On Thu, Oct 15, 2015 at 10:36 PM, Michael Paquier
<michael.paquier@gmail.com> wrote:
> On Fri, Oct 16, 2015 at 9:07 AM, Thomas Munro
> <thomas.munro@enterprisedb.com> wrote:
>> On Wed, Oct 7, 2015 at 8:52 AM, Robert Haas <robertmhaas@gmail.com> wrote:
>>> On Thu, Oct 1, 2015 at 11:01 PM, Michael Paquier
>>> <michael.paquier@gmail.com> wrote:
>>>>> Right, see attached.
>>>>
>>>> It seems to me that we could as well simplify checkpoint.c and
>>>> logical.c. In those files volatile casts are used as well to protect
>>>> from reordering for spinlock operations. See for example 0002 on top
>>>> of 0001 that is Thomas' patch.
>>>
>>> These patches look good to me, so I have committed them.
>>
>> Thanks. Also, spin.h's comment contains an out of date warning about
>> this. Here's a suggested fix for that, and a couple more volatrivia
>> patches.
>
> I have looked at the rest of the code, and it seems that we can get
> rid of volatile in a couple of extra places like in the attached as
> those are used with spin locks. This applies on top of Thomas' set.
OK, committed his, and yours.
I back-patched his spin.h comment fix to 9.5 since that's a factual
error, but the rest of this seems like optimization so I committed it
only to master.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company