diff options
| author | Tiezhu Yang <yangtiezhu@loongson.cn> | 2026-11-10 17:20:30 +0800 |
|---|---|---|
| committer | Huacai Chen <chenhuacai@loongson.cn> | 2026-11-10 17:20:30 +0800 |
| commit | 9496832b824ab5f570c86d65e3b6f08163e6908b (patch) | |
| tree | cd1c5257510dc6c26529cc15da2ac3c20f950064 /include/linux/timerqueue.h | |
| parent | 40d7a3e63c3d88b43e9115d23d9937dd506bbd02 (diff) | |
LoongArch: BPF: Make arena support depend on 128-bit cmpxchg
When testing bpf selftest "jit_insns/cast_to_user", there are different
behaviors on different Loongson machines.
(1) Loongson-3A6000
$ sudo ./test_progs -a jit_insns/cast_to_user
...
Summary: 1/1 PASSED, 0 SKIPPED, 0/0 FAILED
(2) Loongson-3A5000
$ sudo ./test_progs -a jit_insns/cast_to_user
...
libbpf: map 'arena': failed to create: -ENOMEM
run_subtest:FAIL:unexpected_load_failure unexpected error: -12 (errno 12)
...
Summary: 0/0 PASSED, 0 SKIPPED, 1/1 FAILED
This is because the arena range tree uses kmalloc_nolock() API, which
requires 128-bit atomic cmpxchg. This feature is provided by not only
hardware SCQ extension but also assembler support on LoongArch. On CPUs
without the SCQ extension, or if the older versions of assembler do not
support the sc.q instruction, arena map creation fails with a misleading
-ENOMEM error.
Since system_has_cmpxchg128() and arch_cmpxchg128() are only available
when CONFIG_AS_HAS_SCQ_EXTENSION is enabled (need assembler support), we
guard bpf_jit_supports_arena() with an #ifdef block to preserve compile-
time robustness first, and then safely return system_has_cmpxchg128() if
defined, otherwise falling back to false. This ensures that arena map
creation cleanly fails with a -EOPNOTSUPP on the systems without either
the assembler support or the hardware SCQ extension.
With this patch, here is the test result on Loongson-3A5000 without SCQ
extension:
$ sudo ./test_progs -a jit_insns/cast_to_user
...
libbpf: map 'arena': failed to create: -EOPNOTSUPP
run_subtest:FAIL:unexpected_load_failure unexpected error: -95 (errno 95)
...
Summary: 0/0 PASSED, 0 SKIPPED, 1/1 FAILED
It is better to mark the test status as SKIPPED rather than logging a
false-alarm failure if the feature is not supported, but this belongs
to a separate bpf selftest patch.
This is similar to commit 536b523b4073 ("bpf, riscv: Make arena support
depend on ZACAS").
Note that BPF should not depend on a SLUB debug-mode escape hatch, so no
need to consider the slub_debug exception.
Cc: stable@vger.kernel.org
Fixes: 4fdb5dd8aeba ("LoongArch: BPF: Implement bpf_addr_space_cast instruction")
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Diffstat (limited to 'include/linux/timerqueue.h')
0 files changed, 0 insertions, 0 deletions
