Re: [Xen-devel] [Patch V2 10/15] xen: check pre-allocated page tables for conflict with memory map

[Juergen Gross <jgross@xxxxxxxx>]

On 04/09/2015 02:47 PM, David Vrabel wrote:
> On 09/04/2015 07:55, Juergen Gross wrote:
> > Check whether the page tables built by the domain builder are at
> > memory addresses which are in conflict with the target memory map.
> > If this is the case just panic instead of running into problems
> > later.
> >
> > Signed-off-by: Juergen Gross <jgross@xxxxxxxx>
> > ---
> >   arch/x86/xen/mmu.c     | 19 ++++++++++++++++---
> >   arch/x86/xen/setup.c   |  6 ++++++
> >   arch/x86/xen/xen-ops.h |  1 +
> >   3 files changed, 23 insertions(+), 3 deletions(-)
> >
> > diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
> > index 1ca5197..41aeb1c 100644
> > --- a/arch/x86/xen/mmu.c
> > +++ b/arch/x86/xen/mmu.c
> > @@ -116,6 +116,7 @@ static pud_t level3_user_vsyscall[PTRS_PER_PUD] 
> > __page_aligned_bss;
> >   DEFINE_PER_CPU(unsigned long, xen_cr3);        /* cr3 stored as physaddr */
> >   DEFINE_PER_CPU(unsigned long, xen_current_cr3);        /* actual vcpu cr3 */
> >
> > +static phys_addr_t xen_pt_base, xen_pt_size;
> These be __init, but the use of globals in this way is confusing.
How else would you want to do it?

> >   /*
> >    * Just beyond the highest usermode address.  STACK_TOP_MAX has a
> > @@ -1998,7 +1999,9 @@ void __init xen_setup_kernel_pagetable(pgd_t *pgd, 
> > unsigned long max_pfn)
> >                 check_pt_base(&pt_base, &pt_end, addr[i]);
> >
> >         /* Our (by three pages) smaller Xen pagetable that we are using */
> > -       memblock_reserve(PFN_PHYS(pt_base), (pt_end - pt_base) * PAGE_SIZE);
> > +       xen_pt_base = PFN_PHYS(pt_base);
> > +       xen_pt_size = (pt_end - pt_base) * PAGE_SIZE;
> > +       memblock_reserve(xen_pt_base, xen_pt_size);
> Why not provide a xen_memblock_check_and_reserve() call that has the
> xen_is_e820_reserved() check and the memblock_reserve() call?  This may
> also be useful for patch #9 as well.
Uuh, not really. memblock_reserve() for those areas is called much
earlier than the e820 map is constructed.

Thinking more about it, I even have to modify patch 11 and 13:
relocation must be done _after_ doing the memblock_reserve() of all
pre-populated areas to avoid relocating to such an area.


Juergen


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel