From: Jonathan Wakely <jwakely@redhat.com>
To: Richard Henderson <rth@redhat.com>
Cc: libstdc++@gcc.gnu.org, gcc-patches@gcc.gnu.org,
Andrew MacLeod <amacleod@redhat.com>
Subject: Re: [libstdc++/65033] Give alignment info to libatomic
Date: Tue, 31 Mar 2015 15:41:00 -0000 [thread overview]
Message-ID: <20150331154055.GU9755@redhat.com> (raw)
In-Reply-To: <551AB9A8.9000905@redhat.com>
On 31/03/15 08:13 -0700, Richard Henderson wrote:
>On 03/31/2015 08:03 AM, Jonathan Wakely wrote:
>> On 31/03/15 07:54 -0700, Richard Henderson wrote:
>>> On 03/31/2015 06:41 AM, Jonathan Wakely wrote:
>>>> This is the best I've come up with, does anyone have any better ideas
>>>> than the #else branch to hardcode alignment of 16-byte types to 16?
>>>
>>> If there's no 16 byte type, are we convinced this matters? I mean, there isn't
>>> a 16-byte atomic instruction for 32-bit x86 (or any other 32-bit cpu of which I
>>> am aware). So we're forced to use a locking path anyway.
>>
>> The C front end gives struct S { char s[16]; } 16 byte alignment...
>
>Um, I'm pretty sure it doesn't.
>
> struct S { char s[16]; };
> int x = __alignof(struct S);
>
> .type x, @object
> .size x, 4
>x:
> .long 1
>
>What you're interpreting as alignment for that struct is probably optional
>*additional* alignment from LOCAL_ALIGNMENT or LOCAL_DECL_ALIGNMENT or something.
Sorry for not being clear, I meant __alignof(_Atomic struct S) is 16.
>> And it matters most for the integral types, not arbitrary structs.
>>
>>> And if we do want the alignment, do we stop pretending with all the sizeof's
>>> and alignof's and just use power-of-two size alignment for all of them, e.g.
>>>
>>> min_align = ((size & (size - 1)) || size > 16 ? 0 : size)
>>
>> Yeah, I wondered about that too. Joseph indicated there are targets
>> where C gives alignof(_Atomic X) != sizeof(X), which is why the target
>> hook exists, but maybe we can just not worry about those targets for
>> now.
>
>Those targets have alignof < sizeof. So in a way we'd probably be doing them a
>favor. I know for instance that CRIS is in this category, where most data is
>all byte aligned, but atomics must be naturally aligned.
Aha, I wondered why CRIS overrides the atomic_align_for_mode hook when
it seemed to be giving them natural alignment anyway - I didn't
realise non-atomic types are only byte-aligned.
next prev parent reply other threads:[~2015-03-31 15:41 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-02-12 21:23 Richard Henderson
2015-02-18 12:15 ` Jonathan Wakely
2015-03-25 16:22 ` Jonathan Wakely
2015-03-25 18:36 ` Richard Henderson
2015-03-25 18:49 ` Jonathan Wakely
2015-03-25 19:04 ` Richard Henderson
2015-03-26 13:21 ` Jonathan Wakely
2015-03-31 13:41 ` Jonathan Wakely
2015-03-31 14:54 ` Richard Henderson
2015-03-31 15:03 ` Jonathan Wakely
2015-03-31 15:13 ` Richard Henderson
2015-03-31 15:41 ` Jonathan Wakely [this message]
2015-04-06 22:59 ` Hans-Peter Nilsson
2015-04-13 4:45 ` patch fix issue 1 with "[libstdc++/65033] Give alignment info to libatomic" Hans-Peter Nilsson
2015-04-13 11:59 ` Jonathan Wakely
2015-04-13 5:59 ` Issue 2 " Hans-Peter Nilsson
2015-04-13 17:53 ` Joseph Myers
2015-03-25 18:39 ` [libstdc++/65033] Give alignment info to libatomic Richard Henderson
2015-04-03 3:04 ` Hans-Peter Nilsson
2015-03-26 11:54 ` Jonathan Wakely
2015-04-03 3:57 ` Hans-Peter Nilsson
2015-04-03 9:25 ` Hans-Peter Nilsson
2015-04-03 14:13 ` Jonathan Wakely
2015-04-03 19:13 ` Richard Henderson
2015-04-07 13:14 ` Jonathan Wakely
2015-04-09 11:17 ` Jonathan Wakely
2015-04-06 1:07 ` Hans-Peter Nilsson
2015-04-07 9:45 ` Jonathan Wakely
2015-04-07 10:52 ` Hans-Peter Nilsson
2015-04-07 13:12 ` Jonathan Wakely
2015-04-07 14:51 ` Hans-Peter Nilsson
2015-04-07 15:06 ` Jonathan Wakely
2015-04-08 3:58 ` Hans-Peter Nilsson
2015-04-08 9:35 ` Jonathan Wakely
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=20150331154055.GU9755@redhat.com \
--to=jwakely@redhat.com \
--cc=amacleod@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=libstdc++@gcc.gnu.org \
--cc=rth@redhat.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).