From: Richard Biener <richard.guenther@gmail.com>
To: Mikhail Maltsev <maltsevm@gmail.com>
Cc: Bernd Schmidt <bschmidt@redhat.com>,
gcc-patches mailing list <gcc-patches@gcc.gnu.org>,
Jeff Law <law@redhat.com>
Subject: Re: [PATCH 5/9] ENABLE_CHECKING refactoring: pool allocators
Date: Wed, 21 Oct 2015 11:02:00 -0000 [thread overview]
Message-ID: <CAFiYyc2LefXW0Cohvkft3mkbF--qD4Xc8ubeM-EVE9ooV2OBrw@mail.gmail.com> (raw)
In-Reply-To: <562434C4.30004@gmail.com>
On Mon, Oct 19, 2015 at 2:09 AM, Mikhail Maltsev <maltsevm@gmail.com> wrote:
> On 10/06/2015 03:45 PM, Richard Biener wrote:
>> On Tue, Oct 6, 2015 at 2:41 PM, Bernd Schmidt <bschmidt@redhat.com> wrote:
>>> On 10/06/2015 01:32 AM, Mikhail Maltsev wrote:
>>>>
>>>> gcc/ChangeLog:
>>>>
>>>> 2015-10-05 Mikhail Maltsev <maltsevm@gmail.com>
>>>>
>>>> * alloc-pool.h (base_pool_allocator::initialize, ::allocate,
>>>> ::remove): Adjust to use CHECKING_P.
>>>
>>>
>>> Why CHECKING_P for these and not flag_checking as elsewhere?
>>
>> Probably because they are in inline functions and thus possibly would
>> affect optimization. Not sure why they are inline functions in the
>> first place... I'd agree to using flag_checking here.
>>
>> Richard.
>>
>>>
>>> Bernd
>
> Adjusted. Note: I had to include 'options.h' into 'alloc-pool.h' in order to use
> flag_checking.
Ugh (stupid templates).
@@ -387,10 +389,10 @@ base_pool_allocator <TBlockAllocator>::allocate ()
block = m_virgin_free_list;
header = (allocation_pool_list*) allocation_object::get_data (block);
header->next = NULL;
-#ifdef ENABLE_CHECKING
+
/* Mark the element to be free. */
- ((allocation_object*) block)->id = 0;
-#endif
+ if (flag_checking)
+ ((allocation_object*) block)->id = 0;
just set id to zero unconditionally. That'll be faster than checking
flag_checking.
-#ifdef ENABLE_CHECKING
/* Set the ID for element. */
- allocation_object::get_instance (header)->id = m_id;
-#endif
+ if (flag_checking)
+ allocation_object::get_instance (header)->id = m_id;
likewise.
Given that, the pool initialization itself can do with unconditonally
computing the id as well, thus
-#ifdef ENABLE_CHECKING
- /* Increase the last used ID and use it for this pool.
- ID == 0 is used for free elements of pool so skip it. */
- last_id++;
- if (last_id == 0)
- last_id++;
+ if (flag_checking)
+ {
+ /* Increase the last used ID and use it for this pool.
+ ID == 0 is used for free elements of pool so skip it. */
+ last_id++;
+ if (last_id == 0)
+ last_id++;
- m_id = last_id;
-#endif
+ m_id = last_id;
+ }
make that unconditionally enabled as well.
-#ifdef ENABLE_CHECKING
- gcc_assert (object
+ gcc_checking_assert (object
/* Check if we free more than we allocated, which is Bad (TM). */
&& m_elts_free < m_elts_allocated
/* Check whether the PTR was allocated from POOL. */
&& m_id == allocation_object::get_instance (object)->id);
- memset (object, 0xaf, size);
-
- /* Mark the element to be free. */
- allocation_object::get_instance (object)->id = 0;
-#endif
+ if (flag_checking)
+ {
+ memset (object, 0xaf, size);
+ /* Mark the element to be free. */
+ allocation_object::get_instance (object)->id = 0;
+ }
guard the assert with flag_checking and do the id set unconditionally
(for consistency).
Ok with those changes.
Thanks,
Richard.
> --
> Regards,
> Mikhail Maltsev
>
> 2015-10-19 Mikhail Maltsev <maltsevm@gmail.com>
>
> * alloc-pool.h (base_pool_allocator ::initialize, ::allocate,
> ::remove): Adjust to use flag_checking.
