Red Hat Bugzilla – Bug 445348
Please port the mtrr fixes from 2.4.25 so that some hardware can boot with >4GiB memory
Last modified: 2008-05-29 08:28:32 EDT
Description of problem:
I have an Intel G965 based motherboard from Asus named P5B-VM whose BIOS needs a MTRR
workaround from recent mainline kernels to be able to boot when I install more than 4GiB of memory.
I reported this problem for Fedora 8 as #438960 and it was promptly fixed, but now I need to run RHEL
5 to do some work related stuff on my machine and the update kernel 2.6.18-53.1.14.el5 exhibits the
same problems as fedora kernels did before kernel-18.104.22.168-64.fc8
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. get my motherboard
2. install more than 4GiB of memory
3. try to boot
the boot grinds to a halt, as predicted by the upstream fix author (referenced in the original bug).
a regular boot, as exhibited when 4GiB memory or less is installed.
as stated above, the fix from 2.6.25 backported to the 2.6.24 based fedora kernel is available from https://bugzilla.redhat.com/show_bug.cgi?id=438960 as well as pointers to the relevant commits to
the upstream tree.
If you need help testing proposed patches to verify that it solves the problem, I'm more than willing to
help out with that.
Created attachment 305161 [details]
mtrr change from bug 438960
patch from bug 438960 comment#17 by djones.
This is the take2 patch that went into fedora 8.
Please be advised, that Bugzilla is not a support tool. It is an
Engineering and Community tool. So although all changes to Enterprise
Linux go through Bugzilla and Red Hat considers issues directly entered
into Bugzilla valuable feedback, there is no SLA around it.
If this is a production issue, please report it to your Red Hat Support contact.
I am advised. However, my guess is that my hardware is not all that exotic, the fix is nicely contained and
suitable for backporting IMO so integrating it into the RHEL5 kernel would be something that RH would
want to do.
After all, the next customer that hits this problem might not be able to point you to the relevant fix and if
you proactively address this issue now, your own time and customers time might be saved later on.
Well well, in a couple of years we'll have a RH release based on a later upstream kernel (and I will have
new hardware) and inte meantime I'll be able to use Fedora and live with the mismatch between our
development and production environment.