From: David Laight <David.Laight@ACULAB.COM>
To: 'Zhangjin Wu' <falcon@tinylab.org>, "w@1wt.eu" <w@1wt.eu>
Cc: "arnd@arndb.de" <arnd@arndb.de>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"linux-kselftest@vger.kernel.org"
<linux-kselftest@vger.kernel.org>,
"thomas@t-8ch.de" <thomas@t-8ch.de>
Subject: RE: [PATCH v3] tools/nolibc: fix up size inflate regression
Date: Mon, 7 Aug 2023 08:40:27 +0000 [thread overview]
Message-ID: <f51e54bcf470451ea36f24640f000e61@AcuMS.aculab.com> (raw)
In-Reply-To: <8eaab5da2dcbba42e3f3efc2ae686a22c95f84f0.1691386601.git.falcon@tinylab.org>
From: Zhangjin Wu
> Sent: 07 August 2023 06:58
...
> +/* __auto_type is used instead of __typeof__ to workaround the build error
> + * 'error: assignment of read-only variable' when the argument has 'const' in
> + * the type, but __auto_type is a new feature from newer gcc version and it
> + * only works with 'const' from gcc 11.0 (__GXX_ABI_VERSION = 1016)
> + * https://gcc.gnu.org/legacy-ml/gcc-patches/2013-11/msg01378.html
> + */
You can use typeof((x) + 0) to lose the 'const' flag.
The only downside is that char/short become int.
> +
> +#if __GXX_ABI_VERSION >= 1016
> +#define __GXX_HAS_AUTO_TYPE_WITH_CONST_SUPPORT
> +#endif
> +
> +#ifdef __GXX_HAS_AUTO_TYPE_WITH_CONST_SUPPORT
> +#define __sysret(arg) \
> +({ \
> + __auto_type __ret = (arg); \
> + if (__is_signed_type(__typeof__(arg))) { \
> + if (__ret < 0) { \
> + SET_ERRNO(-(long)__ret); \
> + __ret = (__typeof__(arg))(-1L); \
> + } \
> + } else { \
> + if ((unsigned long)__ret >= (unsigned long)-MAX_ERRNO) { \
> + SET_ERRNO(-(long)__ret); \
> + __ret = (__typeof__(arg))(-1L); \
> + } \
> + } \
> + __ret; \
> +})
> +
> +#else /* ! __GXX_HAS_AUTO_TYPE_WITH_CONST_SUPPORT */
> +#define __sysret(arg) \
> +({ \
> + long __ret = (long)(arg); \
> + if (__is_signed_type(__typeof__(arg))) { \
> + if (__ret < 0) { \
> + SET_ERRNO(-__ret); \
> + __ret = -1L; \
> + } \
> + } else { \
> + if ((unsigned long)__ret >= (unsigned long)-MAX_ERRNO) { \
> + SET_ERRNO(-__ret); \
> + __ret = -1L; \
> + } \
> + } \
> + (__typeof__(arg))__ret; \
> +})
> +#endif /* ! __GXX_HAS_AUTO_TYPE_WITH_CONST_SUPPORT */
with (retyped so it may be wrong):
#define is_constexpr(x) sizeof(*(0 ? (void *)((long)(x) * 0) : (int *)0)) == 1)
and (because even (void *)0 isn't completely constant):
#define is_pointer(x) (!is_constexpr((typeof(x))0))
You can probably do:
#define __sysret(arg) \
({ \
typeof((arg) + 0) __ret = arg; \
if (__built_choose_expr(is_pointer(arg), (unsigned long)-(MAX_ERRNO+1), __ret) \
< (__built_choose_expr(is_pointer(arg), (unsigned long)__ret, 0)) { \
SET_ERRNO(-__ret); \
__reg = typeof(ret)-1L; \
} \
__ret; \
})
Apart from the annoyance of having to reverse the conditional
that only has one copy of the check.
Using two __builtin_choose_expr() saves you having to write two
comparisons that are valid for both pointer and integer.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
next prev parent reply other threads:[~2023-08-07 8:40 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-07 5:57 Zhangjin Wu
2023-08-07 8:40 ` David Laight [this message]
2023-08-07 9:53 ` David Laight
2023-08-07 14:12 ` Zhangjin Wu
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=f51e54bcf470451ea36f24640f000e61@AcuMS.aculab.com \
--to=david.laight@aculab.com \
--cc=arnd@arndb.de \
--cc=falcon@tinylab.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=thomas@t-8ch.de \
--cc=w@1wt.eu \
/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
all inboxes | Powered by JetHome®