* [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes
@ 2026-08-03 0:18 Changwoo Min
2026-08-13 4:38 ` Kumar Kartikeya Dwivedi
2026-08-17 9:05 ` Daniel Borkmann
0 siblings, 2 replies; 6+ messages in thread
From: Changwoo Min @ 2026-08-03 0:18 UTC (permalink / raw)
To: Alexei Starovoitov, Daniel Borkmann, Andrii Nakryiko,
Eduard Zingerman, Kumar Kartikeya Dwivedi, Shuah Khan,
Emil Tsalapatis
Cc: Martin KaFai Lau, Song Liu, Yonghong Song, Jiri Olsa, bpf,
sched-ext, linux-kselftest, linux-kernel, changwoo, kernel-dev
bpf_arena_spin_lock.h defines its 64KB qnodes array in the header, so
every translation unit including it emits a copy. __weak makes them all
resolve to one instance, but bpftool gen object merges only the symbols
and concatenates each input's .addr_space.1 bytes, leaving the surplus
copies unreferenced in the linked object.
libarena links ten such units, so nine copies were dead weight (bytes):
object before after
-----------------------------------------------------
.addr_space.1 in libarena.bpf.o 676200 86376
libarena.skel.h 2100123 892371
libarena_asan.skel.h 2641124 1466477
Declare qnodes in the header and let each program define it once:
libarena in src/common.bpf.c, and the arena_spin_lock test beside the
lock it guards.
Tested with test_progs -t arena_spin_lock and -t libarena.
Signed-off-by: Changwoo Min <changwoo@igalia.com>
---
.../selftests/bpf/libarena/include/bpf_arena_spin_lock.h | 7 +------
tools/testing/selftests/bpf/libarena/src/common.bpf.c | 7 +++++++
tools/testing/selftests/bpf/progs/arena_spin_lock.c | 7 +++++++
3 files changed, 15 insertions(+), 6 deletions(-)
diff --git a/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h b/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h
index ae6b72d15bb6..71d9db610263 100644
--- a/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h
+++ b/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h
@@ -103,12 +103,7 @@ struct arena_qnode {
#define _Q_LOCKED_VAL (1U << _Q_LOCKED_OFFSET)
#define _Q_PENDING_VAL (1U << _Q_PENDING_OFFSET)
-/*
- * The qnodes are marked __weak so we can define them in the header
- * while still ensuring all compilation units use the same struct
- * instance.
- */
-struct arena_qnode __weak __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
+extern struct arena_qnode __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
static inline u32 encode_tail(int cpu, int idx)
{
diff --git a/tools/testing/selftests/bpf/libarena/src/common.bpf.c b/tools/testing/selftests/bpf/libarena/src/common.bpf.c
index 50be57213dfb..06481f2a2892 100644
--- a/tools/testing/selftests/bpf/libarena/src/common.bpf.c
+++ b/tools/testing/selftests/bpf/libarena/src/common.bpf.c
@@ -8,6 +8,13 @@ const volatile u32 zero = 0;
struct buddy __arena buddy;
+/*
+ * Storage for the queue nodes declared by bpf_arena_spin_lock.h. Each program
+ * linking the arena spinlock provides exactly one definition, so that the array
+ * is emitted once rather than once per translation unit.
+ */
+struct arena_qnode __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
+
int arena_fls(__u64 word)
{
if (!word)
diff --git a/tools/testing/selftests/bpf/progs/arena_spin_lock.c b/tools/testing/selftests/bpf/progs/arena_spin_lock.c
index cf7cda79c16c..92e75ec3844c 100644
--- a/tools/testing/selftests/bpf/progs/arena_spin_lock.c
+++ b/tools/testing/selftests/bpf/progs/arena_spin_lock.c
@@ -23,6 +23,13 @@ int cs_count;
#if defined(ENABLE_ATOMICS_TESTS) && defined(__BPF_FEATURE_ADDR_SPACE_CAST)
arena_spinlock_t __arena lock;
int test_skip = 1;
+
+/*
+ * Storage for the queue nodes declared by bpf_arena_spin_lock.h. Each program
+ * linking the arena spinlock provides exactly one definition; libarena's lives
+ * in libarena/src/common.bpf.c.
