Date: Sun, 11 Dec 2005 15:14:02 -0500 (EST) From: "Michael L. Squires" <mikes@siralan.org> To: freebsd-stable@freebsd.org Subject: Toshiba 8100 Fix (was Re: 5.4-STABLE changes breaks IDE boot (was Re: 5.3 -> 5.4 breaks ATA (Intel ICH2)) Message-ID: <20051211145338.J56200@familysquires.net> In-Reply-To: <20050925142016.L7868@familysquires.net> References: <87y85nuqhy.fsf@beaker.data-secure.net> <4335D1D2.9060501@leadhill.net> <87r7bdt6o3.fsf@beaker.data-secure.net> <20050925142016.L7868@familysquires.net>
next in thread | previous in thread | raw e-mail | index | archive | help
I have a Toshiba 8100 notebook which runs 5.4-STABLE. Kernels compiled
after some time in September 2005 fail to boot; the machine locks up right
after the ata controller is called in the boot messages.
I have traced the problem to a change to /usr/src/sys/dev/pci.c, (see
below from my cvsup log). I found the problem by cvsup'ing /usr/src for
dates starting with my last working kernel and once I had the date of
problem code tied down to a couple of days adding in each update to see if
it caused the failure to boot. I then added the fix to the latest 5.4
sources to make sure that the fix still worked.
The problem (really the Toshiba 8100 BIOS problem) was
triggered by:
Edit src/sys/dev/pci/pci.c
Add delta 1.264.2.12 2005.09.05.18.09.20 imp
which changes line 835 of pci.c from
if (base == 0)
to
if (base == 0 || map == testval)
The sysctl parameter "hw.pci.enable_io_modes" is supposed to allow PCs
with "lazy" BIOSes to boot, but in the case of the 8100 setting
hw.pci.enable_io_modes: 0
doesn't allow the system to boot.
This is obviously a problem with the Toshiba BIOS, which I know from other
times has other problems; I am putting this here in case someone else has
the same problem and wants to boot FreeBSD 5.4-STABLE with a kernel
revision after 9/5/2005.
Mike Squires
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051211145338.J56200>