From: Hou Tao <houtao1@xxxxxxxxxx> Hi, As reported by syzboot [1] and [2], when trying to read vsyscall page by using bpf_probe_read_kernel() or bpf_probe_read(), oops will happen. Thomas Gleixner had proposed a test patch [3], but it seems that no formal patch is posted after about one month [4], so I post it instead and add an Originally-from tag in patch #2. Patch #1 makes is_vsyscall_vaddr() being a common helper. Patch #2 fixes the problem by disallowing vsyscall page read for copy_from_kernel_nofault(). Patch #3 adds one test case to ensure the read of vsyscall page through bpf is rejected. Although vsyscall page can be disabled by vsyscall=none, but it doesn't affect the reproduce of the problem and the added test. Comments are always welcome. [1]: https://lore.kernel.org/bpf/CAG48ez06TZft=ATH1qh2c5mpS5BT8UakwNkzi6nvK5_djC-4Nw@xxxxxxxxxxxxxx/ [2]: https://lore.kernel.org/bpf/CABOYnLynjBoFZOf3Z4BhaZkc5hx_kHfsjiW+UWLoB=w33LvScw@xxxxxxxxxxxxxx/ [3]: https://lore.kernel.org/bpf/87r0jwquhv.ffs@tglx/ [4]: https://lore.kernel.org/bpf/e24b125c-8ff4-9031-6c53-67ff2e01f316@xxxxxxxxxxxxxxx/ Hou Tao (3): x86/mm: Move is_vsyscall_vaddr() into mm_internal.h x86/mm: Disallow vsyscall page read for copy_from_kernel_nofault() selftest/bpf: Test the read of vsyscall page under x86-64 arch/x86/mm/fault.c | 11 +--- arch/x86/mm/maccess.c | 6 ++ arch/x86/mm/mm_internal.h | 13 ++++ .../selftests/bpf/prog_tests/read_vsyscall.c | 61 +++++++++++++++++++ .../selftests/bpf/progs/read_vsyscall.c | 45 ++++++++++++++ 5 files changed, 127 insertions(+), 9 deletions(-) create mode 100644 tools/testing/selftests/bpf/prog_tests/read_vsyscall.c create mode 100644 tools/testing/selftests/bpf/progs/read_vsyscall.c -- 2.29.2
