diff options
Diffstat (limited to 'Documentation/bpf/standardization')
| -rw-r--r-- | Documentation/bpf/standardization/abi.rst | 3 | ||||
| -rw-r--r-- | Documentation/bpf/standardization/instruction-set.rst | 333 |
2 files changed, 196 insertions, 140 deletions
diff --git a/Documentation/bpf/standardization/abi.rst b/Documentation/bpf/standardization/abi.rst index 0c2e10eeb89a..41514137cb7b 100644 --- a/Documentation/bpf/standardization/abi.rst +++ b/Documentation/bpf/standardization/abi.rst @@ -23,3 +23,6 @@ The BPF calling convention is defined as: R0 - R5 are scratch registers and BPF programs needs to spill/fill them if necessary across calls. + +The BPF program needs to store the return value into register R0 before doing an +``EXIT``. diff --git a/Documentation/bpf/standardization/instruction-set.rst b/Documentation/bpf/standardization/instruction-set.rst index 00c93eb42613..ab820d565052 100644 --- a/Documentation/bpf/standardization/instruction-set.rst +++ b/Documentation/bpf/standardization/instruction-set.rst @@ -5,15 +5,29 @@ BPF Instruction Set Architecture (ISA) ====================================== -eBPF (which is no longer an acronym for anything), also commonly +eBPF, also commonly referred to as BPF, is a technology with origins in the Linux kernel that can run untrusted programs in a privileged context such as an operating system kernel. This document specifies the BPF instruction set architecture (ISA). +As a historical note, BPF originally stood for Berkeley Packet Filter, +but now that it can do so much more than packet filtering, the acronym +no longer makes sense. BPF is now considered a standalone term that +does not stand for anything. The original BPF is sometimes referred to +as cBPF (classic BPF) to distinguish it from the now widely deployed +eBPF (extended BPF). + Documentation conventions ========================= +The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", +"SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and +"OPTIONAL" in this document are to be interpreted as described in +BCP 14 `<https://www.rfc-editor.org/info/rfc2119>`_ +`<https://www.rfc-editor.org/info/rfc8174>`_ +when, and only when, they appear in all capitals, as shown here. + For brevity and consistency, this document refers to families of types using a shorthand syntax and refers to several expository, mnemonic functions when describing the semantics of instructions. @@ -25,7 +39,7 @@ Types This document refers to integer types with the notation `SN` to specify a type's signedness (`S`) and bit width (`N`), respectively. -.. table:: Meaning of signedness notation. +.. table:: Meaning of signedness notation ==== ========= S Meaning @@ -34,7 +48,7 @@ a type's signedness (`S`) and bit width (`N`), respectively. s signed ==== ========= -.. table:: Meaning of bit-width notation. +.. table:: Meaning of bit-width notation ===== ========= N Bit width @@ -52,24 +66,18 @@ numbers. Functions --------- -* htobe16: Takes an unsigned 16-bit number in host-endian format and - returns the equivalent number as an unsigned 16-bit number in big-endian - format. -* htobe32: Takes an unsigned 32-bit number in host-endian format and - returns the equivalent number as an unsigned 32-bit number in big-endian - format. -* htobe64: Takes an unsigned 64-bit number in host-endian format and - returns the equivalent number as an unsigned 64-bit number in big-endian - format. -* htole16: Takes an unsigned 16-bit number in host-endian format and - returns the equivalent number as an unsigned 16-bit number in little-endian - format. -* htole32: Takes an unsigned 32-bit number in host-endian format and - returns the equivalent number as an unsigned 32-bit number in little-endian - format. -* htole64: Takes an unsigned 64-bit number in host-endian format and - returns the equivalent number as an unsigned 64-bit number in little-endian - format. + +The following byteswap functions are direction-agnostic. That is, +the same function is used for conversion in either direction discussed +below. + +* be16: Takes an unsigned 16-bit number and converts it between + host byte order and big-endian + (`IEN137 <https://www.rfc-editor.org/ien/ien137.txt>`_) byte order. +* be32: Takes an unsigned 32-bit number and converts it between + host byte order and big-endian byte order. +* be64: Takes an unsigned 64-bit number and converts it between + host byte order and big-endian byte order. * bswap16: Takes an unsigned 16-bit number in either big- or little-endian format and returns the equivalent number with the same bit width but opposite endianness. @@ -79,7 +87,12 @@ Functions * bswap64: Takes an unsigned 64-bit number in either big- or little-endian format and returns the