search for: migrate_vma_pag

When migrating the special zero page, migrate_vma_pages() calls mmu_notifier_invalidate_range_start() before replacing the zero page PFN in the CPU page tables. This is unnecessary since the range was invalidated in migrate_vma_setup() and the page table entry is checked to be sure it hasn't changed between migrate_vma_setup() and migrate_vma_page...

On Thu, Jul 23, 2020 at 03:30:04PM -0700, Ralph Campbell wrote: > When migrating the special zero page, migrate_vma_pages() calls > mmu_notifier_invalidate_range_start() before replacing the zero page > PFN in the CPU page tables. This is unnecessary since the range was > invalidated in migrate_vma_setup() and the page table entry is checked > to be sure it hasn't changed between migrate_vma_setup()...

On 7/28/20 12:19 PM, Jason Gunthorpe wrote: > On Thu, Jul 23, 2020 at 03:30:04PM -0700, Ralph Campbell wrote: >> When migrating the special zero page, migrate_vma_pages() calls >> mmu_notifier_invalidate_range_start() before replacing the zero page >> PFN in the CPU page tables. This is unnecessary since the range was >> invalidated in migrate_vma_setup() and the page table entry is checked >> to be sure it hasn't changed between migr...

On Tue, Jul 28, 2020 at 03:04:07PM -0700, Ralph Campbell wrote: > > On 7/28/20 12:19 PM, Jason Gunthorpe wrote: > > On Thu, Jul 23, 2020 at 03:30:04PM -0700, Ralph Campbell wrote: > > > When migrating the special zero page, migrate_vma_pages() calls > > > mmu_notifier_invalidate_range_start() before replacing the zero page > > > PFN in the CPU page tables. This is unnecessary since the range was > > > invalidated in migrate_vma_setup() and the page table entry is checked > > > to be sure it hasn...

...igrated. Note those pages were still copied to a new page and thus we wasted -bandwidth but this is considered as a rare event and a price that we are -willing to pay to keep all the code simpler. +engine to perform copy from and to device memory. For this we need a new to +use migrate_vma_setup(), migrate_vma_pages(), and migrate_vma_finalize() +helpers. Memory cgroup (memcg) and rss accounting diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index 345c63cb752a..38416798abd4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/no...

...igrated. Note those pages were still copied to a new page and thus we wasted -bandwidth but this is considered as a rare event and a price that we are -willing to pay to keep all the code simpler. +engine to perform copy from and to device memory. For this we need a new to +use migrate_vma_setup(), migrate_vma_pages(), and migrate_vma_finalize() +helpers. Memory cgroup (memcg) and rss accounting diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index 345c63cb752a..38416798abd4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nou...

...pages were still copied to a new page and thus we wasted > -bandwidth but this is considered as a rare event and a price that we are > -willing to pay to keep all the code simpler. > +engine to perform copy from and to device memory. For this we need a new to > +use migrate_vma_setup(), migrate_vma_pages(), and migrate_vma_finalize() > +helpers. > > > Memory cgroup (memcg) and rss accounting > diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c > index 345c63cb752a..38416798abd4 100644 > --- a/drivers/gpu/drm/nouveau/nouveau_dm...

Hi Jérôme, Ben and Jason, below is a series against the hmm tree which starts revamping the migrate_vma functionality. The prime idea is to export three slightly lower level functions and thus avoid the need for migrate_vma_ops callbacks. Diffstat: 7 files changed, 282 insertions(+), 614 deletions(-) A git tree is also available at: git://git.infradead.org/users/hch/misc.git

...eady been migrated. The approach taken is to introduce a new mmu notifier invalidation event type and use that in the device driver to skip invalidation callbacks from migrate_vma_setup(). The device driver is also then expected to handle device MMU invalidations as part of the migrate_vma_setup(), migrate_vma_pages(), migrate_vma_finalize() process. Note that this is opt-in. A device driver can simply invalidate its MMU in the mmu notifier callback and not handle MMU invalidations in the migration sequence. This series is based on Jason Gunthorpe's HMM tree (linux-5.8.0-rc4). Also, this replaces the n...

Hi Jérôme, Ben and Jason, below is a series against the hmm tree which starts revamping the migrate_vma functionality. The prime idea is to export three slightly lower level functions and thus avoid the need for migrate_vma_ops callbacks. Diffstat: 4 files changed, 285 insertions(+), 602 deletions(-) A git tree is also available at: git://git.infradead.org/users/hch/misc.git

Hi Jérôme, Ben and Jason, below is a series against the hmm tree which starts revamping the migrate_vma functionality. The prime idea is to export three slightly lower level functions and thus avoid the need for migrate_vma_ops callbacks. Diffstat: 5 files changed, 281 insertions(+), 607 deletions(-) A git tree is also available at: git://git.infradead.org/users/hch/misc.git

...umentation/vm/hmm.rst @@ -339,9 +339,8 @@ Migration to and from device memory =================================== Because the CPU cannot access device memory, migration must use the device DMA -engine to perform copy from and to device memory. For this we need a new to -use migrate_vma_setup(), migrate_vma_pages(), and migrate_vma_finalize() -helpers. +engine to perform copy from and to device memory. For this we need to use +migrate_vma_setup(), migrate_vma_pages(), and migrate_vma_finalize() helpers. Memory cgroup (memcg) and rss accounting diff --git a/mm/migrate.c b/mm/migrate.c index 993386cb53...

