diff --git a/arch/arm64/net/bpf_jit_comp.c b/arch/arm64/net/bpf_jit_comp.c index 9eb0ea9a644f171e3484872c8154b07739119002..6bcce848a720fa7a7d0048272f038fa289dd0878 100644 --- a/arch/arm64/net/bpf_jit_comp.c +++ b/arch/arm64/net/bpf_jit_comp.c @@ -153,14 +153,17 @@ static inline void emit_addr_mov_i64(const int reg, const u64 val, } } -static inline int bpf2a64_offset(int bpf_to, int bpf_from, +static inline int bpf2a64_offset(int bpf_insn, int off, const struct jit_ctx *ctx) { - int to = ctx->offset[bpf_to]; - /* -1 to account for the Branch instruction */ - int from = ctx->offset[bpf_from] - 1; - - return to - from; + /* BPF JMP offset is relative to the next instruction */ + bpf_insn++; + /* + * Whereas arm64 branch instructions encode the offset + * from the branch itself, so we must subtract 1 from the + * instruction offset. + */ + return ctx->offset[bpf_insn + off] - (ctx->offset[bpf_insn] - 1); } static void jit_fill_hole(void *area, unsigned int size) @@ -546,7 +549,7 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, /* JUMP off */ case BPF_JMP | BPF_JA: - jmp_offset = bpf2a64_offset(i + off, i, ctx); + jmp_offset = bpf2a64_offset(i, off, ctx); check_imm26(jmp_offset); emit(A64_B(jmp_offset), ctx); break; @@ -573,7 +576,7 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, case BPF_JMP32 | BPF_JSLE | BPF_X: emit(A64_CMP(is64, dst, src), ctx); emit_cond_jmp: - jmp_offset = bpf2a64_offset(i + off, i, ctx); + jmp_offset = bpf2a64_offset(i, off, ctx); check_imm19(jmp_offset); switch (BPF_OP(code)) { case BPF_JEQ: @@ -798,10 +801,21 @@ static int build_body(struct jit_ctx *ctx, bool extra_pass) const struct bpf_prog *prog = ctx->prog; int i; + /* + * - offset[0] offset of the end of prologue, + * start of the 1st instruction. + * - offset[1] - offset of the end of 1st instruction, + * start of the 2nd instruction + * [....] + * - offset[3] - offset of the end of 3rd instruction, + * start of 4th instruction + */ for (i = 0; i < prog->len; i++) { const struct bpf_insn *insn = &prog->insnsi[i]; int ret; + if (ctx->image == NULL) + ctx->offset[i] = ctx->idx; ret = build_insn(insn, ctx, extra_pass); if (ret > 0) { i++; @@ -809,11 +823,16 @@ static int build_body(struct jit_ctx *ctx, bool extra_pass) ctx->offset[i] = ctx->idx; continue; } - if (ctx->image == NULL) - ctx->offset[i] = ctx->idx; if (ret) return ret; } + /* + * offset is allocated with prog->len + 1 so fill in + * the last element with the offset after the last + * instruction (end of program) + */ + if (ctx->image == NULL) + ctx->offset[i] = ctx->idx; return 0; } @@ -889,7 +908,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog) memset(&ctx, 0, sizeof(ctx)); ctx.prog = prog; - ctx.offset = kcalloc(prog->len, sizeof(int), GFP_KERNEL); + ctx.offset = kcalloc(prog->len + 1, sizeof(int), GFP_KERNEL); if (ctx.offset == NULL) { prog = orig_prog; goto out_off; @@ -972,7 +991,12 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog) prog->jited_len = image_size; if (!prog->is_func || extra_pass) { - bpf_prog_fill_jited_linfo(prog, ctx.offset); + int i; + + /* offset[prog->len] is the size of program */ + for (i = 0; i <= prog->len; i++) + ctx.offset[i] *= AARCH64_INSN_SIZE; + bpf_prog_fill_jited_linfo(prog, ctx.offset + 1); out_off: kfree(ctx.offset); kfree(jit_data); diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c index 21faf6d3cc680f559d23cccce047e26ef5632973..0f3125fc4188053f02a61a5d550b5ede86d6e752 100644 --- a/arch/x86/net/bpf_jit_comp.c +++ b/arch/x86/net/bpf_jit_comp.c @@ -394,8 +394,9 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, emit_prologue(&prog, bpf_prog->aux->stack_depth, bpf_prog_was_classic(bpf_prog)); + addrs[0] = prog - temp; - for (i = 0; i < insn_cnt; i++, insn++) { + for (i = 1; i <= insn_cnt; i++, insn++) { const s32 imm32 = insn->imm; u32 dst_reg = insn->dst_reg; u32 src_reg = insn->src_reg; @@ -1130,7 +1131,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog) extra_pass = true; goto skip_init_addrs; } - addrs = kmalloc_array(prog->len, sizeof(*addrs), GFP_KERNEL); + addrs = kmalloc_array(prog->len + 1, sizeof(*addrs), GFP_KERNEL); if (!addrs) { prog = orig_prog; goto out_addrs; @@ -1140,7 +1141,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog) * Before first pass, make a rough estimation of addrs[] * each BPF instruction is translated to less than 64 bytes */ - for (proglen = 0, i = 0; i < prog->len; i++) { + for (proglen = 0, i = 0; i <= prog->len; i++) { proglen += 64; addrs[i] = proglen; } @@ -1205,7 +1206,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog) if (!image || !prog->is_func || extra_pass) { if (image) - bpf_prog_fill_jited_linfo(prog, addrs); + bpf_prog_fill_jited_linfo(prog, addrs + 1); out_addrs: kfree(addrs); kfree(jit_data); diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h index 0b96b1077c6d24a4499a96f3e2fc78754ab00ca9..16f957d08cc5aef0298d4cb09275cb98dd7c9131 100644 --- a/include/linux/bpf_verifier.h +++ b/include/linux/bpf_verifier.h @@ -139,6 +139,8 @@ struct bpf_reg_state { */ s32 subreg_def; enum bpf_reg_liveness live; + /* if (!precise && SCALAR_VALUE) min/max/tnum don't affect safety */ + bool precise; }; enum bpf_stack_slot_type { @@ -190,14 +192,77 @@ struct bpf_func_state { struct bpf_stack_state *stack; }; +struct bpf_idx_pair { + u32 prev_idx; + u32 idx; +}; + #define MAX_CALL_FRAMES 8 struct bpf_verifier_state { /* call stack tracking */ struct bpf_func_state *frame[MAX_CALL_FRAMES]; + struct bpf_verifier_state *parent; + /* + * 'branches' field is the number of branches left to explore: + * 0 - all possible paths from this state reached bpf_exit or + * were safely pruned + * 1 - at least one path is being explored. + * This state hasn't reached bpf_exit + * 2 - at least two paths are being explored. + * This state is an immediate parent of two children. + * One is fallthrough branch with branches==1 and another + * state is pushed into stack (to be explored later) also with + * branches==1. The parent of this state has branches==1. + * The verifier state tree connected via 'parent' pointer looks like: + * 1 + * 1 + * 2 -> 1 (first 'if' pushed into stack) + * 1 + * 2 -> 1 (second 'if' pushed into stack) + * 1 + * 1 + * 1 bpf_exit. + * + * Once do_check() reaches bpf_exit, it calls update_branch_counts() + * and the verifier state tree will look: + * 1 + * 1 + * 2 -> 1 (first 'if' pushed into stack) + * 1 + * 1 -> 1 (second 'if' pushed into stack) + * 0 + * 0 + * 0 bpf_exit. + * After pop_stack() the do_check() will resume at second 'if'. + * + * If is_state_visited() sees a state with branches > 0 it means + * there is a loop. If such state is exactly equal to the current state + * it's an infinite loop. Note states_equal() checks for states + * equvalency, so two states being 'states_equal' does not mean + * infinite loop. The exact comparison is provided by + * states_maybe_looping() function. It's a stronger pre-check and + * much faster than states_equal(). + * + * This algorithm may not find all possible infinite loops or + * loop iteration count may be too high. + * In such cases BPF_COMPLEXITY_LIMIT_INSNS limit kicks in. + */ + u32 branches; u32 insn_idx; u32 curframe; u32 active_spin_lock; bool speculative; + + /* first and last insn idx of this verifier state */ + u32 first_insn_idx; + u32 last_insn_idx; + /* jmp history recorded from first to last. + * backtracking is using it to go from last to first. + * For most states jmp_history_cnt is [0-3]. + * For loops can go up to ~40. + */ + struct bpf_idx_pair *jmp_history; + u32 jmp_history_cnt; }; #define bpf_get_spilled_reg(slot, frame) \ @@ -313,7 +378,9 @@ struct bpf_verifier_env { } cfg; u32 subprog_cnt; /* number of instructions analyzed by the verifier */ - u32 insn_processed; + u32 prev_insn_processed, insn_processed; + /* number of jmps, calls, exits analyzed so far */ + u32 prev_jmps_processed, jmps_processed; /* total verification time */ u64 verification_time; /* maximum number of verifier states kept in 'branching' instructions */ diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h index 8e7658d571b8828c7a0479eb7215d629fb16945a..ed78ecb3c2044e94905c0889268a65e1cda713c6 100644 --- a/include/uapi/linux/bpf.h +++ b/include/uapi/linux/bpf.h @@ -3231,6 +3231,7 @@ struct bpf_sock_addr { __u32 msg_src_ip6[4]; /* Allows 1,2,4-byte read an 4-byte write. * Stored in network byte order. */ + __bpf_md_ptr(struct bpf_sock *, sk); }; /* User bpf_sock_ops struct to access socket values and specify request ops @@ -3282,6 +3283,7 @@ struct bpf_sock_ops { __u32 sk_txhash; __u64 bytes_received; __u64 bytes_acked; + __bpf_md_ptr(struct bpf_sock *, sk); }; /* Definitions for bpf_sock_ops_cb_flags */ diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index d8a9cff087b72e209ab9587d88897615f8c3ac42..c667d9e3b571387eb636eb6e46c20c864321b9c1 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -454,12 +454,12 @@ static void print_verifier_state(struct bpf_verifier_env *env, verbose(env, " R%d", i); print_liveness(env, reg->live); verbose(env, "=%s", reg_type_str[t]); + if (t == SCALAR_VALUE && reg->precise) + verbose(env, "P"); if ((t == SCALAR_VALUE || t == PTR_TO_STACK) && tnum_is_const(reg->var_off)) { /* reg->off should be 0 for SCALAR_VALUE */ verbose(env, "%lld", reg->var_off.value + reg->off); - if (t == PTR_TO_STACK) - verbose(env, ",call_%d", func(env, reg)->callsite); } else { verbose(env, "(id=%d", reg->id); if (reg_type_may_be_refcounted_or_null(t)) @@ -521,11 +521,17 @@ static void print_verifier_state(struct bpf_verifier_env *env, continue; verbose(env, " fp%d", (-i - 1) * BPF_REG_SIZE); print_liveness(env, state->stack[i].spilled_ptr.live); - if (state->stack[i].slot_type[0] == STACK_SPILL) - verbose(env, "=%s", - reg_type_str[state->stack[i].spilled_ptr.type]); - else + if (state->stack[i].slot_type[0] == STACK_SPILL) { + reg = &state->stack[i].spilled_ptr; + t = reg->type; + verbose(env, "=%s", reg_type_str[t]); + if (t == SCALAR_VALUE && reg->precise) + verbose(env, "P"); + if (t == SCALAR_VALUE && tnum_is_const(reg->var_off)) + verbose(env, "%lld", reg->var_off.value + reg->off); + } else { verbose(env, "=%s", types_buf); + } } if (state->acquired_refs && state->refs[0].id) { verbose(env, " refs=%d", state->refs[0].id); @@ -674,6 +680,13 @@ static void free_func_state(struct bpf_func_state *state) kfree(state); } +static void clear_jmp_history(struct bpf_verifier_state *state) +{ + kfree(state->jmp_history); + state->jmp_history = NULL; + state->jmp_history_cnt = 0; +} + static void free_verifier_state(struct bpf_verifier_state *state, bool free_self) { @@ -683,6 +696,7 @@ static void free_verifier_state(struct bpf_verifier_state *state, free_func_state(state->frame[i]); state->frame[i] = NULL; } + clear_jmp_history(state); if (free_self) kfree(state); } @@ -710,8 +724,18 @@ static int copy_verifier_state(struct bpf_verifier_state *dst_state, const struct bpf_verifier_state *src) { struct bpf_func_state *dst; + u32 jmp_sz = sizeof(struct bpf_idx_pair) * src->jmp_history_cnt; int i, err; + if (dst_state->jmp_history_cnt < src->jmp_history_cnt) { + kfree(dst_state->jmp_history); + dst_state->jmp_history = kmalloc(jmp_sz, GFP_USER); + if (!dst_state->jmp_history) + return -ENOMEM; + } + memcpy(dst_state->jmp_history, src->jmp_history, jmp_sz); + dst_state->jmp_history_cnt = src->jmp_history_cnt; + /* if dst has more stack frames then src frame, free them */ for (i = src->curframe + 1; i <= dst_state->curframe; i++) { free_func_state(dst_state->frame[i]); @@ -720,6 +744,10 @@ static int copy_verifier_state(struct bpf_verifier_state *dst_state, dst_state->speculative = src->speculative; dst_state->curframe = src->curframe; dst_state->active_spin_lock = src->active_spin_lock; + dst_state->branches = src->branches; + dst_state->parent = src->parent; + dst_state->first_insn_idx = src->first_insn_idx; + dst_state->last_insn_idx = src->last_insn_idx; for (i = 0; i <= src->curframe; i++) { dst = dst_state->frame[i]; if (!dst) { @@ -735,6 +763,23 @@ static int copy_verifier_state(struct bpf_verifier_state *dst_state, return 0; } +static void update_branch_counts(struct bpf_verifier_env *env, struct bpf_verifier_state *st) +{ + while (st) { + u32 br = --st->branches; + + /* WARN_ON(br > 1) technically makes sense here, + * but see comment in push_stack(), hence: + */ + WARN_ONCE((int)br < 0, + "BUG update_branch_counts:branches_to_explore=%d\n", + br); + if (br) + break; + st = st->parent; + } +} + static int pop_stack(struct bpf_verifier_env *env, int *prev_insn_idx, int *insn_idx) { @@ -788,6 +833,18 @@ static struct bpf_verifier_state *push_stack(struct bpf_verifier_env *env, env->stack_size); goto err; } + if (elem->st.parent) { + ++elem->st.parent->branches; + /* WARN_ON(branches > 2) technically makes sense here, + * but + * 1. speculative states will bump 'branches' for non-branch + * instructions + * 2. is_state_visited() heuristics may decide not to create + * a new state for a sequence of branches and all such current + * and cloned states will be pointing to a single parent state + * which might have large 'branches' count. + */ + } return &elem->st; err: free_verifier_state(env->cur_state, true); @@ -802,7 +859,8 @@ static const int caller_saved[CALLER_SAVED_REGS] = { BPF_REG_0, BPF_REG_1, BPF_REG_2, BPF_REG_3, BPF_REG_4, BPF_REG_5 }; -static void __mark_reg_not_init(struct bpf_reg_state *reg); +static void __mark_reg_not_init(const struct bpf_verifier_env *env, + struct bpf_reg_state *reg); /* Mark the unknown part of a register (variable offset or scalar value) as * known to have the value @imm. @@ -840,7 +898,7 @@ static void mark_reg_known_zero(struct bpf_verifier_env *env, verbose(env, "mark_reg_known_zero(regs, %u)\n", regno); /* Something bad happened, let's kill all regs */ for (regno = 0; regno < MAX_BPF_REG; regno++) - __mark_reg_not_init(regs + regno); + __mark_reg_not_init(env, regs + regno); return; } __mark_reg_known_zero(regs + regno); @@ -938,7 +996,8 @@ static void __mark_reg_unbounded(struct bpf_reg_state *reg) } /* Mark a register as having a completely unknown (scalar) value. */ -static void __mark_reg_unknown(struct bpf_reg_state *reg) +static void __mark_reg_unknown(const struct bpf_verifier_env *env, + struct bpf_reg_state *reg) { /* * Clear type, id, off, and union(map_ptr, range) and @@ -948,6 +1007,8 @@ static void __mark_reg_unknown(struct bpf_reg_state *reg) reg->type = SCALAR_VALUE; reg->var_off = tnum_unknown; reg->frameno = 0; + reg->precise = env->subprog_cnt > 1 || !env->allow_ptr_leaks ? + true : false; __mark_reg_unbounded(reg); } @@ -958,15 +1019,16 @@ static void mark_reg_unknown(struct bpf_verifier_env *env, verbose(env, "mark_reg_unknown(regs, %u)\n", regno); /* Something bad happened, let's kill all regs except FP */ for (regno = 0; regno < BPF_REG_FP; regno++) - __mark_reg_not_init(regs + regno); + __mark_reg_not_init(env, regs + regno); return; } - __mark_reg_unknown(regs + regno); + __mark_reg_unknown(env, regs + regno); } -static void __mark_reg_not_init(struct bpf_reg_state *reg) +static void __mark_reg_not_init(const struct bpf_verifier_env *env, + struct bpf_reg_state *reg) { - __mark_reg_unknown(reg); + __mark_reg_unknown(env, reg); reg->type = NOT_INIT; } @@ -977,10 +1039,10 @@ static void mark_reg_not_init(struct bpf_verifier_env *env, verbose(env, "mark_reg_not_init(regs, %u)\n", regno); /* Something bad happened, let's kill all regs except FP */ for (regno = 0; regno < BPF_REG_FP; regno++) - __mark_reg_not_init(regs + regno); + __mark_reg_not_init(env, regs + regno); return; } - __mark_reg_not_init(regs + regno); + __mark_reg_not_init(env, regs + regno); } #define DEF_NOT_SUBREG (0) @@ -1346,6 +1408,441 @@ static int check_reg_arg(struct bpf_verifier_env *env, u32 regno, return 0; } +/* for any branch, call, exit record the history of jmps in the given state */ +static int push_jmp_history(struct bpf_verifier_env *env, + struct bpf_verifier_state *cur) +{ + u32 cnt = cur->jmp_history_cnt; + struct bpf_idx_pair *p; + + cnt++; + p = krealloc(cur->jmp_history, cnt * sizeof(*p), GFP_USER); + if (!p) + return -ENOMEM; + p[cnt - 1].idx = env->insn_idx; + p[cnt - 1].prev_idx = env->prev_insn_idx; + cur->jmp_history = p; + cur->jmp_history_cnt = cnt; + return 0; +} + +/* Backtrack one insn at a time. If idx is not at the top of recorded + * history then previous instruction came from straight line execution. + */ +static int get_prev_insn_idx(struct bpf_verifier_state *st, int i, + u32 *history) +{ + u32 cnt = *history; + + if (cnt && st->jmp_history[cnt - 1].idx == i) { + i = st->jmp_history[cnt - 1].prev_idx; + (*history)--; + } else { + i--; + } + return i; +} + +/* For given verifier state backtrack_insn() is called from the last insn to + * the first insn. Its purpose is to compute a bitmask of registers and + * stack slots that needs precision in the parent verifier state. + */ +static int backtrack_insn(struct bpf_verifier_env *env, int idx, + u32 *reg_mask, u64 *stack_mask) +{ + const struct bpf_insn_cbs cbs = { + .cb_print = verbose, + .private_data = env, + }; + struct bpf_insn *insn = env->prog->insnsi + idx; + u8 class = BPF_CLASS(insn->code); + u8 opcode = BPF_OP(insn->code); + u8 mode = BPF_MODE(insn->code); + u32 dreg = 1u << insn->dst_reg; + u32 sreg = 1u << insn->src_reg; + u32 spi; + + if (insn->code == 0) + return 0; + if (env->log.level & BPF_LOG_LEVEL) { + verbose(env, "regs=%x stack=%llx before ", *reg_mask, *stack_mask); + verbose(env, "%d: ", idx); + print_bpf_insn(&cbs, insn, env->allow_ptr_leaks); + } + + if (class == BPF_ALU || class == BPF_ALU64) { + if (!