* [PATCH bpf-next v4 1/2] libbpf: do not resolve size on duplicate FUNCs
2024-10-02 6:25 [PATCH bpf-next v4 0/2] BPF static linker: fix linking duplicate extern functions Eric Long via B4 Relay
@ 2024-10-02 6:25 ` Eric Long via B4 Relay
2024-10-02 6:25 ` [PATCH bpf-next v4 2/2] selftests/bpf: test linking with duplicate extern functions Eric Long via B4 Relay
2024-10-08 3:30 ` [PATCH bpf-next v4 0/2] BPF static linker: fix linking " patchwork-bot+netdevbpf
2 siblings, 0 replies; 4+ messages in thread
From: Eric Long via B4 Relay @ 2024-10-02 6:25 UTC (permalink / raw)
To: bpf; +Cc: Andrii Nakryiko, Alexei Starovoitov, netdev, Eric Long
From: Eric Long <i@hack3r.moe>
FUNCs do not have sizes, thus currently btf__resolve_size will fail
with -EINVAL. Add conditions so that we only update size when the BTF
object is not function or function prototype.
Signed-off-by: Eric Long <i@hack3r.moe>
---
tools/lib/bpf/linker.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/tools/lib/bpf/linker.c b/tools/lib/bpf/linker.c
index 81dbbdd79a7c65a4b048b85e1dba99cb5f7cb56b..f83c1c29982c3d9d7f6775cd98a6f3387a4e9c50 100644
--- a/tools/lib/bpf/linker.c
+++ b/tools/lib/bpf/linker.c
@@ -2451,6 +2451,10 @@ static int linker_append_btf(struct bpf_linker *linker, struct src_obj *obj)
if (glob_sym && glob_sym->var_idx >= 0) {
__s64 sz;
+ /* FUNCs don't have size, nothing to update */
+ if (btf_is_func(t))
+ continue;
+
dst_var = &dst_sec->sec_vars[glob_sym->var_idx];
/* Because underlying BTF type might have
* changed, so might its size have changed, so
--
2.46.2
^ permalink raw reply related [flat|nested] 4+ messages in thread* [PATCH bpf-next v4 2/2] selftests/bpf: test linking with duplicate extern functions
2024-10-02 6:25 [PATCH bpf-next v4 0/2] BPF static linker: fix linking duplicate extern functions Eric Long via B4 Relay
2024-10-02 6:25 ` [PATCH bpf-next v4 1/2] libbpf: do not resolve size on duplicate FUNCs Eric Long via B4 Relay
@ 2024-10-02 6:25 ` Eric Long via B4 Relay
2024-10-08 3:30 ` [PATCH bpf-next v4 0/2] BPF static linker: fix linking " patchwork-bot+netdevbpf
2 siblings, 0 replies; 4+ messages in thread
From: Eric Long via B4 Relay @ 2024-10-02 6:25 UTC (permalink / raw)
To: bpf; +Cc: Andrii Nakryiko, Alexei Starovoitov, netdev, Eric Long
From: Eric Long <i@hack3r.moe>
Previously when multiple BPF object files referencing the same extern
function (usually kfunc) are statically linked using `bpftool gen
object`, libbpf tries to get the nonexistent size of BTF_KIND_FUNC_PROTO
and fails. This test ensures it is fixed.
