From: Siddhesh Poyarekar <siddhesh@gotplt.org>
To: "Frédéric Bérat" <fberat@redhat.com>, libc-alpha@sourceware.org
Subject: Re: [PATCH v3 10/16] unistd: Avoid PLT entries with _FORTIFY_SOURCE
Date: Fri, 30 Jun 2023 11:25:44 -0400 [thread overview]
Message-ID: <bf56f16a-9407-8947-d15e-85edc351f45f@gotplt.org> (raw)
In-Reply-To: <20230628084246.778302-11-fberat@redhat.com>
On 2023-06-28 04:42, Frédéric Bérat wrote:
> The change is meant to avoid unwanted PLT entries for the read_chk,
> getdomainname_chk and getlogin_r_chk routines when _FORTIFY_SOURCE is set.
> ---
> debug/getdomainname_chk.c | 1 +
> debug/read_chk.c | 1 +
> include/unistd.h | 4 ++++
> login/getlogin_r_chk.c | 1 +
> posix/bits/unistd-decl.h | 6 +++---
> 5 files changed, 10 insertions(+), 3 deletions(-)
>
> diff --git a/debug/getdomainname_chk.c b/debug/getdomainname_chk.c
> index ccea1b2d9e..c0e6c7bd4b 100644
> --- a/debug/getdomainname_chk.c
> +++ b/debug/getdomainname_chk.c
> @@ -26,3 +26,4 @@ __getdomainname_chk (char *buf, size_t buflen, size_t nreal)
>
> return getdomainname (buf, buflen);
> }
> +libc_hidden_builtin_def (__getdomainname_chk)
getdomainname is not a builtin, so perhaps just use libc_hidden_def? It
doesn't make a difference in practice, but I reckon we may make a
distinction in future and this might slip up.
Builtins are defined in gcc sources, see gcc/builtins.def. As an aside
though, I wonder if we've ever made this distinction, but that's a QoI
issue at the moment.
> diff --git a/debug/read_chk.c b/debug/read_chk.c
> index 17588736a1..dfd87206b2 100644
> --- a/debug/read_chk.c
> +++ b/debug/read_chk.c
> @@ -25,3 +25,4 @@ __read_chk (int fd, void *buf, size_t nbytes, size_t buflen)
>
> return __read (fd, buf, nbytes);
> }
> +libc_hidden_builtin_def (__read_chk)
Likewise, not a builtin.
> diff --git a/include/unistd.h b/include/unistd.h
> index b042a485f9..243e52b471 100644
> --- a/include/unistd.h
> +++ b/include/unistd.h
> @@ -32,6 +32,10 @@ libc_hidden_proto (readlinkat)
> libc_hidden_proto (fsync)
> libc_hidden_proto (fdatasync)
>
> +libc_hidden_builtin_proto (__read_chk)
> +libc_hidden_builtin_proto (__getdomainname_chk)
> +libc_hidden_builtin_proto (__getlogin_r_chk)
Likewise, libc_hidden_proto.
> +
> /* Now define the internal interfaces. */
> extern int __access (const char *__name, int __type);
> libc_hidden_proto (__access)
> diff --git a/login/getlogin_r_chk.c b/login/getlogin_r_chk.c
> index 12a0d12bd4..576567f287 100644
> --- a/login/getlogin_r_chk.c
> +++ b/login/getlogin_r_chk.c
> @@ -26,3 +26,4 @@ __getlogin_r_chk (char *buf, size_t buflen, size_t nreal)
>
> return getlogin_r (buf, buflen);
> }
> +libc_hidden_builtin_def (__getlogin_r_chk)
Likewise.
