kernel_optimize_test/kernel
Daniel Borkmann 5fc6ed1831 bpf: Fix leakage under speculation on mispredicted branches
[ Upstream commit 9183671af6dbf60a1219371d4ed73e23f43b49db ]

The verifier only enumerates valid control-flow paths and skips paths that
are unreachable in the non-speculative domain. And so it can miss issues
under speculative execution on mispredicted branches.

For example, a type confusion has been demonstrated with the following
crafted program:

  // r0 = pointer to a map array entry
  // r6 = pointer to readable stack slot
  // r9 = scalar controlled by attacker
  1: r0 = *(u64 *)(r0) // cache miss
  2: if r0 != 0x0 goto line 4
  3: r6 = r9
  4: if r0 != 0x1 goto line 6
  5: r9 = *(u8 *)(r6)
  6: // leak r9

Since line 3 runs iff r0 == 0 and line 5 runs iff r0 == 1, the verifier
concludes that the pointer dereference on line 5 is safe. But: if the
attacker trains both the branches to fall-through, such that the following
is speculatively executed ...

  r6 = r9
  r9 = *(u8 *)(r6)
  // leak r9

... then the program will dereference an attacker-controlled value and could
leak its content under speculative execution via side-channel. This requires
to mistrain the branch predictor, which can be rather tricky, because the
branches are mutually exclusive. However such training can be done at
congruent addresses in user space using different branches that are not
mutually exclusive. That is, by training branches in user space ...

  A:  if r0 != 0x0 goto line C
  B:  ...
  C:  if r0 != 0x0 goto line D
  D:  ...

... such that addresses A and C collide to the same CPU branch prediction
entries in the PHT (pattern history table) as those of the BPF program's
lines 2 and 4, respectively. A non-privileged attacker could simply brute
force such collisions in the PHT until observing the attack succeeding.

Alternative methods to mistrain the branch predictor are also possible that
avoid brute forcing the collisions in the PHT. A reliable attack has been
demonstrated, for example, using the following crafted program:

  // r0 = pointer to a [control] map array entry
  // r7 = *(u64 *)(r0 + 0), training/attack phase
  // r8 = *(u64 *)(r0 + 8), oob address
  // [...]
  // r0 = pointer to a [data] map array entry
  1: if r7 == 0x3 goto line 3
  2: r8 = r0
  // crafted sequence of conditional jumps to separate the conditional
  // branch in line 193 from the current execution flow
  3: if r0 != 0x0 goto line 5
  4: if r0 == 0x0 goto exit
  5: if r0 != 0x0 goto line 7
  6: if r0 == 0x0 goto exit
  [...]
  187: if r0 != 0x0 goto line 189
  188: if r0 == 0x0 goto exit
  // load any slowly-loaded value (due to cache miss in phase 3) ...
  189: r3 = *(u64 *)(r0 + 0x1200)
  // ... and turn it into known zero for verifier, while preserving slowly-
  // loaded dependency when executing:
  190: r3 &= 1
  191: r3 &= 2
  // speculatively bypassed phase dependency
  192: r7 += r3
  193: if r7 == 0x3 goto exit
  194: r4 = *(u8 *)(r8 + 0)
  // leak r4

As can be seen, in training phase (phase != 0x3), the condition in line 1
turns into false and therefore r8 with the oob address is overridden with
the valid map value address, which in line 194 we can read out without
issues. However, in attack phase, line 2 is skipped, and due to the cache
miss in line 189 where the map value is (zeroed and later) added to the
phase register, the condition in line 193 takes the fall-through path due
to prior branch predictor training, where under speculation, it'll load the
byte at oob address r8 (unknown scalar type at that point) which could then
be leaked via side-channel.

