Messages in this thread | ![/](/images/icornerl.gif) | | Date | Mon, 16 Jan 2023 18:06:54 +0000 | From | "Russell King (Oracle)" <> | Subject | Re: [PATCH v2] net: mtk_sgmii: implement mtk_pcs_ops |
| |
On Mon, Jan 16, 2023 at 06:59:19PM +0100, Bjørn Mork wrote: > "Russell King (Oracle)" <linux@armlinux.org.uk> writes: > > > On Mon, Jan 16, 2023 at 05:45:12PM +0100, Bjørn Mork wrote: > >> Bjørn Mork <bjorn@mork.no> writes: > >> > >> >> You have bmcr=0x1000, but the code sets two bits - SGMII_AN_RESTART and > >> >> SGMII_AN_ENABLE which are bits 9 and 12, so bmcr should be 0x1200, not > >> >> 0x1000. Any ideas why? > >> > > >> > No, not really > >> > >> Doh! Looked over it again, and this was my fault of course. Had an > >> > >> "bmcr = SGMII_AN_ENABLE;" > >> > >> line overwriting the original value from a previous attempt without > >> changing the if condition.. Thanks for spotting that. > >> > >> But this still doesn't work any better: > >> > >> [ 43.019395] mtk_soc_eth 15100000.ethernet wan: Link is Down > >> [ 45.099898] mtk_sgmii_select_pcs: id=1 > >> [ 45.103653] mtk_pcs_config: interface=4 > >> [ 45.107473] offset:0 0x140 > >> [ 45.107476] offset:4 0x4d544950 > >> [ 45.110181] offset:8 0x20 > >> [ 45.113305] forcing AN > >> [ 45.118256] mtk_pcs_config: rgc3=0x0, advertise=0x1 (changed), link_timer=1600000, sgm_mode=0x103, bmcr=0x1200, use_an=1 > >> [ 45.129191] mtk_pcs_link_up: interface=4 > >> [ 45.133100] offset:0 0x81140 > >> [ 45.133102] offset:4 0x4d544950 > >> [ 45.135967] offset:8 0x1 > >> [ 45.139104] mtk_soc_eth 15100000.ethernet wan: Link is Up - 1Gbps/Full - flow control rx/tx > > > > In your _dump_pcs_ctrl() function, please can you dump the > > SGMSYS_SGMII_MODE register as well (offset 0x20), in case this gives > > more clue as to what's going on. > > > [ 49.339410] mtk_soc_eth 15100000.ethernet wan: Link is Down > [ 52.459913] mtk_sgmii_select_pcs: id=1 > [ 52.463673] mtk_pcs_config: interface=4 > [ 52.467494] offset:0 0x140 > [ 52.467496] offset:4 0x4d544950 > [ 52.470199] offset:8 0x20 > [ 52.473325] offset:20 0x10000 > [ 52.475929] forcing AN > [ 52.481232] mtk_pcs_config: rgc3=0x0, advertise=0x1 (changed), link_timer=1600000, sgm_mode=0x103, bmcr=0x1200, use_an=1 > [ 52.492166] mtk_pcs_link_up: interface=4 > [ 52.496072] offset:0 0x81140 > [ 52.496074] offset:4 0x4d544950 > [ 52.498938] offset:8 0x1 > [ 52.502067] offset:20 0x10000 > [ 52.504599] mtk_soc_eth 15100000.ethernet wan: Link is Up - 1Gbps/Full - flow control rx/tx > [ 65.979410] mtk_soc_eth 15100000.ethernet wan: Link is Down > [ 70.139856] mtk_sgmii_select_pcs: id=1 > [ 70.143616] mtk_pcs_config: interface=22 > [ 70.147523] offset:0 0x81140 > [ 70.147525] offset:4 0x4d544950 > [ 70.150402] offset:8 0x1 > [ 70.153526] offset:20 0x10000 > [ 70.156049] mtk_pcs_config: rgc3=0x4, advertise=0x20 (changed), link_timer=10000000, sgm_mode=0x0, bmcr=0x0, use_an=0 > [ 70.169672] mtk_pcs_link_up: interface=22 > [ 70.173664] offset:0 0x40140 > [ 70.173666] offset:4 0x4d544950 > [ 70.176530] offset:8 0x20 > [ 70.179659] offset:20 0x10000 > [ 70.182279] mtk_soc_eth 15100000.ethernet wan: Link is Up - 2.5Gbps/Full - flow control rx/tx
Umm. What's at offset:20 seems to be unprogrammable - it always reads back with only bit 16 set! This probably explains why it's not working, as it looks like it can't be programmed to operate in SGMII mode!
-- RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
| ![\](/images/icornerr.gif) |