...rivers/gpu/drm/nouveau/nouveau_dmem.c @@ -678,7 +678,7 @@ static void nouveau_dmem_migrate_chunk(struct nouveau_drm *drm, } if (!nouveau_fence_new(&fence)) - nouveau_fence_emit(fence, chunk->drm->dmem->migrate.chan); + nouveau_fence_emit(fence, drm->dmem->migrate.chan); migrate_vma_pages(args); nouveau_dmem_fence_done(&fence); nouveau_pfns_map(svmm, args->vma->vm_mm, args->start, pfns, i); base-commit: 82d750e9d2f5d0594c8f7057ce59127e701af781 -- 2.41.0

...+ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -209,8 +209,7 @@ static vm_fault_t nouveau_dmem_migrate_to_ram(struct vm_fault *vmf) goto done; } - if (!nouveau_fence_new(&fence)) - nouveau_fence_emit(fence, dmem->migrate.chan); + nouveau_fence_new(&fence, dmem->migrate.chan); migrate_vma_pages(&args); nouveau_dmem_fence_done(&fence); dma_unmap_page(drm->dev->dev, dma_addr, PAGE_SIZE, DMA_BIDIRECTIONAL); @@ -403,8 +402,7 @@ nouveau_dmem_evict_chunk(struct nouveau_dmem_chunk *chunk) } } - if (!nouveau_fence_new(&fence)) - nouveau_fence_emit(fence, chunk->...

.../* * FIXME what we really want is to find some heuristic to migrate more @@ -281,16 +201,17 @@ static vm_fault_t nouveau_dmem_migrate_to_ram(struct vm_fault *vmf) if (!args.cpages) return 0; - nouveau_dmem_fault_alloc_and_copy(args.vma, src, dst, args.start, - args.end, &fault); - migrate_vma_pages(&args); - nouveau_dmem_fault_finalize_and_map(&fault); + ret = nouveau_dmem_fault_copy_one(drm, vmf, &args, &dma_addr); + if (ret || dst == 0) + goto done; + nouveau_fence_new(dmem->migrate.chan, false, &fence); + migrate_vma_pages(&args); + nouveau_dmem_fence_done(&...

.../* * FIXME what we really want is to find some heuristic to migrate more @@ -281,16 +199,18 @@ static vm_fault_t nouveau_dmem_migrate_to_ram(struct vm_fault *vmf) if (!args.cpages) return 0; - nouveau_dmem_fault_alloc_and_copy(args.vma, src, dst, args.start, - args.end, &fault); - migrate_vma_pages(&args); - nouveau_dmem_fault_finalize_and_map(&fault); + ret = nouveau_dmem_fault_copy_one(drm, vmf->vma, vmf->address, src, + &dst, &dma_addr); + if (ret || dst == 0) + goto done; + nouveau_fence_new(dmem->migrate.chan, false, &fence); + migrate_vma_pages(&a...

...to find some heuristic to migrate more > @@ -281,16 +201,17 @@ static vm_fault_t nouveau_dmem_migrate_to_ram(struct vm_fault *vmf) > if (!args.cpages) > return 0; > > - nouveau_dmem_fault_alloc_and_copy(args.vma, src, dst, args.start, > - args.end, &fault); > - migrate_vma_pages(&args); > - nouveau_dmem_fault_finalize_and_map(&fault); > + ret = nouveau_dmem_fault_copy_one(drm, vmf, &args, &dma_addr); > + if (ret || dst == 0) > + goto done; > > + nouveau_fence_new(dmem->migrate.chan, false, &fence); > + migrate_vma_pages(&am...

...st[i] = nouveau_dmem_migrate_copy_one(drm, args->vma, + addr, args->src[i], &dma_addrs[nr_dma]); + if (args->dst[i]) + nr_dma++; + addr += PAGE_SIZE; + } - nouveau_dmem_fence_done(&migrate->fence); + nouveau_fence_new(drm->dmem->migrate.chan, false, &fence); + migrate_vma_pages(args); + nouveau_dmem_fence_done(&fence); - while (migrate->dma_nr--) { - dma_unmap_page(drm->dev->dev, migrate->dma[migrate->dma_nr], - PAGE_SIZE, PCI_DMA_BIDIRECTIONAL); + while (nr_dma--) { + dma_unmap_page(drm->dev->dev, dma_addrs[nr_dma], PAGE_SIZE, +...

...eady been migrated. The approach taken is to introduce a new mmu notifier invalidation event type and use that in the device driver to skip invalidation callbacks from migrate_vma_setup(). The device driver is also then expected to handle device MMU invalidations as part of the migrate_vma_setup(), migrate_vma_pages(), migrate_vma_finalize() process. Note that this is opt-in. A device driver can simply invalidate its MMU in the mmu notifier callback and not handle MMU invalidations in the migration sequence. This series is based on Jason Gunthorpe's HMM tree (linux-5.8.0-rc4). Also, this replaces the n...

...The approach taken is to introduce a new mmu notifier > invalidation event type and use that in the device driver to skip > invalidation callbacks from migrate_vma_setup(). The device driver is > also then expected to handle device MMU invalidations as part of the > migrate_vma_setup(), migrate_vma_pages(), migrate_vma_finalize() process. > Note that this is opt-in. A device driver can simply invalidate its MMU > in the mmu notifier callback and not handle MMU invalidations in the > migration sequence. > > This series is based on Jason Gunthorpe's HMM tree (linux-5.8.0-rc4). &...