[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH v1 6/8]: PVH ballooning and grant changes
On Fri, 2012-09-21 at 20:19 +0100, Mukesh Rathor wrote: > --- > drivers/xen/balloon.c | 35 ++++++++++++++++++++++++++++------- > drivers/xen/gntdev.c | 3 ++- > drivers/xen/grant-table.c | 25 +++++++++++++++++++++---- > 3 files changed, 51 insertions(+), 12 deletions(-) > > diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c > index 31ab82f..85a6917 100644 > --- a/drivers/xen/balloon.c > +++ b/drivers/xen/balloon.c > @@ -358,10 +358,21 @@ static enum bp_state increase_reservation(unsigned long > nr_pages) > BUG_ON(!xen_feature(XENFEAT_auto_translated_physmap) && > phys_to_machine_mapping_valid(pfn)); > > - set_phys_to_machine(pfn, frame_list[i]); > + if (xen_pv_domain() && > + xen_feature(XENFEAT_auto_translated_physmap)) { > + /* PVH: we just need to update native page table */ Do you not need a PageHighMem check here somewhere? (although this is moot if you follow the suggestion below) > + pte_t *ptep; > + unsigned int level; > + void *addr = __va(pfn << PAGE_SHIFT); > + ptep = lookup_address((unsigned long)addr, &level); > + set_pte(ptep, pfn_pte(pfn, PAGE_KERNEL)); > + } else > + set_phys_to_machine(pfn, frame_list[i]); > > /* Link back into the page tables if not highmem. */ > - if (xen_pv_domain() && !PageHighMem(page)) { > + if (xen_pv_domain() && !PageHighMem(page) && > + !xen_feature(XENFEAT_auto_translated_physmap)) { I think it would be more logical to put the PVH PT update here next to the PV PT update and gate the set_phys_to_machine above on !a_t_pm, like how you've done it in the decrease case. (or perhaps per your comment below set_phys_to_machine isa NOP also on PVH?) > + > int ret; > ret = HYPERVISOR_update_va_mapping( > (unsigned long)__va(pfn << PAGE_SHIFT), > @@ -418,12 +429,21 @@ static enum bp_state decrease_reservation(unsigned long > nr_pages, gfp_t gfp) > scrub_page(page); > > if (xen_pv_domain() && !PageHighMem(page)) { > - ret = HYPERVISOR_update_va_mapping( > - (unsigned long)__va(pfn << PAGE_SHIFT), > - __pte_ma(0), 0); > - BUG_ON(ret); > + if (xen_feature(XENFEAT_auto_translated_physmap)) { > + unsigned int level; > + pte_t *ptep; > + void *addr = __va(pfn << PAGE_SHIFT); > + ptep = lookup_address((unsigned long)addr, > + &level); > + set_pte(ptep, __pte(0)); > + > + } else { > + ret = HYPERVISOR_update_va_mapping( > + (unsigned long)__va(pfn << PAGE_SHIFT), > + __pte_ma(0), 0); > + BUG_ON(ret); > + } > } > - > } > > /* Ensure that ballooned highmem pages don't have kmaps. */ > @@ -433,6 +453,7 @@ static enum bp_state decrease_reservation(unsigned long > nr_pages, gfp_t gfp) > /* No more mappings: invalidate P2M and add to balloon. */ > for (i = 0; i < nr_pages; i++) { > pfn = mfn_to_pfn(frame_list[i]); > + /* PVH note: following will noop for auto translated */ > __set_phys_to_machine(pfn, INVALID_P2M_ENTRY); > balloon_append(pfn_to_page(pfn)); > } > diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c > index 1ffd03b..35d1e3c 100644 > --- a/drivers/xen/gntdev.c > +++ b/drivers/xen/gntdev.c > @@ -802,7 +802,8 @@ static int __init gntdev_init(void) > if (!xen_domain()) > return -ENODEV; > > - use_ptemod = xen_pv_domain(); > + use_ptemod = xen_pv_domain() && > + !xen_feature(XENFEAT_auto_translated_physmap); I keep wondering if #define XENFEAT(x) xen_feature(XENFEAT_#x) might not be useful to keep some of these long lines under control. Probably not, since the bulk is the "x" bit... Ian. _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxx http://lists.xen.org/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |