Hi Menglong, kernel test robot noticed the following build errors:
[auto build test ERROR on bpf-next/master] url: https://github.com/intel-lab-lkp/linux/commits/Menglong-Dong/ftrace-introduce-FTRACE_OPS_FL_JMP/20251117-115243 base: https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git master patch link: https://lore.kernel.org/r/20251117034906.32036-6-dongml2%40chinatelecom.cn patch subject: [PATCH bpf-next v2 5/6] bpf: specify the old and new poke_type for bpf_arch_text_poke config: powerpc64-randconfig-002-20251118 (https://download.01.org/0day-ci/archive/20251118/[email protected]/config) compiler: powerpc64-linux-gcc (GCC) 8.5.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20251118/[email protected]/reproduce) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <[email protected]> | Closes: https://lore.kernel.org/oe-kbuild-all/[email protected]/ All errors (new ones prefixed by >>): arch/powerpc/net/bpf_jit_comp.c: In function 'bpf_arch_text_poke': >> arch/powerpc/net/bpf_jit_comp.c:1135:7: error: 'poke_type' undeclared (first >> use in this function); did you mean 'probe_type'? if (poke_type != BPF_MOD_JUMP) { ^~~~~~~~~ probe_type arch/powerpc/net/bpf_jit_comp.c:1135:7: note: each undeclared identifier is reported only once for each function it appears in vim +1135 arch/powerpc/net/bpf_jit_comp.c d243b62b7bd3d5 Naveen N Rao 2024-10-30 1070 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1071 /* d243b62b7bd3d5 Naveen N Rao 2024-10-30 1072 * A 3-step process for bpf prog entry: d243b62b7bd3d5 Naveen N Rao 2024-10-30 1073 * 1. At bpf prog entry, a single nop/b: d243b62b7bd3d5 Naveen N Rao 2024-10-30 1074 * bpf_func: d243b62b7bd3d5 Naveen N Rao 2024-10-30 1075 * [nop|b] ool_stub d243b62b7bd3d5 Naveen N Rao 2024-10-30 1076 * 2. Out-of-line stub: d243b62b7bd3d5 Naveen N Rao 2024-10-30 1077 * ool_stub: d243b62b7bd3d5 Naveen N Rao 2024-10-30 1078 * mflr r0 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1079 * [b|bl] <bpf_prog>/<long_branch_stub> d243b62b7bd3d5 Naveen N Rao 2024-10-30 1080 * mtlr r0 // CONFIG_PPC_FTRACE_OUT_OF_LINE only d243b62b7bd3d5 Naveen N Rao 2024-10-30 1081 * b bpf_func + 4 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1082 * 3. Long branch stub: d243b62b7bd3d5 Naveen N Rao 2024-10-30 1083 * long_branch_stub: d243b62b7bd3d5 Naveen N Rao 2024-10-30 1084 * .long <branch_addr>/<dummy_tramp> d243b62b7bd3d5 Naveen N Rao 2024-10-30 1085 * mflr r11 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1086 * bcl 20,31,$+4 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1087 * mflr r12 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1088 * ld r12, -16(r12) d243b62b7bd3d5 Naveen N Rao 2024-10-30 1089 * mtctr r12 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1090 * mtlr r11 // needed to retain ftrace ABI d243b62b7bd3d5 Naveen N Rao 2024-10-30 1091 * bctr d243b62b7bd3d5 Naveen N Rao 2024-10-30 1092 * d243b62b7bd3d5 Naveen N Rao 2024-10-30 1093 * dummy_tramp is used to reduce synchronization requirements. d243b62b7bd3d5 Naveen N Rao 2024-10-30 1094 * d243b62b7bd3d5 Naveen N Rao 2024-10-30 1095 * When attaching a bpf trampoline to a bpf prog, we do not need any d243b62b7bd3d5 Naveen N Rao 2024-10-30 1096 * synchronization here since we always have a valid branch