are pinned currently to ensure the host
>>>>>>>>> device memory operations are correct.
>>>>>>>>> The free and unmap operations for userspace can be
>>>>>>>>> handled by MMU notifier this is a simple and basice
>>>>>>>>> SVM feature for this series patches.
>>>>>>>>> The physical PFNS update operations is splited into
>>>>>>>>> two OPs in here. The evicted memories won't be used
>>>>>>>>> anymore but remap into host again to achieve same
>>>>>>>>> effect with hmm_rang_fault.
>>>>>>>>
>>>>>>>> So in my opinion there are two ways to implement userptr that make sense:
>>>>>>>>
>>>>>>>> - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
>>>>>>>>      notifier
>>>>>>>>
>>>>>>>> - unpinnned userptr where you entirely rely on userptr and do not hold any
>>>>>>>>      page references or page pins at all, for full SVM integration. This
>>>>>>>>      should use hmm_range_fault ideally, since that's the version that
>>>>>>>>      doesn't ever grab any page reference pins.
>>>>>>>>
>>>>>>>> All the in-between variants are imo really bad hacks, whether they hold a
>>>>>>>> page reference or a temporary page pin (which seems to be what you're
>>>>>>>> doing here). In much older kernels there was some justification for them,
>>>>>>>> because strange stuff happened over fork(), but with FOLL_LONGTERM this is
>>>>>>>> now all sorted out. So there's really only fully pinned, or true svm left
>>>>>>>> as clean design choices imo.
>>>>>>>>
>>>>>>>> With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
>>>>>>>> you?
>>>>>>>
>>>>>>> +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
>>>>>>> in complexity that pinning everything avoids.  Furthermore, this avoids the
>>>>>>> host having to take action in response to guest memory reclaim requests.
>>>>>>> This avoids additional complexity (and thus attack surface) on the host side.
>>>>>>> Furthermore, since this is for ROCm and not for graphics, I am less concerned
>>>>>>> about supporting systems that require swappable GPU VRAM.
>>>>>>
>>>>>> Hi Sima and Demi,
>>>>>>
>>>>>> I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
>>>>>> version.
>>>>>>
>>>>>> And for the first pin variants implementation, the MMU notifier is also
>>>>>> needed I think.Cause the userptr feature in UMD generally used like this:
>>>>>> the registering of userptr always is explicitly invoked by user code like
>>>>>> "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
>>>>>> there is no explicit API for it, at least in hsakmt/KFD stack. User just
>>>>>> need call system call "free(userptrAddr)", then kernel driver will release
>>>>>> the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
>>>>>> user has been free the userptr except for MMU notifior.And in UMD theres is
>>>>>> no way to get the free() operation is invoked by user.The only way is use
>>>>>> MMU notifier in virtio-GPU driver and free the corresponding data in host by
>>>>>> some virtio CMDs as far as I can see.
>>>>>>
>>>>>> And for the second way that is use hmm_range_fault, there is a predictable
>>>>>> issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
>>>>>> may migrate when GPU/device is working. In bare metal, when memory is
>>>>>> migrating KFD driver will pause the compute work of the device in
>>>>>> mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
>>>>>> memories to GPU then restore the compute work of device to ensure the
>>>>>> correction of the data. But in virtio-GPU driver the migration happen in
>>>>>> guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
>>>>>> can be used for notify host but as lack of mmap_write_lock protection in
>>>>>> host kernel, host will hold invalid data for a short period of time, this
>>>>>> may lead to some issues. And it is hard to fix as far as I can see.
>>>>>>
>>>>>> I will extract some APIs into helper according to your request, and I will
>>>>>> refactor the whole userptr implementation, use some callbacks in page
>>>>>> getting path, let the pin method and hmm_range_fault can be choiced
>>>>>> in this series patches.
>>>>>
>>>>> Ok, so if this is for svm, then you need full blast hmm, or the semantics
>>>>> are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
>>>>> not work.
>>>>>
>>>>> The other option is that hsakmt/kfd api is completely busted, and that's
>>>>> kinda not a kernel problem.
>>>>> -Sima
>>>>
>>>> On further thought, I believe the driver needs to migrate the pages to
>>>> device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
>>>> pin on them.  The reason is that it isn’t possible to migrate these pages
>>>> back to "host" memory without unmapping them from the GPU.  For the reasons
>>>> I mention in [1], I believe that temporarily revoking access to virtio-GPU
>>>> blob objects is not feasible.  Instead, the pages must be treated as if
>>>> they are permanently in device memory until guest userspace unmaps them
>>>> from the GPU, after which they must be migrated back to host memory.
>>>
>>> Discussion on IRC indicates that migration isn't reliable.  This is
>>> because Linux core memory management is largely lock-free for
>>> performance reasons, so there is no way to prevent temporary elevation
>>> of a page's reference count.  A page with an elevated reference count
>>> cannot be migrated.
>>>
>>> The only alternative I can think of is for the hypervisor to perform the
>>> migration.  The hypervisor can revoke the guest's access to the page
>>> without the guest's consent or involvement.  The host can then replace
>>> the page with one of its own pages, which might be on the CPU or GPU.
>>> Further migration between the CPU and GPU is controlled by the host
>>> kernel-mode driver (KMD) and host kernel memory management.  The guest
>>> kernel driver must take a FOLL_LONGTERM pin before telling the host to
>>> use the pages, but that is all.
>>>
>>> On KVM, this should be essentially automatic, as guest memory really is
>>> just host userspace memory.  On Xen, this requires that the backend
>>> domain can revoke fronted access to _any_ frontend page, or at least
>>> frontend pages that have been granted to the backend.  The backend will
>>> then need to be able to handle page faults for the frontend pages, and
>>> replace the frontend pages with its own pages at will.  Furthermore,
>>> revoking pages that the backend has installed into the frontend must
>>> never fail, because the backend will panic if it does fail.
>>>
>>> Sima, is putting guest pages under host kernel control the only option?
>>> I thought that this could be avoided by leaving the pages on the CPU if
>>> migration fails, but that won't work because there will be no way to
>>> migrate them to the GPU later, causing performance problems that would
>>> be impossible to debug.  Is waiting (possibly forever) on migration to
>>> finish an option?  Otherwise, this might mean extra complexity in the
>>> Xen hypervisor, as I do not believe the primitives needed are currently
>>> available.  Specifically, in addition to the primitives discussed at Xen
>>> Project Summit 2024, the backend also needs to intercept access to, and
>>> replace the contents of, arbitrary frontend-controlled pages.
>>
>> Hi Demi,
>>
>> I agree that to achieve the complete SVM feature in virtio-GPU, it is
>> necessary to have the hypervisor deeply involved and add new features.
>> It needs solid design, I saw the detailed reply in a another thread, it
>> is very helpful,looking forward to the response from the Xen/hypervisor
>> experts.
> 
>  From further discussion with Sima, I suspect that virtio-GPU cannot
> support SVM with reasonable performance.  Native contexts have such good
> performance for graphics workloads because graphics workloads very rarely
> perform blocking waits for host GPU operations to complete, so one can
> make all frequently-used operations asynchronous and therefore hide the
> guest <=> host latency.  SVM seems to require many synchronous GPU
> operations, and I believe those will severely harm performance with
> virtio-GPU.
> 
> If you need full SVM for your workloads, I recommend using hardware
> SR-IOV.  This should allow the guest to perform GPU virtual memory
> operations without host involvement, which I expect will be much faster
> than paravirtualizing these operations.  Scalable I/O virtualization
> might also work, but it might also require paravirtualizing the
> performance-critical address-space operations unless the hardware has
> stage 2 translation tables.
> 

Yes I think so, the SR-IOV or some other hardware virtualization are 
clean design for ROCm/compute currently. But actually those hardware 
features supported solution also have their own limitation, like high 
hardware cost and the performance decreasing caused by different guest 
VMs hardware workload schedule. We are trying a low-cost, 
high-performance virtualization solution, it appears to be difficult to 
support full feature VS SR-IOV at present. But it doesn't prevent us 
from enabling part of functions.

