selftests/bpf: Check that JMP32 pointer vs zero jumps are not predictedgrafted

Add jmp32_ptr_vs_zero_jne: the fall-through of the 32-bit compare, which
the verifier used to skip, contains an out of bounds map value access,
hence w/o the previous patch the program is accepted. See previous patch
for detailed description.

Reported-by: Nicholas Carlini <npc@anthropic.com>
Suggested-by: Nicholas Carlini <npc@anthropic.com>
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20260904083325.2083493-6-eddyz87@gmail.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>

1 files changed, 31 insertions, 0 deletions

diff --git a/rust/proc-macro2/location.rs b/rust/proc-macro2/location.rs
new file mode 100644
index 000000000..3a11871e2
--- /dev/null
+++ b/rust/proc-macro2/location.rs
@@ -0,0 +1,31 @@
+// SPDX-License-Identifier: Apache-2.0 OR MIT
+
+use core::cmp::Ordering;
+
+/// A line-column pair representing the start or end of a `Span`.
+///
+/// This type is semver exempt and not exposed by default.
+#[cfg_attr(docsrs, doc(cfg(feature = "span-locations")))]
+#[derive(Copy, Clone, Debug, PartialEq, Eq, Hash)]
+pub struct LineColumn {
+    /// The 1-indexed line in the source file on which the span starts or ends
+    /// (inclusive).
+    pub line: usize,
+    /// The 0-indexed column (in UTF-8 characters) in the source file on which
+    /// the span starts or ends (inclusive).
+    pub column: usize,
+}
+
+impl Ord for LineColumn {
+    fn cmp(&self, other: &Self) -> Ordering {
+        self.line
+            .cmp(&other.line)
+            .then(self.column.cmp(&other.column))
+    }
+}
+
+impl PartialOrd for LineColumn {
+    fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
+        Some(self.cmp(other))
+    }
+}