From patchwork Wed Apr 10 19:50:21 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jiong Wang X-Patchwork-Id: 1083594 X-Patchwork-Delegate: bpf@iogearbox.net Return-Path: X-Original-To: patchwork-incoming-netdev@ozlabs.org Delivered-To: patchwork-incoming-netdev@ozlabs.org Authentication-Results: ozlabs.org; spf=none (mailfrom) smtp.mailfrom=vger.kernel.org (client-ip=209.132.180.67; helo=vger.kernel.org; envelope-from=netdev-owner@vger.kernel.org; receiver=) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=netronome.com Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=netronome-com.20150623.gappssmtp.com header.i=@netronome-com.20150623.gappssmtp.com header.b="UAMKuXC/"; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 44fZXx3S9Wz9s5c for ; Thu, 11 Apr 2019 05:51:29 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726886AbfDJTv2 (ORCPT ); Wed, 10 Apr 2019 15:51:28 -0400 Received: from mail-wr1-f67.google.com ([209.85.221.67]:37140 "EHLO mail-wr1-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726448AbfDJTut (ORCPT ); Wed, 10 Apr 2019 15:50:49 -0400 Received: by mail-wr1-f67.google.com with SMTP id w10so4368670wrm.4 for ; Wed, 10 Apr 2019 12:50:47 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=netronome-com.20150623.gappssmtp.com; s=20150623; h=from:to:cc:subject:date:message-id:in-reply-to:references; bh=NlPoGKLapOA094jH5RaXWl2m1CpdKBwMGl/CNr+Rkc8=; b=UAMKuXC/brrvhsXqVdvgzuv0kznGyJTltWHHwL5pbQhUF/vIIteF1aVA1aecPdTVdI 0wYOh/aiSSsHB3/jdCWP0hQs/UaJuBES81Da97qxeiAFww3+a53/JyoY9ACayggVAanG ZGg1Hkod17cYWh8alVmb51BXoBPMVm5QQ9zcajlqg6qIj3OWb/pKx89pRPAs9574Z/5/ V/rlItavvbYUz+DxI+QlrzGub1GH3isc/NXPANDU33QKVvqyFn3kQQsoDM+cMF+38UuJ gJ01hcJk5Iy9csKQpuRRfu2DiPlueSG+dscm+N+GBPwNdzIwQbU2jbo5s13txLg+U1gA Alpw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references; bh=NlPoGKLapOA094jH5RaXWl2m1CpdKBwMGl/CNr+Rkc8=; b=mKt1a3Auhcg2OVM+Lawfo3iAZtnNuJFGztHTa+VgxagXWwjDYD//8mINFRIm6V0SuG e2IBh7ZlF1c8DAz3qEIe3j9RpZGt1apeVfVA0SSEEUoFyqY/85GrLU8EZ+pTG+KQuwJp STYiEgFnZqYtUN5cuU/iKVLjB5JvNk2gVps5pgL4T1vOzmLNsxyGl74sgtGWNDix8PQ2 axVFhaR2Xp3qsYtfGkQW1NOekw3pJDtSrwTiMBuNmsxrnQ6eMHIHft68bSfF/o6gGJOD h5aLWzofQfvQR4zLpO+Ni/2N8MZ7Kcjj9okIqFLJwf36+v0TXBBOwtVnXGiRILL3Iq5j cdSA== X-Gm-Message-State: APjAAAVOgE90egE4LMUEiQfr7qnP7ZCrGPennJqpbMxWM032dhdmBCWR 5qpFgyZvrW2LrtD1Vgla/VbJyQ== X-Google-Smtp-Source: APXvYqznRfgJgsqo1f54xhUgbQU3y7miXhQnX1088CH1D4cHIF8dr50R2Pi/76zrmp7nAH7pxfayhw== X-Received: by 2002:a5d:428c:: with SMTP id k12mr13003060wrq.279.1554925846818; Wed, 10 Apr 2019 12:50:46 -0700 (PDT) Received: from cbtest28.netronome.com ([217.38.71.146]) by smtp.gmail.com with ESMTPSA id 4sm2288540wmg.12.2019.04.10.12.50.45 (version=TLS1_2 cipher=ECDHE-RSA-AES128-SHA bits=128/128); Wed, 10 Apr 2019 12:50:46 -0700 (PDT) From: Jiong Wang To: alexei.starovoitov@gmail.com, daniel@iogearbox.net Cc: bpf@vger.kernel.org, netdev@vger.kernel.org, oss-drivers@netronome.com, Jiong Wang Subject: [PATCH/RFC v2 bpf-next 07/19] bpf: reduce false alarm by refining helper call arg types Date: Wed, 10 Apr 2019 20:50:21 +0100 Message-Id: <1554925833-7333-8-git-send-email-jiong.wang@netronome.