One way to mitigate these is to 'branch off' an unreachable path, meaning,
the current verification path keeps following the is_branch_taken() path
and we push the other branch to the verification stack. Given this is
unreachable from the non-speculative domain, this branch's vstate is
explicitly marked as speculative. This is needed for two reasons: i) if
this path is solely seen from speculative execution, then we later on still
want the dead code elimination to kick in in order to sanitize these
instructions with jmp-1s, and ii) to ensure that paths walked in the
non-speculative domain are not pruned from earlier walks of paths walked in
the speculative domain. Additionally, for robustness, we mark the registers
which have been part of the conditional as unknown in the speculative path
given there should be no assumptions made on their content.

The fix in here mitigates type confusion attacks described earlier due to
i) all code paths in the BPF program being explored and ii) existing
verifier logic already ensuring that given memory access instruction
references one specific data structure.

An alternative to this fix that has also been looked at in this scope was to
mark aux->alu_state at the jump instruction with a BPF_JMP_TAKEN state as
well as direction encoding (always-goto, always-fallthrough, unknown), such
that mixing of different always-* directions themselves as well as mixing of
always-* with unknown directions would cause a program rejection by the
verifier, e.g. programs with constructs like 'if ([...]) { x = 0; } else
{ x = 1; }' with subsequent 'if (x == 1) { [...] }'. For unprivileged, this
would result in only single direction always-* taken paths, and unknown taken
paths being allowed, such that the former could be patched from a conditional
jump to an unconditional jump (ja). Compared to this approach here, it would
have two downsides: i) valid programs that otherwise are not performing any
pointer arithmetic, etc, would potentially be rejected/broken, and ii) we are
required to turn off path pruning for unprivileged, where both can be avoided
in this work through pushing the invalid branch to the verification stack.

The issue was originally discovered by Adam and Ofek, and later independently
discovered and reported as a result of Benedict and Piotr's research work.

