Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 30 Aug 2005 12:23:20 -0400
From:      Mikhail Teterin <mi+mx@aldan.algebra.com>
To:        Miguel Mendez <flynn@energyhq.es.eu.org>
Cc:        re@freebsd.org, current@freebsd.org
Subject:   Re: installing 6.0-BETA3
Message-ID:  <200508301223.20626.mi%2Bmx@aldan.algebra.com>
In-Reply-To: <20050829191927.6694a969.flynn@energyhq.es.eu.org>
References:  <200508281326.j7SDQGfc073329@blue.virtual-estates.net> <20050829191927.6694a969.flynn@energyhq.es.eu.org>

next in thread | previous in thread | raw e-mail | index | archive | help
> > 	. Something is wrong with the ed-driver (in my case --
> > 	  Kingston's KNE-PC2T, FCC ID L40WST200). Kernel reports
> > 	  being unable to do something with pccard1, then reports
> > 		  ed1: ....
> > 	  and hangs. I let it wait for an hour and it never continued
> > 	  booting. I'll try to investigate, when I'm done installing.
> > 	  The card worked just fine in a newer Pentium-II laptop under
> > 	  FreeBSD-5.x
>
> If the card is a 16bit pcmcia car make sure the bridge is set to 16bit.

How do I do that?

> I had a similar problem on an old Thinkpad when trying to use a 32bit
> cardbus nic while the bridge was set to default to 16bit in the BIOS.

Both -- the hanging ed1 and the working wi0 are 16bit 5volt cards, according 
to their labeling. Prior to the ed1-hang, there is a message about 32 cfe 
allocation failure.

> I don't see any advantage in using the partition-less mode.

Less layers, and (slightly) more diskspace -- just feels neater, I guess. But 
the BIOS would not boot, so out with it in this case.

> I wouldn't blame WITNESS on this. Laptops, especially old ones, have
> incredibly slow disks. The buildworld process is heavily i/o bounded,
> my bet is on the disk subsystem.

According to systat and top, the CPU is never idle -- as would've been the 
case, if the I/O were the limiting factor. The Sys-component of `systat -vm'
was steadily above 50%.

Now that I have a witness-less kernel, the builds are, indeed, I/O bound.

	-mi



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200508301223.20626.mi%2Bmx>