s390/bpf: Add s390 JIT support for timed may_goto
Subsystems (5)
Ilya Leoshkevich
maintainerv2 cover s390/bpf: Add s390 JIT support for timed may_goto
v1: https://lore.kernel.org/bpf/[email protected]/
v1 -> v2: Fix test_stream_errors (caught by CI).
Hi,
This series adds timed may_goto implementation to the s390x JIT. Patch 1 is the implementation itself, patches 2-5 are the associated test changes.
Best regards,
Ilya
Ilya Leoshkevich (5):
s390/bpf: Add s390 JIT support for timed may_goto selftests/bpf: Add a missing newline to the "bad arch spec" message selftests/bpf: Add __arch_s390x macro
selftests/bpf: Enable timed may_goto verifier tests on s390x selftests/bpf: Remove may_goto tests from DENYLIST.s390x
arch/s390/net/Makefile | 2 +- arch/s390/net/bpf_jit_comp.c | 25 +++++++++-- arch/s390/net/bpf_timed_may_goto.S | 45 +++++++++++++++++++ tools/testing/selftests/bpf/DENYLIST.s390x | 1 - .../testing/selftests/bpf/prog_tests/stream.c | 2 +- tools/testing/selftests/bpf/progs/bpf_misc.h | 1 + .../selftests/bpf/progs/verifier_may_goto_1.c | 8 +++- tools/testing/selftests/bpf/test_loader.c | 7 ++- 8 files changed, 81 insertions(+), 10 deletions(-) create mode 100644 arch/s390/net/bpf_timed_may_goto.S
Signature · 1 line
2.50.1
Ilya Leoshkevich
maintainerPATCH v2 1/5 s390/bpf: Add s390 JIT support for timed may_goto
The verifier provides an architecture-independent implementation of the may_goto instruction, which is currently used on s390x, but it has a
downside: there is no way to prevent progs using it from running for a very long time.
The solution to this problem is an alternative timed implementation, which requires architecture-specific bits. Its availability is signaled to the verifier by bpf_jit_supports_timed_may_goto() returning true.
The verifier then emits a call to arch_bpf_timed_may_goto() using a non-standard calling convention. This function must act as a trampoline for bpf_check_timed_may_goto().
Implement bpf_jit_supports_timed_may_goto(), account for the special calling convention in the BPF_CALL implementation, and implement arch_bpf_timed_may_goto().
--- arch/s390/net/Makefile | 2 +- arch/s390/net/bpf_jit_comp.c | 25 ++++++++++++++--- arch/s390/net/bpf_timed_may_goto.S | 45 ++++++++++++++++++++++++++++++ 3 files changed, 67 insertions(+), 5 deletions(-) create mode 100644 arch/s390/net/bpf_timed_may_goto.S
arch/s390/net/Makefile +1 -1
@@ -2,5 +2,5 @@ # # Arch-specific network modules # -obj-$(CONFIG_BPF_JIT) += bpf_jit_comp.o +obj-$(CONFIG_BPF_JIT) += bpf_jit_comp.o bpf_timed_may_goto.o obj-$(CONFIG_HAVE_PNETID) += pnet.o
arch/s390/net/bpf_jit_comp.c +21 -4
@@ -1806,10 +1806,22 @@ static noinline int bpf_jit_insn(struct bpf_jit *jit, struct bpf_prog *fp, } } -/* brasl %r14,func */ -EMIT6_PCREL_RILB_PTR (0xc0050000 ,REG_14 , (void *)func); -/* lgr %b0,%r2: load return value into %b0 */ -EMIT4 (0xb9040000 ,BPF_REG_0 ,REG_2 ); +if ((void *)func == arch_bpf_timed_may_goto) { +/* +* arch_bpf_timed_may_goto() has a special ABI: the +* parameters are in BPF_REG_AX and BPF_REG_10; the +* return value is in BPF_REG_AX; and all GPRs except +* REG_W0, REG_W1, and BPF_REG_AX are callee-saved. +*/ + +/* brasl %r0,func */ +EMIT6_PCREL_RILB_PTR (0xc0050000 ,REG_0 , (void *)func); + }else { +/* brasl %r14,func */ +EMIT6_PCREL_RILB_PTR (0xc0050000 ,REG_14 , (void *)func); +/* lgr %b0,%r2: load return value into %b0 */ +EMIT4 (0xb9040000 ,BPF_REG_0 ,REG_2 ); + }/* * Copy the potentially updated tail call counter back.
