[ANNOUNCE] PUCK Agenda - 2024.01.17 - TDP MMU for IOMMU

[ANNOUNCE] PUCK Agenda - 2024.01.17 - TDP MMU for IOMMU

[Sean Christopherson]

Tomorrow's PUCK topic is utilizing KVM's TDP MMU for IOMMU page tables.

FYI, I am currently without my normal internet (hooray tethering), and we're
supposed to get a healthy dose of freezing rain tonight, i.e. I might lose power
too.  I expect to be able to join even if that happens, but I apologize in
advance if I end up being a no-show.

https://lore.kernel.org/all/20231202091211.13376-1-yan.y.zhao@intel.com

Time:     6am PDT
Video:    https://meet.google.com/vdb-aeqo-knk
Phone:    https://tel.meet/vdb-aeqo-knk?pin=3003112178656

Calendar: https://calendar.google.com/calendar/u/0?cid=Y182MWE1YjFmNjQ0NzM5YmY1YmVkN2U1ZWE1ZmMzNjY5Y2UzMmEyNTQ0YzVkYjFjN2M4OTE3MDJjYTUwOTBjN2Q1QGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20
Drive:    https://drive.google.com/drive/folders/1aTqCrvTsQI9T4qLhhLs_l986SngGlhPH?resourcekey=0-FDy0ykM3RerZedI8R-zj4A&usp=drive_link

Future Schedule:
January 24th - Memtypes for non-coherent DMA
January 31st - Available!
February     - Available!

Re: [ANNOUNCE] PUCK Agenda - 2024.01.17 - TDP MMU for IOMMU

[Michael Roth]

On Tue, Jan 16, 2024 at 05:06:44PM -0800, Sean Christopherson wrote:
> Tomorrow's PUCK topic is utilizing KVM's TDP MMU for IOMMU page tables.
> 
> FYI, I am currently without my normal internet (hooray tethering), and we're
> supposed to get a healthy dose of freezing rain tonight, i.e. I might lose power
> too.  I expect to be able to join even if that happens, but I apologize in
> advance if I end up being a no-show.
> 
> https://lore.kernel.org/all/20231202091211.13376-1-yan.y.zhao@intel.com
> 
> Time:     6am PDT
> Video:    https://meet.google.com/vdb-aeqo-knk
> Phone:    https://tel.meet/vdb-aeqo-knk?pin=3003112178656
> 
> Calendar: https://calendar.google.com/calendar/u/0?cid=Y182MWE1YjFmNjQ0NzM5YmY1YmVkN2U1ZWE1ZmMzNjY5Y2UzMmEyNTQ0YzVkYjFjN2M4OTE3MDJjYTUwOTBjN2Q1QGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20
> Drive:    https://drive.google.com/drive/folders/1aTqCrvTsQI9T4qLhhLs_l986SngGlhPH?resourcekey=0-FDy0ykM3RerZedI8R-zj4A&usp=drive_link
> 
> Future Schedule:
> January 24th - Memtypes for non-coherent DMA
> January 31st - Available!

Hi Sean,

I'd like to propose the following topic for the next available slot:

  "Finalizing internal guest_memfd APIs needed for SNP (TDX?) upstreaming"

There's 2 existing interfaces, gmem_prepare, gmem_invalidate, that are
needed by the current SNP patches, and there's some additional background
about the design decisions here:

  https://lore.kernel.org/kvm/20231016115028.996656-1-michael.roth@amd.com/

There's also another gmem interface that you recently proposed for handling
setting up the initial launch image of SNP guests here that seems like it
would have a lot of potential overlap with how gmem_prepare is implemented:

  https://lore.kernel.org/lkml/ZZ67oJwzAsSvui5U@google.com/

I'd like to try to get some clarity on what these should look like in order
to be considered acceptable for upstreaming of SNP, and potentially any
considerations that need to be taken into account for other users like
TDX/pKVM/etc.

Thanks,

Mike

> February     - Available!
> 

Re: [ANNOUNCE] PUCK Agenda - 2024.01.17 - TDP MMU for IOMMU

[Sean Christopherson]

