On Fri, Feb 26, 2021 at 10:26:47AM +0100, Thomas Zimmermann wrote:

USB devices cannot perform DMA and hence have no dma_mask set in their device structure. Therefore importing dmabuf into a USB-based driver fails, which breaks joining and mirroring of display in X11.

For USB devices, pick the associated USB controller as attachment device. This allows the DRM import helpers to perform the DMA setup. If the DMA controller does not support DMA transfers, we're out of luck and cannot import. Our current USB-based DRM drivers don't use DMA, so the actual DMA device is not important.

Drivers should use DRM_GEM_SHMEM_DROVER_OPS_USB to initialize their instance of struct drm_driver.

Tested by joining/mirroring displays of udl and radeon un der Gnome/X11.

v5:

• provide a helper for USB interfaces (Alan)
• add FIXME item to documentation and TODO list (Daniel)

v4:

• implement workaround with USB helper functions (Greg)
• use struct usb_device->bus->sysdev as DMA device (Takashi)

v3:

• drop gem_create_object
• use DMA mask of USB controller, if any (Daniel, Christian, Noralf)

v2:

• move fix to importer side (Christian, Daniel)
• update SHMEM and CMA helpers for new PRIME callbacks

Signed-off-by: Thomas Zimmermann tzimmermann@suse.de Fixes: 6eb0233ec2d0 ("usb: don't inherity DMA properties for USB devices") Tested-by: Pavel Machek pavel@ucw.cz Acked-by: Christian König christian.koenig@amd.com Acked-by: Daniel Vetter daniel.vetter@ffwll.ch Cc: Christoph Hellwig hch@lst.de Cc: Greg Kroah-Hartman gregkh@linuxfoundation.org Cc: stable@vger.kernel.org # v5.10+

---

Documentation/gpu/todo.rst | 15 ++++++++++ drivers/gpu/drm/drm_prime.c | 45 ++++++++++++++++++++++++++++++ drivers/gpu/drm/tiny/gm12u320.c | 2 +- drivers/gpu/drm/udl/udl_drv.c | 2 +- drivers/usb/core/usb.c | 31 ++++++++++++++++++++ include/drm/drm_gem_shmem_helper.h | 16 +++++++++++ include/drm/drm_prime.h | 5 ++++ include/linux/usb.h | 24 ++++++++++++++++ 8 files changed, 138 insertions(+), 2 deletions(-)

diff --git a/Documentation/gpu/todo.rst b/Documentation/gpu/todo.rst index f872d3d33218..c185e0a2951e 100644 --- a/Documentation/gpu/todo.rst +++ b/Documentation/gpu/todo.rst @@ -617,6 +617,21 @@ Contact: Daniel Vetter Level: Intermediate +Remove automatic page mapping from dma-buf importing +----------------------------------------------------

+When importing dma-bufs, the dma-buf and PRIME frameworks automatically map +imported pages into the importer's DMA area. This is a problem for USB devices, +which do not support DMA operations. By default, importing fails for USB +devices. USB-based drivers work around this problem by employing +drm_gem_prime_import_usb(). To fix the issue, automatic page mappings should +be removed from the buffer-sharing code.

+Contact: Thomas Zimmermann tzimmermann@suse.de, Daniel Vetter

+Level: Advanced

Better Testing

diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c index 2a54f86856af..59013bb1cd4b 100644 --- a/drivers/gpu/drm/drm_prime.c +++ b/drivers/gpu/drm/drm_prime.c @@ -29,6 +29,7 @@ #include <linux/export.h> #include <linux/dma-buf.h> #include <linux/rbtree.h> +#include <linux/usb.h> #include <drm/drm.h> #include <drm/drm_drv.h> @@ -1055,3 +1056,47 @@ void drm_prime_gem_destroy(struct drm_gem_object *obj, struct sg_table *sg) dma_buf_put(dma_buf); } EXPORT_SYMBOL(drm_prime_gem_destroy);

+/**

• • drm_gem_prime_import_usb - helper library implementation of the import callback for USB devices
• • @dev: drm_device to import into
• • @dma_buf: dma-buf object to import
• • This is an implementation of drm_gem_prime_import() for USB-based devices.
• • USB devices cannot perform DMA directly. This function selects the USB host
• • controller as DMA device instead. Drivers can use this as their
• • &drm_driver.gem_prime_import implementation.
• • See also drm_gem_prime_import().
• • FIXME: The dma-buf framework expects to map the exported pages into

• •    the importer's DMA area. USB devices don't support DMA, and
    

• •    importing would fail. Foir the time being, this function provides
    

• •    a workaround by using the USB controller's DMA area. The real
    

• •    solution is to remove page-mapping operations from the dma-buf
    

• •    framework.
    

• • Returns: A GEM object on success, or a pointer-encoder errno value otherwise.
• */

+#ifdef CONFIG_USB +struct drm_gem_object *drm_gem_prime_import_usb(struct drm_device *dev,

• 				struct dma_buf *dma_buf)
  

+{

• struct device *dmadev;
• struct drm_gem_object *obj;
• if (!dev_is_usb(dev->dev))

• return ERR_PTR(-ENODEV);
  

I have resisted the "dev_is_*()" type of function for USB for a long time now, and I really don't want to add it now.

The driver core explicitly was not created with RTI (run type identification), but over time it has been slowly added on a per-bus basis for various reasons, some good and others not good.

In this function, why would a drm device that was NOT a usb device ever call it? Because of that, I don't think dev_is_usb() is needed at all, just don't call this function unless it really is a USB device.

If you need help enforcing it, add a 'struct usb_interface *' to the function parameters but please don't create this function.

Hm, looks like we have 2 extern definitions of usb_bus_type in different .h files, I should go fix that up now to try to prevent this type of thing in the first place...

thanks,

greg k-h