The following cset from xen-3.1-testing.hg accomodates PAE guests that don't support extended-cr3 format. Not entirely sure what the implications of this are but it sounds like something that wouldn't hurt to add to RHEL 5.2 15043 - Ensure that the L3 page table page for a PAE guest which does not support the extended-cr3 format is below the 4G boundary.
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux maintenance release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Update release for currently deployed products. This request is not yet committed for inclusion in an Update release.
This request was previously evaluated by Red Hat Product Management for inclusion in the current Red Hat Enterprise Linux release, but Red Hat was unable to resolve it in time. This request will be reviewed for a future Red Hat Enterprise Linux release.
Well, I have spent some time now looking for some kind of guests not supporting extended-cr3 format but no luck yet. Could anybody guide me so that I can test it after backporting from upstream? I can't test it right now because I lack information about guests not supporting extended-cr3 format and unfortunatelly googling it find nothing relevant for this issue. Anyway if somebody could replicate this and write me steps to replicate, including architecture (i386, x86_64, IA64) and some logs there it would help me much. Thanks.
I would say that if you can't find any info about it online, and no one has explicitly reported a bug, and we've done fine thus far, we just close this as WONTFIX. This was left over from us scouring all xen stable csets between 5.1 and 5.2. This patch sounded scary and important so it was nominated (without really any idea what it entails), but if there isn't any info to be found, it's likely not a huge deal.
Cole, Agreed. I think you would have to look at a NetBSD, FreeBSD, Solaris, or *very* old PV guest (older than any of our RHEL guests) to find something that would need this. With that being the case, I would say WONTFIX; if a customer later on needs this functionality (highly unlikely), then we can open a new bug. Chris Lalancette
Ok, thank you both for advice, closing with WONTFIX reason... Michal Novotny