From: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
To: Christopher Li <sparse@chrisli.org>
Cc: Sparse Mailing-list <linux-sparse@vger.kernel.org>
Subject: Re: [PATCH 01/10] Add the __restrict__ keyword
Date: Wed, 06 Aug 2014 16:21:45 +0100 [thread overview]
Message-ID: <53E24809.5020603@ramsay1.demon.co.uk> (raw)
In-Reply-To: <CANeU7Qkr5oX-_iLRXTt6DKD4pCwV3LoWPy9CuhQ0kk9NLz7s_Q@mail.gmail.com>
On 06/08/14 10:19, Christopher Li wrote:
> On Mon, Aug 4, 2014 at 11:32 AM, Ramsay Jones
> <ramsay@ramsay1.demon.co.uk> wrote:
>>
>> - if (match_idents(token, &restrict_ident, &__restrict_ident, NULL))
>> + if (match_idents(token, &restrict_ident, &__restrict_ident, &__restrict___ident, NULL))
>> token = abstract_array_static_declarator(token->next, &has_static);
>> token = parse_expression(token, &expr);
>> sym->array_size = expr;
>> diff --git a/validation/reserved.c b/validation/reserved.c
>> index caacd21..e5d7af8 100644
>> --- a/validation/reserved.c
>> +++ b/validation/reserved.c
>> @@ -30,6 +30,7 @@ reserved.c:8:12: error: Trying to use reserved word '__const' as identifier
>> reserved.c:9:12: error: Trying to use reserved word '__const__' as identifier
>> reserved.c:10:12: error: Trying to use reserved word 'restrict' as identifier
>> reserved.c:11:12: error: Trying to use reserved word '__restrict' as identifier
>> +reserved.c:12:12: error: Trying to use reserved word '__restrict__' as identifier
>> reserved.c:13:12: error: Trying to use reserved word 'typedef' as identifier
>> reserved.c:14:12: error: Trying to use reserved word '__typeof' as identifier
>> reserved.c:15:12: error: Trying to use reserved word '__typeof__' as identifier
>
> Can you give more test case for __restrict__? Your patch seems suggest that you
> hit some code use "__restruct__" for abstract array declaration. Did
> that actually
> happen? I want to get some example.
No, __restrict__ is only applied to various pointer types in the MinGW
header files (see below). So, I agree that validation/reserved.c is not
exactly representative of the actual use of __restrict__ (or *any* of the
keywords come to that), but that does at least provide a minimal test.
Also, I can't believe that Al did not intend to include __restrict__ in the
original commit d5c9c2431; well actually he *did* include it in the test, he
just didn't notice some of the missing implementation! ;-D
Having said that, I think I see two tests included for the array declaration
usage: abstract-array-declarator-static.c and restrict-array.c.
As for the MinGW headers, I can only talk about the current/old 32-bit
version of MinGW (I read somewhere that the new MinGW-64 project will
actually re-vamp the 32-bit version as well), but my installation was
updated only last September, so they are still being used. One of the
reasons for the MinGW patches being on the back-burner is because I
haven't installed the 64-bit version yet (last time I looked, it wasn't
considered ready for use).
I have included, below, the output of a grep for __restrict__ in the
header files, which returned 50 matching lines (note that the one and
only use of __restrict in the declaration of strtoll which must have
been a typo!).
HTH,
ATB,
Ramsay Jones
-- 8< --
/mingw/include/inttypes.h:264:intmax_t __cdecl __MINGW_NOTHROW strtoimax (const char* __restrict__ nptr,
/mingw/include/inttypes.h:265: char** __restrict__ endptr, int base);
/mingw/include/inttypes.h:266:uintmax_t __cdecl __MINGW_NOTHROW strtoumax (const char* __restrict__ nptr,
/mingw/include/inttypes.h:267: char** __restrict__ endptr, int base);
/mingw/include/inttypes.h:269:intmax_t __cdecl __MINGW_NOTHROW wcstoimax (const wchar_t* __restrict__ nptr,
/mingw/include/inttypes.h:270: wchar_t** __restrict__ endptr, int base);
/mingw/include/inttypes.h:271:uintmax_t __cdecl __MINGW_NOTHROW wcstoumax (const wchar_t* __restrict__ nptr,
/mingw/include/inttypes.h:272: wchar_t** __restrict__ endptr, int base);
/mingw/include/search.h:82:void * __cdecl tdelete (const void * __restrict__, void ** __restrict__,
/mingw/include/stdio.h:333:int __cdecl __MINGW_NOTHROW vscanf (const char * __restrict__, __VALIST);
/mingw/include/stdio.h:334:int __cdecl __MINGW_NOTHROW vfscanf (FILE * __restrict__, const char * __restrict__,
