Re: [HACKERS] xlc atomics

2016-04-26 Thread Noah Misch
On Mon, Apr 25, 2016 at 11:52:04AM -0700, Andres Freund wrote: > On 2016-04-23 21:54:07 -0400, Noah Misch wrote: > > The bug is that the pg_atomic_compare_exchange_*() specifications > > grant "full barrier semantics", but generic-xlc.h provided only the > > semantics of an acquire barrier. > > I

Re: [HACKERS] xlc atomics

2016-04-25 Thread Andres Freund
On 2016-04-23 21:54:07 -0400, Noah Misch wrote: > I missed a second synchronization bug in generic-xlc.h, but the pgbench test > suite caught it promptly after commit 008608b. Nice catch. > The bug is that the pg_atomic_compare_exchange_*() specifications > grant "full barrier semantics", but

Re: [HACKERS] xlc atomics

2016-04-23 Thread Noah Misch
On Mon, Feb 15, 2016 at 02:02:41PM -0500, Tom Lane wrote: > Noah Misch writes: > > That commit (0d32d2e) permitted things to compile and usually pass tests, > > but > > I missed the synchronization bug. Since 2015-10-01, the buildfarm has seen > > sixteen

Re: [HACKERS] xlc atomics

2016-02-15 Thread Craig Ringer
On 5 July 2015 at 06:54, Andres Freund wrote: > I wrote this entirely blindly, as evidenced here by the changes you > needed. At the time somebody had promised to soon put up an aix animal, > but that apparently didn't work out. > Similarly, I asked IBM for XL/C for a

Re: [HACKERS] xlc atomics

2016-02-15 Thread Tom Lane
Noah Misch writes: > That commit (0d32d2e) permitted things to compile and usually pass tests, but > I missed the synchronization bug. Since 2015-10-01, the buildfarm has seen > sixteen duplicate-catalog-OID failures. I'd been wondering about those ... > These suggested

Re: [HACKERS] xlc atomics

2016-02-15 Thread Noah Misch
On Mon, Feb 15, 2016 at 06:33:42PM +0100, Andres Freund wrote: > On 2016-02-15 12:11:29 -0500, Noah Misch wrote: > > These suggested OidGenLock wasn't doing its job. I've seen similar symptoms > > around WALInsertLocks with "IBM XL C/C++ for Linux, V13.1.2 (5725-C73, > > 5765-J08)" for ppc64le.

Re: [HACKERS] xlc atomics

2016-02-15 Thread Andres Freund
On 2016-02-15 12:11:29 -0500, Noah Misch wrote: > These suggested OidGenLock wasn't doing its job. I've seen similar symptoms > around WALInsertLocks with "IBM XL C/C++ for Linux, V13.1.2 (5725-C73, > 5765-J08)" for ppc64le. The problem is generic-xlc.h > pg_atomic_compare_exchange_u32_impl()

Re: [HACKERS] xlc atomics

2016-02-15 Thread Noah Misch
On Sat, Jul 04, 2015 at 08:07:49PM -0400, Noah Misch wrote: > On Sun, Jul 05, 2015 at 12:54:43AM +0200, Andres Freund wrote: > > On 2015-07-04 18:40:41 -0400, Noah Misch wrote: > > > (1) "IBM XL C/C++ for AIX, V12.1 (5765-J02, 5725-C72)". Getting it > > > working > > > required the attached

[HACKERS] xlc atomics

2015-07-04 Thread Noah Misch
On Wed, Jun 25, 2014 at 07:14:34PM +0200, Andres Freund wrote: * gcc, msvc work. acc, xlc, sunpro have blindly written support which should be relatively easy to fix up. I tried this on three xlc configurations. (1) IBM XL C/C++ for AIX, V12.1 (5765-J02, 5725-C72). Getting it working

Re: [HACKERS] xlc atomics

2015-07-04 Thread Andres Freund
On 2015-07-04 18:40:41 -0400, Noah Misch wrote: (1) IBM XL C/C++ for AIX, V12.1 (5765-J02, 5725-C72). Getting it working required the attached patch. None of my xlc configurations have an atomic.h header, and a web search turned up no evidence of one in connection with xlc platforms. Did

Re: [HACKERS] xlc atomics

2015-07-04 Thread Noah Misch
On Sun, Jul 05, 2015 at 12:54:43AM +0200, Andres Freund wrote: On 2015-07-04 18:40:41 -0400, Noah Misch wrote: (1) IBM XL C/C++ for AIX, V12.1 (5765-J02, 5725-C72). Getting it working required the attached patch. Will you apply? Having the ability to test change seems to put you in a much