Re: [RFC PATCH 0/8] TTM shrinker helpers and xe buffer object shrinker
Hi, Amaranath, On Tue, 2024-04-02 at 17:25 +0530, Somalapuram, Amaranath wrote: > > On 3/29/2024 8:26 PM, Thomas Hellström wrote: > > This series implements TTM shrinker / eviction helpers and an xe bo > > shrinker. It builds on two previous series. First > > > > https://www.mail-archive.com/dri-devel@lists.freedesktop.org/msg484425.html > > > > for patch 1-4, which IMO still could be reviewed and pushed as a > > separate series. > > > > Second the previous TTM shrinker series > On the latest drm-misc good amount of conflicts on both patch series, > able to re-base first one, Second on has 16 patches. > If you have latest re-base patches, Please share it. I was a bit unclear here. This series includes a rebased variant of the first series, and the rest of this series is a rework of the remaining patches in the second series below. So this series is self-contained and the links are provided only for reference and context. This series still seems to apply using git am -3 on drm-tip, and with a minor single conflict on drm-misc-next. (Typically we send the patches against drm-tip since that's what CI is using). The below series has not been rebased but I have a version of this series with a direct-swap-cache insertion backend lying around somewhere. I can post it or share it on gitlab if you want. Thanks, /Thomas > > Regards, > S.Amarnath > > https://lore.kernel.org/linux-mm/b7491378-defd-4f1c-31e2-29e4c77e2...@amd.com/T/ > > > > Where the comment about layering > > https://lore.kernel.org/linux-mm/b7491378-defd-4f1c-31e2-29e4c77e2...@amd.com/T/#ma918844aa8a6efe8768fdcda0c6590d5c93850c9 > > > > now addressed, and this version also implements shmem objects for > > backup > > rather than direct swap-cache insertions, which was used in the > > previuos > > series. It turns out that with per-page backup / shrinking, shmem > > objects > > appears to work just as well as direct swap-cache insertions with > > the > > added benefit that was introduced in the previous TTM shrinker > > series to > > avoid running out of swap entries isn't really needed. > > > > In any case, patch 1-4 are better described in their separate > > series. > > (RFC is removed for those). > > > > Patch 5 could in theory be skipped but introduces a possibility to > > easily > > add or test multiple backup backends, like the direct swap-cache > > insertion or even files into fast dedicated nvme storage for for > > example. > > > > Patch 6 introduces helpers in the ttm_pool code for page-by-page > > shrinking > > and recovery. It avoids having to temporarily allocate a huge > > amount of > > memory to be able to shrink a buffer object. It also introduces the > > possibility to immediately write-back pages if needed, since that > > tends > > to be a bit delayed when left to kswapd. > > > > Patch 7 introduces a LRU walk helper for eviction and shrinking. > > It's > > currently xe-only but not xe-specific and can easily be moved to > > TTM when > > used by more than one driver or when eviction is implemented using > > it. > > > > Patch 8 introduces a helper callback for shrinking (Also ready to > > be > > moved to TTM) and an xe-specific shrinker implementation. > > > > Testing: > > ATM I don't think we have good tests to cover the shrinking > > functionality > > The series has been tested with a hack that continously creates > > TTM_TT buffer objects until system memory and swap space is > > exhausted, > > and then reads them back and frees them. However, these tests > > seem to be very slow. > > Ideally a similar test on a machine with very fast solid state or > > similar storage should be set up. Ideally also verifying content > > preservation. > > > > Cc: Somalapuram Amaranath > > Cc: Christian König > > Cc: > > > > Thomas Hellström (8): > > drm/ttm: Allow TTM LRU list nodes of different types > > drm/ttm: Use LRU hitches > > drm/ttm, drm/amdgpu, drm/xe: Consider hitch moves within bulk > > sublist > > moves > > drm/ttm: Allow continued swapout after -ENOSPC falure > > drm/ttm: Add a virtual base class for graphics memory backup > > drm/ttm/pool: Provide a helper to shrink pages. > > drm/xe, drm/ttm: Provide a generic LRU walker helper > > drm/xe: Add a shrinker for xe bos > > > > drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 + > > drivers/gpu/drm/ttm/Makefile | 2 +- > > drivers/gpu/drm/ttm/ttm_backup_shmem.c | 137 + > > drivers/gpu/drm/ttm/ttm_bo.c | 1 + > > drivers/gpu/drm/ttm/ttm_device.c | 33 ++- > > drivers/gpu/drm/ttm/ttm_pool.c | 391 > > - > > drivers/gpu/drm/ttm/ttm_resource.c | 231 --- > > drivers/gpu/drm/ttm/ttm_tt.c | 34 +++ > > drivers/gpu/drm/xe/Makefile | 2 + > > drivers/gpu/drm/xe/xe_bo.c | 123 ++-- > > drivers/gpu/drm/xe/xe_bo.h | 3 + > > drivers/gpu/drm/xe/xe_device.c | 8 + > >