(*reg_mask & dreg)) + return 0; + if (opcode == BPF_MOV) { + if (BPF_SRC(insn->code) == BPF_X) { + /* dreg = sreg + * dreg needs precision after this insn + * sreg needs precision before this insn + */ + *reg_mask &= ~dreg; + *reg_mask |= sreg; + } else { + /* dreg = K + * dreg needs precision after this insn. + * Corresponding register is already marked + * as precise=true in this verifier state. + * No further markings in parent are necessary + */ + *reg_mask &= ~dreg; + } + } else { + if (BPF_SRC(insn->code) == BPF_X) { + /* dreg += sreg + * both dreg and sreg need precision + * before this insn + */ + *reg_mask |= sreg; + } /* else dreg += K + * dreg still needs precision before this insn + */ + } + } else if (class == BPF_LDX) { + if (!(*reg_mask & dreg)) + return 0; + *reg_mask &= ~dreg; + + /* scalars can only be spilled into stack w/o losing precision. + * Load from any other memory can be zero extended. + * The desire to keep that precision is already indicated + * by 'precise' mark in corresponding register of this state. + * No further tracking necessary. + */ + if (insn->src_reg != BPF_REG_FP) + return 0; + if (BPF_SIZE(insn->code) != BPF_DW) + return 0; + + /* dreg = *(u64 *)[fp - off] was a fill from the stack. + * that [fp - off] slot contains scalar that needs to be + * tracked with precision + */ + spi = (-insn->off - 1) / BPF_REG_SIZE; + if (spi >= 64) { + verbose(env, "BUG spi %d\n", spi); + WARN_ONCE(1, "verifier backtracking bug"); + return -EFAULT; + } + *stack_mask |= 1ull << spi; + } else if (class == BPF_STX || class == BPF_ST) { + if (*reg_mask & dreg) + /* stx & st shouldn't be using _scalar_ dst_reg + * to access memory. It means backtracking + * encountered a case of pointer subtraction. + */ + return -ENOTSUPP; + /* scalars can only be spilled into stack */ + if (insn->dst_reg != BPF_REG_FP) + return 0; + if (BPF_SIZE(insn->code) != BPF_DW) + return 0; + spi = (-insn->off - 1) / BPF_REG_SIZE; + if (spi >= 64) { + verbose(env, "BUG spi %d\n", spi); + WARN_ONCE(1, "verifier backtracking bug"); + return -EFAULT; + } + if (!(*stack_mask & (1ull << spi))) + return 0; + *stack_mask &= ~(1ull << spi); + if (class == BPF_STX) + *reg_mask |= sreg; + } else if (class == BPF_JMP || class == BPF_JMP32) { + if (opcode == BPF_CALL) { + if (insn->src_reg == BPF_PSEUDO_CALL) + return -ENOTSUPP; + /* regular helper call sets R0 */ + *reg_mask &= ~1; + if (*reg_mask & 0x3f) { + /* if backtracing was looking for registers R1-R5 + * they should have been found already. + */ + verbose(env, "BUG regs %x\n", *reg_mask); + WARN_ONCE(1, "verifier backtracking bug"); + return -EFAULT; + } + } else if (opcode == BPF_EXIT) { + return -ENOTSUPP; + } + } else if (class == BPF_LD) { + if (!(*reg_mask & dreg)) + return 0; + *reg_mask &= ~dreg; + /* It's ld_imm64 or ld_abs or ld_ind. + * For ld_imm64 no further tracking of precision + * into parent is necessary + */ + if (mode == BPF_IND || mode == BPF_ABS) + /* to be analyzed */ + return -ENOTSUPP; + } + return 0; +} + +/* the scalar precision tracking algorithm: + * . at the start all registers have precise=false. + * . scalar ranges are tracked as normal through alu and jmp insns. + * . once precise value of the scalar register is used in: + * . ptr + scalar alu + * . if (scalar cond K|scalar) + * . helper_call(.., scalar, ...) where ARG_CONST is expected + * backtrack through the verifier states and mark all registers and + * stack slots with spilled constants that these scalar regisers + * should be precise. + * . during state pruning two registers (or spilled stack slots) + * are equivalent if both are not precise. + * + * Note the verifier cannot simply walk register parentage chain, + * since many different registers and stack slots could have been + * used to compute single precise scalar. + * + * The approach of starting with precise=true for all registers and then + * backtrack to mark a register as not precise when the verifier detects + * that program doesn't care about specific value (e.g., when helper + * takes register as ARG_ANYTHING parameter) is not safe. + * + * It's ok to walk single parentage chain of the verifier states. + * It's possible that this backtracking will go all the way till 1st insn. + * All other branches will be explored for needing precision later. + * + * The backtracking needs to deal with cases like: + * R8=map_value(id=0,off=0,ks=4,vs=1952,imm=0) R9_w=map_value(id=0,off=40,ks=4,vs=1952,imm=0) + * r9 -= r8 + * r5 = r9 + * if r5 > 0x79f goto pc+7 + * R5_w=inv(id=0,umax_value=1951,var_off=(0x0; 0x7ff)) + * r5 += 1 + * ... + * call bpf_perf_event_output#25 + * where .arg5_type = ARG_CONST_SIZE_OR_ZERO + * + * and this case: + * r6 = 1 + * call foo // uses callee's r6 inside to compute r0 + * r0 += r6 + * if r0 == 0 goto + * + * to track above reg_mask/stack_mask needs to be independent for each frame. + * + * Also if parent's curframe > frame where backtracking started, + * the verifier need to mark registers in both frames, otherwise callees + * may incorrectly prune callers. This is similar to + * commit 7640ead93924 ("bpf: verifier: make sure callees don't prune with caller differences") + * + * For now backtracking falls back into conservative marking. + */ +static void mark_all_scalars_precise(struct bpf_verifier_env *env, + struct bpf_verifier_state *st) +{ + struct bpf_func_state *func; + struct bpf_reg_state *reg; + int i, j; + + /* big hammer: mark all scalars precise in this path. + * pop_stack may still get !precise scalars. + */ + for (; st; st = st->parent) + for (i = 0; i <= st->curframe; i++) { + func = st->frame[i]; + for (j = 0; j < BPF_REG_FP; j++) { + reg = &func->regs[j]; + if (reg->type != SCALAR_VALUE) + continue; + reg->precise = true; + } + for (j = 0; j < func->allocated_stack / BPF_REG_SIZE; j++) { + if (func->stack[j].slot_type[0] != STACK_SPILL) + continue; + reg = &func->stack[j].spilled_ptr; + if (reg->type != SCALAR_VALUE) + continue; + reg->precise = true; + } + } +} + +static int __mark_chain_precision(struct bpf_verifier_env *env, int frame, int regno, + int spi) +{ + struct bpf_verifier_state *st = env->cur_state; + int first_idx = st->first_insn_idx; + int last_idx = env->insn_idx; + struct bpf_func_state *func; + struct bpf_reg_state *reg; + u32 reg_mask = regno >= 0 ? 1u << regno : 0; + u64 stack_mask = spi >= 0 ? 1ull << spi : 0; + bool skip_first = true; + bool new_marks = false; + int i, err; + + if (!env->allow_ptr_leaks) + /* backtracking is root only for now */ + return 0; + + func = st->frame[frame]; + if (regno >= 0) { + reg = &func->regs[regno]; + if (reg->type != SCALAR_VALUE) { + WARN_ONCE(1, "backtracing misuse"); + return -EFAULT; + } + if (!reg->precise) + new_marks = true; + else + reg_mask = 0; + reg->precise = true; + } + + while (spi >= 0) { + if (func->stack[spi].slot_type[0] != STACK_SPILL) { + stack_mask = 0; + break; + } + reg = &func->stack[spi].spilled_ptr; + if (reg->type != SCALAR_VALUE) { + stack_mask = 0; + break; + } + if (!reg->precise) + new_marks = true; + else + stack_mask = 0; + reg->precise = true; + break; + } + + if (!new_marks) + return 0; + if (!reg_mask && !stack_mask) + return 0; + for (;;) { + DECLARE_BITMAP(mask, 64); + u32 history = st->jmp_history_cnt; + + if (env->log.level & BPF_LOG_LEVEL) + verbose(env, "last_idx %d first_idx %d\n", last_idx, first_idx); + for (i = last_idx;;) { + if (skip_first) { + err = 0; + skip_first = false; + } else { + err = backtrack_insn(env, i, ®_mask, &stack_mask); + } + if (err == -ENOTSUPP) { + mark_all_scalars_precise(env, st); + return 0; + } else if (err) { + return err; + } + if (!reg_mask && !stack_mask) + /* Found assignment(s) into tracked register in this state. + * Since this state is already marked, just return. + * Nothing to be tracked further in the parent state. + */ + return 0; + if (i == first_idx) + break; + i = get_prev_insn_idx(st, i, &history); + if (i >= env->prog->len) { + /* This can happen if backtracking reached insn 0 + * and there are still reg_mask or stack_mask + * to backtrack. + * It means the backtracking missed the spot where + * particular register was initialized with a constant. + */ + verbose(env, "BUG backtracking idx %d\n", i); + WARN_ONCE(1, "verifier backtracking bug"); + return -EFAULT; + } + } + st = st->parent; + if (!st) + break; + + new_marks = false; + func = st->frame[frame]; + bitmap_from_u64(mask, reg_mask); + for_each_set_bit(i, mask, 32) { + reg = &func->regs[i]; + if (reg->type != SCALAR_VALUE) { + reg_mask &= ~(1u << i); + continue; + } + if (!reg->precise) + new_marks = true; + reg->precise = true; + } + + bitmap_from_u64(mask, stack_mask); + for_each_set_bit(i, mask, 64) { + if (i >= func->allocated_stack / BPF_REG_SIZE) { + /* the sequence of instructions: + * 2: (bf) r3 = r10 + * 3: (7b) *(u64 *)(r3 -8) = r0 + * 4: (79) r4 = *(u64 *)(r10 -8) + * doesn't contain jmps. It's backtracked + * as a single block. + * During backtracking insn 3 is not recognized as + * stack access, so at the end of backtracking + * stack slot fp-8 is still marked in stack_mask. + * However the parent state may not have accessed + * fp-8 and it's "unallocated" stack space. + * In such case fallback to conservative. + */ + mark_all_scalars_precise(env, st); + return 0; + } + + if (func->stack[i].slot_type[0] != STACK_SPILL) { + stack_mask &= ~(1ull << i); + continue; + } + reg = &func->stack[i].spilled_ptr; + if (reg->type != SCALAR_VALUE) { + stack_mask &= ~(1ull << i); + continue; + } + if (!reg->precise) + new_marks = true; + reg->precise = true; + } + if (env->log.level & BPF_LOG_LEVEL) { + print_verifier_state(env, func); + verbose(env, "parent %s regs=%x stack=%llx marks\n", + new_marks ? "didn't have" : "already had", + reg_mask, stack_mask); + } + + if (!reg_mask && !stack_mask) + break; + if (!new_marks) + break; + + last_idx = st->last_insn_idx; + first_idx = st->first_insn_idx; + } + return 0; +} + +int mark_chain_precision(struct bpf_verifier_env *env, int regno) +{ + return __mark_chain_precision(env, env->cur_state->curframe, regno, -1); +} + +static int mark_chain_precision_frame(struct bpf_verifier_env *env, int frame, int regno) +{ + return __mark_chain_precision(env, frame, regno, -1); +} + +static int mark_chain_precision_stack_frame(struct bpf_verifier_env *env, int frame, int spi) +{ + return __mark_chain_precision(env, frame, -1, spi); +} + static bool is_spillable_regtype(enum bpf_reg_type type) { switch (type) { @@ -1377,6 +1874,23 @@ static bool register_is_null(struct bpf_reg_state *reg) return reg->type == SCALAR_VALUE && tnum_equals_const(reg->var_off, 0); } +static bool register_is_const(struct bpf_reg_state *reg) +{ + return reg->type == SCALAR_VALUE && tnum_is_const(reg->var_off); +} + +static void save_register_state(struct bpf_func_state *state, + int spi, struct bpf_reg_state *reg) +{ + int i; + + state->stack[spi].spilled_ptr = *reg; + state->stack[spi].spilled_ptr.live |= REG_LIVE_WRITTEN; + + for (i = 0; i < BPF_REG_SIZE; i++) + state->stack[spi].slot_type[i] = STACK_SPILL; +} + /* check_stack_read/write functions track spill/fill of registers, * stack boundary and alignment are checked in check_mem_access() */ @@ -1386,7 +1900,8 @@ static int check_stack_write(struct bpf_verifier_env *env, { struct bpf_func_state *cur; /* state of the current function */ int i, slot = -off - 1, spi = slot / BPF_REG_SIZE, err; - enum bpf_reg_type type; + u32 dst_reg = env->prog->insnsi[insn_idx].dst_reg; + struct bpf_reg_state *reg = NULL; err = realloc_func_state(state, round_up(slot + 1, BPF_REG_SIZE), state->acquired_refs, true); @@ -1403,27 +1918,48 @@ static int check_stack_write(struct bpf_verifier_env *env, } cur = env->cur_state->frame[env->cur_state->curframe]; - if (value_regno >= 0 && - is_spillable_regtype((type = cur->regs[value_regno].type))) { - + if (value_regno >= 0) + reg = &cur->regs[value_regno]; + + if (reg && size == BPF_REG_SIZE && register_is_const(reg) && + !register_is_null(reg) && env->allow_ptr_leaks) { + if (dst_reg != BPF_REG_FP) { + /* The backtracking logic can only recognize explicit + * stack slot address like [fp - 8]. Other spill of + * scalar via different register has to be conervative. + * Backtrack from here and mark all registers as precise + * that contributed into 'reg' being a constant. + */ + err = mark_chain_precision(env, value_regno); + if (err) + return err; + } + save_register_state(state, spi, reg); + } else if (reg && is_spillable_regtype(reg->type)) { /* register containing pointer is being spilled into stack */ if (size != BPF_REG_SIZE) { + verbose_linfo(env, insn_idx, "; "); verbose(env, "invalid size of register spill\n"); return -EACCES; } - if (state != cur && type == PTR_TO_STACK) { + if (state != cur && reg->type == PTR_TO_STACK) { verbose(env, "cannot spill pointers to stack into stack frame of the caller\n"); return -EINVAL; } - /* save register state */ - state->stack[spi].spilled_ptr = cur->regs[value_regno]; - state->stack[spi].spilled_ptr.live |= REG_LIVE_WRITTEN; - - for (i = 0; i < BPF_REG_SIZE; i++) { - if (state->stack[spi].slot_type[i] == STACK_MISC && - !env->allow_ptr_leaks) { + if (!env->allow_ptr_leaks) { + bool sanitize = false; + + if (state->stack[spi].slot_type[0] == STACK_SPILL && + register_is_const(&state->stack[spi].spilled_ptr)) + sanitize = true; + for (i = 0; i < BPF_REG_SIZE; i++) + if (state->stack[spi].slot_type[i] == STACK_MISC) { + sanitize = true; + break; + } + if (sanitize) { int *poff = &env->insn_aux_data[insn_idx].sanitize_stack_off; int soff = (-spi - 1) * BPF_REG_SIZE; @@ -1446,8 +1982,8 @@ static int check_stack_write(struct bpf_verifier_env *env, } *poff = soff; } - state->stack[spi].slot_type[i] = STACK_SPILL; } + save_register_state(state, spi, reg); } else { u8 type = STACK_MISC; @@ -1466,9 +2002,13 @@ static int check_stack_write(struct bpf_verifier_env *env, state->stack[spi].spilled_ptr.live |= REG_LIVE_WRITTEN; /* when we zero initialize stack slots mark them as such */ - if (value_regno >= 0 && - register_is_null(&cur->regs[value_regno])) + if (reg && register_is_null(reg)) { + /* backtracking doesn't work for STACK_ZERO yet. */ + err = mark_chain_precision(env, value_regno); + if (err) + return err; type = STACK_ZERO; + } for (i = 0; i < size; i++) state->stack[spi].slot_type[(slot - i) % BPF_REG_SIZE] = @@ -1484,6 +2024,7 @@ static int check_stack_read(struct bpf_verifier_env *env, struct bpf_verifier_state *vstate = env->cur_state; struct bpf_func_state *state = vstate->frame[vstate->curframe]; int i, slot = -off - 1, spi = slot / BPF_REG_SIZE; + struct bpf_reg_state *reg; u8 *stype; if (reg_state->allocated_stack <= slot) { @@ -1492,11 +2033,21 @@ static int check_stack_read(struct bpf_verifier_env *env, return -EACCES; } stype = reg_state->stack[spi].slot_type; + reg = ®_state->stack[spi].spilled_ptr; if (stype[0] == STACK_SPILL) { if (size != BPF_REG_SIZE) { - verbose(env, "invalid size of register spill\n"); - return -EACCES; + if (reg->type != SCALAR_VALUE) { + verbose_linfo(env, env->insn_idx, "; "); + verbose(env, "invalid size of register fill\n"); + return -EACCES; + } + if (value_regno >= 0) { + mark_reg_unknown(env, state->regs, value_regno); + state->regs[value_regno].live |= REG_LIVE_WRITTEN; + } + mark_reg_read(env, reg, reg->parent, REG_LIVE_READ64); + return 0; } for (i = 1; i < BPF_REG_SIZE; i++) { if (stype[(slot - i) % BPF_REG_SIZE] != STACK_SPILL) { @@ -1507,17 +2058,14 @@ static int check_stack_read(struct bpf_verifier_env *env, if (value_regno >= 0) { /* restore register state from stack */ - state->regs[value_regno] = reg_state->stack[spi].spilled_ptr; + state->regs[value_regno] = *reg; /* mark reg as written since spilled pointer state likely * has its liveness marks cleared by is_state_visited() * which resets stack/reg liveness for state transitions */ state->regs[value_regno].live |= REG_LIVE_WRITTEN; } - mark_reg_read(env, ®_state->stack[spi].spilled_ptr, - reg_state->stack[spi].spilled_ptr.parent, - REG_LIVE_READ64); - return 0; + mark_reg_read(env, reg, reg->parent, REG_LIVE_READ64); } else { int zeros = 0; @@ -1532,23 +2080,32 @@ static int check_stack_read(struct bpf_verifier_env *env, off, i, size); return -EACCES; } - mark_reg_read(env, ®_state->stack[spi].spilled_ptr, - reg_state->stack[spi].spilled_ptr.parent, - REG_LIVE_READ64); + mark_reg_read(env, reg, reg->parent, REG_LIVE_READ64); if (value_regno >= 0) { if (zeros == size) { /* any size read into register is zero extended, * so the whole register == const_zero */ __mark_reg_const_zero(&state->regs[value_regno]); + /* backtracking doesn't support STACK_ZERO yet, + * so mark it precise here, so that later + * backtracking can stop here. + * Backtracking may not need this if this register + * doesn't participate in pointer adjustment. + * Forward propagation of precise flag is not + * necessary either. This mark is only to stop + * backtracking. Any register that contributed + * to const 0 was marked precise before spill. + */ + state->regs[value_regno].precise = true; } else { /* have read misc data from the stack */ mark_reg_unknown(env, state->regs, value_regno); } state->regs[value_regno].live |= REG_LIVE_WRITTEN; } - return 0; } + return 0; } static int check_stack_access(struct bpf_verifier_env *env, @@ -2409,7 +2966,7 @@ static int check_stack_boundary(struct bpf_verifier_env *env, int regno, { struct bpf_reg_state *reg = reg_state(env, regno); struct bpf_func_state *state = func(env, reg); - int err, min_off, max_off, i, slot, spi; + int err, min_off, max_off, i, j, slot, spi; if (reg->type != PTR_TO_STACK) { /* Allow zero-byte read from NULL, regardless of pointer type */ @@ -2497,6 +3054,14 @@ static int check_stack_boundary(struct bpf_verifier_env *env, int regno, *stype = STACK_MISC; goto mark; } + if (state->stack[spi].slot_type[0] == STACK_SPILL && + state->stack[spi].spilled_ptr.type == SCALAR_VALUE) { + __mark_reg_unknown(env, &state->stack[spi].spilled_ptr); + for (j = 0; j < BPF_REG_SIZE; j++) + state->stack[spi].slot_type[j] = STACK_MISC; + goto mark; + } + err: if (tnum_is_const(reg->var_off)) { verbose(env, "invalid indirect read from stack off %d+%d size %d\n", @@ -2847,6 +3412,8 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 regno, err = check_helper_mem_access(env, regno - 1, reg->umax_value, zero_size_allowed, meta); + if (!err) + err = mark_chain_precision(env, regno); } else if (arg_type_is_int_ptr(arg_type)) { int size = int_ptr_type_to_size(arg_type); @@ -3129,7 +3696,7 @@ static void __clear_all_pkt_pointers(struct bpf_verifier_env *env, if (!reg) continue; if (reg_is_pkt_pointer_any(reg)) - __mark_reg_unknown(reg); + __mark_reg_unknown(env, reg); } } @@ -3157,7 +3724,7 @@ static void release_reg_references(struct bpf_verifier_env *env, if (!reg) continue; if (reg->ref_obj_id == ref_obj_id) - __mark_reg_unknown(reg); + __mark_reg_unknown(env, reg); } } @@ -3970,7 +4537,7 @@ static int adjust_ptr_min_max_vals(struct bpf_verifier_env *env, /* Taint dst register if offset had invalid bounds derived from * e.g. dead branches. */ - __mark_reg_unknown(dst_reg); + __mark_reg_unknown(env, dst_reg); return 0; } @@ -4202,13 +4769,13 @@ static int adjust_scalar_min_max_vals(struct bpf_verifier_env *env, /* Taint dst register if offset had invalid bounds derived from * e.g. dead branches. */ - __mark_reg_unknown(dst_reg); + __mark_reg_unknown(env, dst_reg); return 0; } if (!src_known && opcode != BPF_ADD && opcode != BPF_SUB && opcode != BPF_AND) { - __mark_reg_unknown(dst_reg); + __mark_reg_unknown(env, dst_reg); return 0; } @@ -4450,6 +5017,7 @@ static int adjust_reg_min_max_vals(struct bpf_verifier_env *env, struct bpf_reg_state *regs = state->regs, *dst_reg, *src_reg; struct bpf_reg_state *ptr_reg = NULL, off_reg = {0}; u8 opcode = BPF_OP(insn->code); + int err; dst_reg = ®s[insn->dst_reg]; src_reg = NULL; @@ -4476,13 +5044,24 @@ static int adjust_reg_min_max_vals(struct bpf_verifier_env *env, * This is legal, but we have to reverse our * src/dest handling in computing the range */ + err = mark_chain_precision(env, insn->dst_reg); + if (err) + return err; return adjust_ptr_min_max_vals(env, insn, src_reg, dst_reg); } } else if (ptr_reg) { /* pointer += scalar */ + err = mark_chain_precision(env, insn->src_reg); + if (err) + return err; return adjust_ptr_min_max_vals(env, insn, dst_reg, src_reg); + } else if (dst_reg->precise) { + /* if dst_reg is precise, src_reg should be precise as well */ + err = mark_chain_precision(env, insn->src_reg); + if (err) + return err; } } else { /* Pretend the src is a reg with a known value, since we only @@ -5437,6 +6016,13 @@ static int check_cond_jmp_op(struct bpf_verifier_env *env, pred = is_branch_taken(dst_reg, src_reg->var_off.value, opcode, is_jmp32); + if (pred >= 0) { + err = mark_chain_precision(env, insn->dst_reg); + if (BPF_SRC(insn->code) == BPF_X && !err) + err = mark_chain_precision(env, insn->src_reg); + if (err) + return err; + } if (pred == 1) { /* Only follow the goto, ignore fall-through. If needed, push * the fall-through branch for simulation under speculative @@ -5818,7 +6404,8 @@ static void init_explored_state(struct bpf_verifier_env *env, int idx) * w - next instruction * e - edge */ -static int push_insn(int t, int w, int e, struct bpf_verifier_env *env) +static int push_insn(int t, int w, int e, struct bpf_verifier_env *env, + bool loop_ok) { int *insn_stack = env->cfg.insn_stack; int *insn_state = env->cfg.insn_state; @@ -5848,6 +6435,8 @@ static int push_insn(int t, int w, int e, struct bpf_verifier_env *env) insn_stack[env->cfg.cur_stack++] = w; return 1; } else if ((insn_state[w] & 0xF0) == DISCOVERED) { + if (loop_ok && env->allow_ptr_leaks) + return 0; verbose_linfo(env, t, "%d: ", t); verbose_linfo(env, w, "%d: ", w); verbose(env, "back-edge from insn %d to %d\n", t, w); @@ -5899,7 +6488,7 @@ static int check_cfg(struct bpf_verifier_env *env) if (opcode == BPF_EXIT) { goto mark_explored; } else if (opcode == BPF_CALL) { - ret = push_insn(t, t + 1, FALLTHROUGH, env); + ret = push_insn(t, t + 1, FALLTHROUGH, env, false); if (ret == 1) goto peek_stack; else if (ret < 0) @@ -5908,7 +6497,8 @@ static int check_cfg(struct bpf_verifier_env *env) init_explored_state(env, t + 1); if (insns[t].src_reg == BPF_PSEUDO_CALL) { init_explored_state(env, t); - ret = push_insn(t, t + insns[t].imm + 1, BRANCH, env); + ret = push_insn(t, t + insns[t].imm + 1, BRANCH, + env, false); if (ret == 1) goto peek_stack; else if (ret < 0) @@ -5921,11 +6511,16 @@ static int check_cfg(struct bpf_verifier_env *env) } /* unconditional jump with single edge */ ret = push_insn(t, t + insns[t].off + 1, - FALLTHROUGH, env); + FALLTHROUGH, env, true); if (ret == 1) goto peek_stack; else if (ret < 0) goto err_free; + /* unconditional jmp is not a good pruning point, + * but it's marked, since backtracking needs + * to record jmp history in is_state_visited(). + */ + init_explored_state(env, t + insns[t].off + 1); /* tell verifier to check for equivalent states * after every call and jump */ @@ -5934,13 +6529,13 @@ static int check_cfg(struct bpf_verifier_env *env) } else { /* conditional jump with two edges */ init_explored_state(env, t); - ret = push_insn(t, t + 1, FALLTHROUGH, env); + ret = push_insn(t, t + 1, FALLTHROUGH, env, true); if (ret == 1) goto peek_stack; else if (ret < 0) goto err_free; - ret = push_insn(t, t + insns[t].off + 1, BRANCH, env); + ret = push_insn(t, t + insns[t].off + 1, BRANCH, env, true); if (ret == 1) goto peek_stack; else if (ret < 0) @@ -5950,7 +6545,7 @@ static int check_cfg(struct bpf_verifier_env *env) /* all other non-branch instructions with single * fall-through edge */ - ret = push_insn(t, t + 1, FALLTHROUGH, env); + ret = push_insn(t, t + 1, FALLTHROUGH, env, false); if (ret == 1) goto peek_stack; else if (ret < 0) @@ -6315,7 +6910,7 @@ static void clean_func_state(struct bpf_verifier_env *env, /* since the register is unused, clear its state * to make further comparison simpler */ - __mark_reg_not_init(&st->regs[i]); + __mark_reg_not_init(env, &st->regs[i]); } for (i = 0; i < st->allocated_stack / BPF_REG_SIZE; i++) { @@ -6323,7 +6918,7 @@ static void clean_func_state(struct bpf_verifier_env *env, /* liveness must not touch this stack slot anymore */ st->stack[i].spilled_ptr.live |= REG_LIVE_DONE; if (!(live & REG_LIVE_READ)) { - __mark_reg_not_init(&st->stack[i].spilled_ptr); + __mark_reg_not_init(env, &st->stack[i].spilled_ptr); for (j = 0; j < BPF_REG_SIZE; j++) st->stack[i].slot_type[j] = STACK_INVALID; } @@ -6383,6 +6978,8 @@ static void clean_live_states(struct bpf_verifier_env *env, int insn, sl = *explored_state(env, insn); while (sl) { + if (sl->state.branches) + goto next; if (sl->state.insn_idx != insn || sl->state.curframe != cur->curframe) goto next; @@ -6424,6 +7021,8 @@ static bool regsafe(struct bpf_reg_state *rold, struct bpf_reg_state *rcur, switch (rold->type) { case SCALAR_VALUE: if (rcur->type == SCALAR_VALUE) { + if (!rold->precise && !rcur->precise) + return true; /* new val must satisfy old val knowledge */ return range_within(rold, rcur) && tnum_in(rold->var_off, rcur->var_off); @@ -6747,19 +7346,95 @@ static int propagate_liveness(struct bpf_verifier_env *env, return 0; } +/* find precise scalars in the previous equivalent state and + * propagate them into the current state + */ +static int propagate_precision(struct bpf_verifier_env *env, + const struct bpf_verifier_state *old) +{ + struct bpf_reg_state *state_reg; + struct bpf_func_state *state; + int i, err = 0, fr; + + for (fr = old->curframe; fr >= 0; fr--) { + state = old->frame[fr]; + state_reg = state->regs; + for (i = 0; i < BPF_REG_FP; i++, state_reg++) { + if (state_reg->type != SCALAR_VALUE || + !state_reg->precise) + continue; + if (env->log.level & BPF_LOG_LEVEL2) + verbose(env, "frame %d: propagating r%d\n", i, fr); + err = mark_chain_precision_frame(env, fr, i); + if (err < 0) + return err; + } + + for (i = 0; i < state->allocated_stack / BPF_REG_SIZE; i++) { + if (state->stack[i].slot_type[0] != STACK_SPILL) + continue; + state_reg = &state->stack[i].spilled_ptr; + if (state_reg->type != SCALAR_VALUE || + !state_reg->precise) + continue; + if (env->log.level & BPF_LOG_LEVEL2) + verbose(env, "frame %d: propagating fp%d\n", + (-i - 1) * BPF_REG_SIZE, fr); + err = mark_chain_precision_stack_frame(env, fr, i); + if (err < 0) + return err; + } + + } + return 0; +} + +static bool states_maybe_looping(struct bpf_verifier_state *old, + struct bpf_verifier_state *cur) +{ + struct bpf_func_state *fold, *fcur; + int i, fr = cur->curframe; + + if (old->curframe != fr) + return false; + + fold = old->frame[fr]; + fcur = cur->frame[fr]; + for (i = 0; i < MAX_BPF_REG; i++) + if (memcmp(&fold->regs[i], &fcur->regs[i], + offsetof(struct bpf_reg_state, parent))) + return false; + return true; +} + + static int is_state_visited(struct bpf_verifier_env *env, int insn_idx) { struct bpf_verifier_state_list *new_sl; struct bpf_verifier_state_list *sl, **pprev; struct bpf_verifier_state *cur = env->cur_state, *new; int i, j, err, states_cnt = 0; + bool add_new_state = false; + cur->last_insn_idx = env->prev_insn_idx; if (!env->insn_aux_data[insn_idx].prune_point) /* this 'insn_idx' instruction wasn't marked, so we will not * be doing state search here */ return 0; + /* bpf progs typically have pruning point every 4 instructions + * http://vger.kernel.org/bpfconf2019.html#session-1 + * Do not add new state for future pruning if the verifier hasn't seen + * at least 2 jumps and at least 8 instructions. + * This heuristics helps decrease 'total_states' and 'peak_states' metric. + * In tests that amounts to up to 50% reduction into total verifier + * memory consumption and 20% verifier time speedup. + */ + if (env->jmps_processed - env->prev_jmps_processed >= 2 && + env->insn_processed - env->prev_insn_processed >= 8) + add_new_state = true; + pprev = explored_state(env, insn_idx); sl = *pprev; @@ -6769,6 +7444,30 @@ static int is_state_visited(struct bpf_verifier_env *env, int insn_idx) states_cnt++; if (sl->state.insn_idx != insn_idx) goto next; + if (sl->state.branches) { + if (states_maybe_looping(&sl->state, cur) && + states_equal(env, &sl->state, cur)) { + verbose_linfo(env, insn_idx, "; "); + verbose(env, "infinite loop detected at insn %d\n", insn_idx); + return -EINVAL; + } + /* if the verifier is processing a loop, avoid adding new state + * too often, since different loop iterations have distinct + * states and may not help future pruning. + * This threshold shouldn't be too low to make sure that + * a loop with large bound will be rejected quickly. + * The most abusive loop will be: + * r1 += 1 + * if r1 < 1000000 goto pc-2 + * 1M insn_procssed limit / 100 == 10k peak states. + * This threshold shouldn't be too high either, since states + * at the end of the loop are likely to be useful in pruning. + */ + if (env->jmps_processed - env->prev_jmps_processed < 20 && + env->insn_processed - env->prev_insn_processed < 100) + add_new_state = false; + goto miss; + } if (states_equal(env, &sl->state, cur)) { sl->hit_cnt++; /* reached equivalent register/stack state, @@ -6782,11 +7481,27 @@ static int is_state_visited(struct bpf_verifier_env *env, int insn_idx) * this state and will pop a new one. */ err = propagate_liveness(env, &sl->state, cur); + + /* if previous state reached the exit with precision and + * current state is equivalent to it (except precsion marks) + * the precision needs to be propagated back in + * the current state. + */ + err = err ? : push_jmp_history(env, cur); + err = err ? : propagate_precision(env, &sl->state); if (err) return err; return 1; } - sl->miss_cnt++; +miss: + /* when new state is not going to be added do not increase miss count. + * Otherwise several loop iterations will remove the state + * recorded earlier. The goal of these heuristics is to have + * states from some iterations of the loop (some in the beginning + * and some at the end) to help pruning. + */ + if (add_new_state) + sl->miss_cnt++; /* heuristic to determine whether this state is beneficial * to keep checking from state equivalence point of view. * Higher numbers increase max_states_per_insn and verification time, @@ -6798,6 +7513,11 @@ static int is_state_visited(struct bpf_verifier_env *env, int insn_idx) */ *pprev = sl->next; if (sl->state.frame[0]->regs[0].live & REG_LIVE_DONE) { + u32 br = sl->state.branches; + + WARN_ONCE(br, + "BUG live_done but branches_to_explore %d\n", + br); free_verifier_state(&sl->state, false); kfree(sl); env->peak_states--; @@ -6821,20 +7541,27 @@ static int is_state_visited(struct bpf_verifier_env *env, int insn_idx) env->max_states_per_insn = states_cnt; if (!env->allow_ptr_leaks && states_cnt > BPF_COMPLEXITY_LIMIT_STATES) - return 0; + return push_jmp_history(env, cur); - /* there were no equivalent states, remember current one. - * technically the current state is not proven to be safe yet, + if (!add_new_state) + return push_jmp_history(env, cur); + + /* There were no equivalent states, remember the current one. + * Technically the current state is not proven to be safe yet, * but it will either reach outer most bpf_exit (which means it's safe) - * or it will be rejected. Since there are no loops, we won't be + * or it will be rejected. When there are no loops the verifier won't be * seeing this tuple (frame[0].callsite, frame[1].callsite, .. insn_idx) - * again on the way to bpf_exit + * again on the way to bpf_exit. + * When looping the sl->state.branches will be > 0 and this state + * will not be considered for equivalence until branches == 0. */ new_sl = kzalloc(sizeof(struct bpf_verifier_state_list), GFP_KERNEL); if (!new_sl) return -ENOMEM; env->total_states++; env->peak_states++; + env->prev_jmps_processed = env->jmps_processed; + env->prev_insn_processed = env->insn_processed; /* add new state to the head of linked list */ new = &new_sl->state; @@ -6845,6 +7572,12 @@ static int is_state_visited(struct bpf_verifier_env *env, int insn_idx) return err; } new->insn_idx = insn_idx; + WARN_ONCE(new->branches != 1, + "BUG is_state_visited:branches_to_explore=%d insn %d\n", new->branches, insn_idx); + + cur->parent = new; + cur->first_insn_idx = insn_idx; + clear_jmp_history(cur); new_sl->next = *explored_state(env, insn_idx); *explored_state(env, insn_idx) = new_sl; /* connect new state to parentage chain. Current frame needs all @@ -6854,17 +7587,18 @@ static int is_state_visited(struct bpf_verifier_env *env, int insn_idx) * the state of the call instruction (with WRITTEN set), and r0 comes * from callee with its full parentage chain, anyway. */ - for (j = 0; j <= cur->curframe; j++) - for (i = j < cur->curframe ? BPF_REG_6 : 0; i < BPF_REG_FP; i++) - cur->frame[j]->regs[i].parent = &new->frame[j]->regs[i]; /* clear write marks in current state: the writes we did are not writes * our child did, so they don't screen off its reads from us. * (There are no read marks in current state, because reads always mark * their parent and current state never has children yet. Only * explored_states can get read marks.) */ - for (i = 0; i < BPF_REG_FP; i++) - cur->frame[cur->curframe]->regs[i].live = REG_LIVE_NONE; + for (j = 0; j <= cur->curframe; j++) { + for (i = j < cur->curframe ? BPF_REG_6 : 0; i < BPF_REG_FP; i++) + cur->frame[j]->regs[i].parent = &new->frame[j]->regs[i]; + for (i = 0; i < BPF_REG_FP; i++) + cur->frame[j]->regs[i].live = REG_LIVE_NONE; + } /* all stack frames are accessible from callee, clear them all */ for (j = 0; j <= cur->curframe; j++) { @@ -6923,6 +7657,7 @@ static int do_check(struct bpf_verifier_env *env) struct bpf_reg_state *regs; int insn_cnt = env->prog->len; bool do_print_state = false; + int prev_insn_idx = -1; env->prev_linfo = NULL; @@ -6931,6 +7666,7 @@ static int do_check(struct bpf_verifier_env *env) return -ENOMEM; state->curframe = 0; state->speculative = false; + state->branches = 1; state->frame[0] = kzalloc(sizeof(struct bpf_func_state), GFP_KERNEL); if (!state->frame[0]) { kfree(state); @@ -6947,6 +7683,7 @@ static int do_check(struct bpf_verifier_env *env) u8 class; int err; + env->prev_insn_idx = prev_insn_idx; if (env->insn_idx >= insn_cnt) { verbose(env, "invalid insn idx %d insn_cnt %d\n", env->insn_idx, insn_cnt); @@ -7019,6 +7756,7 @@ static int do_check(struct bpf_verifier_env *env) regs = cur_regs(env); sanitize_mark_insn_seen(env); + prev_insn_idx = env->insn_idx; if (class == BPF_ALU || class == BPF_ALU64) { err = check_alu_op(env, insn); @@ -7137,6 +7875,7 @@ static int do_check(struct bpf_verifier_env *env) } else if (class == BPF_JMP || class == BPF_JMP32) { u8 opcode = BPF_OP(insn->code); + env->jmps_processed++; if (opcode == BPF_CALL) { if (BPF_SRC(insn->code) != BPF_K || insn->off != 0 || @@ -7191,7 +7930,6 @@ static int do_check(struct bpf_verifier_env *env) if (state->curframe) { /* exit from nested function */ - env->prev_insn_idx = env->insn_idx; err = prepare_func_exit(env, &env->insn_idx); if (err) return err; @@ -7222,7 +7960,8 @@ static int do_check(struct bpf_verifier_env *env) if (err) return err; process_bpf_exit: - err = pop_stack(env, &env->prev_insn_idx, + update_branch_counts(env, env->cur_state); + err = pop_stack(env, &prev_insn_idx, &env->insn_idx); if (err < 0) { if (err != -ENOENT) diff --git a/net/core/filter.c b/net/core/filter.c index d1a84712ec0b4502f11ebd31e3ce4371f0006c7a..782a1cae2167950a27a553651146bc0812f51032 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -5896,6 +5896,10 @@ sock_addr_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog) case BPF_FUNC_skc_lookup_tcp: return &bpf_sock_addr_skc_lookup_tcp_proto; #endif /* CONFIG_INET */ + case BPF_FUNC_sk_storage_get: + return &bpf_sk_storage_get_proto; + case BPF_FUNC_sk_storage_delete: + return &bpf_sk_storage_delete_proto; default: return bpf_base_func_proto(func_id); } @@ -6113,6 +6117,14 @@ sock_ops_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog) return &bpf_get_local_storage_proto; case BPF_FUNC_perf_event_output: return &bpf_sockopt_event_output_proto; + case BPF_FUNC_sk_storage_get: + return &bpf_sk_storage_get_proto; + case BPF_FUNC_sk_storage_delete: + return &bpf_sk_storage_delete_proto; +#ifdef CONFIG_INET + case BPF_FUNC_tcp_sock: + return &bpf_tcp_sock_proto; +#endif /* CONFIG_INET */ default: return bpf_base_func_proto(func_id); } @@ -6801,6 +6813,13 @@ static bool sock_addr_is_valid_access(int off, int size, if (size != size_default) return false; break; + case offsetof(struct bpf_sock_addr, sk): + if (type != BPF_READ) + return false; + if (size != sizeof(__u64)) + return false; + info->reg_type = PTR_TO_SOCKET; + break; default: if (type == BPF_READ) { if (size != size_default) @@ -6844,6 +6863,11 @@ static bool sock_ops_is_valid_access(int off, int size, if (size != sizeof(__u64)) return false; break; + case offsetof(struct bpf_sock_ops, sk): + if (size != sizeof(__u64)) + return false; + info->reg_type = PTR_TO_SOCKET_OR_NULL; + break; default: if (size != size_default) return false; @@ -7735,6 +7759,11 @@ static u32 sock_addr_convert_ctx_access(enum bpf_access_type type, struct bpf_sock_addr_kern, struct in6_addr, t_ctx, s6_addr32[0], BPF_SIZE(si->code), off, tmp_reg); break; + case offsetof(struct bpf_sock_addr, sk): + *insn++ = BPF_LDX_MEM(BPF_FIELD_SIZEOF(struct bpf_sock_addr_kern, sk), + si->dst_reg, si->src_reg, + offsetof(struct bpf_sock_addr_kern, sk)); + break; } return insn - insn_buf; @@ -7793,6 +7822,43 @@ static u32 sock_ops_convert_ctx_access(enum bpf_access_type type, } \ } while (0) +#define SOCK_OPS_GET_SK() \ + do { \ + int fullsock_reg = si->dst_reg, reg = BPF_REG_9, jmp = 1; \ + if (si->dst_reg == reg || si->src_reg == reg) \ + reg--; \ + if (si->dst_reg == reg || si->src_reg == reg) \ + reg--; \ + if (si->dst_reg == si->src_reg) { \ + *insn++ = BPF_STX_MEM(BPF_DW, si->src_reg, reg, \ + offsetof(struct bpf_sock_ops_kern, \ + temp)); \ + fullsock_reg = reg; \ + jmp += 2; \ + } \ + *insn++ = BPF_LDX_MEM(BPF_FIELD_SIZEOF( \ + struct bpf_sock_ops_kern, \ + is_fullsock), \ + fullsock_reg, si->src_reg, \ + offsetof(struct bpf_sock_ops_kern, \ + is_fullsock)); \ + *insn++ = BPF_JMP_IMM(BPF_JEQ, fullsock_reg, 0, jmp); \ + if (si->dst_reg == si->src_reg) \ + *insn++ = BPF_LDX_MEM(BPF_DW, reg, si->src_reg, \ + offsetof(struct bpf_sock_ops_kern, \ + temp)); \ + *insn++ = BPF_LDX_MEM(BPF_FIELD_SIZEOF( \ + struct bpf_sock_ops_kern, sk),\ + si->dst_reg, si->src_reg, \ + offsetof(struct bpf_sock_ops_kern, sk));\ + if (si->dst_reg == si->src_reg) { \ + *insn++ = BPF_JMP_A(1); \ + *insn++ = BPF_LDX_MEM(BPF_DW, reg, si->src_reg, \ + offsetof(struct bpf_sock_ops_kern, \ + temp)); \ + } \ + } while (0) + #define SOCK_OPS_GET_TCP_SOCK_FIELD(FIELD) \ SOCK_OPS_GET_FIELD(FIELD, FIELD, struct tcp_sock) @@ -8016,6 +8082,9 @@ static u32 sock_ops_convert_ctx_access(enum bpf_access_type type, SOCK_OPS_GET_OR_SET_FIELD(sk_txhash, sk_txhash, struct sock, type); break; + case offsetof(struct bpf_sock_ops, sk): + SOCK_OPS_GET_SK(); + break; } return insn - insn_buf; } diff --git a/tools/bpf/bpftool/common.c b/tools/bpf/bpftool/common.c index cc08743dc6094f7770551fafecb7a322dee98b9b..6a71324be6283eccb0d816a72dd61f66b4527de6 100644 --- a/tools/bpf/bpftool/common.c +++ b/tools/bpf/bpftool/common.c @@ -21,6 +21,7 @@ #include #include +#include /* libbpf_num_possible_cpus */ #include "main.h" @@ -443,57 +444,13 @@ unsigned int get_page_size(void) unsigned int get_possible_cpus(void) { - static unsigned int result; - char buf[128]; - long int n; - char *ptr; - int fd; - - if (result) - return result; - - fd = open("/sys/devices/system/cpu/possible", O_RDONLY); - if (fd < 0) { - p_err("can't open sysfs possible cpus"); - exit(-1); - } - - n = read(fd, buf, sizeof(buf)); - if (n < 2) { - p_err("can't read sysfs possible cpus"); - exit(-1); - } - close(fd); + int cpus = libbpf_num_possible_cpus(); - if (n == sizeof(buf)) { - p_err("read sysfs possible cpus overflow"); + if (cpus < 0) { + p_err("Can't get # of possible cpus: %s", strerror(-cpus)); exit(-1); } - - ptr = buf; - n = 0; - while (*ptr && *ptr != '\n') { - unsigned int a, b; - - if (sscanf(ptr, "%u-%u", &a, &b) == 2) { - n += b - a + 1; - - ptr = strchr(ptr, '-') + 1; - } else if (sscanf(ptr, "%u", &a) == 1) { - n++; - } else { - assert(0); - } - - while (isdigit(*ptr)) - ptr++; - if (*ptr == ',') - ptr++; - } - - result = n; - - return result; + return cpus; } static char * diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h index c40b9ad94621cafe5d844e42b4a5cdee211b8b0f..5b9627d981cf624535580c34baf9d5bc2ab601b0 100644 --- a/tools/include/uapi/linux/bpf.h +++ b/tools/include/uapi/linux/bpf.h @@ -3228,6 +3228,7 @@ struct bpf_sock_addr { __u32 msg_src_ip6[4]; /* Allows 1,2,4-byte read an 4-byte write. * Stored in network byte order. */ + __bpf_md_ptr(struct bpf_sock *, sk); }; /* User bpf_sock_ops struct to access socket values and specify request ops @@ -3279,6 +3280,7 @@ struct bpf_sock_ops { __u32 sk_txhash; __u64 bytes_received; __u64 bytes_acked; + __bpf_md_ptr(struct bpf_sock *, sk); }; /* Definitions for bpf_sock_ops_cb_flags */ @@ -3394,8 +3396,8 @@ struct bpf_raw_tracepoint_args { /* DIRECT: Skip the FIB rules and go to FIB table associated with device * OUTPUT: Do lookup from egress perspective; default is ingress */ -#define BPF_FIB_LOOKUP_DIRECT BIT(0) -#define BPF_FIB_LOOKUP_OUTPUT BIT(1) +#define BPF_FIB_LOOKUP_DIRECT (1U << 0) +#define BPF_FIB_LOOKUP_OUTPUT (1U << 1) enum { BPF_FIB_LKUP_RET_SUCCESS, /* lookup successful */ diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c index 1589c4b24ff39c37aa87bf64d091b57eb1c5df08..4c63a5959fa0c2de4edc55c7e2ecbd8d5c511f1a 100644 --- a/tools/lib/bpf/libbpf.c +++ b/tools/lib/bpf/libbpf.c @@ -3872,3 +3872,64 @@ void bpf_program__bpil_offs_to_addr(struct bpf_prog_info_linear *info_linear) desc->array_offset, addr); } } + +int libbpf_num_possible_cpus(void) +{ + static const char *fcpu = "/sys/devices/system/cpu/possible"; + int len = 0, n = 0, il = 0, ir = 0; + unsigned int start = 0, end = 0; + int tmp_cpus = 0; + static int cpus; + char buf[128]; + int error = 0; + int fd = -1; + + tmp_cpus = READ_ONCE(cpus); + if (tmp_cpus > 0) + return tmp_cpus; + + fd = open(fcpu, O_RDONLY); + if (fd < 0) { + error = errno; + pr_warning("Failed to open file %s: %s\n", + fcpu, strerror(error)); + return -error; + } + len = read(fd, buf, sizeof(buf)); + close(fd); + if (len <= 0) { + error = len ? errno : EINVAL; + pr_warning("Failed to read # of possible cpus from %s: %s\n", + fcpu, strerror(error)); + return -error; + } + if (len == sizeof(buf)) { + pr_warning("File %s size overflow\n", fcpu); + return -EOVERFLOW; + } + buf[len] = '\0'; + + for (ir = 0, tmp_cpus = 0; ir <= len; ir++) { + /* Each sub string separated by ',' has format \d+-\d+ or \d+ */ + if (buf[ir] == ',' || buf[ir] == '\0') { + buf[ir] = '\0'; + n = sscanf(&buf[il], "%u-%u", &start, &end); + if (n <= 0) { + pr_warning("Failed to get # CPUs from %s\n", + &buf[il]); + return -EINVAL; + } else if (n == 1) { + end = start; + } + tmp_cpus += end - start + 1; + il = ir + 1; + } + } + if (tmp_cpus <= 0) { + pr_warning("Invalid #CPUs %d from %s\n", tmp_cpus, fcpu); + return -EINVAL; + } + + WRITE_ONCE(cpus, tmp_cpus); + return tmp_cpus; +} diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h index 1af0d48178c8a0a2129ca0790246fd7b32a7833b..2e594a0fa9617f82ba4c8d6b2324d7fb7516b1c6 100644 --- a/tools/lib/bpf/libbpf.h +++ b/tools/lib/bpf/libbpf.h @@ -454,6 +454,22 @@ bpf_program__bpil_addr_to_offs(struct bpf_prog_info_linear *info_linear); LIBBPF_API void bpf_program__bpil_offs_to_addr(struct bpf_prog_info_linear *info_linear); +/* + * A helper function to get the number of possible CPUs before looking up + * per-CPU maps. Negative errno is returned on failure. + * + * Example usage: + * + * int ncpus = libbpf_num_possible_cpus(); + * if (ncpus < 0) { + * // error handling + * } + * long values[ncpus]; + * bpf_map_lookup_elem(per_cpu_map_fd, key, values); + * + */ +LIBBPF_API int libbpf_num_possible_cpus(void); + #ifdef __cplusplus } /* extern "C" */ #endif diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map index 46dcda89df211289669a88d21522d7833f13fe78..2c6d835620d254ef39b16564487dd0477c0fce8d 100644 --- a/tools/lib/bpf/libbpf.map +++ b/tools/lib/bpf/libbpf.map @@ -172,4 +172,5 @@ LIBBPF_0.0.4 { btf_dump__new; btf__parse_elf; bpf_object__load_xattr; + libbpf_num_possible_cpus; } LIBBPF_0.0.3; diff --git a/tools/testing/selftests/bpf/bpf_endian.h b/tools/testing/selftests/bpf/bpf_endian.h index b25595ea4a78987bbdacb2cd06c8eeaed4fb0ef6..05f036df8a4c51b7c32fd72f65b9766e13ac7689 100644 --- a/tools/testing/selftests/bpf/bpf_endian.h +++ b/tools/testing/selftests/bpf/bpf_endian.h @@ -2,6 +2,7 @@ #ifndef __BPF_ENDIAN__ #define __BPF_ENDIAN__ +#include #include /* LLVM's BPF target selects the endianness of the CPU diff --git a/tools/testing/selftests/bpf/bpf_util.h b/tools/testing/selftests/bpf/bpf_util.h index a29206ebbd1303cd60e9300d57b8ab12514959be..ec219f84e0415715c0d572e209f39f706587d0b5 100644 --- a/tools/testing/selftests/bpf/bpf_util.h +++ b/tools/testing/selftests/bpf/bpf_util.h @@ -6,44 +6,17 @@ #include #include #include +#include /* libbpf_num_possible_cpus */ static inline unsigned int bpf_num_possible_cpus(void) { - static const char *fcpu = "/sys/devices/system/cpu/possible"; - unsigned int start, end, possible_cpus = 0; - char buff[128]; - FILE *fp; - int len, n, i, j = 0; + int possible_cpus = libbpf_num_possible_cpus(); - fp = fopen(fcpu, "r"); - if (!fp) { - printf("Failed to open %s: '%s'!