On 15/12/2023 08:08, Siddharth Vadapalli wrote: >>> err: >>> @@ -395,6 +410,40 @@ struct k3_udma_glue_tx_channel *k3_udma_glue_request_tx_chn(struct device *dev, >>> } >>> EXPORT_SYMBOL_GPL(k3_udma_glue_request_tx_chn); >>> >>> +struct k3_udma_glue_tx_channel * >>> +k3_udma_glue_request_tx_chn_by_id(struct device *dev, struct k3_udma_glue_tx_channel_cfg *cfg, >>> + struct device_node *udmax_np, u32 thread_id) >> >> udmax_np is not dev->of_node ? > > I am not sure I fully understand the question. If you meant to ask if the driver > which uses this API will not have its device's of_node set to udmax_np, then yes > that's correct. > > The driver shall be probed over RPMsg-bus, due to which its device's of_node > will not be udmax_np. Additionally, the udmax_np is the device-tree node of one > of the DMA Controllers described in the device-tree. The driver shall obtain the > reference to the udmax_np node using the API: > of_find_compatible_node() > with the compatible to be passed to the above API being a part of the driver's > data. Thus, it is possible to specify which DMA Controller to use by specifying > the compatible in the driver's data. I hope that I have answered your question. > Please let me know otherwise. I see, thank you for the detailed explanation! > Thank you for reviewing the series. I will rename the API as mentioned above and > if the question you had above regarding the of_node has been addressed, I will > post the v3 series. Kindly let me know. I don't have other open issues, thanks for the updates -- Péter
