| Message ID | 20190718065222.31310-2-jgross@suse.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | xen/gntdev: sanitize user interface handling | expand |
On 18/07/2019 07:52, Juergen Gross wrote: > Today there is a global limit of pages mapped via /dev/xen/gntdev set > to 1 million pages per default. The Xen default limit even for dom0 is 1024 pages * 16 entries per page, which is far lower than this limit. > There is no reason why that limit is > existing, as total number of foreign mappings is limited by the s/foreign/grant/ ? > hypervisor anyway and preferring kernel mappings over userspace ones > doesn't make sense. Its probably also worth stating that this a root-only device, which further brings in to question the user/kernel split. > > Additionally checking of that limit is fragile, as the number of pages > to map via one call is specified in a 32-bit unsigned variable which > isn't tested to stay within reasonable limits (the only test is the > value to be <= zero, which basically excludes only calls without any > mapping requested). So trying to map e.g. 0xffff0000 pages while > already nearly 1000000 pages are mapped will effectively lower the > global number of mapped pages such that a parallel call mapping a > reasonable amount of pages can succeed in spite of the global limit > being violated. > > So drop the global limit and introduce per call limit instead. Its probably worth talking about this new limit. What is it trying to protect? ~Andrew
On 18.07.19 19:36, Andrew Cooper wrote: > On 18/07/2019 07:52, Juergen Gross wrote: >> Today there is a global limit of pages mapped via /dev/xen/gntdev set >> to 1 million pages per default. > > The Xen default limit even for dom0 is 1024 pages * 16 entries per page, > which is far lower than this limit. Actually its 256 entries per page, but this is still lower than the current limit. > >> There is no reason why that limit is >> existing, as total number of foreign mappings is limited by the > > s/foreign/grant/ ? Can do. > >> hypervisor anyway and preferring kernel mappings over userspace ones >> doesn't make sense. > > Its probably also worth stating that this a root-only device, which > further brings in to question the user/kernel split. Yes. > >> >> Additionally checking of that limit is fragile, as the number of pages >> to map via one call is specified in a 32-bit unsigned variable which >> isn't tested to stay within reasonable limits (the only test is the >> value to be <= zero, which basically excludes only calls without any >> mapping requested). So trying to map e.g. 0xffff0000 pages while >> already nearly 1000000 pages are mapped will effectively lower the >> global number of mapped pages such that a parallel call mapping a >> reasonable amount of pages can succeed in spite of the global limit >> being violated. >> >> So drop the global limit and introduce per call limit instead. > > Its probably worth talking about this new limit. What is it trying to > protect? Out-of-bounds allocations. Juergen
diff --git a/drivers/xen/gntdev-common.h b/drivers/xen/gntdev-common.h index 2f8b949c3eeb..0e5d4660e7b8 100644 --- a/drivers/xen/gntdev-common.h +++ b/drivers/xen/gntdev-common.h @@ -87,7 +87,7 @@ void gntdev_add_map(struct gntdev_priv *priv, struct gntdev_grant_map *add); void gntdev_put_map(struct gntdev_priv *priv, struct gntdev_grant_map *map); -bool gntdev_account_mapped_pages(int count); +bool gntdev_test_page_count(unsigned int count); int gntdev_map_grant_pages(struct gntdev_grant_map *map); diff --git a/drivers/xen/gntdev-dmabuf.c b/drivers/xen/gntdev-dmabuf.c index 2c4f324f8626..63f0857bf62d 100644 --- a/drivers/xen/gntdev-dmabuf.c +++ b/drivers/xen/gntdev-dmabuf.c @@ -446,7 +446,7 @@ dmabuf_exp_alloc_backing_storage(struct gntdev_priv *priv, int dmabuf_flags, { struct gntdev_grant_map *map; - if (unlikely(count <= 0)) + if (unlikely(gntdev_test_page_count(count))) return ERR_PTR(-EINVAL); if ((dmabuf_flags & GNTDEV_DMA_FLAG_WC) && @@ -459,11 +459,6 @@ dmabuf_exp_alloc_backing_storage(struct gntdev_priv *priv, int dmabuf_flags, if (!map) return ERR_PTR(-ENOMEM); - if (unlikely(gntdev_account_mapped_pages(count))) { - pr_debug("can't map %d pages: over limit\n", count); - gntdev_put_map(NULL, map); - return ERR_PTR(-ENOMEM); - } return map; } @@ -771,7 +766,7 @@ long gntdev_ioctl_dmabuf_exp_from_refs(struct gntdev_priv *priv, int use_ptemod, if (copy_from_user(&op, u, sizeof(op)) != 0) return -EFAULT; - if (unlikely(op.count <= 0)) + if (unlikely(gntdev_test_page_count(op.count))) return -EINVAL; refs = kcalloc(op.count, sizeof(*refs), GFP_KERNEL); @@ -818,7 +813,7 @@ long gntdev_ioctl_dmabuf_imp_to_refs(struct gntdev_priv *priv, if (copy_from_user(&op, u, sizeof(op)) != 0) return -EFAULT; - if (unlikely(op.count <= 0)) + if (unlikely(gntdev_test_page_count(op.count))) return -EINVAL; gntdev_dmabuf = dmabuf_imp_to_refs(priv->dmabuf_priv, diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c index 4c339c7e66e5..23e21a9aedf7 100644 --- a/drivers/xen/gntdev.c +++ b/drivers/xen/gntdev.c @@ -57,12 +57,10 @@ MODULE_AUTHOR("Derek G. Murray <Derek.Murray@cl.cam.ac.uk>, " "Gerd Hoffmann <kraxel@redhat.com>"); MODULE_DESCRIPTION("User-space granted page access driver"); -static int limit = 1024*1024; -module_param(limit, int, 0644); +static unsigned int limit = 64*1024; +module_param(limit, uint, 0644); MODULE_PARM_DESC(limit, "Maximum number of grants that may be mapped by " - "the gntdev device"); - -static atomic_t pages_mapped = ATOMIC_INIT(0); + "one mapping request"); static int use_ptemod; #define populate_freeable_maps use_ptemod @@ -74,9 +72,9 @@ static struct miscdevice gntdev_miscdev; /* ------------------------------------------------------------------ */ -bool gntdev_account_mapped_pages(int count) +bool gntdev_test_page_count(unsigned int count) { - return atomic_add_return(count, &pages_mapped) > limit; + return !count || count > limit; } static void gntdev_print_maps(struct gntdev_priv *priv, @@ -244,8 +242,6 @@ void gntdev_put_map(struct gntdev_priv *priv, struct gntdev_grant_map *map) if (!refcount_dec_and_test(&map->users)) return; - atomic_sub(map->count, &pages_mapped); - if (map->notify.flags & UNMAP_NOTIFY_SEND_EVENT) { notify_remote_via_evtchn(map->notify.event); evtchn_put(map->notify.event); @@ -677,7 +673,7 @@ static long gntdev_ioctl_map_grant_ref(struct gntdev_priv *priv, if (copy_from_user(&op, u, sizeof(op)) != 0) return -EFAULT; pr_debug("priv %p, add %d\n", priv, op.count); - if (unlikely(op.count <= 0)) + if (unlikely(gntdev_test_page_count(op.count))) return -EINVAL; err = -ENOMEM; @@ -685,12 +681,6 @@ static long gntdev_ioctl_map_grant_ref(struct gntdev_priv *priv, if (!map) return err; - if (unlikely(gntdev_account_mapped_pages(op.count))) { - pr_debug("can't map: over limit\n"); - gntdev_put_map(NULL, map); - return err; - } - if (copy_from_user(map->grants, &u->refs, sizeof(map->grants[0]) * op.count) != 0) { gntdev_put_map(NULL, map);
Today there is a global limit of pages mapped via /dev/xen/gntdev set to 1 million pages per default. There is no reason why that limit is existing, as total number of foreign mappings is limited by the hypervisor anyway and preferring kernel mappings over userspace ones doesn't make sense. Additionally checking of that limit is fragile, as the number of pages to map via one call is specified in a 32-bit unsigned variable which isn't tested to stay within reasonable limits (the only test is the value to be <= zero, which basically excludes only calls without any mapping requested). So trying to map e.g. 0xffff0000 pages while already nearly 1000000 pages are mapped will effectively lower the global number of mapped pages such that a parallel call mapping a reasonable amount of pages can succeed in spite of the global limit being violated. So drop the global limit and introduce per call limit instead. Signed-off-by: Juergen Gross <jgross@suse.com> --- drivers/xen/gntdev-common.h | 2 +- drivers/xen/gntdev-dmabuf.c | 11 +++-------- drivers/xen/gntdev.c | 22 ++++++---------------- 3 files changed, 10 insertions(+), 25 deletions(-)