linux/kernel/bpf
Daniel Borkmann 8ba25a9ef9 bpf: Fix leakage of uninitialized bpf stack under speculation
commit 801c6058d1 upstream.

The current implemented mechanisms to mitigate data disclosure under
speculation mainly address stack and map value oob access from the
speculative domain. However, Piotr discovered that uninitialized BPF
stack is not protected yet, and thus old data from the kernel stack,
potentially including addresses of kernel structures, could still be
extracted from that 512 bytes large window. The BPF stack is special
compared to map values since it's not zero initialized for every
program invocation, whereas map values /are/ zero initialized upon
their initial allocation and thus cannot leak any prior data in either
domain. In the non-speculative domain, the verifier ensures that every
stack slot read must have a prior stack slot write by the BPF program
to avoid such data leaking issue.

However, this is not enough: for example, when the pointer arithmetic
operation moves the stack pointer from the last valid stack offset to
the first valid offset, the sanitation logic allows for any intermediate
offsets during speculative execution, which could then be used to
extract any restricted stack content via side-channel.

Given for unprivileged stack pointer arithmetic the use of unknown
but bounded scalars is generally forbidden, we can simply turn the
register-based arithmetic operation into an immediate-based arithmetic
operation without the need for masking. This also gives the benefit
of reducing the needed instructions for the operation. Given after
the work in 7fedb63a83 ("bpf: Tighten speculative pointer arithmetic
mask"), the aux->alu_limit already holds the final immediate value for
the offset register with the known scalar. Thus, a simple mov of the
immediate to AX register with using AX as the source for the original
instruction is sufficient and possible now in this case.

Reported-by: Piotr Krysiuk <piotras@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Piotr Krysiuk <piotras@gmail.com>
Reviewed-by: Piotr Krysiuk <piotras@gmail.com>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2021-05-07 10:51:37 +02:00
..
arraymap.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2019-06-17 20:20:36 -07:00
bpf_lru_list.c bpf_lru_list: Read double-checked variable once without lock 2021-03-04 10:26:16 +01:00
bpf_lru_list.h treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 206 2019-05-30 11:29:53 -07:00
btf.c bpf: Explicitly memset some bpf info structures declared on the stack 2020-04-02 15:11:01 +02:00
cgroup.c bpf, cgroup: Fix problematic bounds check 2021-02-10 09:25:27 +01:00
core.c bpf: Don't rely on GCC __attribute__((optimize)) to disable GCSE 2020-11-18 19:20:26 +01:00
cpumap.c cpumap: Avoid warning when CONFIG_DEBUG_PER_CPU_MAPS is enabled 2020-05-02 08:48:51 +02:00
devmap.c devmap: Use bpf_map_area_alloc() for allocating hash buckets 2020-06-30 15:36:56 -04:00
disasm.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 295 2019-06-05 17:36:38 +02:00
disasm.h treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 295 2019-06-05 17:36:38 +02:00
hashtab.c bpf: Zero-fill re-used per-cpu map element 2020-11-18 19:20:26 +01:00
helpers.c bpf: Fix helper bpf_map_peek_elem_proto pointing to wrong callback 2021-01-23 15:57:56 +01:00
inode.c bpf: Fix a rcu warning for bpffs map pretty-print 2020-10-01 13:18:19 +02:00
local_storage.c bpf: move memory size checks to bpf_map_charge_init() 2019-05-31 16:52:56 -07:00
lpm_trie.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2019-06-22 08:59:24 -04:00
Makefile bpf: Don't rely on GCC __attribute__((optimize)) to disable GCSE 2020-11-18 19:20:26 +01:00
map_in_map.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 206 2019-05-30 11:29:53 -07:00
map_in_map.h treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 206 2019-05-30 11:29:53 -07:00
offload.c bpf, offload: Replace bitwise AND by logical AND in bpf_prog_offload_info_fill 2020-02-28 17:22:27 +01:00
percpu_freelist.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 206 2019-05-30 11:29:53 -07:00
percpu_freelist.h treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 206 2019-05-30 11:29:53 -07:00
queue_stack_maps.c bpf: move memory size checks to bpf_map_charge_init() 2019-05-31 16:52:56 -07:00
reuseport_array.c bpf: move memory size checks to bpf_map_charge_init() 2019-05-31 16:52:56 -07:00
stackmap.c bpf: Check for integer overflow when using roundup_pow_of_two() 2021-02-17 10:35:16 +01:00
syscall.c bpf: sockmap: Require attach_bpf_fd when detaching a program 2020-08-07 09:34:02 +02:00
sysfs_btf.c bpf: Fix sysfs export of empty BTF section 2020-10-14 10:32:58 +02:00
tnum.c bpf: Fix incorrect verifier simulation of ARSH under ALU32 2020-01-23 08:22:44 +01:00
verifier.c bpf: Fix leakage of uninitialized bpf stack under speculation 2021-05-07 10:51:37 +02:00
xskmap.c bpf/xskmap: Return ERR_PTR for failure case instead of NULL. 2019-09-25 22:14:16 +02:00