unofficial mirror of libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Carlos O'Donell via Libc-alpha <libc-alpha@sourceware.org>
To: "H.J. Lu" <hjl.tools@gmail.com>, Florian Weimer <fweimer@redhat.com>
Cc: Andreas Schwab <schwab@suse.de>,
	"H.J. Lu via Libc-alpha" <libc-alpha@sourceware.org>,
	Alistair Francis <alistair.francis@wdc.com>,
	Joseph Myers <joseph@codesourcery.com>
Subject: Re: [PATCH] Update HP_TIMING_NOW for _ISOMAC in sysdeps/generic/hp-timing.h
Date: Fri, 5 Jun 2020 12:26:26 -0400	[thread overview]
Message-ID: <96f6414d-ea0a-eab7-8603-c66af51e02bb@redhat.com> (raw)
In-Reply-To: <CAMe9rOpYRvBnQe6EoEsNgSUXcCOqqYeCC0NbHiAm=-Ur4DgPRg@mail.gmail.com>

On 5/28/20 10:50 AM, H.J. Lu via Libc-alpha wrote:
> On Thu, May 28, 2020 at 7:29 AM Florian Weimer <fweimer@redhat.com> wrote:
>>
>> * H. J. Lu:
>>
>>> On Thu, May 28, 2020 at 5:58 AM Florian Weimer <fweimer@redhat.com> wrote:
>>>>
>>>> * H. J. Lu via Libc-alpha:
>>>>
>>>>> diff --git a/sysdeps/generic/hp-timing.h b/sysdeps/generic/hp-timing.h
>>>>> index af9d92f7f7..99b12995f1 100644
>>>>> --- a/sysdeps/generic/hp-timing.h
>>>>> +++ b/sysdeps/generic/hp-timing.h
>>>>> @@ -24,6 +24,11 @@
>>>>>  #include <stdint.h>
>>>>>  #include <hp-timing-common.h>
>>>>>
>>>>> +#ifdef _ISOMAC
>>>>> +# define __timespec64 timespec
>>>>> +# define __clock_gettime64 clock_gettime
>>>>> +#endif
>>>>> +
>>>>>  /* It should not be used for ld.so.  */
>>>>>  #define HP_TIMING_INLINE     (0)
>>>>
>>>> To be honest, I would have expected different definitions of the timing
>>>> macros.  This redirection looks a bit iffy for me for _ISOMAC.
>>>
>>> What do you have in mind?
>>
>> Something like this:
>>
>> /* The clock_gettime (CLOCK_MONOTONIC) has unspecified starting time,
>>    nano-second accuracy, and for some architectues is implemented as
>>    vDSO symbol.  */
>> #ifdef _ISOMAC
>> # define HP_TIMING_NOW(var) \
>> ({                                                              \
>>   struct timespec tv;                                           \
>>   clock_gettime (CLOCK_MONOTONIC, &tv);                         \
>>   (var) = (tv.tv_nsec + UINT64_C(1000000000) * tv.tv_sec);      \
>> })
>> #else
>> # define HP_TIMING_NOW(var) \
>> ({                                                              \
>>   struct __timespec64 tv;                                       \
>>   __clock_gettime64 (CLOCK_MONOTONIC, &tv);                     \
>>   (var) = (tv.tv_nsec + UINT64_C(1000000000) * tv.tv_sec);      \
>> })
>> #endif
>>
>> in sysdeps/generic/hp-timing.h.
>>
> 
> Here is the updated patch.  OK for master?

OK for master.

Reviewed-by: Carlos O'Donell <carlos@redhat.com>

> From f312c52fa80828ecdc473594e160ef2f5d2a00d7 Mon Sep 17 00:00:00 2001
> From: "H.J. Lu" <hjl.tools@gmail.com>
> Date: Sat, 23 May 2020 10:04:20 -0700
> Subject: [PATCH] Update HP_TIMING_NOW for _ISOMAC in
>  sysdeps/generic/hp-timing.h
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
> 
> commit e9698175b0b60407db1e89bcf29437ab224bca0b
> Author: Lukasz Majewski <lukma@denx.de>
> Date:   Mon Mar 16 08:31:41 2020 +0100
> 
>     y2038: Replace __clock_gettime with __clock_gettime64
> 
> breaks benchtests with sysdeps/generic/hp-timing.h:

Yes, Patsy Griffin in testing for Fedora Rawhide has seen this for
aarch64, s390x, and armv7hl. Patsy is testing this patch now.

> In file included from ./bench-timing.h:23,
>                  from ./bench-skeleton.c:25,
>                  from
> /export/build/gnu/tools-build/glibc-gitlab/build-x86_64-linux/benchtests/bench-rint.c:45:
> ./bench-skeleton.c: In function ‘main’:
> ../sysdeps/generic/hp-timing.h:37:23: error: storage size of ‘tv’ isn’t known
>    37 |   struct __timespec64 tv;      \
>       |                       ^~
> 
> Define HP_TIMING_NOW with clock_gettime in sysdeps/generic/hp-timing.h
> if _ISOMAC is defined.  Don't define __clock_gettime in bench-timing.h
> since it is no longer needed.