>> So for the current virito-GPU userptr implementation, It can not support the
>> full SVM feature, it just can only let GPU access the user space memory,
>> maybe can be called by userptr feature. I think I will finish this small
>> part firstly and then to try to complete the whole SVM feature.
> 
> I think you will still have problems if the host is able to migrate
> pages in any way.  This requires that the host install an MMU notifier
> for the pages it has received from the guest, which in turn implies that
> the host must be able to prevent the guest from accessing the pages.
> If the pages are used in grant table operations, this isn't possible.
> 
> If you are willing to have the pages be pinned on the host side things
> are much simpler.  Such pages will always be in system memory, and will
> never be able to migrate to VRAM.  This will result in a performance
> penalty and will likely require explicit prefetching by programs using
> ROCm, but this may be acceptable for your use-cases.  The performance
> penalty is the same as that with XNACK disabled, which is the case for
> all RDNA2+ GPUs, so all code that aims to be portable to recent consumer
> hardware will have to account for it anyway.

Totally agreed. Actually memory migrating to VRAM is very common in GFX 
side, but in ROCm/KFD, maybe it can be disabled and not often used as 
far as I know. ROCm/KFD always uses SDMA to transfer or copy data maybe 
this is faster than migrating to VRAM (needs further verification).
But we have some method to workaround it. Really thanks for your reminding.

Regards,
Honglei

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Demi Marie Obenour]

[-- Attachment #1: Type: text/plain, Size: 16077 bytes --]

