public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: "H.J. Lu" <hjl.tools@gmail.com>
To: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cc: GNU C Library <libc-alpha@sourceware.org>,
	Joseph Myers <joseph@codesourcery.com>
Subject: Re: [PATCH v6 3/5] Add GLIBC_ABI_DT_RELR for DT_RELR support
Date: Wed, 30 Mar 2022 07:41:35 -0700	[thread overview]
Message-ID: <CAMe9rOp0ZNLZ1Ye0FH1npf78HRZxRr+Z5PV8YEene2LRk4WS3A@mail.gmail.com> (raw)
In-Reply-To: <066eedff-b1d4-ec3c-3716-5846e2646f9a@linaro.org>

On Wed, Mar 30, 2022 at 7:18 AM Adhemerval Zanella
<adhemerval.zanella@linaro.org> wrote:
>
>
>
> On 29/03/2022 19:29, H.J. Lu wrote:
> > On Tue, Mar 29, 2022 at 9:52 AM Adhemerval Zanella
> > <adhemerval.zanella@linaro.org> wrote:
> >>
> >>
> >>
> >> On 10/03/2022 17:03, H.J. Lu via Libc-alpha wrote:
> >>> The EI_ABIVERSION field of the ELF header in executables and shared
> >>> libraries can be bumped to indicate the minimum ABI requirement on the
> >>> dynamic linker.  However, EI_ABIVERSION in executables isn't checked by
> >>> the Linux kernel ELF loader nor the existing dynamic linker.  Executables
> >>> will crash mysteriously if the dynamic linker doesn't support the ABI
> >>> features required by the EI_ABIVERSION field.  The dynamic linker should
> >>> be changed to check EI_ABIVERSION in executables.
> >>>
> >>> Add a glibc version, GLIBC_ABI_DT_RELR, to indicate DT_RELR support so
> >>> that the existing dynamic linkers will issue an error on executables with
> >>> GLIBC_ABI_DT_RELR dependency.  Issue an error if there is a DT_RELR entry
> >>> without GLIBC_ABI_DT_RELR dependency nor GLIBC_PRIVATE definition.
> >>>
> >>> Support __placeholder_only_for_empty_version_map as the placeholder symbol
> >>> used only for empty version map to generate GLIBC_ABI_DT_RELR without any
> >>> symbols.
> >>
> >> I think it only make sense to add the GLIBC_ABI_DT_RELR if the ABI actually
> >> supports DT_RELR, otherwise if the ABI start to support old glibcs might
> >> wrongly assumes DT_RELR and fails with undefined behavior at runtime
> >> (which this patch essentially tries to avoid).
> >
> > The DT_RELR set enables DT_RELR in glibc for all targets.
> If I understood correctly, if we add GLIBC_ABI_DT_RELR on 2.36 but only enable
> it on some architecture on 2.37 and try to run a binary with DT_RELR enabled
> on glibc 2.36, _dl_check_map_versions won't accuse a failure and it will
> eventually only fail at runtime (since relocation won't be applied).  My
> understanding we are trying to avoid such failures.

My patch set enables DT_RELR and GLIBC_ABI_DT_RELR on glibc 2.36
for all, regardless of binutils version.   There is no per-arch
behavior.    Only
DT_RELR run-time tests are enabled for linkers with -z pack-relative-relocs.

