From patchwork Mon Apr 16 19:33:18 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Martin KaFai Lau X-Patchwork-Id: 898877 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=pass (p=none dis=none) header.from=fb.com Authentication-Results: ozlabs.org; dkim=pass (1024-bit key; unprotected) header.d=fb.com header.i=@fb.com header.b="BV8jqtD5"; dkim-atps=neutral Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 40Pz8K5QQrz9s0W for ; Tue, 17 Apr 2018 05:33:53 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753405AbeDPTdb (ORCPT ); Mon, 16 Apr 2018 15:33:31 -0400 Received: from mx0a-00082601.pphosted.com ([67.231.145.42]:48674 "EHLO mx0a-00082601.pphosted.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752934AbeDPTd2 (ORCPT ); Mon, 16 Apr 2018 15:33:28 -0400 Received: from pps.filterd (m0044010.ppops.net [127.0.0.1]) by mx0a-00082601.pphosted.com (8.16.0.22/8.16.0.22) with SMTP id w3GJTuRm009040 for ; Mon, 16 Apr 2018 12:33:28 -0700 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=fb.com; h=from : to : cc : subject : date : message-id : in-reply-to : references : mime-version : content-type; s=facebook; bh=iuOZSKJ7i0T4t5pmwcgREGqziv0oZpRzbzcL6/DmEgk=; b=BV8jqtD5hYQ8rC3UkfYdITPeWD2JA8g/ShoNTUzSo760DqKASDw4WNrQasYq3/xYG+b4 aBdOssyD4X/m4HqiG2T4/LdffpGlUv0vkewLc4Z3mNG2GDfeCJNHAqM01qzyrMB007jL 0HnzqPomcHUZd172tJTDjzWyswcBmRevIys= Received: from mail.thefacebook.com ([199.201.64.23]) by mx0a-00082601.pphosted.com with ESMTP id 2hd1nqg2ke-2 (version=TLSv1 cipher=ECDHE-RSA-AES256-SHA bits=256 verify=NOT) for ; Mon, 16 Apr 2018 12:33:28 -0700 Received: from mx-out.facebook.com (192.168.52.123) by PRN-CHUB01.TheFacebook.com (192.168.16.11) with Microsoft SMTP Server id 14.3.361.1; Mon, 16 Apr 2018 12:33:27 -0700 Received: by devbig738.prn1.facebook.com (Postfix, from userid 6611) id 598FA450065E; Mon, 16 Apr 2018 12:33:27 -0700 (PDT) Smtp-Origin-Hostprefix: devbig From: Martin KaFai Lau Smtp-Origin-Hostname: devbig738.prn1.facebook.com To: CC: Alexei Starovoitov , Daniel Borkmann , Smtp-Origin-Cluster: prn1c29 Subject: [PATCH bpf-next v3 01/10] bpf: btf: Introduce BPF Type Format (BTF) Date: Mon, 16 Apr 2018 12:33:18 -0700 Message-ID: <20180416193327.477239-2-kafai@fb.com> X-Mailer: git-send-email 2.9.5 In-Reply-To: <20180416193327.477239-1-kafai@fb.com> References: <20180416193327.477239-1-kafai@fb.com> X-FB-Internal: Safe MIME-Version: 1.0 X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:, , definitions=2018-04-16_10:, , signatures=0 X-Proofpoint-Spam-Reason: safe X-FB-Internal: Safe Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org This patch introduces BPF type Format (BTF). BTF (BPF Type Format) is the meta data format which describes the data types of BPF program/map. Hence, it basically focus on the C programming language which the modern BPF is primary using. The first use case is to provide a generic pretty print capability for a BPF map. BTF has its root from CTF (Compact C-Type format). To simplify the handling of BTF data, BTF removes the differences between small and big type/struct-member. Hence, BTF consistently uses u32 instead of supporting both "one u16" and "two u32 (+padding)" in describing type and struct-member. It also raises the number of types (and functions) limit from 0x7fff to 0x7fffffff. Due to the above changes, the format is not compatible to CTF. Hence, BTF starts with a new BTF_MAGIC and version number. This patch does the first verification pass to the BTF. The first pass checks: 1. meta-data size (e.g. It does not go beyond the total btf's size) 2. name_offset is valid 3. Each BTF_KIND (e.g. int, enum, struct....) does its own check of its meta-data. Some other checks, like checking a struct's member is referring to a valid type, can only be done in the second pass. The second verification pass will be implemented in the next patch. Signed-off-by: Martin KaFai Lau Acked-by: Alexei Starovoitov --- include/uapi/linux/btf.h | 132 +++++++ kernel/bpf/Makefile | 1 + kernel/bpf/btf.c | 940 +++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 1073 insertions(+) create mode 100644 include/uapi/linux/btf.h create mode 100644 kernel/bpf/btf.c diff --git a/include/uapi/linux/btf.h b/include/uapi/linux/btf.h new file mode 100644 index 000000000000..30f23ad78eff --- /dev/null +++ b/include/uapi/linux/btf.h @@ -0,0 +1,132 @@ +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ +/* Copyright (c) 2018 Facebook */ +#ifndef _UAPI__LINUX_BTF_H__ +#define _UAPI__LINUX_BTF_H__ + +#include + +#define BTF_MAGIC 0xeB9F +#define BTF_MAGIC_SWAP 0x9FeB +#define BTF_VERSION 1 +#define BTF_FLAGS_COMPR 0x01 + +struct btf_header { + __u16 magic; + __u8 version; + __u8 flags; + + __u32 parent_label; + __u32 parent_name; + + /* All offsets are in bytes relative to the end of this header */ + __u32 label_off; /* offset of label section */ + __u32 object_off; /* offset of data object section*/ + __u32 func_off; /* offset of function section */ + __u32 type_off; /* offset of type section */ + __u32 str_off; /* offset of string section */ + __u32 str_len; /* length of string section */ +}; + +/* Max # of type identifier */ +#define BTF_MAX_TYPE 0x7fffffff +/* Max offset into the string section */ +#define BTF_MAX_NAME_OFFSET 0x7fffffff +/* Max # of struct/union/enum members or func args */ +#define BTF_MAX_VLEN 0xffff + +/* The type id is referring to a parent BTF */ +#define BTF_TYPE_PARENT(id) (((id) >> 31) & 0x1) +#define BTF_TYPE_ID(id) ((id) & BTF_MAX_TYPE) + +/* String is in the ELF string section */ +#define BTF_STR_TBL_ELF_ID(ref) (((ref) >> 31) & 0x1) +#define BTF_STR_OFFSET(ref) ((ref) & BTF_MAX_NAME_OFFSET) + +struct btf_type { + __u32 name; + /* "info" bits arrangement + * bits 0-15: vlen (e.g. # of struct's members) + * bits 16-23: unused + * bits 24-28: kind (e.g. int, ptr, array...etc) + * bits 29-30: unused + * bits 31: root + */ + __u32 info; + /* "size" is used by INT, ENUM, STRUCT and UNION. + * "size" tells the size of the type it is describing. + * + * "type" is used by PTR, TYPEDEF, VOLATILE, CONST and RESTRICT. + * "type" is a type_id referring to another type. + */ + union { + __u32 size; + __u32 type; + }; +}; + +#define BTF_INFO_KIND(info) (((info) >> 24) & 0x1f) +#define BTF_INFO_ISROOT(info) (!!(((info) >> 24) & 0x80)) +#define BTF_INFO_VLEN(info) ((info) & 0xffff) + +#define BTF_KIND_UNKN 0 /* Unknown */ +#define BTF_KIND_INT 1 /* Integer */ +#define BTF_KIND_FLOAT 2 /* Float */ +#define BTF_KIND_PTR 3 /* Pointer */ +#define BTF_KIND_ARRAY 4 /* Array */ +#define BTF_KIND_FUNC 5 /* Function */ +#define BTF_KIND_STRUCT 6 /* Struct */ +#define BTF_KIND_UNION 7 /* Union */ +#define BTF_KIND_ENUM 8 /* Enumeration */ +#define BTF_KIND_FWD 9 /* Forward */ +#define BTF_KIND_TYPEDEF 10 /* Typedef */ +#define BTF_KIND_VOLATILE 11 /* Volatile */ +#define BTF_KIND_CONST 12 /* Const */ +#define BTF_KIND_RESTRICT 13 /* Restrict */ +#define BTF_KIND_MAX 13 +#define NR_BTF_KINDS 14 + +/* For some specific BTF_KIND, "struct btf_type" is immediately + * followed by extra data. + */ + +/* BTF_KIND_INT is followed by a u32 and the following + * is the 32 bits arrangement: + */ +#define BTF_INT_ENCODING(VAL) (((VAL) & 0xff000000) >> 24) +#define BTF_INT_OFFSET(VAL) (((VAL & 0x00ff0000)) >> 16) +#define BTF_INT_BITS(VAL) ((VAL) & 0x0000ffff) + +/* Attributes stored in the BTF_INT_ENCODING */ +#define BTF_INT_SIGNED 0x1 +#define BTF_INT_CHAR 0x2 +#define BTF_INT_BOOL 0x4 +#define BTF_INT_VARARGS 0x8 + +/* BTF_KIND_ENUM is followed by multiple "struct btf_enum". + * The exact number of btf_enum is stored in the vlen (of the + * info in "struct btf_type"). + */ +struct btf_enum { + __u32 name; + __s32 val; +}; + +/* BTF_KIND_ARRAY is followed by one "struct btf_array" */ +struct btf_array { + __u32 type; + __u32 index_type; + __u32 nelems; +}; + +/* BTF_KIND_STRUCT and BTF_KIND_UNION are followed + * by multiple "struct btf_member". The exact number + * of btf_member is stored in the vlen (of the info in + * "struct btf_type"). + */ +struct btf_member { + __u32 name; + __u32 type; + __u32 