On Mon, 30 Oct 2023 02:01:54 +0300
Dmitry Osipenko <dmitry.osipenko@xxxxxxxxxxxxx> wrote:
> To simplify the drm-shmem refcnt handling, we're moving away from
> the implicit get_pages() that is used by get_pages_sgt(). From now on
> drivers will have to pin pages while they use sgt. Panfrost's shrinker
> doesn't support swapping out BOs, hence pages are pinned and sgt is valid
> as long as pages' use-count > 0.
>
> Signed-off-by: Dmitry Osipenko <dmitry.osipenko@xxxxxxxxxxxxx>
> ---
> drivers/gpu/drm/panfrost/panfrost_gem.c | 17 +++++++++++++++++
> drivers/gpu/drm/panfrost/panfrost_mmu.c | 6 ++----
> 2 files changed, 19 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/panfrost/panfrost_gem.c b/drivers/gpu/drm/panfrost/panfrost_gem.c
> index 6b77d8cebcb2..bb9d43cf7c3c 100644
> --- a/drivers/gpu/drm/panfrost/panfrost_gem.c
> +++ b/drivers/gpu/drm/panfrost/panfrost_gem.c
> @@ -47,8 +47,13 @@ static void panfrost_gem_free_object(struct drm_gem_object *obj)
> }
> }
> kvfree(bo->sgts);
> +
> + drm_gem_shmem_put_pages(&bo->base);
> }
>
> + if (!bo->is_heap && !obj->import_attach)
> + drm_gem_shmem_put_pages(&bo->base);
> +
> drm_gem_shmem_free(&bo->base);
> }
>
> @@ -269,6 +274,7 @@ panfrost_gem_create(struct drm_device *dev, size_t size, u32 flags)
> {
> struct drm_gem_shmem_object *shmem;
> struct panfrost_gem_object *bo;
> + int err;
>
> /* Round up heap allocations to 2MB to keep fault handling simple */
> if (flags & PANFROST_BO_HEAP)
> @@ -282,7 +288,18 @@ panfrost_gem_create(struct drm_device *dev, size_t size, u32 flags)
> bo->noexec = !!(flags & PANFROST_BO_NOEXEC);
> bo->is_heap = !!(flags & PANFROST_BO_HEAP);
>
> + if (!bo->is_heap) {
> + err = drm_gem_shmem_get_pages(shmem);
I really hate the fact we request pages here while we call
panfrost_mmu_map() in panfrost_gem_open(), because ultimately, pages
are requested for the MMU mapping. Also hate the quirk we have in shmem
to call free_pages() instead of put_pages_locked() when the BO refcount
dropped to zero, and I was hoping we could get rid of it at some point
by teaching drivers to request pages when they actually need it instead
of tying pages lifetime to the GEM object lifetime.
Maybe what we should do instead is move the get/put_pages() in
panfrost_mmu_map/unmap() (as I suggested), but have a special mapping
panfrost_mmu_evict/restore() helpers that kill/restore the MMU mappings
without releasing/acquiring the pages ref.