> >
> >>> ---
> >>>  elf/Makefile         | 12 ++++++++++--
> >>>  elf/Versions         |  5 +++++
> >>>  elf/dl-version.c     | 33 +++++++++++++++++++++++++++++++--
> >>>  include/link.h       |  6 ++++++
> >>>  scripts/abilist.awk  |  2 ++
> >>>  scripts/versions.awk |  7 ++++++-
> >>>  6 files changed, 60 insertions(+), 5 deletions(-)
> >>>
> >>> diff --git a/elf/Makefile b/elf/Makefile
> >>> index fd462ba315..d3118b9777 100644
> >>> --- a/elf/Makefile
> >>> +++ b/elf/Makefile
> >>> @@ -1113,8 +1113,10 @@ $(eval $(call include_dsosort_tests,dso-sort-tests-1.def))
> >>>  $(eval $(call include_dsosort_tests,dso-sort-tests-2.def))
> >>>  endif
> >>>
> >>> -check-abi: $(objpfx)check-abi-ld.out
> >>> -tests-special += $(objpfx)check-abi-ld.out
> >>> +check-abi: $(objpfx)check-abi-ld.out \
> >>> +        $(objpfx)check-abi-version-libc.out
> >>> +tests-special += $(objpfx)check-abi-ld.out \
> >>> +        $(objpfx)check-abi-version-libc.out
> >>
> >> One line per definition.
> >
> > Fixed in v7.
> >
> >>>  update-abi: update-abi-ld
> >>>  update-all-abi: update-all-abi-ld
> >>>
> >>> @@ -2739,3 +2741,9 @@ $(objpfx)check-tst-relr-pie.out: $(objpfx)tst-relr-pie
> >>>               | sed -ne '/required from libc.so/,$$ p' \
> >>>               | grep GLIBC_ABI_DT_RELR > $@; \
> >>>       $(evaluate-test)
> >>> +
> >>> +$(objpfx)check-abi-version-libc.out: $(common-objpfx)libc.so
> >>> +     LC_ALL=C $(READELF) -V -W $< \
> >>> +             | sed -ne '/.gnu.version_d/, /.gnu.version_r/ p' \
> >>> +             | grep GLIBC_ABI_DT_RELR > $@; \
> >>> +     $(evaluate-test)
> >>
> >> This test should be enable iff for $(have-dt-relr) == yes.
> >
> > This test checks if libc.so has the GLIBC_ABI_DT_RELR
> > version definition.   Since it doesn't require a new linker,
> > there is no need to check $(have-dt-relr).   This test
> > passes with an old linker.
> >
> >>> diff --git a/elf/Versions b/elf/Versions
> >>> index 8bed855d8c..a9ff278de7 100644
> >>> --- a/elf/Versions
> >>> +++ b/elf/Versions
> >>> @@ -23,6 +23,11 @@ libc {
> >>>    GLIBC_2.35 {
> >>>      _dl_find_object;
> >>>    }
> >>> +  GLIBC_ABI_DT_RELR {
> >>> +    # This symbol is used only for empty version map and will be removed
> >>> +    # by scripts/versions.awk.
> >>> +    __placeholder_only_for_empty_version_map;
> >>> +  }
> >>
> >> Same as before. Maybe use the same strategy we do for time64: add a config.h
> >> define to HAVE_DTRELR (set by configure) and use it to set the symbol if
> >> it is defined.
> >>
> >> This define would be used on dl-version.c to also enable the l_abi_version
> >> check.
