linux 2.4.19 and E820 problems CONFIRMED

[Ronald G. Minnich]

I have just confirmed that 2.4.19 won't work at all with the memory split.

I set remap_high to disable in cmos and the node is now workig correctly.

I think I'm going to have a CONFIG option to disable this completely, 
since if it gets turned on accidently I have an unbootable node. Comments?

ron

[steven james]

Greetings,

Unless/until the kernel can handle it, I agree.

G'day,
sjames

On Wed, 4 Dec 2002, Ronald G. Minnich wrote:
> I have just confirmed that 2.4.19 won't work at all with the memory split.
>
> I set remap_high to disable in cmos and the node is now workig correctly.
>
> I think I'm going to have a CONFIG option to disable this completely, 
> since if it gets turned on accidently I have an unbootable node. Comments?
>
> ron
>
> Linuxbios mailing list
> [e-mail]
> http://www.clustermatic.org/mailman/listinfo/linuxbios

steven james  writes:
> Greetings,
>
> Unless/until the kernel can handle it, I agree.

It works on 2.4.18 just fine.  So if it fails on 2.4.19 it is a kernel
regression.  I will take a look at 2.4.19 shortly, and see what the
confusion is.  But if this is a kernel regression, the kernel needs
fixing not LinuxBIOS.  The configuration is 100% valid.

Ron could you provide a little more detailed report of the practical
problem you are seeing?  Where does it blow up?  A console log of a
good and a bad bootup would be ideal. I want to make certain when I
attempt to reproduce this that I actually am seeing the problem you
are seeing.   Unless I borrow one of the nodes on Pink I cannot
exactly replicate what you are seeing...

Eric

"Ronald G. Minnich"  writes:
> I have just confirmed that 2.4.19 won't work at all with the memory split.
>
> I set remap_high to disable in cmos and the node is now workig correctly.
>
> I think I'm going to have a CONFIG option to disable this completely, 
> since if it gets turned on accidently I have an unbootable node. Comments?

A small one.  For the short term this is not necessary as the remap_high option
is not available in fallback mode. So you should still be able to boot problematical
machines.

Eric

[Ronald G. Minnich]

On 4 Dec 2002, Eric W. Biederman wrote:
> > I think I'm going to have a CONFIG option to disable this completely, 
> > since if it gets turned on accidently I have an unbootable node. Comments?
>
> A small one.  For the short term this is not necessary as the remap_high option
> is not available in fallback mode. So you should still be able to boot problematical
> machines.

OK I won't do that option after all, I think this can all be resolved in 
the short term.

ron

[Ronald G. Minnich]

until I can get a console log of good/bad bootup, can you send me dmesg 
output from 2.4.18 booting with the remap_high on?

thanks

ron