On Fri, Feb 07, 2025 at 07:07:11PM +0800, Huang, Honglei1 wrote:
> On 2025/2/7 2:21, Demi Marie Obenour wrote:
> > On Thu, Feb 06, 2025 at 06:53:55PM +0800, Huang, Honglei1 wrote:
> > > On 2025/1/31 8:33, Demi Marie Obenour wrote:
> > > > On Wed, Jan 29, 2025 at 03:54:59PM -0500, Demi Marie Obenour wrote:
> > > > > On 1/8/25 12:05 PM, Simona Vetter wrote:
> > > > > > On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
> > > > > > > 
> > > > > > > On 2024/12/22 9:59, Demi Marie Obenour wrote:
> > > > > > > > On 12/20/24 10:35 AM, Simona Vetter wrote:
> > > > > > > > > On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
> > > > > > > > > > From: Honglei Huang <Honglei1.Huang@amd.com>
> > > > > > > > > > 
> > > > > > > > > > A virtio-gpu userptr is based on HMM notifier.
> > > > > > > > > > Used for let host access guest userspace memory and
> > > > > > > > > > notice the change of userspace memory.
> > > > > > > > > > This series patches are in very beginning state,
> > > > > > > > > > User space are pinned currently to ensure the host
> > > > > > > > > > device memory operations are correct.
> > > > > > > > > > The free and unmap operations for userspace can be
> > > > > > > > > > handled by MMU notifier this is a simple and basice
> > > > > > > > > > SVM feature for this series patches.
> > > > > > > > > > The physical PFNS update operations is splited into
> > > > > > > > > > two OPs in here. The evicted memories won't be used
> > > > > > > > > > anymore but remap into host again to achieve same
> > > > > > > > > > effect with hmm_rang_fault.
> > > > > > > > > 
> > > > > > > > > So in my opinion there are two ways to implement userptr that make sense:
> > > > > > > > > 
> > > > > > > > > - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
> > > > > > > > >      notifier
> > > > > > > > > 
> > > > > > > > > - unpinnned userptr where you entirely rely on userptr and do not hold any
> > > > > > > > >      page references or page pins at all, for full SVM integration. This
> > > > > > > > >      should use hmm_range_fault ideally, since that's the version that
> > > > > > > > >      doesn't ever grab any page reference pins.
> > > > > > > > > 
> > > > > > > > > All the in-between variants are imo really bad hacks, whether they hold a
> > > > > > > > > page reference or a temporary page pin (which seems to be what you're
> > > > > > > > > doing here). In much older kernels there was some justification for them,
> > > > > > > > > because strange stuff happened over fork(), but with FOLL_LONGTERM this is
> > > > > > > > > now all sorted out. So there's really only fully pinned, or true svm left
> > > > > > > > > as clean design choices imo.
> > > > > > > > > 
> > > > > > > > > With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
> > > > > > > > > you?
> > > > > > > > 
> > > > > > > > +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
> > > > > > > > in complexity that pinning everything avoids.  Furthermore, this avoids the
> > > > > > > > host having to take action in response to guest memory reclaim requests.
> > > > > > > > This avoids additional complexity (and thus attack surface) on the host side.
> > > > > > > > Furthermore, since this is for ROCm and not for graphics, I am less concerned
> > > > > > > > about supporting systems that require swappable GPU VRAM.
> > > > > > > 
> > > > > > > Hi Sima and Demi,
> > > > > > > 
> > > > > > > I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
> > > > > > > version.
> > > > > > > 
> > > > > > > And for the first pin variants implementation, the MMU notifier is also
> > > > > > > needed I think.Cause the userptr feature in UMD generally used like this:
> > > > > > > the registering of userptr always is explicitly invoked by user code like
> > > > > > > "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
> > > > > > > there is no explicit API for it, at least in hsakmt/KFD stack. User just
> > > > > > > need call system call "free(userptrAddr)", then kernel driver will release
> > > > > > > the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
> > > > > > > user has been free the userptr except for MMU notifior.And in UMD theres is
> > > > > > > no way to get the free() operation is invoked by user.The only way is use
> > > > > > > MMU notifier in virtio-GPU driver and free the corresponding data in host by
> > > > > > > some virtio CMDs as far as I can see.
> > > > > > > 
> > > > > > > And for the second way that is use hmm_range_fault, there is a predictable
> > > > > > > issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
> > > > > > > may migrate when GPU/device is working. In bare metal, when memory is
> > > > > > > migrating KFD driver will pause the compute work of the device in
> > > > > > > mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
> > > > > > > memories to GPU then restore the compute work of device to ensure the
> > > > > > > correction of the data. But in virtio-GPU driver the migration happen in
> > > > > > > guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
> > > > > > > can be used for notify host but as lack of mmap_write_lock protection in
> > > > > > > host kernel, host will hold invalid data for a short period of time, this
> > > > > > > may lead to some issues. And it is hard to fix as far as I can see.
> > > > > > > 
> > > > > > > I will extract some APIs into helper according to your request, and I will
> > > > > > > refactor the whole userptr implementation, use some callbacks in page
> > > > > > > getting path, let the pin method and hmm_range_fault can be choiced
> > > > > > > in this series patches.
> > > > > > 
> > > > > > Ok, so if this is for svm, then you need full blast hmm, or the semantics
> > > > > > are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
> > > > > > not work.
> > > > > > 
> > > > > > The other option is that hsakmt/kfd api is completely busted, and that's
> > > > > > kinda not a kernel problem.
> > > > > > -Sima
> > > > > 
> > > > > On further thought, I believe the driver needs to migrate the pages to
> > > > > device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
> > > > > pin on them.  The reason is that it isn’t possible to migrate these pages
> > > > > back to "host" memory without unmapping them from the GPU.  For the reasons
> > > > > I mention in [1], I believe that temporarily revoking access to virtio-GPU
> > > > > blob objects is not feasible.  Instead, the pages must be treated as if
> > > > > they are permanently in device memory until guest userspace unmaps them
> > > > > from the GPU, after which they must be migrated back to host memory.
> > > > 
> > > > Discussion on IRC indicates that migration isn't reliable.  This is
> > > > because Linux core memory management is largely lock-free for
> > > > performance reasons, so there is no way to prevent temporary elevation
> > > > of a page's reference count.  A page with an elevated reference count
> > > > cannot be migrated.
> > > > 
> > > > The only alternative I can think of is for the hypervisor to perform the
> > > > migration.  The hypervisor can revoke the guest's access to the page
> > > > without the guest's consent or involvement.  The host can then replace
> > > > the page with one of its own pages, which might be on the CPU or GPU.
> > > > Further migration between the CPU and GPU is controlled by the host
> > > > kernel-mode driver (KMD) and host kernel memory management.  The guest
> > > > kernel driver must take a FOLL_LONGTERM pin before telling the host to
> > > > use the pages, but that is all.
> > > > 
> > > > On KVM, this should be essentially automatic, as guest memory really is
> > > > just host userspace memory.  On Xen, this requires that the backend
> > > > domain can revoke fronted access to _any_ frontend page, or at least
> > > > frontend pages that have been granted to the backend.  The backend will
> > > > then need to be able to handle page faults for the frontend pages, and
> > > > replace the frontend pages with its own pages at will.  Furthermore,
> > > > revoking pages that the backend has installed into the frontend must
> > > > never fail, because the backend will panic if it does fail.
> > > > 
> > > > Sima, is putting guest pages under host kernel control the only option?
> > > > I thought that this could be avoided by leaving the pages on the CPU if
> > > > migration fails, but that won't work because there will be no way to
> > > > migrate them to the GPU later, causing performance problems that would
> > > > be impossible to debug.  Is waiting (possibly forever) on migration to
> > > > finish an option?  Otherwise, this might mean extra complexity in the
> > > > Xen hypervisor, as I do not believe the primitives needed are currently
> > > > available.  Specifically, in addition to the primitives discussed at Xen
> > > > Project Summit 2024, the backend also needs to intercept access to, and
> > > > replace the contents of, arbitrary frontend-controlled pages.
> > > 
> > > Hi Demi,
> > > 
> > > I agree that to achieve the complete SVM feature in virtio-GPU, it is
> > > necessary to have the hypervisor deeply involved and add new features.
> > > It needs solid design, I saw the detailed reply in a another thread, it
> > > is very helpful,looking forward to the response from the Xen/hypervisor
> > > experts.
> > 
> >  From further discussion with Sima, I suspect that virtio-GPU cannot
> > support SVM with reasonable performance.  Native contexts have such good
> > performance for graphics workloads because graphics workloads very rarely
> > perform blocking waits for host GPU operations to complete, so one can
> > make all frequently-used operations asynchronous and therefore hide the
> > guest <=> host latency.  SVM seems to require many synchronous GPU
> > operations, and I believe those will severely harm performance with
> > virtio-GPU.
> > 
> > If you need full SVM for your workloads, I recommend using hardware
> > SR-IOV.  This should allow the guest to perform GPU virtual memory
> > operations without host involvement, which I expect will be much faster
> > than paravirtualizing these operations.  Scalable I/O virtualization
> > might also work, but it might also require paravirtualizing the
> > performance-critical address-space operations unless the hardware has
> > stage 2 translation tables.
> > 
> 
> Yes I think so, the SR-IOV or some other hardware virtualization are clean
> design for ROCm/compute currently. But actually those hardware features
> supported solution also have their own limitation, like high hardware cost
> and the performance decreasing caused by different guest VMs hardware
> workload schedule. We are trying a low-cost, high-performance virtualization
> solution, it appears to be difficult to support full feature VS SR-IOV at
> present. But it doesn't prevent us from enabling part of functions.
> 
> > > So for the current virito-GPU userptr implementation, It can not support the
> > > full SVM feature, it just can only let GPU access the user space memory,
> > > maybe can be called by userptr feature. I think I will finish this small
> > > part firstly and then to try to complete the whole SVM feature.
> > 
> > I think you will still have problems if the host is able to migrate
> > pages in any way.  This requires that the host install an MMU notifier
> > for the pages it has received from the guest, which in turn implies that
> > the host must be able to prevent the guest from accessing the pages.
> > If the pages are used in grant table operations, this isn't possible.
> > 
> > If you are willing to have the pages be pinned on the host side things
> > are much simpler.  Such pages will always be in system memory, and will
> > never be able to migrate to VRAM.  This will result in a performance
> > penalty and will likely require explicit prefetching by programs using
> > ROCm, but this may be acceptable for your use-cases.  The performance
> > penalty is the same as that with XNACK disabled, which is the case for
> > all RDNA2+ GPUs, so all code that aims to be portable to recent consumer
> > hardware will have to account for it anyway.
> 
> Totally agreed. Actually memory migrating to VRAM is very common in GFX
> side, but in ROCm/KFD, maybe it can be disabled and not often used as far as
> I know. ROCm/KFD always uses SDMA to transfer or copy data maybe this is
> faster than migrating to VRAM (needs further verification).
> But we have some method to workaround it. Really thanks for your reminding.

I think you will do okay if you treat virtio-GPU as providing a virtual
GPU with no XNACK support.  XNACK is necessary for migrating pages
between GPU and CPU based on demand, and it is this migration that is
so hard to implement.  Furthermore, I highly doubt that the combination
of AMDKFD and the hardware you are targeting even supports XNACK.

At Xen Project Summit 2024, AMD mentioned that it wanted to enable both
rendering (Vulkan/OpenGL) and compute (ROCm) with virtio-GPU native
contexts under Xen.  The only GPUs for which AMDKFD will enable XNACK
support are GFX9 GPUs, which are GCN and CDNA.  Shipping a GCN GPU in a
new design would be very unusual and CDNA (Instinct) accelerators do not
support graphics, so either AMD is using separate devices for compute
and graphics or the workloads will run with no XNACK support.  Since you
mention HW cost as an important consideration, I suspect the latter.

