linux/kernel/bpf
Kumar Kartikeya Dwivedi d6fefa1105 bpf: Fix state pruning for STACK_DYNPTR stack slots
The root of the problem is missing liveness marking for STACK_DYNPTR
slots. This leads to all kinds of problems inside stacksafe.

The verifier by default inside stacksafe ignores spilled_ptr in stack
slots which do not have REG_LIVE_READ marks. Since this is being checked
in the 'old' explored state, it must have already done clean_live_states
for this old bpf_func_state. Hence, it won't be receiving any more
liveness marks from to be explored insns (it has received REG_LIVE_DONE
marking from liveness point of view).

What this means is that verifier considers that it's safe to not compare
the stack slot if was never read by children states. While liveness
marks are usually propagated correctly following the parentage chain for
spilled registers (SCALAR_VALUE and PTR_* types), the same is not the
case for STACK_DYNPTR.

clean_live_states hence simply rewrites these stack slots to the type
STACK_INVALID since it sees no REG_LIVE_READ marks.

The end result is that we will never see STACK_DYNPTR slots in explored
state. Even if verifier was conservatively matching !REG_LIVE_READ
slots, very next check continuing the stacksafe loop on seeing
STACK_INVALID would again prevent further checks.

Now as long as verifier stores an explored state which we can compare to
when reaching a pruning point, we can abuse this bug to make verifier
prune search for obviously unsafe paths using STACK_DYNPTR slots
thinking they are never used hence safe.

Doing this in unprivileged mode is a bit challenging. add_new_state is
only set when seeing BPF_F_TEST_STATE_FREQ (which requires privileges)
or when jmps_processed difference is >= 2 and insn_processed difference
is >= 8. So coming up with the unprivileged case requires a little more
work, but it is still totally possible. The test case being discussed
below triggers the heuristic even in unprivileged mode.

However, it no longer works since commit
8addbfc7b3 ("bpf: Gate dynptr API behind CAP_BPF").

Let's try to study the test step by step.

Consider the following program (C style BPF ASM):

0  r0 = 0;
1  r6 = &ringbuf_map;
3  r1 = r6;
4  r2 = 8;
5  r3 = 0;
6  r4 = r10;
7  r4 -= -16;
8  call bpf_ringbuf_reserve_dynptr;
9  if r0 == 0 goto pc+1;
10 goto pc+1;
11 *(r10 - 16) = 0xeB9F;
12 r1 = r10;
13 r1 -= -16;
14 r2 = 0;
15 call bpf_ringbuf_discard_dynptr;
16 r0 = 0;
17 exit;

We know that insn 12 will be a pruning point, hence if we force
add_new_state for it, it will first verify the following path as
safe in straight line exploration:
0 1 3 4 5 6 7 8 9 -> 10 -> (12) 13 14 15 16 17

Then, when we arrive at insn 12 from the following path:
0 1 3 4 5 6 7 8 9 -> 11 (12)

We will find a state that has been verified as safe already at insn 12.
Since register state is same at this point, regsafe will pass. Next, in
stacksafe, for spi = 0 and spi = 1 (location of our dynptr) is skipped
seeing !REG_LIVE_READ. The rest matches, so stacksafe returns true.
Next, refsafe is also true as reference state is unchanged in both
states.

The states are considered equivalent and search is pruned.

Hence, we are able to construct a dynptr with arbitrary contents and use
the dynptr API to operate on this arbitrary pointer and arbitrary size +
offset.

To fix this, first define a mark_dynptr_read function that propagates
liveness marks whenever a valid initialized dynptr is accessed by dynptr
helpers. REG_LIVE_WRITTEN is marked whenever we initialize an
uninitialized dynptr. This is done in mark_stack_slots_dynptr. It allows
screening off mark_reg_read and not propagating marks upwards from that
point.

This ensures that we either set REG_LIVE_READ64 on both dynptr slots, or
none, so clean_live_states either sets both slots to STACK_INVALID or
none of them. This is the invariant the checks inside stacksafe rely on.

Next, do a complete comparison of both stack slots whenever they have
STACK_DYNPTR. Compare the dynptr type stored in the spilled_ptr, and
also whether both form the same first_slot. Only then is the later path
safe.

