On Wed, Aug 21, 2024 at 5:30 PM Yonghong Song <yonghong.song@xxxxxxxxx> wrote:
>
>
> On 8/19/24 3:30 PM, Will Hawkins wrote:
> > Make the values of the symbolic constants that define the valid linkages
> > for functions and variables explicit.
> >
> > Signed-off-by: Will Hawkins <hawkinsw@xxxxxx>
> > ---
> > Documentation/bpf/btf.rst | 44 +++++++++++++++++++++++++++++++++++----
> > 1 file changed, 40 insertions(+), 4 deletions(-)
> >
> > diff --git a/Documentation/bpf/btf.rst b/Documentation/bpf/btf.rst
> > index 257a7e1cdf5d..cce03f1e552a 100644
> > --- a/Documentation/bpf/btf.rst
> > +++ b/Documentation/bpf/btf.rst
> > @@ -368,7 +368,7 @@ No additional type data follow ``btf_type``.
> > * ``info.kind_flag``: 0
> > * ``info.kind``: BTF_KIND_FUNC
> > * ``info.vlen``: linkage information (BTF_FUNC_STATIC, BTF_FUNC_GLOBAL
> > - or BTF_FUNC_EXTERN)
> > + or BTF_FUNC_EXTERN - see :ref:`BTF_Function_Linkage_Constants`)
> > * ``type``: a BTF_KIND_FUNC_PROTO type
> >
> > No additional type data follow ``btf_type``.
> > @@ -424,9 +424,9 @@ following data::
> > __u32 linkage;
> > };
> >
> > -``struct btf_var`` encoding:
> > - * ``linkage``: currently only static variable 0, or globally allocated
> > - variable in ELF sections 1
> > +``btf_var.linkage`` may take the values: BTF_VAR_STATIC (for a static variable),
> > +or BTF_VAR_GLOBAL_ALLOCATED (for a globally allocated variable stored in ELF sections 1) -
>
> Let us remove the above '1', just say '(... stored in explicit ELF sections)'.
Great! I only kept that because it was there in the existing documentation!
>
> Actually, for btf_var linkage, we actually have 3 values. See
>
> enum {
> BTF_VAR_STATIC = 0,
> BTF_VAR_GLOBAL_ALLOCATED = 1,
> BTF_VAR_GLOBAL_EXTERN = 2,
> };
>
> See https://github.com/torvalds/linux/blob/master/include/uapi/linux/btf.h#L150-L154
>
Great! I will make that update!
> Similar to BTF_VAR_GLOBAL_ALLOCATED, BTF_VAR_GLOBAL_EXTERN is encoded in datasec only
> if the variable is stored in explicit ELF sections.
>
> Since you are touching this doc, could you add BTF_VAR_GLOBAL_EXTERN as well?
>
Of course!
> > +see :ref:`BTF_Var_Linkage_Constants`.
> >
> > Not all type of global variables are supported by LLVM at this point.
> > The following is currently available:
> > @@ -549,6 +549,42 @@ The ``btf_enum64`` encoding:
> > If the original enum value is signed and the size is less than 8,
> > that value will be sign extended into 8 bytes.
> >
> > +2.3 Constant Values
> > +-------------------
> > +
> > +.. _BTF_Function_Linkage_Constants:
> > +
> > +2.3.1 Function Linkage Constant Values
> > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > +.. list-table::
> > + :widths: 1 1
> > + :header-rows: 1
> > +
> > + * - Name
> > + - Value
> > + * - ``BTF_FUNC_STATIC``
> > + - ``0``
> > + * - ``BTF_FUNC_GLOBAL``
> > + - ``1``
> > + * - ``BTF_FUNC_EXTERN``
> > + - ``2``
> > +
> > +.. _BTF_Var_Linkage_Constants:
> > +
> > +2.3.2 Variable Linkage Constant Values
> > +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > +.. list-table::
> > + :widths: 1 1
> > + :header-rows: 1
> > +
> > + * - Name
> > + - Value
> > + * - ``BTF_VAR_STATIC``
> > + - ``0``
> > + * - ``BTF_VAR_GLOBAL_ALLOCATED``
> > + - ``1``
> > +
> > +
>
> Form the above, could you use similar format as in Documentation/bpf/standardization/instruction-set.rst? For example,
>
> .. 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`_
> ===== ===== =============================== ===================================
>
> I would like we have consistant table presentation between instruction set and btf.
>
Absolutely. I volunteered to edit the document that will ultimately
turn into the IETF spec draft for BTF and this is the first fix that
jumped out to me when I was proofreading.
Thank you for the review! I will have a v2 of the patch in just a few minutes!
Thank you, again!
Will
> > 3. BTF Kernel API
> > =================
> >