Re: [RFC PATCH 0/8] TTM shrinker helpers and xe buffer object shrinker
On 3/29/2024 8:26 PM, Thomas Hellström wrote: This series implements TTM shrinker / eviction helpers and an xe bo shrinker. It builds on two previous series. First https://www.mail-archive.com/dri-devel@lists.freedesktop.org/msg484425.html for patch 1-4, which IMO still could be reviewed and pushed as a separate series. Second the previous TTM shrinker series On the latest drm-misc good amount of conflicts on both patch series, able to re-base first one, Second on has 16 patches. If you have latest re-base patches, Please share it. Regards, S.Amarnath https://lore.kernel.org/linux-mm/b7491378-defd-4f1c-31e2-29e4c77e2...@amd.com/T/ Where the comment about layering https://lore.kernel.org/linux-mm/b7491378-defd-4f1c-31e2-29e4c77e2...@amd.com/T/#ma918844aa8a6efe8768fdcda0c6590d5c93850c9 now addressed, and this version also implements shmem objects for backup rather than direct swap-cache insertions, which was used in the previuos series. It turns out that with per-page backup / shrinking, shmem objects appears to work just as well as direct swap-cache insertions with the added benefit that was introduced in the previous TTM shrinker series to avoid running out of swap entries isn't really needed. In any case, patch 1-4 are better described in their separate series. (RFC is removed for those). Patch 5 could in theory be skipped but introduces a possibility to easily add or test multiple backup backends, like the direct swap-cache insertion or even files into fast dedicated nvme storage for for example. Patch 6 introduces helpers in the ttm_pool code for page-by-page shrinking and recovery. It avoids having to temporarily allocate a huge amount of memory to be able to shrink a buffer object. It also introduces the possibility to immediately write-back pages if needed, since that tends to be a bit delayed when left to kswapd. Patch 7 introduces a LRU walk helper for eviction and shrinking. It's currently xe-only but not xe-specific and can easily be moved to TTM when used by more than one driver or when eviction is implemented using it. Patch 8 introduces a helper callback for shrinking (Also ready to be moved to TTM) and an xe-specific shrinker implementation. Testing: ATM I don't think we have good tests to cover the shrinking functionality The series has been tested with a hack that continously creates TTM_TT buffer objects until system memory and swap space is exhausted, and then reads them back and frees them. However, these tests seem to be very slow. Ideally a similar test on a machine with very fast solid state or similar storage should be set up. Ideally also verifying content preservation. Cc: Somalapuram Amaranath Cc: Christian König Cc: Thomas Hellström (8): drm/ttm: Allow TTM LRU list nodes of different types drm/ttm: Use LRU hitches drm/ttm, drm/amdgpu, drm/xe: Consider hitch moves within bulk sublist moves drm/ttm: Allow continued swapout after -ENOSPC falure drm/ttm: Add a virtual base class for graphics memory backup drm/ttm/pool: Provide a helper to shrink pages. drm/xe, drm/ttm: Provide a generic LRU walker helper drm/xe: Add a shrinker for xe bos drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 + drivers/gpu/drm/ttm/Makefile | 2 +- drivers/gpu/drm/ttm/ttm_backup_shmem.c | 137 + drivers/gpu/drm/ttm/ttm_bo.c | 1 + drivers/gpu/drm/ttm/ttm_device.c | 33 ++- drivers/gpu/drm/ttm/ttm_pool.c | 391 - drivers/gpu/drm/ttm/ttm_resource.c | 231 --- drivers/gpu/drm/ttm/ttm_tt.c | 34 +++ drivers/gpu/drm/xe/Makefile| 2 + drivers/gpu/drm/xe/xe_bo.c | 123 ++-- drivers/gpu/drm/xe/xe_bo.h | 3 + drivers/gpu/drm/xe/xe_device.c | 8 + drivers/gpu/drm/xe/xe_device_types.h | 2 + drivers/gpu/drm/xe/xe_shrinker.c | 237 +++ drivers/gpu/drm/xe/xe_shrinker.h | 18 ++ drivers/gpu/drm/xe/xe_ttm_helpers.c| 224 ++ drivers/gpu/drm/xe/xe_ttm_helpers.h| 63 drivers/gpu/drm/xe/xe_vm.c | 4 + include/drm/ttm/ttm_backup.h | 136 + include/drm/ttm/ttm_device.h | 2 + include/drm/ttm/ttm_pool.h | 4 + include/drm/ttm/ttm_resource.h | 96 +- include/drm/ttm/ttm_tt.h | 19 ++ 23 files changed, 1683 insertions(+), 91 deletions(-) create mode 100644 drivers/gpu/drm/ttm/ttm_backup_shmem.c create mode 100644 drivers/gpu/drm/xe/xe_shrinker.c create mode 100644 drivers/gpu/drm/xe/xe_shrinker.h create mode 100644 drivers/gpu/drm/xe/xe_ttm_helpers.c create mode 100644 drivers/gpu/drm/xe/xe_ttm_helpers.h create mode 100644 include/drm/ttm/ttm_backup.h
