[PATCH bpf-next v2] bpf: make perf_event_read_output accessible from BPF core when available

[PATCH bpf-next v2] bpf: make perf_event_read_output accessible from BPF core when available

[Emil Tsalapatis]

The perf_event_read_event_output helper is currently only available to
tracing protrams, but is useful for other BPF programs like sched_ext
schedulers. When the helper is available, provide its bpf_func_proto
directly from the bpf core.

Signed-off-by: Emil Tsalapatis (Meta) <emil@etsalapatis.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
---
 include/linux/bpf.h      | 2 ++
 kernel/bpf/core.c        | 5 +++++
 kernel/bpf/helpers.c     | 2 ++
 kernel/trace/bpf_trace.c | 5 +++++
 4 files changed, 14 insertions(+)

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 0d7b70124d81..973a88d9b52b 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -2059,6 +2059,8 @@ int bpf_prog_calc_tag(struct bpf_prog *fp);
 const struct bpf_func_proto *bpf_get_trace_printk_proto(void);
 const struct bpf_func_proto *bpf_get_trace_vprintk_proto(void);
 
+const struct bpf_func_proto *bpf_get_perf_event_read_value_proto(void);
+
 typedef unsigned long (*bpf_ctx_copy_t)(void *dst, const void *src,
 					unsigned long off, unsigned long len);
 typedef u32 (*bpf_convert_ctx_access_t)(enum bpf_access_type type,
diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
index 62cb9557ad3b..ba6b6118cf50 100644
--- a/kernel/bpf/core.c
+++ b/kernel/bpf/core.c
@@ -2972,6 +2972,11 @@ const struct bpf_func_proto * __weak bpf_get_trace_vprintk_proto(void)
 	return NULL;
 }
 
+const struct bpf_func_proto * __weak bpf_get_perf_event_read_value_proto(void)
+{
+	return NULL;
+}
+
 u64 __weak
 bpf_event_output(struct bpf_map *map, u64 flags, void *meta, u64 meta_size,
 		 void *ctx, u64 ctx_size, bpf_ctx_copy_t ctx_copy)
diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index 5449756ba102..ddaa41a70676 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -2056,6 +2056,8 @@ bpf_base_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog)
 		return &bpf_task_pt_regs_proto;
 	case BPF_FUNC_trace_vprintk:
 		return bpf_get_trace_vprintk_proto();
+	case BPF_FUNC_perf_event_read_value:
+		return bpf_get_perf_event_read_value_proto();
 	default:
 		return NULL;
 	}
diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 13bef2462e94..6b07fa7081d9 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c
@@ -607,6 +607,11 @@ static const struct bpf_func_proto bpf_perf_event_read_value_proto = {
 	.arg4_type	= ARG_CONST_SIZE,
 };
 
+const struct bpf_func_proto *bpf_get_perf_event_read_value_proto(void)
+{
+	return &bpf_perf_event_read_value_proto;
+}
+
 static __always_inline u64
 __bpf_perf_event_output(struct pt_regs *regs, struct bpf_map *map,
 			u64 flags, struct perf_raw_record *raw,
-- 
2.47.1

Re: [PATCH bpf-next v2] bpf: make perf_event_read_output accessible from BPF core when available

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to bpf/bpf-next.git (master)
by Alexei Starovoitov <ast@kernel.org>:

On Mon, 17 Mar 2025 23:07:53 -0400 you wrote:
> The perf_event_read_event_output helper is currently only available to
> tracing protrams, but is useful for other BPF programs like sched_ext
> schedulers. When the helper is available, provide its bpf_func_proto
> directly from the bpf core.
> 
> Signed-off-by: Emil Tsalapatis (Meta) <emil@etsalapatis.com>
> Acked-by: Jiri Olsa <jolsa@kernel.org>
> 
> [...]

Here is the summary with links:
  - [bpf-next,v2] bpf: make perf_event_read_output accessible from BPF core when available
    https://git.kernel.org/bpf/bpf-next/c/ae0a457f5d33

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html