From: Richard Biener <richard.guenther@gmail.com>
To: Aldy Hernandez <aldyh@redhat.com>,Martin Sebor
<msebor@gmail.com>,gcc-patches <gcc-patches@gcc.gnu.org>
Subject: Re: protected alloca class for malloc fallback
Date: Fri, 05 Aug 2016 17:56:00 -0000 [thread overview]
Message-ID: <A04768E7-68E6-4331-B5C4-1C28003BF383@gmail.com> (raw)
In-Reply-To: <57A4A5E8.90205@redhat.com>
On August 5, 2016 4:42:48 PM GMT+02:00, Aldy Hernandez <aldyh@redhat.com> wrote:
>
>> I was surprised by the always_inline trick. I suppose it makes
>> sense but I wouldn't have expected to be able to rely on it. Out
>> of curiosity I tested it with other compilers. It turns out that
>> it only works with some like GCC and IBM XLC++, but not others
>> like Clang or Sun CC. In recursive calls, they don't seem to
>> hold on to the memory allocated via alloca by the class ctor in
>> the caller.
>
>Well, it was surprising to me as well, hence the comment. I expected
>it
>to just work, and when it didn't I had to hunt in the inliner code to
>find out why it was selectively inlining:
>
> case GIMPLE_CALL:
> /* Refuse to inline alloca call unless user explicitly forced so as
> this may change program's memory overhead drastically when the
> function using alloca is called in loop. In GCC present in
> SPEC2000 inlining into schedule_block cause it to require 2GB of
> RAM instead of 256MB. Don't do so for alloca calls emitted for
> VLA objects as those can't cause unbounded growth (they're always
> wrapped inside stack_save/stack_restore regions. */
>
>As Richi pointed out, if the constructor doesn't get inlined (as you're
>
>seeing in Clang and Sun CC), we could potentially clobber freed memory.
>
> So much for that idea...
>
>>
>> FWIW, rather than creating a new class around alloca and putting
>> effort into changing code to use it I think that investing time
>> into replacing the function's uses with an existing C++ container
>> class (either GCC's own or one from the STL) might be a better
>> way to go.
>
>Yes, I suggested a combination of auto_vec<> (gcc's) and std::string
>down thread.
Please don't use std::string. For string building you can use obstacks.
Richard.
>Thanks for checking out the result from other compilers. Much
>appreciated.
>
>Aldy
next prev parent reply other threads:[~2016-08-05 17:56 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-04 11:30 Aldy Hernandez
2016-08-04 12:58 ` Richard Biener
2016-08-04 15:19 ` Aldy Hernandez
2016-08-04 19:24 ` Jeff Law
2016-08-05 14:37 ` Aldy Hernandez
2016-08-05 15:15 ` Pedro Alves
2016-08-05 16:23 ` Jeff Law
2016-08-05 17:48 ` Richard Biener
2016-08-05 8:17 ` Richard Biener
2016-08-04 19:06 ` Pedro Alves
2016-08-04 19:16 ` Jeff Law
2016-08-04 19:22 ` Pedro Alves
2016-08-04 19:26 ` Jeff Law
2016-08-04 19:31 ` Pedro Alves
2016-08-05 2:10 ` Martin Sebor
2016-08-05 14:42 ` Aldy Hernandez
2016-08-05 17:56 ` Richard Biener [this message]
2016-08-05 18:16 ` Oleg Endo
2016-08-05 20:07 ` Richard Biener
2016-08-06 10:09 ` Aldy Hernandez
2016-08-06 10:15 ` Aldy Hernandez
2016-08-06 15:08 ` Richard Biener
2016-08-08 17:00 ` Jeff Law
2016-08-08 17:32 ` Trevor Saunders
2016-08-08 19:03 ` Richard Biener
2016-08-09 11:34 ` Oleg Endo
2016-08-09 17:34 ` Trevor Saunders
2016-08-10 17:03 ` Oleg Endo
2016-08-11 1:23 ` Trevor Saunders
2016-08-11 12:18 ` Oleg Endo
2016-08-11 17:55 ` Trevor Saunders
2016-08-20 2:29 ` Mike Stump
2016-08-21 20:00 ` C++11? (Re: protected alloca class for malloc fallback) Pedro Alves
2016-08-22 7:10 ` Trevor Saunders
2016-08-22 7:28 ` Richard Biener
2016-08-22 12:02 ` Eric Gallager
2016-08-22 12:58 ` Manuel López-Ibáñez
2016-08-22 22:08 ` Mike Stump
2016-08-23 23:17 ` Eric Gallager
2016-08-09 13:17 ` protected alloca class for malloc fallback Aldy Hernandez
2016-08-09 13:21 ` Bernd Schmidt
2016-08-10 10:04 ` Richard Biener
2016-08-10 10:12 ` Aldy Hernandez
2016-08-10 10:39 ` Richard Biener
2016-08-10 18:00 ` Jeff Law
2016-08-10 18:33 ` Richard Biener
2016-08-16 16:28 ` Jeff Law
2016-08-16 16:44 ` Jakub Jelinek
2016-08-16 16:47 ` Jeff Law
2016-08-16 17:54 ` Martin Sebor
2016-08-17 8:27 ` Richard Biener
2016-08-17 13:39 ` Martin Sebor
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=A04768E7-68E6-4331-B5C4-1C28003BF383@gmail.com \
--to=richard.guenther@gmail.com \
--cc=aldyh@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=msebor@gmail.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).