target regardless d243b62b7bd3d5 Naveen N Rao 2024-10-30 1097 * of the order in which the above stores are seen. dummy_tramp ensures that d243b62b7bd3d5 Naveen N Rao 2024-10-30 1098 * the long_branch stub goes to a valid destination on other cpus, even when d243b62b7bd3d5 Naveen N Rao 2024-10-30 1099 * the branch to the long_branch stub is seen before the updated trampoline d243b62b7bd3d5 Naveen N Rao 2024-10-30 1100 * address. d243b62b7bd3d5 Naveen N Rao 2024-10-30 1101 * d243b62b7bd3d5 Naveen N Rao 2024-10-30 1102 * However, when detaching a bpf trampoline from a bpf prog, or if changing d243b62b7bd3d5 Naveen N Rao 2024-10-30 1103 * the bpf trampoline address, we need synchronization to ensure that other d243b62b7bd3d5 Naveen N Rao 2024-10-30 1104 * cpus can no longer branch into the older trampoline so that it can be d243b62b7bd3d5 Naveen N Rao 2024-10-30 1105 * safely freed. bpf_tramp_image_put() uses rcu_tasks to ensure all cpus d243b62b7bd3d5 Naveen N Rao 2024-10-30 1106 * make forward progress, but we still need to ensure that other cpus d243b62b7bd3d5 Naveen N Rao 2024-10-30 1107 * execute isync (or some CSI) so that they don't go back into the d243b62b7bd3d5 Naveen N Rao 2024-10-30 1108 * trampoline again. d243b62b7bd3d5 Naveen N Rao 2024-10-30 1109 */ 7cc5910294285d Menglong Dong 2025-11-17 1110 int bpf_arch_text_poke(void *ip, enum bpf_text_poke_type old_t, 7cc5910294285d Menglong Dong 2025-11-17 1111 enum bpf_text_poke_type new_t, void *old_addr, 7cc5910294285d Menglong Dong 2025-11-17 1112 void *new_addr) d243b62b7bd3d5 Naveen N Rao 2024-10-30 1113 { d243b62b7bd3d5 Naveen N Rao 2024-10-30 1114 unsigned long bpf_func, bpf_func_end, size, offset; d243b62b7bd3d5 Naveen N Rao 2024-10-30 1115 ppc_inst_t old_inst, new_inst; d243b62b7bd3d5 Naveen N Rao 2024-10-30 1116 int ret = 0, branch_flags; d243b62b7bd3d5 Naveen N Rao 2024-10-30 1117 char name[KSYM_NAME_LEN]; d243b62b7bd3d5 Naveen N Rao 2024-10-30 1118 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1119 if (IS_ENABLED(CONFIG_PPC32)) d243b62b7bd3d5 Naveen N Rao 2024-10-30 1120 return -EOPNOTSUPP; d243b62b7bd3d5 Naveen N Rao 2024-10-30 1121 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1122 bpf_func = (unsigned long)ip; d243b62b7bd3d5 Naveen N Rao 2024-10-30 1123 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1124 /* We currently only support poking bpf programs */ d243b62b7bd3d5 Naveen N Rao 2024-10-30 1125 if (!__bpf_address_lookup(bpf_func, &size, &offset, name)) { d243b62b7bd3d5 Naveen N Rao 2024-10-30 1126 pr_err("%s (0x%lx): kernel/modules are not supported\n", __func__, bpf_func); d243b62b7bd3d5 Naveen N Rao 2024-10-30 1127 return -EOPNOTSUPP; d243b62b7bd3d5 Naveen N Rao 2024-10-30 1128 } d243b62b7bd3d5 Naveen N Rao 2024-10-30 1129 d243b62b7bd3d5 Naveen N Rao 2024-10-30 1130 /* d243b62b7bd3d5 Naveen N Rao 2024-10-30 1131 * If we are not poking at bpf prog entry, then we are simply patching in/out d243b62b7bd3d5 Naveen N Rao 2024-10-30 1132 * an unconditional branch instruction at im->ip_after_call d243b62b7bd3d5 Naveen N Rao 2024-10-30 1133 */ d243b62b7bd3d5 Naveen N Rao 2024-10-30 1134 if (offset) { d243b62b7bd3d5 Naveen N Rao 2024-10-30 @1135 if (poke_type != BPF_MOD_JUMP) { -- 0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