/mingw/include/stdio.h:336:int __cdecl __MINGW_NOTHROW vsscanf (const char * __restrict__,
/mingw/include/stdio.h:337: const char * __restrict__, __VALIST);
/mingw/include/stdio.h:607:int __cdecl __MINGW_NOTHROW vwscanf (const wchar_t * __restrict__, __VALIST);
/mingw/include/stdio.h:608:int __cdecl __MINGW_NOTHROW vfwscanf (FILE * __restrict__,
/mingw/include/stdio.h:609: const wchar_t * __restrict__, __VALIST);
/mingw/include/stdio.h:610:int __cdecl __MINGW_NOTHROW vswscanf (const wchar_t * __restrict__,
/mingw/include/stdio.h:611: const wchar_t * __restrict__, __VALIST);
/mingw/include/stdlib.h:318:strtod (const char* __restrict__ __nptr, char** __restrict__ __endptr)
/mingw/include/stdlib.h:322:float __cdecl __MINGW_NOTHROW strtof (const char * __restrict__, char ** __restrict__);
/mingw/include/stdlib.h:323:long double __cdecl __MINGW_NOTHROW strtold (const char * __restrict__, char ** __restrict__);
/mingw/include/stdlib.h:337:float __cdecl __MINGW_NOTHROW wcstof( const wchar_t * __restrict__, wchar_t ** __restrict__);
/mingw/include/stdlib.h:338:long double __cdecl __MINGW_NOTHROW wcstold (const wchar_t * __restrict__, wchar_t ** __restrict__);
/mingw/include/stdlib.h:518:long long __cdecl __MINGW_NOTHROW strtoll (const char* __restrict__, char** __restrict, int);
/mingw/include/stdlib.h:519:unsigned long long __cdecl __MINGW_NOTHROW strtoull (const char* __restrict__, char** __restrict__, int);
/mingw/include/sys/time.h:39:int __cdecl __MINGW_NOTHROW gettimeofday(struct timeval *__restrict__,
/mingw/include/sys/time.h:40: void *__restrict__ /* tzp (unused) */);
/mingw/include/wchar.h:150:int __cdecl __MINGW_NOTHROW vwscanf (const wchar_t * __restrict__, __VALIST);
/mingw/include/wchar.h:151:int __cdecl __MINGW_NOTHROW vfwscanf (FILE * __restrict__,
/mingw/include/wchar.h:152: const wchar_t * __restrict__, __VALIST);
/mingw/include/wchar.h:153:int __cdecl __MINGW_NOTHROW vswscanf (const wchar_t * __restrict__,
/mingw/include/wchar.h:154: const wchar_t * __restrict__, __VALIST);
/mingw/include/wchar.h:165:float __cdecl __MINGW_NOTHROW wcstof (const wchar_t * __restrict__, wchar_t ** __restrict__);
/mingw/include/wchar.h:166:long double __cdecl __MINGW_NOTHROW wcstold (const wchar_t * __restrict__, wchar_t ** __restrict__);
/mingw/include/wchar.h:281:size_t __cdecl __MINGW_NOTHROW mbrlen(const char * __restrict__, size_t,
/mingw/include/wchar.h:282: mbstate_t * __restrict__);
/mingw/include/wchar.h:283:size_t __cdecl __MINGW_NOTHROW mbrtowc(wchar_t * __restrict__, const char * __restrict__,
/mingw/include/wchar.h:284: size_t, mbstate_t * __restrict__);
/mingw/include/wchar.h:285:size_t __cdecl __MINGW_NOTHROW mbsrtowcs(wchar_t * __restrict__, const char ** __restrict__,
/mingw/include/wchar.h:286: size_t, mbstate_t * __restrict__);
/mingw/include/wchar.h:287:size_t __cdecl __MINGW_NOTHROW wcrtomb(char * __restrict__, wchar_t,
/mingw/include/wchar.h:288: mbstate_t * __restrict__);
/mingw/include/wchar.h:289:size_t __cdecl __MINGW_NOTHROW wcsrtombs(char * __restrict__, const wchar_t ** __restrict__,
/mingw/include/wchar.h:290: size_t, mbstate_t * __restrict__);
/mingw/include/wchar.h:306:wchar_t* __cdecl __MINGW_NOTHROW wmemcpy(wchar_t* __restrict__,
/mingw/include/wchar.h:307: const wchar_t* __restrict__,
/mingw/include/wchar.h:310:long long __cdecl __MINGW_NOTHROW wcstoll(const wchar_t * __restrict__,
/mingw/include/wchar.h:311: wchar_t** __restrict__, int);
/mingw/include/wchar.h:312:unsigned long long __cdecl __MINGW_NOTHROW wcstoull(const wchar_t * __restrict__,
/mingw/include/wchar.h:313: wchar_t ** __restrict__, int);
next prev parent reply other threads:[~2014-08-06 15:21 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-08-04 18:32 [PATCH 01/10] Add the __restrict__ keyword Ramsay Jones
2014-08-06 9:19 ` Christopher Li
2014-08-06 15:21 ` Ramsay Jones [this message]
[not found] ` <CANeU7Qm_pUe55OY0+Nh9OS=ZN5NqFytkB3CaJMTZZOvPqtK7nw@mail.gmail.com>
2014-08-07 3:04 ` Fwd: " Christopher Li
2014-08-07 20:02 ` Ramsay Jones
2014-08-08 11:13 ` Ramsay Jones
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=53E24809.5020603@ramsay1.demon.co.uk \
--to=ramsay@ramsay1.demon.co.uk \
--cc=linux-sparse@vger.kernel.org \
--cc=sparse@chrisli.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