+ */
+struct arena_qnode __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
#else
int test_skip = 2;
#endif
--
2.55.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes
2026-08-03 0:18 [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes Changwoo Min
@ 2026-08-13 4:38 ` Kumar Kartikeya Dwivedi
2026-08-17 9:05 ` Daniel Borkmann
1 sibling, 0 replies; 6+ messages in thread
From: Kumar Kartikeya Dwivedi @ 2026-08-13 4:38 UTC (permalink / raw)
To: Changwoo Min, Alexei Starovoitov, Daniel Borkmann,
Andrii Nakryiko, Eduard Zingerman, Shuah Khan, Emil Tsalapatis
Cc: Martin KaFai Lau, Song Liu, Yonghong Song, Jiri Olsa, bpf,
sched-ext, linux-kselftest, linux-kernel, kernel-dev
On Mon Aug 3, 2026 at 2:18 AM CEST, Changwoo Min wrote:
> bpf_arena_spin_lock.h defines its 64KB qnodes array in the header, so
> every translation unit including it emits a copy. __weak makes them all
> resolve to one instance, but bpftool gen object merges only the symbols
> and concatenates each input's .addr_space.1 bytes, leaving the surplus
> copies unreferenced in the linked object.
>
> libarena links ten such units, so nine copies were dead weight (bytes):
>
> object before after
> -----------------------------------------------------
> .addr_space.1 in libarena.bpf.o 676200 86376
> libarena.skel.h 2100123 892371
> libarena_asan.skel.h 2641124 1466477
>
> Declare qnodes in the header and let each program define it once:
> libarena in src/common.bpf.c, and the arena_spin_lock test beside the
> lock it guards.
>
> Tested with test_progs -t arena_spin_lock and -t libarena.
>
> Signed-off-by: Changwoo Min <changwoo@igalia.com>
> ---
This seems ok to me. I don't have better ideas. Emil, any thoughts?
> .../selftests/bpf/libarena/include/bpf_arena_spin_lock.h | 7 +------
> tools/testing/selftests/bpf/libarena/src/common.bpf.c | 7 +++++++
> tools/testing/selftests/bpf/progs/arena_spin_lock.c | 7 +++++++
> 3 files changed, 15 insertions(+), 6 deletions(-)
>
> diff --git a/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h b/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h
> index ae6b72d15bb6..71d9db610263 100644
> --- a/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h
> +++ b/tools/testing/selftests/bpf/libarena/include/bpf_arena_spin_lock.h
> @@ -103,12 +103,7 @@ struct arena_qnode {
> #define _Q_LOCKED_VAL (1U << _Q_LOCKED_OFFSET)
> #define _Q_PENDING_VAL (1U << _Q_PENDING_OFFSET)
>
> -/*
> - * The qnodes are marked __weak so we can define them in the header
> - * while still ensuring all compilation units use the same struct
> - * instance.
> - */
> -struct arena_qnode __weak __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
> +extern struct arena_qnode __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
>
> static inline u32 encode_tail(int cpu, int idx)
> {
> diff --git a/tools/testing/selftests/bpf/libarena/src/common.bpf.c b/tools/testing/selftests/bpf/libarena/src/common.bpf.c
> index 50be57213dfb..06481f2a2892 100644
> --- a/tools/testing/selftests/bpf/libarena/src/common.bpf.c
> +++ b/tools/testing/selftests/bpf/libarena/src/common.bpf.c
> @@ -8,6 +8,13 @@ const volatile u32 zero = 0;
>
> struct buddy __arena buddy;
>
> +/*
> + * Storage for the queue nodes declared by bpf_arena_spin_lock.h. Each program
> + * linking the arena spinlock provides exactly one definition, so that the array
> + * is emitted once rather than once per translation unit.
> + */
> +struct arena_qnode __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
> +
> int arena_fls(__u64 word)
> {
> if (!word)
> diff --git a/tools/testing/selftests/bpf/progs/arena_spin_lock.c b/tools/testing/selftests/bpf/progs/arena_spin_lock.c
> index cf7cda79c16c..92e75ec3844c 100644
> --- a/tools/testing/selftests/bpf/progs/arena_spin_lock.c
> +++ b/tools/testing/selftests/bpf/progs/arena_spin_lock.c
> @@ -23,6 +23,13 @@ int cs_count;
> #if defined(ENABLE_ATOMICS_TESTS) && defined(__BPF_FEATURE_ADDR_SPACE_CAST)
> arena_spinlock_t __arena lock;
> int test_skip = 1;
> +
> +/*
> + * Storage for the queue nodes declared by bpf_arena_spin_lock.h. Each program
> + * linking the arena spinlock provides exactly one definition; libarena's lives
> + * in libarena/src/common.bpf.c.