equivalent number with the same bit width but opposite endianness. - +* le16: Takes an unsigned 16-bit number and converts it between + host byte order and little-endian byte order. +* le32: Takes an unsigned 32-bit number and converts it between + host byte order and little-endian byte order. +* le64: Takes an unsigned 64-bit number and converts it between + host byte order and little-endian byte order. Definitions ----------- @@ -106,9 +119,9 @@ Conformance groups An implementation does not need to support all instructions specified in this document (e.g., deprecated instructions). Instead, a number of conformance -groups are specified. An implementation must support the base32 conformance -group and may support additional conformance groups, where supporting a -conformance group means it must support all instructions in that conformance +groups are specified. An implementation MUST support the base32 conformance +group and MAY support additional conformance groups, where supporting a +conformance group means it MUST support all instructions in that conformance group. The use of named conformance groups enables interoperability between a runtime @@ -209,7 +222,7 @@ For example:: 07 1 0 00 00 11 22 33 44 r1 += 0x11223344 // big Note that most instructions do not use all of the fields. -Unused fields shall be cleared to zero. +Unused fields SHALL be cleared to zero. Wide instruction encoding -------------------------- @@ -256,18 +269,20 @@ Instruction classes The three least significant bits of the 'opcode' field store the instruction class: -===== ===== =============================== =================================== -class value description reference -===== ===== =============================== =================================== -LD 0x0 non-standard load operations `Load and store instructions`_ -LDX 0x1 load into register operations `Load and store instructions`_ -ST 0x2 store from immediate operations `Load and store instructions`_ -STX 0x3 store from register operations `Load and store instructions`_ -ALU 0x4 32-bit arithmetic operations `Arithmetic and jump instructions`_ -JMP 0x5 64-bit jump operations `Arithmetic and jump instructions`_ -JMP32 0x6 32-bit jump operations `Arithmetic and jump instructions`_ -ALU64 0x7 64-bit arithmetic operations `Arithmetic and jump instructions`_ -===== ===== =============================== =================================== +.. table:: Instruction class + + ===== ===== =============================== =================================== + class value description reference + ===== ===== =============================== =================================== + LD 0x0 non-standard load operations `Load and store instructions`_ + LDX 0x1 load into register operations `Load and store instructions`_ + ST 0x2 store from immediate operations `Load and store instructions`_ + STX 0x3 store from register operations `Load and store instructions`_ + ALU 0x4 32-bit arithmetic operations `Arithmetic and jump instructions`_ + JMP 0x5 64-bit jump operations `Arithmetic and jump instructions`_ + JMP32 0x6 32-bit jump operations `Arithmetic and jump instructions`_ + ALU64 0x7 64-bit arithmetic operations `Arithmetic and jump instructions`_ + ===== ===== =============================== =================================== Arithmetic and jump instructions ================================ @@ -285,12 +300,14 @@ For arithmetic and jump instructions (``ALU``, ``ALU64``, ``JMP`` and **s (source)** the source operand location, which unless otherwise specified is one of: - ====== ===== ============================================== - source value description - ====== ===== ============================================== - K 0 use 32-bit 'imm' value as source operand - X 1 use 'src_reg' register value as source operand - ====== ===== ============================================== + .. table:: Source operand location + + ====== ===== ============================================== + source value description + ====== ===== ============================================== + K 0 use 32-bit 'imm' value as source operand + X 1 use 'src_reg' register value as source operand + ====== ===== ============================================== **instruction class** the instruction class (see `Instruction classes`_) @@ -305,27 +322,29 @@ The 'code' field encodes the operation as below, where 'src' refers to the the source operand and 'dst' refers to the value of the destination register. -===== ===== ======= ========================================================== -name code offset description -===== ===== ======= ========================================================== -ADD 0x0 0 dst += src -SUB 0x1 0 dst -= src -MUL 0x2 0 dst \*= src -DIV 0x3 