public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: libc-alpha@sourceware.org
Subject: Re: [PATCH v3 10/13] aarch64: configure check for pac-ret code generation
Date: Mon, 25 May 2020 18:49:09 -0300	[thread overview]
Message-ID: <17d35e4b-5eed-06e2-48e3-81cfc28dfc27@linaro.org> (raw)
In-Reply-To: <e6d61e792176ff26969e04f70fb6c431461fa13b.1589552055.git.szabolcs.nagy@arm.com>



On 15/05/2020 11:40, Szabolcs Nagy wrote:
> Return address signing requires unwinder support, which is
> present in libgcc since >=gcc-7, however due to bugs the
> support may be broken in <gcc-10 (and similarly there may
> be issues in custom unwinders), so pac-ret is not always
> safe to use. So in assembly code glibc should only use
> pac-ret if the compiler uses it too. Unfortunately there
> is no predefined feature macro for it set by the compiler
> so pac-ret is inferred from the code generation.

OK with this fixes below.

Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>

> ---
>  config.h.in                  |  3 +++
>  sysdeps/aarch64/configure    | 39 ++++++++++++++++++++++++++++++++++++
>  sysdeps/aarch64/configure.ac | 21 +++++++++++++++++++
>  3 files changed, 63 insertions(+)
> 
> diff --git a/config.h.in b/config.h.in
> index 506b0c416c..f441470385 100644
> --- a/config.h.in
> +++ b/config.h.in
> @@ -112,6 +112,9 @@
>  /* AArch64 BTI support enabled.  */
>  #undef HAVE_AARCH64_BTI
>  
> +/* AArch64 PAC-RET code generation is enabled.  */
> +#undef HAVE_AARCH64_PAC_RET
> +
>  /* C-SKY ABI version.  */
>  #undef CSKYABI
>  

Use

  #define HAVE_AARCH64_PAC_RET 0

instead so we can check the define value instead of the its existence.

> diff --git a/sysdeps/aarch64/configure b/sysdeps/aarch64/configure
> index 70477a7fa5..5f1cdf5d9b 100644
> --- a/sysdeps/aarch64/configure
> +++ b/sysdeps/aarch64/configure
> @@ -214,3 +214,42 @@ if test $libc_cv_aarch64_bti = yes; then
>    $as_echo "#define HAVE_AARCH64_BTI 1" >>confdefs.h
>  
>  fi
> +
> +# Check if glibc is built with return address signing, i.e.
> +# if -mbranch-protection=pac-ret is on. We need this because
> +# pac-ret relies on unwinder support so it's not safe to use
> +# it in assembly code unconditionally, but there is no
> +# feature test macro for it in gcc.
> +{ $as_echo "$as_me:${as_lineno-$LINENO}: checking if pac-ret is enabled" >&5
> +$as_echo_n "checking if pac-ret is enabled... " >&6; }
> +if ${libc_cv_aarch64_pac_ret+:} false; then :
> +  $as_echo_n "(cached) " >&6
> +else
> +    cat > conftest.c <<EOF
> +int bar (void);
> +int foo (void) { return bar () + 1; }
> +EOF
> +  libc_cv_aarch64_pac_ret=no
> +  if { ac_try='${CC-cc} $CFLAGS $CPPFLAGS -S -o conftest.s conftest.c'
> +  { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5
> +  (eval $ac_try) 2>&5
> +  ac_status=$?
> +  $as_echo "$as_me:${as_lineno-$LINENO}: \$? = $ac_status" >&5
> +  test $ac_status = 0; }; } \
> +     && { ac_try='grep -q -E '\''(hint( |	)+25|paciasp)'\'' conftest.s'
> +  { { eval echo "\"\$as_me\":${as_lineno-$LINENO}: \"$ac_try\""; } >&5
> +  (eval $ac_try) 2>&5
> +  ac_status=$?
> +  $as_echo "$as_me:${as_lineno-$LINENO}: \$? = $ac_status" >&5
> +  test $ac_status = 0; }; }
> +  then
> +    libc_cv_aarch64_pac_ret=yes
> +  fi
> +  rm -rf conftest.*
> +fi
> +{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $libc_cv_aarch64_pac_ret" >&5
> +$as_echo "$libc_cv_aarch64_pac_ret" >&6; }
> +if test $libc_cv_aarch64_pac_ret = yes; then
> +  $as_echo "#define HAVE_AARCH64_PAC_RET 1" >>confdefs.h
> +
> +fi
> diff --git a/sysdeps/aarch64/configure.ac b/sysdeps/aarch64/configure.ac
> index 798f494740..8248ecf2ed 100644
> --- a/sysdeps/aarch64/configure.ac
> +++ b/sysdeps/aarch64/configure.ac
> @@ -39,3 +39,24 @@ EOF
>  if test $libc_cv_aarch64_bti = yes; then
>    AC_DEFINE(HAVE_AARCH64_BTI)
>  fi
> +
> +# Check if glibc is built with return address signing, i.e.
> +# if -mbranch-protection=pac-ret is on. We need this because

Two spaces after period.

> +# pac-ret relies on unwinder support so it's not safe to use
> +# it in assembly code unconditionally, but there is no
> +# feature test macro for it in gcc.
> +AC_CACHE_CHECK([if pac-ret is enabled], [libc_cv_aarch64_pac_ret], [dnl
> +  cat > conftest.c <<EOF
> +int bar (void);
> +int foo (void) { return bar () + 1; }
> +EOF
> +  libc_cv_aarch64_pac_ret=no
> +  if AC_TRY_COMMAND([${CC-cc} $CFLAGS $CPPFLAGS -S -o conftest.s conftest.c]) \
> +     && AC_TRY_COMMAND([grep -q -E '\''(hint( |	)+25|paciasp)'\'' conftest.s])

I am not sure this expression is doing what it is intended to do,
I think you are matching the gcc output because it also add an
extra comment with 'paciasp'.

I think what you need here is "(hint[[:space:]]+25|paciasp)".

> +  then
> +    libc_cv_aarch64_pac_ret=yes
> +  fi
> +  rm -rf conftest.*])
> +if test $libc_cv_aarch64_pac_ret = yes; then
> +  AC_DEFINE(HAVE_AARCH64_PAC_RET)
> +fi
> 

  reply	other threads:[~2020-05-25 21:49 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-15 14:40 [PATCH v3 00/13] aarch64: branch protection support Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 01/13] elf.h: Add PT_GNU_PROPERTY Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 02/13] elf.h: add aarch64 property definitions Szabolcs Nagy
2020-05-18 15:26   ` Florian Weimer
2020-05-21  8:57     ` Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 03/13] Rewrite abi-note.S in C Szabolcs Nagy
2020-05-18 15:28   ` Florian Weimer
2020-05-20 10:27     ` Szabolcs Nagy
2020-05-20 10:34       ` Florian Weimer
2020-05-21  8:54         ` [PATCH v4] " Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 04/13] aarch64: configure test for BTI support Szabolcs Nagy
2020-05-25 18:41   ` Adhemerval Zanella
2020-05-25 18:48     ` Adhemerval Zanella
2020-05-15 14:40 ` [PATCH v3 05/13] aarch64: Add BTI support to assembly files Szabolcs Nagy
2020-05-25 18:49   ` Adhemerval Zanella
2020-05-15 14:40 ` [PATCH v3 06/13] aarch64: Rename place holder .S files to .c Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 07/13] aarch64: fix swapcontext for BTI Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 08/13] aarch64: fix RTLD_START " Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 09/13] aarch64: enable BTI at runtime Szabolcs Nagy
2020-05-25 19:53   ` Adhemerval Zanella
2020-05-26 11:20     ` Szabolcs Nagy
2020-05-15 14:40 ` [PATCH v3 10/13] aarch64: configure check for pac-ret code generation Szabolcs Nagy
2020-05-25 21:49   ` Adhemerval Zanella [this message]
2020-05-15 14:40 ` [PATCH v3 11/13] aarch64: Add pac-ret support to assembly files Szabolcs Nagy
2020-05-26 11:26   ` Adhemerval Zanella
2020-05-15 14:40 ` [PATCH v3 12/13] aarch64: redefine RETURN_ADDRESS to strip PAC Szabolcs Nagy
2020-05-26 11:29   ` Adhemerval Zanella
2020-05-15 14:40 ` [PATCH v3 13/13] aarch64: fix _mcount for pac-ret Szabolcs Nagy
2020-05-26 11:33   ` Adhemerval Zanella
2020-05-26 18:38     ` Szabolcs Nagy
2020-05-26 19:13       ` Adhemerval Zanella

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=17d35e4b-5eed-06e2-48e3-81cfc28dfc27@linaro.org \
    --to=adhemerval.zanella@linaro.org \
    --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).