[RFC PATCH 0/8] TTM shrinker helpers and xe buffer object shrinker
This series implements TTM shrinker / eviction helpers and an xe bo shrinker. It builds on two previous series. First https://www.mail-archive.com/dri-devel@lists.freedesktop.org/msg484425.html for patch 1-4, which IMO still could be reviewed and pushed as a separate series. Second the previous TTM shrinker series https://lore.kernel.org/linux-mm/b7491378-defd-4f1c-31e2-29e4c77e2...@amd.com/T/ Where the comment about layering https://lore.kernel.org/linux-mm/b7491378-defd-4f1c-31e2-29e4c77e2...@amd.com/T/#ma918844aa8a6efe8768fdcda0c6590d5c93850c9 now addressed, and this version also implements shmem objects for backup rather than direct swap-cache insertions, which was used in the previuos series. It turns out that with per-page backup / shrinking, shmem objects appears to work just as well as direct swap-cache insertions with the added benefit that was introduced in the previous TTM shrinker series to avoid running out of swap entries isn't really needed. In any case, patch 1-4 are better described in their separate series. (RFC is removed for those). Patch 5 could in theory be skipped but introduces a possibility to easily add or test multiple backup backends, like the direct swap-cache insertion or even files into fast dedicated nvme storage for for example. Patch 6 introduces helpers in the ttm_pool code for page-by-page shrinking and recovery. It avoids having to temporarily allocate a huge amount of memory to be able to shrink a buffer object. It also introduces the possibility to immediately write-back pages if needed, since that tends to be a bit delayed when left to kswapd. Patch 7 introduces a LRU walk helper for eviction and shrinking. It's currently xe-only but not xe-specific and can easily be moved to TTM when used by more than one driver or when eviction is implemented using it. Patch 8 introduces a helper callback for shrinking (Also ready to be moved to TTM) and an xe-specific shrinker implementation. Testing: ATM I don't think we have good tests to cover the shrinking functionality The series has been tested with a hack that continously creates TTM_TT buffer objects until system memory and swap space is exhausted, and then reads them back and frees them. However, these tests seem to be very slow. Ideally a similar test on a machine with very fast solid state or similar storage should be set up. Ideally also verifying content preservation. Cc: Somalapuram Amaranath Cc: Christian König Cc: Thomas Hellström (8): drm/ttm: Allow TTM LRU list nodes of different types drm/ttm: Use LRU hitches drm/ttm, drm/amdgpu, drm/xe: Consider hitch moves within bulk sublist moves drm/ttm: Allow continued swapout after -ENOSPC falure drm/ttm: Add a virtual base class for graphics memory backup drm/ttm/pool: Provide a helper to shrink pages. drm/xe, drm/ttm: Provide a generic LRU walker helper drm/xe: Add a shrinker for xe bos drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 + drivers/gpu/drm/ttm/Makefile | 2 +- drivers/gpu/drm/ttm/ttm_backup_shmem.c | 137 + drivers/gpu/drm/ttm/ttm_bo.c | 1 + drivers/gpu/drm/ttm/ttm_device.c | 33 ++- drivers/gpu/drm/ttm/ttm_pool.c | 391 - drivers/gpu/drm/ttm/ttm_resource.c | 231 --- drivers/gpu/drm/ttm/ttm_tt.c | 34 +++ drivers/gpu/drm/xe/Makefile| 2 + drivers/gpu/drm/xe/xe_bo.c | 123 ++-- drivers/gpu/drm/xe/xe_bo.h | 3 + drivers/gpu/drm/xe/xe_device.c | 8 + drivers/gpu/drm/xe/xe_device_types.h | 2 + drivers/gpu/drm/xe/xe_shrinker.c | 237 +++ drivers/gpu/drm/xe/xe_shrinker.h | 18 ++ drivers/gpu/drm/xe/xe_ttm_helpers.c| 224 ++ drivers/gpu/drm/xe/xe_ttm_helpers.h| 63 drivers/gpu/drm/xe/xe_vm.c | 4 + include/drm/ttm/ttm_backup.h | 136 + include/drm/ttm/ttm_device.h | 2 + include/drm/ttm/ttm_pool.h | 4 + include/drm/ttm/ttm_resource.h | 96 +- include/drm/ttm/ttm_tt.h | 19 ++ 23 files changed, 1683 insertions(+), 91 deletions(-) create mode 100644 drivers/gpu/drm/ttm/ttm_backup_shmem.c create mode 100644 drivers/gpu/drm/xe/xe_shrinker.c create mode 100644 drivers/gpu/drm/xe/xe_shrinker.h create mode 100644 drivers/gpu/drm/xe/xe_ttm_helpers.c create mode 100644 drivers/gpu/drm/xe/xe_ttm_helpers.h create mode 100644 include/drm/ttm/ttm_backup.h -- 2.44.0