drm/nouveau/nouveau: fix page fault on device private memory
authorRalph Campbell <rcampbell@nvidia.com>
Fri, 26 Jun 2020 21:03:37 +0000 (14:03 -0700)
committerBen Skeggs <bskeggs@redhat.com>
Wed, 8 Jul 2020 03:30:42 +0000 (13:30 +1000)
commited710a6ed797430026aa5116dd0ab22378798b69
treeb70cc42bd6032679c783ad7bad0bc6244a1d1ba3
parentad61f5f5e02e0aaf667efcd0c44ba5d70ea0b356
drm/nouveau/nouveau: fix page fault on device private memory

If system memory is migrated to device private memory and no GPU MMU
page table entry exists, the GPU will fault and call hmm_range_fault()
to get the PFN for the page. Since the .dev_private_owner pointer in
struct hmm_range is not set, hmm_range_fault returns an error which
results in the GPU program stopping with a fatal fault.
Fix this by setting .dev_private_owner appropriately.

Fixes: 08ddddda667b ("mm/hmm: check the device private page owner in hmm_range_fault()")
Cc: stable@vger.kernel.org
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@mellanox.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
drivers/gpu/drm/nouveau/nouveau_svm.c