0 dst = (src != 0) ? (dst / src) : 0 -SDIV 0x3 1 dst = (src != 0) ? (dst s/ src) : 0 -OR 0x4 0 dst \|= src -AND 0x5 0 dst &= src -LSH 0x6 0 dst <<= (src & mask) -RSH 0x7 0 dst >>= (src & mask) -NEG 0x8 0 dst = -dst -MOD 0x9 0 dst = (src != 0) ? (dst % src) : dst -SMOD 0x9 1 dst = (src != 0) ? (dst s% src) : dst -XOR 0xa 0 dst ^= src -MOV 0xb 0 dst = src -MOVSX 0xb 8/16/32 dst = (s8,s16,s32)src -ARSH 0xc 0 :term:`sign extending<Sign Extend>` dst >>= (src & mask) -END 0xd 0 byte swap operations (see `Byte swap instructions`_ below) -===== ===== ======= ========================================================== +.. table:: Arithmetic instructions + + ===== ===== ======= ========================================================== + name code offset description + ===== ===== ======= ========================================================== + ADD 0x0 0 dst += src + SUB 0x1 0 dst -= src + MUL 0x2 0 dst \*= src + DIV 0x3 0 dst = (src != 0) ? (dst / src) : 0 + SDIV 0x3 1 dst = (src != 0) ? (dst s/ src) : 0 + OR 0x4 0 dst \|= src + AND 0x5 0 dst &= src + LSH 0x6 0 dst <<= (src & mask) + RSH 0x7 0 dst >>= (src & mask) + NEG 0x8 0 dst = -dst + MOD 0x9 0 dst = (src != 0) ? (dst % src) : dst + SMOD 0x9 1 dst = (src != 0) ? (dst s% src) : dst + XOR 0xa 0 dst ^= src + MOV 0xb 0 dst = src + MOVSX 0xb 8/16/32 dst = (s8,s16,s32)src + ARSH 0xc 0 :term:`sign extending<Sign Extend>` dst >>= (src & mask) + END 0xd 0 byte swap operations (see `Byte swap instructions`_ below) + ===== ===== ======= ========================================================== Underflow and overflow are allowed during arithmetic operations, meaning the 64-bit or 32-bit value will wrap. If BPF program execution would @@ -374,7 +393,7 @@ interpreted as a 64-bit signed value. Note that there are varying definitions of the signed modulo operation when the dividend or divisor are negative, where implementations often vary by language such that Python, Ruby, etc. differ from C, Go, Java, -etc. This specification requires that signed modulo use truncated division +etc. This specification requires that signed modulo MUST use truncated division (where -13 % 3 == -1) as implemented in C, Go, etc.:: a % n = a - n * trunc(a / n) @@ -386,6 +405,19 @@ The ``MOVSX`` instruction does a move operation with sign extension. operands into 64-bit operands. Unlike other arithmetic instructions, ``MOVSX`` is only defined for register source operands (``X``). +``{MOV, K, ALU64}`` means:: + + dst = (s64)imm + +``{MOV, X, ALU}`` means:: + + dst = (u32)src + +``{MOVSX, X, ALU}`` with 'offset' 8 means:: + + dst = (u32)(s32)(s8)src + + The ``NEG`` instruction is only defined when the source bit is clear (``K``). @@ -404,15 +436,17 @@ only and do not use a separate source register or immediate value. For ``ALU``, the 1-bit source operand field in the opcode is used to select what byte order the operation converts from or to. For ``ALU64``, the 1-bit source operand field in the opcode is reserved -and must be set to 0. +and MUST be set to 0. + +.. table:: Byte swap instructions -===== ======== ===== ================================================= -class source value description -===== ======== ===== ================================================= -ALU TO_LE 0 convert between host byte order and little endian -ALU TO_BE 1 convert between host byte order and big endian -ALU64 Reserved 0 do byte swap unconditionally -===== ======== ===== ================================================= + ===== ======== ===== ================================================= + class source value description + ===== ======== ===== ================================================= + ALU LE 0 convert between host byte order and little endian + ALU BE 1 convert between host byte order and big endian + ALU64 Reserved 0 do byte swap unconditionally + ===== ======== ===== ================================================= The 'imm' field encodes the width of the swap operations. The following widths are supported: 16, 32 and 64. Width 64 operations belong to the base64 @@ -421,19 +455,19 @@ conformance group. Examples: -``{END, TO_LE, ALU}`` with 'imm' = 16/32/64 means:: +``{END, LE, ALU}`` with 'imm' = 16/32/64 means:: - dst = htole16(dst) - dst = htole32(dst) - dst = htole64(dst) + dst = le16(dst) + dst = le32(dst) + dst = le64(dst) -``{END, TO_BE, ALU}`` with 'imm' = 16/32/64 means:: +``{END, BE, ALU}`` with 'imm' = 16/32/64 means:: - dst = htobe16(dst) - dst = htobe32(dst) - dst = htobe64(dst) + dst = be16(dst) + dst = be32(dst) + dst = be64(dst) -``{END, TO_LE, ALU64}`` with 'imm' = 16/32/64 means:: +``{END, TO, ALU64}`` with 'imm' = 16/32/64 means:: dst = bswap16(dst) dst = bswap32(dst) @@ -448,27 +482,29 @@ otherwise identical operations, and