Re: [hail patch 2/3] fix 32/64 wire interoperability

2010-08-05 Thread Jeff Garzik
On 08/04/2010 07:16 PM, Pete Zaitcev wrote: Testing found that tabled and chunkd running on CPUs with different word length cannot talk to each other. The bug was introduced by commit ea5d20bc22aeed077312c9c1824e84651af17a16. The fix is to add named padding that takes the place of the

Re: [hail patch 2/3] fix 32/64 wire interoperability

2010-08-05 Thread Pete Zaitcev
On Thu, 05 Aug 2010 17:25:30 -0400 Jeff Garzik j...@garzik.org wrote: I wonder if we shouldn't switch to attribute(packed) for safety, though. I dunno, as long as we test this... I irrationally hate the gcc attributes, although probably any viable competitor will have to support them. Talk

[tabled patch 2/3] Fix the flood of read obj(N)

2010-08-05 Thread Pete Zaitcev
Sometimes during operation tabled starts flooding a message read obj. I thought it was something about replication thread fetching incomplete objects, but actually it is caused by attempts to read from a closed object. It is purely dictated by load levels. This fix closes the obvious avenue, but