I believe that the Instinct accelerators that support XNACK also support
SR-IOV, but if you wish to combine XNACK and virtio-GPU, this should be
possible subject to caveats.  The main caveat is that under no
circumstances can the host's Xen driver install an MMU notifier for
pages that the guest can use in grant table operations or DMA.  A driver
that installs an MMU notifier promises that it can block access to
pages in a bounded amount of time, and if the guest can DMA to the pages
or grant them to other domains this is not possible.  Without the Xen
driver installing an MMU notifier, there is no way for the pages to be
migrated to the GPU without risking use-after-free or at least data
corruption.  Instead, one of the following options will be needed:

1. hipMallocManaged() allocates the memory from the backend using the
   Map primitive discussed elsewhere.  Such memory is not mappable in
   the IOMMU (if there is an assigned PCI device) and cannot be used for
   grant table operations.  Memory allocated via system allocators
   (anonymous pages) is not able to be migrated.

2. The frontend uses shadow buffers for all I/O.  This allows the
   backend to use a new Steal primitive to revoke the guest's accesses
   to anonymous pages and handle page faults accordingly.

3. Same as 2 except that the frontend allocates all memory (except
   bounce buffers) from the backend, just like a KVM guest does, rather
   than from the Xen toolstack.

4. The frontend tries to migrate the pages to backend-provided ones, and
   falls back to leaving them pinned on the CPU.  The frontend's MMU
   notifier tells the backend to stop accessing the pages, blocking
   until the backend confirms this.  The frontend then moves the pages
   to its own memory and returns from the notifier.  This may require
   new AMDKFD APIs.

5. Same as 4 except that the frontend uses hmm_range_fault to move the
   pages to the backend in response to GPU page faults.  This requires a
   frontend <-> backend round-trip for each fault (slooooow) so a new
   fast mechanism for this might be needed.
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Demi Marie Obenour]

[-- Attachment #1: Type: text/plain, Size: 16752 bytes --]