offset; /* offset in bits */ +}; + +#endif /* _UAPI__LINUX_BTF_H__ */ diff --git a/kernel/bpf/Makefile b/kernel/bpf/Makefile index a713fd23ec88..35c485fa9ea3 100644 --- a/kernel/bpf/Makefile +++ b/kernel/bpf/Makefile @@ -4,6 +4,7 @@ obj-y := core.o obj-$(CONFIG_BPF_SYSCALL) += syscall.o verifier.o inode.o helpers.o tnum.o obj-$(CONFIG_BPF_SYSCALL) += hashtab.o arraymap.o percpu_freelist.o bpf_lru_list.o lpm_trie.o map_in_map.o obj-$(CONFIG_BPF_SYSCALL) += disasm.o +obj-$(CONFIG_BPF_SYSCALL) += btf.o ifeq ($(CONFIG_NET),y) obj-$(CONFIG_BPF_SYSCALL) += devmap.o obj-$(CONFIG_BPF_SYSCALL) += cpumap.o diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c new file mode 100644 index 000000000000..905be2f48a35 --- /dev/null +++ b/kernel/bpf/btf.c @@ -0,0 +1,940 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* Copyright (c) 2018 Facebook */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +/* BTF (BPF Type Format) is the meta data format which describes + * the data types of BPF program/map. Hence, it basically focus + * on the C programming language which the modern BPF is primary + * using. + * + * ELF Section: + * ~~~~~~~~~~~ + * The BTF data is stored under the ".BTF" ELF section + * + * struct btf_type: + * ~~~~~~~~~~~~~~~ + * Each 'struct btf_type' object describes a C data type. + * Depending on the type it is describing, a 'struct btf_type' + * object may be followed by more data. F.e. + * To describe an array, 'struct btf_type' is followed by + * 'struct btf_array'. + * + * 'struct btf_type' and any extra data following it are + * 4 bytes aligned. + * + * Type section: + * ~~~~~~~~~~~~~ + * The BTF type section contains a list of 'struct btf_type' objects. + * Each one describes a C type. Recall from the above section + * that a 'struct btf_type' object could be immediately followed by extra + * data in order to desribe some particular C types. + * + * type_id: + * ~~~~~~~ + * Each btf_type object is identified by a type_id. The type_id + * is implicitly implied by the location of the btf_type object in + * the BTF type section. The first one has type_id 1. The second + * one has type_id 2...etc. Hence, an earlier btf_type has + * a smaller type_id. + * + * A btf_type object may refer to another btf_type object by using + * type_id (i.e. the "type" in the "struct btf_type"). + * + * NOTE that we cannot assume any reference-order. + * A btf_type object can refer to an earlier btf_type object + * but it can also refer to a later btf_type object. + * + * For example, to describe "const void *". A btf_type + * object describing "const" may refer to another btf_type + * object describing "void *". This type-reference is done + * by specifying type_id: + * + * [1] CONST (anon) type_id=2 + * [2] PTR (anon) type_id=0 + * + * The above is the btf_verifier debug log: + * - Each line started with "[?]" is a btf_type object + * - [?] is the type_id of the btf_type object. + * - CONST/PTR is the BTF_KIND_XXX + * - "(anon)" is the name of the type. It just + * happens that CONST and PTR has no name. + * - type_id=XXX is the 'u32 type' in btf_type + * + * NOTE: "void" has type_id 0 + * + * String section: + * ~~~~~~~~~~~~~~ + * The BTF string section contains the names used by the type section. + * Each string is referred by an "offset" from the beginning of the + * string section. + * + * Each string is '\0' terminated. + * + * The first character in the string section must be '\0' + * which is used to mean 'anonymous'. Some btf_type may not + * have a name. + */ + +/* BTF verification: + * + * To verify BTF data, two passes are needed. + * + * Pass #1 + * ~~~~~~~ + * The first pass is to collect all btf_type objects to + * an array: "btf->types". + * + * Depending on the C type that a btf_type is describing, + * a btf_type may be followed by extra data. We don't know + * how many btf_type is there, and more importantly we don't + * know where each btf_type is located in the type section. + * + * Without knowing the location of each type_id, most verifications + * cannot be done. e.g. an earlier btf_type may refer to a later + * btf_type (recall the "const void *" above), so we cannot + * check this type-reference in the first pass. + * + * In the first pass, it still does some verifications (e.g. + * checking the name is a valid offset to the string section). + */ + +#define BITS_PER_U64 (sizeof(u64) * BITS_PER_BYTE) +#define BITS_PER_BYTE_MASK (BITS_PER_BYTE - 1) +#define BITS_PER_BYTE_MASKED(bits) ((bits) & BITS_PER_BYTE_MASK) +#define BITS_ROUNDDOWN_BYTES(bits) ((bits) >> 3) +#define BITS_ROUNDUP_BYTES(bits) \ + (BITS_ROUNDDOWN_BYTES(bits) + !!BITS_PER_BYTE_MASKED(bits)) + +/* 16MB for 64k structs and each has 16 members and + * a few MB spaces for the string section. + * The hard limit is S32_MAX. + */ +#define BTF_MAX_SIZE (16 * 1024 * 1024) +/* 64k. We can raise it later. The hard limit is S32_MAX. */ +#define BTF_MAX_NR_TYPES 65535 + +#define for_each_member(i, struct_type, member) \ + for (i = 0, member = btf_type_member(struct_type); \ + i < btf_type_vlen(struct_type); \ + i++, member++) + +struct btf { + union { + struct btf_header *hdr; + void *data; + }; + struct btf_type **types; + const char *strings; + void *nohdr_data; + u32 nr_types; + u32 types_size; + u32 data_size; +}; + +struct btf_verifier_env { + struct btf *btf; + struct bpf_verifier_log log; + u32 log_type_id; +}; + +static const char * const btf_kind_str[NR_BTF_KINDS] = { + [BTF_KIND_UNKN] = "UNKNOWN", + [BTF_KIND_INT] = "INT", + [BTF_KIND_FLOAT] = "FLOAT", + [BTF_KIND_PTR] = "PTR", + [BTF_KIND_ARRAY] = "ARRAY", + [BTF_KIND_FUNC] = "FUNC", + [BTF_KIND_STRUCT] = "STRUCT", + [BTF_KIND_UNION] = "UNION", + [BTF_KIND_ENUM] = "ENUM", + [BTF_KIND_FWD] = "FWD", + [BTF_KIND_TYPEDEF] = "TYPEDEF", + [BTF_KIND_VOLATILE] = "VOLATILE", + [BTF_KIND_CONST] = "CONST", + [BTF_KIND_RESTRICT] = "RESTRICT", +}; + +struct btf_kind_operations { + s32 (*check_meta)(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left); + void (*log_details)(struct btf_verifier_env *env, + const struct btf_type *t); +}; + +static const struct btf_kind_operations * const kind_ops[NR_BTF_KINDS]; +static struct btf_type btf_void; + +static const char *btf_int_encoding_str(u8 encoding) +{ + if (encoding == 0) + return "(none)"; + else if (encoding == BTF_INT_SIGNED) + return "SIGNED"; + else if (encoding == BTF_INT_CHAR) + return "CHAR"; + else if (encoding == BTF_INT_BOOL) + return "BOOL"; + else if (encoding == BTF_INT_VARARGS) + return "VARARGS"; + else + return "UNKN"; +} + +static u16 btf_type_vlen(const struct btf_type *t) +{ + return BTF_INFO_VLEN(t->info); +} + +static u32 btf_type_int(const struct btf_type *t) +{ + return *(u32 *)(t + 1); +} + +static const struct btf_array *btf_type_array(const struct btf_type *t) +{ + return (const struct btf_array *)(t + 1); +} + +static const struct btf_member *btf_type_member(const struct btf_type *t) +{ + return (const struct btf_member *)(t + 1); +} + +static const struct btf_enum *btf_type_enum(const struct btf_type *t) +{ + return (const struct btf_enum *)(t + 1); +} + +static const struct btf_kind_operations *btf_type_ops(const struct btf_type *t) +{ + return kind_ops[BTF_INFO_KIND(t->info)]; +} + +static bool btf_name_offset_valid(const struct btf *btf, u32 offset) +{ + return !BTF_STR_TBL_ELF_ID(offset) && + BTF_STR_OFFSET(offset) < btf->hdr->str_len; +} + +static const char *btf_name_by_offset(const struct btf *btf, u32 offset) +{ + if (!BTF_STR_OFFSET(offset)) + return "(anon)"; + else if (BTF_STR_OFFSET(offset) < btf->hdr->str_len) + return &btf->strings[BTF_STR_OFFSET(offset)]; + else + return "(invalid-name-offset)"; +} + +__printf(2, 3) static void __btf_verifier_log(struct bpf_verifier_log *log, + const char *fmt, ...) +{ + va_list args; + + va_start(args, fmt); + bpf_verifier_vlog(log, fmt, args); + va_end(args); +} + +__printf(2, 3) static void btf_verifier_log(struct btf_verifier_env *env, + const char *fmt, ...) +{ + struct bpf_verifier_log *log = &env->log; + va_list args; + + if (!bpf_verifier_log_needed(log)) + return; + + va_start(args, fmt); + bpf_verifier_vlog(log, fmt, args); + va_end(args); +} + +__printf(4, 5) static void __btf_verifier_log_type(struct btf_verifier_env *env, + const struct btf_type *t, + bool log_details, + const char *fmt, ...) +{ + struct bpf_verifier_log *log = &env->log; + u8 kind = BTF_INFO_KIND(t->info); + struct btf *btf = env->btf; + va_list args; + + if (!bpf_verifier_log_needed(log)) + return; + + __btf_verifier_log(log, "[%u] %s %s%s", + env->log_type_id, + btf_kind_str[kind], + btf_name_by_offset(btf, t->name), + log_details ? " " : ""); + + if (log_details) + btf_type_ops(t)->log_details(env, t); + + if (fmt && *fmt) { + __btf_verifier_log(log, " "); + va_start(args, fmt); + bpf_verifier_vlog(log, fmt, args); + va_end(args); + } + + __btf_verifier_log(log, "\n"); +} + +#define btf_verifier_log_type(env, t, ...) \ + __btf_verifier_log_type((env), (t), true, __VA_ARGS__) +#define btf_verifier_log_basic(env, t, ...) \ + __btf_verifier_log_type((env), (t), false, __VA_ARGS__) + +__printf(4, 5) +static void btf_verifier_log_member(struct btf_verifier_env *env, + const struct btf_type *struct_type, + const struct btf_member *member, + const char *fmt, ...) +{ + struct bpf_verifier_log *log = &env->log; + struct btf *btf = env->btf; + va_list args; + + if (!bpf_verifier_log_needed(log)) + return; + + __btf_verifier_log(log, "\t%s type_id=%u bits_offset=%u", + btf_name_by_offset(btf, member->name), + member->type, member->offset); + + if (fmt && *fmt) { + __btf_verifier_log(log, " "); + va_start(args, fmt); + bpf_verifier_vlog(log, fmt, args); + va_end(args); + } + + __btf_verifier_log(log, "\n"); +} + +static void btf_verifier_log_hdr(struct btf_verifier_env *env) +{ + struct bpf_verifier_log *log = &env->log; + const struct btf *btf = env->btf; + const struct btf_header *hdr; + + if (!bpf_verifier_log_needed(log)) + return; + + hdr = btf->hdr; + __btf_verifier_log(log, "magic: 0x%x\n", hdr->magic); + __btf_verifier_log(log, "version: %u\n", hdr->version); + __btf_verifier_log(log, "flags: 0x%x\n", hdr->flags); + __btf_verifier_log(log, "parent_label: %u\n", hdr->parent_label); + __btf_verifier_log(log, "parent_name: %u\n", hdr->parent_name); + __btf_verifier_log(log, "label_off: %u\n", hdr->label_off); + __btf_verifier_log(log, "object_off: %u\n", hdr->object_off); + __btf_verifier_log(log, "func_off: %u\n", hdr->func_off); + __btf_verifier_log(log, "type_off: %u\n", hdr->type_off); + __btf_verifier_log(log, "str_off: %u\n", hdr->str_off); + __btf_verifier_log(log, "str_len: %u\n", hdr->str_len); + __btf_verifier_log(log, "btf_total_size: %u\n", btf->data_size); +} + +static int btf_add_type(struct btf_verifier_env *env, struct btf_type *t) +{ + struct btf *btf = env->btf; + + /* < 2 because +1 for btf_void which is always in btf->types[0]. + * btf_void is not accounted in btf->nr_types because btf_void + * does not come from the BTF file. + */ + if (btf->types_size - btf->nr_types < 2) { + /* Expand 'types' array */ + + struct btf_type **new_types; + u32 expand_by, new_size; + + if (btf->types_size == BTF_MAX_NR_TYPES) { + btf_verifier_log(env, "Exceeded max num of types"); + return -E2BIG; + } + + expand_by = max_t(u32, btf->types_size >> 2, 16); + new_size = min_t(u32, BTF_MAX_NR_TYPES, + btf->types_size + expand_by); + + new_types = kvzalloc(new_size * sizeof(*new_types), + GFP_KERNEL | __GFP_NOWARN); + if (!new_types) + return -ENOMEM; + + if (btf->nr_types == 0) + new_types[0] = &btf_void; + else + memcpy(new_types, btf->types, + sizeof(*btf->types) * (btf->nr_types + 1)); + + kvfree(btf->types); + btf->types = new_types; + btf->types_size = new_size; + } + + btf->types[++(btf->nr_types)] = t; + + return 0; +} + +static void btf_free(struct btf *btf) +{ + kvfree(btf->types); + kvfree(btf->data); + kfree(btf); +} + +static void btf_verifier_env_free(struct btf_verifier_env *env) +{ + kfree(env); +} + +static int btf_df_check_meta(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left) +{ + btf_verifier_log_basic(env, t, "Unsupported