> >>
> >>>    GLIBC_PRIVATE {
> >>>      # functions used in other libraries
> >>>      __libc_early_init;
> >>> diff --git a/elf/dl-version.c b/elf/dl-version.c
> >>> index b47bd91727..720ec596a5 100644
> >>> --- a/elf/dl-version.c
> >>> +++ b/elf/dl-version.c
> >>> @@ -214,12 +214,20 @@ _dl_check_map_versions (struct link_map *map, int verbose, int trace_mode)
> >>>             while (1)
> >>>               {
> >>>                 /* Match the symbol.  */
> >>> +               const char *string = strtab + aux->vna_name;
> >>>                 result |= match_symbol (DSO_FILENAME (map->l_name),
> >>>                                         map->l_ns, aux->vna_hash,
> >>> -                                       strtab + aux->vna_name,
> >>> -                                       needed->l_real, verbose,
> >>> +                                       string, needed->l_real, verbose,
> >>>                                         aux->vna_flags & VER_FLG_WEAK);
> >>>
> >>> +               if (map->l_abi_version == lav_none
> >>> +                   /* 0xfd0e42: _dl_elf_hash ("GLIBC_ABI_DT_RELR").  */
> >>> +                   && aux->vna_hash == 0xfd0e42
> >>> +                   && __glibc_likely (strcmp (string,
> >>> +                                              "GLIBC_ABI_DT_RELR")
> >>> +                                      == 0))
> >>> +                 map->l_abi_version = lav_dt_relr_ref;
> >>> +
> >>>                 /* Compare the version index.  */
> >>>                 if ((unsigned int) (aux->vna_other & 0x7fff) > ndx_high)
> >>>                   ndx_high = aux->vna_other & 0x7fff;
> >>> @@ -253,6 +261,16 @@ _dl_check_map_versions (struct link_map *map, int verbose, int trace_mode)
> >>>        ent = (ElfW(Verdef) *) (map->l_addr + def->d_un.d_ptr);
> >>>        while (1)
> >>>       {
> >>> +       /* 0x0963cf85: _dl_elf_hash ("GLIBC_PRIVATE").  */
> >>> +       if (ent->vd_hash == 0x0963cf85)
> >>> +         {
> >>> +           ElfW(Verdaux) *aux = (ElfW(Verdaux) *) ((char *) ent
> >>> +                                                   + ent->vd_aux);
> >>> +           if (__glibc_likely (strcmp ("GLIBC_PRIVATE",
> >>> +                                       strtab + aux->vda_name) == 0))
> >>> +             map->l_abi_version = lav_private_def;
> >>> +         }
> >>> +
> >>>         if ((unsigned int) (ent->vd_ndx & 0x7fff) > ndx_high)
> >>>           ndx_high = ent->vd_ndx & 0x7fff;
> >>>
> >>> @@ -352,6 +370,17 @@ _dl_check_map_versions (struct link_map *map, int verbose, int trace_mode)
> >>>       }
> >>>      }
> >>>
> >>> +  /* Issue an error if there is a DT_RELR entry without GLIBC_ABI_DT_RELR
> >>> +     dependency nor GLIBC_PRIVATE definition.  */
> >>> +  if (map->l_info[DT_RELR] != NULL
> >>> +      && __glibc_unlikely (map->l_abi_version == lav_none))
> >>> +    {
> >>> +      _dl_exception_create
> >>> +     (&exception, DSO_FILENAME (map->l_name),
> >>> +      N_("DT_RELR without GLIBC_ABI_DT_RELR dependency"));
> >>> +      goto call_error;
> >>> +    }
> >>> +
> >>>    return result;
> >>>  }
> >>>
> >>> diff --git a/include/link.h b/include/link.h
> >>> index 03db14c7b0..8ec5e35cf2 100644
> >>> --- a/include/link.h
> >>> +++ b/include/link.h
> >>> @@ -177,6 +177,12 @@ struct link_map
> >>>       lt_library,             /* Library needed by main executable.  */
> >>>       lt_loaded               /* Extra run-time loaded shared object.  */
> >>>        } l_type:2;
> >>> +    enum                     /* ABI dependency of this object.  */
> >>> +      {
> >>> +     lav_none,               /* No ABI dependency.  */
> >>> +     lav_dt_relr_ref,        /* Need GLIBC_ABI_DT_RELR.  */
> >>> +     lav_private_def         /* Define GLIBC_PRIVATE.  */
> >>> +      } l_abi_version:2;
> >>>      unsigned int l_relocated:1;      /* Nonzero if object's relocations done.  */
> >>>      unsigned int l_init_called:1; /* Nonzero if DT_INIT function called.  */
> >>>      unsigned int l_global:1; /* Nonzero if object in _dl_global_scope.  */
> >>> diff --git a/scripts/abilist.awk b/scripts/abilist.awk
> >>> index 24a34ccbed..6cc7af6ac8 100644
> >>> --- a/scripts/abilist.awk
> >>> +++ b/scripts/abilist.awk
> >>> @@ -55,6 +55,8 @@ $2 == "g" || $2 == "w" && (NF == 7 || NF == 8) {
> >>>    # caused STV_HIDDEN symbols to appear in .dynsym, though that is useless.