OK.

> ---
>  benchtests/bench-timing.h   |  1 -
>  sysdeps/generic/hp-timing.h | 11 ++++++++++-
>  2 files changed, 10 insertions(+), 2 deletions(-)
> 
> diff --git a/benchtests/bench-timing.h b/benchtests/bench-timing.h
> index a0d6f82465..d0176fb76e 100644
> --- a/benchtests/bench-timing.h
> +++ b/benchtests/bench-timing.h
> @@ -18,7 +18,6 @@
>  
>  #undef attribute_hidden
>  #define attribute_hidden
> -#define __clock_gettime __clock_gettime64

OK.

>  #include <hp-timing.h>
>  #include <stdint.h>
>  
> diff --git a/sysdeps/generic/hp-timing.h b/sysdeps/generic/hp-timing.h
> index af9d92f7f7..1960a71c62 100644
> --- a/sysdeps/generic/hp-timing.h
> +++ b/sysdeps/generic/hp-timing.h
> @@ -32,11 +32,20 @@ typedef uint64_t hp_timing_t;
>  /* The clock_gettime (CLOCK_MONOTONIC) has unspecified starting time,
>     nano-second accuracy, and for some architectues is implemented as
>     vDSO symbol.  */
> -#define HP_TIMING_NOW(var) \
> +#ifdef _ISOMAC
> +# define HP_TIMING_NOW(var) \
> +({								\
> +  struct timespec tv;						\
> +  clock_gettime (CLOCK_MONOTONIC, &tv);				\
> +  (var) = (tv.tv_nsec + UINT64_C(1000000000) * tv.tv_sec);	\
> +})
> +#else
> +# define HP_TIMING_NOW(var) \

OK.

>  ({								\
>    struct __timespec64 tv;						\
>    __clock_gettime64 (CLOCK_MONOTONIC, &tv);			\
>    (var) = (tv.tv_nsec + UINT64_C(1000000000) * tv.tv_sec);	\
>  })
> +#endif
>  
>  #endif	/* hp-timing.h */
> -- 
> 2.26.2
> 

-- 
Cheers,
Carlos.


      reply	other threads:[~2020-06-05 16:26 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-05 18:16 [PATCH v3] y2038: Replace __clock_gettime with __clock_gettime64 Lukasz Majewski
2020-05-19 19:51 ` Lukasz Majewski
2020-05-20 10:29 ` Andreas Schwab
2020-05-21  1:09 ` H.J. Lu via Libc-alpha
2020-05-21 10:24   ` Lukasz Majewski
2020-05-21 11:11     ` H.J. Lu via Libc-alpha
2020-05-23 20:32       ` [PATCH] Fix __clock_gettime64 for sysdeps/generic/hp-timing.h H.J. Lu via Libc-alpha
2020-05-24  9:03         ` Lukasz Majewski
2020-05-24 11:39           ` H.J. Lu via Libc-alpha
2020-05-24 15:30             ` Lukasz Majewski
2020-05-24 16:13               ` H.J. Lu via Libc-alpha
2020-05-24 23:34                 ` V2 " H.J. Lu via Libc-alpha
2020-05-25 10:40                   ` Lukasz Majewski
2020-05-25 10:57   ` [PATCH v3] y2038: Replace __clock_gettime with __clock_gettime64 Andreas Schwab
2020-05-27 12:44     ` Florian Weimer via Libc-alpha
2020-05-27 12:47       ` H.J. Lu via Libc-alpha
2020-05-27 12:50         ` Florian Weimer via Libc-alpha
2020-05-28 12:28           ` [PATCH] Fix __clock_gettime64 with _ISOMAC in sysdeps/generic/hp-timing.h H.J. Lu via Libc-alpha
2020-05-28 12:58             ` Florian Weimer via Libc-alpha
2020-05-28 13:07               ` H.J. Lu via Libc-alpha
2020-05-28 14:28                 ` Florian Weimer via Libc-alpha
2020-05-28 14:50                   ` [PATCH] Update HP_TIMING_NOW for " H.J. Lu via Libc-alpha
2020-06-05 16:26                     ` Carlos O'Donell via Libc-alpha [this message]

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

  List information: https://www.gnu.org/software/libc/involved.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=96f6414d-ea0a-eab7-8603-c66af51e02bb@redhat.com \
    --to=libc-alpha@sourceware.org \
    --cc=alistair.francis@wdc.com \
    --cc=carlos@redhat.com \
    --cc=fweimer@redhat.com \
    --cc=hjl.tools@gmail.com \
    --cc=joseph@codesourcery.com \
    --cc=schwab@suse.de \
    /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).