Skip to contentSkip to conversation

bpf, docs: Add docs on extended 64-bit immediate instructions

PATCH Applied Dave Thaler 3 participants 3 messages Mar 26, 2023 – Apr 3, 2023

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v4 1/1 bpf, docs: Add docs on extended 64-bit immediate instructions

From: Dave Thaler <[email protected]>

Add docs on extended 64-bit immediate instructions, including six instructions previously undocumented. Include a brief description of maps and variables, as used by those instructions.

V1 -> V2: rebased on top of latest master

V2 -> V3: addressed comments from Alexei

V3 -> V4: addressed comments from David Vernet

---
 Documentation/bpf/instruction-set.rst | 57 +++++++++++++++++++++++----
 Documentation/bpf/linux-notes.rst     | 22 +++++++++++
 2 files changed, 71 insertions(+), 8 deletions(-)
Documentation/bpf/instruction-set.rst +49 -8
index b4464058905..b3efa4b74ec 100644
@@ -401,14 +401,55 @@ and loaded back to ``R0``.
 -----------------------------
 
 Instructions with the ``BPF_IMM`` 'mode' modifier use the wide instruction
-encoding for an extra imm64 value.
-
-There is currently only one such instruction.
-
-``BPF_LD | BPF_DW | BPF_IMM`` means::
-
-  dst = imm64
-
+encoding defined in `Instruction encoding`_, and use the 'src' field of the
+basic instruction to hold an opcode subtype.
+
+The following table defines a set of ``BPF_IMM | BPF_DW | BPF_LD`` instructions
+with opcode subtypes in the 'src' field, using new terms such as "map"
+defined further below:
+
+=========================  ======  ===  =========================================  ===========  ==============
+opcode construction        opcode  src  pseudocode                                 imm type     dst type
+=========================  ======  ===  =========================================  ===========  ==============
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x0  dst = imm64                                integer      integer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x1  dst = map_by_fd(imm)                       map fd       map
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x2  dst = map_val(map_by_fd(imm)) + next_imm   map fd       data pointer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x3  dst = var_addr(imm)                        variable id  data pointer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x4  dst = code_addr(imm)                       integer      code pointer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x5  dst = map_by_idx(imm)                      map index    map
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x6  dst = map_val(map_by_idx(imm)) + next_imm  map index    data pointer
+=========================  ======  ===  =========================================  ===========  ==============
+
+where
+
+* map_by_fd(imm) means to convert a 32-bit POSIX file descriptor into an address of a map (see `Maps`_)
+* map_by_idx(imm) means to convert a 32-bit index into an address of a map
+* map_val(map) gets the address of the first value in a given map
+* var_addr(imm) gets the address of a platform variable (see `Platform Variables`_) with a given id
+* code_addr(imm) gets the address of the instruction at a specified relative offset in number of (64-bit) instructions
+* the 'imm type' can be used by disassemblers for display
+* the 'dst type' can be used for verification and JIT compilation purposes
+
+Maps
+~~~~
+
+Maps are shared memory regions accessible by eBPF programs on some platforms.
+A map can have various semantics as defined in a separate document, and may or may not have a single
+contiguous memory region, but the 'map_val(map)' is currently only defined for maps that do have a single
+contiguous memory region.
+
+Each map can have a POSIX file descriptor (fd) if supported by the platform,
+where 'map_by_fd(imm)' means to get the map with the specified file descriptor.
+Each BPF program can also be defined to use a set of maps associated with the program
+at load time, and 'map_by_idx(imm)' means to get the map with the given index in the set
+associated with the BPF program containing the instruction.
+
+Platform Variables
+~~~~~~~~~~~~~~~~~~
+
+Platform variables are memory regions, identified by integer ids, exposed by the runtime and accessible by BPF programs on
+some platforms.  The 'var_addr(imm)' operation means to get the address of the memory region
+identified by the given id.
 
 Legacy BPF Packet access instructions
 -------------------------------------
Documentation/bpf/linux-notes.rst +22 -0
index f43b9c797bc..508d009d3be 100644
@@ -20,6 +20,28 @@ integer would be read from a specified register, is not currently supported
 by the verifier.  Any programs with this instruction will fail to load
 until such support is added.
 
+Maps
+====
+
+Linux only supports the 'map_val(map)' operation on array maps with a single element.
+
+Linux uses an fd_array to store maps associated with a BPF program. Thus,
+map_by_idx(imm) uses the fd at that index in the array.
+
+Variables
+=========
+
+The following 64-bit immediate instruction specifies that a variable address,
+which corresponds to some integer stored in the 'imm' field, should be loaded:
+
+=========================  ======  ===  =========================================  ===========  ==============
+opcode construction        opcode  src  pseudocode                                 imm type     dst type
+=========================  ======  ===  =========================================  ===========  ==============
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x3  dst = var_addr(imm)                        variable id  data pointer
+=========================  ======  ===  =========================================  ===========  ==============
+
+On Linux, this integer is a BTF ID.
+
 Legacy BPF Packet access instructions
 =====================================
 
Signature · 1 line
2.33.4
#1 lore raw
Quoted 47 lines · Dave Thaler

From: Dave Thaler <[email protected]>

Add docs on extended 64-bit immediate instructions, including six instructions previously undocumented. Include a brief description of maps and variables, as used by those instructions.

V1 -> V2: rebased on top of latest master

V2 -> V3: addressed comments from Alexei

V3 -> V4: addressed comments from David Vernet

---
 Documentation/bpf/instruction-set.rst | 57 +++++++++++++++++++++++----
 Documentation/bpf/linux-notes.rst     | 22 +++++++++++
 2 files changed, 71 insertions(+), 8 deletions(-)
Documentation/bpf/instruction-set.rst +12 -8
index b4464058905..b3efa4b74ec 100644
@@ -401,14 +401,55 @@ and loaded back to ``R0``.
 -----------------------------
 
 Instructions with the ``BPF_IMM`` 'mode' modifier use the wide instruction
-encoding for an extra imm64 value.
-
-There is currently only one such instruction.
-
-``BPF_LD | BPF_DW | BPF_IMM`` means::
-
-  dst = imm64
-
+encoding defined in `Instruction encoding`_, and use the 'src' field of the
+basic instruction to hold an opcode subtype.
+
+The following table defines a set of ``BPF_IMM | BPF_DW | BPF_LD`` instructions
+with opcode subtypes in the 'src' field, using new terms such as "map"
+defined further below:
+
+=========================  ======  ===  =========================================  ===========  ==============
+opcode construction        opcode  src  pseudocode                                 imm type     dst type
+=========================  ======  ===  =========================================  ===========  ==============
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x0  dst = imm64                                integer      integer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x1  dst = map_by_fd(imm)                       map fd       map

Documentation/bpf/instruction-set.rst and loaded back to ``R0``. · line 425

Quoted 10 lines
Documentation/bpf/instruction-set.rst +10 -0
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x2  dst = map_val(map_by_fd(imm)) + next_imm   map fd       data pointer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x3  dst = var_addr(imm)                        variable id  data pointer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x4  dst = code_addr(imm)                       integer      code pointer
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x5  dst = map_by_idx(imm)                      map index    map
+BPF_IMM | BPF_DW | BPF_LD  0x18    0x6  dst = map_val(map_by_idx(imm)) + next_imm  map index    data pointer
+=========================  ======  ===  =========================================  ===========  ==============
+
+where
+
+* map_by_fd(imm) means to convert a 32-bit POSIX file descriptor into an address of a map (see `Maps`_)

I removed the screaming "POSIX" word, since "file descriptor" is descriptive enough.

Quoted 4 lines
Documentation/bpf/instruction-set.rst +4 -0
+* map_by_idx(imm) means to convert a 32-bit index into an address of a map
+* map_val(map) gets the address of the first value in a given map
+* var_addr(imm) gets the address of a platform variable (see `Platform Variables`_) with a given id
+* code_addr(imm) gets the address of the instruction at a specified relative offset in number of (64-bit) instructions

Documentation/bpf/instruction-set.rst and loaded back to ``R0``. · line 439

Quoted 10 lines
Documentation/bpf/instruction-set.rst +10 -0
+* the 'imm type' can be used by disassemblers for display
+* the 'dst type' can be used for verification and JIT compilation purposes
+
+Maps
+~~~~
+
+Maps are shared memory regions accessible by eBPF programs on some platforms.
+A map can have various semantics as defined in a separate document, and may or may not have a single
+contiguous memory region, but the 'map_val(map)' is currently only defined for maps that do have a single
+contiguous memory region.

Reformatted these sections to fit 80 char.
And applied.

#2 lore raw
In reply to Dave Thaler
This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Alexei Starovoitov <[email protected]>:

Quoted 9 lines · On Sun, 26 Mar 2023 05:49:46 +0000 you wrote:

From: Dave Thaler <[email protected]>

Add docs on extended 64-bit immediate instructions, including six instructions previously undocumented. Include a brief description of maps and variables, as used by those instructions.

V1 -> V2: rebased on top of latest master

[...]

Here is the summary with links:
- [bpf-next,v4] bpf, docs: Add docs on extended 64-bit immediate instructions

    https://git.kernel.org/bpf/bpf-next/c/16b7c970cc81

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html