On Fri, Feb 07, 2025 at 09:30:45PM -0500, Demi Marie Obenour wrote:
> On Fri, Feb 07, 2025 at 07:07:11PM +0800, Huang, Honglei1 wrote:
> > On 2025/2/7 2:21, Demi Marie Obenour wrote:
> > > On Thu, Feb 06, 2025 at 06:53:55PM +0800, Huang, Honglei1 wrote:
> > > > On 2025/1/31 8:33, Demi Marie Obenour wrote:
> > > > > On Wed, Jan 29, 2025 at 03:54:59PM -0500, Demi Marie Obenour wrote:
> > > > > > On 1/8/25 12:05 PM, Simona Vetter wrote:
> > > > > > > On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
> > > > > > > > 
> > > > > > > > On 2024/12/22 9:59, Demi Marie Obenour wrote:
> > > > > > > > > On 12/20/24 10:35 AM, Simona Vetter wrote:
> > > > > > > > > > On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
> > > > > > > > > > > From: Honglei Huang <Honglei1.Huang@amd.com>
> > > > > > > > > > > 
> > > > > > > > > > > A virtio-gpu userptr is based on HMM notifier.
> > > > > > > > > > > Used for let host access guest userspace memory and
> > > > > > > > > > > notice the change of userspace memory.
> > > > > > > > > > > This series patches are in very beginning state,
> > > > > > > > > > > User space are pinned currently to ensure the host
> > > > > > > > > > > device memory operations are correct.
> > > > > > > > > > > The free and unmap operations for userspace can be
> > > > > > > > > > > handled by MMU notifier this is a simple and basice
> > > > > > > > > > > SVM feature for this series patches.
> > > > > > > > > > > The physical PFNS update operations is splited into
> > > > > > > > > > > two OPs in here. The evicted memories won't be used
> > > > > > > > > > > anymore but remap into host again to achieve same
> > > > > > > > > > > effect with hmm_rang_fault.
> > > > > > > > > > 
> > > > > > > > > > So in my opinion there are two ways to implement userptr that make sense:
> > > > > > > > > > 
> > > > > > > > > > - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
> > > > > > > > > >      notifier
> > > > > > > > > > 
> > > > > > > > > > - unpinnned userptr where you entirely rely on userptr and do not hold any
> > > > > > > > > >      page references or page pins at all, for full SVM integration. This
> > > > > > > > > >      should use hmm_range_fault ideally, since that's the version that
> > > > > > > > > >      doesn't ever grab any page reference pins.
> > > > > > > > > > 
> > > > > > > > > > All the in-between variants are imo really bad hacks, whether they hold a
> > > > > > > > > > page reference or a temporary page pin (which seems to be what you're
> > > > > > > > > > doing here). In much older kernels there was some justification for them,
> > > > > > > > > > because strange stuff happened over fork(), but with FOLL_LONGTERM this is
> > > > > > > > > > now all sorted out. So there's really only fully pinned, or true svm left
> > > > > > > > > > as clean design choices imo.
> > > > > > > > > > 
> > > > > > > > > > With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
> > > > > > > > > > you?
> > > > > > > > > 
> > > > > > > > > +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
> > > > > > > > > in complexity that pinning everything avoids.  Furthermore, this avoids the
> > > > > > > > > host having to take action in response to guest memory reclaim requests.
> > > > > > > > > This avoids additional complexity (and thus attack surface) on the host side.
> > > > > > > > > Furthermore, since this is for ROCm and not for graphics, I am less concerned
> > > > > > > > > about supporting systems that require swappable GPU VRAM.
> > > > > > > > 
> > > > > > > > Hi Sima and Demi,
> > > > > > > > 
> > > > > > > > I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
> > > > > > > > version.
> > > > > > > > 
> > > > > > > > And for the first pin variants implementation, the MMU notifier is also
> > > > > > > > needed I think.Cause the userptr feature in UMD generally used like this:
> > > > > > > > the registering of userptr always is explicitly invoked by user code like
> > > > > > > > "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
> > > > > > > > there is no explicit API for it, at least in hsakmt/KFD stack. User just
> > > > > > > > need call system call "free(userptrAddr)", then kernel driver will release
> > > > > > > > the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
> > > > > > > > user has been free the userptr except for MMU notifior.And in UMD theres is
> > > > > > > > no way to get the free() operation is invoked by user.The only way is use
> > > > > > > > MMU notifier in virtio-GPU driver and free the corresponding data in host by
> > > > > > > > some virtio CMDs as far as I can see.
> > > > > > > > 
> > > > > > > > And for the second way that is use hmm_range_fault, there is a predictable
> > > > > > > > issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
> > > > > > > > may migrate when GPU/device is working. In bare metal, when memory is
> > > > > > > > migrating KFD driver will pause the compute work of the device in
> > > > > > > > mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
> > > > > > > > memories to GPU then restore the compute work of device to ensure the
> > > > > > > > correction of the data. But in virtio-GPU driver the migration happen in
> > > > > > > > guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
> > > > > > > > can be used for notify host but as lack of mmap_write_lock protection in
> > > > > > > > host kernel, host will hold invalid data for a short period of time, this
> > > > > > > > may lead to some issues. And it is hard to fix as far as I can see.
> > > > > > > > 
> > > > > > > > I will extract some APIs into helper according to your request, and I will
> > > > > > > > refactor the whole userptr implementation, use some callbacks in page
> > > > > > > > getting path, let the pin method and hmm_range_fault can be choiced
> > > > > > > > in this series patches.
> > > > > > > 
> > > > > > > Ok, so if this is for svm, then you need full blast hmm, or the semantics
> > > > > > > are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
> > > > > > > not work.
> > > > > > > 
> > > > > > > The other option is that hsakmt/kfd api is completely busted, and that's
> > > > > > > kinda not a kernel problem.
> > > > > > > -Sima
> > > > > > 
> > > > > > On further thought, I believe the driver needs to migrate the pages to
> > > > > > device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
> > > > > > pin on them.  The reason is that it isn’t possible to migrate these pages
> > > > > > back to "host" memory without unmapping them from the GPU.  For the reasons
> > > > > > I mention in [1], I believe that temporarily revoking access to virtio-GPU
> > > > > > blob objects is not feasible.  Instead, the pages must be treated as if
> > > > > > they are permanently in device memory until guest userspace unmaps them
> > > > > > from the GPU, after which they must be migrated back to host memory.
> > > > > 
> > > > > Discussion on IRC indicates that migration isn't reliable.  This is
> > > > > because Linux core memory management is largely lock-free for
> > > > > performance reasons, so there is no way to prevent temporary elevation
> > > > > of a page's reference count.  A page with an elevated reference count
> > > > > cannot be migrated.
> > > > > 
> > > > > The only alternative I can think of is for the hypervisor to perform the
> > > > > migration.  The hypervisor can revoke the guest's access to the page
> > > > > without the guest's consent or involvement.  The host can then replace
> > > > > the page with one of its own pages, which might be on the CPU or GPU.
> > > > > Further migration between the CPU and GPU is controlled by the host
> > > > > kernel-mode driver (KMD) and host kernel memory management.  The guest
> > > > > kernel driver must take a FOLL_LONGTERM pin before telling the host to
> > > > > use the pages, but that is all.
> > > > > 
> > > > > On KVM, this should be essentially automatic, as guest memory really is
> > > > > just host userspace memory.  On Xen, this requires that the backend
> > > > > domain can revoke fronted access to _any_ frontend page, or at least
> > > > > frontend pages that have been granted to the backend.  The backend will
> > > > > then need to be able to handle page faults for the frontend pages, and
> > > > > replace the frontend pages with its own pages at will.  Furthermore,
> > > > > revoking pages that the backend has installed into the frontend must
> > > > > never fail, because the backend will panic if it does fail.
> > > > > 
> > > > > Sima, is putting guest pages under host kernel control the only option?
> > > > > I thought that this could be avoided by leaving the pages on the CPU if
> > > > > migration fails, but that won't work because there will be no way to
> > > > > migrate them to the GPU later, causing performance problems that would
> > > > > be impossible to debug.  Is waiting (possibly forever) on migration to
> > > > > finish an option?  Otherwise, this might mean extra complexity in the
> > > > > Xen hypervisor, as I do not believe the primitives needed are currently
> > > > > available.  Specifically, in addition to the primitives discussed at Xen
> > > > > Project Summit 2024, the backend also needs to intercept access to, and
> > > > > replace the contents of, arbitrary frontend-controlled pages.
> > > > 
> > > > Hi Demi,
> > > > 
> > > > I agree that to achieve the complete SVM feature in virtio-GPU, it is
> > > > necessary to have the hypervisor deeply involved and add new features.
> > > > It needs solid design, I saw the detailed reply in a another thread, it
> > > > is very helpful,looking forward to the response from the Xen/hypervisor
> > > > experts.
> > > 
> > >  From further discussion with Sima, I suspect that virtio-GPU cannot
> > > support SVM with reasonable performance.  Native contexts have such good
> > > performance for graphics workloads because graphics workloads very rarely
> > > perform blocking waits for host GPU operations to complete, so one can
> > > make all frequently-used operations asynchronous and therefore hide the
> > > guest <=> host latency.  SVM seems to require many synchronous GPU
> > > operations, and I believe those will severely harm performance with
> > > virtio-GPU.
> > > 
> > > If you need full SVM for your workloads, I recommend using hardware
> > > SR-IOV.  This should allow the guest to perform GPU virtual memory
> > > operations without host involvement, which I expect will be much faster
> > > than paravirtualizing these operations.  Scalable I/O virtualization
> > > might also work, but it might also require paravirtualizing the
> > > performance-critical address-space operations unless the hardware has
> > > stage 2 translation tables.
> > > 
> > 
> > Yes I think so, the SR-IOV or some other hardware virtualization are clean
> > design for ROCm/compute currently. But actually those hardware features
> > supported solution also have their own limitation, like high hardware cost
> > and the performance decreasing caused by different guest VMs hardware
> > workload schedule. We are trying a low-cost, high-performance virtualization
> > solution, it appears to be difficult to support full feature VS SR-IOV at
> > present. But it doesn't prevent us from enabling part of functions.
> > 
> > > > So for the current virito-GPU userptr implementation, It can not support the
> > > > full SVM feature, it just can only let GPU access the user space memory,
> > > > maybe can be called by userptr feature. I think I will finish this small
> > > > part firstly and then to try to complete the whole SVM feature.
> > > 
> > > I think you will still have problems if the host is able to migrate
> > > pages in any way.  This requires that the host install an MMU notifier
> > > for the pages it has received from the guest, which in turn implies that
> > > the host must be able to prevent the guest from accessing the pages.
> > > If the pages are used in grant table operations, this isn't possible.
> > > 
> > > If you are willing to have the pages be pinned on the host side things
> > > are much simpler.  Such pages will always be in system memory, and will
> > > never be able to migrate to VRAM.  This will result in a performance
> > > penalty and will likely require explicit prefetching by programs using
> > > ROCm, but this may be acceptable for your use-cases.  The performance
> > > penalty is the same as that with XNACK disabled, which is the case for
> > > all RDNA2+ GPUs, so all code that aims to be portable to recent consumer
> > > hardware will have to account for it anyway.
> > 
> > Totally agreed. Actually memory migrating to VRAM is very common in GFX
> > side, but in ROCm/KFD, maybe it can be disabled and not often used as far as
> > I know. ROCm/KFD always uses SDMA to transfer or copy data maybe this is
> > faster than migrating to VRAM (needs further verification).
> > But we have some method to workaround it. Really thanks for your reminding.
> 
> I think you will do okay if you treat virtio-GPU as providing a virtual
> GPU with no XNACK support.  XNACK is necessary for migrating pages
> between GPU and CPU based on demand, and it is this migration that is
> so hard to implement.  Furthermore, I highly doubt that the combination
> of AMDKFD and the hardware you are targeting even supports XNACK.
> 
> At Xen Project Summit 2024, AMD mentioned that it wanted to enable both
> rendering (Vulkan/OpenGL) and compute (ROCm) with virtio-GPU native
> contexts under Xen.  The only GPUs for which AMDKFD will enable XNACK
> support are GFX9 GPUs, which are GCN and CDNA.  Shipping a GCN GPU in a
> new design would be very unusual and CDNA (Instinct) accelerators do not
> support graphics, so either AMD is using separate devices for compute
> and graphics or the workloads will run with no XNACK support.  Since you
> mention HW cost as an important consideration, I suspect the latter.
> 
> I believe that the Instinct accelerators that support XNACK also support
> SR-IOV, but if you wish to combine XNACK and virtio-GPU, this should be
> possible subject to caveats.  The main caveat is that under no
> circumstances can the host's Xen driver install an MMU notifier for
> pages that the guest can use in grant table operations or DMA.  A driver
> that installs an MMU notifier promises that it can block access to
> pages in a bounded amount of time, and if the guest can DMA to the pages
> or grant them to other domains this is not possible.  Without the Xen
> driver installing an MMU notifier, there is no way for the pages to be
> migrated to the GPU without risking use-after-free or at least data
> corruption.  Instead, one of the following options will be needed:
> 
> 1. hipMallocManaged() allocates the memory from the backend using the
>    Map primitive discussed elsewhere.  Such memory is not mappable in
>    the IOMMU (if there is an assigned PCI device) and cannot be used for
>    grant table operations.  Memory allocated via system allocators
>    (anonymous pages) is not able to be migrated.
> 
> 2. The frontend uses shadow buffers for all I/O.  This allows the
>    backend to use a new Steal primitive to revoke the guest's accesses
>    to anonymous pages and handle page faults accordingly.
> 
> 3. Same as 2 except that the frontend allocates all memory (except
>    bounce buffers) from the backend, just like a KVM guest does, rather
>    than from the Xen toolstack.
> 
> 4. The frontend tries to migrate the pages to backend-provided ones, and
>    falls back to leaving them pinned on the CPU.  The frontend's MMU
>    notifier tells the backend to stop accessing the pages, blocking
>    until the backend confirms this.  The frontend then moves the pages
>    to its own memory and returns from the notifier.  This may require
>    new AMDKFD APIs.
> 
> 5. Same as 4 except that the frontend uses hmm_range_fault to move the
>    pages to the backend in response to GPU page faults.  This requires a
>    frontend <-> backend round-trip for each fault (slooooow) so a new
>    fast mechanism for this might be needed.
> -- 
> Sincerely,
> Demi Marie Obenour (she/her/hers)
> Invisible Things Lab

