bpf, docs: Move legacy packet instructions to a separate file

Move legacy packet instructions to a separate file.

Signed-off-by: Dave Thaler <dthaler@microsoft.com>
Link: https://lore.kernel.org/r/20220927185958.14995-1-dthaler1968@googlemail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

2 files changed, 68 insertions, 35 deletions

diff --git a/Documentation/bpf/instruction-set.rst b/Documentation/bpf/instruction-set.rst
index 1b0e6711dec971..352f25a1eb17e7 100644
--- a/Documentation/bpf/instruction-set.rst
+++ b/Documentation/bpf/instruction-set.rst
@@ -282,8 +282,6 @@ arithmetic operations in the imm field to encode the atomic operation:
 
   *(u64 *)(dst_reg + off16) += src_reg
 
-``BPF_XADD`` is a deprecated name for ``BPF_ATOMIC | BPF_ADD``.
-
 In addition to the simple atomic operations, there also is a modifier and
 two complex atomic operations:
 
@@ -331,36 +329,6 @@ There is currently only one such instruction.
 Legacy BPF Packet access instructions
 -------------------------------------
 
-eBPF has special instructions for access to packet data that have been
-carried over from classic BPF to retain the performance of legacy socket
-filters running in the eBPF interpreter.
-
-The instructions come in two forms: ``BPF_ABS | <size> | BPF_LD`` and
-``BPF_IND | <size> | BPF_LD``.
-
-These instructions are used to access packet data and can only be used when
-the program context is a pointer to networking packet.  ``BPF_ABS``
-accesses packet data at an absolute offset specified by the immediate data
-and ``BPF_IND`` access packet data at an offset that includes the value of
-a register in addition to the immediate data.
-
-These instructions have seven implicit operands:
-
- * Register R6 is an implicit input that must contain pointer to a
-   struct sk_buff.
- * Register R0 is an implicit output which contains the data fetched from
-   the packet.
- * Registers R1-R5 are scratch registers that are clobbered after a call to
-   ``BPF_ABS | BPF_LD`` or ``BPF_IND | BPF_LD`` instructions.
-
-These instructions have an implicit program exit condition as well. When an
-eBPF program is trying to access the data beyond the packet boundary, the
-program execution will be aborted.
-
-``BPF_ABS | BPF_W | BPF_LD`` means::
-
-  R0 = ntohl(*(u32 *) (((struct sk_buff *) R6)->data + imm32))
-
-``BPF_IND | BPF_W | BPF_LD`` means::
-
-  R0 = ntohl(*(u32 *) (((struct sk_buff *) R6)->data + src_reg + imm32))
+eBPF previously introduced special instructions for access to packet data that were
+carried over from classic BPF. However, these instructions are
+deprecated and should no longer be used.
diff --git a/Documentation/bpf/linux-notes.rst b/Documentation/bpf/linux-notes.rst
new file mode 100644
index 00000000000000..93c01386d92c7b
--- /dev/null
+++ b/Documentation/bpf/linux-notes.rst
@@ -0,0 +1,65 @@
+.. contents::
+.. sectnum::
+
+==========================
+Linux implementation notes
+==========================
+
+This document provides more details specific to the Linux kernel implementation of the eBPF instruction set.
+
+Legacy BPF Packet access instructions
+=====================================
+
+As mentioned in the `ISA standard documentation <instruction-set.rst#legacy-bpf-packet-access-instructions>`_,
+Linux has special eBPF instructions for access to packet data that have been
+carried over from classic BPF to retain the performance of legacy socket
+filters running in the eBPF interpreter.
+
+The instructions come in two forms: ``BPF_ABS | <size> | BPF_LD`` and
+``BPF_IND | <size> | BPF_LD``.
+
+These instructions are used to access packet data and can only be used when
+the program context is a pointer to a networking packet.  ``BPF_ABS``
+accesses packet data at an absolute offset specified by the immediate data
+and ``BPF_IND`` access packet data at an offset that includes the value of
+a register in addition to the immediate data.
+
+These instructions have seven implicit operands:
+
+* Register R6 is an implicit input that must contain a pointer to a
+  struct sk_buff.
+* Register R0 is an implicit output which contains the data fetched from
+  the packet.
+* Registers R1-R5 are scratch registers that are clobbered by the
+  instruction.
+
+These instructions have an implicit program exit condition as well. If an
+eBPF program attempts access data beyond the packet boundary, the
+program execution will be aborted.
+
+``BPF_ABS | BPF_W | BPF_LD`` (0x20) means::
+
+  R0 = ntohl(*(u32 *) ((struct sk_buff *) R6->data + imm))
+
+where ``ntohl()`` converts a 32-bit value from network byte order to host byte order.
+
+``BPF_IND | BPF_W | BPF_LD`` (0x40) means::
+
+  R0 = ntohl(*(u32 *) ((struct sk_buff *) R6->data + src + imm))
+
+Appendix
+========
+
+For reference, the following table lists legacy Linux-specific opcodes in order by value.
+
+======  ====  ===================================================  =============
+opcode  imm   description                                          reference
+======  ====  ===================================================  =============
+0x20    any   dst = ntohl(\*(uint32_t \*)(R6->data + imm))         `Legacy BPF Packet access instructions`_
+0x28    any   dst = ntohs(\*(uint16_t \*)(R6->data + imm))         `Legacy BPF Packet access instructions`_
+0x30    any   dst = (\*(uint8_t \*)(R6->data + imm))               `Legacy BPF Packet access instructions`_
+0x38    any   dst = ntohll(\*(uint64_t \*)(R6->data + imm))        `Legacy BPF Packet access instructions`_
+0x40    any   dst = ntohl(\*(uint32_t \*)(R6->data + src + imm))   `Legacy BPF Packet access instructions`_
+0x48    any   dst = ntohs(\*(uint16_t \*)(R6->data + src + imm))   `Legacy BPF Packet access instructions`_
+0x50    any   dst = \*(uint8_t \*)(R6->data + src + imm))          `Legacy BPF Packet access instructions`_
+0x58    any   dst = ntohll(\*(uint64_t \*)(R6->data + src + imm))  `Legacy BPF Packet access instructions`_