Based on:
Subj: libbpf: Helpers to access XDP hints based on BTF definitions
https://lore.kernel.org/all/20210803010331.39453-15-ederson.desouza@intel.com/
From: Ederson de Souza <ederson.desouza@intel.com>
Ederson says:
A new set of functions to help get the BTF definition of XDP hints
structure and get the information based on it.
`xsk_umem__btf_id` helps retrieve the BTF id of XDP metadata.
`xsk_btf__init` sets up a context based on the BTF, including a hashmap,
so that subsequent queries are faster.
`xsk_btf__read` returns a pointer to the position in the XDP metadata
containing a given field.
`xsk_btf__has_field` checks the presence of a field in the BTF.
`xsk_btf__free` frees up the context.
Besides those, a macro `XSK_BTF_READ_INTO` acts as a convenient helper
to read the field contents into a given variable.
Note that currently, the hashmap used to speed-up offset location into
the BTF doesn't use the field name as a string as key to the hashmap. It
directly uses the pointer value instead, as it is expected that most of
time, field names will be addressed by a shared constant string residing
on read-only memory, thus saving some time. If this assumption is not
entirely true, this optimisation needs to be rethought (or discarded
altogether).
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
This is a copy of the kernels/libbpf hashmap
tools/lib/bpf/hashmap.{h,c}
The plan is to prototype an AF_XDP/xsk userspace API for accessing
BTF information, that should be moved to libbpf (or libxdp). Thus,
this hashmap code will become avail if successful.
Original kernel commit 553db8ba73df ("libbpf: add resizable non-thread
safe internal hashmap") (Author: Andrii Nakryiko). Thus, giving Andrii
author credit in this git commit.
Andrii Nakryiko said:
libbpf: add resizable non-thread safe internal hashmap
There is a need for fast point lookups inside libbpf for multiple use
cases (e.g., name resolution for BTF-to-C conversion, by-name lookups in
BTF for upcoming BPF CO-RE relocation support, etc). This patch
implements simple resizable non-thread safe hashmap using single linked
list chains.
Four different insert strategies are supported:
- HASHMAP_ADD - only add key/value if key doesn't exist yet;
- HASHMAP_SET - add key/value pair if key doesn't exist yet; otherwise,
update value;
- HASHMAP_UPDATE - update value, if key already exists; otherwise, do
nothing and return -ENOENT;
- HASHMAP_APPEND - always add key/value pair, even if key already exists.
This turns hashmap into a multimap by allowing multiple values to be
associated with the same key. Most useful read API for such hashmap is
hashmap__for_each_key_entry() iteration. If hashmap__find() is still
used, it will return last inserted key/value entry (first in a bucket
chain).
For HASHMAP_SET and HASHMAP_UPDATE, old key/value pair is returned, so
that calling code can handle proper memory management, if necessary.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Don't allow loading the default xsk BPF-prog if not specifying
any BPF-prog. We have a need for our own BPf-prog.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
To verify the contents of the incomming packets add a
function print_pkt_info() that decode part of the packet
headers and print IP-header src+dst (both IPv4 and IPv6).
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
The default frame_headroom seems to be zero as
define XSK_UMEM__DEFAULT_FRAME_HEADROOM is zero.
This patch does explicit xsk_umem_config setup.
Comment taken from kernel/samples/bpf/xdpsock_user.c.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
This only works in skb-mode, other modes fail the bpf_xdp_adjust_meta
call and result in XDP_ABORTED. This only occurred on mlx5.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
BPF program title (bpf_program__title) is getting deprecated
in libbpf use bpf_program__section_name() instead.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
In-practice only 'xdp_stats_map' definition gets associated BTF,
as it uses the proper '__type' construct.
The AF_XDP map BPF_MAP_TYPE_XSKMAP doesn't know about BTF yet.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
When not loading an XDP-prog the XSK library bits will load a
default XDP-prog.
In our use-case we always want to load our XDP-prog, which is
stored in af_xdp_kern.o.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Something in fishy as it seems this BPF-object is not getting used.
Use old maps SEC defines until figuring out what is wrong,
but keep new dot-maps defines.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
The BPF-helper is now defined in bpf_helper_defs.h header:
lib/libbpf-install/usr/include/bpf/bpf_helper_defs.h
Update and keep comment as a help for people doing kernel
devel on BPF and want to test BPF-helpers not yet upstream.
Accepted in kernel v5.12
- commit 34b2021cc616 ("bpf: Add BPF-helper for MTU checking")
- https://git.kernel.org/torvalds/c/34b2021cc616
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
This USER_TARGETS_OBJS_DEPS allows makefiles to define
dependencies for USER_TARGETS_OBJS targets.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
This allows including Makefile to define USER_TARGETS_OBJS
that contain code that USER_TARGETS depend on and needs
to be compiled as objects and later linked with.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
The variable substitution used to parse the clang version was a bashism, so
it broke on Debian/Ubuntu. So let's just use sed instead.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Seems some distributions will change the clang version output to be
something like 'Ubuntu clang version
12.0.1-++20210918042554+fed41342a82f-1~exp1~20210918143322.141'. Fix the
version parsing in configure to not barf on such weird version strings.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Add rewriting of ICMP headers to nat64. This is specified in RFC6145, and
the implementation here follows that. The support is only partial, in
particular, in that the payload of ICMP error messages is not rewritten,
even though the RFC specifies that they should be.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
IPv6 doesn't support fragmentation, so make sure IPv4 packets are not
fragmented in-flight either.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
We insert a v4-via-v6 route on the interface to direct packets to the v4
subnet to the right interface, where it will be rewritten by the BPF
program. We also create a fake neighbour entry so the kernel won't do
neighbour resolution when sending the pre-rewrite packet. The egress BPF
program will use bpf_redirect_neigh() to do proper neighbour resolution for
the actual destination after rewriting the packet.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Turns out the atomics just needed to operate on a 64-bit variable, we
didn't actually need a newer compiler.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Endianness fixes for IP header length, need to also rewrite Ethernet
protocol. Also use direct packet access instead of skb_store_bytes().
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
This adds an initial version of a NAT64 translator in BPF. It compiles and
loads, but doesn't actually appear to work yet.
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