Fixes: b2157399cc ("bpf: prevent out-of-bounds speculation")
Reported-by: Adam Morrison <mad@cs.tau.ac.il>
Reported-by: Ofek Kirzner <ofekkir@gmail.com>
Reported-by: Benedict Schlueter <benedict.schlueter@rub.de>
Reported-by: Piotr Krysiuk <piotras@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: John Fastabend <john.fastabend@gmail.com>
Reviewed-by: Benedict Schlueter <benedict.schlueter@rub.de>
Reviewed-by: Piotr Krysiuk <piotras@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2021-06-23 14:42:45 +02:00
..
bpf bpf: Fix leakage under speculation on mispredicted branches 2021-06-23 14:42:45 +02:00
cgroup cgroup1: don't allow '\n' in renaming 2021-06-16 12:01:40 +02:00
configs
debug kgdb: fix to kill breakpoints on initmem after boot 2021-03-04 11:38:46 +01:00
dma swiotlb: Fix the type of index 2021-05-19 10:13:04 +02:00
entry x86/entry: Move nmi entry/exit into common code 2021-03-17 17:06:36 +01:00
events perf: Fix data race between pin_count increment/decrement 2021-06-16 12:01:45 +02:00
gcov gcov: re-fix clang-11+ support 2021-04-14 08:41:58 +02:00
irq genirq/matrix: Prevent allocation counter corruption 2021-05-11 14:47:17 +02:00
kcsan kcsan: Fix debugfs initcall return type 2021-05-26 12:06:54 +02:00
livepatch kernel/: fix repeated words in comments 2020-10-16 11:11:19 -07:00
locking locking/mutex: clear MUTEX_FLAGS if wait_list is empty due to signal 2021-05-26 12:06:50 +02:00
power PM: EM: postpone creating the debugfs dir till fs_initcall 2021-03-30 14:32:04 +02:00
printk printk: fix deadlock when kernel panic 2021-03-04 11:38:41 +01:00
rcu rcu: Remove spurious instrumentation_end() in rcu_nmi_enter() 2021-05-14 09:50:22 +02:00
sched sched/fair: Fix util_est UTIL_AVG_UNCHANGED handling 2021-06-16 12:01:46 +02:00
time posix-timers: Preserve return value in clock_adjtime32() 2021-05-11 14:47:16 +02:00
trace tracing: Correct the length check which causes memory corruption 2021-06-16 12:01:47 +02:00
.gitignore kbuild: update config_data.gz only when the content of .config is changed 2021-05-11 14:47:37 +02:00
acct.c kernel: acct.c: fix some kernel-doc nits 2020-10-16 11:11:19 -07:00
async.c
audit_fsnotify.c fsnotify: generalize handle_inode_event() 2020-12-30 11:54:18 +01:00
audit_tree.c fsnotify: generalize handle_inode_event() 2020-12-30 11:54:18 +01:00
audit_watch.c fsnotify: generalize handle_inode_event() 2020-12-30 11:54:18 +01:00
audit.c audit: Remove redundant null check 2020-08-26 09:10:39 -04:00
audit.h
auditfilter.c
auditsc.c
backtracetest.c
bounds.c
capability.c LSM: Signal to SafeSetID when setting group IDs 2020-10-13 09:17:34 -07:00
compat.c
configs.c
context_tracking.c
cpu_pm.c notifier: Fix broken error handling pattern 2020-09-01 09:58:03 +02:00
cpu.c kernel/cpu: add arch override for clear_tasks_mm_cpumask() mm handling 2020-11-27 00:10:39 +11:00
crash_core.c
crash_dump.c
cred.c
delayacct.c
dma.c
exec_domain.c
exit.c kernel/io_uring: cancel io_uring before task works 2021-01-30 13:55:18 +01:00
extable.c
fail_function.c fail_function: Remove a redundant mutex unlock 2020-11-19 11:58:16 -08:00
fork.c mm/fork: clear PASID for new mm 2021-03-30 14:31:52 +02:00
freezer.c Revert "kernel: freezer should treat PF_IO_WORKER like PF_KTHREAD for freezing" 2021-04-07 15:00:14 +02:00
futex.c futex: Do not apply time namespace adjustment on FUTEX_LOCK_PI 2021-05-11 14:47:37 +02:00
gen_kheaders.sh
groups.c LSM: Signal to SafeSetID when setting group IDs 2020-10-13 09:17:34 -07:00
hung_task.c kernel/hung_task.c: make type annotations consistent 2020-11-02 12:14:19 -08:00
iomem.c
irq_work.c
jump_label.c static_call: Fix static_call_update() sanity check 2021-03-25 09:04:18 +01:00
kallsyms.c treewide: Convert macro and uses of __section(foo) to __section("foo") 2020-10-25 14:51:49 -07:00
kcmp.c exec: Transform exec_update_mutex into a rw_semaphore 2021-01-09 13:46:24 +01:00
Kconfig.freezer
Kconfig.hz
Kconfig.locks
Kconfig.preempt
kcov.c
kexec_core.c kernel: kexec: remove the lock operation of system_transition_mutex 2021-02-03 23:28:37 +01:00
