| Message ID | 20230118155630.2762921-1-leitao@debian.org (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | io_uring: Enable KASAN for request cache | expand |
On 1/18/23 15:56, Breno Leitao wrote: > Every io_uring request is represented by struct io_kiocb, which is > cached locally by io_uring (not SLAB/SLUB) in the list called > submit_state.freelist. This patch simply enabled KASAN for this free > list. > > This list is initially created by KMEM_CACHE, but later, managed by > io_uring. This patch basically poisons the objects that are not used > (i.e., they are the free list), and unpoisons it when the object is > allocated/removed from the list. > > Touching these poisoned objects while in the freelist will cause a KASAN > warning. Doesn't apply cleanly to for-6.3/io_uring, but otherwise looks good Reviewed-by: Pavel Begunkov <asml.silence@gmail.com> > Suggested-by: Jens Axboe <axboe@kernel.dk> > Signed-off-by: Breno Leitao <leitao@debian.org> > --- > io_uring/io_uring.c | 3 ++- > io_uring/io_uring.h | 11 ++++++++--- > 2 files changed, 10 insertions(+), 4 deletions(-) > > diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c > index 2ac1cd8d23ea..8cc0f12034d1 100644 > --- a/io_uring/io_uring.c > +++ b/io_uring/io_uring.c > @@ -151,7 +151,7 @@ static void io_move_task_work_from_local(struct io_ring_ctx *ctx); > static void __io_submit_flush_completions(struct io_ring_ctx *ctx); > static __cold void io_fallback_tw(struct io_uring_task *tctx); > > -static struct kmem_cache *req_cachep; > +struct kmem_cache *req_cachep; > > struct sock *io_uring_get_socket(struct file *file) > { > @@ -230,6 +230,7 @@ static inline void req_fail_link_node(struct io_kiocb *req, int res) > static inline void io_req_add_to_cache(struct io_kiocb *req, struct io_ring_ctx *ctx) > { > wq_stack_add_head(&req->comp_list, &ctx->submit_state.free_list); > + kasan_poison_object_data(req_cachep, req); > } > > static __cold void io_ring_ctx_ref_free(struct percpu_ref *ref) > diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h > index ab4b2a1c3b7e..0ccf62a19b65 100644 > --- a/io_uring/io_uring.h > +++ b/io_uring/io_uring.h > @@ -3,6 +3,7 @@ > > #include <linux/errno.h> > #include <linux/lockdep.h> > +#include <linux/kasan.h> > #include <linux/io_uring_types.h> > #include <uapi/linux/eventpoll.h> > #include "io-wq.h" > @@ -379,12 +380,16 @@ static inline bool io_alloc_req_refill(struct io_ring_ctx *ctx) > return true; > } > > +extern struct kmem_cache *req_cachep; > + > static inline struct io_kiocb *io_alloc_req(struct io_ring_ctx *ctx) > { > - struct io_wq_work_node *node; > + struct io_kiocb *req; > > - node = wq_stack_extract(&ctx->submit_state.free_list); > - return container_of(node, struct io_kiocb, comp_list); > + req = container_of(ctx->submit_state.free_list.next, struct io_kiocb, comp_list); > + kasan_unpoison_object_data(req_cachep, req); > + wq_stack_extract(&ctx->submit_state.free_list); > + return req; > } > > static inline bool io_allowed_run_tw(struct io_ring_ctx *ctx)
On 1/20/23 8:09 AM, Pavel Begunkov wrote: > On 1/18/23 15:56, Breno Leitao wrote: >> Every io_uring request is represented by struct io_kiocb, which is >> cached locally by io_uring (not SLAB/SLUB) in the list called >> submit_state.freelist. This patch simply enabled KASAN for this free >> list. >> >> This list is initially created by KMEM_CACHE, but later, managed by >> io_uring. This patch basically poisons the objects that are not used >> (i.e., they are the free list), and unpoisons it when the object is >> allocated/removed from the list. >> >> Touching these poisoned objects while in the freelist will cause a KASAN >> warning. > > Doesn't apply cleanly to for-6.3/io_uring, but otherwise looks good > > Reviewed-by: Pavel Begunkov <asml.silence@gmail.com> I ran testing on this yesterday and noticed the same thing, just a trivial fuzz reject. I can fix it up while applying. Thanks for reviewing!
On Wed, 18 Jan 2023 07:56:30 -0800, Breno Leitao wrote: > Every io_uring request is represented by struct io_kiocb, which is > cached locally by io_uring (not SLAB/SLUB) in the list called > submit_state.freelist. This patch simply enabled KASAN for this free > list. > > This list is initially created by KMEM_CACHE, but later, managed by > io_uring. This patch basically poisons the objects that are not used > (i.e., they are the free list), and unpoisons it when the object is > allocated/removed from the list. > > [...] Applied, thanks! [1/1] io_uring: Enable KASAN for request cache commit: 9f61fe5e410444ccecb429e69536ecd981c73c08 Best regards,
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 2ac1cd8d23ea..8cc0f12034d1 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -151,7 +151,7 @@ static void io_move_task_work_from_local(struct io_ring_ctx *ctx); static void __io_submit_flush_completions(struct io_ring_ctx *ctx); static __cold void io_fallback_tw(struct io_uring_task *tctx); -static struct kmem_cache *req_cachep; +struct kmem_cache *req_cachep; struct sock *io_uring_get_socket(struct file *file) { @@ -230,6 +230,7 @@ static inline void req_fail_link_node(struct io_kiocb *req, int res) static inline void io_req_add_to_cache(struct io_kiocb *req, struct io_ring_ctx *ctx) { wq_stack_add_head(&req->comp_list, &ctx->submit_state.free_list); + kasan_poison_object_data(req_cachep, req); } static __cold void io_ring_ctx_ref_free(struct percpu_ref *ref) diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h index ab4b2a1c3b7e..0ccf62a19b65 100644 --- a/io_uring/io_uring.h +++ b/io_uring/io_uring.h @@ -3,6 +3,7 @@ #include <linux/errno.h> #include <linux/lockdep.h> +#include <linux/kasan.h> #include <linux/io_uring_types.h> #include <uapi/linux/eventpoll.h> #include "io-wq.h" @@ -379,12 +380,16 @@ static inline bool io_alloc_req_refill(struct io_ring_ctx *ctx) return true; } +extern struct kmem_cache *req_cachep; + static inline struct io_kiocb *io_alloc_req(struct io_ring_ctx *ctx) { - struct io_wq_work_node *node; + struct io_kiocb *req; - node = wq_stack_extract(&ctx->submit_state.free_list); - return container_of(node, struct io_kiocb, comp_list); + req = container_of(ctx->submit_state.free_list.next, struct io_kiocb, comp_list); + kasan_unpoison_object_data(req_cachep, req); + wq_stack_extract(&ctx->submit_state.free_list); + return req; } static inline bool io_allowed_run_tw(struct io_ring_ctx *ctx)
Every io_uring request is represented by struct io_kiocb, which is cached locally by io_uring (not SLAB/SLUB) in the list called submit_state.freelist. This patch simply enabled KASAN for this free list. This list is initially created by KMEM_CACHE, but later, managed by io_uring. This patch basically poisons the objects that are not used (i.e., they are the free list), and unpoisons it when the object is allocated/removed from the list. Touching these poisoned objects while in the freelist will cause a KASAN warning. Suggested-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Breno Leitao <leitao@debian.org> --- io_uring/io_uring.c | 3 ++- io_uring/io_uring.h | 11 ++++++++--- 2 files changed, 10 insertions(+), 4 deletions(-)