Re: 16-bit page checksums for 9.2

On Fri, Feb 17, 2012 at 10:13 PM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> On Thu, Feb 16, 2012 at 1:53 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>> On Thu, Feb 16, 2012 at 6:16 AM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
>>> v8 attached
>>
>> It's hard to believe that this version has been tested terribly
>> thoroughly, because it doesn't compile.
>
> I'm just back home from a few days off grid.
>
> It's possible it doesn't compile against current HEAD, though it
> certainly does compile and work against my last git pull.
>
> I will look into your comments in detail tomorrow morning. Thank you
> for looking at the patch.

It does appear there is a chunk of code introduced somewhere between
my testing and sending the patch, so you're right. Looking at the
chunk below it looks like something hit some keys, adding a "]" and
CR, which are adjacent on my keyboard. Trying to develop with 3 kids
and a dog around isn't easy, I guess.

With this chunk removed we're back to the version I've tested - I
think - will carry on checking.

@@ -465,7 +466,9 @@ ReadBuffer_common(SMgrRelation smgr, char
relpersistence, ForkNumber forkNum,
{
/* Only need to adjust flags */
bufHdr->flags |= BM_VALID;
- }
+ [
+
+}
else
{
/* Set BM_VALID, terminate IO, and wake up any waiters */

For testing, I've been running server with shared_buffers = 16, to
force blocks in and out of cache frequently to stress test the
checksumming.

--
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services