On Thu, 6 Jan 2011, Stanislaw Gruszka wrote:

On Wed, Jan 05, 2011 at 10:21:21PM -0800, Russell Whitaker wrote:

quoted

quoted

For some unknown reason resources needed by amd76xrom driver can not be
available. Instead of return error, driver crash the kernel with
messages like below. Patch fix that.

amd76xrom" amd76xrom_init_one(): Unable to register resource 0x00000000ffc00000-0x00000000ffffffff - kernel bug?

Sorry, in my case that's the wrong bug. First found the bug using
Slackware-current. So I stripped down his config file by removing
most of what I can't use, recompiled, and the bug was still there.

I'm not using amd76xrom so went into a MTD submenu and changed it to
not set, etc, and found the bug is still there.

I do not see any other problems except that one:

	BUG: unable to handle kernel paging request at f862aaa0
	IP: [<f8120282>] simple_map_write+0x82/0xbb [map_funcs]

in dmesg's you provided in the first email. That BUG is from amd75xrom
driver. Does dmesg without MTD compiled show something different?

The " - kernel bug?" line is gone.

And this is also not clear for me from your first email, crash happened
every time with pata or randomly?

There are two bugs. In order to reduce the problem to only one bug I
complied 2.6.35.10-smp without MTD's 4 drivers where the name starts
with "bios" and also compiled with pata.

Result, first four attempts at booting: oops 3, success 1.

To verify the problem is in pata I then replaced the pata code with the
ATA (depreciated) code, with no other changes, and recompiled.

Result, all attempts at booting: success every time.

(note: changing the kernel also means changing /etc/fstab to match)

   Russ

Stanislaw