On Wed, Mar 03, 2021 at 01:54:52PM +0800, Kevin Lo wrote: > On Tue, Mar 02, 2021 at 08:23:38PM -0800, Nick Owens wrote: > > > > On Tue, Mar 2, 2021 at 7:00 PM Kevin Lo <ke...@kevlo.org> wrote: > > > > > > On Tue, Mar 02, 2021 at 02:20:12PM -0800, Nick Owens wrote: > > > > > > > > >Synopsis: multiple rge(4) nics causes hang > > > > >Category: kernel > > > > >Environment: > > > > System : OpenBSD 6.8 > > > > Details : OpenBSD 6.8 (GENERIC.MP) #5: Mon Feb 22 04:36:10 > > > > MST 2021 > > > > > > > > r...@syspatch-68-amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP > > > > > > > > Architecture: OpenBSD.amd64 > > > > Machine : amd64 > > > > >Description: > > > > i have odroid h2+ [1] with the 'net card' [2]. it has two > > > > onboard rge(4) nics and four more on the net card. > > > > > > > > using only rge0/1, the system seems stable. but when i try to > > > > bring up more nics, e.g. ifconfig rge2/3/4/5 up, the system will hang. i > > > > cannot do anything, except pressing ^C on the console prints a ^C back > > > > very slowly but doesn't do anything else. it seems like ifconfig is > > > > stuck in kernel, and the system also stops doing anything else like > > > > routing the other working nics. > > > > > > > > current setup is rge0 for WAN with dhcpcd from ports for IPv6 > > > > access, and bridge(4) + vether(4) + rge1/2/3/4/5 with dhcpd/rad for LAN. > > > > for full config, you can find the ansible scrips at > > > > https://github.com/mischief/ansible/. have never used ansible before so > > > > i am sure it is crappy :) > > > > > > > > questions: > > > > 1. how can i debug what is happening when i up the nic > > > > with ifconfig? > > > > 2. would someone on openbsd team (maybe kevin lo) like a copy of > > > > this hardware donated? > > > > > > Hi, > > > > > > otto@ fixed interrupt storm on rge(4): > > > https://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/dev/pci/if_rgereg.h.diff?r1=1.4&r2=1.5 > > > > > > Please apply the patch or try to run the latest OpenBSD snapshot, thanks. > > > > hi! thanks! > > > > i updated to a snapshot and up'ing the interfaces no longer hangs my > > system, and the nics all seem to work. my original bug was fixed > > months ago :-D > > > > one last concern, this seems to happen often: > > > > fugu# dmesg| grep 'watchdog timeout' > > rge4: watchdog timeout > > rge5: watchdog timeout > > rge2: watchdog timeout > > rge2: watchdog timeout > > > > is this anything to worry about? > > Generally watchdog timeout means driver didn't get TX completion interrupt > for a given time which in turn indicates TX MAC is in stuck condition. > RX MAC could be also in the stuck condition at that moment > (i.e. complete MAC lock-up). > > If this happens often, it's definitely a bug.
not happening that much, it maybe only happened during nic up or when i was replugging the physical cables. i will keep an eye on it. thank you for developing this driver so that this hardware works as an awesome router!
