On 1/10/19 8:54 PM, Simon Goldschmidt wrote:
Hi Marek,
Am 10.01.2019 um 20:49 schrieb Simon Goldschmidt:
This driver was written for Arria10, but it applies to Gen5, too.
The main difference is that Gen5 has 2 MACs (Arria10 has 3) and the syscon bits are encoded in the same register, thus an offset is needed.
This offset is already read from the devicetree, but for Arria10 it is always 0, which is probably why it has been ignored. By using this offset when writing the phy mode into the syscon regiter, we can use this driver to set the phy mode for both of the MACs on Gen5.
Tested on socfpga_socrates (where the 2nd MAC is connected, so a shift offset is required).
Signed-off-by: Simon Goldschmidt simon.k.r.goldschmidt@gmail.com
drivers/net/dwmac_socfpga.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)
diff --git a/drivers/net/dwmac_socfpga.c b/drivers/net/dwmac_socfpga.c index 08fc9677c4..309da69647 100644 --- a/drivers/net/dwmac_socfpga.c +++ b/drivers/net/dwmac_socfpga.c @@ -27,6 +27,7 @@ struct dwmac_socfpga_platdata { struct dw_eth_pdata dw_eth_pdata; enum dwmac_type type; void *phy_intf; + u32 reg_shift; }; static int dwmac_socfpga_ofdata_to_platdata(struct udevice *dev) @@ -63,6 +64,7 @@ static int dwmac_socfpga_ofdata_to_platdata(struct udevice *dev) } pdata->phy_intf = range + args.args[0]; + pdata->reg_shift = args.args[1]; /* * Sadly, the Altera DT bindings don't have SoC-specific compatibles, @@ -88,9 +90,11 @@ static int dwmac_socfpga_probe(struct udevice *dev) struct eth_pdata *edata = &pdata->dw_eth_pdata.eth_pdata; struct reset_ctl_bulk reset_bulk; int ret; - u8 modereg; + u32 modereg; + u32 modemask; - if (pdata->type == DWMAC_SOCFPGA_ARRIA10) { + if (pdata->type == DWMAC_SOCFPGA_ARRIA10 || + pdata->type == DWMAC_SOCFPGA_GEN5) {
I just checked: the Linux driver does not seem to have special handling for Gen5/Arria10/Stratix10. Since Gen5 and Arria10 now both work and the registers for Stratix10 seem the same as for Arria10, could we maybe drop this special handling?
That would mean we could drop this whole 'type' and the 'enum dwmac_type'...
I seem to remember the register layout for selecting the PHY mode on Gen5 and Arria10 was different, please check the datasheets. Of course, the less special-casing, the better.