On Wed, 2007-03-14 at 21:04 +0900, Isaku Yamahata wrote:
[PATCH][RFC]
fix ia64 grant_table destruction.
fix type count of grant mapped page for xen/ia64.
This patch is necessary to keep page_info.type_info
correctly. However I don't think that this patch is acceptable.
Another idea is to modify guest_physmap_remove_page() be aware
of grant table mapped page.
Hi Isaku,
I agree, this doesn't feel like the best solution. It's interesting
that the comment on cset 8409 (that commented out the line this patch
adds back) adds gnttab_release_mappings() after
domain_relinquish_resources(). Makes me wonder if the implicit put_page
comment is still entirely valid for x86. Thanks,
Alex