From 6668ed271eaefaa63e686bdfbedaeb7b8e492722 Mon Sep 17 00:00:00 2001 From: Kumar Kartikeya Dwivedi Date: Fri, 4 Sep 2026 10:43:19 +0200 Subject: selftests/bpf: Reject graph kptr use after RCU unlock Add a sleepable verifier test that loads a graph-node local kptr in an explicit RCU read-side critical section, then passes its node to bpf_rbtree_remove() after the section ends. Before the verifier fix, the stale NON_OWN_REF flag makes the node look like a live borrowed reference and the program is accepted. After the fix, the pointer is demoted without NON_OWN_REF and the graph kfunc argument is rejected. Also exercise a graph kptr loaded while a spin lock provides implicit RCU protection. The pointer must be invalidated when the lock is released, which guards the required ordering between non-owning-reference invalidation and RCU demotion. Update the existing fault-protected load test state description. The post-unlock pointer no longer carries NON_OWN_REF, but remains readable because the load is rewritten to use BPF_PROBE_MEM. Signed-off-by: Kumar Kartikeya Dwivedi Link: https://lore.kernel.org/r/20260904084325.52250-7-memxor@gmail.com Signed-off-by: Alexei Starovoitov --- arch/hexagon/mm/uaccess.c | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+) create mode 100644 arch/hexagon/mm/uaccess.c (limited to 'arch/hexagon/mm/uaccess.c') diff --git a/arch/hexagon/mm/uaccess.c b/arch/hexagon/mm/uaccess.c new file mode 100644 index 000000000..3204e9ba6 --- /dev/null +++ b/arch/hexagon/mm/uaccess.c @@ -0,0 +1,37 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2010-2011, The Linux Foundation. All rights reserved. + */ + +/* + * Support for user memory access from kernel. This will + * probably be inlined for performance at some point, but + * for ease of debug, and to a lesser degree for code size, + * we implement here as subroutines. + */ +#include +#include +#include + +/* + * For clear_user(), exploit previously defined copy_to_user function + * and the fact that we've got a handy zero page defined in kernel/head.S + * + * dczero here would be even faster. + */ +__kernel_size_t __clear_user_hexagon(void __user *dest, unsigned long count) +{ + long uncleared; + + while (count > PAGE_SIZE) { + uncleared = raw_copy_to_user(dest, &empty_zero_page, PAGE_SIZE); + if (uncleared) + return count - (PAGE_SIZE - uncleared); + count -= PAGE_SIZE; + dest += PAGE_SIZE; + } + if (count) + count = raw_copy_to_user(dest, &empty_zero_page, count); + + return count; +} -- cgit v1.3.1