Thanks for the patch, it all looks good to me. Feel free to add my r-b: Reviewed-by: Robert Foss <[email protected]>
On Thu, 21 Jan 2021 at 21:34, Takashi Iwai <[email protected]> wrote: > > The URB buffer allocation of pwc driver involves with the > dma_map_single(), and it needs to pass the right device. Currently it > passes usb_device.dev, but it's no real device that manages the DMA. > Since the passed device has no DMA mask set up, now the pwc driver > hits the WARN_ON_ONCE() check in dma_map_page_attrs() (that was > introduced in 5.10), resulting in an error at URB allocations. > Eventually this ended up with the black output. > > This patch fixes the bug by passing the proper device, the bus > controller, to make the URB allocation and map working again. > > BugLink: https://bugzilla.suse.com/show_bug.cgi?id=1181133 > Cc: <[email protected]> > Signed-off-by: Takashi Iwai <[email protected]> > --- > drivers/media/usb/pwc/pwc-if.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/media/usb/pwc/pwc-if.c b/drivers/media/usb/pwc/pwc-if.c > index 61869636ec61..d771160bb168 100644 > --- a/drivers/media/usb/pwc/pwc-if.c > +++ b/drivers/media/usb/pwc/pwc-if.c > @@ -461,7 +461,7 @@ static int pwc_isoc_init(struct pwc_device *pdev) > urb->pipe = usb_rcvisocpipe(udev, pdev->vendpoint); > urb->transfer_flags = URB_ISO_ASAP | URB_NO_TRANSFER_DMA_MAP; > urb->transfer_buffer_length = ISO_BUFFER_SIZE; > - urb->transfer_buffer = pwc_alloc_urb_buffer(&udev->dev, > + urb->transfer_buffer = > pwc_alloc_urb_buffer(udev->bus->controller, > > urb->transfer_buffer_length, > > &urb->transfer_dma); > if (urb->transfer_buffer == NULL) { > @@ -524,7 +524,7 @@ static void pwc_iso_free(struct pwc_device *pdev) > if (urb) { > PWC_DEBUG_MEMORY("Freeing URB\n"); > if (urb->transfer_buffer) > - pwc_free_urb_buffer(&urb->dev->dev, > + pwc_free_urb_buffer(urb->dev->bus->controller, > > urb->transfer_buffer_length, > urb->transfer_buffer, > urb->transfer_dma); > -- > 2.26.2 >
