Merge branch 'bpf-support-to-track-bpf_jne'

Menglong Dong says:

====================
bpf: support to track BPF_JNE

For now, the reg bounds is not handled for BPF_JNE case, which can cause
the failure of following case:

  /* The type of "a" is u32 */
  if (a > 0 && a < 100) {
    /* the range of the register for a is [0, 99], not [1, 99],
     * and will cause the following error:
     *
     *   invalid zero-sized read
     *
     * as a can be 0.
     */
    bpf_skb_store_bytes(skb, xx, xx, a, 0);
  }

In the code above, "a > 0" will be compiled to "if a == 0 goto xxx". In
the TRUE branch, the dst_reg will be marked as known to 0. However, in the
fallthrough(FALSE) branch, the dst_reg will not be handled, which makes
the [min, max] for a is [0, 99], not [1, 99].

In the 1st patch, we reduce the range of the dst reg if the src reg is a
const and is exactly the edge of the dst reg For BPF_JNE.

In the 2nd patch, we remove reduplicated s32 casting in "crafted_cases".

In the 3rd patch, we just activate the test case for this logic in
range_cond(), which is committed by Andrii in the
commit 8863238993e2 ("selftests/bpf: BPF register range bounds tester").

In the 4th patch, we convert the case above to a testcase and add it to
verifier_bounds.c.

Changes since v4:
- add the 2nd patch
- add "{U32, U32, {0, U32_MAX}, {U32_MAX, U32_MAX}}" that we missed in the
  3rd patch
- add some comments to the function that we add in the 4th patch
- add reg_not_equal_const() in the 4th patch

Changes since v3:
- do some adjustment to the crafted cases that we added in the 2nd patch
- add the 3rd patch

Changes since v2:
- fix a typo in the subject of the 1st patch
- add some comments to the 1st patch, as Eduard advised
- add some cases to the "crafted_cases"

Changes since v1:
- simplify the code in the 1st patch
- introduce the 2nd patch for the testing
====================

Link: https://lore.kernel.org/r/20231219134800.1550388-1-menglong8.dong@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

1 files changed, 37 insertions, 1 deletions

diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index 4ceec8c2a484..df1cae459c77 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -14336,7 +14336,43 @@ again:
 		}
 		break;
 	case BPF_JNE:
-		/* we don't derive any new information for inequality yet */
+		if (!is_reg_const(reg2, is_jmp32))
+			swap(reg1, reg2);
+		if (!is_reg_const(reg2, is_jmp32))
+			break;
+
+		/* try to recompute the bound of reg1 if reg2 is a const and
+		 * is exactly the edge of reg1.
+		 */
+		val = reg_const_value(reg2, is_jmp32);
+		if (is_jmp32) {
+			/* u32_min_value is not equal to 0xffffffff at this point,
+			 * because otherwise u32_max_value is 0xffffffff as well,
+			 * in such a case both reg1 and reg2 would be constants,
+			 * jump would be predicted and reg_set_min_max() won't
+			 * be called.
+			 *
+			 * Same reasoning works for all {u,s}{min,max}{32,64} cases
+			 * below.
+			 */
+			if (reg1->u32_min_value == (u32)val)
+				reg1->u32_min_value++;
+			if (reg1->u32_max_value == (u32)val)
+				reg1->u32_max_value--;
+			if (reg1->s32_min_value == (s32)val)
+				reg1->s32_min_value++;
+			if (reg1->s32_max_value == (s32)val)
+				reg1->s32_max_value--;
+		} else {
+			if (reg1->umin_value == (u64)val)
+				reg1->umin_value++;
+			if (reg1->umax_value == (u64)val)
+				reg1->umax_value--;
+			if (reg1->smin_value == (s64)val)
+				reg1->smin_value++;
+			if (reg1->smax_value == (s64)val)
+				reg1->smax_value--;
+		}
 		break;
 	case BPF_JSET:
 		if (!is_reg_const(reg2, is_jmp32))