Date: Sun, 9 Sep 2012 23:17:06 -0700 From: John-Mark Gurney <jmg@funkthat.com> To: Tim Kientzle <kientzle@FreeBSD.org> Cc: freebsd-arm@FreeBSD.org Subject: Re: -current on BeableBone successful Message-ID: <20120910061705.GO58312@funkthat.com> In-Reply-To: <4448530B-7916-4642-BD32-6A96AD5DB22B@FreeBSD.org> References: <20120910014813.GL58312@funkthat.com> <4448530B-7916-4642-BD32-6A96AD5DB22B@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Tim Kientzle wrote this message on Sun, Sep 09, 2012 at 20:16 -0700:
> On Sep 9, 2012, at 6:48 PM, John-Mark Gurney wrote:
>
> > Well, I have successfully booted FreeBSD-current on the BeagleBone:
> > FreeBSD beaglebone 10.0-CURRENT FreeBSD 10.0-CURRENT #0 r240251: Wed Nov 2 14:26:08 PDT 2011 jmg@pcbsd-779:/usr/obj/arm.armv6/usr/src.HEAD/sys/BEAGLEBONE arm
> >
> > Though there are a few issues... I'm using Kientzle's script to build
> > the bootable image, so one is related to his script, and the others are
> > more generic issues..
> >
> > The first off is that for some reason, the uenv.txt file doesn't load..
> > I'm not sure what is wrong, but if I type in:
> > fatload mmc 0:1 0x88000000 ubldr;bootelf 0x88000000
>
> I just ran into this myself. I think the Arago sources
> broke environment loading.
>
> I started to debug the Arago U-Boot build this weekend
> but decided it would likely be more productive to switch
> to using the Denx sources. Denx publishes periodic
> snapshots which should make things a little more stable.
Sounds good...
> > manually, it boots fine...
> >
> > The second is that it looks like we have some left over debugging in
> > SIOC[SG]IFMEDIA that we need to remove, and a fall through comment that
> > doesn't look like it is applicable any more... Should I just remove
> > these?
>
> The cpsw driver still needs work. I've fixed a couple of problems, but there are some deeper issues:
> * Under memory pressure, the rx handler populates the receive buffer list with NULLs then panics when it tries to process them.
> * I've seen the driver stall repeatedly; I'm not sure why yet but I suspect the driver is mis-handling queue restarts.
> * The driver doesn't detach properly (which results in panics if you build the driver as a module and try to unload it)
>
> > Third is that I get this error:
> > ip length 328 disagrees with bytes received 330.
> > accepting packet with data after udp payload.
> >
> > This appeard to be from sbin/dhclient/packet.c... Not sure exactly why
> > we are returning a large packet to userland?
>
> I haven't seen this one.
I'll try to get a tcpdump from both the server and the client side to
see what comes up...
> > Thanks for the work. In a couple days, I'll figure out how well the
> > VLANs work on this system.
> >
> > Thanks to Tim for his script that made building the image really easy,
> > and to the others for getting this working!
>
> Here are my priorities for the BeagleBone, in roughly this order:
> * Fix the cpsw Ethernet driver
> * Get buildworld/buildkernel working correctly natively. (30+ hour builds make this tedious to test; last time I tested the builds completed but installing world crashed things badly)
> * Work on performance for the MMC driver
> * USB host driver
>
> Apart from Ethernet-related panics, the system has
> been remarkably stable. I've been building ports and
> otherwise hammering the system for days on end without
> any apparent ill effects.
Cool...
It's been a while, but I can look a bit at the ethernet driver if you'd
like some help...
Thanks again for your help.
--
John-Mark Gurney Voice: +1 415 225 5579
"All that I will do, has been done, All that I have, has not."
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20120910061705.GO58312>