5 messages in com.xensource.lists.xen-ia64-devel[Xen-ia64-devel] Good projects for so...| From | Sent On | Attachments |
|---|---|---|
| Magenheimer, Dan (HP Labs Fort Collins) | 22 Dec 2005 10:33 | |
| Tristan Gingold | 03 Jan 2006 03:11 | |
| Tristan Gingold | 03 Jan 2006 08:08 | .diffs |
| Magenheimer, Dan (HP Labs Fort Collins) | 03 Jan 2006 11:58 | .bootlog |
| Tristan Gingold | 04 Jan 2006 02:16 |
| Subject: | [Xen-ia64-devel] Good projects for someone![]() |
|---|---|
| From: | Magenheimer, Dan (HP Labs Fort Collins) (dan....@hp.com) |
| Date: | 12/22/2005 10:33:30 AM |
| List: | com.xensource.lists.xen-ia64-devel |
Here's a couple of projects that would be useful for someone to work on:
1) While working on the domU fix earlier in the week, I discovered that Xen/ia64 still ignores any machine memory at an address > 4GB (see xen/arch/ia64/linux-xen/efi.c and search for "hack"). I think nearly all Itanium machines leave some portion of 1GB-4GB unpopulated so most memory is getting wasted! I think CONFIG_VIRTUAL_MEM_MAP needs to be implemented to fix this. (Removing the hack causes a "Not enough memory for frametable" panic.)
2) Turning CONFIG_SMP on doesn't currently work because the (new) routine flush_tlb_mask is unimplemented and panics. Once SMP is working again, perhaps we should turn it on by default?
Also, I still periodically see a problem in dom0 (and I think the same problem in domU) every few HOURS of execution, for example only once in 20 Linux compiles. This problem manifests as a bad pointer dereference, sometimes zero and sometimes a random number. This kills the current process, usually with a segmentation error. I suspect that this is a stale mapping problem, but do not know for sure.
Tracking this down will take a lot of time and some very detailed instrumentation (printfs or hardware debugger), but finding and fixing it would be very helpful.
Dan
_______________________________________________ Xen-ia64-devel mailing list Xen-...@lists.xensource.com http://lists.xensource.com/xen-ia64-devel





.diffs