check_meta"); + return -ENOTSUPP; +} + +static void btf_df_log(struct btf_verifier_env *env, + const struct btf_type *t) +{ + btf_verifier_log(env, "Unsupported log_details"); +} + +static struct btf_kind_operations df_ops = { + .check_meta = btf_df_check_meta, + .log_details = btf_df_log, +}; + +static s32 btf_int_check_meta(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left) +{ + u32 int_data, nr_bits, meta_needed = sizeof(int_data); + u32 nr_bytes; + u16 encoding; + + if (meta_left < meta_needed) { + btf_verifier_log_basic(env, t, + "meta_left:%u meta_needed:%u", + meta_left, meta_needed); + return -EINVAL; + } + + if (btf_type_vlen(t)) { + btf_verifier_log_type(env, t, "vlen != 0"); + return -EINVAL; + } + + int_data = btf_type_int(t); + nr_bits = BTF_INT_BITS(int_data) + BTF_INT_OFFSET(int_data); + nr_bytes = t->size; + + if (nr_bits > BITS_PER_U64) { + btf_verifier_log_type(env, t, "nr_bits exceeds %zu", + BITS_PER_U64); + return -EINVAL; + } + + if (BITS_ROUNDUP_BYTES(nr_bits) > t->size) { + btf_verifier_log_type(env, t, "nr_bits exceeds type_size"); + return -EINVAL; + } + + encoding = BTF_INT_ENCODING(int_data); + if (encoding && + encoding != BTF_INT_SIGNED && + encoding != BTF_INT_CHAR && + encoding != BTF_INT_BOOL && + encoding != BTF_INT_VARARGS) { + btf_verifier_log_type(env, t, "Unsupported encoding"); + return -ENOTSUPP; + } + + btf_verifier_log_type(env, t, NULL); + + return meta_needed; +} + +static void btf_int_log(struct btf_verifier_env *env, + const struct btf_type *t) +{ + int int_data = btf_type_int(t); + + btf_verifier_log(env, + "size=%u bits_offset=%u nr_bits=%u encoding=%s", + t->size, BTF_INT_OFFSET(int_data), + BTF_INT_BITS(int_data), + btf_int_encoding_str(BTF_INT_ENCODING(int_data))); +} + +static const struct btf_kind_operations int_ops = { + .check_meta = btf_int_check_meta, + .log_details = btf_int_log, +}; + +static int btf_ref_type_check_meta(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left) +{ + if (btf_type_vlen(t)) { + btf_verifier_log_type(env, t, "vlen != 0"); + return -EINVAL; + } + + if (BTF_TYPE_PARENT(t->type)) { + btf_verifier_log_type(env, t, "Invalid type_id"); + return -EINVAL; + } + + btf_verifier_log_type(env, t, NULL); + + return 0; +} + +static void btf_ref_type_log(struct btf_verifier_env *env, + const struct btf_type *t) +{ + btf_verifier_log(env, "type_id=%u", t->type); +} + +static struct btf_kind_operations modifier_ops = { + .check_meta = btf_ref_type_check_meta, + .log_details = btf_ref_type_log, +}; + +static struct btf_kind_operations ptr_ops = { + .check_meta = btf_ref_type_check_meta, + .log_details = btf_ref_type_log, +}; + +static struct btf_kind_operations fwd_ops = { + .check_meta = btf_ref_type_check_meta, + .log_details = btf_ref_type_log, +}; + +static s32 btf_array_check_meta(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left) +{ + const struct btf_array *array = btf_type_array(t); + u32 meta_needed = sizeof(*array); + + if (meta_left < meta_needed) { + btf_verifier_log_basic(env, t, + "meta_left:%u meta_needed:%u", + meta_left, meta_needed); + return -EINVAL; + } + + if (btf_type_vlen(t)) { + btf_verifier_log_type(env, t, "vlen != 0"); + return -EINVAL; + } + + /* We are a little forgiving on array->index_type since + * the kernel is not using it. + */ + /* Array elem cannot be in type void, + * so !array->type is not allowed. + */ + if (!array->type || BTF_TYPE_PARENT(array->type)) { + btf_verifier_log_type(env, t, "Invalid type_id"); + return -EINVAL; + } + + btf_verifier_log_type(env, t, NULL); + + return meta_needed; +} + +static void btf_array_log(struct btf_verifier_env *env, + const struct btf_type *t) +{ + const struct btf_array *array = btf_type_array(t); + + btf_verifier_log(env, "type_id=%u index_type_id=%u nr_elems=%u", + array->type, array->index_type, array->nelems); +} + +static struct btf_kind_operations array_ops = { + .check_meta = btf_array_check_meta, + .log_details = btf_array_log, +}; + +static s32 