From: "Emil Tsalapatis" <emil@etsalapatis.com>
To: "Jiayuan Chen" <jiayuan.chen@linux.dev>, <bpf@vger.kernel.org>
Cc: "Alexei Starovoitov" <ast@kernel.org>,
"Daniel Borkmann" <daniel@iogearbox.net>,
"Andrii Nakryiko" <andrii@kernel.org>,
"Eduard Zingerman" <eddyz87@gmail.com>,
"Kumar Kartikeya Dwivedi" <memxor@gmail.com>,
"Martin KaFai Lau" <martin.lau@linux.dev>,
"Song Liu" <song@kernel.org>,
"Yonghong Song" <yonghong.song@linux.dev>,
"Jiri Olsa" <jolsa@kernel.org>,
"Emil Tsalapatis" <emil@etsalapatis.com>,
"Shuah Khan" <shuah@kernel.org>,
"Woojin Ji" <random6.xyz@gmail.com>,
"Puranjay Mohan" <puranjay@kernel.org>,
"Saket Kumar Bhaskar" <skb99@linux.ibm.com>,
"Ihor Solodrai" <ihor.solodrai@linux.dev>,
<linux-kernel@vger.kernel.org>, <linux-kselftest@vger.kernel.org>
Subject: Re: [PATCH bpf-next v2 4/4] selftests/bpf: Add tests for memory usage for arena
Date: Fri, 17 Jul 2026 19:36:25 -0400 [thread overview]
Message-ID: <DK18OQD8NUIY.DCLKET294PNM@etsalapatis.com> (raw)
In-Reply-To: <20260717114117.350851-5-jiayuan.chen@linux.dev>
On Fri Jul 17, 2026 at 7:37 AM EDT, Jiayuan Chen wrote:
> Allocate and free arena pages, both from BPF and via user-space fault-in,
> and check that the map's memlock in fdinfo tracks the number of pages that
> are actually populated.
>
> Like the other arena tests it runs serially.
>
> test:
> ./test_progs -a arena_mem_usage
> #5 arena_mem_usage:OK
> Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED
>
> Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
The Sashiko nit is worth addressing, but overall looks fine.
> ---
> .../bpf/prog_tests/arena_mem_usage.c | 122 ++++++++++++++++++
> .../selftests/bpf/progs/arena_mem_usage.c | 40 ++++++
> 2 files changed, 162 insertions(+)
> create mode 100644 tools/testing/selftests/bpf/prog_tests/arena_mem_usage.c
> create mode 100644 tools/testing/selftests/bpf/progs/arena_mem_usage.c
>
> diff --git a/tools/testing/selftests/bpf/prog_tests/arena_mem_usage.c b/tools/testing/selftests/bpf/prog_tests/arena_mem_usage.c
> new file mode 100644
> index 000000000000..14c2d1a1d673
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/prog_tests/arena_mem_usage.c
> @@ -0,0 +1,122 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <test_progs.h>
> +#include <sys/user.h>
> +#ifndef PAGE_SIZE /* on some archs it comes in sys/user.h */
> +#include <unistd.h>
> +#define PAGE_SIZE getpagesize()
> +#endif
> +
> +#include "arena_mem_usage.skel.h"
> +
> +/*
> + * arena_map_mem_usage() is surfaced to user space through the map's
> + * /proc/<pid>/fdinfo/<fd> "memlock:" line (the same value bpftool map show
> + * prints). Read it directly so the test has no external dependency.
> + */
> +static long map_memlock(int map_fd)
> +{
> + char path[64], line[128];
> + long memlock = -1;
> + FILE *f;
> +
> + snprintf(path, sizeof(path), "/proc/self/fdinfo/%d", map_fd);
> + f = fopen(path, "r");
> + if (!ASSERT_OK_PTR(f, "open_fdinfo"))
> + return -1;
> + while (fgets(line, sizeof(line), f)) {
> + if (sscanf(line, "memlock:\t%ld", &memlock) == 1)
> + break;
> + }
> + fclose(f);
> + ASSERT_NEQ(memlock, -1, "parse_memlock");
> + return memlock;
> +}
> +
> +static int run(struct bpf_program *prog, const char *name)
> +{
> + LIBBPF_OPTS(bpf_test_run_opts, opts);
> + int err = bpf_prog_test_run_opts(bpf_program__fd(prog), &opts);
> +
> + if (!ASSERT_OK(err, name))
> + return -1;
> + if (!ASSERT_OK(opts.retval, name))
> + return -1;
> + return 0;
> +}
> +
> +void serial_test_arena_mem_usage(void)
> +{
> + struct arena_mem_usage *skel;
> + const long ps = PAGE_SIZE;
> + char *base;
> + size_t sz;
> + int fd, i;
> +
> + skel = arena_mem_usage__open_and_load();
> + if (!ASSERT_OK_PTR(skel, "open_load"))
> + return;
> + fd = bpf_map__fd(skel->maps.arena);
> +
> + /* Fresh arena: no data pages, and the scratch page is not counted. */
> + ASSERT_EQ(map_memlock(fd), 0, "initial");
> +
> + /* BPF-side allocation of 17 pages. */
> + skel->bss->alloc_cnt = 17;
> + if (run(skel->progs.alloc, "alloc"))
> + goto out;
> + /*
> + * A NULL ptr means bpf_arena_alloc_pages() itself failed (e.g. the host
> + * is under memory pressure), not a miscount -- flag it distinctly so a
> + * red CI run is not mistaken for a counting bug.