indicates the base64 conformance group unless otherwise specified. The 'code' field encodes the operation as below: -======== ===== ======= ================================= =================================================== -code value src_reg description notes -======== ===== ======= ================================= =================================================== -JA 0x0 0x0 PC += offset {JA, K, JMP} only -JA 0x0 0x0 PC += imm {JA, K, JMP32} only -JEQ 0x1 any PC += offset if dst == src -JGT 0x2 any PC += offset if dst > src unsigned -JGE 0x3 any PC += offset if dst >= src unsigned -JSET 0x4 any PC += offset if dst & src -JNE 0x5 any PC += offset if dst != src -JSGT 0x6 any PC += offset if dst > src signed -JSGE 0x7 any PC += offset if dst >= src signed -CALL 0x8 0x0 call helper function by static ID {CALL, K, JMP} only, see `Helper functions`_ -CALL 0x8 0x1 call PC += imm {CALL, K, JMP} only, see `Program-local functions`_ -CALL 0x8 0x2 call helper function by BTF ID {CALL, K, JMP} only, see `Helper functions`_ -EXIT 0x9 0x0 return {CALL, K, JMP} only -JLT 0xa any PC += offset if dst < src unsigned -JLE 0xb any PC += offset if dst <= src unsigned -JSLT 0xc any PC += offset if dst < src signed -JSLE 0xd any PC += offset if dst <= src signed -======== ===== ======= ================================= =================================================== +.. table:: Jump instructions + + ======== ===== ======= ================================= =================================================== + code value src_reg description notes + ======== ===== ======= ================================= =================================================== + JA 0x0 0x0 PC += offset {JA, K, JMP} only + JA 0x0 0x0 PC += imm {JA, K, JMP32} only + JEQ 0x1 any PC += offset if dst == src + JGT 0x2 any PC += offset if dst > src unsigned + JGE 0x3 any PC += offset if dst >= src unsigned + JSET 0x4 any PC += offset if dst & src + JNE 0x5 any PC += offset if dst != src + JSGT 0x6 any PC += offset if dst > src signed + JSGE 0x7 any PC += offset if dst >= src signed + CALL 0x8 0x0 call helper function by static ID {CALL, K, JMP} only, see `Helper functions`_ + CALL 0x8 0x1 call PC += imm {CALL, K, JMP} only, see `Program-local functions`_ + CALL 0x8 0x2 call helper function by BTF ID {CALL, K, JMP} only, see `Helper functions`_ + EXIT 0x9 0x0 return {CALL, K, JMP} only + JLT 0xa any PC += offset if dst < src unsigned + JLE 0xb any PC += offset if dst <= src unsigned + JSLT 0xc any PC += offset if dst < src signed + JSLE 0xd any PC += offset if dst <= src signed + ======== ===== ======= ================================= =================================================== where 'PC' denotes the program counter, and the offset to increment by is in units of 64-bit instructions relative to the instruction following @@ -476,9 +512,6 @@ the jump instruction. Thus 'PC += 1' skips execution of the next instruction if it's a basic instruction or results in undefined behavior if the next instruction is a 128-bit wide instruction. -The BPF program needs to store the return value into register R0 before doing an -``EXIT``. - Example: ``{JSGE, X, JMP32}`` means:: @@ -487,6 +520,10 @@ Example: where 's>=' indicates a signed '>=' comparison. +``{JLE, K, JMP}`` means:: + + if dst <= (u64)(s64)imm goto +offset + ``{JA, K, JMP32}`` means:: gotol +imm @@ -510,19 +547,25 @@ Helper functions are a concept whereby BPF programs can call into a set of function calls exposed by the underlying platform. Historically, each helper function was identified by a static ID -encoded in the 'imm' field. The available helper functions may differ -for each program type, but static IDs are unique across all program types. +encoded in the 'imm' field. Further documentation of helper functions +is outside the scope of this document and standardization is left for +future work, but use is widely deployed and more information can be +found in platform-specific documentation (e.g., Linux kernel documentation). Platforms that support the BPF Type Format (BTF) support identifying a helper function by a BTF ID encoded in the 'imm' field, where the BTF ID -identifies the helper name and type. +identifies the helper name and type. Further documentation of BTF +is outside the scope of this document and standardization is left for +future work, but use is widely deployed and more information can be +found in platform-specific documentation (e.g., Linux kernel documentation). Program-local functions ~~~~~~~~~~~~~~~~~~~~~~~ Program-local functions are functions exposed by the same BPF program as the -caller, and are referenced by offset from the call instruction, similar to -``JA``. The offset is encoded in the 