CC Simona Vetter
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [RFC PATCH 3/3] drm/virtio: implement blob userptr resource object

[Demi Marie Obenour]

[-- Attachment #1: Type: text/plain, Size: 20664 bytes --]

On Sat, Feb 08, 2025 at 05:44:14PM +0800, Huang, Honglei1 wrote:
> On 2025/2/8 10:43, Demi Marie Obenour wrote:
> > On Fri, Feb 07, 2025 at 09:30:45PM -0500, Demi Marie Obenour wrote:
> > > On Fri, Feb 07, 2025 at 07:07:11PM +0800, Huang, Honglei1 wrote:
> > > > On 2025/2/7 2:21, Demi Marie Obenour wrote:
> > > > > On Thu, Feb 06, 2025 at 06:53:55PM +0800, Huang, Honglei1 wrote:
> > > > > > On 2025/1/31 8:33, Demi Marie Obenour wrote:
> > > > > > > On Wed, Jan 29, 2025 at 03:54:59PM -0500, Demi Marie Obenour wrote:
> > > > > > > > On 1/8/25 12:05 PM, Simona Vetter wrote:
> > > > > > > > > On Fri, Dec 27, 2024 at 10:24:29AM +0800, Huang, Honglei1 wrote:
> > > > > > > > > > 
> > > > > > > > > > On 2024/12/22 9:59, Demi Marie Obenour wrote:
> > > > > > > > > > > On 12/20/24 10:35 AM, Simona Vetter wrote:
> > > > > > > > > > > > On Fri, Dec 20, 2024 at 06:04:09PM +0800, Honglei Huang wrote:
> > > > > > > > > > > > > From: Honglei Huang <Honglei1.Huang@amd.com>
> > > > > > > > > > > > > 
> > > > > > > > > > > > > A virtio-gpu userptr is based on HMM notifier.
> > > > > > > > > > > > > Used for let host access guest userspace memory and
> > > > > > > > > > > > > notice the change of userspace memory.
> > > > > > > > > > > > > This series patches are in very beginning state,
> > > > > > > > > > > > > User space are pinned currently to ensure the host
> > > > > > > > > > > > > device memory operations are correct.
> > > > > > > > > > > > > The free and unmap operations for userspace can be
> > > > > > > > > > > > > handled by MMU notifier this is a simple and basice
> > > > > > > > > > > > > SVM feature for this series patches.
> > > > > > > > > > > > > The physical PFNS update operations is splited into
> > > > > > > > > > > > > two OPs in here. The evicted memories won't be used
> > > > > > > > > > > > > anymore but remap into host again to achieve same
> > > > > > > > > > > > > effect with hmm_rang_fault.
> > > > > > > > > > > > 
> > > > > > > > > > > > So in my opinion there are two ways to implement userptr that make sense:
> > > > > > > > > > > > 
> > > > > > > > > > > > - pinned userptr with pin_user_pages(FOLL_LONGTERM). there is not mmu
> > > > > > > > > > > >       notifier
> > > > > > > > > > > > 
> > > > > > > > > > > > - unpinnned userptr where you entirely rely on userptr and do not hold any
> > > > > > > > > > > >       page references or page pins at all, for full SVM integration. This
> > > > > > > > > > > >       should use hmm_range_fault ideally, since that's the version that
> > > > > > > > > > > >       doesn't ever grab any page reference pins.
> > > > > > > > > > > > 
> > > > > > > > > > > > All the in-between variants are imo really bad hacks, whether they hold a
> > > > > > > > > > > > page reference or a temporary page pin (which seems to be what you're
> > > > > > > > > > > > doing here). In much older kernels there was some justification for them,
> > > > > > > > > > > > because strange stuff happened over fork(), but with FOLL_LONGTERM this is
> > > > > > > > > > > > now all sorted out. So there's really only fully pinned, or true svm left
> > > > > > > > > > > > as clean design choices imo.
> > > > > > > > > > > > 
> > > > > > > > > > > > With that background, why does pin_user_pages(FOLL_LONGTERM) not work for
> > > > > > > > > > > > you?
> > > > > > > > > > > 
> > > > > > > > > > > +1 on using FOLL_LONGTERM.  Fully dynamic memory management has a huge cost
> > > > > > > > > > > in complexity that pinning everything avoids.  Furthermore, this avoids the
> > > > > > > > > > > host having to take action in response to guest memory reclaim requests.
> > > > > > > > > > > This avoids additional complexity (and thus attack surface) on the host side.
> > > > > > > > > > > Furthermore, since this is for ROCm and not for graphics, I am less concerned
> > > > > > > > > > > about supporting systems that require swappable GPU VRAM.
> > > > > > > > > > 
> > > > > > > > > > Hi Sima and Demi,
> > > > > > > > > > 
> > > > > > > > > > I totally agree the flag FOLL_LONGTERM is needed, I will add it in next
> > > > > > > > > > version.
> > > > > > > > > > 
> > > > > > > > > > And for the first pin variants implementation, the MMU notifier is also
> > > > > > > > > > needed I think.Cause the userptr feature in UMD generally used like this:
> > > > > > > > > > the registering of userptr always is explicitly invoked by user code like
> > > > > > > > > > "registerMemoryToGPU(userptrAddr, ...)", but for the userptr release/free,
> > > > > > > > > > there is no explicit API for it, at least in hsakmt/KFD stack. User just
> > > > > > > > > > need call system call "free(userptrAddr)", then kernel driver will release
> > > > > > > > > > the userptr by MMU notifier callback.Virtio-GPU has no other way to know if
> > > > > > > > > > user has been free the userptr except for MMU notifior.And in UMD theres is
> > > > > > > > > > no way to get the free() operation is invoked by user.The only way is use
> > > > > > > > > > MMU notifier in virtio-GPU driver and free the corresponding data in host by
> > > > > > > > > > some virtio CMDs as far as I can see.
> > > > > > > > > > 
> > > > > > > > > > And for the second way that is use hmm_range_fault, there is a predictable
> > > > > > > > > > issues as far as I can see, at least in hsakmt/KFD stack. That is the memory
> > > > > > > > > > may migrate when GPU/device is working. In bare metal, when memory is
> > > > > > > > > > migrating KFD driver will pause the compute work of the device in
> > > > > > > > > > mmap_wirte_lock then use hmm_range_fault to remap the migrated/evicted
> > > > > > > > > > memories to GPU then restore the compute work of device to ensure the
> > > > > > > > > > correction of the data. But in virtio-GPU driver the migration happen in
> > > > > > > > > > guest kernel, the evict mmu notifier callback happens in guest, a virtio CMD
> > > > > > > > > > can be used for notify host but as lack of mmap_write_lock protection in
> > > > > > > > > > host kernel, host will hold invalid data for a short period of time, this
> > > > > > > > > > may lead to some issues. And it is hard to fix as far as I can see.
> > > > > > > > > > 
> > > > > > > > > > I will extract some APIs into helper according to your request, and I will
> > > > > > > > > > refactor the whole userptr implementation, use some callbacks in page
> > > > > > > > > > getting path, let the pin method and hmm_range_fault can be choiced
> > > > > > > > > > in this series patches.
> > > > > > > > > 
> > > > > > > > > Ok, so if this is for svm, then you need full blast hmm, or the semantics
> > > > > > > > > are buggy. You cannot fake svm with pin(FOLL_LONGTERM) userptr, this does
> > > > > > > > > not work.
> > > > > > > > > 
> > > > > > > > > The other option is that hsakmt/kfd api is completely busted, and that's
> > > > > > > > > kinda not a kernel problem.
> > > > > > > > > -Sima
> > > > > > > > 
> > > > > > > > On further thought, I believe the driver needs to migrate the pages to
> > > > > > > > device memory (really a virtio-GPU blob object) *and* take a FOLL_LONGTERM
> > > > > > > > pin on them.  The reason is that it isn’t possible to migrate these pages
> > > > > > > > back to "host" memory without unmapping them from the GPU.  For the reasons
> > > > > > > > I mention in [1], I believe that temporarily revoking access to virtio-GPU
> > > > > > > > blob objects is not feasible.  Instead, the pages must be treated as if
> > > > > > > > they are permanently in device memory until guest userspace unmaps them
> > > > > > > > from the GPU, after which they must be migrated back to host memory.
> > > > > > > 
> > > > > > > Discussion on IRC indicates that migration isn't reliable.  This is
> > > > > > > because Linux core memory management is largely lock-free for
> > > > > > > performance reasons, so there is no way to prevent temporary elevation
> > > > > > > of a page's reference count.  A page with an elevated reference count
> > > > > > > cannot be migrated.
> > > > > > > 
> > > > > > > The only alternative I can think of is for the hypervisor to perform the
> > > > > > > migration.  The hypervisor can revoke the guest's access to the page
> > > > > > > without the guest's consent or involvement.  The host can then replace
> > > > > > > the page with one of its own pages, which might be on the CPU or GPU.
> > > > > > > Further migration between the CPU and GPU is controlled by the host
> > > > > > > kernel-mode driver (KMD) and host kernel memory management.  The guest
> > > > > > > kernel driver must take a FOLL_LONGTERM pin before telling the host to
> > > > > > > use the pages, but that is all.
> > > > > > > 
> > > > > > > On KVM, this should be essentially automatic, as guest memory really is
> > > > > > > just host userspace memory.  On Xen, this requires that the backend
> > > > > > > domain can revoke fronted access to _any_ frontend page, or at least