> + */
> +struct arena_qnode __arena __hidden qnodes[_Q_MAX_CPUS][_Q_MAX_NODES];
> #else
> int test_skip = 2;
> #endif
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes
2026-08-03 0:18 [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes Changwoo Min
2026-08-13 4:38 ` Kumar Kartikeya Dwivedi
@ 2026-08-17 9:05 ` Daniel Borkmann
2026-08-17 16:00 ` Changwoo Min
1 sibling, 1 reply; 6+ messages in thread
From: Daniel Borkmann @ 2026-08-17 9:05 UTC (permalink / raw)
To: Changwoo Min, Alexei Starovoitov, Andrii Nakryiko,
Eduard Zingerman, Kumar Kartikeya Dwivedi, Shuah Khan,
Emil Tsalapatis
Cc: Martin KaFai Lau, Song Liu, Yonghong Song, Jiri Olsa, bpf,
sched-ext, linux-kselftest, linux-kernel, kernel-dev
On 8/3/26 2:18 AM, Changwoo Min wrote:
> bpf_arena_spin_lock.h defines its 64KB qnodes array in the header, so
> every translation unit including it emits a copy. __weak makes them all
> resolve to one instance, but bpftool gen object merges only the symbols
> and concatenates each input's .addr_space.1 bytes, leaving the surplus
> copies unreferenced in the linked object.
>
> libarena links ten such units, so nine copies were dead weight (bytes):
>
> object before after
> -----------------------------------------------------
> .addr_space.1 in libarena.bpf.o 676200 86376
> libarena.skel.h 2100123 892371
> libarena_asan.skel.h 2641124 1466477
>
> Declare qnodes in the header and let each program define it once:
> libarena in src/common.bpf.c, and the arena_spin_lock test beside the
> lock it guards.
>
> Tested with test_progs -t arena_spin_lock and -t libarena.
>
> Signed-off-by: Changwoo Min <changwoo@igalia.com>
This doesn't apply cleanly, could you respin against latest bpf-next?
Thanks,
Daniel
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes
2026-08-17 9:05 ` Daniel Borkmann
@ 2026-08-17 16:00 ` Changwoo Min
2026-08-17 17:14 ` Emil Tsalapatis
0 siblings, 1 reply; 6+ messages in thread
From: Changwoo Min @ 2026-08-17 16:00 UTC (permalink / raw)
To: Daniel Borkmann, Alexei Starovoitov, Andrii Nakryiko,
Eduard Zingerman, Kumar Kartikeya Dwivedi, Shuah Khan,
Emil Tsalapatis
Cc: Martin KaFai Lau, Song Liu, Yonghong Song, Jiri Olsa, bpf,
sched-ext, linux-kselftest, linux-kernel, kernel-dev
Hi Daniel,
On 8/17/26 6:05 PM, Daniel Borkmann wrote:
> On 8/3/26 2:18 AM, Changwoo Min wrote:
>> bpf_arena_spin_lock.h defines its 64KB qnodes array in the header, so
>> every translation unit including it emits a copy. __weak makes them all
>> resolve to one instance, but bpftool gen object merges only the symbols
>> and concatenates each input's .addr_space.1 bytes, leaving the surplus
>> copies unreferenced in the linked object.
>>
>> libarena links ten such units, so nine copies were dead weight (bytes):
>>
>> object before after
>> -----------------------------------------------------
>> .addr_space.1 in libarena.bpf.o 676200 86376
>> libarena.skel.h 2100123 892371
>> libarena_asan.skel.h 2641124 1466477
>>
>> Declare qnodes in the header and let each program define it once:
>> libarena in src/common.bpf.c, and the arena_spin_lock test beside the
>> lock it guards.
>>
>> Tested with test_progs -t arena_spin_lock and -t libarena.
>>
>> Signed-off-by: Changwoo Min <changwoo@igalia.com>
> This doesn't apply cleanly, could you respin against latest bpf-next?
Sure. I will send the v2 rebased on the HEAD of the the bpf-next tree soon.
Regards.
Changwoo Min
>
> Thanks,
> Daniel
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes
2026-08-17 16:00 ` Changwoo Min
@ 2026-08-17 17:14 ` Emil Tsalapatis
2026-08-17 18:05 ` Kumar Kartikeya Dwivedi
0 siblings, 1 reply; 6+ messages in thread
From: Emil Tsalapatis @ 2026-08-17 17:14 UTC (permalink / raw)
To: Changwoo Min
Cc: Daniel Borkmann, Alexei Starovoitov, Andrii Nakryiko,
Eduard Zingerman, Kumar Kartikeya Dwivedi, Shuah Khan,
Emil Tsalapatis, Martin KaFai Lau, Song Liu, Yonghong Song,
Jiri Olsa, bpf, sched-ext, linux-kselftest, linux-kernel,
kernel-dev
On Mon, Aug 17, 2026 at 9:02 AM Changwoo Min <changwoo@igalia.com> wrote:
>
> >
> Hi Daniel,
>
> On 8/17/26 6:05 PM, Daniel Borkmann wrote:
> > On 8/3/26 2:18 AM, Changwoo Min wrote:
> >> bpf_arena_spin_lock.h defines its 64KB qnodes array in the header, so
> >> every translation unit including it emits a copy. __weak makes them all
> >> resolve to one instance, but bpftool gen object merges only the symbols
> >> and concatenates each input's .addr_space.1 bytes, leaving the surplus
> >> copies unreferenced in the linked object.
