Thank you so much for the patch. This did seem to fix my problem, but then I found that it seems to work for some Beaglebone Blacks but not others. Even boards from the same manufacturer (but potentially different lots). What causes this issue? Is there something that would explain why the patch would fix the problem on some boards but not others? Are there some manufacturers that are likely to be better than others?
Thanks, Dave On Monday, March 19, 2018 at 4:56:08 PM UTC-4, RobertCNelson wrote: > > On Mon, Mar 19, 2018 at 3:47 PM, Dave Tucker <[email protected] > <javascript:>> wrote: > > Ethernet on my Beaglebone Black is failing to come online about 3% of > the > > time after power up. Attached are dmesg output from times when ethernet > > works (good_dmesg.txt) and the more rare times when it fails > > (bad_dmesg.txt). I've also attached full_bad_boot.txt which is the > debug > > serial output from a boot that failed. Here is a diff of the two > dmesgs, > > which I think might highlight the relevant areas: > > 70c70 > > < raid6: int32x1 xor() 140 MB/s > > --- > >> raid6: int32x1 xor() 143 MB/s > > 76c76 > > < raid6: int32x8 xor() 118 MB/s > > --- > >> raid6: int32x8 xor() 117 MB/s > > 124c124 > > < davinci_mdio 4a101000.mdio: detected phy mask fffffffe > > --- > >> davinci_mdio 4a101000.mdio: detected phy mask fffffffb > > 126c126 > > < davinci_mdio 4a101000.mdio: phy[0]: device 4a101000.mdio:00, driver > SMSC > > LAN8710/LAN8720 > > --- > >> davinci_mdio 4a101000.mdio: phy[2]: device 4a101000.mdio:02, driver > SMSC > >> LAN8710/LAN8720 > > 169a170,172 > >> EXT4-fs (mmcblk1p3): INFO: recovery required on readonly filesystem > >> EXT4-fs (mmcblk1p3): write access will be enabled during recovery > >> EXT4-fs (mmcblk1p3): recovery complete > > 174a178 > >> EXT4-fs (mmcblk1p2): recovery complete > > 175a180 > >> EXT4-fs (mmcblk1p4): recovery complete > > 183c188,189 > > < SMSC LAN8710/LAN8720 4a101000.mdio:00: attached PHY driver [SMSC > > LAN8710/LAN8720] (mii_bus:phy_addr=4a101000.mdio:00, irq=-1) > > --- > >> libphy: PHY 4a101000.mdio:00 not found > >> net eth0: phy "4a101000.mdio:00" not found on slave 0, err -19 > > 185,186c191 > > < cpsw 4a100000.ethernet eth0: Link is Up - 100Mbps/Full - flow control > > rx/tx > > < IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready > > --- > >> random: crng init done > > > > I don't know what any of this means. I do know I'm not doing anything > > different for the times that fail. I've automated the entire process of > > turning it on and off and seeing if the network interface works. > > > > I'm running a custom build put together with Yocto (Rocko, 2.4), and > I've > > compiled u-boot separately. I can provide my device tree dts and > uEnv.txt > > if needed. > > > > Is this a known problem? What can I do to fix this? > > Yes... Apply this patch, and it'll fix it 95% of the time... > > > https://github.com/RobertCNelson/bb-kernel/blob/am33x-v4.12/patches/drivers/ti/cpsw/0001-cpsw-search-for-phy.patch > > > Later boards (am43/etc) added a gpio to reset the phy, due to what was > seen on this design.. > > Regards, > > -- > Robert Nelson > https://rcn-ee.com/ > -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/beagleboard/b4abca94-0d49-4cc7-af5c-59833281c15c%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
