On Tue, 28 May 2024 at 12:48, Alexander Potapenko <glider@xxxxxxxxxx> wrote: > > As noticed by Brian, KMSAN should not be zeroing the origin when > unpoisoning parts of a four-byte uninitialized value, e.g.: > > char a[4]; > kmsan_unpoison_memory(a, 1); > > This led to false negatives, as certain poisoned values could receive zero > origins, preventing those values from being reported. > > To fix the problem, check that kmsan_internal_set_shadow_origin() writes > zero origins only to slots which have zero shadow. > > Reported-by: Brian Johannesmeyer <bjohannesmeyer@xxxxxxxxx> > Link: https://lore.kernel.org/lkml/20240524232804.1984355-1-bjohannesmeyer@xxxxxxxxx/T/ > Fixes: f80be4571b19 ("kmsan: add KMSAN runtime core") > Signed-off-by: Alexander Potapenko <glider@xxxxxxxxxx> > --- > mm/kmsan/core.c | 15 +++++++++++---- > 1 file changed, 11 insertions(+), 4 deletions(-) > > diff --git a/mm/kmsan/core.c b/mm/kmsan/core.c > index cf2d70e9c9a5f..95f859e38c533 100644 > --- a/mm/kmsan/core.c > +++ b/mm/kmsan/core.c > @@ -196,8 +196,7 @@ void kmsan_internal_set_shadow_origin(void *addr, size_t size, int b, > u32 origin, bool checked) > { > u64 address = (u64)addr; > - void *shadow_start; > - u32 *origin_start; > + u32 *shadow_start, *origin_start; > size_t pad = 0; > > KMSAN_WARN_ON(!kmsan_metadata_is_contiguous(addr, size)); > @@ -225,8 +224,16 @@ void kmsan_internal_set_shadow_origin(void *addr, size_t size, int b, > origin_start = > (u32 *)kmsan_get_metadata((void *)address, KMSAN_META_ORIGIN); > > - for (int i = 0; i < size / KMSAN_ORIGIN_SIZE; i++) > - origin_start[i] = origin; > + /* > + * If the new origin is non-zero, assume that the shadow byte is also non-zero, > + * and unconditionally overwrite the old origin slot. > + * If the new origin is zero, overwrite the old origin slot iff the > + * corresponding shadow slot is zero. > + */ > + for (int i = 0; i < size / KMSAN_ORIGIN_SIZE; i++) { > + if (origin || !shadow_start[i]) > + origin_start[i] = origin; > + } Reviewed-by: Marco Elver <elver@xxxxxxxxxx>