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1554925833-7333-1-git-send-email-jiong.wang@netronome.com> References: <1554925833-7333-1-git-send-email-jiong.wang@netronome.com> Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org Unlike BPF to BPF function call, BPF helper call is calls to native insns that verifier can't walk. So, verifier needs helper proto type descriptions for data-flow purpose. There is such information already, but it is not differentiate sub-register read with full register read. This patch split "enum bpf_arg_type" for sub-register read, and updated descriptions for several functions that shown frequent usage in one Cilium benchmark. "is_reg64" then taught about these new arg types. Reviewed-by: Jakub Kicinski Signed-off-by: Jiong Wang --- include/linux/bpf.h | 3 +++ kernel/bpf/core.c | 2 +- kernel/bpf/helpers.c | 2 +- kernel/bpf/verifier.c | 61 +++++++++++++++++++++++++++++++++++++++++++-------- net/core/filter.c | 28 +++++++++++------------ 5 files changed, 71 insertions(+), 25 deletions(-) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 65f7094..42f3345 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -197,9 +197,12 @@ enum bpf_arg_type { ARG_CONST_SIZE, /* number of bytes accessed from memory */ ARG_CONST_SIZE_OR_ZERO, /* number of bytes accessed from memory or 0 */ + ARG_CONST_SIZE32, /* Likewise, but size fits into 32-bit */ + ARG_CONST_SIZE32_OR_ZERO, /* Ditto */ ARG_PTR_TO_CTX, /* pointer to context */ ARG_ANYTHING, /* any (initialized) argument is ok */ + ARG_ANYTHING32, /* Likewise, but it is a 32-bit argument */ ARG_PTR_TO_SPIN_LOCK, /* pointer to bpf_spin_lock */ ARG_PTR_TO_SOCK_COMMON, /* pointer to sock_common */ }; diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c index ace8c22..2792eda 100644 --- a/kernel/bpf/core.c +++ b/kernel/bpf/core.c @@ -2067,7 +2067,7 @@ const struct bpf_func_proto bpf_tail_call_proto = { .ret_type = RET_VOID, .arg1_type = ARG_PTR_TO_CTX, .arg2_type = ARG_CONST_MAP_PTR, - .arg3_type = ARG_ANYTHING, + .arg3_type = ARG_ANYTHING32, }; /* Stub for JITs that only support cBPF. eBPF programs are interpreted. diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c index a411fc1..6b7453e 100644 --- a/kernel/bpf/helpers.c +++ b/kernel/bpf/helpers.c @@ -218,7 +218,7 @@ const struct bpf_func_proto bpf_get_current_comm_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_UNINIT_MEM, - .arg2_type = ARG_CONST_SIZE, + .arg2_type = ARG_CONST_SIZE32, }; #if defined(CONFIG_QUEUED_SPINLOCKS) || defined(CONFIG_BPF_ARCH_SPINLOCK) diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index b478b90..431fca9 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -1174,12 +1174,47 @@ static int mark_reg_read(struct bpf_verifier_env *env, return 0; } +static bool helper_call_arg64(struct bpf_verifier_env *env, int func_id, + u32 regno) +{ + /* get_func_proto must succeed, other it should have been rejected + * early inside check_helper_call. + */ + const struct bpf_func_proto *fn = + env->ops->get_func_proto(func_id, env->prog); + enum bpf_arg_type arg_type; + + switch (regno) { + case BPF_REG_1: + arg_type = fn->arg1_type; + break; + case BPF_REG_2: + arg_type = fn->arg2_type; + break; + case BPF_REG_3: + arg_type = fn->arg3_type; + break; + case BPF_REG_4: + arg_type = fn->arg4_type; + break; + case BPF_REG_5: + arg_type = fn->arg5_type; + break; + default: + arg_type = ARG_DONTCARE; + } + + return arg_type != ARG_CONST_SIZE32 && + arg_type != ARG_CONST_SIZE32_OR_ZERO && + arg_type != ARG_ANYTHING32; +} + /* This function is supposed to be used by the following 32-bit optimization * code only. It returns TRUE if the source or destination register operates * on 64-bit, otherwise return FALSE. */ -static bool is_reg64(struct bpf_insn *insn, u32 regno, - struct bpf_reg_state *reg, enum reg_arg_type t) +static bool is_reg64(struct bpf_verifier_env *env, struct bpf_insn *insn, + u32 regno, struct bpf_reg_state *reg, enum reg_arg_type t) { u8 code, class, op; @@ -1201,9 +1236,12 @@ static bool is_reg64(struct bpf_insn *insn, u32 regno, if (insn->src_reg == BPF_PSEUDO_CALL) return false; /* Helper call will reach here because of arg type - * check. Conservatively marking all args as 64-bit. + * check. */ - return true; + if (t == SRC_OP) + return helper_call_arg64(env, insn->imm, regno); + + return false; } } @@ -1283,7 +1321,7 @@ static int check_reg_arg(struct bpf_verifier_env *env, u32 regno, } reg = ®s[regno]; - rw64 = is_reg64(insn, regno, reg, t); + rw64 = is_reg64(env, insn, regno, reg, t); if (t == SRC_OP) { /* check whether register used as source operand can be read */ if (reg->type == NOT_INIT) { @@ -2575,7 +2613,9 @@ static bool arg_type_is_mem_ptr(enum bpf_arg_type type) static bool arg_type_is_mem_size(enum bpf_arg_type type) { return type == ARG_CONST_SIZE || - type == ARG_CONST_SIZE_OR_ZERO; + type == ARG_CONST_SIZE_OR_ZERO || + type == ARG_CONST_SIZE32 || + type == ARG_CONST_SIZE32_OR_ZERO; } static int check_func_arg(struct bpf_verifier_env *env, u32 regno, @@ -2593,7 +2633,7 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 regno, if (err) return err; - if (arg_type == ARG_ANYTHING) { + if (arg_type == ARG_ANYTHING || arg_type == ARG_ANYTHING32) { if (is_pointer_value(env, regno)) { verbose(env, "R%d leaks addr into helper function\n", regno); @@ -2616,7 +2656,9 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 regno, type != expected_type) goto err_type; } else if (arg_type == ARG_CONST_SIZE || - arg_type == ARG_CONST_SIZE_OR_ZERO) { + arg_type == ARG_CONST_SIZE_OR_ZERO || + arg_type == ARG_CONST_SIZE32 || + arg_type == ARG_CONST_SIZE32_OR_ZERO) { expected_type = SCALAR_VALUE; if (type != expected_type) goto err_type; @@ -2710,7 +2752,8 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 regno, meta->map_ptr->value_size, false, meta); } else if (arg_type_is_mem_size(arg_type)) { - bool zero_size_allowed = (arg_type == ARG_CONST_SIZE_OR_ZERO); + bool zero_size_allowed = (arg_type == ARG_CONST_SIZE_OR_ZERO || + arg_type == ARG_CONST_SIZE32_OR_ZERO); /* remember the mem_size which may be used later * to refine return values. diff --git a/net/core/filter.c b/net/core/filter.c index 22eb2ed..3f6d8af 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -1693,9 +1693,9 @@ static const struct bpf_func_proto bpf_skb_store_bytes_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, - .arg2_type = ARG_ANYTHING, + .arg2_type = ARG_ANYTHING32, .arg3_type = ARG_PTR_TO_MEM, - .arg4_type = ARG_CONST_SIZE, + .arg4_type = ARG_CONST_SIZE32, .arg5_type = ARG_ANYTHING, }; @@ -1724,9 +1724,9 @@ static const struct bpf_func_proto bpf_skb_load_bytes_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, - .arg2_type = ARG_ANYTHING, + .arg2_type = ARG_ANYTHING32, .arg3_type = ARG_PTR_TO_UNINIT_MEM, - .arg4_type = ARG_CONST_SIZE, + .arg4_type = ARG_CONST_SIZE32, }; BPF_CALL_5(bpf_skb_load_bytes_relative, const struct sk_buff *, skb, @@ -1875,7 +1875,7 @@ static const struct bpf_func_proto bpf_l3_csum_replace_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, - .arg2_type = ARG_ANYTHING, + .arg2_type = ARG_ANYTHING32, .arg3_type = ARG_ANYTHING, .arg4_type = ARG_ANYTHING, .arg5_type = ARG_ANYTHING, @@ -1928,7 +1928,7 @@ static const struct bpf_func_proto bpf_l4_csum_replace_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, - .arg2_type = ARG_ANYTHING, + .arg2_type = ARG_ANYTHING32, .arg3_type = ARG_ANYTHING, .arg4_type = ARG_ANYTHING, .arg5_type = ARG_ANYTHING, @@ -1967,9 +1967,9 @@ static const struct bpf_func_proto bpf_csum_diff_proto = { .pkt_access = true, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_MEM_OR_NULL, - .arg2_type = ARG_CONST_SIZE_OR_ZERO, + .arg2_type = ARG_CONST_SIZE32_OR_ZERO, .arg3_type = ARG_PTR_TO_MEM_OR_NULL, - .arg4_type = ARG_CONST_SIZE_OR_ZERO, + .arg4_type = ARG_CONST_SIZE32_OR_ZERO, .arg5_type = ARG_ANYTHING, }; @@ -2150,7 +2150,7 @@ static const struct bpf_func_proto bpf_redirect_proto = { .func = bpf_redirect, .gpl_only = false, .ret_type = RET_INTEGER, - .arg1_type = ARG_ANYTHING, + .arg1_type = ARG_ANYTHING32, .arg2_type = ARG_ANYTHING, }; @@ -2928,7 +2928,7 @@ static const struct bpf_func_proto bpf_skb_change_proto_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, - .arg2_type = ARG_ANYTHING, + .arg2_type = ARG_ANYTHING32, .arg3_type = ARG_ANYTHING, }; @@ -2948,7 +2948,7 @@ static const struct bpf_func_proto bpf_skb_change_type_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, - .arg2_type = ARG_ANYTHING, + .arg2_type = ARG_ANYTHING32, }; static u32 bpf_skb_net_base_len(const struct sk_buff *skb) @@ -3236,7 +3236,7 @@ static const struct bpf_func_proto bpf_skb_change_tail_proto = { .gpl_only = false, .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, - .arg2_type = ARG_ANYTHING, + .arg2_type = ARG_ANYTHING32, .arg3_type = ARG_ANYTHING, }; @@ -3832,7 +3832,7 @@ static const struct bpf_func_proto bpf_skb_get_tunnel_key_proto = { .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, .arg2_type = ARG_PTR_TO_UNINIT_MEM, - .arg3_type = ARG_CONST_SIZE, + .arg3_type = ARG_CONST_SIZE32, .arg4_type = ARG_ANYTHING, }; @@ -3941,7 +3941,7 @@ static const struct bpf_func_proto bpf_skb_set_tunnel_key_proto = { .ret_type = RET_INTEGER, .arg1_type = ARG_PTR_TO_CTX, .arg2_type = ARG_PTR_TO_MEM, - .arg3_type = ARG_CONST_SIZE, + .arg3_type = ARG_CONST_SIZE32, .arg4_type = ARG_ANYTHING, };