Date: Tue, 04 Feb 2014 10:53:32 -0700 From: Ian Lepore <ian@FreeBSD.org> To: Iain Young <iain@g7iii.net> Cc: freebsd-arm@FreeBSD.org Subject: Re: Turning WITNESS off on Beaglebone Black Message-ID: <1391536412.1196.4.camel@revolution.hippie.lan> In-Reply-To: <52F09836.5070505@g7iii.net> References: <52F09836.5070505@g7iii.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 2014-02-04 at 07:35 +0000, Iain Young wrote: > Hi Folks, > > Anyone else having issues on the Beaglebone Black when disabling > WITNESS ? This is on 11.0-CURRENT, r261200. I saw it a few weeks > ago, as well, but got busy and decided to wait for the snapshots > and try again. > > On mine, it gets as far as this line, and waits a few tens of seconds: > > ustorage_fs0: <USB Mass Storage Interface> on usbus0 > > And then prints this: > > random: unblocking device. > > > Followed by absolutely nothing until I reboot. On a WITNESS > enabled kernel, it then proceeds to warn me about WITNESS and > performance, before going on to mount the root file system, and > bring up user space: > > WARNING: WITNESS option enabled, expect reduced performance. > Trying to mount root from ufs:/dev/mmcsd0s2a [rw,noatime]... > > > Curiously, it looks like the non-WITNESS kernel is also not > finding the onboard MMC, so that might be relevant. In my > case, I don't actually use it, so if it doesn't see it, it's > not an issue for me > > Anyone know what I'm not doing right ? I -thought- all I had > to do to turn WITNESS off was comment out the options line > in my conf file, and I'm sure it worked on the BBW a year or > so back. > > I've attached the two dmesg outputs as well as the conf > used, but as I say, it's stock, with the exception of > WITNESS being turned off. > > > Iain > > PS, About a year ago, we mentioned IEEE1588 and the cpsw driver, > a quick grep through /usr/src/sys showed the following drivers > had the "IEEE1588" string in them: > > e1000 (/dev/e1000/*.h) > IXGBE (/dev/ixgbe/*) > NLM (/dev/mipd/nlm/board.[cv] > > No idea how functional each one is, but if anyone was thinking > about adding 1588 support to the cpsw, might be useful - Unfortunately > it's a little beyond my skillset I can't reproduce this on a BBW at the same os rev, but I use an nfs root, which is probably a significant difference. The fact that yours is failing to find the eMMC in the fail case may be significant too... if the sdhci controller is hung up somehow you'd fail to mount root (but then why doesn't that time out and give an error?) -- Ian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1391536412.1196.4.camel>