Fixes: 97e03f5210 ("bpf: Add verifier support for dynptrs")
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20230121002241.2113993-2-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2023-01-20 17:55:02 -08:00
..
preload bpf: iterators: Build and use lightweight bootstrap version of bpftool 2022-07-15 12:01:30 -07:00
arraymap.c bpf: Do btf_record_free outside map_free callback 2022-11-17 19:11:31 -08:00
bloom_filter.c treewide: use get_random_u32() when possible 2022-10-11 17:42:58 -06:00
bpf_cgrp_storage.c bpf: Fix a compilation failure with clang lto build 2022-11-30 17:13:25 -08:00
bpf_inode_storage.c bpf: Fix a compilation failure with clang lto build 2022-11-30 17:13:25 -08:00
bpf_iter.c bpf: Initialize the bpf_run_ctx in bpf_iter_run_prog() 2022-08-18 17:06:13 -07:00
bpf_local_storage.c bpf: Reduce smap->elem_size 2022-12-21 15:42:39 +01:00
bpf_lru_list.c bpf_lru_list: Read double-checked variable once without lock 2021-02-10 15:54:26 -08:00
bpf_lru_list.h printk: stop including cache.h from printk.h 2022-05-13 07:20:07 -07:00
bpf_lsm.c bpf: Define sock security related BTF IDs under CONFIG_SECURITY_NETWORK 2022-12-19 22:02:17 +01:00
bpf_struct_ops_types.h bpf: Add dummy BPF STRUCT_OPS for test purpose 2021-11-01 14:10:00 -07:00
bpf_struct_ops.c mm: Introduce set_memory_rox() 2022-12-15 10:37:26 -08:00
bpf_task_storage.c bpf: Fix a compilation failure with clang lto build 2022-11-30 17:13:25 -08:00
btf.c bpf: btf: limit logging of ignored BTF mismatches 2023-01-10 15:58:30 -08:00
cgroup_iter.c bpf: Pin the start cgroup in cgroup_iter_seq_init() 2022-11-21 17:40:42 +01:00
cgroup.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2022-10-03 17:44:18 -07:00
core.c Including fixes from bpf, netfilter and can. 2022-12-21 08:41:32 -08:00
cpumap.c bpf: Expand map key argument of bpf_redirect_map to u64 2022-11-15 09:00:27 -08:00
devmap.c bpf: Expand map key argument of bpf_redirect_map to u64 2022-11-15 09:00:27 -08:00
disasm.c bpf: Relicense disassembler as GPL-2.0-only OR BSD-2-Clause 2021-09-02 14:49:23 +02:00
disasm.h bpf: Relicense disassembler as GPL-2.0-only OR BSD-2-Clause 2021-09-02 14:49:23 +02:00
dispatcher.c bpf: Synchronize dispatcher update with bpf_dispatcher_xdp_func 2022-12-14 12:02:14 -08:00
hashtab.c bpf: Do btf_record_free outside map_free callback 2022-11-17 19:11:31 -08:00
helpers.c bpf: rename list_head -> graph_root in field info types 2022-12-28 20:14:22 -08:00
inode.c bpf: Convert bpf_preload.ko to use light skeleton. 2022-02-10 23:31:51 +01:00
Kconfig rcu: Make the TASKS_RCU Kconfig option be selected 2022-04-20 16:52:58 -07:00
link_iter.c bpf: Add bpf_link iterator 2022-05-10 11:20:45 -07:00
local_storage.c bpf: Consolidate spin_lock, timer management into btf_record 2022-11-03 22:19:40 -07:00
lpm_trie.c bpf: Use bpf_map_area_alloc consistently on bpf map creation 2022-08-10 11:50:43 -07:00
Makefile bpf: Implement cgroup storage available to non-cgroup-attached bpf progs 2022-10-25 23:19:19 -07:00
map_in_map.c bpf: Add comments for map BTF matching requirement for bpf_list_head 2022-11-17 19:22:14 -08:00
map_in_map.h
map_iter.c bpf: Introduce MEM_RDONLY flag 2021-12-18 13:27:41 -08:00
memalloc.c bpf: Skip rcu_barrier() if rcu_trace_implies_rcu_gp() is true 2022-12-08 17:50:17 -08:00
mmap_unlock_work.h bpf: Introduce helper bpf_find_vma 2021-11-07 11:54:51 -08:00
net_namespace.c net: Add includes masked by netdevice.h including uapi/bpf.h 2021-12-29 20:03:05 -08:00
offload.c bpf: Use bpf_map_area_alloc consistently on bpf map creation 2022-08-10 11:50:43 -07:00
percpu_freelist.c bpf: Initialize same number of free nodes for each pcpu_freelist 2022-11-11 12:05:14 -08:00
percpu_freelist.h
prog_iter.c
queue_stack_maps.c bpf: Remove unneeded memset in queue_stack_map creation 2022-08-10 11:48:22 -07:00
reuseport_array.c net: Fix suspicious RCU usage in bpf_sk_reuseport_detach() 2022-08-17 16:42:59 -07:00
ringbuf.c bpf: Rename MEM_ALLOC to MEM_RINGBUF 2022-11-14 21:52:45 -08:00
stackmap.c perf/bpf: Always use perf callchains if exist 2022-09-13 15:03:22 +02:00
syscall.c bpf-next-for-netdev 2023-01-04 20:21:25 -08:00
sysfs_btf.c bpf: Load and verify kernel module BTFs 2020-11-10 15:25:53 -08:00
task_iter.c bpf: keep a reference to the mm, in case the task is dead. 2022-12-28 14:11:48 -08:00
tnum.c bpf, tnums: Provably sound, faster, and more precise algorithm for tnum_mul 2021-06-01 13:34:15 +02:00
trampoline.c bpf: Fix panic due to wrong pageattr of im->image 2022-12-28 13:46:28 -08:00
verifier.c bpf: Fix state pruning for STACK_DYNPTR stack slots 2023-01-20 17:55:02 -08:00