\n", fcpu, strerror(errno)); + if (possible_cpus < 0) { + printf("Failed to get # of possible cpus: '%s'!\n", + strerror(-possible_cpus)); exit(1); } - - if (!fgets(buff, sizeof(buff), fp)) { - printf("Failed to read %s!\n", fcpu); - exit(1); - } - - len = strlen(buff); - for (i = 0; i <= len; i++) { - if (buff[i] == ',' || buff[i] == '\0') { - buff[i] = '\0'; - n = sscanf(&buff[j], "%u-%u", &start, &end); - if (n <= 0) { - printf("Failed to retrieve # possible CPUs!\n"); - exit(1); - } else if (n == 1) { - end = start; - } - possible_cpus += end - start + 1; - j = i + 1; - } - } - - fclose(fp); - return possible_cpus; } diff --git a/tools/testing/selftests/bpf/progs/loop1.c b/tools/testing/selftests/bpf/progs/loop1.c new file mode 100644 index 0000000000000000000000000000000000000000..dea395af9ea9f57bb61028d082b13deb1c2d968c --- /dev/null +++ b/tools/testing/selftests/bpf/progs/loop1.c @@ -0,0 +1,28 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook +#include +#include +#include +#include +#include +#include +#include "bpf_helpers.h" + +char _license[] SEC("license") = "GPL"; + +SEC("raw_tracepoint/kfree_skb") +int nested_loops(volatile struct pt_regs* ctx) +{ + int i, j, sum = 0, m; + + for (j = 0; j < 300; j++) + for (i = 0; i < j; i++) { + if (j & 1) + m = ctx->rax; + else + m = j; + sum += i * m; + } + + return sum; +} diff --git a/tools/testing/selftests/bpf/progs/loop2.c b/tools/testing/selftests/bpf/progs/loop2.c new file mode 100644 index 0000000000000000000000000000000000000000..0637bd8e8bcf5827144fca281b08d9dbc88bf76f --- /dev/null +++ b/tools/testing/selftests/bpf/progs/loop2.c @@ -0,0 +1,28 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook +#include +#include +#include +#include +#include +#include +#include "bpf_helpers.h" + +char _license[] SEC("license") = "GPL"; + +SEC("raw_tracepoint/consume_skb") +int while_true(volatile struct pt_regs* ctx) +{ + int i = 0; + + while (true) { + if (ctx->rax & 1) + i += 3; + else + i += 7; + if (i > 40) + break; + } + + return i; +} diff --git a/tools/testing/selftests/bpf/progs/loop3.c b/tools/testing/selftests/bpf/progs/loop3.c new file mode 100644 index 0000000000000000000000000000000000000000..30a0f6cba080b9b5c69d0d073d568182287de074 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/loop3.c @@ -0,0 +1,22 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook +#include +#include +#include +#include +#include +#include +#include "bpf_helpers.h" + +char _license[] SEC("license") = "GPL"; + +SEC("raw_tracepoint/consume_skb") +int while_true(volatile struct pt_regs* ctx) +{ + __u64 i = 0, sum = 0; + do { + i++; + sum += ctx->rax; + } while (i < 0x100000000ULL); + return sum; +} diff --git a/tools/testing/selftests/bpf/progs/pyperf.h b/tools/testing/selftests/bpf/progs/pyperf.h index 0cc5e4ee90bd7d7c5f40170537dd80266452a391..6b0781391be5d83934a1eb1281f0100f695d765f 100644 --- a/tools/testing/selftests/bpf/progs/pyperf.h +++ b/tools/testing/selftests/bpf/progs/pyperf.h @@ -220,7 +220,11 @@ static inline __attribute__((__always_inline__)) int __on_event(struct pt_regs * int32_t* symbol_counter = bpf_map_lookup_elem(&symbolmap, &sym); if (symbol_counter == NULL) return 0; -#pragma unroll +#ifdef NO_UNROLL +#pragma clang loop unroll(disable) +#else +#pragma clang loop unroll(full) +#endif /* Unwind python stack */ for (int i = 0; i < STACK_MAX_LEN; ++i) { if (frame_ptr && get_frame_data(frame_ptr, pidData, &frame, &sym)) { diff --git a/tools/testing/selftests/bpf/progs/pyperf600.c b/tools/testing/selftests/bpf/progs/pyperf600.c new file mode 100644 index 0000000000000000000000000000000000000000..cb49b89e37cd4783abdf9c481a013f62f5467dd4 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/pyperf600.c @@ -0,0 +1,9 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook +#define STACK_MAX_LEN 600 +/* clang will not unroll the loop 600 times. + * Instead it will unroll it to the amount it deemed + * appropriate, but the loop will still execute 600 times. + * Total program size is around 90k insns + */ +#include "pyperf.h" diff --git a/tools/testing/selftests/bpf/progs/pyperf600_nounroll.c b/tools/testing/selftests/bpf/progs/pyperf600_nounroll.c new file mode 100644 index 0000000000000000000000000000000000000000..6beff7502f4d7bb0528dfdee5c786a1a61c70bf1 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/pyperf600_nounroll.c @@ -0,0 +1,8 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook +#define STACK_MAX_LEN 600 +#define NO_UNROLL +/* clang will not unroll at all. + * Total program size is around 2k insns + */ +#include "pyperf.h" diff --git a/tools/testing/selftests/bpf/progs/sockmap_parse_prog.c b/tools/testing/selftests/bpf/progs/sockmap_parse_prog.c index 0f92858f6226a37feb921eeea494902a2520d8bc..b061dfb014dab3e63db8bdd918485053017950bb 100644 --- a/tools/testing/selftests/bpf/progs/sockmap_parse_prog.c +++ b/tools/testing/selftests/bpf/progs/sockmap_parse_prog.c @@ -1,6 +1,5 @@ #include #include "bpf_helpers.h" -#include "bpf_util.h" #include "bpf_endian.h" int _version SEC("version") = 1; diff --git a/tools/testing/selftests/bpf/progs/sockmap_tcp_msg_prog.c b/tools/testing/selftests/bpf/progs/sockmap_tcp_msg_prog.c index 12a7b5c82ed62331503d44f2f2efe96f50118f8d..d8663e95b221d52a04d668e6a535a32e37f24964 100644 --- a/tools/testing/selftests/bpf/progs/sockmap_tcp_msg_prog.c +++ b/tools/testing/selftests/bpf/progs/sockmap_tcp_msg_prog.c @@ -1,6 +1,6 @@ #include + #include "bpf_helpers.h" -#include "bpf_util.h" #include "bpf_endian.h" int _version SEC("version") = 1; diff --git a/tools/testing/selftests/bpf/progs/sockmap_verdict_prog.c b/tools/testing/selftests/bpf/progs/sockmap_verdict_prog.c index 2ce7634a4012a19c6b35c3bdf35d4518c5ae5c46..49c9a6cf6825c7b510c90b386cf0a0886c2cc48d 100644 --- a/tools/testing/selftests/bpf/progs/sockmap_verdict_prog.c +++ b/tools/testing/selftests/bpf/progs/sockmap_verdict_prog.c @@ -1,6 +1,5 @@ #include #include "bpf_helpers.h" -#include "bpf_util.h" #include "bpf_endian.h" int _version SEC("version") = 1; diff --git a/tools/testing/selftests/bpf/progs/strobemeta.c b/tools/testing/selftests/bpf/progs/strobemeta.c new file mode 100644 index 0000000000000000000000000000000000000000..77b50b7041843b013aaa243744445bef344390f0 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/strobemeta.c @@ -0,0 +1,9 @@ +// SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) +// Copyright (c) 2019 Facebook + +#define STROBE_MAX_INTS 2 +#define STROBE_MAX_STRS 25 +#define STROBE_MAX_MAPS 100 +#define STROBE_MAX_MAP_ENTRIES 20 +/* full unroll by llvm #undef NO_UNROLL */ +#include "strobemeta.h" diff --git a/tools/testing/selftests/bpf/progs/strobemeta.h b/tools/testing/selftests/bpf/progs/strobemeta.h new file mode 100644 index 0000000000000000000000000000000000000000..82ee72bc53db734623d2de3a9d67689dd675e4c5 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/strobemeta.h @@ -0,0 +1,531 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook + +#include +#include +#include +#include +#include +#include +#include +#include "bpf_helpers.h" + +typedef uint32_t pid_t; +struct task_struct {}; + +#define TASK_COMM_LEN 16 +#define PERF_MAX_STACK_DEPTH 127 + +#define STROBE_TYPE_INVALID 0 +#define STROBE_TYPE_INT 1 +#define STROBE_TYPE_STR 2 +#define STROBE_TYPE_MAP 3 + +#define STACK_TABLE_EPOCH_SHIFT 20 +#define STROBE_MAX_STR_LEN 1 +#define STROBE_MAX_CFGS 32 +#define STROBE_MAX_PAYLOAD \ + (STROBE_MAX_STRS * STROBE_MAX_STR_LEN + \ + STROBE_MAX_MAPS * (1 + STROBE_MAX_MAP_ENTRIES * 2) * STROBE_MAX_STR_LEN) + +struct strobe_value_header { + /* + * meaning depends on type: + * 1. int: 0, if value not set, 1 otherwise + * 2. str: 1 always, whether value is set or not is determined by ptr + * 3. map: 1 always, pointer points to additional struct with number + * of entries (up to STROBE_MAX_MAP_ENTRIES) + */ + uint16_t len; + /* + * _reserved might be used for some future fields/flags, but we always + * want to keep strobe_value_header to be 8 bytes, so BPF can read 16 + * bytes in one go and get both header and value + */ + uint8_t _reserved[6]; +}; + +/* + * strobe_value_generic is used from BPF probe only, but needs to be a union + * of strobe_value_int/strobe_value_str/strobe_value_map + */ +struct strobe_value_generic { + struct strobe_value_header header; + union { + int64_t val; + void *ptr; + }; +}; + +struct strobe_value_int { + struct strobe_value_header header; + int64_t value; +}; + +struct strobe_value_str { + struct strobe_value_header header; + const char* value; +}; + +struct strobe_value_map { + struct strobe_value_header header; + const struct strobe_map_raw* value; +}; + +struct strobe_map_entry { + const char* key; + const char* val; +}; + +/* + * Map of C-string key/value pairs with fixed maximum capacity. Each map has + * corresponding int64 ID, which application can use (or ignore) in whatever + * way appropriate. Map is "write-only", there is no way to get data out of + * map. Map is intended to be used to provide metadata for profilers and is + * not to be used for internal in-app communication. All methods are + * thread-safe. + */ +struct strobe_map_raw { + /* + * general purpose unique ID that's up to application to decide + * whether and how to use; for request metadata use case id is unique + * request ID that's used to match metadata with stack traces on + * Strobelight backend side + */ + int64_t id; + /* number of used entries in map */ + int64_t cnt; + /* + * having volatile doesn't change anything on BPF side, but clang + * emits warnings for passing `volatile const char *` into + * bpf_probe_read_str that expects just `const char *` + */ + const char* tag; + /* + * key/value entries, each consisting of 2 pointers to key and value + * C strings + */ + struct strobe_map_entry entries[STROBE_MAX_MAP_ENTRIES]; +}; + +/* Following values define supported values of TLS mode */ +#define TLS_NOT_SET -1 +#define TLS_LOCAL_EXEC 0 +#define TLS_IMM_EXEC 1 +#define TLS_GENERAL_DYN 2 + +/* + * structure that universally represents TLS location (both for static + * executables and shared libraries) + */ +struct strobe_value_loc { + /* + * tls_mode defines what TLS mode was used for particular metavariable: + * - -1 (TLS_NOT_SET) - no metavariable; + * - 0 (TLS_LOCAL_EXEC) - Local Executable mode; + * - 1 (TLS_IMM_EXEC) - Immediate Executable mode; + * - 2 (TLS_GENERAL_DYN) - General Dynamic mode; + * Local Dynamic mode is not yet supported, because never seen in + * practice. Mode defines how offset field is interpreted. See + * calc_location() in below for details. + */ + int64_t tls_mode; + /* + * TLS_LOCAL_EXEC: offset from thread pointer (fs:0 for x86-64, + * tpidr_el0 for aarch64). + * TLS_IMM_EXEC: absolute address of GOT entry containing offset + * from thread pointer; + * TLS_GENERAL_DYN: absolute addres of double GOT entry + * containing tls_index_t struct; + */ + int64_t offset; +}; + +struct strobemeta_cfg { + int64_t req_meta_idx; + struct strobe_value_loc int_locs[STROBE_MAX_INTS]; + struct strobe_value_loc str_locs[STROBE_MAX_STRS]; + struct strobe_value_loc map_locs[STROBE_MAX_MAPS]; +}; + +struct strobe_map_descr { + uint64_t id; + int16_t tag_len; + /* + * cnt <0 - map value isn't set; + * 0 - map has id set, but no key/value entries + */ + int16_t cnt; + /* + * both key_lens[i] and val_lens[i] should be >0 for present key/value + * entry + */ + uint16_t key_lens[STROBE_MAX_MAP_ENTRIES]; + uint16_t val_lens[STROBE_MAX_MAP_ENTRIES]; +}; + +struct strobemeta_payload { + /* req_id has valid request ID, if req_meta_valid == 1 */ + int64_t req_id; + uint8_t req_meta_valid; + /* + * mask has Nth bit set to 1, if Nth metavar was present and + * successfully read + */ + uint64_t int_vals_set_mask; + int64_t int_vals[STROBE_MAX_INTS]; + /* len is >0 for present values */ + uint16_t str_lens[STROBE_MAX_STRS]; + /* if map_descrs[i].cnt == -1, metavar is not present/set */ + struct strobe_map_descr map_descrs[STROBE_MAX_MAPS]; + /* + * payload has compactly packed values of str and map variables in the + * form: strval1\0strval2\0map1key1\0map1val1\0map2key1\0map2val1\0 + * (and so on); str_lens[i], key_lens[i] and val_lens[i] determines + * value length + */ + char payload[STROBE_MAX_PAYLOAD]; +}; + +struct strobelight_bpf_sample { + uint64_t ktime; + char comm[TASK_COMM_LEN]; + pid_t pid; + int user_stack_id; + int kernel_stack_id; + int has_meta; + struct strobemeta_payload metadata; + /* + * makes it possible to pass ( + 1) as data size to + * perf_submit() to avoid perf_submit's paranoia about passing zero as + * size, as it deduces that might be + * **theoretically** zero + */ + char dummy_safeguard; +}; + +struct bpf_map_def SEC("maps") samples = { + .type = BPF_MAP_TYPE_PERF_EVENT_ARRAY, + .key_size = sizeof(int), + .value_size = sizeof(int), + .max_entries = 32, +}; + +struct bpf_map_def SEC("maps") stacks_0 = { + .type = BPF_MAP_TYPE_STACK_TRACE, + .key_size = sizeof(uint32_t), + .value_size = sizeof(uint64_t) * PERF_MAX_STACK_DEPTH, + .max_entries = 16, +}; + +struct bpf_map_def SEC("maps") stacks_1 = { + .type = BPF_MAP_TYPE_STACK_TRACE, + .key_size = sizeof(uint32_t), + .value_size = sizeof(uint64_t) * PERF_MAX_STACK_DEPTH, + .max_entries = 16, +}; + +struct bpf_map_def SEC("maps") sample_heap = { + .type = BPF_MAP_TYPE_PERCPU_ARRAY, + .key_size = sizeof(uint32_t), + .value_size = sizeof(struct strobelight_bpf_sample), + .max_entries = 1, +}; + +struct bpf_map_def SEC("maps") strobemeta_cfgs = { + .type = BPF_MAP_TYPE_PERCPU_ARRAY, + .key_size = sizeof(pid_t), + .value_size = sizeof(struct strobemeta_cfg), + .max_entries = STROBE_MAX_CFGS, +}; + +/* Type for the dtv. */ +/* https://github.com/lattera/glibc/blob/master/nptl/sysdeps/x86_64/tls.h#L34 */ +typedef union dtv { + size_t counter; + struct { + void* val; + bool is_static; + } pointer; +} dtv_t; + +/* Partial definition for tcbhead_t */ +/* https://github.com/bminor/glibc/blob/master/sysdeps/x86_64/nptl/tls.h#L42 */ +struct tcbhead { + void* tcb; + dtv_t* dtv; +}; + +/* + * TLS module/offset information for shared library case. + * For x86-64, this is mapped onto two entries in GOT. + * For aarch64, this is pointed to by second GOT entry. + */ +struct tls_index { + uint64_t module; + uint64_t offset; +}; + +static inline __attribute__((always_inline)) +void *calc_location(struct strobe_value_loc *loc, void *tls_base) +{ + /* + * tls_mode value is: + * - -1 (TLS_NOT_SET), if no metavar is present; + * - 0 (TLS_LOCAL_EXEC), if metavar uses Local Executable mode of TLS + * (offset from fs:0 for x86-64 or tpidr_el0 for aarch64); + * - 1 (TLS_IMM_EXEC), if metavar uses Immediate Executable mode of TLS; + * - 2 (TLS_GENERAL_DYN), if metavar uses General Dynamic mode of TLS; + * This schema allows to use something like: + * (tls_mode + 1) * (tls_base + offset) + * to get NULL for "no metavar" location, or correct pointer for local + * executable mode without doing extra ifs. + */ + if (loc->tls_mode <= TLS_LOCAL_EXEC) { + /* static executable is simple, we just have offset from + * tls_base */ + void *addr = tls_base + loc->offset; + /* multiply by (tls_mode + 1) to get NULL, if we have no + * metavar in this slot */ + return (void *)((loc->tls_mode + 1) * (int64_t)addr); + } + /* + * Other modes are more complicated, we need to jump through few hoops. + * + * For immediate executable mode (currently supported only for aarch64): + * - loc->offset is pointing to a GOT entry containing fixed offset + * relative to tls_base; + * + * For general dynamic mode: + * - loc->offset is pointing to a beginning of double GOT entries; + * - (for aarch64 only) second entry points to tls_index_t struct; + * - (for x86-64 only) two GOT entries are already tls_index_t; + * - tls_index_t->module is used to find start of TLS section in + * which variable resides; + * - tls_index_t->offset provides offset within that TLS section, + * pointing to value of variable. + */ + struct tls_index tls_index; + dtv_t *dtv; + void *tls_ptr; + + bpf_probe_read(&tls_index, sizeof(struct tls_index), + (void *)loc->offset); + /* valid module index is always positive */ + if (tls_index.module > 0) { + /* dtv = ((struct tcbhead *)tls_base)->dtv[tls_index.module] */ + bpf_probe_read(&dtv, sizeof(dtv), + &((struct tcbhead *)tls_base)->dtv); + dtv += tls_index.module; + } else { + dtv = NULL; + } + bpf_probe_read(&tls_ptr, sizeof(void *), dtv); + /* if pointer has (void *)-1 value, then TLS wasn't initialized yet */ + return tls_ptr && tls_ptr != (void *)-1 + ? tls_ptr + tls_index.offset + : NULL; +} + +static inline __attribute__((always_inline)) +void read_int_var(struct strobemeta_cfg *cfg, size_t idx, void *tls_base, + struct strobe_value_generic *value, + struct strobemeta_payload *data) +{ + void *location = calc_location(&cfg->int_locs[idx], tls_base); + if (!location) + return; + + bpf_probe_read(value, sizeof(struct strobe_value_generic), location); + data->int_vals[idx] = value->val; + if (value->header.len) + data->int_vals_set_mask |= (1 << idx); +} + +static inline __attribute__((always_inline)) +uint64_t read_str_var(struct strobemeta_cfg* cfg, size_t idx, void *tls_base, + struct strobe_value_generic *value, + struct strobemeta_payload *data, void *payload) +{ + void *location; + uint32_t len; + + data->str_lens[idx] = 0; + location = calc_location(&cfg->str_locs[idx], tls_base); + if (!location) + return 0; + + bpf_probe_read(value, sizeof(struct strobe_value_generic), location); + len = bpf_probe_read_str(payload, STROBE_MAX_STR_LEN, value->ptr); + /* + * if bpf_probe_read_str returns error (<0), due to casting to + * unsinged int, it will become big number, so next check is + * sufficient to check for errors AND prove to BPF verifier, that + * bpf_probe_read_str won't return anything bigger than + * STROBE_MAX_STR_LEN + */ + if (len > STROBE_MAX_STR_LEN) + return 0; + + data->str_lens[idx] = len; + return len; +} + +static inline __attribute__((always_inline)) +void *read_map_var(struct strobemeta_cfg *cfg, size_t idx, void *tls_base, + struct strobe_value_generic *value, + struct strobemeta_payload* data, void *payload) +{ + struct strobe_map_descr* descr = &data->map_descrs[idx]; + struct strobe_map_raw map; + void *location; + uint32_t len; + int i; + + descr->tag_len = 0; /* presume no tag is set */ + descr->cnt = -1; /* presume no value is set */ + + location = calc_location(&cfg->map_locs[idx], tls_base); + if (!location) + return payload; + + bpf_probe_read(value, sizeof(struct strobe_value_generic), location); + if (bpf_probe_read(&map, sizeof(struct strobe_map_raw), value->ptr)) + return payload; + + descr->id = map.id; + descr->cnt = map.cnt; + if (cfg->req_meta_idx == idx) { + data->req_id = map.id; + data->req_meta_valid = 1; + } + + len = bpf_probe_read_str(payload, STROBE_MAX_STR_LEN, map.tag); + if (len <= STROBE_MAX_STR_LEN) { + descr->tag_len = len; + payload += len; + } + +#ifdef NO_UNROLL +#pragma clang loop unroll(disable) +#else +#pragma unroll +#endif + for (int i = 0; i < STROBE_MAX_MAP_ENTRIES; ++i) { + if (i >= map.cnt) + break; + + descr->key_lens[i] = 0; + len = bpf_probe_read_str(payload, STROBE_MAX_STR_LEN, + map.entries[i].key); + if (len <= STROBE_MAX_STR_LEN) { + descr->key_lens[i] = len; + payload += len; + } + descr->val_lens[i] = 0; + len = bpf_probe_read_str(payload, STROBE_MAX_STR_LEN, + map.entries[i].val); + if (len <= STROBE_MAX_STR_LEN) { + descr->val_lens[i] = len; + payload += len; + } + } + + return payload; +} + +/* + * read_strobe_meta returns NULL, if no metadata was read; otherwise returns + * pointer to *right after* payload ends + */ +static inline __attribute__((always_inline)) +void *read_strobe_meta(struct task_struct* task, + struct strobemeta_payload* data) { + pid_t pid = bpf_get_current_pid_tgid() >> 32; + struct strobe_value_generic value = {0}; + struct strobemeta_cfg *cfg; + void *tls_base, *payload; + + cfg = bpf_map_lookup_elem(&strobemeta_cfgs, &pid); + if (!cfg) + return NULL; + + data->int_vals_set_mask = 0; + data->req_meta_valid = 0; + payload = data->payload; + /* + * we don't have struct task_struct definition, it should be: + * tls_base = (void *)task->thread.fsbase; + */ + tls_base = (void *)task; + +#ifdef NO_UNROLL +#pragma clang loop unroll(disable) +#else +#pragma unroll +#endif + for (int i = 0; i < STROBE_MAX_INTS; ++i) { + read_int_var(cfg, i, tls_base, &value, data); + } +#ifdef NO_UNROLL +#pragma clang loop unroll(disable) +#else +#pragma unroll +#endif + for (int i = 0; i < STROBE_MAX_STRS; ++i) { + payload += read_str_var(cfg, i, tls_base, &value, data, payload); + } +#ifdef NO_UNROLL +#pragma clang loop unroll(disable) +#else +#pragma unroll +#endif + for (int i = 0; i < STROBE_MAX_MAPS; ++i) { + payload = read_map_var(cfg, i, tls_base, &value, data, payload); + } + /* + * return pointer right after end of payload, so it's possible to + * calculate exact amount of useful data that needs to be sent + */ + return payload; +} + +SEC("raw_tracepoint/kfree_skb") +int on_event(struct pt_regs *ctx) { + pid_t pid = bpf_get_current_pid_tgid() >> 32; + struct strobelight_bpf_sample* sample; + struct task_struct *task; + uint32_t zero = 0; + uint64_t ktime_ns; + void *sample_end; + + sample = bpf_map_lookup_elem(&sample_heap, &zero); + if (!sample) + return 0; /* this will never happen */ + + sample->pid = pid; + bpf_get_current_comm(&sample->comm, TASK_COMM_LEN); + ktime_ns = bpf_ktime_get_ns(); + sample->ktime = ktime_ns; + + task = (struct task_struct *)bpf_get_current_task(); + sample_end = read_strobe_meta(task, &sample->metadata); + sample->has_meta = sample_end != NULL; + sample_end = sample_end ? : &sample->metadata; + + if ((ktime_ns >> STACK_TABLE_EPOCH_SHIFT) & 1) { + sample->kernel_stack_id = bpf_get_stackid(ctx, &stacks_1, 0); + sample->user_stack_id = bpf_get_stackid(ctx, &stacks_1, BPF_F_USER_STACK); + } else { + sample->kernel_stack_id = bpf_get_stackid(ctx, &stacks_0, 0); + sample->user_stack_id = bpf_get_stackid(ctx, &stacks_0, BPF_F_USER_STACK); + } + + uint64_t sample_size = sample_end - (void *)sample; + /* should always be true */ + if (sample_size < sizeof(struct strobelight_bpf_sample)) + bpf_perf_event_output(ctx, &samples, 0, sample, 1 + sample_size); + return 0; +} + +char _license[] SEC("license") = "GPL"; diff --git a/tools/testing/selftests/bpf/progs/strobemeta_nounroll1.c b/tools/testing/selftests/bpf/progs/strobemeta_nounroll1.c new file mode 100644 index 0000000000000000000000000000000000000000..f0a1669e11d64fffb4575c4d7905ba68d779e23b --- /dev/null +++ b/tools/testing/selftests/bpf/progs/strobemeta_nounroll1.c @@ -0,0 +1,9 @@ +// SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) +// Copyright (c) 2019 Facebook + +#define STROBE_MAX_INTS 2 +#define STROBE_MAX_STRS 25 +#define STROBE_MAX_MAPS 13 +#define STROBE_MAX_MAP_ENTRIES 20 +#define NO_UNROLL +#include "strobemeta.h" diff --git a/tools/testing/selftests/bpf/progs/strobemeta_nounroll2.c b/tools/testing/selftests/bpf/progs/strobemeta_nounroll2.c new file mode 100644 index 0000000000000000000000000000000000000000..4291a7d642e775bc8d481cbda56ac47a1f84bf02 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/strobemeta_nounroll2.c @@ -0,0 +1,9 @@ +// SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) +// Copyright (c) 2019 Facebook + +#define STROBE_MAX_INTS 2 +#define STROBE_MAX_STRS 25 +#define STROBE_MAX_MAPS 30 +#define STROBE_MAX_MAP_ENTRIES 20 +#define NO_UNROLL +#include "strobemeta.h" diff --git a/tools/testing/selftests/bpf/progs/test_seg6_loop.c b/tools/testing/selftests/bpf/progs/test_seg6_loop.c new file mode 100644 index 0000000000000000000000000000000000000000..463964d79f73d782c170a86a02e1402b81550571 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/test_seg6_loop.c @@ -0,0 +1,261 @@ +#include +#include +#include +#include +#include +#include "bpf_helpers.h" +#include "bpf_endian.h" + +/* Packet parsing state machine helpers. */ +#define cursor_advance(_cursor, _len) \ + ({ void *_tmp = _cursor; _cursor += _len; _tmp; }) + +#define SR6_FLAG_ALERT (1 << 4) + +#define htonll(x) ((bpf_htonl(1)) == 1 ? (x) : ((uint64_t)bpf_htonl((x) & \ + 0xFFFFFFFF) << 32) | bpf_htonl((x) >> 32)) +#define ntohll(x) ((bpf_ntohl(1)) == 1 ? (x) : ((uint64_t)bpf_ntohl((x) & \ + 0xFFFFFFFF) << 32) | bpf_ntohl((x) >> 32)) +#define BPF_PACKET_HEADER __attribute__((packed)) + +struct ip6_t { + unsigned int ver:4; + unsigned int priority:8; + unsigned int flow_label:20; + unsigned short payload_len; + unsigned char next_header; + unsigned char hop_limit; + unsigned long long src_hi; + unsigned long long src_lo; + unsigned long long dst_hi; + unsigned long long dst_lo; +} BPF_PACKET_HEADER; + +struct ip6_addr_t { + unsigned long long hi; + unsigned long long lo; +} BPF_PACKET_HEADER; + +struct ip6_srh_t { + unsigned char nexthdr; + unsigned char hdrlen; + unsigned char type; + unsigned char segments_left; + unsigned char first_segment; + unsigned char flags; + unsigned short tag; + + struct ip6_addr_t segments[0]; +} BPF_PACKET_HEADER; + +struct sr6_tlv_t { + unsigned char type; + unsigned char len; + unsigned char value[0]; +} BPF_PACKET_HEADER; + +static __attribute__((always_inline)) struct ip6_srh_t *get_srh(struct __sk_buff *skb) +{ + void *cursor, *data_end; + struct ip6_srh_t *srh; + struct ip6_t *ip; + uint8_t *ipver; + + data_end = (void *)(long)skb->data_end; + cursor = (void *)(long)skb->data; + ipver = (uint8_t *)cursor; + + if ((void *)ipver + sizeof(*ipver) > data_end) + return NULL; + + if ((*ipver >> 4) != 6) + return NULL; + + ip = cursor_advance(cursor, sizeof(*ip)); + if ((void *)ip + sizeof(*ip) > data_end) + return NULL; + + if (ip->next_header != 43) + return NULL; + + srh = cursor_advance(cursor, sizeof(*srh)); + if ((void *)srh + sizeof(*srh) > data_end) + return NULL; + + if (srh->type != 4) + return NULL; + + return srh; +} + +static __attribute__((always_inline)) +int update_tlv_pad(struct __sk_buff *skb, uint32_t new_pad, + uint32_t old_pad, uint32_t pad_off) +{ + int err; + + if (new_pad != old_pad) { + err = bpf_lwt_seg6_adjust_srh(skb, pad_off, + (int) new_pad - (int) old_pad); + if (err) + return err; + } + + if (new_pad > 0) { + char pad_tlv_buf[16] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0}; + struct sr6_tlv_t *pad_tlv = (struct sr6_tlv_t *) pad_tlv_buf; + + pad_tlv->type = SR6_TLV_PADDING; + pad_tlv->len = new_pad - 2; + + err = bpf_lwt_seg6_store_bytes(skb, pad_off, + (void *)pad_tlv_buf, new_pad); + if (err) + return err; + } + + return 0; +} + +static __attribute__((always_inline)) +int is_valid_tlv_boundary(struct __sk_buff *skb, struct ip6_srh_t *srh, + uint32_t *tlv_off, uint32_t *pad_size, + uint32_t *pad_off) +{ + uint32_t srh_off, cur_off; + int offset_valid = 0; + int err; + + srh_off = (char *)srh - (char *)(long)skb->data; + // cur_off = end of segments, start of possible TLVs + cur_off = srh_off + sizeof(*srh) + + sizeof(struct ip6_addr_t) * (srh->first_segment + 1); + + *pad_off = 0; + + // we can only go as far as ~10 TLVs due to the BPF max stack size + #pragma clang loop unroll(disable) + for (int i = 0; i < 100; i++) { + struct sr6_tlv_t tlv; + + if (cur_off == *tlv_off) + offset_valid = 1; + + if (cur_off >= srh_off + ((srh->hdrlen + 1) << 3)) + break; + + err = bpf_skb_load_bytes(skb, cur_off, &tlv, sizeof(tlv)); + if (err) + return err; + + if (tlv.type == SR6_TLV_PADDING) { + *pad_size = tlv.len + sizeof(tlv); + *pad_off = cur_off; + + if (*tlv_off == srh_off) { + *tlv_off = cur_off; + offset_valid = 1; + } + break; + + } else if (tlv.type == SR6_TLV_HMAC) { + break; + } + + cur_off += sizeof(tlv) + tlv.len; + } // we reached the padding or HMAC TLVs, or the end of the SRH + + if (*pad_off == 0) + *pad_off = cur_off; + + if (*tlv_off == -1) + *tlv_off = cur_off; + else if (!offset_valid) + return -EINVAL; + + return 0; +} + +static __attribute__((always_inline)) +int add_tlv(struct __sk_buff *skb, struct ip6_srh_t *srh, uint32_t tlv_off, + struct sr6_tlv_t *itlv, uint8_t tlv_size) +{ + uint32_t srh_off = (char *)srh - (char *)(long)skb->data; + uint8_t len_remaining, new_pad; + uint32_t pad_off = 0; + uint32_t pad_size = 0; + uint32_t partial_srh_len; + int err; + + if (tlv_off != -1) + tlv_off += srh_off; + + if (itlv->type == SR6_TLV_PADDING || itlv->type == SR6_TLV_HMAC) + return -EINVAL; + + err = is_valid_tlv_boundary(skb, srh, &tlv_off, &pad_size, &pad_off); + if (err) + return err; + + err = bpf_lwt_seg6_adjust_srh(skb, tlv_off, sizeof(*itlv) + itlv->len); + if (err) + return err; + + err = bpf_lwt_seg6_store_bytes(skb, tlv_off, (void *)itlv, tlv_size); + if (err) + return err; + + // the following can't be moved inside update_tlv_pad because the + // bpf verifier has some