'imm' field of the call instruction. -An ``EXIT`` within the program-local function will return to the caller. +caller, and are referenced by offset from the instruction following the call +instruction, similar to ``JA``. The offset is encoded in the 'imm' field of +the call instruction. An ``EXIT`` within the program-local function will +return to the caller. Load and store instructions =========================== @@ -537,6 +580,8 @@ For load and store instructions (``LD``, ``LDX``, ``ST``, and ``STX``), the **mode** The mode modifier is one of: + .. table:: Mode modifier + ============= ===== ==================================== ============= mode modifier value description reference ============= ===== ==================================== ============= @@ -551,6 +596,8 @@ For load and store instructions (``LD``, ``LDX``, ``ST``, and ``STX``), the **sz (size)** The size modifier is one of: + .. table:: Size modifier + ==== ===== ===================== size value description ==== ===== ===================== @@ -619,14 +666,16 @@ The 'imm' field is used to encode the actual atomic operation. Simple atomic operation use a subset of the values defined to encode arithmetic operations in the 'imm' field to encode the atomic operation: -======== ===== =========== -imm value description -======== ===== =========== -ADD 0x00 atomic add -OR 0x40 atomic or -AND 0x50 atomic and -XOR 0xa0 atomic xor -======== ===== =========== +.. table:: Simple atomic operations + + ======== ===== =========== + imm value description + ======== ===== =========== + ADD 0x00 atomic add + OR 0x40 atomic or + AND 0x50 atomic and + XOR 0xa0 atomic xor + ======== ===== =========== ``{ATOMIC, W, STX}`` with 'imm' = ADD means:: @@ -640,13 +689,15 @@ XOR 0xa0 atomic xor In addition to the simple atomic operations, there also is a modifier and two complex atomic operations: -=========== ================ =========================== -imm value description -=========== ================ =========================== -FETCH 0x01 modifier: return old value -XCHG 0xe0 | FETCH atomic exchange -CMPXCHG 0xf0 | FETCH atomic compare and exchange -=========== ================ =========================== +.. table:: Complex atomic operations + + =========== ================ =========================== + imm value description + =========== ================ =========================== + FETCH 0x01 modifier: return old value + XCHG 0xe0 | FETCH atomic exchange + CMPXCHG 0xf0 | FETCH atomic compare and exchange + =========== ================ =========================== The ``FETCH`` modifier is optional for simple atomic operations, and always set for the complex atomic operations. If the ``FETCH`` flag @@ -673,17 +724,19 @@ The following table defines a set of ``{IMM, DW, LD}`` instructions with opcode subtypes in the 'src_reg' field, using new terms such as "map" defined further below: -======= ========================================= =========== ============== -src_reg pseudocode imm type dst type -======= ========================================= =========== ============== -0x0 dst = (next_imm << 32) | imm integer integer -0x1 dst = map_by_fd(imm) map fd map -0x2 dst = map_val(map_by_fd(imm)) + next_imm map fd data address -0x3 dst = var_addr(imm) variable id data address -0x4 dst = code_addr(imm) integer code address -0x5 dst = map_by_idx(imm) map index map -0x6 dst = map_val(map_by_idx(imm)) + next_imm map index data address -======= ========================================= =========== ============== +.. table:: 64-bit immediate instructions + + ======= ========================================= =========== ============== + src_reg pseudocode imm type dst type + ======= ========================================= =========== ============== + 0x0 dst = (next_imm << 32) | imm integer integer + 0x1 dst = map_by_fd(imm) map fd map + 0x2 dst = map_val(map_by_fd(imm)) + next_imm map fd data address + 0x3 dst = var_addr(imm) variable id data address + 0x4 dst = code_addr(imm) integer code address + 0x5 dst = map_by_idx(imm) map index map + 0x6 dst = map_val(map_by_idx(imm)) + next_imm map index data address + ======= ========================================= =========== ============== where @@ -725,5 +778,5 @@ carried over from classic BPF. These instructions used an instruction class of ``LD``, a size modifier of ``W``, ``H``, or ``B``, and a mode modifier of ``ABS`` or ``IND``. The 'dst_reg' and 'offset' fields were set to zero, and 'src_reg' was set to zero for ``ABS``. However, these -instructions are deprecated and should no longer be used. All legacy packet +instructions are deprecated and SHOULD no longer be used. All legacy packet access instructions belong to the "packet" conformance group. |