> >>
> >> libarena links ten such units, so nine copies were dead weight (bytes):
> >>
> >> object before after
> >> -----------------------------------------------------
> >> .addr_space.1 in libarena.bpf.o 676200 86376
> >> libarena.skel.h 2100123 892371
> >> libarena_asan.skel.h 2641124 1466477
> >>
> >> Declare qnodes in the header and let each program define it once:
> >> libarena in src/common.bpf.c, and the arena_spin_lock test beside the
> >> lock it guards.
> >>
> >> Tested with test_progs -t arena_spin_lock and -t libarena.
> >>
> >> Signed-off-by: Changwoo Min <changwoo@igalia.com>
Hi Changwoo,
Sorry for the late reply, for the next version feel free to add:
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
> > This doesn't apply cleanly, could you respin against latest bpf-next?
>
> Sure. I will send the v2 rebased on the HEAD of the the bpf-next tree soon.
>
> Regards.
> Changwoo Min
>
> >
> > Thanks,
> > Daniel
> >
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes
2026-08-17 17:14 ` Emil Tsalapatis
@ 2026-08-17 18:05 ` Kumar Kartikeya Dwivedi
0 siblings, 0 replies; 6+ messages in thread
From: Kumar Kartikeya Dwivedi @ 2026-08-17 18:05 UTC (permalink / raw)
To: Emil Tsalapatis, Changwoo Min
Cc: Daniel Borkmann, Alexei Starovoitov, Andrii Nakryiko,
Eduard Zingerman, Shuah Khan, Emil Tsalapatis, Martin KaFai Lau,
Song Liu, Yonghong Song, Jiri Olsa, bpf, sched-ext,
linux-kselftest, linux-kernel, kernel-dev
On Mon Aug 17, 2026 at 7:14 PM CEST, Emil Tsalapatis wrote:
> On Mon, Aug 17, 2026 at 9:02 AM Changwoo Min <changwoo@igalia.com> wrote:
>>
>> >
>> Hi Daniel,
>>
>> On 8/17/26 6:05 PM, Daniel Borkmann wrote:
>> > On 8/3/26 2:18 AM, Changwoo Min wrote:
>> >> bpf_arena_spin_lock.h defines its 64KB qnodes array in the header, so
>> >> every translation unit including it emits a copy. __weak makes them all
>> >> resolve to one instance, but bpftool gen object merges only the symbols
>> >> and concatenates each input's .addr_space.1 bytes, leaving the surplus
>> >> copies unreferenced in the linked object.
>> >>
>> >> libarena links ten such units, so nine copies were dead weight (bytes):
>> >>
>> >> object before after
>> >> -----------------------------------------------------
>> >> .addr_space.1 in libarena.bpf.o 676200 86376
>> >> libarena.skel.h 2100123 892371
>> >> libarena_asan.skel.h 2641124 1466477
>> >>
>> >> Declare qnodes in the header and let each program define it once:
>> >> libarena in src/common.bpf.c, and the arena_spin_lock test beside the
>> >> lock it guards.
>> >>
>> >> Tested with test_progs -t arena_spin_lock and -t libarena.
>> >>
>> >> Signed-off-by: Changwoo Min <changwoo@igalia.com>
>
> Hi Changwoo,
>
> Sorry for the late reply, for the next version feel free to add:
>
> Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
>
Added the tag to v2 and applied, thanks.
>> > This doesn't apply cleanly, could you respin against latest bpf-next?
>>
>> Sure. I will send the v2 rebased on the HEAD of the the bpf-next tree soon.
>>
>> Regards.
>> Changwoo Min
>>
>> >
>> > Thanks,
>> > Daniel
>> >
>>
>>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2026-08-17 18:05 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-08-03 0:18 [PATCH] selftests/bpf: Remove duplicate copies of the arena spinlock qnodes Changwoo Min
2026-08-13 4:38 ` Kumar Kartikeya Dwivedi
2026-08-17 9:05 ` Daniel Borkmann
2026-08-17 16:00 ` Changwoo Min
2026-08-17 17:14 ` Emil Tsalapatis
2026-08-17 18:05 ` Kumar Kartikeya Dwivedi
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox