bpf: Fix tnum_overlap to check for zero mask first
Subsystems (4)
cover bpf: Fix tnum_overlap to check for zero mask first
This small patchset is about avoid verifier bug warning when tnum_overlap() is called with zero mask.
---
KaFai Wan (2):
bpf: Fix tnum_overlap to check for zero mask first selftests/bpf: Range analysis test case for JEQ
kernel/bpf/tnum.c | 2 ++ .../selftests/bpf/progs/verifier_bounds.c | 23 +++++++++++++++++++ 2 files changed, 25 insertions(+)
Signature · 1 line
2.43.0
PATCH 1/2 bpf: Fix tnum_overlap to check for zero mask first
Syzbot reported a kernel warning due to a range invariant violation in the BPF verifier. The issue occurs when tnum_overlap() fails to detect that two tnums don't have any overlapping bits.
The problematic BPF program:
0: call bpf_get_prandom_u32
1: r6 = r0
2: r6 &= 0xFFFFFFFFFFFFFFF0
3: r7 = r0
4: r7 &= 0x07
5: r7 -= 0xFF
6: if r6 == r7 goto <exit>
After instruction 5, R7 has the range:
R7: u64=[0xffffffffffffff01, 0xffffffffffffff08] var_off=(0xffffffffffffff00; 0xf)
R6 and R7 don't overlap since they have no agreeing bits. However, is_branch_taken() fails to recognize this, causing the verifier to refine register bounds and end up with inconsistent bounds:
6: if r6 == r7 goto <exit>
R6: u64=[0xffffffffffffff01, 0xffffffffffffff00] var_off=(0xffffffffffffff00, 0x0)
R7: u64=[0xffffffffffffff01, 0xffffffffffffff00] var_off=(0xffffffffffffff00, 0x0)
The root cause is that tnum_overlap() doesn't properly handle the case where the masks have no overlapping bits.
Fix this by adding an early check for zero mask intersection in tnum_overlap().
--- kernel/bpf/tnum.c | 2 ++ 1 file changed, 2 insertions(+)
kernel/bpf/tnum.c +2 -0
@@ -163,6 +163,8 @@ bool tnum_overlap(struct tnum a, struct tnum b) {u64 mu; +if ((a.mask & b.mask) ==0 ) +return false; mu = ~a.mask & ~b.mask;return (a.value & mu) == (b.value & mu); }
Signature · 1 line
2.43.0
Sorry, this patch is wrong, please ignore.
Quoted 49 lines · KaFai Wan
Syzbot reported a kernel warning due to a range invariant violation in the BPF verifier. The issue occurs when
tnum_overlap()fails to detect that two tnums don't have any overlapping bits.The problematic BPF program:
0: callbpf_get_prandom_u32
1: r6 = r0
2: r6 &=0xFFFFFFFFFFFFFFF0
3: r7 = r0
4: r7 &=0x07
5: r7 -=0xFF
6: if r6 == r7 goto <exit>After instruction 5, R7 has the range:
R7: u64=[0xffffffffffffff01,0xffffffffffffff08] var_off=(0xffffffffffffff00;0xf)R6 and R7 don't overlap since they have no agreeing bits. However,
is_branch_taken()fails to recognize this, causing the verifier to refine register bounds and end up with inconsistent bounds:6: if r6 == r7 goto <exit>
R6: u64=[0xffffffffffffff01,0xffffffffffffff00] var_off=(0xffffffffffffff00,0x0)
R7: u64=[0xffffffffffffff01,0xffffffffffffff00] var_off=(0xffffffffffffff00,0x0)The root cause is that
tnum_overlap()doesn't properly handle the case where the masks have no overlapping bits.Fix this by adding an early check for zero mask intersection in
tnum_overlap().Reported-by [email protected]Fixes f41345f47fb2 ("bpf: Use tnums for JEQ/JNE is_branch_taken logic")Signed-off-by KaFai Wan <[email protected]>---
kernel/bpf/tnum.c | 2 ++1 file changed, 2 insertions(+)kernel/bpf/tnum.c +0
-0@@ -163,6 +163,8 @@ bool tnum_overlap(struct tnum a, struct tnum b) { u64 mu;kernel/bpf/tnum.c +2
-0+if ((a.mask & b.mask) ==0 ) +return false;mu = ~a.mask & ~b.mask;
return (a.value & mu) == (b.value & mu);
}
Signature · 2 lines
Thanks, KaFai
PATCH 2/2 selftests/bpf: Range analysis test case for JEQ
In reply to KaFai WanThis patch adds coverage for the warning detected by syzkaller and fixed in the previous patch. Without the previous patch, this test fails with:
verifier bug: REG INVARIANTS VIOLATION (true_reg1): range bounds violation u64=[0xffffffffffffff01, 0xffffffffffffff00] s64=[0xffffffffffffff01, 0xffffffffffffff00] u32=[0xffffff01, 0xffffff00] s32=[0xffffff00, 0xffffff00] var_off=(0xffffffffffffff00, 0x0)
verifier bug: REG INVARIANTS VIOLATION (true_reg2): range bounds violation u64=[0xffffffffffffff01, 0xffffffffffffff00] s64=[0xffffffffffffff01, 0xffffffffffffff00] u32=[0xffffff01, 0xffffff00] s32=[0xffffff01, 0xffffff00] var_off=(0xffffffffffffff00, 0x0)
--- .../selftests/bpf/progs/verifier_bounds.c | 23 +++++++++++++++++++ 1 file changed, 23 insertions(+)
tools/testing/selftests/bpf/progs/verifier_bounds.c +23 -0
@@ -1550,6 +1550,29 @@ l0_%=: r0 = 0; \ : __clobber_all); } +SEC("socket") +__description("dead branch on jeq, does not result in invariants violation error") +__success __log_level(2) +__retval(0) __flag(BPF_F_TEST_REG_INVARIANTS) +__naked void jeq_range_analysis(void) +{ + asmvolatile (" \ +call %[bpf_get_prandom_u32]; \ +r6 = r0; \ +r6 &= 0xFFFFFFFFFFFFFFF0; \ +r7 = r0; \ +r7 &= 0x07; \ +r7 -= 0xFF; \ +if r6 == r7 goto l1_%=; \ +l0_%=: r0 = 0; \ +exit; \ +l1_%=: r0 = 1; \ +exit; \ +" : + : __imm(bpf_get_prandom_u32) + : __clobber_all); +} +/* This test covers the bounds deduction on 64bits when the s64 and u64 ranges * overlap on the negative side. At instruction 7, the ranges look as follows: *
Signature · 1 line
2.43.0