issues with it + pad_off += sizeof(*itlv) + itlv->len; + partial_srh_len = pad_off - srh_off; + len_remaining = partial_srh_len % 8; + new_pad = 8 - len_remaining; + + if (new_pad == 1) // cannot pad for 1 byte only + new_pad = 9; + else if (new_pad == 8) + new_pad = 0; + + return update_tlv_pad(skb, new_pad, pad_size, pad_off); +} + +// Add an Egress TLV fc00::4, add the flag A, +// and apply End.X action to fc42::1 +SEC("lwt_seg6local") +int __add_egr_x(struct __sk_buff *skb) +{ + unsigned long long hi = 0xfc42000000000000; + unsigned long long lo = 0x1; + struct ip6_srh_t *srh = get_srh(skb); + uint8_t new_flags = SR6_FLAG_ALERT; + struct ip6_addr_t addr; + int err, offset; + + if (srh == NULL) + return BPF_DROP; + + uint8_t tlv[20] = {2, 18, 0, 0, 0xfd, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, + 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x4}; + + err = add_tlv(skb, srh, (srh->hdrlen+1) << 3, + (struct sr6_tlv_t *)&tlv, 20); + if (err) + return BPF_DROP; + + offset = sizeof(struct ip6_t) + offsetof(struct ip6_srh_t, flags); + err = bpf_lwt_seg6_store_bytes(skb, offset, + (void *)&new_flags, sizeof(new_flags)); + if (err) + return BPF_DROP; + + addr.lo = htonll(lo); + addr.hi = htonll(hi); + err = bpf_lwt_seg6_action(skb, SEG6_LOCAL_ACTION_END_X, + (void *)&addr, sizeof(addr)); + if (err) + return BPF_DROP; + return BPF_REDIRECT; +} +char __license[] SEC("license") = "GPL"; diff --git a/tools/testing/selftests/bpf/progs/test_sysctl_loop1.c b/tools/testing/selftests/bpf/progs/test_sysctl_loop1.c new file mode 100644 index 0000000000000000000000000000000000000000..ac1c0ef3c46cb170ac2edcc994cc67a8ba5116b2 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/test_sysctl_loop1.c @@ -0,0 +1,71 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook + +#include +#include + +#include +#include + +#include "bpf_helpers.h" + +#ifndef ARRAY_SIZE +#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) +#endif + +/* tcp_mem sysctl has only 3 ints, but this test is doing TCP_MEM_LOOPS */ +#define TCP_MEM_LOOPS 28 /* because 30 doesn't fit into 512 bytes of stack */ +#define MAX_ULONG_STR_LEN 7 +#define MAX_VALUE_STR_LEN (TCP_MEM_LOOPS * MAX_ULONG_STR_LEN) + +const char tcp_mem_name[] = "net/ipv4/tcp_mem/very_very_very_very_long_pointless_string"; +static __always_inline int is_tcp_mem(struct bpf_sysctl *ctx) +{ + unsigned char i; + char name[sizeof(tcp_mem_name)]; + int ret; + + memset(name, 0, sizeof(name)); + ret = bpf_sysctl_get_name(ctx, name, sizeof(name), 0); + if (ret < 0 || ret != sizeof(tcp_mem_name) - 1) + return 0; + +#pragma clang loop unroll(disable) + for (i = 0; i < sizeof(tcp_mem_name); ++i) + if (name[i] != tcp_mem_name[i]) + return 0; + + return 1; +} + +SEC("cgroup/sysctl") +int sysctl_tcp_mem(struct bpf_sysctl *ctx) +{ + unsigned long tcp_mem[TCP_MEM_LOOPS] = {}; + char value[MAX_VALUE_STR_LEN]; + unsigned char i, off = 0; + int ret; + + if (ctx->write) + return 0; + + if (!is_tcp_mem(ctx)) + return 0; + + ret = bpf_sysctl_get_current_value(ctx, value, MAX_VALUE_STR_LEN); + if (ret < 0 || ret >= MAX_VALUE_STR_LEN) + return 0; + +#pragma clang loop unroll(disable) + for (i = 0; i < ARRAY_SIZE(tcp_mem); ++i) { + ret = bpf_strtoul(value + off, MAX_ULONG_STR_LEN, 0, + tcp_mem + i); + if (ret <= 0 || ret > MAX_ULONG_STR_LEN) + return 0; + off += ret & MAX_ULONG_STR_LEN; + } + + return tcp_mem[0] < tcp_mem[1] && tcp_mem[1] < tcp_mem[2]; +} + +char _license[] SEC("license") = "GPL"; diff --git a/tools/testing/selftests/bpf/progs/test_sysctl_loop2.c b/tools/testing/selftests/bpf/progs/test_sysctl_loop2.c new file mode 100644 index 0000000000000000000000000000000000000000..55251046c9b73b751127b28b23af1972197ccc6d --- /dev/null +++ b/tools/testing/selftests/bpf/progs/test_sysctl_loop2.c @@ -0,0 +1,72 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook + +#include +#include + +#include +#include + +#include "bpf_helpers.h" + +#ifndef ARRAY_SIZE +#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) +#endif + +/* tcp_mem sysctl has only 3 ints, but this test is doing TCP_MEM_LOOPS */ +#define TCP_MEM_LOOPS 20 /* because 30 doesn't fit into 512 bytes of stack */ +#define MAX_ULONG_STR_LEN 7 +#define MAX_VALUE_STR_LEN (TCP_MEM_LOOPS * MAX_ULONG_STR_LEN) + +const char tcp_mem_name[] = "net/ipv4/tcp_mem/very_very_very_very_long_pointless_string_to_stress_byte_loop"; +static __attribute__((noinline)) int is_tcp_mem(struct bpf_sysctl *ctx) +{ + unsigned char i; + char name[sizeof(tcp_mem_name)]; + int ret; + + memset(name, 0, sizeof(name)); + ret = bpf_sysctl_get_name(ctx, name, sizeof(name), 0); + if (ret < 0 || ret != sizeof(tcp_mem_name) - 1) + return 0; + +#pragma clang loop unroll(disable) + for (i = 0; i < sizeof(tcp_mem_name); ++i) + if (name[i] != tcp_mem_name[i]) + return 0; + + return 1; +} + + +SEC("cgroup/sysctl") +int sysctl_tcp_mem(struct bpf_sysctl *ctx) +{ + unsigned long tcp_mem[TCP_MEM_LOOPS] = {}; + char value[MAX_VALUE_STR_LEN]; + unsigned char i, off = 0; + int ret; + + if (ctx->write) + return 0; + + if (!is_tcp_mem(ctx)) + return 0; + + ret = bpf_sysctl_get_current_value(ctx, value, MAX_VALUE_STR_LEN); + if (ret < 0 || ret >= MAX_VALUE_STR_LEN) + return 0; + +#pragma clang loop unroll(disable) + for (i = 0; i < ARRAY_SIZE(tcp_mem); ++i) { + ret = bpf_strtoul(value + off, MAX_ULONG_STR_LEN, 0, + tcp_mem + i); + if (ret <= 0 || ret > MAX_ULONG_STR_LEN) + return 0; + off += ret & MAX_ULONG_STR_LEN; + } + + return tcp_mem[0] < tcp_mem[1] && tcp_mem[1] < tcp_mem[2]; +} + +char _license[] SEC("license") = "GPL"; diff --git a/tools/testing/selftests/bpf/progs/test_sysctl_prog.c b/tools/testing/selftests/bpf/progs/test_sysctl_prog.c index a295cad805d7d4ccdc30277376737e57718fffd3..5cbbff416998c78fb889129bf90a5f7aafc53a7c 100644 --- a/tools/testing/selftests/bpf/progs/test_sysctl_prog.c +++ b/tools/testing/selftests/bpf/progs/test_sysctl_prog.c @@ -8,7 +8,6 @@ #include #include "bpf_helpers.h" -#include "bpf_util.h" /* Max supported length of a string with unsigned long in base 10 (pow2 - 1). */ #define MAX_ULONG_STR_LEN 0xF @@ -16,6 +15,10 @@ /* Max supported length of sysctl value string (pow2). */ #define MAX_VALUE_STR_LEN 0x40 +#ifndef ARRAY_SIZE +#define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) +#endif + static __always_inline int is_tcp_mem(struct bpf_sysctl *ctx) { char tcp_mem_name[] = "net/ipv4/tcp_mem"; diff --git a/tools/testing/selftests/bpf/progs/test_tcpbpf_kern.c b/tools/testing/selftests/bpf/progs/test_tcpbpf_kern.c index 4b7fd540cea9dd89d506bfe693be0414e79b6d90..236f7649bd9cdeb3a82ff99e6a9a29a7ed6fb4be 100644 --- a/tools/testing/selftests/bpf/progs/test_tcpbpf_kern.c +++ b/tools/testing/selftests/bpf/progs/test_tcpbpf_kern.c @@ -45,12 +45,53 @@ int _version SEC("version") = 1; SEC("sockops") int bpf_testcb(struct bpf_sock_ops *skops) { + struct bpf_sock_ops *reuse = skops; int rv = -1; int bad_call_rv = 0; int good_call_rv = 0; int op; int v = 0; + /* Test reading fields in bpf_sock_ops using single register */ + asm volatile ( + "%[reuse] = *(u32 *)(%[reuse] +96)" + : [reuse] "+r"(reuse) + :); + + asm volatile ( + "%[op] = *(u32 *)(%[skops] +96)" + : [op] "+r"(op) + : [skops] "r"(skops) + :); + + asm volatile ( + "r9 = %[skops];\n" + "r8 = *(u32 *)(r9 +164);\n" + "*(u32 *)(r9 +164) = r8;\n" + :: [skops] "r"(skops) + : "r9", "r8"); + + asm volatile ( + "r1 = %[skops];\n" + "r1 = *(u64 *)(r1 +184);\n" + "if r1 == 0 goto +1;\n" + "r1 = *(u32 *)(r1 +4);\n" + :: [skops] "r"(skops):"r1"); + + asm volatile ( + "r9 = %[skops];\n" + "r9 = *(u64 *)(r9 +184);\n" + "if r9 == 0 goto +1;\n" + "r9 = *(u32 *)(r9 +4);\n" + :: [skops] "r"(skops):"r9"); + + asm volatile ( + "r1 = %[skops];\n" + "r2 = *(u64 *)(r1 +184);\n" + "if r2 == 0 goto +1;\n" + "r2 = *(u32 *)(r2 +4);\n" + :: [skops] "r"(skops):"r1", "r2"); + op = (int) skops->op; update_event_map(op); diff --git a/tools/testing/selftests/bpf/progs/test_xdp_loop.c b/tools/testing/selftests/bpf/progs/test_xdp_loop.c new file mode 100644 index 0000000000000000000000000000000000000000..7fa4677df22e17fff0f99eae030a67f2584cca78 --- /dev/null +++ b/tools/testing/selftests/bpf/progs/test_xdp_loop.c @@ -0,0 +1,231 @@ +// SPDX-License-Identifier: GPL-2.0 +// Copyright (c) 2019 Facebook +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "bpf_helpers.h" +#include "bpf_endian.h" +#include "test_iptunnel_common.h" + +int _version SEC("version") = 1; + +struct bpf_map_def SEC("maps") rxcnt = { + .type = BPF_MAP_TYPE_PERCPU_ARRAY, + .key_size = sizeof(__u32), + .value_size = sizeof(__u64), + .max_entries = 256, +}; + +struct bpf_map_def SEC("maps") vip2tnl = { + .type = BPF_MAP_TYPE_HASH, + .key_size = sizeof(struct vip), + .value_size = sizeof(struct iptnl_info), + .max_entries = MAX_IPTNL_ENTRIES, +}; + +static __always_inline void count_tx(__u32 protocol) +{ + __u64 *rxcnt_count; + + rxcnt_count = bpf_map_lookup_elem(&rxcnt, &protocol); + if (rxcnt_count) + *rxcnt_count += 1; +} + +static __always_inline int get_dport(void *trans_data, void *data_end, + __u8 protocol) +{ + struct tcphdr *th; + struct udphdr *uh; + + switch (protocol) { + case IPPROTO_TCP: + th = (struct tcphdr *)trans_data; + if (th + 1 > data_end) + return -1; + return th->dest; + case IPPROTO_UDP: + uh = (struct udphdr *)trans_data; + if (uh + 1 > data_end) + return -1; + return uh->dest; + default: + return 0; + } +} + +static __always_inline void set_ethhdr(struct ethhdr *new_eth, + const struct ethhdr *old_eth, + const struct iptnl_info *tnl, + __be16 h_proto) +{ + memcpy(new_eth->h_source, old_eth->h_dest, sizeof(new_eth->h_source)); + memcpy(new_eth->h_dest, tnl->dmac, sizeof(new_eth->h_dest)); + new_eth->h_proto = h_proto; +} + +static __always_inline int handle_ipv4(struct xdp_md *xdp) +{ + void *data_end = (void *)(long)xdp->data_end; + void *data = (void *)(long)xdp->data; + struct iptnl_info *tnl; + struct ethhdr *new_eth; + struct ethhdr *old_eth; + struct iphdr *iph = data + sizeof(struct ethhdr); + __u16 *next_iph; + __u16 payload_len; + struct vip vip = {}; + int dport; + __u32 csum = 0; + int i; + + if (iph + 1 > data_end) + return XDP_DROP; + + dport = get_dport(iph + 1, data_end, iph->protocol); + if (dport == -1) + return XDP_DROP; + + vip.protocol = iph->protocol; + vip.family = AF_INET; + vip.daddr.v4 = iph->daddr; + vip.dport = dport; + payload_len = bpf_ntohs(iph->tot_len); + + tnl = bpf_map_lookup_elem(&vip2tnl, &vip); + /* It only does v4-in-v4 */ + if (!tnl || tnl->family != AF_INET) + return XDP_PASS; + + if (bpf_xdp_adjust_head(xdp, 0 - (int)sizeof(struct iphdr))) + return XDP_DROP; + + data = (void *)(long)xdp->data; + data_end = (void *)(long)xdp->data_end; + + new_eth = data; + iph = data + sizeof(*new_eth); + old_eth = data + sizeof(*iph); + + if (new_eth + 1 > data_end || + old_eth + 1 > data_end || + iph + 1 > data_end) + return XDP_DROP; + + set_ethhdr(new_eth, old_eth, tnl, bpf_htons(ETH_P_IP)); + + iph->version = 4; + iph->ihl = sizeof(*iph) >> 2; + iph->frag_off = 0; + iph->protocol = IPPROTO_IPIP; + iph->check = 0; + iph->tos = 0; + iph->tot_len = bpf_htons(payload_len + sizeof(*iph)); + iph->daddr = tnl->daddr.v4; + iph->saddr = tnl->saddr.v4; + iph->ttl = 8; + + next_iph = (__u16 *)iph; +#pragma clang loop unroll(disable) + for (i = 0; i < sizeof(*iph) >> 1; i++) + csum += *next_iph++; + + iph->check = ~((csum & 0xffff) + (csum >> 16)); + + count_tx(vip.protocol); + + return XDP_TX; +} + +static __always_inline int handle_ipv6(struct xdp_md *xdp) +{ + void *data_end = (void *)(long)xdp->data_end; + void *data = (void *)(long)xdp->data; + struct iptnl_info *tnl; + struct ethhdr *new_eth; + struct ethhdr *old_eth; + struct ipv6hdr *ip6h = data + sizeof(struct ethhdr); + __u16 payload_len; + struct vip vip = {}; + int dport; + + if (ip6h + 1 > data_end) + return XDP_DROP; + + dport = get_dport(ip6h + 1, data_end, ip6h->nexthdr); + if (dport == -1) + return XDP_DROP; + + vip.protocol = ip6h->nexthdr; + vip.family = AF_INET6; + memcpy(vip.daddr.v6, ip6h->daddr.s6_addr32, sizeof(vip.daddr)); + vip.dport = dport; + payload_len = ip6h->payload_len; + + tnl = bpf_map_lookup_elem(&vip2tnl, &vip); + /* It only does v6-in-v6 */ + if (!tnl || tnl->family != AF_INET6) + return XDP_PASS; + + if (bpf_xdp_adjust_head(xdp, 0 - (int)sizeof(struct ipv6hdr))) + return XDP_DROP; + + data = (void *)(long)xdp->data; + data_end = (void *)(long)xdp->data_end; + + new_eth = data; + ip6h = data + sizeof(*new_eth); + old_eth = data + sizeof(*ip6h); + + if (new_eth + 1 > data_end || old_eth + 1 > data_end || + ip6h + 1 > data_end) + return XDP_DROP; + + set_ethhdr(new_eth, old_eth, tnl, bpf_htons(ETH_P_IPV6)); + + ip6h->version = 6; + ip6h->priority = 0; + memset(ip6h->flow_lbl, 0, sizeof(ip6h->flow_lbl)); + ip6h->payload_len = bpf_htons(bpf_ntohs(payload_len) + sizeof(*ip6h)); + ip6h->nexthdr = IPPROTO_IPV6; + ip6h->hop_limit = 8; + memcpy(ip6h->saddr.s6_addr32, tnl->saddr.v6, sizeof(tnl->saddr.v6)); + memcpy(ip6h->daddr.s6_addr32, tnl->daddr.v6, sizeof(tnl->daddr.v6)); + + count_tx(vip.protocol); + + return XDP_TX; +} + +SEC("xdp_tx_iptunnel") +int _xdp_tx_iptunnel(struct xdp_md *xdp) +{ + void *data_end = (void *)(long)xdp->data_end; + void *data = (void *)(long)xdp->data; + struct ethhdr *eth = data; + __u16 h_proto; + + if (eth + 1 > data_end) + return XDP_DROP; + + h_proto = eth->h_proto; + + if (h_proto == bpf_htons(ETH_P_IP)) + return handle_ipv4(xdp); + else if (h_proto == bpf_htons(ETH_P_IPV6)) + + return handle_ipv6(xdp); + else + return XDP_DROP; +} + +char _license[] SEC("license") = "GPL"; diff --git a/tools/testing/selftests/bpf/test_progs.c b/tools/testing/selftests/bpf/test_progs.c index 02d9e9eab03328023559d5a63d25ebf600c23dc8..fb5566eb06808c1b372c9956a2769d009567cc74 100644 --- a/tools/testing/selftests/bpf/test_progs.c +++ b/tools/testing/selftests/bpf/test_progs.c @@ -2374,7 +2374,7 @@ static int libbpf_debug_print_verifier_scale(enum libbpf_print_level level, const char *format, va_list args) { if (level != LIBBPF_DEBUG) - return 0; + return vfprintf(stderr, format, args); if (!strstr(format, "verifier log")) return 0; @@ -2384,7 +2384,7 @@ static int libbpf_debug_print_verifier_scale(enum libbpf_print_level level, static int check_load(const char *file, enum bpf_prog_type type) { struct bpf_prog_load_attr attr; - struct bpf_object *obj; + struct bpf_object *obj = NULL; int err, prog_fd; memset(&attr, 0, sizeof(struct bpf_prog_load_attr)); @@ -2400,26 +2400,71 @@ static int check_load(const char *file, enum bpf_prog_type type) void test_bpf_verif_scale(void) { - const char *scale[] = { - "./test_verif_scale1.o", "./test_verif_scale2.o", "./test_verif_scale3.o" + const char *sched_cls[] = { + "./test_verif_scale1.o", "./test_verif_scale2.o", "./test_verif_scale3.o", }; - const char *pyperf[] = { - "./pyperf50.o", "./pyperf100.o", "./pyperf180.o" + const char *raw_tp[] = { + /* full unroll by llvm */ + "./pyperf50.o", "./pyperf100.o", "./pyperf180.o", + + /* partial unroll. llvm will unroll loop ~150 times. + * C loop count -> 600. + * Asm loop count -> 4. + * 16k insns in loop body. + * Total of 5 such loops. Total program size ~82k insns. + */ + "./pyperf600.o", + + /* no unroll at all. + * C loop count -> 600. + * ASM loop count -> 600. + * ~110 insns in loop body. + * Total of 5 such loops. Total program size ~1500 insns. + */ + "./pyperf600_nounroll.o", + + "./loop1.o", "./loop2.o", + + /* partial unroll. 