diff options
| author | Kumar Kartikeya Dwivedi <memxor@gmail.com> | 2026-09-04 10:43:19 +0200 |
|---|---|---|
| committer | Alexei Starovoitov <ast@kernel.org> | 2026-09-04 07:58:36 -0700 |
| commit | 6668ed271eaefaa63e686bdfbedaeb7b8e492722 (patch) | |
| tree | 58e8288fead484f6755ad0815fc74c3742663761 /arch/mips/mm/cex-oct.S | |
| download | linux-stable-6668ed271eaefaa63e686bdfbedaeb7b8e492722.tar.gz linux-stable-6668ed271eaefaa63e686bdfbedaeb7b8e492722.zip | |
selftests/bpf: Reject graph kptr use after RCU unlockgrafted
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 <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260904084325.52250-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Diffstat (limited to 'arch/mips/mm/cex-oct.S')
| -rw-r--r-- | arch/mips/mm/cex-oct.S | 70 |
1 files changed, 70 insertions, 0 deletions
diff --git a/arch/mips/mm/cex-oct.S b/arch/mips/mm/cex-oct.S new file mode 100644 index 000000000..9029092aa --- /dev/null +++ b/arch/mips/mm/cex-oct.S @@ -0,0 +1,70 @@ +/* + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file "COPYING" in the main directory of this archive + * for more details. + * + * Copyright (C) 2006 Cavium Networks + * Cache error handler + */ + +#include <asm/asm.h> +#include <asm/regdef.h> +#include <asm/mipsregs.h> +#include <asm/stackframe.h> + +/* + * Handle cache error. Indicate to the second level handler whether + * the exception is recoverable. + */ + LEAF(except_vec2_octeon) + + .set push + .set mips64r2 + .set noreorder + .set noat + + + /* due to an errata we need to read the COP0 CacheErr (Dcache) + * before any cache/DRAM access */ + + rdhwr k0, $0 /* get core_id */ + PTR_LA k1, cache_err_dcache + sll k0, k0, 3 + PTR_ADDU k1, k0, k1 /* k1 = &cache_err_dcache[core_id] */ + + dmfc0 k0, CP0_CACHEERR, 1 + sd k0, (k1) + dmtc0 $0, CP0_CACHEERR, 1 + + /* check whether this is a nested exception */ + mfc0 k1, CP0_STATUS + andi k1, k1, ST0_EXL + beqz k1, 1f + nop + j cache_parity_error_octeon_non_recoverable + nop + + /* exception is recoverable */ +1: j handle_cache_err + nop + + .set pop + END(except_vec2_octeon) + + /* We need to jump to handle_cache_err so that the previous handler + * can fit within 0x80 bytes. We also move from 0xFFFFFFFFAXXXXXXX + * space (uncached) to the 0xFFFFFFFF8XXXXXXX space (cached). */ + LEAF(handle_cache_err) + .set push + .set noreorder + .set noat + + SAVE_ALL + KMODE + jal cache_parity_error_octeon_recoverable + nop + j ret_from_exception + nop + + .set pop + END(handle_cache_err) |