@@ -2993,3 +3005,8 @@ void arch_bpf_stack_walk(bool (*consume_fn)(void *, u64, u64, u64), prev_addr = addr; } } + +bool bpf_jit_supports_timed_may_goto (void ) +{ +return true; +}
arch/s390/net/bpf_timed_may_goto.S new file +45 -0
45 lines not on this page · raw message
Signature · 1 line
2.50.1
Ilya Leoshkevich
maintainerPATCH v2 2/5 selftests/bpf: Add a missing newline to the "bad arch spec" message
In reply to Ilya LeoshkevichFix error messages like this one:
parse_test_spec:FAIL:569 bad arch spec: 's390x'process_subtest:FAIL:1153 Can't parse test spec for program 'may_goto_simple'
--- tools/testing/selftests/bpf/test_loader.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/test_loader.c +1 -1
@@ -566,7 +566,7 @@ static int parse_test_spec(struct test_loader *tester, }else if (strcmp (val,"RISCV64" ) ==0 ) { arch =ARCH_RISCV64 ; }else { -PRINT_FAIL ("bad arch spec: '%s'" , val); +PRINT_FAIL ("bad arch spec: '%s'\n" , val); err = -EINVAL ; goto cleanup; }
Signature · 1 line
2.50.1
Ilya Leoshkevich
maintainerPATCH v2 3/5 selftests/bpf: Add __arch_s390x macro
In reply to Ilya LeoshkevichMake it possible to limit certain tests to s390x, just like it's already done for x86_64, arm64, and riscv64.
--- tools/testing/selftests/bpf/progs/bpf_misc.h | 1 + tools/testing/selftests/bpf/test_loader.c | 5 +++++ 2 files changed, 6 insertions(+)
tools/testing/selftests/bpf/progs/bpf_misc.h +1 -0
@@ -136,6 +136,7 @@#define __arch_x86_64 __arch("X86_64")#define __arch_arm64 __arch("ARM64")#define __arch_riscv64 __arch("RISCV64") +#define __arch_s390x __arch("s390x")#define __caps_unpriv (caps) __attribute__((btf_decl_tag("comment:test_caps_unpriv=" EXPAND_QUOTE(caps))))#define __load_if_JITed () __attribute__((btf_decl_tag("comment:load_mode=jited")))#define __load_if_no_JITed () __attribute__((btf_decl_tag("comment:load_mode=no_jited")))
tools/testing/selftests/bpf/test_loader.c +5 -0
@@ -374,6 +374,7 @@ enum arch {ARCH_X86_64 =0x2 ,ARCH_ARM64 =0x4 ,ARCH_RISCV64 =0x8 , + ARCH_S390X = 0x10, };static int get_current_arch (void )
@@ -384,6 +385,8 @@ static int get_current_arch(void)return ARCH_ARM64 ;#elif defined(__riscv) && __riscv_xlen ==64 return ARCH_RISCV64 ; +#elif defined(__s390x__) +return ARCH_S390X ;#endif return ARCH_UNKNOWN ; }
@@ -565,6 +568,8 @@ static int parse_test_spec(struct test_loader *tester, arch =ARCH_ARM64 ; }else if (strcmp (val,"RISCV64" ) ==0 ) { arch =ARCH_RISCV64 ; + }else if (strcmp (val,"s390x" ) ==0 ) { + arch =ARCH_S390X ; }else {PRINT_FAIL ("bad arch spec: '%s'\n" , val); err = -EINVAL ;
Signature · 1 line
2.50.1
Ilya Leoshkevich
maintainerPATCH v2 4/5 selftests/bpf: Enable timed may_goto verifier tests on s390x
In reply to Ilya LeoshkevichNow that the timed may_goto implementation is available on s390x, enable the respective verifier tests.
