linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Masahiro Yamada <yamada.masahiro@socionext.com>
To: paul.walmsley@sifive.com
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Linux Kbuild mailing list <linux-kbuild@vger.kernel.org>,
	Russell King <linux@armlinux.org.uk>,
	jimw@sifive.com, Michal Marek <michal.lkml@markovi.net>,
	Paul Walmsley <paul@pwsan.com>
Subject: Re: [PATCH 2/2] modpost: skip ELF local symbols by default during section mismatch check
Date: Tue, 13 Nov 2018 11:19:15 +0900	[thread overview]
Message-ID: <CAK7LNARC2+BLFay_-CtaJXKCu3M0NZzxbeff1XandeH58BbUcg@mail.gmail.com> (raw)
In-Reply-To: <20181020131911.22443-3-paul.walmsley@sifive.com>

Hi Paul,


On Sat, Oct 20, 2018 at 10:21 PM Paul Walmsley <paul.walmsley@sifive.com> wrote:
>
> During development of a serial console driver with a RISC-V toolchain,
> the following modpost warning appeared:
>
> ----
> WARNING: vmlinux.o(.data+0x19b10): Section mismatch in reference from the variable .LANCHOR1 to the function .init.text:sifive_serial_console_setup()
> The variable .LANCHOR1 references
> the function __init sifive_serial_console_setup()
> If the reference is valid then annotate the
> variable with __init* or __refdata (see linux/init.h) or name the variable:
> *_template, *_timer, *_sht, *_ops, *_probe, *_probe_one, *_console
> ----

Could you provide me a little more information to reproduce it?

I tried your dev/paulw/serial-v4.19-rc7,
but I could not get that warning.

I used risc64-linux-gcc (GCC 7.3, 8.1) from kernel.org





> ".LANCHOR1" is an ELF local symbol, automatically created by gcc's section
> anchor generation code:
>
> https://gcc.gnu.org/onlinedocs/gccint/Anchored-Addresses.html
>
> https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/varasm.c;h=cd9591a45617464946dcf9a126dde277d9de9804;hb=9fb89fa845c1b2e0a18d85ada0b077c84508ab78#l7473
>
> This was verified by compiling the kernel with -fno-section-anchors.
> The serial driver code idiom triggering the warning is standard serial
> driver practice, and one that has a specific whitelist inclusion in
> modpost.c.
>
> I'm neither a modpost nor an ELF expert, but naively, it doesn't seem
> useful for modpost to report section mismatch warnings caused by ELF
> local symbols by default.  Local symbols have compiler-generated
> names, and thus bypass modpost's whitelisting algorithm, which relies
> on the presence of a non-autogenerated symbol name.  This increases
> the likelihood that false positive warnings will be generated (as in
> the above case).
>
> Thus, disable section mismatch reporting on ELF local symbols by
> default.  The rationale here is similar to that of commit
> 2e3a10a1551d6ceea005e6a62ca58183b8976217 ("ARM: avoid ARM binutils
> leaking ELF local symbols") and of similar code already present in
> modpost.c:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/scripts/mod/modpost.c?h=v4.19-rc4&id=7876320f88802b22d4e2daf7eb027dd14175a0f8#n1256
>
> Since it might be useful for some static analysis runs to detect
> whether any new symbols have been added that result in any section
> mismatches, even from ELF local symbols, this change can be disabled
> (along with other tests that can often generate false positives) by
> passing the '-P' switch on the modpost command line.
>
> Cc: Russell King <linux@armlinux.org.uk>
> Cc: Jim Wilson <jimw@sifive.com>
> Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
> Cc: Michal Marek <michal.lkml@markovi.net>
> Cc: linux-kbuild@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
> Signed-off-by: Paul Walmsley <paul@pwsan.com>
> ---
>  scripts/mod/modpost.c | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
>
> diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c
> index 38fc1bd47926..2b8c960ece66 100644
> --- a/scripts/mod/modpost.c
> +++ b/scripts/mod/modpost.c
> @@ -1255,6 +1255,17 @@ static inline int is_arm_mapping_symbol(const char *str)
>                && (str[2] == '\0' || str[2] == '.');
>  }
>
> +/*
> + * If a symbol name follows the convention for ELF-local symbols (i.e., the
> + * name begins with a ".L"), return true; otherwise false.  This is used to
> + * skip section mismatch reporting on ELF-local symbols, due to the risk of
> + * false positives.
> + */
> +static inline int is_local_symbol(const char *str)
> +{
> +       return str[0] == '.' && str[1] == 'L';
> +}
> +
>  /*
>   * If there's no name there, ignore it; likewise, ignore it if it's
>   * one of the magic symbols emitted used by current ARM tools.
> @@ -1537,6 +1548,9 @@ static void default_mismatch_handler(const char *modname, struct elf_info *elf,
>         if (strstarts(fromsym, "reference___initcall"))
>                 return;
>
> +       if (!accept_falsepos_risk && is_local_symbol(fromsym))
> +               return;
> +
>         tosec = sec_name(elf, get_secindex(elf, sym));
>         to = find_elf_symbol(elf, r->r_addend, sym);
>         tosym = sym_name(elf, to);
> --
> 2.19.1
>


-- 
Best Regards
Masahiro Yamada

  reply	other threads:[~2018-11-13  2:20 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-20 13:19 [PATCH 0/2] modpost: skip section mismatch warnings on ELF local symbols by default Paul Walmsley
2018-10-20 13:19 ` [PATCH 1/2] modpost: add switch to skip symbol exclusions likely to generate false positives Paul Walmsley
2018-10-20 13:19 ` [PATCH 2/2] modpost: skip ELF local symbols by default during section mismatch check Paul Walmsley
2018-11-13  2:19   ` Masahiro Yamada [this message]
2018-11-14 19:30     ` Paul Walmsley
2018-10-20 14:08 ` [PATCH 0/2] modpost: skip section mismatch warnings on ELF local symbols by default Sam Ravnborg
2018-11-15  0:48   ` Paul Walmsley

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=CAK7LNARC2+BLFay_-CtaJXKCu3M0NZzxbeff1XandeH58BbUcg@mail.gmail.com \
    --to=yamada.masahiro@socionext.com \
    --cc=jimw@sifive.com \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux@armlinux.org.uk \
    --cc=michal.lkml@markovi.net \
    --cc=paul.walmsley@sifive.com \
    --cc=paul@pwsan.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;
as well as URLs for NNTP newsgroup(s).