Hi Andrew On Tue, Aug 6, 2024 at 4:13 PM Andrew Lunn <andrew@xxxxxxx> wrote: > > On Mon, Aug 05, 2024 at 05:36:30PM -0700, Jitendra Vegiraju wrote: > > Hi Andrew, > > On Fri, Aug 2, 2024 at 3:59 PM Andrew Lunn <andrew@xxxxxxx> wrote: > > > > > > > + user_ver = stmmac_get_user_version(priv, GMAC4_VERSION); > > > > + if (priv->synopsys_id == DWXGMAC_CORE_4_00 && > > > > + user_ver == DWXGMAC_USER_VER_X22) > > > > + mac->dma = &dwxgmac400_dma_ops; > > > > > > I know nothing about this hardware.... > > > > > > Does priv->synopsys_id == DWXGMAC_CORE_4_0 not imply > > > dwxgmac400_dma_ops? > > > > > > Could a user synthesise DWXGMAC_CORE_4_00 without using > > > dwxgmac400_dma_ops? Could dwxgmac500_dma_ops or dwxgmac100_dma_ops be > > > used? > > Yes, the user can choose between Enhanced DMA , Hyper DMA , Normal DMA. > > This SoC support has chosen Hyper DMA for future expandability. > > Is there a register which describes the synthesis configuration? It is > much better that the hardware tells us what it is, rather than having > to expand this condition for every new devices which gets added. > Sorry, for the delayed response. We got confirmation that HDMA capability can be identified by the presence of DWXGMAC_CORE_4_xx and device_id 0x55. > Also, what is the definition of user_ver. Can we guarantee this is > unique and can actually be used to determine what DMA variant has been > synthesised? The initial information I got on this is that user versions are allocated and reserved by Synopsis. We probably don't need to key off this information now. > > Andrew