> > > > > > > frontend pages that have been granted to the backend.  The backend will
> > > > > > > then need to be able to handle page faults for the frontend pages, and
> > > > > > > replace the frontend pages with its own pages at will.  Furthermore,
> > > > > > > revoking pages that the backend has installed into the frontend must
> > > > > > > never fail, because the backend will panic if it does fail.
> > > > > > > 
> > > > > > > Sima, is putting guest pages under host kernel control the only option?
> > > > > > > I thought that this could be avoided by leaving the pages on the CPU if
> > > > > > > migration fails, but that won't work because there will be no way to
> > > > > > > migrate them to the GPU later, causing performance problems that would
> > > > > > > be impossible to debug.  Is waiting (possibly forever) on migration to
> > > > > > > finish an option?  Otherwise, this might mean extra complexity in the
> > > > > > > Xen hypervisor, as I do not believe the primitives needed are currently
> > > > > > > available.  Specifically, in addition to the primitives discussed at Xen
> > > > > > > Project Summit 2024, the backend also needs to intercept access to, and
> > > > > > > replace the contents of, arbitrary frontend-controlled pages.
> > > > > > 
> > > > > > Hi Demi,
> > > > > > 
> > > > > > I agree that to achieve the complete SVM feature in virtio-GPU, it is
> > > > > > necessary to have the hypervisor deeply involved and add new features.
> > > > > > It needs solid design, I saw the detailed reply in a another thread, it
> > > > > > is very helpful,looking forward to the response from the Xen/hypervisor
> > > > > > experts.
> > > > > 
> > > > >   From further discussion with Sima, I suspect that virtio-GPU cannot
> > > > > support SVM with reasonable performance.  Native contexts have such good
> > > > > performance for graphics workloads because graphics workloads very rarely
> > > > > perform blocking waits for host GPU operations to complete, so one can
> > > > > make all frequently-used operations asynchronous and therefore hide the
> > > > > guest <=> host latency.  SVM seems to require many synchronous GPU
> > > > > operations, and I believe those will severely harm performance with
> > > > > virtio-GPU.
> > > > > 
> > > > > If you need full SVM for your workloads, I recommend using hardware
> > > > > SR-IOV.  This should allow the guest to perform GPU virtual memory
> > > > > operations without host involvement, which I expect will be much faster
> > > > > than paravirtualizing these operations.  Scalable I/O virtualization
> > > > > might also work, but it might also require paravirtualizing the
> > > > > performance-critical address-space operations unless the hardware has
> > > > > stage 2 translation tables.
> > > > > 
> > > > 
> > > > Yes I think so, the SR-IOV or some other hardware virtualization are clean
> > > > design for ROCm/compute currently. But actually those hardware features
> > > > supported solution also have their own limitation, like high hardware cost
> > > > and the performance decreasing caused by different guest VMs hardware
> > > > workload schedule. We are trying a low-cost, high-performance virtualization
> > > > solution, it appears to be difficult to support full feature VS SR-IOV at
> > > > present. But it doesn't prevent us from enabling part of functions.
> > > > 
> > > > > > So for the current virito-GPU userptr implementation, It can not support the
> > > > > > full SVM feature, it just can only let GPU access the user space memory,
> > > > > > maybe can be called by userptr feature. I think I will finish this small
> > > > > > part firstly and then to try to complete the whole SVM feature.
> > > > > 
> > > > > I think you will still have problems if the host is able to migrate
> > > > > pages in any way.  This requires that the host install an MMU notifier
> > > > > for the pages it has received from the guest, which in turn implies that
> > > > > the host must be able to prevent the guest from accessing the pages.
> > > > > If the pages are used in grant table operations, this isn't possible.
> > > > > 
> > > > > If you are willing to have the pages be pinned on the host side things
> > > > > are much simpler.  Such pages will always be in system memory, and will
> > > > > never be able to migrate to VRAM.  This will result in a performance
> > > > > penalty and will likely require explicit prefetching by programs using
> > > > > ROCm, but this may be acceptable for your use-cases.  The performance
> > > > > penalty is the same as that with XNACK disabled, which is the case for
> > > > > all RDNA2+ GPUs, so all code that aims to be portable to recent consumer
> > > > > hardware will have to account for it anyway.
> > > > 
> > > > Totally agreed. Actually memory migrating to VRAM is very common in GFX
> > > > side, but in ROCm/KFD, maybe it can be disabled and not often used as far as
> > > > I know. ROCm/KFD always uses SDMA to transfer or copy data maybe this is
> > > > faster than migrating to VRAM (needs further verification).
> > > > But we have some method to workaround it. Really thanks for your reminding.
> > > 
> > > I think you will do okay if you treat virtio-GPU as providing a virtual
> > > GPU with no XNACK support.  XNACK is necessary for migrating pages
> > > between GPU and CPU based on demand, and it is this migration that is
> > > so hard to implement.  Furthermore, I highly doubt that the combination
> > > of AMDKFD and the hardware you are targeting even supports XNACK.
> 
> Yes the goal of this patch set is to support functions without memory
> migration related. It seems like XNACK is hard to support at present.
> 
> > > At Xen Project Summit 2024, AMD mentioned that it wanted to enable both
> > > rendering (Vulkan/OpenGL) and compute (ROCm) with virtio-GPU native
> > > contexts under Xen.  The only GPUs for which AMDKFD will enable XNACK
> > > support are GFX9 GPUs, which are GCN and CDNA.  Shipping a GCN GPU in a
> > > new design would be very unusual and CDNA (Instinct) accelerators do not
> > > support graphics, so either AMD is using separate devices for compute
> > > and graphics or the workloads will run with no XNACK support.  Since you
> > > mention HW cost as an important consideration, I suspect the latter.
> > > 
> > > I believe that the Instinct accelerators that support XNACK also support
> > > SR-IOV, but if you wish to combine XNACK and virtio-GPU, this should be
> > > possible subject to caveats.  The main caveat is that under no
> > > circumstances can the host's Xen driver install an MMU notifier for
> > > pages that the guest can use in grant table operations or DMA.  A driver
> > > that installs an MMU notifier promises that it can block access to
> > > pages in a bounded amount of time, and if the guest can DMA to the pages
> > > or grant them to other domains this is not possible.  Without the Xen
> > > driver installing an MMU notifier, there is no way for the pages to be
> > > migrated to the GPU without risking use-after-free or at least data
> > > corruption.  Instead, one of the following options will be needed:
> > > 
> > > 1. hipMallocManaged() allocates the memory from the backend using the
> > >     Map primitive discussed elsewhere.  Such memory is not mappable in
> > >     the IOMMU (if there is an assigned PCI device) and cannot be used for
> > >     grant table operations.  Memory allocated via system allocators
> > >     (anonymous pages) is not able to be migrated.
> > > 
> > > 2. The frontend uses shadow buffers for all I/O.  This allows the
> > >     backend to use a new Steal primitive to revoke the guest's accesses
> > >     to anonymous pages and handle page faults accordingly.
> > > 
> > > 3. Same as 2 except that the frontend allocates all memory (except
> > >     bounce buffers) from the backend, just like a KVM guest does, rather
> > >     than from the Xen toolstack.
> > > 
> > > 4. The frontend tries to migrate the pages to backend-provided ones, and
> > >     falls back to leaving them pinned on the CPU.  The frontend's MMU
> > >     notifier tells the backend to stop accessing the pages, blocking
> > >     until the backend confirms this.  The frontend then moves the pages
> > >     to its own memory and returns from the notifier.  This may require
> > >     new AMDKFD APIs.
> 
> This step needs new AMDKFD userspace APIs, KFD has some internel APIs for
> it, need exported into UMD. But actually the most challenging parts are
> adding hypervisior primitive, and adding the frontend <-> backend sync
> solution. KFD needs mmap_write_lock to handle migrate/update operations,
> this lock needs be removed or the sync between frontend and backend is hard
> to implement. It maybe needs refactor the KFD SVM, only for this task needs
> a lot of work, and there may be other work that needs to do I haven't
> discovered it yet.
> Before starting all of it, we may need a solid design or another
> clever/compromise solution that may reduce some of the workload. Your
> reminding are professional and detailed, many things I haven't noticed,
> really thanks a lot.

