search for: nouveau_dmem_chunk

In nouveau_dmem_init(), a number of struct nouveau_dmem_chunk are allocated and put on the dmem->chunk_empty list. Then in nouveau_dmem_pages_alloc(), a nouveau_dmem_chunk is removed from the list and GPU memory is allocated. However, the nouveau_dmem_chunk is never removed from the chunk_empty list nor placed on the chunk_free or chunk_full lists. This re...

On Tue, Apr 21, 2020 at 04:11:07PM -0700, Ralph Campbell wrote: > In nouveau_dmem_init(), a number of struct nouveau_dmem_chunk are allocated > and put on the dmem->chunk_empty list. Then in nouveau_dmem_pages_alloc(), > a nouveau_dmem_chunk is removed from the list and GPU memory is allocated. > However, the nouveau_dmem_chunk is never removed from the chunk_empty > list nor placed on the chunk_free or chunk...

...au/nouveau_dmem.c index e696157f771e..d469bc334438 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -102,6 +102,14 @@ struct nouveau_migrate { unsigned long dma_nr; }; +static unsigned long nouveau_dmem_page_addr(struct page *page) +{ + struct nouveau_dmem_chunk *chunk = page->zone_device_data; + unsigned long idx = page_to_pfn(page) - chunk->pfn_first; + + return (idx << PAGE_SHIFT) + chunk->bo->bo.offset; +} + static void nouveau_dmem_page_free(struct page *page) { struct nouveau_dmem_chunk *chunk = page->zone_device_data; @@ -16...

...rtions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index 12feecf71e7..6fb65b01d77 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -378,9 +378,9 @@ nouveau_dmem_evict_chunk(struct nouveau_dmem_chunk *chunk) dma_addr_t *dma_addrs; struct nouveau_fence *fence; - src_pfns = kcalloc(npages, sizeof(*src_pfns), GFP_KERNEL); - dst_pfns = kcalloc(npages, sizeof(*dst_pfns), GFP_KERNEL); - dma_addrs = kcalloc(npages, sizeof(*dma_addrs), GFP_KERNEL); + src_pfns = kvcalloc(npages, sizeof(*src_pfns),...

...s two patches I posted earlier. The first is queued in Ben Skegg's nouveau tree and the second is still pending review/not queued. [1] ("nouveau/hmm: map pages after migration") https://lore.kernel.org/linux-mm/20200304001339.8248-5-rcampbell at nvidia.com/ [2] ("nouveau/hmm: fix nouveau_dmem_chunk allocations") https://lore.kernel.org/lkml/20200421231107.30958-1-rcampbell at nvidia.com/ drivers/gpu/drm/nouveau/nouveau_dmem.c | 75 ++++++++++++++++++++++---- 1 file changed, 66 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/n...

...au/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index aa9fec80492d..900a302b7ce9 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -100,12 +100,10 @@ static void nouveau_dmem_free(struct hmm_devmem *devmem, struct page *page) { struct nouveau_dmem_chunk *chunk; - struct nouveau_drm *drm; unsigned long idx; chunk = (void *)hmm_devmem_page_get_drvdata(page); idx = page_to_pfn(page) - chunk->pfn_first; - drm = chunk->drm; /* * FIXME:

...> diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c > index 12feecf71e7..6fb65b01d77 100644 > --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c > +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c > @@ -378,9 +378,9 @@ nouveau_dmem_evict_chunk(struct nouveau_dmem_chunk *chunk) > dma_addr_t *dma_addrs; > struct nouveau_fence *fence; > > - src_pfns = kcalloc(npages, sizeof(*src_pfns), GFP_KERNEL); > - dst_pfns = kcalloc(npages, sizeof(*dst_pfns), GFP_KERNEL); > - dma_addrs = kcalloc(npages, sizeof(*dma_addrs), GFP_KERNEL); > + src_pfns...

...letions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index 12feecf71e7..9a578262c6d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -374,13 +374,13 @@ static void nouveau_dmem_evict_chunk(struct nouveau_dmem_chunk *chunk) { unsigned long i, npages = range_len(&chunk->pagemap.range) >> PAGE_SHIFT; - unsigned long *src_pfns, *dst_pfns; - dma_addr_t *dma_addrs; + unsigned long src_pfns[npages], dst_pfns[npages]; + dma_addr_t dma_addrs[npages]; struct nouveau_fence *fence; - src_pfns = kcallo...