> + */
> + if (!ASSERT_OK_PTR(skel->bss->ptr, "arena_alloc_pages"))
> + goto out;
> + ASSERT_EQ(map_memlock(fd), 17 * ps, "after_alloc");
> +
> + /* Free a single page (arena_free_pages page_cnt==1 path). */
> + skel->bss->free_byte_off = 0;
> + skel->bss->free_cnt = 1;
> + if (run(skel->progs.free_pages, "free_one"))
> + goto out;
> + ASSERT_EQ(map_memlock(fd), 16 * ps, "after_free_one");
> +
> + /* Free ten pages in one call (bulk path); only the freed pages count. */
> + skel->bss->free_byte_off = 1 * ps;
> + skel->bss->free_cnt = 10;
> + if (run(skel->progs.free_pages, "free_bulk"))
> + goto out;
> + ASSERT_EQ(map_memlock(fd), 6 * ps, "after_free_bulk");
> +
> + /* Free the remaining six -> arena empty again. */
> + skel->bss->free_byte_off = 11 * ps;
> + skel->bss->free_cnt = 6;
> + if (run(skel->progs.free_pages, "free_rest"))
> + goto out;
> + ASSERT_EQ(map_memlock(fd), 0, "after_free_rest");
> +
> + /*
> + * User-space fault-in: touching unallocated arena pages allocates them
> + * through arena_vm_fault(). libbpf mmap()s the arena at map_extra during
> + * load, so bpf_map__initial_value() hands back that base.
> + */
> + base = bpf_map__initial_value(skel->maps.arena, &sz);
> + if (!ASSERT_OK_PTR(base, "arena_base"))
> + goto out;
> + for (i = 0; i < 8; i++)
> + base[i * ps] = 1;
> + ASSERT_EQ(map_memlock(fd), 8 * ps, "after_faultin");
> +
> + /*
> + * Free the faulted-in pages from BPF. They are mapped into the user vma
> + * (elevated refcount), so this also exercises the zap path.
> + */
> + skel->bss->ptr = base;
> + skel->bss->free_byte_off = 0;
> + skel->bss->free_cnt = 8;
> + if (run(skel->progs.free_pages, "free_faulted"))
> + goto out;
> + ASSERT_EQ(map_memlock(fd), 0, "after_free_faulted");
> +out:
> + arena_mem_usage__destroy(skel);
> +}
> diff --git a/tools/testing/selftests/bpf/progs/arena_mem_usage.c b/tools/testing/selftests/bpf/progs/arena_mem_usage.c
> new file mode 100644
> index 000000000000..455ecd669a5a
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/progs/arena_mem_usage.c
> @@ -0,0 +1,40 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <vmlinux.h>
> +#include <bpf/bpf_helpers.h>
> +#include "bpf_arena_common.h"
> +
> +struct {
> + __uint(type, BPF_MAP_TYPE_ARENA);
> + __uint(map_flags, BPF_F_MMAPABLE);
> + __uint(max_entries, 1000); /* number of pages */
> +#ifdef __TARGET_ARCH_arm64
> + __ulong(map_extra, 0x1ull << 32); /* start of mmap() region */
> +#else
> + __ulong(map_extra, 0x1ull << 44); /* start of mmap() region */
> +#endif
> +} arena SEC(".maps");
> +
> +void __arena *ptr;
> +int alloc_cnt; /* in: pages to allocate */
> +long free_byte_off; /* in: byte offset within ptr to start freeing */
> +int free_cnt; /* in: pages to free */
> +
> +SEC("syscall")
> +int alloc(void *ctx)
> +{
> + ptr = bpf_arena_alloc_pages(&arena, NULL, alloc_cnt, NUMA_NO_NODE, 0);
> + /* Success/failure is checked from user space via skel->bss->ptr. */
> + return 0;
> +}
> +
> +SEC("syscall")
> +int free_pages(void *ctx)
> +{
> + if (!ptr)
> + return 1;
> + bpf_arena_free_pages(&arena, (char __arena *)ptr + free_byte_off, free_cnt);
> + return 0;
> +}
> +
> +char _license[] SEC("license") = "GPL";
prev parent reply other threads:[~2026-07-17 23:36 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-17 11:37 [PATCH bpf-next v2 0/4] bpf: Add memory usage for arena and selftest Jiayuan Chen
2026-07-17 11:37 ` [PATCH bpf-next v2 1/4] bpf: Pass arena instead of scratch_page to the pte callbacks Jiayuan Chen
2026-07-17 23:32 ` Emil Tsalapatis
2026-07-17 11:37 ` [PATCH bpf-next v2 2/4] bpf: Add memory usage for arena Jiayuan Chen
2026-07-17 23:33 ` Emil Tsalapatis
2026-07-17 11:37 ` [PATCH bpf-next v2 3/4] selftests/bpf: Run arena tests serially Jiayuan Chen
2026-07-17 23:34 ` Emil Tsalapatis
2026-07-17 11:37 ` [PATCH bpf-next v2 4/4] selftests/bpf: Add tests for memory usage for arena Jiayuan Chen
2026-07-17 23:36 ` Emil Tsalapatis [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=DK18OQD8NUIY.DCLKET294PNM@etsalapatis.com \
--to=emil@etsalapatis.com \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=eddyz87@gmail.com \
--cc=ihor.solodrai@linux.dev \
--cc=jiayuan.chen@linux.dev \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=martin.lau@linux.dev \
--cc=memxor@gmail.com \
--cc=puranjay@kernel.org \
--cc=random6.xyz@gmail.com \
--cc=shuah@kernel.org \
--cc=skb99@linux.ibm.com \
--cc=song@kernel.org \
--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