On Thu 03-01-19 15:59:52, Roman Penyaev wrote: > area->size can include adjacent guard page but get_vm_area_size() > returns actual size of the area. > > This fixes possible kernel crash when userspace tries to map area > on 1 page bigger: size check passes but the following vmalloc_to_page() > returns NULL on last guard (non-existing) page. > > Signed-off-by: Roman Penyaev <[email protected]> > Cc: Andrew Morton <[email protected]> > Cc: Michal Hocko <[email protected]> > Cc: Andrey Ryabinin <[email protected]> > Cc: Joe Perches <[email protected]> > Cc: "Luis R. Rodriguez" <[email protected]> > Cc: [email protected] > Cc: [email protected] > Cc: [email protected]
Forgot to add Acked-by: Michal Hocko <[email protected]> Although I am not really sure the stable backport is really needed as I haven't seen any explicit example of a buggy kernel code to trigger the issue. > --- > mm/vmalloc.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/mm/vmalloc.c b/mm/vmalloc.c > index 871e41c55e23..2cd24186ba84 100644 > --- a/mm/vmalloc.c > +++ b/mm/vmalloc.c > @@ -2248,7 +2248,7 @@ int remap_vmalloc_range_partial(struct vm_area_struct > *vma, unsigned long uaddr, > if (!(area->flags & VM_USERMAP)) > return -EINVAL; > > - if (kaddr + size > area->addr + area->size) > + if (kaddr + size > area->addr + get_vm_area_size(area)) > return -EINVAL; > > do { > -- > 2.19.1 -- Michal Hocko SUSE Labs
