From: Jonathan Wakely <jwakely@redhat.com>
To: Richard Biener <richard.guenther@gmail.com>
Cc: Hans-Peter Nilsson <hp@bitrange.com>,
Thiago Macieira <thiago.macieira@intel.com>,
Andreas Schwab <schwab@linux-m68k.org>,
libstdc++ <libstdc++@gcc.gnu.org>,
GCC Patches <gcc-patches@gcc.gnu.org>
Subject: Re: [PATCH 1/5] std::latch: reduce internal implementation from ptrdiff_t to int
Date: Wed, 3 Mar 2021 15:10:19 +0000 [thread overview]
Message-ID: <YD+m24iqKEpZqM3B@redhat.com> (raw)
In-Reply-To: <YD+jk81Z8dU7T1jA@redhat.com>
On 03/03/21 14:56 +0000, Jonathan Wakely wrote:
>On 01/03/21 09:56 +0100, Richard Biener via Libstdc++ wrote:
>>On Sun, Feb 28, 2021 at 10:53 PM Hans-Peter Nilsson <hp@bitrange.com> wrote:
>>>
>>>
>>>
>>>On Fri, 26 Feb 2021, Thiago Macieira via Gcc-patches wrote:
>>>
>>>> On Friday, 26 February 2021 11:31:00 PST Andreas Schwab wrote:
>>>> > On Feb 26 2021, Thiago Macieira wrote:
>>>> > > On Friday, 26 February 2021 10:14:42 PST Andreas Schwab wrote:
>>>> > >> On Feb 26 2021, Thiago Macieira via Gcc-patches wrote:
>>>> > >> > - alignas(__alignof__(ptrdiff_t)) ptrdiff_t _M_a;
>>>> > >> > + alignas(__alignof__(int)) int _M_a;
>>>> > >>
>>>> > >> Futexes must be aligned to 4 bytes.
>>>> > >
>>>> > > Agreed, but doesn't this accomplish that?
>>>> >
>>>> > No. It uses whatever alignment the type already has, and is an
>>>> > elaborate no-op.
>>>>
>>>> I thought so too when I read the original line. But I expected it was written
>>>> like that for a reason, especially since the same pattern appears in other
>>>> places.
>>>>
>>>> I can change to "alignas(4)" (which is a GCC extension, I believe). Is that
>>>> the correct solution?
It's not a GCC extensions. You're thinking of alignas(obj) which is a
GCC extension.
>>>IMNSHO make use of the corresponding atomic type. Then there'd
>>>be no need for separate what's-the-right-align-curse games.
>
>That won't work though, because we need direct access to the integer
>object, not to a std::atomic<int> which contains it.
>
>>Or align as the corresponding atomic type (in case using an actual
>>std::atomic<int> is undesirable). OTOH the proposed code isn't
>>any more bogus than the previous ...
>
>The previous code accounts for the fact that ptrdiff_t is a typedef
>for an unspecified type, and that some ABIs allow struct members to have
>weaker alignment than they would have otherwise.
>
>e.g. __alignof__(long long) != alignof(long long) on x86.
>
>Yes, I know ptrdiff_t isn't long long on x86, but since it's a typedef
>for some target-specific type, it's still possible that
>alignof != __alignof__. So alignas(__alignof__(T)) is not necessarily
>a no-op. So not bogus.
>
>For int, there shouldn't be any need to force the alignment. I don't
>think any ABI supported by GCC allows int members to be aligned to
>less than __alignof__(int). Users could break it by compiling with
>#pragma pack, but I have no sympathy for such silliness.
Jakub said on IRC that m68k might have alignof(int) == 2, so we'd need
to increase that alignment to use it as a futex.
N.B. std::atomic and std::atomic_ref don't use alignas(__alignof__(T))
they do this instead:
static_assert(is_trivially_copyable_v<_Tp>);
// 1/2/4/8/16-byte types must be aligned to at least their size.
static constexpr int _S_min_alignment
= (sizeof(_Tp) & (sizeof(_Tp) - 1)) || sizeof(_Tp) > 16
? 0 : sizeof(_Tp);
public:
static constexpr size_t required_alignment
= _S_min_alignment > alignof(_Tp) ? _S_min_alignment : alignof(_Tp);
Using std::atomic_ref<T>::required_alignment would give that value.
