* [PATCH] newlib: undefine _FORTIFY_SOURCE
@ 2021-11-09 2:56 Mike Frysinger
2021-11-09 2:59 ` [PATCH alternative] newlib: ignore _FORTIFY_SOURCE when building newlib Mike Frysinger
2021-11-09 3:15 ` [PATCH] newlib: undefine _FORTIFY_SOURCE Keith Packard
0 siblings, 2 replies; 6+ messages in thread
From: Mike Frysinger @ 2021-11-09 2:56 UTC (permalink / raw)
To: newlib
Some distros enable _FORTIFY_SOURCE by default which upsets building
newlib which itself implements the logic for this define. For example,
building gets.c fails because the includes set up a gets() macro which
expands in the definition.
Since newlib isn't prepared to build itself with _FORTIFY_SOURCE, and
it's not clear if it's even useful, disable it when building the code.
This also matches what glibc is doing.
---
newlib/configure.host | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/newlib/configure.host b/newlib/configure.host
index e737f73f5f75..cb0c5d01f874 100644
--- a/newlib/configure.host
+++ b/newlib/configure.host
@@ -54,7 +54,7 @@
# have_init_fini have init/fini ("yes" or "no", set to "yes" by default)
# noinclude list of include files to not install
-newlib_cflags="-D_COMPILING_NEWLIB"
+newlib_cflags="-D_COMPILING_NEWLIB -U_FORTIFY_SOURCE"
libm_machine_dir=
machine_dir=
shared_machine_dir=
--
2.33.0
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH alternative] newlib: ignore _FORTIFY_SOURCE when building newlib
2021-11-09 2:56 [PATCH] newlib: undefine _FORTIFY_SOURCE Mike Frysinger
@ 2021-11-09 2:59 ` Mike Frysinger
2021-11-12 10:28 ` Corinna Vinschen
2021-11-09 3:15 ` [PATCH] newlib: undefine _FORTIFY_SOURCE Keith Packard
1 sibling, 1 reply; 6+ messages in thread
From: Mike Frysinger @ 2021-11-09 2:59 UTC (permalink / raw)
To: newlib
Some distros enable _FORTIFY_SOURCE by default which upsets building
newlib which itself implements the logic for this define. For example,
building gets.c fails because the includes set up a gets() macro which
expands in the definition.
Since newlib isn't prepared to build itself with _FORTIFY_SOURCE, and
it's not clear if it's even useful, ignore it when building the code.
This also matches what glibc is doing.
---
NB: This is an alternative approach instead of passing -U_FORTIFY_SOURCE.
I think the -U approach is better, but throwing this up for discussion.
newlib/libc/include/sys/features.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/newlib/libc/include/sys/features.h b/newlib/libc/include/sys/features.h
index 2188071785f6..6b4999e83482 100644
--- a/newlib/libc/include/sys/features.h
+++ b/newlib/libc/include/sys/features.h
@@ -320,7 +320,8 @@ extern "C" {
#endif
#if _FORTIFY_SOURCE > 0 && !defined(__cplusplus) && !defined(__lint__) && \
- (__OPTIMIZE__ > 0 || defined(__clang__)) && __GNUC_PREREQ__(4, 1)
+ (__OPTIMIZE__ > 0 || defined(__clang__)) && __GNUC_PREREQ__(4, 1) && \
+ !defined(_COMPILING_NEWLIB)
# if _FORTIFY_SOURCE > 1
# define __SSP_FORTIFY_LEVEL 2
# else
--
2.33.0
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] newlib: undefine _FORTIFY_SOURCE
2021-11-09 2:56 [PATCH] newlib: undefine _FORTIFY_SOURCE Mike Frysinger
2021-11-09 2:59 ` [PATCH alternative] newlib: ignore _FORTIFY_SOURCE when building newlib Mike Frysinger
@ 2021-11-09 3:15 ` Keith Packard
2021-11-09 11:41 ` Corinna Vinschen
2021-11-10 2:07 ` Mike Frysinger
1 sibling, 2 replies; 6+ messages in thread
From: Keith Packard @ 2021-11-09 3:15 UTC (permalink / raw)
To: Mike Frysinger, newlib
[-- Attachment #1: Type: text/plain, Size: 692 bytes --]
Mike Frysinger <vapier@gentoo.org> writes:
> Since newlib isn't prepared to build itself with _FORTIFY_SOURCE, and
> it's not clear if it's even useful, disable it when building the code.
It is useful; I've caught a few library bugs when I turned this on for
picolibc. Fixing newlib to handle -D_FORTIFY_SOURCE could probably start
with the patches I did in picolibc, which mostly just #undef the various
mem and str functions before defining them.