> >>>    if (NF > 7 && $7 == ".hidden") next;
> >>>
> >>> +  if (version ~ /^GLIBC_ABI_/ && !include_abi_version) next;
> >>> +
> >>>    if (version == "GLIBC_PRIVATE" && !include_private) next;
> >>>
> >>>    desc = "";
> >>> diff --git a/scripts/versions.awk b/scripts/versions.awk
> >>> index 357ad1355e..d70b07bd1a 100644
> >>> --- a/scripts/versions.awk
> >>> +++ b/scripts/versions.awk
> >>> @@ -185,8 +185,13 @@ END {
> >>>       closeversion(oldver, veryoldver);
> >>>       veryoldver = oldver;
> >>>        }
> >>> -      printf("%s {\n  global:\n", $2) > outfile;
> >>>        oldver = $2;
> >>> +      # Skip the placeholder symbol used only for empty version map.
> >>> +      if ($3 == "__placeholder_only_for_empty_version_map;") {
> >>> +     printf("%s {\n", $2) > outfile;
> >>> +     continue;
> >>> +      }
> >>> +      printf("%s {\n  global:\n", $2) > outfile;
> >>>      }
> >>>      printf("   ") > outfile;
> >>>      for (n = 3; n <= NF; ++n) {
> >
> >
> >



-- 
H.J.

  reply	other threads:[~2022-03-30 14:42 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-10 20:03 [PATCH v6 0/5] Support DT_RELR relative relocation format H.J. Lu
2022-03-10 20:03 ` [PATCH v6 1/5] elf: Support DT_RELR relative relocation format [BZ #27924] H.J. Lu
2022-03-29 16:34   ` Adhemerval Zanella
2022-03-29 22:34     ` H.J. Lu
2022-03-10 20:03 ` [PATCH v6 2/5] elf: Properly handle zero DT_RELA/DT_REL values H.J. Lu
2022-03-29 16:38   ` Adhemerval Zanella
2022-03-29 22:30     ` H.J. Lu
2022-03-10 20:03 ` [PATCH v6 3/5] Add GLIBC_ABI_DT_RELR for DT_RELR support H.J. Lu
2022-03-29 16:52   ` Adhemerval Zanella
2022-03-29 22:29     ` H.J. Lu
2022-03-30 14:18       ` Adhemerval Zanella
2022-03-30 14:41         ` H.J. Lu [this message]
2022-03-30 17:17           ` Adhemerval Zanella
2022-03-30 17:32             ` H.J. Lu
2022-03-30 17:39               ` Adhemerval Zanella
2022-03-30 19:22             ` Joseph Myers
2022-03-30 20:38               ` Adhemerval Zanella
2022-03-30 21:32                 ` Fangrui Song
2022-03-10 20:03 ` [PATCH v6 4/5] Add --disable-default-dt-relr H.J. Lu
2022-03-29 17:19   ` Adhemerval Zanella
2022-03-29 22:25     ` H.J. Lu
2022-03-10 20:03 ` [PATCH v6 5/5] NEWS: Mention DT_RELR support H.J. Lu
2022-03-29 17:25   ` Adhemerval Zanella
2022-03-29 22:22     ` H.J. Lu
2022-03-29 23:34       ` Fangrui Song
2022-03-30 14:20         ` Adhemerval Zanella
2022-03-10 20:09 ` [PATCH v6 0/5] Support DT_RELR relative relocation format Fangrui Song

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=CAMe9rOp0ZNLZ1Ye0FH1npf78HRZxRr+Z5PV8YEene2LRk4WS3A@mail.gmail.com \
    --to=hjl.tools@gmail.com \
    --cc=adhemerval.zanella@linaro.org \
    --cc=joseph@codesourcery.com \
    --cc=libc-alpha@sourceware.org \
    /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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).