On Thu, Mar 19, 2015 at 06:22:06PM +0100, Juergen Gross wrote:
> On 03/19/2015 05:21 PM, Daniel Kiper wrote:
> >On Thu, Mar 19, 2015 at 03:31:02PM +0100, Juergen Gross wrote:
> >>Commit 25b884a83d487fd62c3de7ac1ab5549979188482 ("x86/xen: set
> >>regions above the end of RAM as 1:1") introduced a regression.
> >>
> >>To be able to add memory pages which were added via memory hotplug to
> >>a pv domain, the pages must be "invalid" instead of "identity" in the
> >>p2m list before they can be added.
> >>
> >>Suggested-by: David Vrabel <david.vra...@citrix.com>
> >>Signed-off-by: Juergen Gross <jgr...@suse.com>
> >>---
> >> drivers/xen/balloon.c | 13 +++++++++++--
> >> 1 file changed, 11 insertions(+), 2 deletions(-)
> >>
> >>diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c
> >>index 0b52d92..52e331f 100644
> >>--- a/drivers/xen/balloon.c
> >>+++ b/drivers/xen/balloon.c
> >>@@ -221,15 +221,24 @@ static bool balloon_is_inflated(void)
> >>
> >> static enum bp_state reserve_additional_memory(long credit)
> >> {
> >>- int nid, rc;
> >>+ int nid, rc = 0;
> >> u64 hotplug_start_paddr;
> >> unsigned long balloon_hotplug = credit;
> >>+ unsigned long pfn;
> >>
> >> hotplug_start_paddr = PFN_PHYS(SECTION_ALIGN_UP(max_pfn));
> >> balloon_hotplug = round_up(balloon_hotplug, PAGES_PER_SECTION);
> >> nid = memory_add_physaddr_to_nid(hotplug_start_paddr);
> >>
> >>- rc = add_memory(nid, hotplug_start_paddr, balloon_hotplug <<
> >>PAGE_SHIFT);
> >>+ for (pfn = PFN_DOWN(hotplug_start_paddr);
> >>+ !rc && pfn < PFN_DOWN(hotplug_start_paddr) + balloon_hotplug;
> >>+ pfn++)
> >>+ if (!set_phys_to_machine(pfn, INVALID_P2M_ENTRY))
> >
> >rc = set_phys_to_machine(pfn, INVALID_P2M_ENTRY)?
>
> Not really. set_phys_to_machine returns false on failure...
>
> >
> >>+ rc = 1;
> >
> >I do not think that this stuff is needed for HVM or PVH guests.
>
> True.
>
> >
> >>+ if (!rc)
> >>+ rc = add_memory(nid, hotplug_start_paddr,
> >>+ balloon_hotplug << PAGE_SHIFT);
> >>
> >> if (rc) {
> >> pr_warn("Cannot add additional memory (%i)\n", rc);
> >
> >It will be nice to know what part of infrastructure failed.
> >Could you create separate pr_warn() message for set_phys_to_machine()?
>
> Value 1 for rc is the indicator for that case.
Well... Personally I prefer explicit messages telling what happened than
something which requires digging in a code to understand a problem.
Additionally, I think that we should use negative numbers (as David
suggested) to signal an error. Most of kernel stuff work in that way.
Daniel
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