kexec_elf.c
kexec_file.c kernel: kexec_file: fix error return code of kexec_calculate_store_digests() 2021-05-19 10:13:09 +02:00
kexec_internal.h
kexec.c LSM: Introduce kernel_post_load_data() hook 2020-10-05 13:37:03 +02:00
kheaders.c
kmod.c
kprobes.c kprobes: Fix to delay the kprobes jump optimization 2021-03-04 11:38:35 +01:00
ksysfs.c
kthread.c kthread: Extract KTHREAD_IS_PER_CPU 2021-02-07 15:37:17 +01:00
latencytop.c
Makefile kbuild: update config_data.gz only when the content of .config is changed 2021-05-11 14:47:37 +02:00
module_signature.c module: harden ELF info handling 2021-03-25 09:04:11 +01:00
module_signing.c module: harden ELF info handling 2021-03-25 09:04:11 +01:00
module-internal.h
module.c module: harden ELF info handling 2021-03-25 09:04:11 +01:00
notifier.c notifier: Fix broken error handling pattern 2020-09-01 09:58:03 +02:00
nsproxy.c
padata.c padata: fix possible padata_works_lock deadlock 2020-09-04 17:51:55 +10:00
panic.c panic: don't dump stack twice on warn 2020-11-14 11:26:04 -08:00
params.c params: Replace zero-length array with flexible-array member 2020-10-29 17:22:59 -05:00
pid_namespace.c kernel/: fix repeated words in comments 2020-10-16 11:11:19 -07:00
pid.c exec: Transform exec_update_mutex into a rw_semaphore 2021-01-09 13:46:24 +01:00
profile.c
ptrace.c ptrace: make ptrace() fail if the tracee changed its pid unexpectedly 2021-05-26 12:06:49 +02:00
range.c kernel.h: split out min()/max() et al. helpers 2020-10-16 11:11:19 -07:00
reboot.c reboot: fix overflow parsing reboot cpu number 2020-11-14 11:26:03 -08:00
regset.c
relay.c kernel/relay.c: drop unneeded initialization 2020-10-16 11:11:22 -07:00
resource.c kernel/resource: make walk_mem_res() find all busy IORESOURCE_MEM resources 2021-05-19 10:13:09 +02:00
rseq.c
scftorture.c
scs.c
seccomp.c seccomp: Refactor notification handler to prepare for new semantics 2021-06-03 09:00:31 +02:00
signal.c ptrace: fix task_join_group_stop() for the case when current is traced 2020-11-02 12:14:19 -08:00
smp.c smp: Fix smp_call_function_single_async prototype 2021-05-14 09:50:46 +02:00
smpboot.c kthread: Extract KTHREAD_IS_PER_CPU 2021-02-07 15:37:17 +01:00
smpboot.h
softirq.c softirq: Add debug check to __raise_softirq_irqoff() 2020-09-16 15:18:56 +02:00
stackleak.c stackleak: let stack_erasing_sysctl take a kernel pointer buffer 2020-09-19 13:13:39 -07:00
stacktrace.c stacktrace: Remove reliable argument from arch_stack_walk() callback 2020-09-18 14:24:16 +01:00
static_call.c static_call: Align static_call_is_init() patching condition 2021-04-07 15:00:06 +02:00
stop_machine.c stop_machine, rcu: Mark functions as notrace 2020-10-26 12:12:27 +01:00
sys_ni.c mm/madvise: introduce process_madvise() syscall: an external memory hinting API 2020-10-18 09:27:10 -07:00
sys.c kernel/sys.c: fix prototype of prctl_get_tid_address() 2020-10-25 11:44:16 -07:00
sysctl-test.c
sysctl.c sysctl.c: fix underflow value setting risk in vm_table 2021-03-17 17:06:25 +01:00
task_work.c task_work: cleanup notification modes 2020-10-17 15:05:30 -06:00
taskstats.c taskstats: move specifying netlink policy back to ops 2020-10-02 19:11:12 -07:00
test_kprobes.c
torture.c
tracepoint.c tracepoint: Do not fail unregistering a probe due to memory failure 2021-03-04 11:38:03 +01:00
tsacct.c
ucount.c
uid16.c
uid16.h
umh.c usermodehelper: reset umask to default before executing user process 2020-10-06 10:31:52 -07:00
up.c smp: Fix smp_call_function_single_async prototype 2021-05-14 09:50:46 +02:00
user_namespace.c capabilities: require CAP_SETFCAP to map uid 0 2021-05-07 11:04:31 +02:00
user-return-notifier.c
user.c
usermode_driver.c bpf: Fix umd memory leak in copy_process() 2021-03-30 14:32:03 +02:00
utsname_sysctl.c
utsname.c
watch_queue.c
watchdog_hld.c
watchdog.c watchdog: fix barriers when printing backtraces from all CPUs 2021-05-19 10:13:00 +02:00
workqueue_internal.h
workqueue.c wq: handle VM suspension in stall detection 2021-06-16 12:01:36 +02:00