[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAErzpmtFEudhMtYvmhb8gs0-YeA+_iET4azNML6fd9-8M9r1iQ@mail.gmail.com>
Date: Wed, 29 Oct 2025 09:57:14 +0800
From: Donglin Peng <dolinux.peng@...il.com>
To: Alexei Starovoitov <alexei.starovoitov@...il.com>
Cc: Alexei Starovoitov <ast@...nel.org>, LKML <linux-kernel@...r.kernel.org>,
bpf <bpf@...r.kernel.org>, Eduard Zingerman <eddyz87@...il.com>,
Andrii Nakryiko <andrii.nakryiko@...il.com>, Alan Maguire <alan.maguire@...cle.com>,
Song Liu <song@...nel.org>, pengdonglin <pengdonglin@...omi.com>
Subject: Re: [RFC PATCH v3 3/3] btf: Reuse libbpf code for BTF type sorting
verification and binary search
On Tue, Oct 28, 2025 at 3:56 AM Alexei Starovoitov
<alexei.starovoitov@...il.com> wrote:
>
> On Mon, Oct 27, 2025 at 6:54 AM Donglin Peng <dolinux.peng@...il.com> wrote:
> >
> > +
> > +/*
> > + * btf_sort.c is included directly to avoid function call overhead
> > + * when accessing BTF private data, as this file is shared between
> > + * libbpf and kernel and may be called frequently (especially when
> > + * funcgraph-args or func-args tracing options are enabled).
> > + */
> > +#include "../../tools/lib/bpf/btf_sort.c"
>
> function call overhead? I don't believe it's measurable.
Thank you. Since the overhead is primarily due to the function graph tracer,
I have reverted to the v2 approach and added the notrace attribute to the
relevant functions.
>
> Don't do it on libbpf side either.
Yes.
>
> pw-bot: cr
Powered by blists - more mailing lists