next prev parent reply other threads:[~2015-10-21 10:57 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-05 23:27 [PATCH 1/9] ENABLE_CHECKING refactoring Mikhail Maltsev
2015-10-05 23:29 ` [PATCH 2/9] ENABLE_CHECKING refactoring: libcpp Mikhail Maltsev
2015-10-06 12:40 ` Bernd Schmidt
2015-10-12 20:57 ` Jeff Law
2015-10-19 1:18 ` Mikhail Maltsev
2015-10-21 22:29 ` Jeff Law
2015-10-21 21:19 ` Jeff Law
2015-10-05 23:30 ` [PATCH 3/9] ENABLE_CHECKING refactoring: Java and Ada Mikhail Maltsev
2015-10-22 19:25 ` Jeff Law
2015-10-05 23:31 ` [PATCH 4/9] ENABLE_CHECKING refactoring: Fortran Mikhail Maltsev
2015-10-23 22:38 ` Jeff Law
2015-10-05 23:32 ` [PATCH 5/9] ENABLE_CHECKING refactoring: pool allocators Mikhail Maltsev
2015-10-06 12:41 ` Bernd Schmidt
2015-10-06 12:45 ` Richard Biener
2015-10-19 0:47 ` Mikhail Maltsev
2015-10-21 11:02 ` Richard Biener [this message]
2015-10-26 2:07 ` Mikhail Maltsev
2015-10-26 9:48 ` Richard Biener
2015-10-26 10:57 ` Mikhail Maltsev
2015-10-05 23:34 ` [PATCH 6/9] ENABLE_CHECKING refactoring: generators Mikhail Maltsev
2015-10-06 12:57 ` Richard Biener
2015-10-19 0:09 ` Mikhail Maltsev
2015-10-21 10:57 ` Richard Biener
2015-10-28 16:32 ` Jeff Law
2015-10-29 16:31 ` Jeff Law
2015-10-05 23:39 ` [PATCH 7/9] ENABLE_CHECKING refactoring: middle-end, LTO FE Mikhail Maltsev
2015-10-06 12:46 ` Bernd Schmidt
2015-10-06 12:59 ` Richard Biener
2015-10-19 0:56 ` Mikhail Maltsev
2015-10-19 12:19 ` Bernd Schmidt
2015-10-26 17:04 ` Jeff Law
2015-10-26 17:15 ` Bernd Schmidt
2015-10-26 19:05 ` Jeff Law
2015-10-28 1:17 ` Jeff Law
2015-10-28 2:12 ` Trevor Saunders
2015-10-06 12:59 ` Richard Biener
2015-10-05 23:40 ` [PATCH 8/9] ENABLE_CHECKING refactoring: target-specific parts Mikhail Maltsev
2015-10-06 12:48 ` Bernd Schmidt
2015-10-29 19:43 ` Jeff Law
2015-10-29 21:23 ` Jeff Law
2015-10-30 4:13 ` Jeff Law
2015-10-30 4:20 ` Jeff Law
2015-10-06 12:53 ` [PATCH 1/9] ENABLE_CHECKING refactoring Richard Biener
2015-10-12 20:48 ` Jeff Law
2015-10-13 21:33 ` Jeff Law
2015-10-18 8:25 ` Mikhail Maltsev
2015-10-19 11:14 ` Bernd Schmidt
2015-10-19 13:54 ` Mikhail Maltsev
2015-10-21 15:59 ` Jeff Law
2015-10-21 16:06 ` Bernd Schmidt
2015-10-21 16:18 ` Richard Biener
2015-10-21 16:28 ` Jeff Law
2015-11-07 22:42 ` Gerald Pfeifer
2015-10-21 16:19 ` Jeff Law
2015-10-21 16:22 ` Bernd Schmidt
2015-10-21 16:44 ` Jakub Jelinek
2015-10-22 7:58 ` Richard Biener
2015-10-21 20:06 ` Jeff Law
2015-10-20 16:14 ` Jeff Law
2015-10-21 21:17 ` Jeff Law
2015-11-01 14:58 ` [PATCH 9/9] ENABLE_CHECKING refactoring: C family front ends Mikhail Maltsev
2015-11-02 23:34 ` Jeff Law
2015-11-04 14:41 ` Mikhail Maltsev
2015-11-01 20:19 ` [PATCH 10/9] ENABLE_CHECKING refactoring: remove remaining occurrences Mikhail Maltsev
2015-11-02 23:35 ` Jeff Law
2015-11-04 15:03 ` Mikhail Maltsev
2016-02-23 15:21 ` Richard Biener
2016-02-24 14:17 ` Martin Liška
2016-02-24 14:27 ` Michael Matz
2016-02-24 14:53 ` Martin Liška
2016-02-24 15:43 ` Pierre-Marie de Rodat
2016-02-25 9:24 ` Richard Biener
2016-02-25 10:14 ` Pierre-Marie de Rodat
2016-02-25 10:15 ` Martin Liška
2016-02-25 10:16 ` Pierre-Marie de Rodat
2016-02-25 9:24 ` Richard Biener
[not found] ` <C5BB0125-FB5F-46C6-B16D-74C3D0F07C10@gmail.com>
2015-11-08 15:37 ` Mikhail Maltsev
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=CAFiYyc2LefXW0Cohvkft3mkbF--qD4Xc8ubeM-EVE9ooV2OBrw@mail.gmail.com \
--to=richard.guenther@gmail.com \
--cc=bschmidt@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=law@redhat.com \
--cc=maltsevm@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).