On Fri, Jan 12, 2024 at 5:13 AM Daniel Vetter <daniel@xxxxxxxx> wrote: > > On Sun, Dec 24, 2023 at 02:29:24AM +0800, Jason-JH.Lin wrote: > > Add secure buffer control flow to mtk_drm_gem. > > > > When user space takes DRM_MTK_GEM_CREATE_ENCRYPTED flag and size > > to create a mtk_drm_gem object, mtk_drm_gem will find a matched size > > dma buffer from secure dma-heap and bind it to mtk_drm_gem object. > > > > Signed-off-by: Jason-JH.Lin <jason-jh.lin@xxxxxxxxxxxx> > > Apologies for jumping rather late, but last year was a mess here. > > There's the fundamental issue that this is new uapi, and it needs open > userspace, and I haven't seen that. The open userspace for it is currently in minigbm for ChromeOS here: https://chromium.googlesource.com/chromiumos/platform/minigbm/+/main/mediatek.c#322 Does that satisfy that particular requirement? > > What's more, this is a pure kms api so there's no precedent at all for > adding special ioctl to those - all the existing support for > protected/restricted content buffers in upstream has used render nodes and > EGL_EXT_protected_content in mesa3d to enable this feature on the drm/kms > side. So I'm not exactly sure what your plan here is, but you need one, > and it needs to be more than a testcase/demo. I'm probably not understanding something here, but for the Intel protected content allocation there was a specific structure (drm_i915_gem_create_ext_protected_content) added to the DRM_IOCTL_I915_GEM_CREATE_EXT ioctl in order to enable allocation of protected buffers. So wouldn't that be precedent for using an ioctl like this to allocate a GEM object? > > The other issue, and the reason I've looked into the mtk code, is that the > dma-buf implementation breaks the dma-buf api. So that needs to be > changed. Yeah, agreed here...I do see now that attaching dma_bufs for content that is completely inaccessible by the kernel will be a no-go. > > Finally I think the primary way to display a protected content buffer on a > pure kms driver should be by using prime fd2handle buffer importing. > Because you're adding a dma-buf heap it's already possible for userspace > to use this path (or at least try), and so we have to make this path work > anyway. Is what you're getting at here having MTK implement their own gem_prime_import in order to work around having to do the dma_buf attach operation? (from looking at the code, this appears to be the only place in non-vendor specific code that dma_buf_attach is called) > > Once we have the prime import path working correctly for protected content > buffers (which should shake out all the dma-api issues I've explained in > the dma-buf heaps thread), it should be possible to implement the direct > allocation function in a generic helper: > > struct drm_gem_object * drm_gem_create_object_from_heap(struct drm_device *dev, > struct drm_file *file, > struct drm_buf_heap *heap); > > Which does roughly: > > - allocate a dma-buf from the desired heap > - import that dma-buf into the device using prime for the drm_file > - using the already implemented driver import code for special cases like > protected content buffers > > There should be no need to hand-roll all this code here, and especially > not have any special-casing for the heap string name or things like that. > That all must be handled in the dma-buf prime import code. > > Cheers, Sima > > > --- > > drivers/gpu/drm/mediatek/mtk_drm_gem.c | 85 +++++++++++++++++++++++++- > > drivers/gpu/drm/mediatek/mtk_drm_gem.h | 4 ++ > > 2 files changed, 88 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/gpu/drm/mediatek/mtk_drm_gem.c b/drivers/gpu/drm/mediatek/mtk_drm_gem.c > > index 30e347adcbe9..858f34a735f8 100644 > > --- a/drivers/gpu/drm/mediatek/mtk_drm_gem.c > > +++ b/drivers/gpu/drm/mediatek/mtk_drm_gem.c > > @@ -4,6 +4,8 @@ > > */ > > > > #include <linux/dma-buf.h> > > +#include <linux/dma-heap.h> > > +#include <uapi/linux/dma-heap.h> > > #include <drm/mediatek_drm.h> > > > > #include <drm/drm.h> > > @@ -55,6 +57,81 @@ static struct mtk_drm_gem_obj *mtk_drm_gem_init(struct drm_device *dev, > > return mtk_gem_obj; > > } > > > > +struct mtk_drm_gem_obj *mtk_drm_gem_create_from_heap(struct drm_device *dev, > > + const char *heap, size_t size) > > +{ > > + struct mtk_drm_private *priv = dev->dev_private; > > + struct mtk_drm_gem_obj *mtk_gem; > > + struct drm_gem_object *obj; > > + struct dma_heap *dma_heap; > > + struct dma_buf *dma_buf; > > + struct dma_buf_attachment *attach; > > + struct sg_table *sgt; > > + struct iosys_map map = {}; > > + int ret; > > + > > + mtk_gem = mtk_drm_gem_init(dev, size); > > + if (IS_ERR(mtk_gem)) > > + return ERR_CAST(mtk_gem); > > + > > + obj = &mtk_gem->base; > > + > > + dma_heap = dma_heap_find(heap); > > + if (!dma_heap) { > > + DRM_ERROR("heap find fail\n"); > > + goto err_gem_free; > > + } > > + dma_buf = dma_heap_buffer_alloc(dma_heap, size, > > + O_RDWR | O_CLOEXEC, DMA_HEAP_VALID_HEAP_FLAGS); > > + if (IS_ERR(dma_buf)) { > > + DRM_ERROR("buffer alloc fail\n"); > > + dma_heap_put(dma_heap); > > + goto err_gem_free; > > + } > > + dma_heap_put(dma_heap); > > + > > + attach = dma_buf_attach(dma_buf, priv->dma_dev); > > + if (IS_ERR(attach)) { > > + DRM_ERROR("attach fail, return\n"); > > + dma_buf_put(dma_buf); > > + goto err_gem_free; > > + } > > + > > + sgt = dma_buf_map_attachment(attach, DMA_BIDIRECTIONAL); > > + if (IS_ERR(sgt)) { > > + DRM_ERROR("map failed, detach and return\n"); > > + dma_buf_detach(dma_buf, attach); > > + dma_buf_put(dma_buf); > > + goto err_gem_free; > > + } > > + obj->import_attach = attach; > > + mtk_gem->dma_addr = sg_dma_address(sgt->sgl); > > + mtk_gem->sg = sgt; > > + mtk_gem->size = dma_buf->size; > > + > > + if (!strcmp(heap, "mtk_svp") || !strcmp(heap, "mtk_svp_cma")) { > > + /* secure buffer can not be mapped */ > > + mtk_gem->secure = true; > > + } else { > > + ret = dma_buf_vmap(dma_buf, &map); > > + mtk_gem->kvaddr = map.vaddr; > > + if (ret) { > > + DRM_ERROR("map failed, ret=%d\n", ret); > > + dma_buf_unmap_attachment(attach, sgt, DMA_BIDIRECTIONAL); > > + dma_buf_detach(dma_buf, attach); > > + dma_buf_put(dma_buf); > > + mtk_gem->kvaddr = NULL; > > + } > > + } > > + > > + return mtk_gem; > > + > > +err_gem_free: > > + drm_gem_object_release(obj); > > + kfree(mtk_gem); > > + return ERR_PTR(-ENOMEM); > > +} > > + > > struct mtk_drm_gem_obj *mtk_drm_gem_create(struct drm_device *dev, > > size_t size, bool alloc_kmap) > > { > > @@ -225,7 +302,9 @@ struct drm_gem_object *mtk_gem_prime_import_sg_table(struct drm_device *dev, > > if (IS_ERR(mtk_gem)) > > return ERR_CAST(mtk_gem); > > > > + mtk_gem->secure = !sg_page(sg->sgl); > > mtk_gem->dma_addr = sg_dma_address(sg->sgl); > > + mtk_gem->size = attach->dmabuf->size; > > mtk_gem->sg = sg; > > > > return &mtk_gem->base; > > @@ -301,7 +380,11 @@ int mtk_gem_create_ioctl(struct drm_device *dev, void *data, > > struct drm_mtk_gem_create *args = data; > > int ret; > > > > - mtk_gem = mtk_drm_gem_create(dev, args->size, false); > > + if (args->flags & DRM_MTK_GEM_CREATE_ENCRYPTED) > > + mtk_gem = mtk_drm_gem_create_from_heap(dev, "mtk_svp_cma", args->size); > > + else > > + mtk_gem = mtk_drm_gem_create(dev, args->size, false); > > + > > if (IS_ERR(mtk_gem)) > > return PTR_ERR(mtk_gem); > > > > diff --git a/drivers/gpu/drm/mediatek/mtk_drm_gem.h b/drivers/gpu/drm/mediatek/mtk_drm_gem.h > > index 90f3d2916ec5..8fd5ce827d4f 100644 > > --- a/drivers/gpu/drm/mediatek/mtk_drm_gem.h > > +++ b/drivers/gpu/drm/mediatek/mtk_drm_gem.h > > @@ -27,9 +27,11 @@ struct mtk_drm_gem_obj { > > void *cookie; > > void *kvaddr; > > dma_addr_t dma_addr; > > + size_t size; > > unsigned long dma_attrs; > > struct sg_table *sg; > > struct page **pages; > > + bool secure; > > }; > > > > #define to_mtk_gem_obj(x) container_of(x, struct mtk_drm_gem_obj, base) > > @@ -37,6 +39,8 @@ struct mtk_drm_gem_obj { > > void mtk_drm_gem_free_object(struct drm_gem_object *gem); > > struct mtk_drm_gem_obj *mtk_drm_gem_create(struct drm_device *dev, size_t size, > > bool alloc_kmap); > > +struct mtk_drm_gem_obj *mtk_drm_gem_create_from_heap(struct drm_device *dev, > > + const char *heap, size_t size); > > int mtk_drm_gem_dumb_create(struct drm_file *file_priv, struct drm_device *dev, > > struct drm_mode_create_dumb *args); > > struct sg_table *mtk_gem_prime_get_sg_table(struct drm_gem_object *obj); > > -- > > 2.18.0 > > > > -- > Daniel Vetter > Software Engineer, Intel Corporation > http://blog.ffwll.ch
