xen-ia64-devel
Re: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable
So you're talking problem from when booting domU? You're right about the
x86 behavior, but that's different on current ia64 model. Dan prefers to
adopt a transparent virtualization mechanism, so minimal changes are
required to touch common linux code. If we also place those system pages
after the kernel image, you have to modify efi syb-system to ignore
these pages.
Currently I just put the last page allocated to xenU as the store page,
not in the middle. But yes, there's still one problem in dom_fw_init,
where all allocated pages are exposed to domain. Instead we should
reserve the last page and let store page out of domain's scope. Check
whether following changes solves your issue:
diff -r 8799d14bef77 xen/arch/ia64/dom_fw.c
--- a/xen/arch/ia64/dom_fw.c Thu Aug 25 22:53:20 2005
+++ b/xen/arch/ia64/dom_fw.c Fri Sep 2 10:38:11 2005
@@ -512,6 +512,11 @@
return 0;
}
*/
+
+ /* Last page is for xenstore, and not exported to domain */
+ if (d != dom0)
+ maxmem = (d->max_pages - 1) * PAGE_SIZE;
+
memset(fw_mem, 0, fw_mem_size);
#ifdef XEN
Thanks,
Kevin
With all the various patches, things work fine and I get a login prompt.
I knew the x86 model was different, what I didn't know was what you did
with this patch was legal. It looks to me as if you have the operating
system accessing a page the VM really shouldn't let it have. I was
trying to figure out how to reserve the page in the EFI memory map: add
it to the end of the hypercall page or something. Maybe its fortunate I
had to go home so you could come up with the more straightforward fix.
Spoke a little too soon, again. Looks like dom0 just crashed hard:
(XEN) +01+01Heartbeat... iip=f0000000040092d0,psr.i=0,pend=1
(XEN) +01ia64_fault: General Exception: IA-64 Reserved Register/Field
fault (data access): reflecting
I've no clue what the problem might be, things were idle. Still, things
are better.
Thanks,
John
_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel
|
<Prev in Thread] |
Current Thread |
[Next in Thread>
|
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, (continued)
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Magenheimer, Dan (HP Labs Fort Collins)
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- Re: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable,
John Byrne <=
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Magenheimer, Dan (HP Labs Fort Collins)
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Magenheimer, Dan (HP Labs Fort Collins)
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Magenheimer, Dan (HP Labs Fort Collins)
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Magenheimer, Dan (HP Labs Fort Collins)
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
- RE: [Xen-ia64-devel] Community effort neededtocatch upwithxen-unstable, Tian, Kevin
|
|
|