Re: [Nouveau] [PATCH] nouveau/hmm: fix migrate zero page to GPU

2020-05-21 Thread Ben Skeggs
On Thu, 21 May 2020 at 07:05, Ralph Campbell wrote: > > > On 5/20/20 12:20 PM, Jason Gunthorpe wrote: > > On Wed, May 20, 2020 at 11:36:52AM -0700, Ralph Campbell wrote: > >> When calling OpenCL clEnqueueSVMMigrateMem() on a region of memory that > >> is backed by pte_none() or zero pages, migrate

Re: [PATCH] nouveau/hmm: fix migrate zero page to GPU

2020-05-20 Thread Ralph Campbell
On 5/20/20 12:20 PM, Jason Gunthorpe wrote: On Wed, May 20, 2020 at 11:36:52AM -0700, Ralph Campbell wrote: When calling OpenCL clEnqueueSVMMigrateMem() on a region of memory that is backed by pte_none() or zero pages, migrate_vma_setup() will fill the source PFN array with an entry indicating

Re: [PATCH] nouveau/hmm: fix migrate zero page to GPU

2020-05-20 Thread Jason Gunthorpe
On Wed, May 20, 2020 at 11:36:52AM -0700, Ralph Campbell wrote: > When calling OpenCL clEnqueueSVMMigrateMem() on a region of memory that > is backed by pte_none() or zero pages, migrate_vma_setup() will fill the > source PFN array with an entry indicating the source page is zero. > Use this to opt

[PATCH] nouveau/hmm: fix migrate zero page to GPU

2020-05-20 Thread Ralph Campbell
When calling OpenCL clEnqueueSVMMigrateMem() on a region of memory that is backed by pte_none() or zero pages, migrate_vma_setup() will fill the source PFN array with an entry indicating the source page is zero. Use this to optimize migration to device private memory by allocating GPU memory and ze