For something being used as a futex we should just use alignas(4)
though, since that's what the kernel requires.
next prev parent reply other threads:[~2021-03-03 15:10 UTC|newest]
Thread overview: 73+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-25 22:50 C++2a synchronisation inefficient in GCC 11 Thiago Macieira
2021-02-26 11:19 ` Jonathan Wakely
2021-02-26 17:37 ` Thiago Macieira
2021-02-26 18:29 ` Jonathan Wakely
2021-02-26 19:30 ` Ville Voutilainen
2021-02-26 21:17 ` Jonathan Wakely
2021-02-26 21:18 ` Ville Voutilainen
2021-02-26 21:39 ` Jonathan Wakely
2021-02-26 18:47 ` Ville Voutilainen
2021-02-26 23:53 ` Thiago Macieira
2021-02-26 23:58 ` Ville Voutilainen
2021-02-27 0:11 ` Thiago Macieira
2021-02-27 0:18 ` Ville Voutilainen
2021-02-27 0:36 ` Thiago Macieira
2021-02-27 0:44 ` Ville Voutilainen
2021-02-27 0:53 ` Thiago Macieira
2021-02-27 1:03 ` Ville Voutilainen
2021-03-03 14:30 ` Jonathan Wakely
2021-03-03 17:07 ` Thiago Macieira
2021-03-03 17:14 ` Jonathan Wakely
2021-02-27 0:22 ` Marc Glisse
2021-02-27 0:30 ` Ville Voutilainen
2021-02-27 0:43 ` Thiago Macieira
2021-03-03 14:24 ` Jonathan Wakely
2021-03-03 17:12 ` Thiago Macieira
2021-02-26 15:59 ` [PATCH 1/5] std::latch: reduce internal implementation from ptrdiff_t to int Thiago Macieira
2021-02-26 15:59 ` [PATCH 2/5] Atomic __platform_wait: accept any 32-bit type, not just int Thiago Macieira
2021-03-03 14:34 ` Jonathan Wakely
2021-03-03 16:21 ` Jonathan Wakely
2021-03-03 17:27 ` Thiago Macieira
2021-03-03 17:34 ` Ville Voutilainen
2021-03-03 17:41 ` Jonathan Wakely
2021-02-26 15:59 ` [PATCH 3/5] std::__atomic_wait: don't use __detail::__waiter with futex Thiago Macieira
2021-02-26 15:59 ` [PATCH 4/5] barrier: use int instead of unsigned char for the phase state Thiago Macieira
2021-02-28 15:05 ` Hans-Peter Nilsson
2021-03-01 16:28 ` Thomas Rodgers
2021-03-01 17:24 ` Thiago Macieira
2021-03-01 17:38 ` Thomas Rodgers
2021-03-01 17:40 ` Thomas Rodgers
2021-03-01 18:06 ` Thiago Macieira
2021-03-01 19:08 ` Thomas Rodgers
2021-03-01 18:12 ` Ville Voutilainen
2021-03-01 19:44 ` Thiago Macieira
2021-03-01 20:35 ` Ville Voutilainen
2021-03-01 21:54 ` Thiago Macieira
2021-03-01 22:04 ` Ville Voutilainen
2021-03-01 22:21 ` Thiago Macieira
2021-03-01 22:31 ` Ville Voutilainen
2021-03-01 22:40 ` Thiago Macieira
2021-02-26 15:59 ` [PATCH 5/5] barrier: optimise by not having the hasher in a loop Thiago Macieira
2021-03-03 14:36 ` Jonathan Wakely
2021-02-26 18:14 ` [PATCH 1/5] std::latch: reduce internal implementation from ptrdiff_t to int Andreas Schwab
2021-02-26 19:08 ` Thiago Macieira
2021-02-26 19:31 ` Andreas Schwab
2021-02-27 0:13 ` Thiago Macieira
2021-02-28 21:31 ` Hans-Peter Nilsson
2021-03-01 8:56 ` Richard Biener
2021-03-03 14:56 ` Jonathan Wakely
2021-03-03 15:02 ` Andreas Schwab
2021-03-03 15:10 ` Jonathan Wakely [this message]
2021-03-03 15:37 ` Hans-Peter Nilsson
2021-03-01 16:32 ` Thomas Rodgers
2021-03-03 14:34 ` Jonathan Wakely
2021-03-03 17:14 ` Thiago Macieira
2021-03-03 17:18 ` Jonathan Wakely
2021-02-27 1:13 ` C++2a synchronisation inefficient in GCC 11 Thomas Rodgers
2021-02-27 1:29 ` Thomas Rodgers
2021-02-27 3:01 ` Thomas Rodgers
2021-03-01 17:46 ` Thomas Rodgers
2021-03-01 18:00 ` Thiago Macieira
2021-03-01 18:34 ` Thomas Rodgers
2021-03-01 19:11 ` Thiago Macieira
2021-02-27 2:02 ` Ville Voutilainen
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=YD+m24iqKEpZqM3B@redhat.com \
--to=jwakely@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=hp@bitrange.com \
--cc=libstdc++@gcc.gnu.org \
--cc=richard.guenther@gmail.com \
--cc=schwab@linux-m68k.org \
--cc=thiago.macieira@intel.com \
/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).