> diff --git a/posix/bits/unistd-decl.h b/posix/bits/unistd-decl.h
> index 2d0f7a4456..2603039767 100644
> --- a/posix/bits/unistd-decl.h
> +++ b/posix/bits/unistd-decl.h
> @@ -26,7 +26,7 @@
> extern ssize_t __read_chk (int __fd, void *__buf, size_t __nbytes,
> size_t __buflen)
> __wur __attr_access ((__write_only__, 2, 3));
> -extern ssize_t __REDIRECT (__read_alias, (int __fd, void *__buf,
> +extern ssize_t __REDIRECT_FORTIFY (__read_alias, (int __fd, void *__buf,
> size_t __nbytes), read)
> __wur __attr_access ((__write_only__, 2, 3));
> extern ssize_t __REDIRECT (__read_chk_warn,
> @@ -157,7 +157,7 @@ extern int __REDIRECT_NTH (__ttyname_r_chk_warn,
> #ifdef __USE_POSIX199506
> extern int __getlogin_r_chk (char *__buf, size_t __buflen, size_t __nreal)
> __nonnull ((1)) __attr_access ((__write_only__, 1, 2));
> -extern int __REDIRECT (__getlogin_r_alias, (char *__buf, size_t __buflen),
> +extern int __REDIRECT_FORTIFY (__getlogin_r_alias, (char *__buf, size_t __buflen),
> getlogin_r) __nonnull ((1));
> extern int __REDIRECT (__getlogin_r_chk_warn,
> (char *__buf, size_t __buflen, size_t __nreal),
> @@ -184,7 +184,7 @@ extern int __REDIRECT_NTH (__gethostname_chk_warn,
> #if defined __USE_MISC || (defined __USE_XOPEN && !defined __USE_UNIX98)
> extern int __getdomainname_chk (char *__buf, size_t __buflen, size_t __nreal)
> __THROW __nonnull ((1)) __wur __attr_access ((__write_only__, 1, 2));
> -extern int __REDIRECT_NTH (__getdomainname_alias, (char *__buf,
> +extern int __REDIRECT_FORTIFY_NTH (__getdomainname_alias, (char *__buf,
> size_t __buflen),
> getdomainname) __nonnull ((1))
> __wur __attr_access ((__write_only__, 1, 2));
next prev parent reply other threads:[~2023-06-30 15:25 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-28 8:42 [PATCH v3 00/16] Allow glibc to be built " Frédéric Bérat
2023-06-28 8:42 ` [PATCH v3 01/16] " Frédéric Bérat
2023-06-28 14:48 ` Joseph Myers
2023-06-28 8:42 ` [PATCH v3 02/16] Exclude routines from fortification Frédéric Bérat
2023-06-30 14:55 ` Siddhesh Poyarekar
2023-07-03 15:16 ` Frederic Berat
2023-07-04 16:04 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 03/16] sysdeps: Ensure ieee128*_chk routines to be properly named Frédéric Bérat
2023-06-30 14:58 ` Siddhesh Poyarekar
2023-06-30 15:55 ` Paul E Murphy
2023-06-30 15:57 ` Frederic Berat
2023-06-28 8:42 ` [PATCH v3 04/16] string: Ensure *_chk routines have their hidden builtin definition available Frédéric Bérat
2023-06-30 15:06 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 05/16] stdio: " Frédéric Bérat
2023-06-30 15:09 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 06/16] asprintf_chk: Ensure compatibility for both s390x and ppc64le Frédéric Bérat
2023-06-30 15:11 ` Siddhesh Poyarekar
2023-06-30 16:08 ` Rajalakshmi Srinivasaraghavan
2023-06-30 17:51 ` Paul E Murphy
2023-07-03 5:35 ` Frederic Berat
2023-06-28 8:42 ` [PATCH v3 07/16] misc/sys/cdefs.h: Create FORTIFY redirects for internal calls Frédéric Bérat
2023-06-30 15:13 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 08/16] wchar: Avoid PLT entries with _FORTIFY_SOURCE Frédéric Bérat
2023-06-30 15:17 ` Siddhesh Poyarekar
2023-06-30 15:26 ` Frederic Berat
2023-06-28 8:42 ` [PATCH v3 09/16] posix/bits/unistd.h: Clearly separate declaration from definitions Frédéric Bérat
2023-06-30 15:19 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 10/16] unistd: Avoid PLT entries with _FORTIFY_SOURCE Frédéric Bérat
2023-06-30 15:25 ` Siddhesh Poyarekar [this message]
2023-06-28 8:42 ` [PATCH v3 11/16] misc/bits/select2.h: Clearly separate declaration from definitions Frédéric Bérat
2023-06-30 15:26 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 12/16] misc/bits/syslog.h: Clearly separate declaration from definition Frédéric Bérat
2023-06-30 15:28 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 13/16] libio/bits/stdio2.h: Clearly separate declaration from definitions Frédéric Bérat
2023-06-30 15:29 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 14/16] libio/bits/stdio2-decl.h: Avoid PLT entries with _FORTIFY_SOURCE Frédéric Bérat
2023-06-30 15:30 ` Siddhesh Poyarekar
2023-06-30 15:38 ` Frederic Berat
2023-06-30 15:48 ` Siddhesh Poyarekar
2023-06-30 17:08 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 15/16] sysdeps/ieee754/ldbl-128ibm-compat: Fix warn unused result Frédéric Bérat
2023-06-30 15:33 ` Siddhesh Poyarekar
2023-06-28 8:42 ` [PATCH v3 16/16] Add --enable-fortify-source option Frédéric Bérat
2023-06-30 13:51 ` Siddhesh Poyarekar
2023-07-03 8:50 ` Andreas Schwab
2023-07-03 12:51 ` Adhemerval Zanella Netto
2023-07-04 12:40 ` Frederic Berat
2023-07-04 15:59 ` Siddhesh Poyarekar
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=bf56f16a-9407-8947-d15e-85edc351f45f@gotplt.org \
--to=siddhesh@gotplt.org \
--cc=fberat@redhat.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).