| Message ID | 1497915397-93805-18-git-send-email-keescook@chromium.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Jun 19, 2017 at 04:36:31PM -0700, Kees Cook wrote: > From: David Windsor <dave@nullcore.net> > > When a dentry name is short enough, it can be stored directly in > the dentry itself. These dentry short names, stored in struct > dentry.d_iname and therefore contained in the dentry_cache slab cache, > need to be coped to/from userspace. > > In support of usercopy hardening, this patch defines a region in > the dentry_cache slab cache in which userspace copy operations > are allowed. > > This region is known as the slab cache's usercopy region. Slab > caches can now check that each copy operation involving cache-managed > memory falls entirely within the slab's usercopy region. > > This patch is modified from Brad Spengler/PaX Team's PAX_USERCOPY > whitelisting code in the last public patch of grsecurity/PaX based on my > understanding of the code. Changes or omissions from the original code are > mine and don't reflect the original grsecurity/PaX code. > For all these patches please mention *where* the data is being copied to/from userspace. > diff --git a/include/linux/slab.h b/include/linux/slab.h > index a48f54238273..97f4a0117b3b 100644 > --- a/include/linux/slab.h > +++ b/include/linux/slab.h > @@ -151,6 +151,11 @@ void memcg_destroy_kmem_caches(struct mem_cgroup *); > sizeof(struct __struct), __alignof__(struct __struct),\ > (__flags), NULL) > > +#define KMEM_CACHE_USERCOPY(__struct, __flags, __field) kmem_cache_create_usercopy(#__struct,\ > + sizeof(struct __struct), __alignof__(struct __struct),\ > + (__flags), offsetof(struct __struct, __field),\ > + sizeof_field(struct __struct, __field), NULL) > + This helper macro should be added in the patch which adds kmem_cache_create_usercopy(), not in this one. - Eric
On Mon, Jun 19, 2017 at 9:08 PM, Eric Biggers <ebiggers3@gmail.com> wrote: > On Mon, Jun 19, 2017 at 04:36:31PM -0700, Kees Cook wrote: >> From: David Windsor <dave@nullcore.net> >> >> When a dentry name is short enough, it can be stored directly in >> the dentry itself. These dentry short names, stored in struct >> dentry.d_iname and therefore contained in the dentry_cache slab cache, >> need to be coped to/from userspace. >> >> In support of usercopy hardening, this patch defines a region in >> the dentry_cache slab cache in which userspace copy operations >> are allowed. >> >> This region is known as the slab cache's usercopy region. Slab >> caches can now check that each copy operation involving cache-managed >> memory falls entirely within the slab's usercopy region. >> >> This patch is modified from Brad Spengler/PaX Team's PAX_USERCOPY >> whitelisting code in the last public patch of grsecurity/PaX based on my >> understanding of the code. Changes or omissions from the original code are >> mine and don't reflect the original grsecurity/PaX code. >> > > For all these patches please mention *where* the data is being copied to/from > userspace. Can you explain what you mean here? The field being copied is already mentioned in the commit log; do you mean where in the kernel source does the copy happen? > >> diff --git a/include/linux/slab.h b/include/linux/slab.h >> index a48f54238273..97f4a0117b3b 100644 >> --- a/include/linux/slab.h >> +++ b/include/linux/slab.h >> @@ -151,6 +151,11 @@ void memcg_destroy_kmem_caches(struct mem_cgroup *); >> sizeof(struct __struct), __alignof__(struct __struct),\ >> (__flags), NULL) >> >> +#define KMEM_CACHE_USERCOPY(__struct, __flags, __field) kmem_cache_create_usercopy(#__struct,\ >> + sizeof(struct __struct), __alignof__(struct __struct),\ >> + (__flags), offsetof(struct __struct, __field),\ >> + sizeof_field(struct __struct, __field), NULL) >> + > > This helper macro should be added in the patch which adds > kmem_cache_create_usercopy(), not in this one. It got moved here since this was the only user of this function and there was already enough happening in the first patch. But yes, probably it should stay with the first patch. It can be moved. -Kees
On Wed, Jun 28, 2017 at 09:44:13AM -0700, Kees Cook wrote: > On Mon, Jun 19, 2017 at 9:08 PM, Eric Biggers <ebiggers3@gmail.com> wrote: > > On Mon, Jun 19, 2017 at 04:36:31PM -0700, Kees Cook wrote: > >> From: David Windsor <dave@nullcore.net> > >> > >> When a dentry name is short enough, it can be stored directly in > >> the dentry itself. These dentry short names, stored in struct > >> dentry.d_iname and therefore contained in the dentry_cache slab cache, > >> need to be coped to/from userspace. > >> > >> In support of usercopy hardening, this patch defines a region in > >> the dentry_cache slab cache in which userspace copy operations > >> are allowed. > >> > >> This region is known as the slab cache's usercopy region. Slab > >> caches can now check that each copy operation involving cache-managed > >> memory falls entirely within the slab's usercopy region. > >> > >> This patch is modified from Brad Spengler/PaX Team's PAX_USERCOPY > >> whitelisting code in the last public patch of grsecurity/PaX based on my > >> understanding of the code. Changes or omissions from the original code are > >> mine and don't reflect the original grsecurity/PaX code. > >> > > > > For all these patches please mention *where* the data is being copied to/from > > userspace. > > Can you explain what you mean here? The field being copied is already > mentioned in the commit log; do you mean where in the kernel source > does the copy happen? > Yes, for the ones where it isn't obvious, mentioning a syscall or ioctl might be sufficient. Others may need more explanation. Eric
diff --git a/fs/dcache.c b/fs/dcache.c index f7f3c4114baa..bae2e148946c 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -3580,8 +3580,9 @@ static void __init dcache_init(void) * but it is probably not worth it because of the cache nature * of the dcache. */ - dentry_cache = KMEM_CACHE(dentry, - SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|SLAB_MEM_SPREAD|SLAB_ACCOUNT); + dentry_cache = KMEM_CACHE_USERCOPY(dentry, + SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|SLAB_MEM_SPREAD|SLAB_ACCOUNT, + d_iname); /* Hash may have been set up in dcache_init_early */ if (!hashdist) diff --git a/include/linux/slab.h b/include/linux/slab.h index a48f54238273..97f4a0117b3b 100644 --- a/include/linux/slab.h +++ b/include/linux/slab.h @@ -151,6 +151,11 @@ void memcg_destroy_kmem_caches(struct mem_cgroup *); sizeof(struct __struct), __alignof__(struct __struct),\ (__flags), NULL) +#define KMEM_CACHE_USERCOPY(__struct, __flags, __field) kmem_cache_create_usercopy(#__struct,\ + sizeof(struct __struct), __alignof__(struct __struct),\ + (__flags), offsetof(struct __struct, __field),\ + sizeof_field(struct __struct, __field), NULL) + /* * Common kmalloc functions provided by all allocators */