From: Martin KaFai Lau <martin.lau@linux.dev>
To: Andrey Grafin <conquistador@yandex-team.ru>
Cc: andrii@kernel.org, bpf@vger.kernel.org,
Yonghong Song <yonghong.song@linux.dev>
Subject: Re: [PATCH bpf v2] libbpf: Apply map_set_def_max_entries() for inner_maps on creation
Date: Fri, 12 Jan 2024 11:47:48 -0800 [thread overview]
Message-ID: <ee2f4fb3-12b5-41ef-b3bf-b67a73f0105d@linux.dev> (raw)
In-Reply-To: <20240112121051.17325-1-conquistador@yandex-team.ru>
On 1/12/24 4:10 AM, Andrey Grafin wrote:
> This patch allows to create BPF_MAP_TYPE_ARRAY_OF_MAPS and
> BPF_MAP_TYPE_HASH_OF_MAPS with values of BPF_MAP_TYPE_PERF_EVENT_ARRAY.
>
> Previous behaviour created a zero filled btf_map_def for inner maps and
> tried to use it for a map creation but the linux kernel forbids to create
> a BPF_MAP_TYPE_PERF_EVENT_ARRAY map with max_entries=0.
>
> A simple bpf snippet to reproduce:
> struct inner_map {
> __uint(type, BPF_MAP_TYPE_PERF_EVENT_ARRAY);
> __uint(key_size, sizeof(int));
> __uint(value_size, sizeof(u32));
> } inner_map0 SEC(".maps"), inner_map1 SEC(".maps");
>
> struct {
> __uint(type, BPF_MAP_TYPE_ARRAY_OF_MAPS);
> __uint(max_entries, 2);
> __type(key, u32);
> __array(values, struct inner_map);
> } outer_map SEC(".maps") = {
> .values = {&inner_map0, &inner_map1}};
> ...
>
> Previous behaviour:
> # sudo bpftool prog load ./bpf_sample.elf /sys/fs/bpf/test
> libbpf: map 'outer_map': failed to create inner map: -22
> libbpf: map 'outer_map': failed to create: Invalid argument(-22)
> libbpf: failed to load object './bpf_sample.elf'
> Error: failed to load object file
>
> # sudo strace -e bpf bpftool prog load ./bpf_sample.elf /sys/fs/bpf/test
> ...
> bpf(BPF_MAP_CREATE, {map_type=BPF_MAP_TYPE_PERF_EVENT_ARRAY, key_size=4,
> value_size=4, max_entries=16, map_flags=0, inner_map_fd=0,
> map_name="inner_map0", map_ifindex=0, btf_fd=0, btf_key_type_id=0,
> btf_value_type_id=0, btf_vmlinux_value_type_id=0, map_extra=0}, 72) = 4
> bpf(BPF_MAP_CREATE, {map_type=BPF_MAP_TYPE_PERF_EVENT_ARRAY, key_size=4,
> value_size=4, max_entries=16, map_flags=0, inner_map_fd=0,
> map_name="inner_map1", map_ifindex=0, btf_fd=0, btf_key_type_id=0,
> btf_value_type_id=0, btf_vmlinux_value_type_id=0, map_extra=0}, 72) = 5
> bpf(BPF_MAP_CREATE, {map_type=BPF_MAP_TYPE_PERF_EVENT_ARRAY, key_size=4,
> value_size=4, max_entries=0, map_flags=0, inner_map_fd=0,
> map_name="outer_map.inner", map_ifindex=0, btf_fd=0,
> btf_key_type_id=0, btf_value_type_id=0, btf_vmlinux_value_type_id=0,
> map_extra=0}, 72) = -1 EINVAL (Invalid argument)
The change makes sense. Please help to add a real selftest to catch future
regression. I believe it is what Yonghong has already asked in v1. Some of the
existing "progs/*map_in_map*.c" may be a good candidate to add this test case.
pw-bot: cr
prev parent reply other threads:[~2024-01-12 19:47 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-12 12:10 [PATCH bpf v2] libbpf: Apply map_set_def_max_entries() for inner_maps on creation Andrey Grafin
2024-01-12 19:39 ` Yonghong Song
2024-01-12 19:47 ` Martin KaFai Lau [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ee2f4fb3-12b5-41ef-b3bf-b67a73f0105d@linux.dev \
--to=martin.lau@linux.dev \
--cc=andrii@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=conquistador@yandex-team.ru \
--cc=yonghong.song@linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).