s390/bpf: Replace ly instruction with llgf
Subsystems (5)
cover s390/bpf: Replace ly instruction with llgf
Sashiko found a bug [1] in 9012cf2491e3, where upper bits of a 64-bit register are not zero-extended after ly instruction in used.
Tested:
./test_progs -v -t get_smp_processor_id
[1] https://sashiko.dev/#/patchset/20260414142930.528751-1-max%40linux.ibm.com
Maxim Khmelevskii (2):
s390/bpf: Replace ly instruction with llgf selftests/bpf: Add test for bpf_get_smp_processor_id
arch/s390/net/bpf_jit_comp.c | 4 +- .../bpf/prog_tests/get_smp_processor_id.c | 45 +++++++++++++++++++ .../bpf/progs/get_smp_processor_id.c | 20 +++++++++ 3 files changed, 67 insertions(+), 2 deletions(-) create mode 100644 tools/testing/selftests/bpf/prog_tests/get_smp_processor_id.c create mode 100644 tools/testing/selftests/bpf/progs/get_smp_processor_id.c
Signature · 1 line
2.54.0
PATCH 1/2 s390/bpf: Replace ly instruction with llgf
cpu_nr is a 32 bit value and BPF_REG_0 is a 64 bit register, when ly loads the cpu_nr into BPF_REG_0 it does not zero the upper bits, but llgf does.
--- arch/s390/net/bpf_jit_comp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
arch/s390/net/bpf_jit_comp.c +2 -2
@@ -1783,8 +1783,8 @@ static noinline int bpf_jit_insn(struct bpf_jit *jit, struct bpf_prog *fp, insn->imm == BPF_FUNC_get_smp_processor_id) {const u32 *cpu_nr = &get_lowcore ()->cpu_nr; -/* ly %b0, cpu_nr */ -EMIT6_DISP_LH (0xe3000000 ,0x0058 ,BPF_REG_0 ,REG_0 ,REG_0 , +/* llgf %b0, cpu_nr */ +EMIT6_DISP_LH (0xe3000000 ,0x0016 ,BPF_REG_0 ,REG_0 ,REG_0 , (unsigned long )cpu_nr);break ; }
Signature · 1 line
2.54.0
PATCH 2/2 selftests/bpf: Add test for bpf_get_smp_processor_id
In reply to Maxim KhmelevskiiAdd a test which checks on each CPU that thebpf_get_smp_processor_id BPF helper is returning
the correct CPU number.
--- .../bpf/prog_tests/get_smp_processor_id.c | 45 +++++++++++++++++++ .../bpf/progs/get_smp_processor_id.c | 20 +++++++++ 2 files changed, 65 insertions(+) create mode 100644 tools/testing/selftests/bpf/prog_tests/get_smp_processor_id.c create mode 100644 tools/testing/selftests/bpf/progs/get_smp_processor_id.c
tools/testing/selftests/bpf/prog_tests/get_smp_processor_id.c new file +45 -0
@@ -0,0 +1,45 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <test_progs.h> +#include "bpf/libbpf_internal.h" +#include "get_smp_processor_id.skel.h" + +void test_get_smp_processor_id (void ) +{ +LIBBPF_OPTS (bpf_test_run_opts, opts, + .flags = BPF_F_TEST_RUN_ON_CPU, + .cpu = 0, + ); +struct get_smp_processor_id *skel; +int prog_fd, err, online_cpu_nr, i; +bool *online =NULL ; + + err =parse_cpu_mask_file ("/sys/devices/system/cpu/online" , + &online, &online_cpu_nr); +if (!ASSERT_OK (err,"parse_cpu_mask_file" )) +return ; + + skel =get_smp_processor_id__open_and_load (); +if (!ASSERT_OK_PTR (skel,"get_smp_processor_id__open_and_load" )) + goto cleanup; + + prog_fd =bpf_program__fd (skel->progs.call_bpf_get_smp_processor_id); + +for (i =0 ; i < online_cpu_nr; i++) { +if (!online[i]) +continue ; + + opts.cpu = i; + skel->bss->cpu_nr_result =-1 ; + + err =bpf_prog_test_run_opts (prog_fd, &opts); +if (!ASSERT_OK (err,"bpf_prog_test_run_opts" )) + goto cleanup; + +ASSERT_EQ (skel->bss->cpu_nr_result, opts.cpu,"cpu_nr_result" ); + } + +cleanup: +free (online); +get_smp_processor_id__destroy (skel); +}
tools/testing/selftests/bpf/progs/get_smp_processor_id.c new file +20 -0
@@ -0,0 +1,20 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <linux/bpf.h> +#include <bpf/bpf_helpers.h> +#include "bpf_misc.h" + +__u64 cpu_nr_result; + +SEC("raw_tp") +void call_bpf_get_smp_processor_id (void ) +{ + register __u64 r0 asm("r0") = -1; + asmvolatile ("call %[bpf_get_smp_processor_id];" + :"+r" (r0) + : __imm(bpf_get_smp_processor_id) + : "r1", "r2", "r3", "r4", "r5", "memory"); + cpu_nr_result = r0; +} + +char _license[] SEC("license") = "GPL";
Signature · 1 line
2.54.0
This series was applied to bpf/bpf-next.git (master) by Daniel Borkmann <[email protected]>:
Hello:
This series was applied to bpf/bpf-next.git (master) by Daniel Borkmann <[email protected]>:
Quoted 9 lines · On Fri, 3 Jul 2026 14:51:34 +0200 you wrote:
Sashiko found a bug [1] in
9012cf2491e3, where upper bits of a 64-bit register are not zero-extended after ly instruction in used.Tested:
./test_progs -v -tget_smp_processor_id[1] https://sashiko.dev/#/patchset/20260414142930.528751-1-max%40linux.ibm.com
[...]
Here is the summary with links:
- [bpf-next,1/2] s390/bpf: Replace ly instruction with llgf
https://git.kernel.org/bpf/bpf-next/c/4abb1bdb600b - [bpf-next,2/2] selftests/bpf: Add test for bpf_get_smp_processor_id https://git.kernel.org/bpf/bpf-next/c/f71f6c98d1fe
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html