Date: Wed, 1 Dec 2004 15:19:13 +0100 (CET) From: Barry Bouwsma <freebsd-misuser@remove-NOSPAM-to-reply.NOSPAM.dyndns.dk> To: freebsd-usb@freebsd.org Subject: Re: USB Ethernet (aue) panics on 4.10 (kern/69319) Message-ID: <200412011419.iB1EJDs10868@Mail.NOSPAM.DynDNS.dK> References: <20041115000815.GD57138@tuatara.fishballoon.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[drop me from replies and I'll catch up via the archives, thankz] Apologies for not finding this until a couple weeks late... On Mon, 15 Nov 2004 00:08:15 +0000, Scott Mitchell wrote: > Is anyone else using USB Ethernet - particularly anything that uses the > aue(4) driver - on FreeBSD 4.10, with or without success? I have a > reproducible panic when aue(4) devices are first brought up, ie. on the > first ifconfig(8) of the device, with two different adapters on two Ah, I'm actually trying to get an axe-type device working based on code lifted from -current, as I picked up one of those type of adapters recently. Is there any work/plan to MFC if_axe into 4.x before sending it to its demise? The panic I see (with both OHCI and EHCI and FreeBSD 4.x) is the same as the panic you reported in the PR -- thanks! For all I know, the axe(4) device I have -- a USB2-capable unit -- already is in 4.x, but I found a relevant PR about what I experience on my 4.x when attaching via UHCI: kern/73000 Unfortunately, when I built my latest -current-based 4.x USB kernel modules, the patch in PR 73000 failed to solve the problem of eternal UHCI attach. Perhaps there's something in 5.x/-current that's missing from my 4.x; alternatively the one-month MT4 that was announced has probably passed so I should check the latest source. I experience the panic when I `ifconfig axe0 something' the interface -- after a second or two, an endless stream of messages, no debugger, at best auto-reboot. Anyway, I'll report back when I bring my 4.x system a bit more up-to-speed. I've connected the device to NetBSD- current and if I recall, it works with EHCI (don't remember if I actually did any networking with it), but I don't remember if I saw the same eternal-attach with UHCI -- if so, I probably put it down to incompatibility of the adapter with USB1. One more note, if I remember, the if_axe code has been added to DragonFly, though I haven't gotten around to check if that code is cleaner than my hack, in case anyone is looking to add if_axe to 4.11. (And this is really the wrong place to ask, but are there any plans -- if not already done -- to merge the not-USB if_re for gigabit ethernet, now that those cards are dirt- cheap, into 4.11? I've done hacks that I think get it to work for me, and I think again DFly has added if_re as well.) > The attached patch restores the tsleep hack and fixes a couple of other > problems I found while looking into this: [ ... ] > - Fix ioctl handling to return EINVAL for commands that don't make sense > for this device. Otherwise ifconfig(8) prints a lot of irrelevant data, > such as wireless settings. I also had to hack this in if_axe and 4.x, FWIW... If I don't find that this has all already been done, I'll give a report about my hacks, if there's interest in adding if_axe to 4.11. thanks barry bouwsma (oh, did I ask about merging if_axe into 4.x at all? ... )
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200412011419.iB1EJDs10868>