...first is queued in Ben Skegg's nouveau >> tree and the second is still pending review/not queued. >> [1] ("nouveau/hmm: map pages after migration") >> https://lore.kernel.org/linux-mm/20200304001339.8248-5-rcampbell at nvidia.com/ >> [2] ("nouveau/hmm: fix nouveau_dmem_chunk allocations") >> https://lore.kernel.org/lkml/20200421231107.30958-1-rcampbell at nvidia.com/ > > It would be best if it goes through Ben's tree if it doesn't have > conflicts with the hunks I have in the hmm tree.. Is it the case? > > Jason I think there might...

...e the new flag. >> >> Note that this series depends on a patch queued in Ben Skeggs' nouveau >> tree ("nouveau/hmm: map pages after migration") and the patches queued >> in Jason's HMM tree. >> There is also a patch outstanding ("nouveau/hmm: fix nouveau_dmem_chunk >> allocations") that is independent of the above and could be applied >> before or after. > > Did Christoph and Matt's remarks get addressed here? Both questioned the need to add the HMM_PFN_COMPOUND flag to the hmm_range_fault() output array saying that the PFN can be...

...ed earlier. The first is queued in Ben Skegg's nouveau > tree and the second is still pending review/not queued. > [1] ("nouveau/hmm: map pages after migration") > https://lore.kernel.org/linux-mm/20200304001339.8248-5-rcampbell at nvidia.com/ > [2] ("nouveau/hmm: fix nouveau_dmem_chunk allocations") > https://lore.kernel.org/lkml/20200421231107.30958-1-rcampbell at nvidia.com/ It would be best if it goes through Ben's tree if it doesn't have conflicts with the hunks I have in the hmm tree.. Is it the case? Jason

...s are updated to use the new flag. > > Note that this series depends on a patch queued in Ben Skeggs' nouveau > tree ("nouveau/hmm: map pages after migration") and the patches queued > in Jason's HMM tree. > There is also a patch outstanding ("nouveau/hmm: fix nouveau_dmem_chunk > allocations") that is independent of the above and could be applied > before or after. Did Christoph and Matt's remarks get addressed here? I think ODP could use something like this, currently it checks every page to get back to the huge page size and this flag would optimze that...

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

From: Yonatan Maman <Ymaman at Nvidia.com> This patch series aims to enable Peer-to-Peer (P2P) DMA access in GPU-centric applications that utilize RDMA and private device pages. This enhancement is crucial for minimizing data transfer overhead by allowing the GPU to directly expose device private page data to devices such as NICs, eliminating the need to traverse system RAM, which is the

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

...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->drm->dmem->migrate.chan); + nouveau_fence_new(&fence, chunk->drm->dmem->migrate.chan); migrate_device_pages(src_pfns, dst_pfns, npages); nouveau_dmem_fence_done(&fence); migr...

...struct nouveau_dmem, pagemap); } -struct nouveau_migrate { - struct vm_area_struct *vma; - struct nouveau_drm *drm; - struct nouveau_fence *fence; - unsigned long npages; - dma_addr_t *dma; - unsigned long dma_nr; -}; - static unsigned long nouveau_dmem_page_addr(struct page *page) { struct nouveau_dmem_chunk *chunk = page->zone_device_data; @@ -569,131 +558,67 @@ nouveau_dmem_init(struct nouveau_drm *drm) drm->dmem = NULL; } -static void -nouveau_dmem_migrate_alloc_and_copy(struct vm_area_struct *vma, - const unsigned long *src_pfns, - unsigned long *dst_pfns, - unsigned...

From: Yonatan Maman <Ymaman at Nvidia.com> Based on: Provide a new two step DMA mapping API patchset https://lore.kernel.org/kvm/20241114170247.GA5813 at lst.de/T/#t This patch series aims to enable Peer-to-Peer (P2P) DMA access in GPU-centric applications that utilize RDMA and private device pages. This enhancement reduces data transfer overhead by allowing the GPU to directly expose

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

...Nouveau and the HMM tests are updated to use the new flag. Note that this series depends on a patch queued in Ben Skeggs' nouveau tree ("nouveau/hmm: map pages after migration") and the patches queued in Jason's HMM tree. There is also a patch outstanding ("nouveau/hmm: fix nouveau_dmem_chunk allocations") that is independent of the above and could be applied before or after. Ralph Campbell (6): nouveau/hmm: map pages after migration nouveau: make nvkm_vmm_ctor() and nvkm_mmu_ptp_get() static nouveau/hmm: fault one page at a time mm/hmm: add output flag for compound page...