On Fri, 22 Nov 2019 at 12:26, <glider@xxxxxxxxxx> wrote: > > KMSAN is going to use 3/4 of existing vmalloc space to hold the > metadata, therefore we lower VMALLOC_END to make sure vmalloc() doesn't > allocate past the first 1/4. > > Signed-off-by: Alexander Potapenko <glider@xxxxxxxxxx> > To: Alexander Potapenko <glider@xxxxxxxxxx> > Cc: Vegard Nossum <vegard.nossum@xxxxxxxxxx> > Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> > Cc: Dmitry Vyukov <dvyukov@xxxxxxxxxx> > Cc: linux-mm@xxxxxxxxx > > --- > > Change-Id: Iaa5e8e0fc2aa66c956f937f5a1de6e5ef40d57cc > --- > arch/x86/include/asm/pgtable_64_types.h | 15 +++++++++++++++ > 1 file changed, 15 insertions(+) > > diff --git a/arch/x86/include/asm/pgtable_64_types.h b/arch/x86/include/asm/pgtable_64_types.h > index 52e5f5f2240d..586629e20436 100644 > --- a/arch/x86/include/asm/pgtable_64_types.h > +++ b/arch/x86/include/asm/pgtable_64_types.h > @@ -139,7 +139,22 @@ extern unsigned int ptrs_per_p4d; > # define VMEMMAP_START __VMEMMAP_BASE_L4 > #endif /* CONFIG_DYNAMIC_MEMORY_LAYOUT */ > > +#ifndef CONFIG_KMSAN I think it might be more readable if this was non-negative, i.e. '#ifdef CONFIG_KMSAN'. > #define VMALLOC_END (VMALLOC_START + (VMALLOC_SIZE_TB << 40) - 1) > +#else > +/* > + * In KMSAN builds vmalloc area is four times smaller, and the remaining 3/4 > + * are used to keep the metadata for virtual pages. > + */ > +#define VMALLOC_QUARTER_SIZE ((VMALLOC_SIZE_TB << 40) >> 2) > +#define VMALLOC_END (VMALLOC_START + VMALLOC_QUARTER_SIZE - 1) > +#define VMALLOC_SHADOW_OFFSET VMALLOC_QUARTER_SIZE > +#define VMALLOC_ORIGIN_OFFSET (VMALLOC_QUARTER_SIZE * 2) > +#define VMALLOC_META_END (VMALLOC_END + VMALLOC_ORIGIN_OFFSET) > +#define MODULES_SHADOW_START (VMALLOC_META_END + 1) > +#define MODULES_ORIGIN_START (MODULES_SHADOW_START + MODULES_LEN) > +#define MODULES_ORIGIN_END (MODULES_ORIGIN_START + MODULES_LEN) Are all these, except VMALLOC_END, KMSAN specific? For readability and avoid further confusion, would it make sense to prefix these with 'KMSAN_' ? This file is for x86 only -- would other architectures define these similarly? If so, maybe some of this could be moved into a helper, such as include/asm-generic/kmsan_pgtable.h? Thanks, -- Marco > +#endif > > #define MODULES_VADDR (__START_KERNEL_map + KERNEL_IMAGE_SIZE) > /* The module sections ends with the start of the fixmap */ > -- > 2.24.0.432.g9d3f5f5b63-goog >