General code:
https://github.com/picolibc/picolibc/commit/c2c593afa60b64bff0a958ad930e0336463db5bf
Arm specific code:
https://github.com/picolibc/picolibc/commit/b3946840fb3a677125bc9232ca93ea7533375188
--
-keith
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] newlib: undefine _FORTIFY_SOURCE
2021-11-09 3:15 ` [PATCH] newlib: undefine _FORTIFY_SOURCE Keith Packard
@ 2021-11-09 11:41 ` Corinna Vinschen
2021-11-10 2:07 ` Mike Frysinger
1 sibling, 0 replies; 6+ messages in thread
From: Corinna Vinschen @ 2021-11-09 11:41 UTC (permalink / raw)
To: newlib
On Nov 8 19:15, Keith Packard wrote:
> Mike Frysinger <vapier@gentoo.org> writes:
>
> > Since newlib isn't prepared to build itself with _FORTIFY_SOURCE, and
> > it's not clear if it's even useful, disable it when building the code.
>
> It is useful; I've caught a few library bugs when I turned this on for
> picolibc. Fixing newlib to handle -D_FORTIFY_SOURCE could probably start
> with the patches I did in picolibc, which mostly just #undef the various
> mem and str functions before defining them.
Please send patches.
Thanks,
Corinna
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] newlib: undefine _FORTIFY_SOURCE
2021-11-09 3:15 ` [PATCH] newlib: undefine _FORTIFY_SOURCE Keith Packard
2021-11-09 11:41 ` Corinna Vinschen
@ 2021-11-10 2:07 ` Mike Frysinger
1 sibling, 0 replies; 6+ messages in thread
From: Mike Frysinger @ 2021-11-10 2:07 UTC (permalink / raw)
To: Keith Packard; +Cc: newlib
[-- Attachment #1: Type: text/plain, Size: 649 bytes --]
On 08 Nov 2021 19:15, Keith Packard wrote:
> Mike Frysinger <vapier@gentoo.org> writes:
> > Since newlib isn't prepared to build itself with _FORTIFY_SOURCE, and
> > it's not clear if it's even useful, disable it when building the code.
>
> It is useful; I've caught a few library bugs when I turned this on for
> picolibc. Fixing newlib to handle -D_FORTIFY_SOURCE could probably start
> with the patches I did in picolibc, which mostly just #undef the various
> mem and str functions before defining them.
i'm not against making it work, but it doesn't now, so we should disable the
logic if/until it's made to actually work.
-mike
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH alternative] newlib: ignore _FORTIFY_SOURCE when building newlib
2021-11-09 2:59 ` [PATCH alternative] newlib: ignore _FORTIFY_SOURCE when building newlib Mike Frysinger
@ 2021-11-12 10:28 ` Corinna Vinschen
0 siblings, 0 replies; 6+ messages in thread
From: Corinna Vinschen @ 2021-11-12 10:28 UTC (permalink / raw)
To: newlib
On Nov 8 21:59, Mike Frysinger wrote:
> Some distros enable _FORTIFY_SOURCE by default which upsets building
> newlib which itself implements the logic for this define. For example,
> building gets.c fails because the includes set up a gets() macro which
> expands in the definition.
>
> Since newlib isn't prepared to build itself with _FORTIFY_SOURCE, and
> it's not clear if it's even useful, ignore it when building the code.
> This also matches what glibc is doing.
> ---
> NB: This is an alternative approach instead of passing -U_FORTIFY_SOURCE.
> I think the -U approach is better, but throwing this up for discussion.
>
> newlib/libc/include/sys/features.h | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/newlib/libc/include/sys/features.h b/newlib/libc/include/sys/features.h
> index 2188071785f6..6b4999e83482 100644
> --- a/newlib/libc/include/sys/features.h
> +++ b/newlib/libc/include/sys/features.h
> @@ -320,7 +320,8 @@ extern "C" {
> #endif
>
> #if _FORTIFY_SOURCE > 0 && !defined(__cplusplus) && !defined(__lint__) && \
> - (__OPTIMIZE__ > 0 || defined(__clang__)) && __GNUC_PREREQ__(4, 1)
> + (__OPTIMIZE__ > 0 || defined(__clang__)) && __GNUC_PREREQ__(4, 1) && \
> + !defined(_COMPILING_NEWLIB)
> # if _FORTIFY_SOURCE > 1
> # define __SSP_FORTIFY_LEVEL 2
> # else
> --
> 2.33.0
I like this one more. Most people have no problems to look into
source, but many shy away from the convoluted build system.
Thanks,
Corinna
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2021-11-12 10:28 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-11-09 2:56 [PATCH] newlib: undefine _FORTIFY_SOURCE Mike Frysinger
2021-11-09 2:59 ` [PATCH alternative] newlib: ignore _FORTIFY_SOURCE when building newlib Mike Frysinger
2021-11-12 10:28 ` Corinna Vinschen
2021-11-09 3:15 ` [PATCH] newlib: undefine _FORTIFY_SOURCE Keith Packard
2021-11-09 11:41 ` Corinna Vinschen
2021-11-10 2:07 ` Mike Frysinger
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).