linux-kernel - Re: [RFC PATCH v7 3/7] tools/resolve_btfids: Add --btf

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAErzpmsgWK_RBv-DFa=jBvF8vqxUwU86yJhizXyYKc07k6Ys6w@mail.gmail.com>
Date: Mon, 24 Nov 2025 20:14:28 +0800
From: Donglin Peng <dolinux.peng@...il.com>
To: Eduard Zingerman <eddyz87@...il.com>
Cc: ast@...nel.org, andrii.nakryiko@...il.com, zhangxiaoqin@...omi.com, 
	linux-kernel@...r.kernel.org, bpf@...r.kernel.org, 
	Donglin Peng <pengdonglin@...omi.com>, Alan Maguire <alan.maguire@...cle.com>, 
	Song Liu <song@...nel.org>
Subject: Re: [RFC PATCH v7 3/7] tools/resolve_btfids: Add --btf_sort option
 for BTF name sorting

On Fri, Nov 21, 2025 at 8:18 AM Eduard Zingerman <eddyz87@...il.com> wrote:
>
> On Wed, 2025-11-19 at 11:15 +0800, Donglin Peng wrote:
>
> [...]
>
> > diff --git a/scripts/Makefile.btf b/scripts/Makefile.btf
> > index db76335dd917..d5eb4ee70e88 100644
> > --- a/scripts/Makefile.btf
> > +++ b/scripts/Makefile.btf
> > @@ -27,6 +27,7 @@ pahole-flags-$(call test-ge, $(pahole-ver), 130) += --btf_features=attributes
> >
> >  ifneq ($(KBUILD_EXTMOD),)
> >  module-pahole-flags-$(call test-ge, $(pahole-ver), 128) += --btf_features=distilled_base
> > +module-resolve_btfid-flags-y = --distilled_base
>                                   ^^^^^^^^^^^^^^^^
> This flag should be guarded by pahole version as well.  However, I'd
> suggest not adding this flag at all, but instead modify resolve_btfids
> to check for BTF_BASE_ELF_SEC section existence and acting accordingly.

Thanks, I will remove it in the next version.

>
> >  endif
> >
> >  endif
> > @@ -35,3 +36,4 @@ pahole-flags-$(CONFIG_PAHOLE_HAS_LANG_EXCLUDE)              += --lang_exclude=rust
> >
> >  export PAHOLE_FLAGS := $(pahole-flags-y)
> >  export MODULE_PAHOLE_FLAGS := $(module-pahole-flags-y)
> > +export MODULE_RESOLVE_BTFID_FLAGS := $(module-resolve_btfid-flags-y)
> > diff --git a/scripts/Makefile.modfinal b/scripts/Makefile.modfinal
> > index 542ba462ed3e..4481dda2f485 100644
> > --- a/scripts/Makefile.modfinal
> > +++ b/scripts/Makefile.modfinal
> > @@ -40,6 +40,7 @@ quiet_cmd_btf_ko = BTF [M] $@
> >               printf "Skipping BTF generation for %s due to unavailability of vmlinux\n" $@ 1>&2; \
> >       else                                                            \
> >               LLVM_OBJCOPY="$(OBJCOPY)" $(PAHOLE) -J $(PAHOLE_FLAGS) $(MODULE_PAHOLE_FLAGS) --btf_base $(objtree)/vmlinux $@; \
> > +             $(RESOLVE_BTFIDS) -b $(objtree)/vmlinux $(MODULE_RESOLVE_BTFID_FLAGS) --btf_sort $@;    \
>                                                                               ^^^^^^^^^^
>                                              Agree with Ihor (and off-list discussion with Alexei),
>                                              this flag appears redundant. Are there situations when
>                                              one would like to avoid sorting kernel/module BTF?

Thanks, I will make sorting unconditional.

>
> >               $(RESOLVE_BTFIDS) -b $(objtree)/vmlinux $@;             \
> >       fi;
>
> [...]
>
> > diff --git a/tools/bpf/resolve_btfids/main.c b/tools/bpf/resolve_btfids/main.c
> > index d47191c6e55e..dc0badd6f375 100644
> > --- a/tools/bpf/resolve_btfids/main.c
> > +++ b/tools/bpf/resolve_btfids/main.c
>
> [...]
>
> > +static int update_btf_section(const char *path, const struct btf *btf,
> > +                               const char *btf_secname)
> > +{
> > +     GElf_Shdr shdr_mem, *shdr;
> > +     Elf_Data *btf_data = NULL;
> > +     Elf_Scn *scn = NULL;
> > +     Elf *elf = NULL;
> > +     const void *raw_btf_data;
> > +     uint32_t raw_btf_size;
> > +     int fd, err = -1;
> > +     size_t strndx;
> > +
> > +     fd = open(path, O_RDWR);
> > +     if (fd < 0) {
> > +             pr_err("FAILED to open %s\n", path);
> > +             return -1;
> > +     }
> > +
> > +     if (elf_version(EV_CURRENT) == EV_NONE) {
> > +             pr_err("FAILED to set libelf version");
> > +             goto out;
> > +     }
> > +
> > +     elf = elf_begin(fd, ELF_C_RDWR, NULL);
> > +     if (elf == NULL) {
> > +             pr_err("FAILED to update ELF file");
> > +             goto out;
> > +     }
> > +
> > +     elf_flagelf(elf, ELF_C_SET, ELF_F_LAYOUT);
> > +
> > +     elf_getshdrstrndx(elf, &strndx);
> > +     while ((scn = elf_nextscn(elf, scn)) != NULL) {
> > +             char *secname;
> > +
> > +             shdr = gelf_getshdr(scn, &shdr_mem);
> > +             if (shdr == NULL)
> > +                     continue;
> > +             secname = elf_strptr(elf, strndx, shdr->sh_name);
> > +             if (strcmp(secname, btf_secname) == 0) {
> > +                     btf_data = elf_getdata(scn, btf_data);
> > +                     break;
> > +             }
> > +     }
> > +
> > +     raw_btf_data = btf__raw_data(btf, &raw_btf_size);
>
> `btf__raw_data()` can return NULL.

Thanks, I will fix it.

>
> > +
> > +     if (btf_data) {
> > +             if (raw_btf_size != btf_data->d_size) {
> > +                     pr_err("FAILED: size mismatch");
> > +                     goto out;
> > +             }
> > +
> > +             btf_data->d_buf = (void *)raw_btf_data;
> > +             btf_data->d_type = ELF_T_WORD;
> > +             elf_flagdata(btf_data, ELF_C_SET, ELF_F_DIRTY);
> > +
> > +             if (elf_update(elf, ELF_C_WRITE) >= 0)
> > +                     err = 0;
> > +     }
> > +
> > +out:
> > +     if (fd != -1)
> > +             close(fd);
> > +     if (elf)
> > +             elf_end(elf);
> > +     return err;
> > +}
>
> [...]