Re: [PATCH bpf v2] selftests/bpf: Fix "missing ENDBR" BUG for destructor kfunc

From: Jiri Olsa
Date: Tue Nov 22 2022 - 08:48:18 EST


On Mon, Nov 21, 2022 at 11:32:43PM -0800, Chen Hu wrote:
> With CONFIG_X86_KERNEL_IBT enabled, the test_verifier triggers the
> following BUG:
>
> traps: Missing ENDBR: bpf_kfunc_call_test_release+0x0/0x30
> ------------[ cut here ]------------
> kernel BUG at arch/x86/kernel/traps.c:254!
> invalid opcode: 0000 [#1] PREEMPT SMP
> <TASK>
> asm_exc_control_protection+0x26/0x50
> RIP: 0010:bpf_kfunc_call_test_release+0x0/0x30
> Code: 00 48 c7 c7 18 f2 e1 b4 e8 0d ca 8c ff 48 c7 c0 00 f2 e1 b4 c3
> 0f 1f 44 00 00 66 0f 1f 00 0f 1f 44 00 00 0f 0b 31 c0 c3 66 90
> <66> 0f 1f 00 0f 1f 44 00 00 48 85 ff 74 13 4c 8d 47 18 b8 ff ff ff
> bpf_map_free_kptrs+0x2e/0x70
> array_map_free+0x57/0x140
> process_one_work+0x194/0x3a0
> worker_thread+0x54/0x3a0
> ? rescuer_thread+0x390/0x390
> kthread+0xe9/0x110
> ? kthread_complete_and_exit+0x20/0x20
>
> This is because there are no compile-time references to the destructor
> kfuncs, bpf_kfunc_call_test_release() for example. So objtool marked
> them sealable and ENDBR in the functions were sealed (converted to NOP)
> by apply_ibt_endbr().
>
> This fix creates dummy compile-time references to destructor kfuncs so
> ENDBR stay there.
>
> Fixes: 05a945deefaa ("selftests/bpf: Add verifier tests for kptr")
> Signed-off-by: Chen Hu <hu1.chen@xxxxxxxxx>
> Tested-by: Pengfei Xu <pengfei.xu@xxxxxxxxx>
> ---
> v2:
> - Use generic macro name and place the macro after function body as
> - suggested by Jiri Olsa
>
> v1: https://lore.kernel.org/all/20221121085113.611504-1-hu1.chen@xxxxxxxxx/
>
> include/linux/btf_ids.h | 7 +++++++
> net/bpf/test_run.c | 4 ++++
> 2 files changed, 11 insertions(+)
>
> diff --git a/include/linux/btf_ids.h b/include/linux/btf_ids.h
> index 2aea877d644f..db02691b506d 100644
> --- a/include/linux/btf_ids.h
> +++ b/include/linux/btf_ids.h
> @@ -266,4 +266,11 @@ MAX_BTF_TRACING_TYPE,
>
> extern u32 btf_tracing_ids[];
>
> +#if defined(CONFIG_X86_KERNEL_IBT) && !defined(__DISABLE_EXPORTS)
> +#define FUNC_IBT_NOSEAL(name) \
> + asm(IBT_NOSEAL(#name));
> +#else
> +#define FUNC_IBT_NOSEAL(name)
> +#endif /* CONFIG_X86_KERNEL_IBT */

hum, IBT_NOSEAL is x86 specific, so this will probably fail build
on other archs.. I think we could ifdef it with CONFIG_X86, but
it should go to some IBT related header? surely not to btf_ids.h

cc-ing Peter and Josh

thanks,
jirka


> +
> #endif
> diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
> index 13d578ce2a09..07263b7cc12d 100644
> --- a/net/bpf/test_run.c
> +++ b/net/bpf/test_run.c
> @@ -597,10 +597,14 @@ noinline void bpf_kfunc_call_test_release(struct prog_test_ref_kfunc *p)
> refcount_dec(&p->cnt);
> }
>
> +FUNC_IBT_NOSEAL(bpf_kfunc_call_test_release)
> +
> noinline void bpf_kfunc_call_memb_release(struct prog_test_member *p)
> {
> }
>
> +FUNC_IBT_NOSEAL(bpf_kfunc_call_memb_release)
> +
> noinline void bpf_kfunc_call_memb1_release(struct prog_test_member1 *p)
> {
> WARN_ON_ONCE(1);
> --
> 2.34.1
>