Re: USB networking MAC regression (aka the usb0 -> eth0 change)
On Sun, May 24, 2009 at 16:34, Nicolas Cavallari wrote: > I don't want to try SHR-unstable right now (i use my FR as my main phone), > but hacking the > g_ether.sh script to do the same thing also fix the problem. Well, for me SHR-unstable is more usable as daily phone than SHR-testing, but that's only my opinion and off-topic :) ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: USB networking MAC regression (aka the usb0 -> eth0 change)
>> Johny Tenfinger wrote : >>> On Sat, May 23, 2009 at 17:05, Nicolas Cavallari wrote: This worked fine during the usb0 phase. But since the distro started using the official MAC address, This wouldn't work anymore. >>> Which distro? In SHR-unstable it should be fixed for long time, but in >>> others, where kernel is using arguments passed by Qi (dunno if Om2009 >>> is using Qi or SHR way for settings MAC address...) it isn't, because >>> of bug in Qi. > On Sat, May 23, 2009 at 19:33, Nicolas Cavallari wrote: >> It's the latest SHR Testing. >> but my Qi installation may not be up to date... Screw that, Qi was up to date. Johny Tenfinger wrote : > > Try latest SHR-unstable. It should be really better than -testing, and > it should have it fixed (not depending to bootloader). >From what i saw, unstable does just increment the host mac adress if the >device and the host mac are the same, just like Paul Fertser's Qi patch on the kernel@ mailling list. I don't want to try SHR-unstable right now (i use my FR as my main phone), but hacking the g_ether.sh script to do the same thing also fix the problem. I also created a ticket on trac, as suggested : https://docs.openmoko.org/trac/ticket/2290 ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: USB networking MAC regression (aka the usb0 -> eth0 change)
Hi, Nicolas Cavallari writes: > For a while, I used to use USB networking to connect my FR to the interwebs. [snip] > The bridge fails because it consider the host mac address to be a local mac > (brctl showmacs) and > so ignore any host with the same mac. Changing the host mac address (with > ifconfig ethXXX hw ether) > solves the problem, but it clearly not a acceptable solution. Thanks for reporting. I've just sent a patch to the kernel ML that should fix the issue. -- Be free, use free (http://www.gnu.org/philosophy/free-sw.html) software! mailto:fercer...@gmail.com ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: USB networking MAC regression (aka the usb0 -> eth0 change)
On Sat, May 23, 2009 at 19:33, Nicolas Cavallari wrote: > Johny Tenfinger a écrit : >> On Sat, May 23, 2009 at 17:05, Nicolas Cavallari wrote: >>> This worked fine during the usb0 phase. But since the distro started using >>> the official >>> MAC address, This wouldn't work anymore. >> >> Which distro? In SHR-unstable it should be fixed for long time, but in >> others, where kernel is using arguments passed by Qi (dunno if Om2009 >> is using Qi or SHR way for settings MAC address...) it isn't, because >> of bug in Qi. > > It's the latest SHR Testing. but my Qi installation may not be up to date... Try latest SHR-unstable. It should be really better than -testing, and it should have it fixed (not depending to bootloader). ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: USB networking MAC regression (aka the usb0 -> eth0 change)
Johny Tenfinger a écrit : > On Sat, May 23, 2009 at 17:05, Nicolas Cavallari wrote: >> This worked fine during the usb0 phase. But since the distro started using >> the official >> MAC address, This wouldn't work anymore. > > Which distro? In SHR-unstable it should be fixed for long time, but in > others, where kernel is using arguments passed by Qi (dunno if Om2009 > is using Qi or SHR way for settings MAC address...) it isn't, because > of bug in Qi. It's the latest SHR Testing. but my Qi installation may not be up to date... ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: USB networking MAC regression (aka the usb0 -> eth0 change)
On Sat, May 23, 2009 at 17:05, Nicolas Cavallari wrote: > This worked fine during the usb0 phase. But since the distro started using > the official > MAC address, This wouldn't work anymore. Which distro? In SHR-unstable it should be fixed for long time, but in others, where kernel is using arguments passed by Qi (dunno if Om2009 is using Qi or SHR way for settings MAC address...) it isn't, because of bug in Qi. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
USB networking MAC regression (aka the usb0 -> eth0 change)
For a while, I used to use USB networking to connect my FR to the interwebs. I wanted to be able to connect my moko to any computer and let my entire LAN talk to it, so instead of using masquerading or routing, i used the much simpler bridge mechanism, to bridge my FR with my network. This worked fine during the usb0 phase. But since the distro started using the official MAC address, This wouldn't work anymore. I traced this problem down to the fact that both the host NIC and the device NIC share the same mac address. I wonder if this is a deliberate change (i hope not), because this not only break bridging, but it would break many other advanced network configurations. The bridge fails because it consider the host mac address to be a local mac (brctl showmacs) and so ignore any host with the same mac. Changing the host mac address (with ifconfig ethXXX hw ether) solves the problem, but it clearly not a acceptable solution. What should i do instead ? Should i report this as a bug ? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community