btf_struct_check_meta(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left) +{ + bool is_union = BTF_INFO_KIND(t->info) == BTF_KIND_UNION; + const struct btf_member *member; + struct btf *btf = env->btf; + u32 struct_size = t->size; + u32 meta_needed; + u16 i; + + meta_needed = btf_type_vlen(t) * sizeof(*member); + if (meta_left < meta_needed) { + btf_verifier_log_basic(env, t, + "meta_left:%u meta_needed:%u", + meta_left, meta_needed); + return -EINVAL; + } + + btf_verifier_log_type(env, t, NULL); + + for_each_member(i, t, member) { + if (!btf_name_offset_valid(btf, member->name)) { + btf_verifier_log_member(env, t, member, + "Invalid member name_offset:%u", + member->name); + return -EINVAL; + } + + /* A member cannot be in type void */ + if (!member->type || BTF_TYPE_PARENT(member->type)) { + btf_verifier_log_member(env, t, member, + "Invalid type_id"); + return -EINVAL; + } + + if (is_union && member->offset) { + btf_verifier_log_member(env, t, member, + "Invalid member bits_offset"); + return -EINVAL; + } + + if (BITS_ROUNDUP_BYTES(member->offset) > struct_size) { + btf_verifier_log_member(env, t, member, + "Memmber bits_offset exceeds its struct size"); + return -EINVAL; + } + + btf_verifier_log_member(env, t, member, NULL); + } + + return meta_needed; +} + +static void btf_struct_log(struct btf_verifier_env *env, + const struct btf_type *t) +{ + btf_verifier_log(env, "size=%u vlen=%u", t->size, btf_type_vlen(t)); +} + +static struct btf_kind_operations struct_ops = { + .check_meta = btf_struct_check_meta, + .log_details = btf_struct_log, +}; + +static s32 btf_enum_check_meta(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left) +{ + const struct btf_enum *enums = btf_type_enum(t); + struct btf *btf = env->btf; + u16 i, nr_enums; + u32 meta_needed; + + nr_enums = btf_type_vlen(t); + meta_needed = nr_enums * sizeof(*enums); + + if (meta_left < meta_needed) { + btf_verifier_log_basic(env, t, + "meta_left:%u meta_needed:%u", + meta_left, meta_needed); + return -EINVAL; + } + + if (t->size != sizeof(int)) { + btf_verifier_log_type(env, t, "Expected size:%zu", + sizeof(int)); + return -EINVAL; + } + + btf_verifier_log_type(env, t, NULL); + + for (i = 0; i < nr_enums; i++) { + if (!btf_name_offset_valid(btf, enums[i].name)) { + btf_verifier_log(env, "\tInvalid name_offset:%u", + enums[i].name); + return -EINVAL; + } + + btf_verifier_log(env, "\t%s val=%d\n", + btf_name_by_offset(btf, enums[i].name), + enums[i].val); + } + + return meta_needed; +} + +static void btf_enum_log(struct btf_verifier_env *env, + const struct btf_type *t) +{ + btf_verifier_log(env, "size=%u vlen=%u", t->size, btf_type_vlen(t)); +} + +static struct btf_kind_operations enum_ops = { + .check_meta = btf_enum_check_meta, + .log_details = btf_enum_log, +}; + +static const struct btf_kind_operations * const kind_ops[NR_BTF_KINDS] = { + [BTF_KIND_INT] = &int_ops, + [BTF_KIND_FLOAT] = &df_ops, + [BTF_KIND_PTR] = &ptr_ops, + [BTF_KIND_ARRAY] = &array_ops, + [BTF_KIND_FUNC] = &df_ops, + [BTF_KIND_STRUCT] = &struct_ops, + [BTF_KIND_UNION] = &struct_ops, + [BTF_KIND_ENUM] = &enum_ops, + [BTF_KIND_FWD] = &fwd_ops, + [BTF_KIND_TYPEDEF] = &modifier_ops, + [BTF_KIND_VOLATILE] = &modifier_ops, + [BTF_KIND_CONST] = &modifier_ops, + [BTF_KIND_RESTRICT] = &modifier_ops, +}; + +static s32 btf_check_meta(struct btf_verifier_env *env, + const struct btf_type *t, + u32 meta_left) +{ + u32 saved_meta_left = meta_left; + s32 var_meta_size; + + if (meta_left < sizeof(*t)) { + btf_verifier_log(env, "[%u] meta_left:%u meta_needed:%zu", + env->log_type_id, meta_left, sizeof(*t)); + return -EINVAL; + } + meta_left -= sizeof(*t); + + if (BTF_INFO_KIND(t->info) > BTF_KIND_MAX || + BTF_INFO_KIND(t->info) == BTF_KIND_UNKN) { + btf_verifier_log(env, "[%u] Invalid kind:%u", + env->log_type_id, BTF_INFO_KIND(t->info)); + return -EINVAL; + } + + if (!btf_name_offset_valid(env->btf, t->name)) { + btf_verifier_log(env, "[%u] Invalid name_offset:%u", + env->log_type_id, t->name); + return -EINVAL; + } + + var_meta_size = btf_type_ops(t)->check_meta(env, t, meta_left); + if (var_meta_size < 0) + return var_meta_size; + + meta_left -= var_meta_size; + + return saved_meta_left - meta_left; +} + +static int btf_check_all_metas(struct btf_verifier_env *env) +{ + struct btf *btf = env->btf; + struct btf_header *hdr; + void *cur, *end; + + hdr = btf->hdr; + cur = btf->nohdr_data + hdr->type_off; + end = btf->nohdr_data + hdr->str_off; + + env->log_type_id = 1; + while (cur < end) { + struct btf_type *t = cur; + s32 meta_size; + + meta_size = btf_check_meta(env, t, end - cur); + if (meta_size < 0) + return meta_size; + + btf_add_type(env, t); + cur += meta_size; + env->log_type_id++; + } + + return 0; +} + +static int btf_parse_type_sec(struct btf_verifier_env *env) +{ + return btf_check_all_metas(env); +} + +static int btf_parse_str_sec(struct btf_verifier_env *env) +{ + const struct btf_header *hdr; + struct btf *btf = env->btf; + const char *start, *end; + + hdr = btf->hdr; + start = btf->nohdr_data + hdr->str_off; + end = start + hdr->str_len; + + if (!hdr->str_len || hdr->str_len - 1 > BTF_MAX_NAME_OFFSET || + start[0] || end[-1]) { + btf_verifier_log(env, "Invalid string section"); + return -EINVAL; + } + + btf->strings = start; + + return 0; +} + +static int btf_parse_hdr(struct btf_verifier_env *env) +{ + const struct btf_header *hdr; + struct btf *btf = env->btf; + u32 meta_left; + + if (btf->data_size < sizeof(*hdr)) { + btf_verifier_log(env, "btf_header not found"); + return -EINVAL; + } + + btf_verifier_log_hdr(env); + + hdr = btf->hdr; + if (hdr->magic != BTF_MAGIC) { + btf_verifier_log(env, "Invalid magic"); + return -EINVAL; + } + + if (hdr->version != BTF_VERSION) { + btf_verifier_log(env, "Unsupported version"); + return -ENOTSUPP; + } + + if (hdr->flags) { + btf_verifier_log(env, "Unsupported flags"); + return -ENOTSUPP; + } + + meta_left = btf->data_size - sizeof(*hdr); + if (!meta_left) { + btf_verifier_log(env, "No data"); + return -EINVAL; + } + + if (meta_left < hdr->type_off || hdr->str_off <= hdr->type_off || + /* Type section must align to 4 bytes */ + hdr->type_off & (sizeof(u32) - 1)) { + btf_verifier_log(env, "Invalid type_off"); + return -EINVAL; + } + + if (meta_left < hdr->str_off || + meta_left - hdr->str_off < hdr->str_len) { + btf_verifier_log(env, "Invalid str_off or str_len"); + return -EINVAL; + } + + btf->nohdr_data = btf->hdr + 1; + + return 0; +} + +static struct btf *btf_parse(void __user *btf_data, u32 btf_data_size, + u32 log_level, char __user *log_ubuf, u32 log_size) +{ + struct btf_verifier_env *env = NULL; + struct bpf_verifier_log *log; + struct btf *btf = NULL; + u8 *data; + int err; + + if (btf_data_size > BTF_MAX_SIZE) + return ERR_PTR(-E2BIG); + + env = kzalloc(sizeof(*env), GFP_KERNEL | __GFP_NOWARN); + if (!env) + return ERR_PTR(-ENOMEM); + + log = &env->log; + if (log_level || log_ubuf || log_size) { + /* user requested verbose verifier output + * and supplied buffer to store the verification trace + */ + log->level = log_level; + log->ubuf = log_ubuf; + log->len_total = log_size; + + /* log attributes have to be sane */ + if (log->len_total < 128 || log->len_total > UINT_MAX >> 8 || + !log->level || !log->ubuf) { + err = -EINVAL; + goto errout; + } + } + + btf = kzalloc(sizeof(*btf), GFP_KERNEL | __GFP_NOWARN); + if (!btf) { + err = -ENOMEM; + goto errout; + } + + data = kvmalloc(btf_data_size, GFP_KERNEL | __GFP_NOWARN); + if (!data) { + err = -ENOMEM; + goto errout; + } + + btf->data = data; + btf->data_size = btf_data_size; + + if (copy_from_user(data, btf_data, btf_data_size)) { + err = -EFAULT; + goto errout; + } + + env->btf = btf; + + err = btf_parse_hdr(env); + if (err) + goto errout; + + err = btf_parse_str_sec(env); + if (err) + goto errout; + + err = btf_parse_type_sec(env); + if (err) + goto errout; + + if (!err && log->level && bpf_verifier_log_full(log)) { + err = -ENOSPC; + goto errout; + } + + if (!err) { + btf_verifier_env_free(env); + return btf; + } + +errout: + btf_verifier_env_free(env); + if (btf) + btf_free(btf); + return ERR_PTR(err); +}