19k insn in a loop. + * Total program size 20.8k insn. + * ~350k processed_insns + */ + "./strobemeta.o", + + /* no unroll, tiny loops */ + "./strobemeta_nounroll1.o", + "./strobemeta_nounroll2.o", + }; + const char *cg_sysctl[] = { + "./test_sysctl_loop1.o", "./test_sysctl_loop2.o", }; int err, i; if (verifier_stats) libbpf_set_print(libbpf_debug_print_verifier_scale); - for (i = 0; i < ARRAY_SIZE(scale); i++) { - err = check_load(scale[i], BPF_PROG_TYPE_SCHED_CLS); - printf("test_scale:%s:%s\n", scale[i], err ? "FAIL" : "OK"); + err = check_load("./loop3.o", BPF_PROG_TYPE_RAW_TRACEPOINT); + printf("test_scale:loop3:%s\n", err ? (error_cnt--, "OK") : "FAIL"); + + for (i = 0; i < ARRAY_SIZE(sched_cls); i++) { + err = check_load(sched_cls[i], BPF_PROG_TYPE_SCHED_CLS); + printf("test_scale:%s:%s\n", sched_cls[i], err ? "FAIL" : "OK"); } - for (i = 0; i < ARRAY_SIZE(pyperf); i++) { - err = check_load(pyperf[i], BPF_PROG_TYPE_RAW_TRACEPOINT); - printf("test_scale:%s:%s\n", pyperf[i], err ? "FAIL" : "OK"); + for (i = 0; i < ARRAY_SIZE(raw_tp); i++) { + err = check_load(raw_tp[i], BPF_PROG_TYPE_RAW_TRACEPOINT); + printf("test_scale:%s:%s\n", raw_tp[i], err ? "FAIL" : "OK"); } + + for (i = 0; i < ARRAY_SIZE(cg_sysctl); i++) { + err = check_load(cg_sysctl[i], BPF_PROG_TYPE_CGROUP_SYSCTL); + printf("test_scale:%s:%s\n", cg_sysctl[i], err ? "FAIL" : "OK"); + } + err = check_load("./test_xdp_loop.o", BPF_PROG_TYPE_XDP); + printf("test_scale:test_xdp_loop:%s\n", err ? "FAIL" : "OK"); + + err = check_load("./test_seg6_loop.o", BPF_PROG_TYPE_LWT_SEG6LOCAL); + printf("test_scale:test_seg6_loop:%s\n", err ? "FAIL" : "OK"); } static void test_global_data_number(struct bpf_object *obj, __u32 duration) diff --git a/tools/testing/selftests/bpf/test_verifier.c b/tools/testing/selftests/bpf/test_verifier.c index 44497572ce9a7429e2914fee951ec7c35f70cdc8..35ffb64c981352af845947443002f32cd601f9fd 100644 --- a/tools/testing/selftests/bpf/test_verifier.c +++ b/tools/testing/selftests/bpf/test_verifier.c @@ -237,10 +237,10 @@ static void bpf_fill_scale1(struct bpf_test *self) insn[i++] = BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_6, -8 * (k % 64 + 1)); } - /* every jump adds 1 step to insn_processed, so to stay exactly - * within 1m limit add MAX_TEST_INSNS - MAX_JMP_SEQ - 1 MOVs and 1 EXIT + /* is_state_visited() doesn't allocate state for pruning for every jump. + * Hence multiply jmps by 4 to accommodate that heuristic */ - while (i < MAX_TEST_INSNS - MAX_JMP_SEQ - 1) + while (i < MAX_TEST_INSNS - MAX_JMP_SEQ * 4) insn[i++] = BPF_ALU64_IMM(BPF_MOV, BPF_REG_0, 42); insn[i] = BPF_EXIT_INSN(); self->prog_len = i + 1; @@ -269,10 +269,7 @@ static void bpf_fill_scale2(struct bpf_test *self) insn[i++] = BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_6, -8 * (k % (64 - 4 * FUNC_NEST) + 1)); } - /* every jump adds 1 step to insn_processed, so to stay exactly - * within 1m limit add MAX_TEST_INSNS - MAX_JMP_SEQ - 1 MOVs and 1 EXIT - */ - while (i < MAX_TEST_INSNS - MAX_JMP_SEQ - 1) + while (i < MAX_TEST_INSNS - MAX_JMP_SEQ * 4) insn[i++] = BPF_ALU64_IMM(BPF_MOV, BPF_REG_0, 42); insn[i] = BPF_EXIT_INSN(); self->prog_len = i + 1; @@ -947,7 +944,8 @@ static struct bpf_test tests[] = { BPF_JMP_IMM(BPF_JA, 0, 0, -1), BPF_EXIT_INSN(), }, - .errstr = "back-edge", + .errstr = "unreachable insn 1", + .errstr_unpriv = "back-edge", .result = REJECT, }, { @@ -959,20 +957,183 @@ static struct bpf_test tests[] = { BPF_JMP_IMM(BPF_JA, 0, 0, -4), BPF_EXIT_INSN(), }, - .errstr = "back-edge", + .errstr = "unreachable insn 4", + .errstr_unpriv = "back-edge", .result = REJECT, }, { "conditional loop", .insns = { - BPF_MOV64_REG(BPF_REG_1, BPF_REG_0), + BPF_MOV64_REG(BPF_REG_0, BPF_REG_1), BPF_MOV64_REG(BPF_REG_2, BPF_REG_0), BPF_MOV64_REG(BPF_REG_3, BPF_REG_0), BPF_JMP_IMM(BPF_JEQ, BPF_REG_1, 0, -3), BPF_EXIT_INSN(), }, + .errstr = "infinite loop detected", + .errstr_unpriv = "back-edge", + .result = REJECT, + }, + { + "bounded loop, count to 4", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 4, -2), + BPF_EXIT_INSN(), + }, + .result = ACCEPT, + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + .retval = 4, + }, + { + "bounded loop, count to 20", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 3), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 20, -2), + BPF_EXIT_INSN(), + }, + .result = ACCEPT, + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + }, + { + "bounded loop, count from positive unknown to 4", + .insns = { + BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0, BPF_FUNC_get_prandom_u32), + BPF_JMP_IMM(BPF_JSLT, BPF_REG_0, 0, 2), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 4, -2), + BPF_EXIT_INSN(), + }, + .result = ACCEPT, + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + .retval = 4, + }, + { + "bounded loop, count from totally unknown to 4", + .insns = { + BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0, BPF_FUNC_get_prandom_u32), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 4, -2), + BPF_EXIT_INSN(), + }, + .result = ACCEPT, + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + }, + { + "bounded loop, count to 4 with equality", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JNE, BPF_REG_0, 4, -2), + BPF_EXIT_INSN(), + }, + .result = ACCEPT, + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + }, + { + "bounded loop, start in the middle", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_JMP_A(1), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 4, -2), + BPF_EXIT_INSN(), + }, + .result = REJECT, + .errstr = "back-edge", + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + .retval = 4, + }, + { + "bounded loop containing a forward jump", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_JMP_REG(BPF_JEQ, BPF_REG_0, BPF_REG_0, 0), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 4, -3), + BPF_EXIT_INSN(), + }, + .result = ACCEPT, + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + .retval = 4, + }, + { + "bounded loop that jumps out rather than in", + .insns = { + BPF_MOV64_IMM(BPF_REG_6, 0), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_6, 1), + BPF_JMP_IMM(BPF_JGT, BPF_REG_6, 10000, 2), + BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 0, 0, BPF_FUNC_get_prandom_u32), + BPF_JMP_A(-4), + BPF_EXIT_INSN(), + }, + .result = ACCEPT, + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + }, + { + "infinite loop after a conditional jump", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 5), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 4, 2), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_JMP_A(-2), + BPF_EXIT_INSN(), + }, + .result = REJECT, + .errstr = "program is too large", + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + }, + { + "bounded recursion", + .insns = { + BPF_MOV64_IMM(BPF_REG_1, 0), + BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 1, 0, 1), + BPF_EXIT_INSN(), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, 1), + BPF_MOV64_REG(BPF_REG_0, BPF_REG_1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_1, 4, 1), + BPF_EXIT_INSN(), + BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 1, 0, -5), + BPF_EXIT_INSN(), + }, + .result = REJECT, .errstr = "back-edge", + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + }, + { + "infinite loop in two jumps", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_JMP_A(0), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 4, -2), + BPF_EXIT_INSN(), + }, + .result = REJECT, + .errstr = "loop detected", + .prog_type = BPF_PROG_TYPE_TRACEPOINT, + }, + { + "infinite loop: three-jump trick", + .insns = { + BPF_MOV64_IMM(BPF_REG_0, 0), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_ALU64_IMM(BPF_AND, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 2, 1), + BPF_EXIT_INSN(), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_ALU64_IMM(BPF_AND, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 2, 1), + BPF_EXIT_INSN(), + BPF_ALU64_IMM(BPF_ADD, BPF_REG_0, 1), + BPF_ALU64_IMM(BPF_AND, BPF_REG_0, 1), + BPF_JMP_IMM(BPF_JLT, BPF_REG_0, 2, -11), + BPF_EXIT_INSN(), + }, .result = REJECT, + .errstr = "loop detected", + .prog_type = BPF_PROG_TYPE_TRACEPOINT, }, { "read uninitialized register", @@ -4205,7 +4366,8 @@ static struct bpf_test tests[] = { offsetof(struct __sk_buff, data)), BPF_LDX_MEM(BPF_W, BPF_REG_3, BPF_REG_1, offsetof(struct __sk_buff, data_end)), - BPF_MOV64_IMM(BPF_REG_0, 0xffffffff), + BPF_LDX_MEM(BPF_W, BPF_REG_0, BPF_REG_1, + offsetof(struct __sk_buff, mark)), BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_0, -8), BPF_LDX_MEM(BPF_DW, BPF_REG_0, BPF_REG_10, -8), BPF_ALU64_IMM(BPF_AND, BPF_REG_0, 0xffff), @@ -7451,9 +7613,9 @@ static struct bpf_test tests[] = { { "helper access to variable memory: stack, bitwise AND, zero included", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), - BPF_MOV64_IMM(BPF_REG_2, 16), BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, -128), BPF_ALU64_IMM(BPF_AND, BPF_REG_2, 64), @@ -7468,9 +7630,9 @@ static struct bpf_test tests[] = { { "helper access to variable memory: stack, bitwise AND + JMP, wrong max", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), - BPF_MOV64_IMM(BPF_REG_2, 16), BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, -128), BPF_ALU64_IMM(BPF_AND, BPF_REG_2, 65), @@ -7544,9 +7706,9 @@ static struct bpf_test tests[] = { { "helper access to variable memory: stack, JMP, bounds + offset", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), - BPF_MOV64_IMM(BPF_REG_2, 16), BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, -128), BPF_JMP_IMM(BPF_JGT, BPF_REG_2, 64, 5), @@ -7565,9 +7727,9 @@ static struct bpf_test tests[] = { { "helper access to variable memory: stack, JMP, wrong max", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), - BPF_MOV64_IMM(BPF_REG_2, 16), BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, -128), BPF_JMP_IMM(BPF_JGT, BPF_REG_2, 65, 4), @@ -7585,9 +7747,9 @@ static struct bpf_test tests[] = { { "helper access to variable memory: stack, JMP, no max check", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), - BPF_MOV64_IMM(BPF_REG_2, 16), BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, -128), BPF_MOV64_IMM(BPF_REG_4, 0), @@ -7605,9 +7767,9 @@ static struct bpf_test tests[] = { { "helper access to variable memory: stack, JMP, no min check", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), - BPF_MOV64_IMM(BPF_REG_2, 16), BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, -128), BPF_JMP_IMM(BPF_JGT, BPF_REG_2, 64, 3), @@ -7623,9 +7785,9 @@ static struct bpf_test tests[] = { { "helper access to variable memory: stack, JMP (signed), no min check", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), - BPF_MOV64_IMM(BPF_REG_2, 16), BPF_STX_MEM(BPF_DW, BPF_REG_1, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, -128), BPF_JMP_IMM(BPF_JSGT, BPF_REG_2, 64, 3), @@ -7667,6 +7829,7 @@ static struct bpf_test tests[] = { { "helper access to variable memory: map, JMP, wrong max", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_6, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_2, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_2, -8), BPF_ST_MEM(BPF_DW, BPF_REG_2, 0, 0), @@ -7674,7 +7837,7 @@ static struct bpf_test tests[] = { BPF_EMIT_CALL(BPF_FUNC_map_lookup_elem), BPF_JMP_IMM(BPF_JEQ, BPF_REG_0, 0, 10), BPF_MOV64_REG(BPF_REG_1, BPF_REG_0), - BPF_MOV64_IMM(BPF_REG_2, sizeof(struct test_val)), + BPF_MOV64_REG(BPF_REG_2, BPF_REG_6), BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_10, -128), BPF_JMP_IMM(BPF_JSGT, BPF_REG_2, @@ -7686,7 +7849,7 @@ static struct bpf_test tests[] = { BPF_MOV64_IMM(BPF_REG_0, 0), BPF_EXIT_INSN(), }, - .fixup_map_hash_48b = { 3 }, + .fixup_map_hash_48b = { 4 }, .errstr = "invalid access to map value, value_size=48 off=0 size=49", .result = REJECT, .prog_type = BPF_PROG_TYPE_TRACEPOINT, @@ -7721,6 +7884,7 @@ static struct bpf_test tests[] = { { "helper access to variable memory: map adjusted, JMP, wrong max", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_6, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_2, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_2, -8), BPF_ST_MEM(BPF_DW, BPF_REG_2, 0, 0), @@ -7729,7 +7893,7 @@ static struct bpf_test tests[] = { BPF_JMP_IMM(BPF_JEQ, BPF_REG_0, 0, 11), BPF_MOV64_REG(BPF_REG_1, BPF_REG_0), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, 20), - BPF_MOV64_IMM(BPF_REG_2, sizeof(struct test_val)), + BPF_MOV64_REG(BPF_REG_2, BPF_REG_6), BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_10, -128), BPF_JMP_IMM(BPF_JSGT, BPF_REG_2, @@ -7741,7 +7905,7 @@ static struct bpf_test tests[] = { BPF_MOV64_IMM(BPF_REG_0, 0), BPF_EXIT_INSN(), }, - .fixup_map_hash_48b = { 3 }, + .fixup_map_hash_48b = { 4 }, .errstr = "R1 min value is outside of the array range", .result = REJECT, .prog_type = BPF_PROG_TYPE_TRACEPOINT, @@ -7763,8 +7927,8 @@ static struct bpf_test tests[] = { { "helper access to variable memory: size > 0 not allowed on NULL (ARG_PTR_TO_MEM_OR_NULL)", .insns = { + BPF_LDX_MEM(BPF_W, BPF_REG_2, BPF_REG_1, 0), BPF_MOV64_IMM(BPF_REG_1, 0), - BPF_MOV64_IMM(BPF_REG_2, 1), BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_10, -128), BPF_ALU64_IMM(BPF_AND, BPF_REG_2, 64), @@ -7991,6 +8155,7 @@ static struct bpf_test tests[] = { { "helper access to variable memory: 8 bytes leak", .insns = { + BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_1, 8), BPF_MOV64_REG(BPF_REG_1, BPF_REG_10), BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, -64), BPF_MOV64_IMM(BPF_REG_0, 0), @@ -8001,7 +8166,6 @@ static struct bpf_test tests[] = { BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_0, -24), BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_0, -16), BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_0, -8), - BPF_MOV64_IMM(BPF_REG_2, 1), BPF_STX_MEM(BPF_DW, BPF_REG_10, BPF_REG_2, -128), BPF_LDX_MEM(BPF_DW, BPF_REG_2, BPF_REG_10, -128), BPF_ALU64_IMM(BPF_AND, BPF_REG_2, 63), @@ -11358,9 +11522,11 @@ static struct bpf_test tests[] = { BPF_MOV64_IMM(BPF_REG_0, 3), BPF_JMP_IMM(BPF_JA, 0, 0, -6), }, - .prog_type = BPF_PROG_TYPE_TRACEPOINT, - .errstr = "back-edge from insn", - .result = REJECT, + .prog_type = BPF_PROG_TYPE_SOCKET_FILTER, + .errstr_unpriv = "back-edge from insn", + .result_unpriv = REJECT, + .result = ACCEPT, + .retval = 1, }, { "calls: conditional call 4", @@ -11393,22 +11559,24 @@ static struct bpf_test tests[] = { BPF_MOV64_IMM(BPF_REG_0, 3), BPF_EXIT_INSN(), }, - .prog_type = BPF_PROG_TYPE_TRACEPOINT, - .errstr = "back-edge from insn", - .result = REJECT, + .prog_type = BPF_PROG_TYPE_SCHED_CLS, + .result = ACCEPT, + .retval = 1, }, { "calls: conditional call 6", .insns = { + BPF_MOV64_REG(BPF_REG_6, BPF_REG_1), + BPF_MOV64_REG(BPF_REG_1, BPF_REG_6), BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, 1, 0, 2), - BPF_JMP_IMM(BPF_JEQ, BPF_REG_0, 0, -2), + BPF_JMP_IMM(BPF_JEQ, BPF_REG_0, 0, -3), BPF_EXIT_INSN(), BPF_LDX_MEM(BPF_W, BPF_REG_0, BPF_REG_1, offsetof(struct __sk_buff, mark)), BPF_EXIT_INSN(), }, - .prog_type = BPF_PROG_TYPE_TRACEPOINT, - .errstr = "back-edge from insn", + .prog_type = BPF_PROG_TYPE_SCHED_CLS, + .errstr = "infinite loop detected", .result = REJECT, }, {