Signed-off-by: Eric Long <i@hack3r.moe>
---
tools/testing/selftests/bpf/progs/linked_funcs1.c | 8 ++++++++
tools/testing/selftests/bpf/progs/linked_funcs2.c | 8 ++++++++
2 files changed, 16 insertions(+)
diff --git a/tools/testing/selftests/bpf/progs/linked_funcs1.c b/tools/testing/selftests/bpf/progs/linked_funcs1.c
index cc79dddac182c20da69a1a57fa39bc81004184f1..049a1f78de3f835e7658dde6f2d03161b6a5a07f 100644
--- a/tools/testing/selftests/bpf/progs/linked_funcs1.c
+++ b/tools/testing/selftests/bpf/progs/linked_funcs1.c
@@ -63,6 +63,8 @@ extern int set_output_val2(int x);
/* here we'll force set_output_ctx2() to be __hidden in the final obj file */
__hidden extern void set_output_ctx2(__u64 *ctx);
+void *bpf_cast_to_kern_ctx(void *obj) __ksym;
+
SEC("?raw_tp/sys_enter")
int BPF_PROG(handler1, struct pt_regs *regs, long id)
{
@@ -86,4 +88,10 @@ int BPF_PROG(handler1, struct pt_regs *regs, long id)
return 0;
}
+/* Generate BTF FUNC record and test linking with duplicate extern functions */
+void kfunc_gen1(void)
+{
+ bpf_cast_to_kern_ctx(0);
+}
+
char LICENSE[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/progs/linked_funcs2.c b/tools/testing/selftests/bpf/progs/linked_funcs2.c
index 942cc5526ddf02004bf82d1f72c74cfba6eaf486..96850759fd8d0074249bbf1f743aab08e20de0fc 100644
--- a/tools/testing/selftests/bpf/progs/linked_funcs2.c
+++ b/tools/testing/selftests/bpf/progs/linked_funcs2.c
@@ -63,6 +63,8 @@ extern int set_output_val1(int x);
/* here we'll force set_output_ctx1() to be __hidden in the final obj file */
__hidden extern void set_output_ctx1(__u64 *ctx);
+void *bpf_cast_to_kern_ctx(void *obj) __ksym;
+
SEC("?raw_tp/sys_enter")
int BPF_PROG(handler2, struct pt_regs *regs, long id)
{
@@ -86,4 +88,10 @@ int BPF_PROG(handler2, struct pt_regs *regs, long id)
return 0;
}
+/* Generate BTF FUNC record and test linking with duplicate extern functions */
+void kfunc_gen2(void)
+{
+ bpf_cast_to_kern_ctx(0);
+}
+
char LICENSE[] SEC("license") = "GPL";
--
2.46.2
^ permalink raw reply related [flat|nested] 4+ messages in thread* Re: [PATCH bpf-next v4 0/2] BPF static linker: fix linking duplicate extern functions
2024-10-02 6:25 [PATCH bpf-next v4 0/2] BPF static linker: fix linking duplicate extern functions Eric Long via B4 Relay
2024-10-02 6:25 ` [PATCH bpf-next v4 1/2] libbpf: do not resolve size on duplicate FUNCs Eric Long via B4 Relay
2024-10-02 6:25 ` [PATCH bpf-next v4 2/2] selftests/bpf: test linking with duplicate extern functions Eric Long via B4 Relay
@ 2024-10-08 3:30 ` patchwork-bot+netdevbpf
2 siblings, 0 replies; 4+ messages in thread
From: patchwork-bot+netdevbpf @ 2024-10-08 3:30 UTC (permalink / raw)
To: Eric Long via B4 Relay; +Cc: bpf, andrii, ast, netdev, i
Hello:
This series was applied to bpf/bpf-next.git (master)
by Andrii Nakryiko <andrii@kernel.org>:
On Wed, 02 Oct 2024 14:25:05 +0800 you wrote:
> Currently, if `bpftool gen object` tries to link two objects that
> contains the same extern function prototype, libbpf will try to get
> their (non-existent) size by calling bpf__resolve_size like extern
> variables and fail with:
>
> libbpf: global 'whatever': failed to resolve size of underlying type: -22
>
> [...]
Here is the summary with links:
- [bpf-next,v4,1/2] libbpf: do not resolve size on duplicate FUNCs
https://git.kernel.org/bpf/bpf-next/c/4b146e95da87
- [bpf-next,v4,2/2] selftests/bpf: test linking with duplicate extern functions
https://git.kernel.org/bpf/bpf-next/c/3c591de28543
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 4+ messages in thread