--- tools/testing/selftests/bpf/prog_tests/stream.c | 2 +- tools/testing/selftests/bpf/progs/verifier_may_goto_1.c | 8 ++++++-- 2 files changed, 7 insertions(+), 3 deletions(-)
tools/testing/selftests/bpf/prog_tests/stream.c +1 -1
@@ -77,7 +77,7 @@ void test_stream_errors(void)ASSERT_OK (ret,"ret" );ASSERT_OK (opts.retval,"retval" ); -#if !defined(__x86_64__) +#if !defined(__x86_64__) && !defined(__s390x__)ASSERT_TRUE (1 ,"Timed may_goto unsupported, skip." );if (i ==0 ) { ret =bpf_prog_stream_read (prog_fd,2 , buf,sizeof (buf), &ropts);
tools/testing/selftests/bpf/progs/verifier_may_goto_1.c +6 -2
@@ -9,6 +9,7 @@ SEC("raw_tp") __description("may_goto 0") __arch_x86_64 +__arch_s390x __xlated("0: r0 = 1")__xlated ("1: exit" ) __success
@@ -27,6 +28,7 @@ __naked void may_goto_simple(void) SEC("raw_tp") __description("batch 2 of may_goto 0") __arch_x86_64 +__arch_s390x __xlated("0: r0 = 1")__xlated ("1: exit" ) __success
@@ -47,6 +49,7 @@ __naked void may_goto_batch_0(void) SEC("raw_tp") __description("may_goto batch with offsets 2/1/0") __arch_x86_64 +__arch_s390x __xlated("0: r0 = 1")__xlated ("1: exit" ) __success
@@ -69,8 +72,9 @@ __naked void may_goto_batch_1(void) } SEC("raw_tp") -__description("may_goto batch with offsets 2/0 - x86_64") +__description("may_goto batch with offsets 2/0 - x86_64 and s390x") __arch_x86_64 +__arch_s390x __xlated("0: *(u64 *)(r10 -16) = 65535") __xlated("1: *(u64 *)(r10 -8) = 0")__xlated ("2: r11 = *(u64 *)(r10 -16)" )
@@ -84,7 +88,7 @@ __xlated("9: r0 = 1") __xlated("10: r0 = 2") __xlated("11: exit") __success -__naked void may_goto_batch_2_x86_64 (void ) +__naked void may_goto_batch_2_x86_64_s390x (void ) { asmvolatile ( ".8byte %[may_goto1];"
Signature · 1 line
2.50.1
Ilya Leoshkevich
maintainerPATCH v2 5/5 selftests/bpf: Remove may_goto tests from DENYLIST.s390x
In reply to Ilya LeoshkevichThe may_goto instruction is now fully supported on s390x, including the timed implementation, so remove the respective test from the denylist.
--- tools/testing/selftests/bpf/DENYLIST.s390x | 1 - 1 file changed, 1 deletion(-)
tools/testing/selftests/bpf/DENYLIST.s390x +0 -1
@@ -2,4 +2,3 @@ # Alphabetical order get_stack_raw_tp # user_stack corrupted user stack (no backchain userspace) stacktrace_build_id # compare_map_keys stackid_hmap vs. stackmap err -2 errno 2 (?) -verifier_iterating_callbacks
Signature · 1 line
2.50.1
This series was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:
Hello:
This series was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:
Quoted 10 lines · On Thu, 21 Aug 2025 13:25:54 +0200 you wrote:
v1: https://lore.kernel.org/bpf/[email protected]/
v1 -> v2: Fixtest_stream_errors(caught by CI).Hi,
This series adds timed
may_gotoimplementation to the s390x JIT. Patch 1 is the implementation itself, patches 2-5 are the associated test changes.[...]
Here is the summary with links:
- [bpf-next,v2,1/5] s390/bpf: Add s390 JIT support for timed may_goto
https://git.kernel.org/bpf/bpf-next/c/b8efa810c1db - [bpf-next,v2,2/5] selftests/bpf: Add a missing newline to the "bad arch spec" message https://git.kernel.org/bpf/bpf-next/c/b68dfcc12a32 - [bpf-next,v2,3/5] selftests/bpf: Add __arch_s390x macro https://git.kernel.org/bpf/bpf-next/c/1e4e6b9e260d - [bpf-next,v2,4/5] selftests/bpf: Enable timed may_goto verifier tests on s390x https://git.kernel.org/bpf/bpf-next/c/7197dbcba230 - [bpf-next,v2,5/5] selftests/bpf: Remove may_goto tests from DENYLIST.s390x https://git.kernel.org/bpf/bpf-next/c/21bce5694054
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html