You're welcome.  I am glad to have helped.

I believe the best first step would be to implement Map and Revoke as
described in https://lore.kernel.org/xen-devel/Z6U7yOrMyLZWqPA4@itl-email/T/,
and to submit the code for upstream review.  These primitives should be
sufficient for graphics, and they are are also prerequisites for future
SVM work.  By submitting the code upstream without waiting for
virtio-GPU ROCm support, you will be able to get review of Xen-related
code to happen in parallel with the ROCm work.  This will shorten the
critical path for upstreaming, and will reduce this risk of spending
development resources on designs that would not be acceptable upstream.
Sending the current patches would be a good idea, even if they are known
to have bugs, as this will allow upstream to help you fix them.

Once Map and Revoke are implemented, you should be able to implement
full support for virtio-GPU native contexts in Xen+QEMU.  This unlocks
OpenGL and Vulkan, along with OpenCL via rusticl.  This work can happen
in parallel with the kernel and hypervisor patches being reviewed.  I
also expect supporting ROCm with no XNACK to be feasible here, though I
am not certain as I am not familiar with the details of the AMDKFD
driver.

Finally, XNACK support can be implemented.  This should be the last
step, as it depends on all of the other work and will be a signfiicant
amount of effort in its own right.  It also might not be necessary in
practice, as if one is willing to abandon fine-grained cache coherency
I believe one can achieve equivalent performance without it, at the
expense of additional effort from the programmer.  Due to the limited
hardware support for XNACK, I expect that programs intended to be run
on end-user devices (as opposed to servers) will not require it, and it
might not be possible to achieve reasonable performance with virtio-GPU.
Since my understanding is that AMD's use-cases for virtio-GPU are
primarily in the automotive sector, I am not sure that AMD will gain any
value from enabling it, unless either UDNA devices will support XNACK in
ROCm or automotive OEMs will be including Instinct accelerators in the
cars they make.
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)
Invisible Things Lab

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]