Daniel, On Tue, Aug 12, 2014 at 12:15 PM, Daniel Mack <zonque@xxxxxxxxx> wrote: > On 08/12/2014 07:05 PM, Bin Liu wrote: >>> Don't you need to check for (urb->transfer_flags & URB_ZERO_PACKET) >>> somewhere here? I'd not expect an extra 0-byte packet at the end of a >>> transfer unless this flag is set. >> >> mode is set to 1 when (urb->transfer_flags & URB_ZERO_PACKET), in >> musb_tx_dma_program in musb_host.c. > > Hmm, yes, but only for (!CONFIG_USB_INVENTRA_DMA && > !CONFIG_USB_UX500_DMA). Otherwise, mode will be 1 'if (length > > pkt_size)', so on those platforms, the driver will now send terminating > ZLPs whenever a transfer spans across multiple packets. Or am I missing > anything? 633 #if defined(CONFIG_USB_INVENTRA_DMA) || defined(CONFIG_USB_UX500_DMA) ... 658 #else 659 if (!is_cppi_enabled() && !tusb_dma_omap()) 660 return false; ... 668 mode = (urb->transfer_flags & URB_ZERO_PACKET) ? 1 : 0; 669 #endif So mode is only set to 1 for CPPI DMA when URB_ZERO_PACKET is set, then musb_cppi41 driver will transmit the ZLP. For other DMA engine (INVERNTRA or UX500), its driver should have its own channel_program() callback, cppi41 driver should not be called. Regards, -Bin. > > You might need to pass a pointer to the urb or its transfer_flags down > to channel_program(). > > > Thanks, > Daniel -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
