BPF List
 help / color / mirror / Atom feed
From: Eduard Zingerman <eddyz87@gmail.com>
To: Mykyta Yatsenko <mykyta.yatsenko5@gmail.com>,
	bpf@vger.kernel.org,  ast@kernel.org, andrii@kernel.org,
	daniel@iogearbox.net, kafai@meta.com,  kernel-team@meta.com,
	memxor@gmail.com
Cc: Mykyta Yatsenko <yatsenko@meta.com>
Subject: Re: [PATCH bpf-next v3 08/10] bpf: verifier: refactor kfunc specialization
Date: Mon, 20 Oct 2025 16:38:31 -0700	[thread overview]
Message-ID: <e6388c88bcb09404209a956e65f4d0510aa13294.camel@gmail.com> (raw)
In-Reply-To: <20251020222538.932915-9-mykyta.yatsenko5@gmail.com>

On Mon, 2025-10-20 at 23:25 +0100, Mykyta Yatsenko wrote:

[...]

> @@ -3375,18 +3366,25 @@ static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16 offset)
>  			return err;
>  	}
>  
> +	err = btf_distill_func_proto(&env->log, desc_btf,
> +				     func_proto, func_name,
> +				     &func_model);
> +	if (err)
> +		return err;
> +
> +	err = kfunc_call_imm(env, addr, func_id, &call_imm);
> +	if (err)
> +		return err;
> +

Sorry, I should have asked in v1/v2. Is there a reason to call this
function two times? In other words, it looks like doing the following
on top of your changes is sufficient and removes the need to call
kfunc_call_imm twice:

    ---- 8< -----------------------------------------

    diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
    index 0418768d13e4..f509f9e0383d 100644
    --- a/kernel/bpf/verifier.c
    +++ b/kernel/bpf/verifier.c
    @@ -3373,13 +3373,8 @@ static int add_kfunc_call(struct bpf_verifier_env *env, u32 func_id, s16 offset)
            if (err)
                    return err;
     
    -       err = kfunc_call_imm(env, addr, func_id, &call_imm);
    -       if (err)
    -               return err;
    -
            desc = &tab->descs[tab->nr_descs++];
            desc->func_id = func_id;
    -       desc->imm = call_imm;
            desc->offset = offset;
            desc->addr = addr;
            desc->func_model = func_model;
    @@ -21892,8 +21887,10 @@ static int specialize_kfunc(struct bpf_verifier_env *env, struct bpf_kfunc_desc
            unsigned long addr = 0;
            int err;
     
    +       addr = desc->addr;
    +
            if (offset) /* return if module BTF is used */
    -               return 0;
    +               goto fill_imm;
     
            if (bpf_dev_bound_kfunc_id(func_id)) {
                    xdp_kfunc = bpf_dev_bound_resolve_kfunc(prog, func_id);
    @@ -21922,9 +21919,7 @@ static int specialize_kfunc(struct bpf_verifier_env *env, struct bpf_kfunc_desc
                            addr = (unsigned long)bpf_dynptr_from_file_sleepable;
            }
     
    -       if (!addr) /* Nothing to patch with */
    -               return 0;
    -
    +fill_imm:
            err = kfunc_call_imm(env, addr, func_id, &desc->imm);
            if (err)
                    return err;

    ----------------------------------------- >8 ----

The desc->imm field is used only from:
- kfunc_desc_cmp_by_imm_off():
  - invoked from do_misc_fixups() at the very end,
    after all specialize_kfunc() calls are done.
- bpf_jit_find_kfunc_model(), from some jits.

So, that should be safe.
Selftests are passing after this change as well.

>  	desc = &tab->descs[tab->nr_descs++];
>  	desc->func_id = func_id;
>  	desc->imm = call_imm;
>  	desc->offset = offset;
>  	desc->addr = addr;
> -	err = btf_distill_func_proto(&env->log, desc_btf,
> -				     func_proto, func_name,
> -				     &desc->func_model);
> -	if (!err)
> -		sort(tab->descs, tab->nr_descs, sizeof(tab->descs[0]),
> -		     kfunc_desc_cmp_by_id_off, NULL);
> -	return err;
> +	desc->func_model = func_model;
> +	sort(tab->descs, tab->nr_descs, sizeof(tab->descs[0]),
> +	     kfunc_desc_cmp_by_id_off, NULL);
> +	return 0;
>  }


[...]

  reply	other threads:[~2025-10-20 23:38 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-20 22:25 [PATCH bpf-next v3 00/10] bpf: Introduce file dynptr Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 01/10] selftests/bpf: remove unnecessary kfunc prototypes Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 02/10] bpf: widen dynptr size/offset to 64 bit Mykyta Yatsenko
2025-10-20 23:01   ` Eduard Zingerman
2025-10-21 11:59     ` Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 03/10] lib: move freader into buildid.h Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 04/10] lib/freader: support reading more than 2 folios Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 05/10] bpf: verifier: centralize const dynptr check in unmark_stack_slots_dynptr() Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 06/10] bpf: add plumbing for file-backed dynptr Mykyta Yatsenko
2025-10-20 23:08   ` Eduard Zingerman
2025-10-20 22:25 ` [PATCH bpf-next v3 07/10] bpf: add kfuncs and helpers support for file dynptrs Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 08/10] bpf: verifier: refactor kfunc specialization Mykyta Yatsenko
2025-10-20 23:38   ` Eduard Zingerman [this message]
2025-10-21 13:03     ` Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 09/10] bpf: dispatch to sleepable file dynptr Mykyta Yatsenko
2025-10-20 22:25 ` [PATCH bpf-next v3 10/10] selftests/bpf: add file dynptr tests Mykyta Yatsenko
2025-10-21  0:45   ` Eduard Zingerman
2025-10-21 13:55     ` Mykyta Yatsenko
2025-10-21 16:40       ` Eduard Zingerman

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=e6388c88bcb09404209a956e65f4d0510aa13294.camel@gmail.com \
    --to=eddyz87@gmail.com \
    --cc=andrii@kernel.org \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=kafai@meta.com \
    --cc=kernel-team@meta.com \
    --cc=memxor@gmail.com \
    --cc=mykyta.yatsenko5@gmail.com \
    --cc=yatsenko@meta.com \
    /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