On Mon, Jan 22, 2024, Michael Roth wrote:
> On Tue, Jan 16, 2024 at 05:06:44PM -0800, Sean Christopherson wrote:
> > Tomorrow's PUCK topic is utilizing KVM's TDP MMU for IOMMU page tables.
> > 
> > FYI, I am currently without my normal internet (hooray tethering), and we're
> > supposed to get a healthy dose of freezing rain tonight, i.e. I might lose power
> > too.  I expect to be able to join even if that happens, but I apologize in
> > advance if I end up being a no-show.
> > 
> > https://lore.kernel.org/all/20231202091211.13376-1-yan.y.zhao@intel.com
> > 
> > Time:     6am PDT
> > Video:    https://meet.google.com/vdb-aeqo-knk
> > Phone:    https://tel.meet/vdb-aeqo-knk?pin=3003112178656
> > 
> > Calendar: https://calendar.google.com/calendar/u/0?cid=Y182MWE1YjFmNjQ0NzM5YmY1YmVkN2U1ZWE1ZmMzNjY5Y2UzMmEyNTQ0YzVkYjFjN2M4OTE3MDJjYTUwOTBjN2Q1QGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20
> > Drive:    https://drive.google.com/drive/folders/1aTqCrvTsQI9T4qLhhLs_l986SngGlhPH?resourcekey=0-FDy0ykM3RerZedI8R-zj4A&usp=drive_link
> > 
> > Future Schedule:
> > January 24th - Memtypes for non-coherent DMA
> > January 31st - Available!
> 
> Hi Sean,
> 
> I'd like to propose the following topic for the next available slot:
> 
>   "Finalizing internal guest_memfd APIs needed for SNP (TDX?) upstreaming"
> 
> There's 2 existing interfaces, gmem_prepare, gmem_invalidate, that are
> needed by the current SNP patches, and there's some additional background
> about the design decisions here:
> 
>   https://lore.kernel.org/kvm/20231016115028.996656-1-michael.roth@amd.com/
> 
> There's also another gmem interface that you recently proposed for handling
> setting up the initial launch image of SNP guests here that seems like it
> would have a lot of potential overlap with how gmem_prepare is implemented:
> 
>   https://lore.kernel.org/lkml/ZZ67oJwzAsSvui5U@google.com/
> 
> I'd like to try to get some clarity on what these should look like in order
> to be considered acceptable for upstreaming of SNP, and potentially any
> considerations that need to be taken into account for other users like
> TDX/pKVM/etc.

I penciled this in for the 31st, let me know if that works for you.

Re: [ANNOUNCE] PUCK Agenda - 2024.01.17 - TDP MMU for IOMMU

[Michael Roth]

On Wed, Jan 24, 2024 at 09:30:35AM -0800, Sean Christopherson wrote:
> On Mon, Jan 22, 2024, Michael Roth wrote:
> > On Tue, Jan 16, 2024 at 05:06:44PM -0800, Sean Christopherson wrote:
> > > Tomorrow's PUCK topic is utilizing KVM's TDP MMU for IOMMU page tables.
> > > 
> > > FYI, I am currently without my normal internet (hooray tethering), and we're
> > > supposed to get a healthy dose of freezing rain tonight, i.e. I might lose power
> > > too.  I expect to be able to join even if that happens, but I apologize in
> > > advance if I end up being a no-show.
> > > 
> > > https://lore.kernel.org/all/20231202091211.13376-1-yan.y.zhao@intel.com
> > > 
> > > Time:     6am PDT
> > > Video:    https://meet.google.com/vdb-aeqo-knk
> > > Phone:    https://tel.meet/vdb-aeqo-knk?pin=3003112178656
> > > 
> > > Calendar: https://calendar.google.com/calendar/u/0?cid=Y182MWE1YjFmNjQ0NzM5YmY1YmVkN2U1ZWE1ZmMzNjY5Y2UzMmEyNTQ0YzVkYjFjN2M4OTE3MDJjYTUwOTBjN2Q1QGdyb3VwLmNhbGVuZGFyLmdvb2dsZS5jb20
> > > Drive:    https://drive.google.com/drive/folders/1aTqCrvTsQI9T4qLhhLs_l986SngGlhPH?resourcekey=0-FDy0ykM3RerZedI8R-zj4A&usp=drive_link
> > > 
> > > Future Schedule:
> > > January 24th - Memtypes for non-coherent DMA
> > > January 31st - Available!
> > 
> > Hi Sean,
> > 
> > I'd like to propose the following topic for the next available slot:
> > 
> >   "Finalizing internal guest_memfd APIs needed for SNP (TDX?) upstreaming"
> > 
> > There's 2 existing interfaces, gmem_prepare, gmem_invalidate, that are
> > needed by the current SNP patches, and there's some additional background
> > about the design decisions here:
> > 
> >   https://lore.kernel.org/kvm/20231016115028.996656-1-michael.roth@amd.com/
> > 
> > There's also another gmem interface that you recently proposed for handling
> > setting up the initial launch image of SNP guests here that seems like it
> > would have a lot of potential overlap with how gmem_prepare is implemented:
> > 
> >   https://lore.kernel.org/lkml/ZZ67oJwzAsSvui5U@google.com/
> > 
> > I'd like to try to get some clarity on what these should look like in order
> > to be considered acceptable for upstreaming of SNP, and potentially any
> > considerations that need to be taken into account for other users like
> > TDX/pKVM/etc.
> 
> I penciled this in for the 31st, let me know if that works for you.

That would be perfect. Thanks!

-Mike